POLITECNICO DI MILANO Mattia... · luogo, di mettere in pratica le conoscenze acquisite migliorando...

POLITECNICO DI MILANO

School of Industrial and Information Engineering

Degree Programme of:

Energy Engineering Laurea Magistrale (Equivalent To Master Of Science)

Track: Energy Engineering for an Environmentally Sustainable World (EE-ESW)

LCA and optimization: an application to biorefineries

Supervisor: Prof. Mario GROSSO

Assistant Supervisor: Ing. Lucia RIGAMONTI

Candidate:

Mattia SISCA Matr. 842768

Academic Year 2015 - 2016

Mattia Sisca: LCA and optimization: an application to biore�neries | Master Thesisin Energy Engineering, Politecnico di Milano.c© Copyright April 2017.

Politecnico di Milano:www.polimi.it

School of Industrial and Information Engineering:www.ingindinf.polimi.it

http://www.polimi.it

http://www.ingindinf.polimi.it

Acknowledgements

First of all, I would to thank Alberto Bezama and Maik Budzinski from theBioenergy Systems department of the DBFZ/UFZ, for the opportunity that I wasgiven of carrying out the Master Thesis at the premises of the research center inLeipzig and for the continuous supervision and support throughout the whole periodof my staying.I would like to thank Professor Mario Grosso and Lucia Rigamonti for their precioushelp and support during the preparation of the Thesis and the time they have givento it.Special thanks go to all my colleagues for the mutual encouragement during theseyears.Last but not least, I would like to thank all my friends and especially my family,for the constant support during my studies that made possible the educational paththat I have undertaken and �nally concluded.

Piacenza, April 2017 M. S.

iii

"If we could change ourselves, the tendencies in the world would also change. As aman changes his own nature, so does the attitude of the world change towards him.

[. . . ] We need not wait to see what others do"Mahatma Gandhi

Contents

Introduction 10.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Goal of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

I The Biore�nery System. LCA. Optimization 7

1 The Biore�nery as an energy system 91.1 State of the art in biofuel production . . . . . . . . . . . . . . . . . . 91.2 Potential for bioproducts . . . . . . . . . . . . . . . . . . . . . . . . . 111.3 The biore�nery concept . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.3.1 De�nition and perspectives . . . . . . . . . . . . . . . . . . . . 121.3.2 Feedstocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1.3.2.1 Carbohydrates and lignocellulose . . . . . . . . . . . 161.3.2.2 Triglycerides . . . . . . . . . . . . . . . . . . . . . . 171.3.2.3 Mixed organic residues . . . . . . . . . . . . . . . . . 18

1.3.3 Technological processes in biore�nery . . . . . . . . . . . . . . 181.3.3.1 Thermochemical processes . . . . . . . . . . . . . . . 181.3.3.2 Biochemical processes . . . . . . . . . . . . . . . . . 201.3.3.3 Mechanical processes . . . . . . . . . . . . . . . . . . 201.3.3.4 Chemical processes . . . . . . . . . . . . . . . . . . . 21

1.3.4 Biore�nery processing strategies . . . . . . . . . . . . . . . . . 221.3.5 Guidelines for future biore�neries . . . . . . . . . . . . . . . . 23

1.4 From oil re�nery to bior�nery . . . . . . . . . . . . . . . . . . . . . . 241.4.1 Carbohydrates vs. hydrocarbons . . . . . . . . . . . . . . . . 241.4.2 Current chemical platforms in oil re�nery . . . . . . . . . . . . 261.4.3 Expected chemical platforms in biore�nery . . . . . . . . . . . 27

1.5 Biore�nery products . . . . . . . . . . . . . . . . . . . . . . . . . . . 281.5.1 Biomass vs. fossils as raw materials . . . . . . . . . . . . . . . 281.5.2 The role of green chemistry . . . . . . . . . . . . . . . . . . . 29

2 The Life Cycle Assessment methodology 312.1 LCA: an hystorical perspective . . . . . . . . . . . . . . . . . . . . . 312.2 LCA applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

vii

viii CONTENTS

2.2.1 Levels of sophistication in LCA for di�erent applications . . . 322.3 Methodological framework . . . . . . . . . . . . . . . . . . . . . . . . 34

2.3.1 Goal and scope de�nition . . . . . . . . . . . . . . . . . . . . 352.3.2 Inventory analysis . . . . . . . . . . . . . . . . . . . . . . . . . 39

3 Optimization and linear programming 43

II Case study 45

4 The system under study: three biore�nery concepts 474.1 Products Portfolio . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474.2 Biore�nery Concept 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.3 Biore�nery Concept 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 504.4 Biore�nery Concept 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5 LCA results 555.1 Goal and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.2 The Inventory Analysis model . . . . . . . . . . . . . . . . . . . . . . 55

5.2.1 Representation of processes and �ows . . . . . . . . . . . . . . 565.2.2 The solution of the inventory problem . . . . . . . . . . . . . 595.2.3 Solving multi-functionality and allocation . . . . . . . . . . . . 615.2.4 Impact assessment . . . . . . . . . . . . . . . . . . . . . . . . 635.2.5 Contribution analysis . . . . . . . . . . . . . . . . . . . . . . . 655.2.6 Avoided Burdens . . . . . . . . . . . . . . . . . . . . . . . . . 655.2.7 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.2.7.1 Demand and scaling vectors . . . . . . . . . . . . . . 675.2.7.2 Emissions and impacts . . . . . . . . . . . . . . . . . 685.2.7.3 Avoided emissions . . . . . . . . . . . . . . . . . . . 685.2.7.4 Contribution analysis . . . . . . . . . . . . . . . . . . 69

6 Results of the optimization model 756.1 Problem description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 756.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766.3 Mixed Integer Linear Programming (MILP) . . . . . . . . . . . . . . 78

7 Conclusions and outlook 85

A Results for the Bioethylene concept 89

B Results for the Lactic Acid concept 99

C A focus on the Organosolv process 109

List of Figures

1 Statistical data about world energy consumption and emissions re-lated to the transport sector . . . . . . . . . . . . . . . . . . . . . . . 2

1.1 Potentials for biofuels and bioproducts . . . . . . . . . . . . . . . . . 12

1.2 Biore�neries: the centre of a new economic system (Source: ePure) . 13

1.3 Examples of industrial clusters . . . . . . . . . . . . . . . . . . . . . . 15

1.4 Cellulose structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

1.5 Biore�nery Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

1.6 Thermochemical platform �owchart . . . . . . . . . . . . . . . . . . 21

1.7 Bioconversion platform �owchart . . . . . . . . . . . . . . . . . . . . 21

1.8 High-level representation of pathways via the sugar platform . . . . . 22

1.9 Oil re�nery versus biore�nery . . . . . . . . . . . . . . . . . . . . . . 25

1.10 Schematic overview �ow-chart: biomass-to-products . . . . . . . . . . 27

2.1 Life cycle assessment framework - phases of an LCA (ISO, 1997a). . . 35

4.1 Scheme for the biore�nefy concept 1 . . . . . . . . . . . . . . . . . . . 48

4.2 Flowchart used for the LCA of the biore�nery concept 1. T=transport;M=mixing; upside-down trapeziodal blocks=displaced material (avoidedburden methodology) . . . . . . . . . . . . . . . . . . . . . . . . . . . 50





5.1 Comparison between a biore�nery and BAU production (Business AsUsual) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

5.2 Analysis of the contribution of each process to the environmentalinterventions g (per hour) . . . . . . . . . . . . . . . . . . . . . . . . 70

5.3 Relative contribution of each process to the total environmental in-terventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

ix

x LIST OF FIGURES

5.4 Analysis of the contribution of each process to the impact categoriesg (per hour) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5.5 Relative contribution of each process to the total impact categories . 73

6.1 Most promising german regions in terms of beech wood potential . . 766.2 Process Matrix for the optimization model . . . . . . . . . . . . . . . 796.3 Process Matrix for the optimization model . . . . . . . . . . . . . . . 80

7.1 Impacts deriving from the treatment of 1 kg of beech wood . . . . . . 867.2 Impacts deriving from the biore�nery systems (blue) compared to the

avoided impacts (red)) . . . . . . . . . . . . . . . . . . . . . . . . . . 877.3 Normalized impacts considering the avoided emissions deriving from

the displacement of �nal products equivalent . . . . . . . . . . . . . . 87

A.1 Comparison between a biore�nery and BAU production (Business AsUsual) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

A.2 Analysis of the contribution of each process to the environmentalinterventions g (per hour) . . . . . . . . . . . . . . . . . . . . . . . . 94

A.3 Relative contribution of each process to the total environmental in-terventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

A.4 Analysis of the contribution of each process to the impact categoriesg (per hour) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

A.5 Relative contribution of each process to the total impact categories . 97

B.1 Comparison between a biore�nery and BAU production (Business AsUsual) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

B.2 Analysis of the contribution of each process to the environmentalinterventions g (per hour) . . . . . . . . . . . . . . . . . . . . . . . . 104

B.3 Relative contribution of each process to the total environmental in-terventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

B.4 Analysis of the contribution of each process to the impact categoriesg (per hour) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

B.5 Relative contribution of each process to the total impact categories . 107

C.1 Organosolv-based lignocellulosic biore�nery . . . . . . . . . . . . . . . 110

List of Tables

1.1 Overview of the alternative fuels for transport . . . . . . . . . . . . . 111.2 Examples of lignocellulosic biore�neries . . . . . . . . . . . . . . . . . 231.3 Examples of checmical and elemental composition . . . . . . . . . . . 26

2.1 Level of detail in some applications of LCA . . . . . . . . . . . . . . . 33

5.1 Process matrix for the biore�nery concept 1. Matrix A: blue part,Matrix B: violet part . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

5.2 Partitioning method . . . . . . . . . . . . . . . . . . . . . . . . . . . 635.3 Characterization matrix, Q . . . . . . . . . . . . . . . . . . . . . . . 645.4 Avoided burdens for the biore�nery concept 1 . . . . . . . . . . . . . 665.5 Final demand vector, f and scaling vector s . . . . . . . . . . . . . . 675.6 Environmental interventions vector, g . . . . . . . . . . . . . . . . . . 685.7 Impact vector h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685.8 Avoided impacts h' . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685.9 Normalized impacts h* . . . . . . . . . . . . . . . . . . . . . . . . . . 69

6.1 Characteristics of the regions under study . . . . . . . . . . . . . . . 786.2 Results from the optimization model: GWP savings . . . . . . . . . . 81

7.1 Summary of the results of the LCA calculations . . . . . . . . . . . . 85

A.1 Process matrix for the biore�nery concept 2. Matrix A: blue part,Matrix B: violet part . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

A.2 Partitioning method . . . . . . . . . . . . . . . . . . . . . . . . . . . 91A.3 Avoided burdens for the biore�nery concept 2 . . . . . . . . . . . . . 91A.4 Final demand vector, f and scaling vector s . . . . . . . . . . . . . . 92A.5 Environmental interventions vector, g . . . . . . . . . . . . . . . . . . 92A.6 Impact vector h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93A.7 Avoided impacts h' . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93A.8 Normalized impacts h* . . . . . . . . . . . . . . . . . . . . . . . . . . 93

B.1 Process matrix for the biore�nery concept 3. Matrix A: blue part,Matrix B: violet part . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

B.2 Partitioning method . . . . . . . . . . . . . . . . . . . . . . . . . . . 101B.3 Avoided burdens for the biore�nery concept 3 . . . . . . . . . . . . . 101

xi

xii LIST OF TABLES

B.4 Final demand vector, f and scaling vector s . . . . . . . . . . . . . . 102B.5 Environmental interventions vector, g . . . . . . . . . . . . . . . . . . 102B.6 Impact vector h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103B.7 Avoided impacts h' . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103B.8 Normalized impacts h* . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Sommario

A livello regionale, nazionale e mondiale ci sono tre fattori principali per l'utilizzodi biomasse in biora�nerie per la produzione di bioenergia, biocarburanti e prodottibiochimici, e cioè: il cambiamento climatico, la sicurezza energetica e lo svilupporurale.

La lignina è il secondo polimero naturale più abbondante sulla Terra e la suastruttura aromatica la rende una piattaforma di partenza adatta per produrre sostanzechimiche provenienti da fonti rinnovabili.

Lo sviluppo di biora�nerie lignocellulosiche è stato identi�cato come un passopromettente per ridurre l'impatto delle attività industriali sull'ambiente; in generale,le biora�nerie sono caratterizzate da impianti tecnologici integrati e multifunzionaliche combinano una varietà di processi di conversione e di separazione per produrrecombustibili, energia elettrica, calore e sostanze chimiche a partire da biomassa.Gli impatti ambientali di questi nuovi concept tecnologici sono stati determinatiprincipalmente utilizzando la metodologia del Life Cycle Assessment (LCA), �nal-izzata sia ad identi�care i processi più impattanti dal punto di vista ambientale siaconfrontandoli con le tecnologie (fossili) di riferimento.

Tuttavia, uno dei limiti del LCA è che non include modi sistematici per individ-uare le migliori opzioni di fronte a problemi decisionali più complessi come la limitatadisponibilità di materia prima, la fattibilità economica, la capacità dell'impianto ela domanda di mercato. Per far fronte a questi problemi, l'approccio che è statoutilizzato è stato quello di incorporare la metodologia LCA in un contesto di ot-timizzazione.

Gli obiettivi della Tesi sono due: in primo luogo, per valutare gli aspetti piùrilevanti in termini di uno studio LCA per i tre concepts di biora�nerie; in secondoluogo, di mettere in pratica le conoscenze acquisite migliorando un caso di studio diLCA esistente di biora�nerie per mezzo di un modello di ottimizzazione, che è ingrado di trovare la migliore combinazione di tecnologie per soddisfare una speci�cadomanda �nale, soggetto a certi vincoli, riducendo al minimo gli impatti ambientaliin termini di potenziale di riscaldamento globale (GWP).

Parole chiave: Biora�neria, Biocombustibili, Bioplastiche, LCA, IOA, Ottimiz-zazione

xiii

Abstract

At regional, national and global levels there are three main drivers for usingbiomass in biore�neries for the production of bioenergy, biofuels and biochemicals,namely: climate change, energy security and rural development.

Lignin is the second most abundant natural polymer on Earth. The aromaticstructure of lignin makes it a suitable platform for biobased chemicals.

The development of lignocellulosic biore�neries has been identi�ed as a promis-ing step to reduce the impact of industrial activities on the environment; generallyspeaking, biore�neries are characterized by integrated and multifunctional techno-logical plants that combine a variety of conversion and separation processes to pro-duce fuels, power, heat and value-added chemicals starting from biomass resources.The environmental impacts of these new technological concepts has been mainlydetermined using the Life Cycle Assessment (LCA) methodology, aimed either toidentifying environmental key processes of the new technology or comparing it to(fossil) reference technologies.

However, one of the limitations of LCA is that it does not include systematic waysto identify best options using biomass resources when faced with more complex de-cision problems considering issues such as feedstock availability, cost-e�ectiveness,plant capacity and consumer demand.To circumvent these shortcomings, the ap-proach that was applied is to incorporate LCA into optimization frameworks.

The goals of the master thesis are two: �rstly, to evaluate the most relevantaspects in terms of a LCA study for three biore�nery concepts; secondly, to putthe acquired knowledge into practice by enhancing an existing LCA case studyof biore�neries by means of an optimization model which is able to �nd the bestcombination of technologies to match a speci�ed �nal demand subjected to certainconstraints, minimizing the environmental impacts in terms of Global WarmingPotential (GWP).

Keywords: Biore�nery, Biofuels, Bioplastics, LCA, IOA, Optimization

xiv

Introduction

0.1 Background

The future energy economy will likely be based on a wide range of alternativeenergy platforms including wind, water, sun, nuclear as well as biomass. Similarly,the production of chemicals will increasingly depend on biomass, particularly plantbiomass.

Nowadays, our society strongly depends on fossil fuels due to the intensive useand consumption of petroleum derivatives which, coupled with diminishing oil re-sources, causes environmental concerns and is a crucial part of most country's po-litical agenda. It has been scienti�cally proven that emissions of greenhouse gases(GHG), such as carbon dioxide (CO2) [Fig.1a], methane (CH4) and nitrous oxide(N2O), arising from fossil fuel combustion and land-use change as a result of humanactivities, are negativey in�uencing the climate of the Earth. The IntergovernmentalPanel on Climate Change (IPCC) Fifth Assessment Report showed that the world'sgrowing population and per capita energy demand are leading to the rapid increasein greenhouse gas (GHG) emissions. In particular, over the past 10 years, transporthas shown the highest rates of growth in GHG emissions in any sector.

The world's primary source of energy for the transport sector (and production ofchemicals as well) is oil. World use of petroleum and other liquid fuels is expected togrow from 90 million barrels per day (b/d) in 2012 to 100 million b/d in 2020 and to121 million b/d in 2040 [Fig.1b]. Most of the growth in liquid fuels consumption is inthe transportation and industrial sectors. In the transportation sector, in particular,liquid fuels continue to provide most of the energy consumed. Although advances innonliquids-based transportation technologies are anticipated (i.e. elctricity), theyare not enough to o�set the rising demand for transportation services worldwide.Liquid fuels consumed for transportation increases by an average of 1.1%/year from2012 to 2040, and the transportation sector accounts for 62% of the total increasein delivered liquid fuels use. While the transport sector continues to expand incountries like the US and Europe, growth in the emerging economies like India andChina is predicted to be even greater, increasing by at least 3% per year. For whatconcerns chemicals, their dependence on fossil resources is even more pronounced:the majority of chemical products are produced in oil re�neries and almost 4% ofoil is used worldwide for chemical and plastic production.

In order to reduce the oil dependence and mitigate climate change in both trans-port and chemical sectors at the same time, alternative production patterns are

1

2 Introduction

(a) World energy-related carbon dioxideemissions by fuel type (billion metric tons)

(b) World transportation sector delivered energy con-sumption by energy source (quadrillion Btu)

(c) World Energy consumption by region (quadrillion Btu)

Figure 1: Statistical data about world energy consumption and emissions related to thetransport sector

Ref.[60]

0.1. Background 3

needed. It is widely recognized that a single solution to these problems is not ex-isting at the moment but rather combined actions are needed, including changesin behavior,vehicle technologies, improvement of public transport and adoption ofinnovative fuels and technologies.

Only recently, society began to recognize the opportunities embodied by a fu-ture sustainable economy based on renewable sources and a great number of R&Dactivities have been �nanced for its implementation. It is increasingly acknowledgedon global scale that plant-based raw materials (i.e. biomass) have the potential toreplace a large fraction of fossil resources as feedstocks for industrial productions,involved in both the energy and non-energy (i.e. chemicals and materials) sectors.

The cost of using starch as a feedstock is high because grain and sugar cropsare expensive. Consequently, there has been an increase in lignocellulosic biomassprocessing research, focusing particularly on agricultural and forestry residues, sincethese are low in cost, abundant, readily available and renewable.

At regional, national and global levels there are three main drivers for usingbiomass in biore�neries for the production of bioenergy, biofuels and biochemicals,namely: climate change, energy security and rural development. The political meansto support renewable sources of energy and chemicals arise from each individualdriver or combinations. It is worth to stress out that policies designed to targetone speci�c driver can be detrimental to another. For instance, policies aimed atensuring energy security may result in increased GHG emissions if local coal reservesare preferred to be exploited instead of imported oil or gas. Furthermore, electricityand heat can be generated by a variety of renewable alternatives (wind, sun, water,biomass, . . . ), while biomass will most probably be the only viable alternative tofossil resources for production of transportation fuels and chemicals, because it isthe only carbon-rich material source available on the planet, besides fossils. Asa consequence, the sustainability aspect in biomass production is a crucial issue,especially concerning a possible land competition with food and feed industries.

The use of biomass as a resource for energy and fuel production will be limitedby maximum production rates and the supply of biomass rather than the demandfor energy and fuel[Fig. 1c]. The relatively low energy content, seasonality anddiscrete geographic availability of biomass feedstocks have been noted as barriers tothe large volume demands for energy and fuel.

With the abundance of biomass waste available, the development of new tech-nologies that will make use of biomass for materials production beyond biofuelsrepresents an important opportunity to fully utilize our resources. Development ofe�cient techniques to fractionate lignocellulosic biomass into its core componentswill facilitate research on the production of speci�c biomass-derived sugars, buildingblock chemicals and ultimately value-added commodity chemicals while preservingthe concept of the biore�nery approach by promoting e�ective utilization of all feed-stock fractions.

4 Introduction

0.2 Goal of the Thesis

The present thesis has been elaborated in the framework of a broader researchproject, the "BioEconomy Cluster", at the premises of the Bioenergy Systems de-partment shared between DBFZ1 and UFZ 2.

The vision of the BioEconomy Cluster is to establish the world's �rst bioeconomyon a regional scale. With this goal in mind, the cluster's strategy is to develop theregion's economy in the context of a bioeconomy and to create new impulses forgrowth. Its aim is to balance the material and energy use of biomass. This can beachieved using the scarce resource of biomass in the most intelligent and mutuallybene�cial way. The focus of the leading edge cluster is on biomass that is not neededfor the food and feed industries. In practice this means cleverly cascading utilisationand re-connecting strong industries in the region like forestry, timber processing, thechemical and plastics industries, and bioenergy.

The need to investigate the topic rises from the concern on the recent escalation ofthe climate change issues due to the use of fossil-based fuels and materials worldwide.

This work tries, at �st, to assess the environmental impacts deriving form threedi�erent biore�nery concepts through the adoption of the Life Cycle Assessment(LCA) mehodology. However, one of the limitations of LCA is that it does notinclude systematic ways to identify best options using biomass resources when facedwith more complex decision problems considering issues such as feedstock availabil-ity, cost-e�ectiveness, plant capacity and consumer demand; the approach that wasapplied is to incorporate LCA into optimization frameworks and it consists in themain contribute of this work to the scienti�c research.

To conclude, in this thesis an optimization model was developed, which is able to�nd the best combination of technologies to match a speci�ed �nal demand subjectedto certain constraints, minimizing the environmental impacts in terms of GlobalWarming Potential (GWP).

1Deutsches Biomasseforschungszentrum2Helmholtz-Zentrum für Umweltforschung

0.3. Outline 5

0.3 Outline

The outline of the thesis is as follows:

In the �rst part some preliminary information regarding the Biore�nery System,the LCA methodology and the Optimization techniques are presented.

In the �rst chapter the Biore�nery concept is described, focusing on he generallayout of the system and on the possible feedstocks available and the relativeplant con�gurations.

In the second chapter the LCA framework is shortly illustrated, highlighting thesingle steps of the methodology, in particular the goal and scope de�nition andthe inventory analysis.

In the third chapter the Linear Programming opimization technique is brie�yreported.

In the second part the case study is analyzed and the �ndings are presented anddiscussed.

In the fourth chapter the three biore�nery concepts are described and the dataare listed.

In the �fth chapter the methodology used to perform the LCA is presented andthe results are reported and discussed.

In the sixth chapter the optimization model is illustrated and the results arepresented, according to the assumptions previously made.

Part I

The Biore�nery System. LCA.

Optimization

7

Chapter 1

The Biore�nery as an energy system

1.1 State of the art in biofuel production

At present, transportation fuels based on biomass (i.e. biofuels) are categorizedunder 1st and 2nd generation biofuels. First generation biofuels usually refer to bio-fuels produced from raw materials in competition with food and feed industries.Because of this competition, these biofuels give rise to ethical, political and environ-mental concerns. In order to overcome these issues, production of second generationbiofuels (i.e. from raw materials based on waste, residues or non-food crop biomass)gained an increasing worldwide interest in the last few years as a possible "greener"alternative to fossil fuels and conventional biofuels. As a development of 2nd gen-eration biofuel production, the use of biomass in biore�nery complexes is expectedto ensure additional environmental bene�ts and implement national energy security,thanks to the coproduction of both bioenergy and high value chemicals.

First generation biofuels are the ones produced from sugar, starch, vegetableoil or animal fats using conventional technologies. The basic feedstocks are oftenrepresented by seeds and grains such as wheat, corn and rapeseed. The most com-mon 1st generation biofuels are bioethanol, biodiesel and biogas, but also straightvegetable oils, biomethanol and bioethers may be included in this category.

Bioethanol is recovered from biomass feedstocks such as sugarcane, sugar beetand starch crops (mainly corn and wheat). In 2015, the total world productionreached 97,2 billion liters. USA is currently the largest producer of bioethanolwith a production of 56,05 billion liters per year, with corn as primary feedstock.Sugarcane is used as primary feedstock in Brazil, currently the world's second largestproducer (26,85 billion liters per year). The European Union produces 5,25 billionliters of bioethanol, mainly from sugar beet and starch crops.

Biodiesel is produced from oil-based crops such as rapeseed, sun�ower, soybeanbut also from palm oil and waste edible oils. World biodiesel production reached23 billion litres in 2015. The USA led biodiesel production in 2015, producing 4,8billion litres. Brazil is the second largest producer with 4,1 billion litres; Germanyis at the third place, producing 2,8 billion liters mainly from rapeseed and sun�owerbut other countries (France, Italy, Austria) are increasing their production.

9

10 Chapter 1. The Biore�nery as an energy system

Biogas is produced through anaerobic digestion of mixtures of feedstocks andby-products such as corn derived starch, manure, organic waste and grasses. Whenis mainly derived from waste and residues it can be categorized as 2nd generationbiofuel, because its feedstock is not in competition with the food and feed industry.The production of biogas is widespread in most countries, and in the last few yearsit has been extensively implemented in countries where economic subsidies for elec-tricity generation from biogas are in force (especially European countries). In somecountries (such as Germany and Sweden), biogas is also used as road transportationbiofuel, after upgrading it to biomethane. For instance, Sweden leads the world inautomotive biogas production, with a total �eet of approximately 47000 vehicleswith 50% of its fuel supplied by biomethane.

The main advantages of 1st generation biofuels are identi�ed in the high sugaror oil content of the raw materials and their easy conversion into biofuel.

One important aspect is that 1st generation biofuels are in competition with foodand feed industries for the use of biomass and agricultural land, giving rise to ethicalimplications. In fact, as prices for fossil fuels increase:

- a larger proportion of cereals or agricultural land will be dedicated to biofuelproduction (instead of using it to produce food),

- as a result, food prices will increase accordingly.

In conclusion, 1st generation biofuels currently produced from sugars, starchesand vegetable oils may cause several concerns: these production chains compete withfood for their feedstock and fertile land; their potential availability is limited by soilfertility and speci�c (i.e. per hectare) yields; the net savings of CO2 emissions andfossil energy consumption are limited by the high energy input required for cropcultivation and conversion. These limitations are expected to be partially overcomeby developing the so-called 2nd generation biofuels.

Second generation biofuels are derived from a variety of non-food crops.These include lignocellulosic materials, such as residues from agriculture, forestryand industry and dedicated crops.

As opposed to 1st generation biofuels, where the utilized fraction (grains andseeds) represents only a small portion of the harvested biomass, 2nd generationbiofuels can rely on the whole plant for bioenergy production. For example, rapeseedgrain yield is about 3,4 t/ha but the oil content of the grain is only 40%, thus the�e�ective� yield is reduced to 1,35 t/ha.

2nd generation biofuels (e.g. Fisher Tropsch(FT)-diesel from biomass and bioethanolfrom lignocellulosic feedstock) ensure advantages over 1st generation biofuels interms of land-use e�ciency and environmental performance, according to most ofthe LCA studies published in the literature. Thanks to technology development, en-vironmental performances of 2nd generation biofuels could bene�t of the use of highquantities of lignocellulosic residues and waste which are already available: theycan constitute the main raw material sources, which can be also supplemented withnon-food crops such as perennial grasses, and short-rotation forestry. On the oneside the characteristics of the raw materials are optimum: widespread, relatively

1.2. Potential for bioproducts 11

cheap and easily available; on the other side, their use could allow the co-productionof valuable biofuels, chemical compounds as well as electricity and heat, leading tobetter energy, environmental and economic performances through the developmentof integrated biore�nery concepts.

1.2 Potential for bioproducts

As shown in Tab.1.1 it is clear that in the near future biofuels will represent one ofthe most suitable alternatives to fossil-derived fuels such as gasoline or diesel. Eventhough we are witnessing a relatively high increase in the number of elecrtic veichles,it is interesting to notice that biofuels are the only ones to have the potential forsubstituting aviation fuels, a mean of transport that is likely to increse or at leastbe stable in the future.

Table 1.1: Overview of the alternative fuels for transport

Road Air Rail Water

Bus/Car High Duty Veihcle

Range Urban Medium Long Short Medium Long Inland Short Sea Maritime

Natural Gas G# G# # G# G# G#

Electricity G# # # # # # # #

Biofuels

Hydrogen # # # #

: feasible option #: unfeaseable option G#: option feasible under certain conditions (e.g.for Natural Gas the requirement of Lique�ed Natural Gas (LNG))Source: M. Steen (JRC: 2014)

Fig.1.1a reports trends regarding the biomass products market demand in theEU. It is possible to notice that the biggest potential for bioproducts is representedby aviation fuels, biogas, bioethanol and bioplastics.

This, combined with the increasing market shares of most of the biofuels (seeFig.1.1b) is the reason for such a high momentum of bio-based processes, in general.


(a) Estimated Biomass products market demand in EU up to 2030 (BIO-TIC project).Source: BEUR (2011)

(b) Shares of alternative fuels compared to the total auto-motive fuel consumption in the world [19]

Figure 1.1: Potentials for biofuels and bioproducts

This, combined with the increasing market shares of most of the biofuels (seeFig.1.1b) is the reason for such a high momentum of bio-based processes, in general.

1.3 The biore�nery concept

1.3.1 De�nition and perspectives

The de�nition of the term "biore�nery" has been subject to debate, but the over-all goal of the biore�nery production approach is the generation of a variety of goodsstarting from di�erent biomass feedstocks, within a particular biore�nery, through

1.3. The biore�nery concept 13

a combination of technologies. Ideally, a biore�nery should integrate biomass con-version processes to produce a range of fuels, power, materials, and chemicals frombiomass. The term biore�nery is derived both from the raw material feedstock whichis renewable biomass and also from the bioconversion processes often applied in thetreatment and processing of the raw materials.

According to the de�nition stated by the IEA Bioenergy Task 42 �Biore�neries�:�Biore�ning is the sustainable processing of biomass into a spectrum of marketableproducts and energy�(Fig.1.2).

Figure 1.2: Biore�neries: the centre of a new economic system (Source: ePure)

The biore�nery concept regards a wide range of technologies able to sunderbiomass resources (wood, grasses, corn. . . ) into their building blocks (carbohy-drates, proteins, triglycerides. . . ) which can then be converted to value-addedproducts, biofuels and chemicals. A biore�nery is a facility (or network of facilities)that integrates biomass conversion processes and equipment to produce transporta-tion biofuels, power, and chemicals from biomass. This concept is analogous to apetroleum re�nery, where multiple fuels and products are produced starting frompetroleum.

A farsighted approach would be to perform a gradual conversion of large sectorsof the global economy/industry into a sustainable biobased society having bioenergy,biofuels and bio-based products as main pillars and biore�neries as the basis. Sucha replacement of oil with biomass will require some breakthrough changes in thepresent production of goods and services: biological and chemical sciences will playa leading role in the generation of future industries and new synergies of biological,physical, chemical and technical sciences must be developed.


The e�cient production of transportation biofuels is seen as one of the main pro-moting factors for the future development of biore�neries: the transportation sectoris growing steadily and the demand for renewable (bio-)fuels grows accordingly. Forthis reason, the main challenge for biore�nery development seems to be the e�cientand cost-e�ective production of transportation biofuels, considering that from theco-produced biomaterials and biochemicals additional economic and environmentalbene�ts can be gained.

Today, the main bio-based products are obtained from the conversion of biomassto basic products like starch, oil, and cellulose. In addition, chemicals like lacticacid and amino acids are produced and used in the food industry. Other biobasedproducts which are already commercially available include: adhesives, cleaning com-pounds, detergents, dielectric �uids, dyes, hydraulic �uids, inks, lubricants, packag-ing materials, paints and coatings, paper and box board, plastic �llers, polymers,solvents, and sorbents.

The majority of the existing biofuels and biochemicals are currently obtainedthrough single production chains and not within a biore�nery concept, and usuallyrequire materials in competition with the food and feed industry. Their exploitationis thereby limited. Conversely, lignocellulosic crops reduce the competition for fertileland, since they can be grown also on land which is not suitable for agricultural crops.Moreover, biore�neries based on lignocellulosic feedstocks can rely on larger biomassper hectare yields, since the entire crop is available as feedstock.

Concerning the conversion plant, consumption of non-renewable energy resourcesduring biore�nery processing should be minimized, along with its related environ-mental impacts, while the complete and e�cient biomass use should be maximized.This ecological perspective requires:

• analyses of three important agricultural and forestry cycles, namely carbon(respiration, photosynthesis, and organic matter decomposition), water (pre-cipitation, evaporation, in�ltration, and run-o�) and nitrogen (N �xation,mineralization, denitri�cation) and their interdependencies,

• system performance evaluations at plant scale,

• environmental impact evaluations carried out by means of Life Cycle Assess-ment.

Biore�nery industries are expected to develop as dispersed industrial complexesable to revive rural areas. Unlike oil re�nery, which entails the construction of verylarge plants, biore�neries will most probably comprise a whole range of di�erent-sized installations. In this context, several bio-industries can combine their material�ows in order to reach a complete exploitation of all biomass components: theresidue from one bio-industry (e.g. lignin from a lignocellulosic ethanol productionplant) becomes an input for other industries, giving rise to a real integrated bio-industrial systems[Fig.1.3]. Furthermore, biomass resources are locally available inmany countries and their use may contribute to reduce national dependence on fossilfuels imports.


(a) "InfraLeuna" industrial cluster (Germany)

(b) "Le Shoettes" industrial cluster (France)

Figure 1.3: Examples of industrial clusters


1.3.2 Feedstocks

The term �feedstock� refers to the raw materials which are used in the biore�ner-ies. The biomass is originally synthesized via the photosynthetic process convertingatmospheric carbon dioxide and water into sugars; plants use these sugars to syn-thesize the complex materials that are generically named biomass. An importantissue in biore�nery system is represented by the provision of a renewable, consistentand regular supply of feedstock. Initial processing may be required to increase itsenergy density and to reduce transport, handling and storage costs.

Renewable, carbon-based raw materials for biore�nery are generally providedfrom four di�erent sectors:

1. agriculture (dedicated crops and residues),

2. forestry,

3. industries (process residues and leftovers) and households (municipal solidwaste and wastewaters),

4. aquaculture (algae and seaweeds).

A further distinction can be done between those feedstocks which come fromdedicated crops and residues from agricultural, forestry and industrial activities,that are available without upstream concerns. The main biomass feedstocks canbe grouped in 3 wide categories: carbohydrates and lignin, triglycerides and mixedorganic residues.

1.3.2.1 Carbohydrates and lignocellulose

Carbohydrates (from starch, cellulose and hemicellulose) are molecules of car-bon, hydrogen, and oxygen and are by far the most common biomass componentfound in plant feedstocks. Six-carbon, single-molecule (�monosaccharide�) sugars(C6H12O6) include glucose, galactose and mannose, while the most common �ve-carbon sugars (C5H10O5) are xylose and arabinose. The two most important sugarcrops are sugar cane and sugar beet which, together with corn (a starch crop), supplyalmost all the ethanol that is being produced today.

Starch (C6H10O5)n is a very large polymer molecule composed of many (hun-dreds or thousands) glucose molecules (polysaccharides), which have to be brokendown into one or two molecule pieces in order to be fermented. The most widespreadstarch crops are wheat and corn. Once sugars have been depolymerized (for starchcrops) or extracted (for sugar crops) they can be easily fermented to ethanol or usedas a substrate for chemical reactions leading to a wide range of chemical products.

Lignocellulosic biomass has three major components with di�erent shares inpercentage of the total dry matter: cellulose (30�50%), hemicellulose (20�40%)and lignin (15-�25%) [Fig.1.4].

Cellulose (C6H10O6)n has a strong molecular structure made by long chains ofglucose molecules (C6 sugar). The di�erence with starch is given by the con�guration


of the bonds formed across the oxygen molecule that joins two hexose units. Starchcan be readily hydrolyzed by enzymes or acid attack to the single sugar monomers,while cellulose is much more di�cult to hydrolyze and set free individual glucosemonomer.

Figure 1.4: Cellulose structure

Hemicellulose (C5H8O5)n is a relatively amorphous component that is easierto break down with chemicals and/or heat than cellulose; it contains a mix of C6

and C5 sugars.Lignin (C9H10O2(OCH3))n, is essentially the glue that provides the overall rigid-

ity to the structure of plants and trees and is made of phenolic polymers. Whilecellulose and hemicellulose are polysaccharides that can be hydrolyzed to sugars andthen fermented to ethanol, is not possible to use lignin in a fermentation processes,but it may be useful for other purposes (chemical extraction or energy generation).Lignin is the largest non-carbohydrate fraction of lignocellulose.

Lignocellulosic biomass can be provided both as a crop or as a residue. Largeamounts of cellulosic biomass can be produced via dedicated crops like perennialherbaceous plant species, or short rotation woody crops. Other sources of lignocel-lulosic biomass are waste and residues, like straw from agriculture, wood waste fromthe pulp and paper industry and forestry residues. The use of waste biomass o�ersa way of creating value for society, displacing fossil fuels with material that wouldtypically decompose, with no additional land use for its production.

1.3.2.2 Triglycerides

Oils and fats are triglycerides which typically consist of glycerin and saturatedand unsaturated fatty acids (their chain length ranges between C8 and C20, but16, 18 and 20 carbons are the most common). The sources of oils and fats are


a variety of vegetable and animal raw materials. Soybean, palm, rapeseed andsun�ower oil are the most important in terms of worldwide production. Vegetableoils are nowadays used for production of biodiesel by reacting with an alcohol, usuallymethanol. However, they can also be used as a substrate for chemical reactionsthanks to two chemically reactive sites: the double bond in the unsaturated fatty acidchain and the acid group of the fatty acid chain. Like sugar and starch crops, oilseedcrops are characterized by low yield and high use of inputs (i.e. energy). In thefuture, non-edible crops like Jatropha curcas and Pongamia pinnata, which requirelower inputs and are suited to marginal lands, may become the most widespread oilcrops for biore�nery purposes, especially in dry and semiarid regions. Other sourcesof vegetable oil for biofuel conversion can be found in waste streams of food industry,where used fried oil (UFO) is mainly generated from commercial services and foodprocessing plants such as restaurants, fast food chains and households.

1.3.2.3 Mixed organic residues

Other types of biomass sources that do not fall within the previous categoriesare: organic fraction of the Municipal Solid Waste (MSW), manure, wild fruits andcrops, proteins and residues from fresh fruit and vegetable industries. The phys-ical and chemical caracteristics of this wide spectrum of biomass resources varylargely. Certain streams such as sewage sludge, manure from dairy and swine farmsand residues from food processing are very wet, with moisture contents over 70%.Therefore, these feedstocks are more suited for an anaerobic digestion process togenerate biogas. Other streams, such as organic MSW, may be more or less con-taminated with heavy metals or other elements, but represents a high potential forenergy recovery. Clearly, the di�erent properties and characteristic of the biomasswaste require the application of di�erent conversion technologies.

1.3.3 Technological processes in biore�nery

The aim of technological process in biore�nery is to depolymerize and deoxy-genate the biomass components. In order to convert biomass feedstock into valuableproducts within a biore�nery system, several technological processes must be jointlyapplied. They can be divided in four main groups: thermochemical, biochemical,mechanical/physical and chemical processes (see Fig.1.5).

A biore�nery approach involves multi-step processes in which the �rst step,following feedstock selection, typically involves treating the precursor-containingbiomass to make it more amenable for further processing. This step is convention-ally referred to as pretreatment. Following pretreatment, the biomass componentsare subject to a combination of biological and/or chemical treatments.

1.3.3.1 Thermochemical processes

There are two main thermochemical processes for converting biomass into energyand chemical products.


(a)

(b)

Figure

1.5:Biore�neryProcesses


The �rst is gasi�cation, which consists in keeping biomass at high temperature(> 700◦C) with low oxygen levels to produce syngas: a mixture of H2, CO, CO2 andCH4. Syngas can be used directly as a stationary biofuel or can be a chemical inter-mediate (platform chemical) for the production of fuels (FT-fuels, dimethyl ether,ethanol, isobutene. . . ) or chemicals (alcohols, organic acids, ammonia, methanoland so on).

The second thermochemical pathway for converting biomass is pyrolysis, whichuses intermediate temperatures (300 − 600◦C) in the absence of oxygen to convertthe feedstock into pyrolytic liquid (or bio-oil), solid charcoal and light gases similarto syngas. Their yields vary with the process conditions; for biore�nery purposesthe treatment which maximizes the production of liquid bio-oil is the most desirable(obtained with a �ash pyrolysis process). The application of bio-oil as a transporta-tion biofuel is nowadays problematic and its use as a source of chemicals is stillunder development. Together with charcoal, it is generally best suited as a fuel forstationary electric power or thermal energy plants.

In addition to gasi�cation and pyrolysis, direct combustion is also includedamong the thermochemical processes. This is the most common and oldest formof biomass conversion that involves burning biomass in an oxygen-rich environmentmainly for the production of heat.

1.3.3.2 Biochemical processes

Unlike thermochemical processes, biochemical processes occur at lower temper-atures and have lower reaction rates. The most common types of biochemical pro-cesses are fermentation and anaerobic digestion.

Fermentation uses microorganisms and/or enzymes to convert a fermentablesubstrate into recoverable products (usually alcohols or organic acids). Ethanol iscurrently the most required fermentation product, but the production of many otherchemical compounds (e.g. hydrogen, methanol, succinic acid, among others) is nowa-days object of many research and development activities. Hexoses (mainly glucose)are the most frequent fermentation substrates, while pentoses (sugars from hemi-cellulose), glycerol and other hydrocarbons require the development of customizedfermentation organisms to enable their conversion to ethanol.

Anaerobic digestion involves the bacterial breakdown of biodegradable organicmaterial in the absence of oxygen over a temperature range from about 30 to 65 ◦C.The main end product of these processes is biogas (a gas mixture made of methane,CO2 and other impurities), which can be upgraded up to > 97% methane contentand used as a substitute of natural gas.

1.3.3.3 Mechanical processes

Mechanical processes do not change the state or the composition of biomass,but only perform a size reduction or a separation of the feedstock components.In the biore�nery processes chain, they are usually applied �rst, because the fol-lowing biomass utilization requires reduction of the material size within speci�c


Figure 1.6: Thermochemical platform�owchart

Figure 1.7: Bioconversion platform�owchart

ranges, depending on feedstock species, handling and further conversion processes.Biomass size reduction is a mechanical treatment that refers to either cutting orcommuting processes that signi�cantly change the particles size, shape and bulkdensity of biomass. Separation processes involve the separation of the substrateinto its components, while with extraction methods valuable compounds are ex-tracted and concentrated from a bulk and inhomogeneous substrate. Lignocellulosicpre-treatment methods (e.g. the split of lignocellulosic biomass into cellulose, hemi-cellulose and lignin) fall within this category, even if some of hemicellulose is alsohydrolyzed to single sugars.

1.3.3.4 Chemical processes

Chemical processes entail a change in the chemical structure of the molecule byreacting with other substances. The most common chemical processes in biomassconversion are hydrolysis and transesteri�cation, but this group also includes thewide class of chemical reactions where a change in the molecular formula occurs.Hydrolysis uses acids, alkalis or enzymes to depolymerise polysaccharides and pro-teins into their component sugars (e.g. glucose from cellulose) or derivate chemicals(e.g. levulinic acid from glucose). Transesteri�cation is the most common methodto produce biodiesel today and is a chemical process by which vegetable oils can beconverted to methyl or ethyl esters of fatty acids, also called biodiesel. This processinvolves the co-production of glycerine, a chemical compound with multiple com-


mercial uses. Other important chemical reactions in biore�ning are Fisher�Tropschsynthesis, methanisation and steam reforming, among many others.

1.3.4 Biore�nery processing strategies

A critical juncture in the area of processing biomass into value-added productsis whether or not to take a direct or indirect product substitution approach. Inthis context, direct product substitution occurs when an existing product on themarket is produced through the development of a new processing method. Thus,direct product substitution involves producing commodities with an existing marketvalue, but using biomass as a feedstock for their production and using new processingapproaches. These products would have the advantage of increased likelihood ofproduct acceptance in the marketplace due to their established familiarity. However,existing petroleum-based infrastructure and optimized large-scale facilities make itdi�cult for biomass-derived products to compete with these established productson an economic basis.

Indirect product substitution occurs when a new product is developed that isunique to the market, but serves a similar function as an existing product. Indirectproduct substitution involves the use of new chemicals to perform similar functionsas existing chemicals, without duplicating their molecular structures. The complexchemical diversity of biomass feedstocks o�ers the opportunity to generate a widerange of new polymers and process intermediate. Indirectly substituted products donot di�er from existing products solely on their production feedstock (i.e. lignocellu-lose instead of petroleum). For indirect product substitution to be successful, theseproducts would have to perform similar functions as existing chemicals at a lowercost, or have unique properties that cannot be obtained with existing chemicals, inaddition to performing similar functions.

Figure 1.8: High-level representation of pathways via the sugar platform

Perhaps one of the most promising and realistic approaches, with a higher like-lihood of acceptance by established chemical producers, is the development of pro-


cesses to make bio-sourced intermediate chemicals(Fig.1.8). These intermediates(e.g. acrylic acid) could be used as �drop in� components to existing chemical pro-duction facilities to make higher value-added chemicals (e.g. acrylic esters), withoutrequiring additional capital investment or expensive changes to an already optimizedchemical manufacturing process. In terms of bulk chemical production on an indus-trial scale, a feedstock that is inexpensive and readily available is required. Theserequirements are met by lignocellulosic biomass. In Tab.1.2 is possible to have anoverview on some of the most important lignocellulosic biore�neries currently inoperation.

Table 1.2: Examples of lignocellulosic biore�neries

Company Feedstock FractionationProcess

Scale MainProduct

LigninUse

Beta Renewables(Italy)

agriculturalresidues

steam explosion commercial bioethanol fuel

ABNT(USA)

corn stover,wheat straw,switch grass

steam explosion commercial bioethanol fuel for steamand elecricity

POET - DSM(USA)


enzimatichydrolysis

commercial bioethanol fuel

ABNT(Spain)

wheat straw steam explosion demo bioethanol fuel,feed additives

CHEMPOLIS(Finland)


organosolv demo bioethanol fuel

INBICON(Denmark)


hydrothermal demo bioethanol fuel

CIMV(France)

agricultural residues (e.g.wheat and rice straw),hardwoods

organic acidand organosolv

pilot cellulose, C5sugars and lignin

performancematerials

SEKAB/EPAB(Sweden)

wood chips, bagasse,wheat straw, energygrass, corn stover

one step diluteacid enzimaticpretreatment

pilot bioethanol fuel and perfor-mance materials

Clariant(Germany)

wheat straw pressurized steamtreatment andenzimatic hydrolysis

pilot bioethanol fuel

1.3.5 Guidelines for future biore�neries

The choice of feedstock and �nal products are important in biore�nery designdue to the large-scale production implications. Initial feedstock availability and itspotential use in multiple production streams both need to be considered.

A biore�nery, similarly to what happens in an oil re�nery, should perform afeedstock upgrading processes, in which raw materials are continuously upgradedand re�ned. At �rst, the biomass feedstock components should be split, and then,by means of a chain of di�erent processes, a high concentration of pure chemicalspecies (e.g. ethanol) or a high concentration of molecules having similar, wellidenti�ed functions (e.g. the mixture of C alkanes in FT-fuels) should be obtained.For this reason, a certain feedstock cannot be directly burnt without any previoustreatment, since the aim of a biore�nery is to increase the value of the di�erentbiomass components as material and energy source: the most desirable option would


be to send to combustion, for heat and electricity production, only the residuesand leftovers of previous technological treatments and conversion processes. Theprevious concept leads to the following remarks:

• a biore�nery should produce at least one high value chemical or material prod-uct, besides low-grade and high-volume products (like animal feed and fertil-izers), according to the speci�cations given above,

• a biore�nery should produce at least one energy product beside heat and elec-tricity; as a consequence, the production of at least one biofuel (liquid, solidor gaseous) is required.

A biore�nery plant should also aim to run sustainably: all the energy require-ments of the processes should be internally supplied by the production of heat andelectricity from combustion of residues. For example, in a lignocellulosic ethanolplant, lignin, after separation from cellulose and hemicellulose, can be burnt to pro-vide the heat and electricity required by the plant. Since lignin can also be used toproduce chemicals and polymers, direct external fossil energy inputs are allowed ifthey ensure overall economic bene�ts and do not over burden the life-cycle environ-mental concerns.

Similarly, also solid, liquid and gaseous waste should be minimized. This targetcan be achieved in two ways:

• using all the di�erent biomass components for producing a wide spectrum ofmultiple products in one location;

• setting up industrial �bio-clusters� where material �ow exchanges among dif-ferent plants are promoted in order to transform a downstream residue of aplant into an upstream raw material for another plant.

This allows for the development of systems that ideally attempt to render theterm �waste�, in its application to biomass processing, obsolete as each productionstream has the potential to be converted into a by-product stream rather than wastestreams.

1.4 From oil re�nery to bior�nery

1.4.1 Carbohydrates vs. hydrocarbons

The raw materials utilized in biore�neries are totally di�erent from those of thecurrent oil re�nery (Fig.1.9). In fact, the crude oil is a mixture of many di�erentorganic hydrocarbon compounds. The �rst step of oil re�nery is to remove waterand impurities, then distil the crude oil into its various fractions as gasoline, dieselfuel, kerosene, lubricating oils and asphalts. Finally, these fractions can be furtherprocessed into other various industrial chemicals and products.

Unlike petroleum, biomass composition is not homogeneous, because the biomassfeedstock might be made of grains, wood, grass, biological waste and so on, and the

1.4. From oil re�nery to bior�nery 25

Figure 1.9: Oil re�nery versus biore�nery

elemental composition is a mixture of C, H and O (plus other minor componentssuch as N, S and other mineral compounds).As an example, chemical and elemen-tal composition of some lignocellulosic biomass feedstocks is reported in Table 1.3.When compared to petroleum, biomass generally has a lower hydrogen content, toomuch oxygen, and a lower fraction of carbon. The large variety in biomass feed-stocks composition is both an advantage and a drawback: biore�neries can processmore types of products that can petroleum re�neries and can rely on a wider rangeof raw materials; a relatively larger range of processing technologies is needed, andmost of these technologies are still not mature for commercialization.

In order to be used for biofuels and chemicals production, biomass needs to bedepolymerized and deoxygenated: deoxygenation is required because the presence ofoxygen in biofuels reduces the heat content of molecules and usually gives them highpolarity, which hinders blending with existing fossil fuels. Chemical applicationsmay require much less deoxygenation, since the presence of oxygen often providesvaluable physical and chemical properties to the product.

A major di�erence when compared to petroleum is that biomass is subject toseasonal changes, since harvesting is not possible all year round. A shift from crudeoil to biomass may require a change in the capacity of chemical industries, involvingthe generation of materials and chemicals on a seasonal basis. Alternatively, inorder to ensure continuous operation, the possibility of a long-term storage could beimplemented, given that the biomass undergoes a stabilization process �rst.

A biore�nery represents a change from the traditional oil re�nery relying on largeuse of natural resources and large waste production towards integrated systemsin which all resources are exploited. As an example it is worth to mention howthe existing corn wet-milling industry historically evolved. Initially, starch was the


Table 1.3: Examples of checmical and elemental composition

Parameter Unit(dry) Softwood Switchgrass Corn stover Wheat straw Petroleum

Water % 15 15 15 15LHV MJ/kg 19,6 18,6 18,5 17,6Cellulose % 44,5 35,4 38,1 32,6Glucan (C6) % 44,5 35,4 38,1 32,6

Hemicellulose % 21,9 26,5 25,3 22,6Xylan (C5) % 6,3 22,4 20,2 19,2Arabinan (C5) % 1,6 2,73 2,03 2,35Galactan (C6) % 2,56 0,96 0,74 0,75Mannan (C6) % 11,4 0,39 0,41 0,31

Lignin % 27,7 18,2 20,2 16,8Acids % 2,67 2,15 4,84 2,24Extractives % 2,88 11,5 4,78 12,9Ash % 0,32 4,28 8,59 10,2

C % 50,3 46,9 46,7 43,9 83-87H % 5,98 5,54 5,49 5,26 10-14O % 42,1 42 38,4 38,7 0,1-1,5N % 0,03 0,62 0,67 0,63 0,1-2S % 0,01 0,7 0,1 0,16 0,5-6

Source: [13]

major product but as technology developed and the need for higher value productsdrove the growth of the industry, the product portfolio expanded from various starchderivatives such as glucose and maltose syrups to high fructose corn syrup. Lateron, fermentation products derived from the starch and glucose such as citric acid,gluconic acid, lactic acid, lysine, threonine and ethanol were added. Many otherby-products, such as corn gluten, corn oil, corn �ber and animal feed are now beingproduced. The overall picture is that the development of the technical, commercialand political infrastructure of a biomass re�nery (biore�nery) makes it similar tothe current oil re�nery concept.

1.4.2 Current chemical platforms in oil re�nery

Nowadays chemical industry processes crude oil into a limited number of basefractions. Using numerous cracking and re�ning catalysts and using distillation asthe dominant separation process, crude oil is re�ned into fractions such as naphtha,gasoline, kerosene, gas oil and residues. The relative volumes of the fractions formeddepend on the processing conditions and the composition of the crude oil. Thenaphtha fraction is subsequently used as a feedstock for the production of just afew platform chemicals from which all the major bulk chemicals are subsequentlyderived. The majority of bulk chemicals can be produced starting from these fewplatform chemicals:

• ethylene,

• propylene,

1.4. From oil re�nery to bior�nery 27

• C4-ole�nes,

• the aromatics benzene, toluene and xylene (often referred to as BTX).

These hydrogen- and carbon-containing platform chemicals are subsequentlyused, for instance as solvents (benzene, toluene),starting material for polymers (ethy-lene, propylene, butadiene) or are further functionalized via the introduction of el-ements such as oxygen, nitrogen or chlorine.

1.4.3 Expected chemical platforms in biore�nery

A biore�nery industry with the aim of producing bulk chemicals from biomasswill be based on a di�erent selection of simple platforms if compared to those cur-rently used in the petrochemical industry. Given the chemical complexity of biomass,there is some choice of which platform chemicals to produce since, within limits, dif-ferent processing strategies of the same material can lead to di�erent �nal products(Fig.1.10).

Figure 1.10: Schematic overview �ow-chart: biomass-to-products

The future biore�neries are expected to be based on a limited number of plat-forms, from which all the other commodity and bulk chemicals can be derived. Inparticular, the carbohydrate fraction of biomass feedstock is expected to play thebiggest role as a renewable carbon source for biochemical products. In fact, biomasspolysaccharides can be e�ectively hydrolyzed to monosaccharides (e.g., glucose, fruc-tose and xylose) which can then be converted, via fermentations or chemical syn-thesis, to an array of bio Platform Molecules (bPM � building block chemicals with


potential use in the production of numerous value-added chemicals), analogous tothe petro-platform molecules of the current oil re�nery.

In comparison to oil-derived platform molecules, bPMs have much higher oxy-gen content. This will result in an interesting shift in chemistries from the oftenharsh and environmentally damaging oxidation procedures to largely greener reduc-tion chemistry. As opposed to adding functionality, as it normally occurs in thepetroleum-based chemical industry, there will be a switch to where a large part ofthe desired functionality or pre-functionality is already present in the substrate.

1.5 Biore�nery products

1.5.1 Biomass vs. fossils as raw materials

The products of biore�nery systems can be grouped in two broad categories:material products and energy products. Energy products are those products whichare used in light of their energy content, providing electricity, heat or transportationservice. On the other hand, material products are not used for an energy generationpurpose but for their chemical or physical properties. In some cases, a furtherdistinction for the characterization of products is needed because some products, likebiohydrogen or bioethanol, might be used either as fuels or as chemical compound inchemical synthesis. In these cases, it is necessary to identify the addressed markets,for instance the transportation sector for H2 and bioethanol.

The products of a biore�nery must be able to replace fossil fuel based productscoming from oil re�nery, both as chemicals and energy carriers. Concerning thechemicals, this objective can be met by producing the same chemical species frombiomass instead of from fossils (e.g. phenols), or producing a molecule having adi�erent structure but an equivalent function. Concerning the fuels, a biore�nerymust replace conventional fossil fuels (mainly gasoline, diesel, heavy oil, coal andnatural gas) with biofuels coming from biomass upgrading.

The most important energy products which can be produced in biore�neries are:

• gaseous biofuels (biogas, syngas, hydrogen, biomethane),

• solid biofuels (pellets, lignin, charcoal),

• liquid biofuels for transportation (bioethanol, biodiesel, FT-fuels, bio-oil).

The most important chemical and material products are the following:

• chemicals (�ne chemicals, building blocks, bulk chemicals),

• organic acids (succinic, lactic, itaconic and other sugar derivatives),

• polymers and resins (starch-based plastics, phenol resins, furan resins),

• biomaterials (wood panels, pulp, paper, cellulose),

1.5. Biore�nery products 29

• food and animal feed,

• fertilizers.

1.5.2 The role of green chemistry

Green chemistry can be considered as a set of principles for the manufacture andapplication of products whose goal is eliminating the use, or generation, of envi-ronmentally harmful and hazardous chemicals. It o�ers a range of techniques andunderlying principles that any researcher could, and should, apply when developingthe next generation of biore�neries. The overall goal of green chemistry combinedwith a biore�nery is the production of genuinely green and sustainable chemicalproducts.

During chemical product manufacture, and indeed during the whole product lifecycle, energy demands should be minimized, safer processes used, and hazardouschemical use and production avoided. The �nal product should be non-toxic, degrad-able into unharmful chemicals and with minimum production of waste.

In addition, there are numerous natural polymers directly available from biomasswith potential for physical and chemical modi�cations. These include starches,cellulose, hemicellulose, lignin,proteins and lipids. Modi�cation of natural polymersis of extreme interest, since they can replace fossil derived polymers like plasticsand textiles. For instance, lignin is a highly complex matrix of aromatic unitsand can be a renewable source of aromatic compounds so widely used in chemicalindustry. Breakdown of lignin to individual aromatic units, such as vanillin, is objectof research and development activities.

The range of chemicals and materials that future biore�neries could produce isextensive, and with further research the selection will become larger. As argued byClark et al., environmental impact evaluation methodologies like LCA and metricssuch as atom economy should become as important in measuring the sustainabilityof a chemical process as yield and selectivity are today.

Chapter 2

The Life Cycle Assessment

methodology

According to Jim Fava, one of the fathers of LCA, �life cycle assessment hasbecome a recognized instrument to assess the ecological burdens and human healthimpacts connected with the complete life cycle of products, processes and activities,enabling the practitioner to model the entire system from which products are derivedor in which processes and activities operate.�

2.1 LCA: an hystorical perspective

The �rst studies to look at life cycle aspects of products and materials date fromthe late sixties and early seventies, and focused on issues such as energy e�ciency,the consumption of raw materials and, to some extent, waste disposal.

The �rst well-known environmental study was conducted in 1969 by Coca-Cola.The study compared beverage containers and showed that all container materialshad a real environmental impact and some materials had a greater impact thanothers. Coca-Cola acted not by removing the worst-performing materials from theirproducts but by working with local authorities to develop a take-back scheme andrecycling infrastructure to collect aluminum cans. In doing so, the company realizeda 90% reduction in the energy used throughout the can's lifetime.

Meanwhile, in Europe, a similar inventory approach was being developed, laterknown as the `Ecobalance'. In 1972, in the UK, Ian Boustead calculated the total en-ergy used in the production of various types of beverage containers, including glass,plastic, steel, and aluminium. Over the next few years, Boustead consolidated hismethodology to make it applicable to a variety of materials, and in 1979, publishedthe Handbook of Industrial Energy Analysis.

Initially, energy use was considered a higher priority than waste and outputs.Because of this, there was little distinction, at the time, between inventory devel-opment (resources going into a product) and the interpretation of total associatedimpacts. But after the oil crisis subsided, energy issues declined in prominence.

During the 1970s and '80s, many approaches to reducing environmental harm

31

32 Chapter 2. The Life Cycle Assessment methodology

included regulatory control of point-source waste releases. Because these approacheswere based on a single stage of a product's life, such as production, or a single issue,such as wastewater, they were not particularly e�ective in achieving net environ-mental bene�ts. What they did achieve was a change in the way people thoughtabout business and environmental management.

In 1979, the Society of Environmental Toxicology and Chemistry (SETAC) wasfounded to serve as a non-pro�t professional society to promote multi-disciplinaryapproaches to the study of environmental issues. SETAC's other founding principlesinclude multidisciplinary approaches to solving environmental problems; tripartitebalance among academia, business and government and science-based objectivity.

At the �rst SETAC-sponsored international workshop in 1990, the term �life cycleassessment� (LCA) was coined. The advantage of LCA over point-source regulation,is that it avoids shifting a product's environmental burden to other life cycle stagesor to other parts of the product system.

Beginning in 1993, the International Organization for Standardization (ISO)tasked a small group of SETAC LCA experts with making a recommendation re-garding the need to standardize LCA. The group's recommendation was to proceedwith standardization, and by 1997 the ISO14040 standard for Life cycle assessment� Principles and framework was complete. A number of additional standards weredeveloped and ultimately reviewed and compiled in 2006 in the form of ISO 14044Life cycle assessment � Requirements and guidelines.

2.2 LCA applications

LCA methodologies were originally developed to create decision support toolsfor distinguishing between products, product systems, or services on environmentalgrounds. During the evolution of LCA, a number of related applications emerged,of which some examples are given below:

• Internal industrial use in product development and improvement

• Internal strategic planning and policy decision support in industry,

• External industrial use for marketing purposes, and

• Governmental policy making in the areas of ecolabelling, green procurementand waste management opportunities.

2.2.1 Levels of sophistication in LCA for di�erent applica-

tions

Conceptual LCA - Life Cycle Thinking is the �rst and simplest level of LCA.At this level the life cycle approach is used to make an assessment of environmentalaspects based on a limited and usually qualitative inventory.

2.2. LCA applications 33

Table 2.1: Level of detail in some applications of LCA

Level of detail in LCA

Application Conceptual Simpli�ed Detailed Remarks

Design for Environment # No formal links to LCA

Product development # # Large variation in sophistication

Product improvement # Often based on already existing products

Environmental claims(ISO type II-labelling)

Seldom based on LCA

Ecolabelling(ISO type I-labelling)

# Only criteria development requires anLCA

Environmental declaration(ISO type III-labelling)

# Inventory and/or impact assessment

Organisation marketing # Inclusion of LCA in environmental report-ing

Strategic planning Gradual development of LCA knowledge

Green procurement # LCA not as detailed as in ecolabelling

Deposit/refund schemes # Reduced number of parameters in theLCA is often su�cient

Environmental("green") taxes # Reduced number of parameters in the

LCA is often su�cient

Choice betweenpackaging systems # Detailed inventory, Scope disputed LCA

results not the only information

indicates the most frequently used level. Ref. [8]

The results of a conceptual LCA can for instance be presented using qualitativestatements or simple scoring systems, indicating which components or materialshave the largest environmental impacts, and why.

It is obvious from the requirements of the ISO standard that conceptual LCAsare not suitable for marketing purposes or other public dissemination of the results.However, a conceptual LCA may help the decision maker identify which productshave a competitive advantage in terms of reduced environmental impacts. Subse-quent simpli�ed or detailed LCAs ful�lling the requirements of a standard can beestablished and used for public information.

Instead of �Conceptual LCA�, the SETAC EUROPE LCA Screening and Stream-lining Working Group uses the term �Life Cycle Thinking�:

Life Cycle Thinking is a mostly qualitative discussion to identify stages of thelife cycle and/or the potential environmental impacts of greatest signi�cance e.g.for use in a design brief or in an introductory discussion of policy measures. Thegreatest bene�t is that it helps focus consideration of the full life cycle of the productor system; data are typically qualitative (statements) or very general and available-by-heart quantitative data.

Simpli�ed LCA is an application of the LCA methodology for a comprehensivescreening assessment i.e. covering the whole life cycle but super�cial e.g. usinggeneric data (qualitative and/or quantitative), standard modules for transportation


or energy production, followed by a simpli�ed assessment i.e. focusing on the mostimportant environmental aspects and/or potential environmental impacts and/orstages of the life cycle and/or phases of the LCA and a thorough assessment of thereliability of the results.

The aim of simplifying LCA is to provide essentially the same results as a detailedLCA, but with a signi�cant reduction in expenses and time used. Simpli�cationpresents a dilemma, however, since it is likely to a�ect the accuracy and reliabilityof the results of the LCA. Thus, the primary object of simpli�cation is to identifythe areas within the LCA which can be omitted or simpli�ed without signi�cantlycompromising the overall result.

Simpli�cation of LCA consists of three stages which are iteratively interlinked:

• Screening: identifying those parts of the system (life cycle) or of the elementary�ows that are either important or have data gaps;

• Simplifying: using the �ndings of the screening in order to focus further workon the important parts of the system or the elementary �ows;

• Assessing reliability: checking that simplifying does not signi�cantly reducethe reliability of the overall result.

Simpli�ed LCAs may be used externally if reported in accordance with the re-quirements in the ISO standard (ISO 14040). However, most simpli�ed LCAs areused for internal purposes without formal requirements for reporting. To avoid mis-interpretation of the results, the user of the LCA should be made explicitly aware ofthe limitations of the study, e.g. by stating all simplifying methods applied in theLCA.

The level of detail in some of the applications is shown in Table 2.1.

2.3 Methodological framework

As shown in Figure 2.1 the life cycle assessment framework is described by fourphases:

• goal and scope de�nitions

• inventory analysis

• impact assessment

• interpretation

The double arrows between the phases indicate the interactive nature of LCAas illustrated by the following examples: when doing the impact assessment it canbecome clear that certain information is missing which means that the inventoryanalysis must be improved, or the interpretation of the results might be insu�cientto ful�l the needs required by the actual application which means that the goal andscope de�nition must be revised.

2.3. Methodological framework 35

Figure 2.1: Life cycle assessment framework - phases of an LCA (ISO, 1997a).

2.3.1 Goal and scope de�nition

Goal and scope de�nition is the �rst phase in a life cycle assessment containingthe following main issues:

• goal

• scope

• functional unit

• system boundaries

• data quality

The de�nition of the goal and scope is the critical parts of an LCA due to thestrong in�uence on the result of the LCA. The following minimum decisions andde�nitions that need to be made are listed:

- the purpose and intended application

- the function of the studied systems(s) and a de�ned functional unit

- the studied product group and chosen alternatives, if relevant

- the system boundaries applied


- the data quality needed

- the validation or critical review process needed

Goal. The de�nition of the purpose of the life cycle assessment is an importantpart of the goal de�nition.

The goal of an LCA study shall unambiguously state the intended application,including the reasons for carrying out the study and the intended audience, i.e. towhom the results of the study are intended to be communicated.

The goal de�nition also has to de�ne the intended use of the results and usersof the result. The practitioner, who has to reach the goal, needs to understandthe detailed purpose of the study in order to make proper decisions throughout thestudy. Examples of goals of a life cycle assessment are:

- to compare two or more di�erent products ful�lling the same function with thepurpose of using the information in marketing of the products or regulatingthe use of the products

- to identify improvement possibilities in further development of existing prod-ucts or in innovation and design of new products

- to identify areas, steps etc. in the life cycle of a product where criteria can beLife Cycle Assessment (LCA) set up as part of the ecolabelling criteria to beused by e.g. the ecolabelling board

The goal de�nition determine the level of sophistication of the study and therequirements to reporting. Transparency is essential for all kind of LCA studies.The target group of the LCA study is also important to have in mind in the choiceof reporting method.

The goal can be rede�ned as a result of the �ndings throughout the study e.g.as a part of the interpretation.

Scope. The de�nition of the scope of the life cycle assessment sets the bordersof the assessment: what is included in the system and what detailed assessmentmethods are to be used.

In de�ning the scope of an LCA study, the following items shall be consideredand clearly described:

- the functions of the system, or in the case of comparative studies, systems;

- the functional unit;

- the system to be studied;

- the system boundaries;

- allocation procedures;

- the types of impact and the methodology of impact assessment and subsequentinterpretation to be used;


- data requirement;

- assumptions;

- limitations;

- the initial data quality requirements;

- the type of critical review, if any;

- the type and format of the report required for the study

The scope should be su�ciently well de�ned to ensure that the breadth, thedepth and the detail of the study are compatible and su�cient to address the statedgoal.

LCA is an iterative technique. Therefore, the scope of the study may need to bemodi�ed while the study is being conducted as additional information is collected.

Functional Unit. De�nition of the functional unit or performance character-istics is the foundation of an LCA because the functional unit sets the scale forcomparison of two or more products including improvement to one product (sys-tem). All data collected in the inventory phase will be related to the functionalunit. When comparing di�erent products ful�lling the same function, de�nition ofthe functional unit is of particular importance.

One of the main purposes for a functional unit is to provide a reference to whichthe input and output data are normalised. A functional unit of the system shall beclearly de�ned and measurable. The result of the measurement of the performanceis the reference �ow.

Comparisons between systems shall be done on the basis of the same function,measured by the same functional unit in the form of equivalent reference �ows.

Three aspects have to be taken into account when de�ning the functional unit:

- the e�ciency of the product

- the durability of the product

- the performance quality standard

When performing an assessment of more complicated systems (e.g. multi-functional)systems special attention has to be paid to byproducts.

If additional functions of one or other of the systems are not taken into accountin the comparison of functional units then these omissions shall be documented.For example, systems A and B perform functions x and y which are representedby the selected functional unit, but system A also performs function z which is notrepresented in the functional unit. As an alternative, systems associated with thedelivery of function z may be added to the boundary of system B to make the systemsmore comparable. In these cases, the selected processes shall be documented andjusti�ed.


System boundaries.The system boundaries de�ne the processes/ operations(e.g. manufacturing, transport, and waste management processes), and the inputsand outputs to be taken into account in the LCA. The input can be the overallinput to a production as well as input to a single process - and the same is true forthe output. The de�nition of system boundaries is a quite subjective operation andincludes the following boundaries: geographical boundaries, life cycle boundaries(i.e. limitations in the life cycle) and boundaries between the technosphere andbiosphere. Due to the subjectivity of de�nition of system boundaries, transparencyof the de�ning process and the assumptions are extremely important.

The initial system boundary de�nes the unit processes which will be included inthe system to be modelled. Ideally, the product system should be modelled in such amanner that the inputs and outputs at its boundary are elementary �ows. However,as a practical matter, there typically will not be su�cient time, data, or resources toconduct such a comprehensive study. Decisions must be made regarding which unitprocesses will be modelled by the study and the level of detail to which these unitprocesses will be studied. Resources need not be expended on the quanti�cation ofminor or negligible inputs and outputs that will not signi�cantly change the overallconclusions of the study. Decisions must also be made regarding which releases tothe environment will be evaluated and the level of detail of this evaluation. Thedecision rules used to assist in the choice of inputs and outputs should be clearlyunder stood and described.

Any omission of life cycle stages, processes or data needs should be clearly statedand justi�ed. Ultimately, the sole criterion used in setting the system boundaries isthe degree of con�dence that the results of the study have not been compromisedand that the goal of a given study has been met.

Data quality. The quality of the data used in the life cycle inventory is naturallyre�ected in the quality of the �nal LCA. The data quality can be described andassessed in di�erent ways. It is important that the data quality is described andassessed in a systematic way that allows others to understand and control for theactual data quality.

Initial data quality requirements shall be established which de�ne the followingparameters:

- Time-related coverage: the desired age (e.g. within last 5 years) and theminimum length of time (e.g. annual);

- Geographical coverage: geographic area from which data for unit processesshould be collected to satisfy the goal of the study (e.g. local, regional, na-tional, continental, global);

- Technology coverage: nature of the technology mix (e.g. weighted average ofthe actual process mix, best available technology or worst operating unit).

Further descriptions which de�ne the nature of the data collected from speci�csites versus data from published sources, and whether the data should be measured,calculated or estimated shall also be considered.


Data from speci�c sites should be used for those unit processes that contributethe majority of the mass and energy �ows in the systems being studied as determinedin the sensitivity analysis . . . . Data from speci�c sites should also be used for unitprocesses that are considered to have environmentally relevant emissions.

In all studies, the following additional data quality indicators shall be taken intoconsideration in a level of detail depending on goal and scope de�nition:

- Precision: measure of the variability of the data values for each data categoryexpressed (e.g. variance);

- Completeness: percentage of locations reporting primary data from the po-tential number in existence for each data category in a unit process;

- Representativeness: qualitative assessment of the degree to which the data setre�ects the true population of interest (i.e. geo graphic and time period andtechnology coverage);

- Consistency: qualitative assessment of how uniformly the study methodologyis applied to the various components of the analysis;

- Reproducibility: qualitative assessment of the extent to which informationabout the methodology and data values allows an independent practitioner toreproduce the results reported in the study.

Where a study is used to support a comparative assertion that is disclosed tothe public, the above mentioned data quality indicators shall be included.

2.3.2 Inventory analysis

Inventory analysis is the second phase in a life cycle containing the followingmain issues:

• data collection

• re�ning system boundaries

• calculation

• validation of data

• relating data to the speci�c system

• allocation

Data collection. The inventory analysis includes collection and treatment ofdata to be used in preparation of a material consumption, waste and emission pro�lefor all the phases in the life cycle, but also for the whole life cycle. The data can besite speci�c e.g. from speci�c companies, speci�c areas and from speci�c countriesbut also more general e.g. data from more general sources e.g. trade organisations,


public surveys etc. The data have to be collected from all single processes in thelife cycle. These data can be quantitative or qualitative. The quantitative dataare important in comparisons of processes or materials, but often the quantitativedata are missing or the quality is poor (too old or not technologically representativeetc.). The more descriptive qualitative data can be used for environmental aspectsor single steps in the life cycle that cannot be quanti�ed, or if the goal and scopede�nition allow a nonquantitative description of the conditions.

Inventory analysis involves data collection and calculation procedures to quantifyrelevant inputs and outputs of a product system. These inputs and outputs mayinclude the use of resources and releases to air, water and land associated with thesystem. Interpretation may be drawn from these data, depending on the goals andscope of the LCA. These data also constitute the input to the life cycle impactassessment.

The process of conducting an inventory analysis is iterative. As data are collectedand more is learned about the system, new data requirements or limitations may beidenti�ed that require a change in the datacollection procedures so that the goals ofthe study will still be met. Sometimes, issues may be identi�ed that require revisionsto the goal or scope of the study.

The qualitative and quantitative data for inclusion in the inventory shall becollected for each unit process that is included within the system boundaries. Theprocedures used for data collection may vary depending on the scope, unit processor intended application of the study. Data collection can be a resource intensiveprocess. Practical constraints on data collection should be considered in the scopeand documented in the report.

Some signi�cant calculation considerations are outlined in the following:

- allocation procedures are needed when dealing with systems involving multi-ple products (e.g. multiple products from a biore�nery). The materials andenergy �ows as well as associated environmental releases shall be allocated tothe di�erent products according to clearly state procedures, which shall bedocumented and justi�ed;

- the calculation of energy �ow should take into account the di�erent fuels andelectricity sources used, the e�ciency of conversion and distribution of energy�ow as well as the inputs and outputs associated with the generation and useof that energy �ow.

Data collection is often the most work intensive part of a life cycle assessment,especially if site speci�c data are required for all the single processes in the lifecycle. In many cases average data from the literature (often previous investigationsof the same or similar products or materials) or data from trade organisations areused. A number of European trade organisations have published or plan to publish�cradle-to-gate� data that include information on inputs and outputs for materialsthrough production of semi-manufactured product to �nal products.

The average data can be used in the conceptual or simpli�ed LCA to get a �rstimpression of the potential inputs and outputs from producing speci�c materials.


When doing a detailed LCA site speci�c data must be preferred. Average dataare often some years old and therefore do not represent the latest in technologicaldevelopment.

Re�ning system boundaries. The system boundaries are de�ned as a part ofthe scope de�nition procedure. After the initial data collection, the system bound-aries can be re�ned e.g. as a result of decisions of exclusion life stages or sub-systems,exclusion of material �ows or inclusion of new unit processes shown to be signi�cantaccording to the sensitivity analysis.

Re�ecting the iterative nature of LCA, decisions regarding the data to be in-cluded shall be based on a sensitivity analysis to determine their signi�cance, therebyverifying the initial analysis. The initial product system boundary shall be revisedin accordance with the cut-o� criteria established in the scope de�nition.

The results of this re�ning process and the sensitivity analysis shall be docu-mented. This analysis serves to limit the subsequent data handling to those inputand output data which are determined to be signi�cant to the goal of the LCA study.

Calculation procedures. No formal demands exist for calculation in life cycleassessment except the demands for allocation procedures. Due to the amount ofdata it is recommended as a minimum to develop a spreadsheet for the speci�c pur-pose. A number of general PC programs/ software for calculation are available e.g.spreadsheets/spreadsheet applications (EXCEL, etc.), together with many softwareprograms developed specially for life cycle assessment. The appropriate programcan be chosen depending on the kind and amount of data to be handled.

Relating data. For each unit process, an appropriate reference �ow shall bedetermined (e.g. one kilogram of material or one megajoule for energy). The quan-titative input and output data of the unit process shall be calculated in relation tothis reference �ow.

Based on the re�ned �ow chart and systems boundary, unit processes are inter-connected to allow calculations of the complete system. This is accomplished bynormalising the inputs and outputs of a unit process in the system to the functionalunit and then normalising all upstream and downstream unit processes accordingly.The calculation should result in all system input and output data being referencedto the functional unit. Care should be taken when aggregating the inputs and out-puts in the product system. The level of aggregation should be su�cient to satisfythe goal of the study.

Data categories should only be aggregated if they are related to equivalent sub-stances and to similar environmental impacts. If more detailed aggregation rulesare required, they should be justi�ed in the goal and scope de�nition phase of thestudy or this should be left to a subsequent impact assessment phase.

The reference �ow or functional unit shall be de�ned in order to describe andcover the actual production/function of the considered product e.g. by number ofhours the actual machinery is in action per week or the actual emission of wastewaterfrom the process. If this is not the case it will not be possible to relate data to theactual product.

Allocation and recycling. When performing a life cycle assessment of a com-plex system, it may not be possible to handle all the impacts and outputs inside the


system boundaries. This problem can be solved by:

1. expanding the system boundaries to include all the inputs and outputs;

2. allocating the relevant environmental impacts to the studied system

When avoiding allocation by e.g. expanding the system boundaries there is arisk of making the system too complex. The data collection, impact assessment andinterpretation can then become too expensive and unrealistic in time and money.Allocation may be a better alternative, if an appropriate method can be found forsolving the actual problem.

Allocation can be necessary when dealing with:

- Multi-output �black box� processes, i.e. when more than one product isproduced and some of those product �ows are crossing the system boundaries.

- Multi-input processes, such as waste treatment, where a strict quantitativecausality between inputs and emissions etc. seldom exists.

- Open-loop recycling, where a waste material leaving the system boundariesis used as a raw material by another system, outside the boundaries of thestudied system.

On the basis of the principles presented above, the following descending order ofallocation procedures is recommended:

1. Wherever possible, allocation should be avoided or minimised. This may beachieved by subdividing the unit process into two or more sub-processes, someof which can be excluded from the system under study. Transport and mate-rials handling are examples of processes which can sometimes be partitionedin this way. For systems which deliver more than one product or function,or involve recycle streams, allocation may be avoided or reduced by includ-ing further unit processes thereby expanding the system boundaries so thatinputs, outputs or recycles remain within the system.

2. Where allocation cannot be avoided, the system inputs and outputs should bepartitioned between its di�erent products or functions in a way which re�ectsthe underlying physical relationships between them; i.e. they must re�ect theway in which the inputs and outputs are changed by quantitative changes inthe products or functions delivered by the system. These �causal relationships�between �ows into and out of the system may be represented by a processmodel, which can also represent the economic relationship of the system. Theresulting allocation will not necessarily be in proportion to any simple measuresuch as mass or molar �ows of co-products.

3. Where physical relationship cannot be established or used as the basis forallocation the inputs should be allocated between the products and functionsin a way which re�ects economic relationships between them. For example,burdens might be allocated between co-products in proportion tothe economic value of the products.

Chapter 3

Optimization and linear

programming

Linear programming (LP) (also called linear optimization) is a method to achievethe best outcome (such as maximum pro�t or lowest cost) in a mathematical modelwhose requirements are represented by linear relationships. Linear programming isa special case of mathematical programming (mathematical optimization).

More formally, linear programming is a technique for the optimization of a linearobjective function, subject to linear equality and linear inequality constraints. Itsfeasible region is a convex polytope, which is a set de�ned as the intersection ofmany �nite half spaces, each of which is de�ned by a linear inequality. Its objectivefunction is a real-valued a�ne (linear) function de�ned on this polyhedron. A linearprogramming algorithm �nds a point in the polyhedron where this function has thesmallest (or largest) value, if such a point exists.

Linear programs are problems that can be expressed in canonical form as:

maximize cTxsubject to Ax ≤ band x ≥ 0

where x represents the vector of variables (to be determined), c and b are vectorsof (known) coe�cients, A is a (known) matrix of coe�cients, and ()T is the matrixtranspose.

The expression to be maximized or minimized is called the objective function(cTx in this case).

The inequalities Ax ≤ b and x ≥ 0 are the constraints which specify a convexpolytope over which the objective function is to be optimized. In this context, twovectors are comparable when they have the same dimensions. If every entry in the�rst is less-than or equal-to the corresponding entry in the second, then we can saythe �rst vector is less-than or equal-to the second vector.

Linear programming can be applied to various �elds of study. It is widely used inbusiness and economics, and is also utilized for some engineering problems. Indus-

43

44 Chapter 3. Optimization and linear programming

tries that use linear programming models include transportation, energy, telecom-munications, and manufacturing. It has proved useful in modeling diverse types ofproblems in planning, routing, scheduling, assignment, and design.

Part II

Case study

45

Chapter 4

The system under study: three

biore�nery concepts

The goal of the analysis is twofold. At �rst the aim was to assess which biore�neryconcept is the one that guarantees the higher advantage in terms of environmentalemission savings (through an LCA study). The second step was to integrate an LCAinto an optimization framework and to �nd the best combination (lowest GWP) ofpossible concepts constructions in a certain region. In this case, three regions inGermany were considered, with a su�cient beech wood availability[43].

Based on the de�ned frame conditions and the objective of minimal costs, the op-timal reaction routes and appropriate �owsheet con�guration, consisting of variousprocess units and the corresponding operating levels are engineered. For that pur-pose and for mass and energy balances, the process simulation software AspenP lusr

was used [12] [45]. A capacity to process 400,000 dry tonnes of beech wood annuallywas assumed. The plant operates 8000 hours per year while the remaining 760 hoursare for maintenance, repair and restarting operations.

It is assumed that fresh beech wood with a water content (w) of 50% is processedwithin the re�nery. A seen in chapter 1, beech wood consists mainly of cellulose,hemicellulose and lignin and some minor components like starch, ash, lipids andproteins; for this reason a pre-treatment is necessary in order to split the wood intoits components. From an operational standpoint, beech wood logs are stored in suchan amount to ensure a continuous operation of the biore�nery.

4.1 Products Portfolio

From a systematic point of view, the design of biore�neries is an open andcomplex problem due to the various raw materials, the potential products and thenovelty of processes and technologies. Furthermore, biore�neries will always becompared to and have to compete with conventional systems. It follows that theywill have to achieve maximum e�ciency with a better design and process integration.For the design of biore�neries, the speci�cation of the feedstock as well as the mainproducts and by-products was found to be a practical approach. In this study, beech

47

48 Chapter 4. The system under study: three biore�nery concepts

wood was chosen as an exemplary feedstock, since it was identi�ed as a promisingraw material for the production of platform chemicals. In order to decide on asuitable product portfolio, the following �ve criteria were formulated:

• High theoretical product yields from substrate

• Market interest in the product as an end product or as an industrially impor-tant intermediate

• High production volume (current or potential)

• Nonfood use of the product

• Ability to be biologically synthesized from the common sugars derived fromvarious forms of biomass

4.2 Biore�nery Concept 1

Ethanol produced on a large scale was selected as the primary product of thebiore�nery concept 1. The second main product is organosolv lignin. Compared toother types of lignin (e.g. kraft lignin, ligninsulfonate), it has a signi�cantly higherquality and purity. Organosolv lignin has a great market potential in the �eld ofbio-based binding agents as well as thermosetting and thermoplastic compounds.However, one precondition for introducing organosolv lignin as an alternative com-modity is its stable availability and consistent quality. Biomethane is produced fromthe residual streams of downstream processes and can be fed into the gas pipelinenetwork or used for the supply of biofuels. The solid by-product hydrolysis lignin(accumulated during the conversion of cellulose to sugars) is assumed to be used asa solid fuel. (See the scheme in Fig. 4.1)

Figure 4.1: Scheme for the biore�nefy concept 1

4.2. Biore�nery Concept 1 49

Firstly, the wood logs are chipped and transported by a conveyor to the organo-solv process during which the lignocellulosic material is fractionated. Then thewashed �bers from the pre-treatment (mainly cellulose and minor portions of hemi-cellulose and lignin) are enzymatically hydrolyzed to hexoses (glucose) and pentoses(xylose) by adding the enzymes cellulase (Celluclast 1.5 L) and beta-glucosidase(Novozyme 188 ).Following the hydrolysis reactor, a screw press removes the solids that have not beenconverted (so-called hydrolysis lignin) from the liquid sugar solution (mash). Themash is then sent for alcoholic fermentation and the hydrolysis lignin is washed withwater to remove remaining sugars.Subsequently, mash, yeast (Saccharomyces Cerevisiae) and nutrients are fed to thefermentation reactor where beside the formation of ethanol various side productsare formed during fermentation. The liquid fraction of the reactor, the so-calledfermentation broth, has an ethanol concentration of 12,2 vol%.The fermentation broth consists of ethanol, a variety of undesired by-products anda large share of water, therefore distillation and recti�cation are used to recoverethanol from the fermentation broth with a purity of 95 vol%. The liquid stream atthe bottom, the so-called stillage, is delivered to a biogas plant.The C5 fraction from the organosolv process and stillage from the ethanol recoverycontain valuable by-products such as xylose, acetic acid and furans. To exploit thewhole biomass potential, the C5 fraction and stillage are then sent to an anaerobicdigester where nutrients, bu�er solution and trace elements are added to ensure astable biogas production. To feed the biogas into the gas pipeline network or to useit as a biofuel, a CH4 content of 98 vol% (corresponding to natural gas type H) isnecessary. In order to ful�ll this requirement, the biogas is upgraded.In Fig. 4.2 is illustrated the �owchart for the biore�nery concept 1. It shows thestreams incoming and outgoing from the biore�nery, including the upstream activ-ities deriving from the production of a certain material. For the Life Cycle Assess-ment, the biore�nery system was considered as a "black box": mass and energy �owsin and out the biore�nery were derived from a computer simulation; starting fromthese results it was possible to expand the system boundaries including at �rst allthe upstream activities and then even further by the avoided burden methodology.In the chart all the main feedstocks for the biore�nery are included, plus the addi-tional materials, services and energy �uxes.Also the transportation process of the materials was included, to take into accountthe relative environmental interventions; in the �owchart this is indicated with theletter T.Since the nutrients input in the biore�nery was given as mixed intake, rather thanthat of single materials, a mixing process was needed in order to be consistent withthe model that will be further used for the LCA calculations.It can be also noticed that on the right-hand side of the �owchart, downstream thebiore�nery process,an indication is given on the hypotheses made for the materialsdisplaced by the biore�nery products.


Figure 4.2: Flowchart used for the LCA of the biore�nery concept 1. T=transport;M=mixing; upside-down trapeziodal blocks=displaced material (avoided bur-den methodology)


Ethylene was selected as the primary product of the biore�nery concept 2; theconversion pathway to produce ethylene from lignocellulosic biomass is relativelysimple and well known. For what concerns the other products same considerationsas for the concept 1 can be done. Therefore the other main products are still:organosolv lignin, biomethane and the by-product hydrolysis lignin.



The process blocks for the biore�nery concept 2 are mainly the same as theprevious concept, except for the production of ethylene from ethanol through dehy-dration, which is a catalytic endothermic reaction. The use of (adiabatic) reactorsto keep the operating temperature within a certain range (335 − 500◦C ) is thennecessary (Fig.4.3).

To obtain marketable polymer-grade ethylene, several puri�cation steps are nec-essary. Firstly, the water present in the outlet gas stream is condensed in a quenchtower, by cooling the gas with spray water from the top. The dried gas is then com-pressed to enable a su�cient pressure level and favorable operating temperatures forfurther down streaming units. Thereafter, CO2 is absorbed in a caustic tower, bywashing the gas with sodium hydroxide (NaOH) in a packed column. To remove anyremaining NaOH, the gas is washed with spray water from the top of the column.The remaining water in the ethylene-rich gas is then removed using a molecularsieve. For the separation of heavier impurities, the ethylene-rich gas is cooled tosaturation and sent to a cryogenic distillation column. Ethylene has to meet theserequirements in order to obtain desirable product qualities for polymerization andto avoid stress to the oxygen-sensitive polyethylene catalyst.

In Fig.4.4 the �owchart for the biore�nery concept 2 is reported. It can beseen that the main di�erence when compared to the previous case is the need for arefrigeration medium and extra energy (heat) due to the dehydration process.




In the third biore�nery concept a much di�erent scenario in terms of outputswas considered, focusing on the possible bio-plastic end-products deriving from lig-nocellulosic feedstock, rather than energy or platform chemical use.

The primary product of the biore�norey concept 3 was chosen to be lactic acid,which could be easily further processed into poli-lactic acid (PLA) to produce bio-plastic based materials. Similarly to the other two biore�nery concepts, the otherproducts are biomethane, hydrolysis lignin and organosolv lignin. Furthermore,some calcium sulphate is produced as a by-product in the fermentation process andrepresents another marketable product.



From the process structure standpoint, the big di�erence in this biore�nery con-cept is represented by the fermentation unit. In this speci�c case, the fermenter hasto be optimized and tuned in order to produce lactic acid, so some crucial changesare found with respect to the previous two concepts. There is a need for a so called"Inoculum" which is mainly composed of microorganisms (Lactobacilli) capable ofcarrying out the fermentation; it has to be growth in a laboratory, with the additionof some nutrients and then it is fed to the fermenter along with other components(e.g. calcium hydroxide).Downstream the fermenter a concentrator unit is needed to obtain the requestedquality of the marketable output.

As seen for the previous cases, a �owchart in terms of materials and energy �owsis available (Fig. 4.6) and is the one that will be used for the LCA study.

Chapter 5

LCA results

Consequential LCA is depicted as probably the most indicated tool for the assess-ment of environmental impacts of new bio-based technologies. Consequential LCAattempts to identify in which way �ows are going to change due to the consequencesof a certain decision. In contrast, attributional LCA focuses on the established lifecycle of a product and the attribution of the impacts to this speci�c product.

5.1 Goal and Scope

The objective of the LCA study is, at �rst, to assess the impacts associated witheach of the three biore�nery concepts (attributional LCA) and then, to evaluatewhat are the emissions in case the biore�nery products are used instead of the onesderiving from current production processes (consequential LCA).

The functional unit that is considered in the study is the amount of treated beechwood from each biore�nery (400000 t(dry)/year). This choice was made in orderto be able to compare the results even if the three concepts have rather di�erentproducts portfolios. Further on, the results can be found to be expressed in termsof emissions per hour, per year and per kg of treated beech wood.

The boundaries set for the attributional LCA study include all the upstreamactivities concerning the production of the feedstocks, the construction of the biore-�nery plant, etc. up to the �nal products (Cradle-to-Gate); they are then extendedfor the consequential LCA by using the avoided burden method (described in thefollowing sections).

5.2 The Inventory Analysis model

Life Cycle Inventory analysis (LCI) is de�ned as a phase of Life Cycle Assessment(LCA) involving the compilation and quanti�cation of inputs and outputs for a givenproduct system throughout its life cycle.

It will be here presented the general model for the inventory analysis methodologywhich was used to carry out the LCA computational steps for the case studiesreported in the previous chapter.

55

56 Chapter 5. LCA results

For the sake of simplicity, all the following examples will involve the biore�neryconcept 1 (Bioethanol), but the same framework can be extended to all the othersconcepts, for which we refer the reader to the Appendixes A and B.

The computational structure presented below is suggested by Heijungs and Suh[50] and adapted to the case under study accordigly and consistently.

5.2.1 Representation of processes and �ows

A �rst step in a formalized treatment is the construction of a suitable systemfor the representation of quanti�ed �ows in connection with process units. For thisreason, the notion of linear space is introduced: a linear space is an abstract conceptwhich allows us to uniquely represent a multi-dimensional data point as a simplevector with a de�nite value of each of the coordinates. For instance, consider aprocess units (or process, in short) which uses a certain amount of input materialsto produce some speci�c quantities of products. Moreover, in doing so, it emits acertain quantities of pollutants and uses a certain amount of resources. A linearspace can now help us to describe this process unit in a very concise notation. Wewill consider as a clarifying example the Biore�nery concept 1.

In terms of linear spaces, the basis (left side) and the coordinates of the processunit, with respect to this process (right side) are:

MJ of energy from natural gaskg of magnesium sulphate

kg of phosphoruskg of sulfur

kg of nutrients (P, S,Mg)kg of transported nutrients

kg of woodchipskg of transported wood chips

kg of yeastkg of transported yeast

kg of enzymeskg of transported enzymeskg of sodium hydroxide

kg of transported sodium hydroxidekg of lime

kg of transported limepieces of refinery building

kg of sulfuric acidkg of transportedsulfuric acidm3 of treated waste watertkm of transportation

kg of bioethanolkg of biomethane

kg hydrolysis ligninkg of organosolv lignin

−−kg of carbon dioxidekg of dinitrogen oxide

kg of methanekg of hard coal used

kg of soft coal (lignite) usedNm3 of natural gas used

kg of crude oil usedm2 per year of land used

m3 of water used

=

−345860, 80000

−133, 80

−500000

−289, 70−1390−630−4, 8

−4, 2E − 060−470−394, 8

04469, 98722, 622684, 27770, 6−−155000

1056, 600000

128461, 3

(~)

The negative sign represents a conventional indication for the direction of the�ow. Here, the negative coordinates indicate an input, while the other positive

5.2. The Inventory Analysis model 57

coordinates indicate outputs with the exception of the last nine rows of the basiswhere the negative sign is implicit in the utilization of a certain resource.

Also notice that the vector that represents the process unit of the biore�nery con-cept 1 has thirty-four coordinates in a de�nite order. Is not possible to interchangethe elements of the vector, unless we change the order of the basis accordingly.

A third type of convention is related to the choice of units. We could change theunit of measurement of a certain element of the basis, provided that we scale therespective coordinate of the process vector as to keep the consistency.

We will be dealing with large systems comprising many di�erent process units,a second step is therefore the representation of such a system of process units. Ingeneral terms, the system will incorporate n processes vector corresponding to thesame number of process units:

p1, p2, p3, . . . , pn.

Which can be represented in a concise notation as:

P = (p1|p2| . . . |pn)

The P matrix will be referred to as the process matrix. Observe that a newconvention is needed to express the fact that the columns order is univocal: eachcolumn corresponds to the associated process unit which is being represented.

A third step is to partition the process matrix into two distinct parts: one repre-senting the �ows within the economic system, referred to as economic �ows, and onerepresenting the �ows from and into the environment, referred to as environmental�ows or interventions for short. With the ISO terminology these two groups arecalled product �ows and elementary �ows respectively. The partitioning leads to apartitioned matrix (Tab.5.1) :

P =

A−B

Although this partitioning is not strictly needed for the representation of process

unit or entire systems of process units, it is a convenient step which will turn outto be needed in the following procedure. The matrix A that represents the �owswithin the economic system will be referred to as the technology matrix; matrixB will be called the intervention matrix, because it represents the environmentalinterventions of process units.

58 Chapter 5. LCA resultsTable

5.1:Process

matrixforthebiore�neryconcept1.

Matrix

A:bluepart,Matrix

B:violetpart

CHP(Naturalgas)

MagnesiumSulphate

Phosphorus(white,liquid)

Sulfur

Nutrients(Mix)

TransportofNutrients

Woodchips

Transportofwoodchips

Yeast(fodder)

Transportofyeast

Enzymes(Cellulase)

Transportofenzymes

SodiumHydroxide

Transportofsodiumhydroxide

Lime(milled)

Transportoflime

Re�neryConstruction

SulfuricAcid

Transportofsulfuricacid

Wastewatertreatment

Transportation

BioEthanol

Biomethane

Lignin(H)

Lignin(O)

Energy

from

NaturalGas

[MJ]

10

00

00

00

00

00

00

00

00

00

0-1,64E

+05

-4,61E

+04

-2,72E

+04

-1,09E

+05

Magnesium

Sulphate[kg]

01

00

-39,67

00

00

00

00

00

00

00

00

00

00

Phosphorus[kg]

00

10

-39,67

00

00

00

00

00

00

00

00

00

00

Sulfur[kg]

00

01

-39,67

00

00

00

00

00

00

00

00

00

00

Nutrients

(P,S,M

g)[kg]

00

00

119

-10

00

00

00

00

00

00

00

00

00

TransportedNutrients

[kg]

00

00

01

00

00

00

00

00

00

00

0-6,34E

+01

-1,78E

+01

-1,05E

+01

-4,20E

+01

WoodChips[kg]

00

00

00

1-1

00

00

00

00

00

00

00

00

0

Transportedwoodchips[kg]

00

00

00

01

00

00

00

00

00

00

0-2,37E

+04

-6,66E

+03

-3,93E

+03

-1,57E

+04

Yeast

[kg]

00

00

00

00

1-1

00

00

00

00

00

00

00

0

Transportedyeast[kg]

00

00

00

00

01

00

00

00

00

00

0-1,37E

+02

-3,86E

+01

-2,28E

+01

-9,10E

+01

Enzymes

[kg]

00

00

00

00

00

1-1

00

00

00

00

00

00

0

Transportedenzymes

[kg]

00

00

00

00

00

01

00

00

00

00

0-6,59E

+01

-1,85E

+01

-1,09E

+01

-4,36E

+01

Sodium

Hydroxide[kg]

00

00

00

00

00

00

1-1

00

00

00

00

00

0

Transportedsodium

hydroxide[kg]

00

00

00

00

00

00

01

00

00

00

0-2,99E

+01

-8,40E

+00

-4,95E

+00

-1,98E

+01

Lim

e[kg]

00

00

00

00

00

00

00

1-1

00

00

00

00

0

Transportedlime[kg]

00

00

00

00

00

00

00

01

00

00

0-2,28E

+00

-6,40E

-01

-3,77E

-01

-1,51E

+00

Building[pcs]

00

00

00

00

00

00

00

00

10

00

0-1,98E

-06

-5,55E

-07

-3,27E

-07

-1,31E

-06

SulfuricAcid[kg]

00

00

00

00

00

00

00

00

01

-10

00

00

0

Transportedsulfuricacid

[kg]

00

00

00

00

00

00

00

00

00

10

0-2,23E

+02

-6,26E

+01

-3,69E

+01

-1,48E

+02

Treated

waste

Water

[m3]

00

00

00

00

00

00

00

00

00

01

0-1,87E

+02

-5,26E

+01

-3,10E

+01

-1,24E

+02

Transportation

[tkm]

00

00

0-0,15

0-0,15

0-0,15

0-0,15

0-0,15

0-0,2

00

-0,15

01

00

00

Bioethanol

[kg]

00

00

00

00

00

00

00

00

00

00

04,47E+03

00

0

Biomethane[kg]

00

00

00

00

00

00

00

00

00

00

00

8,72E+03

00

Lignin

(Hydrolysis)

[kg]

00

00

00

00

00

00

00

00

00

00

00

02,27E+04

0

Lignin

(Organosolv)[kg]

00

00

00

00

00

00

00

00

00

00

00

00

7,77E+03

CO2[kg]

0,06

0,27

8,85

0,24

00

-1,75

0-15,02

05,00

01,23

00,01

01,28E+07

0,10

075,02

0,16

7,35E+03

2,07E+03

1,22E+03

4,87E+03

N2O

[kg]

0,00

0,00

0,00

0,00

00

0,00

00,01

00

00,00

00,00

03,31E+02

0,00

00,00

0,00

00

00

CH4[kg]

0,00

0,00

0,04

0,00

00

0,00

00,24

00

00,00

00,00

04,62E+04

0,00

00,20

0,00

5,01E+02

1,41E+02

8,30E+01

3,32E+02

Coal,hard[kg]

0,00

0,06

3,80

0,02

00

0,00

01,35

00

00,39

00,00

04,77E+06

0,02

00,75

0,01

00

00

Coal,soft,lignite[kg]

0,00

0,07

1,91

0,00

00

0,00

00,23

00

00,08

00,00

06,97E+05

0,01

00,88

0,00

00

00

NaturalGas

[Nm3]

0,03

0,03

1,08

0,13

00

0,00

00,87

00

00,12

00,00

07,16E+05

0,03

00,61

0,00

00

00

Oil,crude[kg]

0,00

0,01

0,45

0,52

00

0,01

01,22

00

00,06

00,00

06,76E+05

0,11

00,22

0,05

00

00

LandUse

[m2*yr]

0,08

0,02

0,92

0,01

00

2,69

018,17

02,00

00,08

00,01

05,79E+06

0,01

01,26

0,01

00

00

Water

Use

[m3]

0,07

2,12

1,00

0,30

00

0,04

00,79

00

00,17

00,88

07,47E+07

0,48

00

0,09

6,09E+04

1,71E+04

1,01E+04

4,03E+04

Thevalues

are

expressed

onanhourlybasis


A fourth step is more related to the goal and scope de�nition rather than to theinventory analysis. It involves the speci�cation of the required performance of thesystem. In general, a reference �ow ϕ will be determined as one way of ful�lling afunctional unit that is quite arbitrarily chosen. For instance, a reference �ow for thebiore�nery concept could be a certain amount of �nal products. The vector f thusrepresents the set of economic �ows, even though only one (or some speci�c ones) ofthese �ows is the reference �ow. The rationale of using a coordinate system requiresthat we reserve an entry for every economic �ow. In general, the only non-zeroelement(s), say the rth, is the reference �ow:

f(i) =

{ϕ, if i = r

0, otherwise

Vector f will be referred to as the �nal (or external) demand vector, because it isan exougenously de�ned set of economic �ows of which we impose that the systemproduces exactly the given amount.

Lastly, we will proceed to de�ne:

g =

g1g2...gn

as a vector of environmental interventions, the inventory vector g.

5.2.2 The solution of the inventory problem

So far, we have only discussed the representation of process units, systems ofprocess units, reference �ows, and so on. We did not calculate anyhing yet. Inparticular, we did not yet discuss how to obtain the values of g1,g2,. . . , gn. Atreatment of this leads to a discussion of what we will call the inventory problem.

First, we introduce a vector with scaling factors, the scaling vector: we will in-dicate this vector by s:

s =

s1s2...sn


For the �rst economic �ow, energy from natural gas, a balance equation can beset up:

a11 × s1 + a12 × s2 + · · ·+ a1n × sn = f1

Then, another balance equation is available, for the second economic �ow, mag-nesium sulphate :

a21 × s1 + a22 × s2 + · · ·+ a2n × sn = f2

And so on for every economic �ow present in the system under study. A �nalstep towards a generally applicable treatment is in terms of matrix solution. thesystem of equations:

a11 × s1 + a12 × s2 + · · ·+ a1n × sn = f1

a21 × s1 + a22 × s2 + · · ·+ a2n × sn = f2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

an1 × s1 + an2 × s2 + · · ·+ ann × sn = fn

can be written as:

a11 a12 . . . a1na21 a22 . . . a2n...

.... . .

...an1 an2 . . . ann

s1s2...sn

=

f1f2...fn

or even more concisely as:

As = f (1)

Given that the technology matrix A is known and that the �nal demand vectorf is known, the balance equations can be solved, under certain restrictions, to yieldthe scaling vector s :

s = A−1f (2)


where A−1 denotes the inverse matrix of the technology matrix A.So, we have found a procedure to calculate the scaling vector for the unit pro-

cesses in a system,such that the system-wide aggregation of economic �ows exactlymatches with the �nal demand vector that represents th predetermined reference�ow of the system. However, the inventory problem has not been solved completely,because the values of the system-wide aggregated environmental �ows are still to befound.

The scaling vector provides a direct clue to the �nal step in solving the inventoryproblem: we must recognize that scaling of a unit process a�ects both the economic�ows and the environmental �ows. For the �rst environmental �ow, carbon dioxide,we have:

g1 = b11 × s1 + b12 × s2 + · · ·+ b1n × sn

More generally: g1 = b11 × s1 + b12 × s2 + · · ·+ b1n × sn

g2 = b21 × s1 + b22 × s2 + · · ·+ b1n × sn

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

gn = bn1 × s1 + bn2 × s2 + · · ·+ bnn × sn

or, in matrix notation:

g = Bs (3)

In principle, the inventory problem is now solved: there is a rule (2) that yieldsthe scaling vector given a technology matrix and �nal demand vector; and there isa second rule (3) that yields the inventory vector given the intervention matrix andthe scaling vector.

5.2.3 Solving multi-functionality and allocation

The basic model for solving the inventory problem was above presented in ageneralized form using matrix notation. The main idea has been the systematicconstruction of a set of linear balance equations, one for each economic �ow, with anumber of scaling factors, one for each process units.

Matrix inversion has been introduced as a way to solve such a system of linearequations, but this requires that the technology matrix is square and invertible.This is not automatically the case involving multifunctional process units, such asthe biore�nery concepts under study.


The approach toward obtaining an invertible technology matrix starts from therecognition that is the multifunctionality of a process units which causes the prob-lem in the �rst place. Hence, splitting a multifunctional process into a number ofindipendent monofunctional processes should provide a solution.

In the partitioning method, a multifunctional process unit ~ is divided into anumber of process units, such that each of the resulting process units is monofunc-tional. In the case of the biore�nery concept 1, four process units will be formed, onefor each biore�nery product: bioethanol, biomethane, hydrolysis lignin and organo-solv lignin (see also the last four columns in Tab 5.1).

MJof

energyfrom

naturalgas

kgof

magnesium

sulphate

kgof

phosph

orus

kgof

sulfur

kgof

nutrients

(P,S

,Mg)

kgof

transportednutrients

kgof

woodchips

kgof

transportedwoodchips

kgof

yeast

kgof

transportedyeast

kgof

enzy

mes

kgof

transporteden

zymes

kgof

sodium

hydroxide

kgof

transportedsodium

hydroxide

kgof

lime

kgof

transportedlime

pieces

ofrefinerybuilding

kgof

sulfuricacid

kgof

transportedsu

lfuricacid

m3of

treatedwastewater

tkm

oftransportation

kgof

bioethanol

kgof

biom

ethane

kghydrolysislignin

kgof

organosolvlignin

kgof

carbon

dioxide

kgof

dinitrogen

oxide

kgof

methane

kgof

hard

coalused

kgof

softcoal(lignite)

used

Nm

3of

naturalgasused

kgof

crudeoilused

m2peryearof

landused

m3of

waterused

−345860,8

0 0 0 0−133,8

0−50000

0−289,7

0−139

0 −63 0−4,8

−4,2E−

060−470

−394,85

04469,9

8722,6

22684,2

7770,6

15500

01056,624

0 0 0 0 0128461,3

=

−1,64E

+05

0 0 0 0−6,34E

+01

0−2,37E

+04

0−1,37E

+02

0−6,59E

+01

0−2,99E

+01

0−2,28E

+00

−1,98E−

060

−2,23E

+02

−1,87E

+02

04,47E

+03

0 0 07,35E

+03

05,01E

+02

0 0 0 0 06,09E

+04

+

−4,61E

+04

0 0 0 0−1,78E

+01

0−6,66E

+03

0−3,86E

+01

0−1,85E

+01

0−8,40E

+00

0−6,40E−

01−5,55E−

070

−6,26E

+01

−5,26E

+01

0 08,72E

+03

0 02,07E

+03

01,41E

+02

0 0 0 0 01,71E

+04

+

−2,72E

+04

0 0 0 0−1,05E

+01

0−3,93E

+03

0−2,28E

+01

0−1,09E

+01

0−4,95E

+00

0−3,77E−

01−3,27E−

070

−3,69E

+01

−3,10E

+01

0 0 02,27E

+04

01,22E

+03

08,30E

+01

0 0 0 0 01,01E

+04

+

−1,09E

+05

0 0 0 0−4,20E

+01

0−1,57E

+04

0−9,10E

+01

0−4,36E

+01

0−1,98E

+01

0−1,51E

+00

−1,31E−

060

−1,48E

+02

−1,24E

+02

0 0 0 07,77E

+03

4,87E

+03

03,32E

+02

0 0 0 0 04,03E

+04


Notice that only the values in the row corrispondent to the monofunctional prod-uct were left untouched and split in the four columns (4469,9; 8722,6; 22684,2;7770,6) while the values in the other rows were scaled according to the economicvalue of the �nal products (See Tab.B.2). The allocation factors are chosen such thatthey lie between 0 and 1; this is often referred to as the 100%-rule: the sum of thepartitioned monofunctional processes is equal to the unpartitioned multifunctionalprocess from which they are derived. A further simpli�cation is that the allocationfactors are almost always equal within one individual monofuncional process unit.

Table 5.2: Partitioning method

Product Economic Value[e/ kg]

Economic Value[e/ h]

AllocationFactors [%]

Bioethanol 0,85 7392,4 47,4

Biomethane 0,46 2077,7 13,3

Lignin (Hydrolysis) 0,05 1225 7,8

Lignin (Organosolv) 0,63 4895,5 31,4

Total 15590,5 100

5.2.4 Impact assessment

Once the inventory problem is solved, pertinent problems arise in impact assess-ment: which impact categories should be chosen, which category indicators providean environmentally relevant, yet feasible, representation of environmental impact,etc. The computational structure of impact assessment, however, is much easierthan that of the inventory analysis. The inventory analysis can be seen as a model,which includes all types of model complications: multifunctionality, and so on. Im-pact assessment, in contrast, is a much more mechanical procedure that uses theresults of all sorts of complicated models. In other words, the model content ofimpact assessment is much lower, because it is assumed that the characterizationmodels that are used in impact assessment cover such modeling technicalities.

In a methodological sense, the selection of impact categories is controversial;the choice that was made in this study was to consider one category for each envi-ronmental sphere, namely: air, soil, water and energy with particular focus on theclimate change (which will be also used in the optimization model).

From a mathematical perspective, however, it is easy. We can just assume thati categories have been chosen, and that associated with these impact categories, icategory indicators have been identi�ed. It is possible then to set up a linear spacewith a i -dimensional basis:

h =

climate change

agricultural land occupationwater depletion

abiotic depletion, fossil fuels

→kg CO2 eqm2 · yrm3

MJ


We will refer to h as the impact vector.

The subsequent step is represented by the characterization, which involves theconversion of the LCI (Life Cycle Impacts) results to common units and, possibly,the aggregation of the converted results within the impact categories.

Conversion is implied to mean multiplication; a general formula for characteri-zation is:

hi =∑j

(qi)j gj

or:

hi = qi g

where j is equal to the number of environmental interventions (e.g.: carbondioxide, methane, etc.) and the vectors qi represent the characterization vectors forthe i impact categories.

When the characterization vectors for several impact categories are justaposed,we can form the characterization matrix Q (see Tab.5.3):

Q =(q1|q2| . . . |qi

)

Table 5.3: Characterization matrix, Q

CO2[kg]

N2O

[kg]

CH4[kg]

Coal,hard[kg]


NaturalGas

[Nm3]

Oil,crude[kg]

LandUse

[m2*yr]

Water

Use

[m3]

climate change [kg CO2 eq] 1 298 25 0 0 0 0 0 0agricultural land occupation [m2*yr] 0 0 0 0 0 0 0 1 0water depletion [m3] 0 0 0 0 0 0 0 0 1abiotic depletion, fossil fuels [MJ] 0 0 0 27,91 13,96 38,84 41,87 0 0

The formula for the characterization into various impact categories assumes theform:

h = Qg (4)


5.2.5 Contribution analysis

When the discussion is con�ned to computational aspects, inventory results aresummarized in the inventory vector g. The formula for it was derived in the previoussection: by substituting equation (2) in equation (3) we have:

g = BA−1f (5)

which is a concise notation for:

∀k : gk =∑j

∑i

bkj(A−1)jifi

Here the sum over j represents the aggregation over all process units, and thesum over i the aggregation over all economic �ows that link these processes. In somecases it is useful to have partial aggrgation or no aggregation at all; such situationsinclude:

• to investigate the contribution of a process unit, e.g. production of bioethanol,to a particular intervention, e.g. the emission of CO2 ;

• to investigate the contribution of a set of process units, e.g. those belonging tothe material production phase, to a particular intervention, e.g. the emissionof CO2 .

Going beyond the inventory analysis, we proceed to the characterization. Theformula for the impact vector can be written as:

h = QBA−1f (6)

which stands for:

∀l : hl =∑k

∑j

∑i

qlkbkj(A−1)jifi

Here the sum over k represents the aggregation over all environmental interven-tions and qlk the characterization factor linking intervention k and impact categoryl .

5.2.6 Avoided Burdens

Until now we have been considering the processes involving the production of thebiore�nery feedstocks and the biore�nery process itself; in strict terms, then, it is a


case of cradle-to-gate LCA. This means that we follow the life cycle of the materials,from the raw matrials extraction and transformation to the �nal processing in thebiore�nery as marketable products ("gate").

To �nalize the procedure toward a more complete LCA (cradle-to-grave) a furtherstep must be undertaken. It is necessary, in fact, to model the end-of-life of suchmarketable products deriving form te biore�nery plant in order to include all theinterventions associated with these stages of a product's life cycle.

When this is not possible, or it may result too di�cult due to lack of data,etc. another route is viable. This is represented by the so called "avoided burdens"method. It consists in making an hypothesis on the product, based on current pro-duction technologies, which will be most probably displaced by a certain biore�neryproduct.

For example, considering the bioethanol produced in biore�nery concept 1, theassumption was to consider that it would be mixed with gasoline for use in veichles;some examples of commercially available ethanol blends are: E10, E15, E85, wherethe �gures represent the percentage of ethanol in the blend. In this way, the productdisplaced by the bioethanol would be petrol (or gasoline, the terms will be usedinterchangeably). All the assumptions made for the biore�nery concepts can beeasily extrapolated form the respective �owcharts, where the product displaced byeach marketable product is indicated (Fig.4.2, Fig.4.4, Fig.4.6 ).

Considering the biore�nery concept 1, the burdens associated with the displacedproducts1 are reported in Table 5.4:

Table 5.4: Avoided burdens for the biore�nery concept 1

petrol production,low-sulfur [1 kg]

natural gas, highpressure [1 m3]

pulverised ligniteproduction [1 MJ]

phenolproduction [1 kg]

CO2 [kg] 1,091 0,414 0,300 4,189N2O [kg] 0,104 0,008 0,008 0,440CH4 [kg] 0,524 0,477 0,767 0,365Coal, hard [kg] 0,046 0,004 0,004 1,237Coal, soft, lignite [kg] 0,023 0,004 0,121 0,211Natural Gas [Nm3] 0,071 1,084 0,606 0,950Oil, crude [kg] 1,219 0,001 0,000 1,158Land Use [m2*yr] 3,204 0,609 1,823 1,635Water Use [m3] 0,934 0,702 0,716 8,687

Source: ecoinvent 3.2

5.2.7 Results

Below are reported, in form of graphs and tables, the results for the biore�neryconcept 1 using the methodology described in the previous sections.

1Notice that the production process of lignite is expressed in the database as MJ while theproduced lignin form the biore�nery is expressed as kg; in this case a conversion factor is neededand it is represented by the Lower Heating Value (LHV) of lignin, indicated in the database:21, 7 MJ/kg. A similar consideration has been made for the natural gas production, where thevalue of the density equal to 0, 8 kg/m3 was used in order to perform the conversion


5.2.7.1 Demand and scaling vectors

Table 5.5: Final demand vector, f and scaling vector s

fFinal

dem

andvector

sScalingvector

Energy from Natural Gas [MJ] 0 345860,792

Magnesium Sulphate [kg] 0 44,5992504

Phosphorus [kg] 0 44,5992504

Sulfur [kg] 0 44,5992504

Nutrients (P,S,Mg) [kg] 0 1,12436975

Transported Nutrients [kg] 0 133,8

Wood Chips [kg] 0 50000

Transported wood chips [kg] 0 50000

Yeast [kg] 0 289,7

Transported yeast [kg] 0 289,7

Enzymes [kg] 0 139

Transported enzymes [kg] 0 139

Sodium Hydroxide [kg] 0 63

Transported sodium hydroxide [kg] 0 63

Lime [kg] 0 4,8

Transported lime [kg] 0 4,8

Building [pcs] 0 4,1667E-06

Sulfuric Acid [kg] 0 470

Transported sulfuric acid [kg] 0 470

Treated waste Water [m3] 0 394,85

Transportation [tkm] 0 7665,045

Bioethanol [kg] 4469,92195 1

Biomethane [kg] 8722,6 1

Lignin (Hydrolysis) [kg] 22684,2 1

Lignin (Organosolv) [kg] 7770,6 1

The non-zero values in the �nal demand vector (f) were chosen to match exactlythe biore�nery concept 1 production, in order to be able to consider as a reference theinput of beech wood: it is as if we consider the entire utilization and transformationof the amount of wood treated by the biore�nery.

As a proof that the above presented model is consistent, it is possible to noticethat in the scaling vector (s), the values corresponding to the reference �ows areall equal to 1: this means that no scaling is needed for this �ows, since these arealready the desired quantities of the products.


5.2.7.2 Emissions and impacts

Table 5.6: Environmental interventions vector, g

Inventory of Emissios, g [-/hour]

CO2 [kg] -22496,4

N2O [kg] 4,4

CH4 [kg] 1276,8

Coal, hard [kg] 1527,4

Coal, soft, lignite [kg] 835,9

Natural Gas [Nm3] 10884,8

Oil, crude [kg] 1658,1

Land Use [m2*yr] 167962,9

Water Use [m3] 156934

Note that the value for the carbon dioxide emissions is negative; tis means thatthere is a net negative balance due to the carbon dioxide absorbed by the woodtrees during the growth. It is important to stress, however, that the climate changeimpact category (expressed in kg of CO2eq.) is a weighted sum of CO2, N2O andCH4 which, as shown in Tab.5.7 has an overall positive value.

Table 5.7: Impact vector h

Impact vector, h [-/year] Impact vector, h [-/kg wood]

climate change [kg CO2 eq] 86067510 0,21

agricultural land occupation [m2*yr] 1343703529 3,35

water depletion [m3] 1255472771 3,13

abiotic depletion, fossil fuels [MJ] 4371991227 10,9

5.2.7.3 Avoided emissions

The previous values in the impact vector h should be updated with the avoidedemissions deriving from the displaced products following the avoided burden method-ology (see Tab.5.8).

Table 5.8: Avoided impacts h'

Reference impactvector, h' [-/hour]

Reference impactvector, h' [-/year]

Reference impact vec-tor, h' [-/kg wood]

climate change[kg CO2 eq]

1474587 11796697314 29,5

agricultural landoccupation [m2*yr]

34248 273985306 0,68

water depletion[m3]

78543 628342429 1,57

abiotic depletion,fossil fuels [MJ]

1597652 12781216708 31,95


The following step is needed to �nd the new values of the impact categories, netof the avoided emissions:

h∗ = h− h′

Which gives the following results:

Table 5.9: Normalized impacts h*

Normalized impacts, h* [-/year] Normalized impacts, h* [-/kg wood]

climate change [kg CO2 eq] -11710629804 -29,27



abiotic depletion, fossil fuels [MJ] -8409225481 -21,02

It is interesting at this point to have a graphical comparison concerning the im-pacts deriving from the production of the same amount of �nal products, both withthe biore�nery and with the current production processes:

Figure 5.1: Comparison between a biore�nery and BAU production (Business As Usual)

It is possible to see that the biore�nery system has much lower impacts in termsof climate change and abiotic depletion while greater impacts for what concernsagricultural land occupation and water depletion. This is due to the land neededfor the wood growth and the intensive water use in the biore�nery.

5.2.7.4 Contribution analysis

The following charts show the contribution of each single process to the envi-ronmental interventions of the overall system; resuts are showed both aggregatedfor each impact category group of interventions (Fig.5.2) and in terms of relativeamount (percentage of the total) (Fig.5.3) .


(a)

(b)

(c)

(d)

Figure

5.2:Analysisof

thecontribution

ofeach

process

totheenvironmentalinterventionsg(per

hour)


Figure 5.3: Relative contribution of each process to the total environmental interventions

The results show that the wood chips production process is the one which con-tributes to the highest extent in carbon dioxide absorption, while the main sourcesof emission in this sense are represented by the wastewater treatment and the com-bined heat and power (CHP) plant. The latter is also responsible for the highestfossil resources utilization, in particular of natural gas which is by far greater thanthe one deriving from the crude oil used for transportation and for the harvestingand processing of the wood.

For what concerns the land use, it can be seen that the main contribution, asexpected, is associated with the wood chips production process; the second highestcontribution comes from the CHP plant due to the high amount of natural gasneeded for the energy consumption of the biore�nery, which entails a non negligibleamount of land resource for the extraction and the distribution infrastructure of thenatural gas.

In regards of the water use, the major contribution is caused by the biore�n-ery itself, where the amounts corresponding to each product are the result of theaformentioned allocation procedure.

The same procedure can be also applied to the impact categories and the resultscan be seen on the following Fig.5.4 and Fig.5.5:


(a)

(b)

(c)

(d)

Figure

5.4:Analysisof

thecontribution

ofeach

process

totheim

pactcategories

g(per

hour)


Figure 5.5: Relative contribution of each process to the total impact categories

As seen from the results of the environmental interventions contributions, similarconsiderations can be made also concerning the impact categories.

It is possible to see that, as expected from the previous contributions in terms ofenvironmental interventions, the biggest share of fossil fuel depletion is due to theCHP plant. The wood chips production on one hand leads to a very high agriculturalland occupation on the other other hand it grants huge negative impacts (savings) interms of climate change, which are counterbalanced by the emissions deriving fromthe CHP plant and the biore�nery itself (whose impacts are split into the di�erentshares of the products).

Chapter 6

Results of the optimization model

6.1 Problem description

A major complication for life cycle assessment in general is the multifunctionalproblem. This condition arises if a process provides more than one output productand this case the environmental burdens associated with the process have to beallocated over various products. Several procedures have been developed to solve themultifunctional problem: consequential LCA approaches prefer to use substitutionmethod to deal with that problem whereas attributional LCA entails the use of thepartitioning method.

A further limitation of standard LCA might be the fact that no optimal produc-tion mixes can be delivered. Particularly in consequential LCA this could representa problem. In certain cases not only the consequences of a potential decision needto be described, but rather the optimal decision itself shall be determined.

Many research e�orts [58] aim at an extension of life-cycle assessment (LCA)in order to increase its spatial or temporal detail or to enlarge its scope but fromthe application standpoint other options are available to obtain more detailed infor-mation in a life-cycle perspective. In literature we can mainly �nd three di�erentstrategies to reach these aims:

1) extension of LCA: one consistent model;

2) use of a tool-box: separate models used in combination;

3) hybrid analysis: combination of models with data �ows between them.

Extension of LCA o�ers the most consistent solution. Developments in LCA aremoving toward greater spatial detail and temporal resolution and the inclusion ofsocial issues. Creating a super tool with too many data and resource requirementsis, however, a risk. Moreover, a number of social issues are not easily modeled inrelation to a functional unit.

The development of a toolbox o�ers the most �exibility regarding spatial andtemporal information and regarding the inclusion of other types of impacts. Therigid structure of LCA no longer sets limits; every aspect can be dealt with according

75

76 Chapter 6. Results of the optimization model

to the logic of the relevant tool. The results lack consistency, however, preventingfurther formal integration.

The third strategy, hybrid analysis, takes up an intermediate position betweenthe other two. This strategy is more �exible than extension of LCA and moreconsistent than a toolbox. Hybrid analysis thus has the potential to combine thestrong points of the other two strategies. It o�ers an interesting path for furtherdiscovery, broader than the already well-known combination of process-LCA andinput-output-LCA.

In this chapter we will demonstrate how LCA can be extended by the use ofmixed integer linear programming :

i) to determine the optimal production mix of new technologies and

ii) to deal with the multifunctional problem.

A mixed integer linear program was implemented in Matlab R©and applied toa case study in which three wood-based biore�nery options (Biore�nery concepts1, 2 and 3) are analyzed in terms of optimal consequences for the total systemunder study. In particular three construction areas were identi�ed as possible siteswhere the wood availability was high enough to guarantee the construction of suchbiore�neries (Fig. 6.1) [43]. The model evaluates the best mix of biore�nery thatcould be constructed and in which area.

Figure 6.1: Most promising german regions in terms of beech wood potential

6.2 Methods

The computational structure of LCA has been comprehensively described inchapter 5 and will be here used again in order to build a new process matrix. Math-ematically, the �rst step of life cycle inventory (LCI) analysis is the determinationof the scaling vector s. The factors of this vector scale processes of a system upor down such that the output of unit processes exactly matches the �nal demand.

6.2. Methods 77

The �nal demand vector includes the reference �ows of the functional unit, whichis de�ned within the goal and scope of a LCA study.

in Fig. 6.2 (left side) and Fig. 6.3 (right side) is showed the obtained processmatrix. The matrix has to be seen as a unique one which, due to layout issues, hadto be split on two pages.

The rows represent the materials which are �owing in and out of the systemwhile the columns represent the single production processes and are split into threegroups: alternative processes, current processes and additional processes.

The alternative processes embody the new technology system under study andare represented by the three biore�nery concepts decribed in chapter 4. There is afurther subdivision into three construction areas (CA1, CA2, CA3) to identify thethree di�erent options in terms of possible construction sites which were identi�edin a previous study.

The current production processes are the one relative to the actual productionof the required goods in the �nal demand, without relying on the new technologyavailable which is represented by the biore�nery systems. These data were takenfrom a database (ecoinvent Version 3.2 ) and it has to be said that a strong assump-tion was made here: by considering the values from the database it was assumedthat the mix of all the processes for the production of a certain material was con-sidered to be the one implicit in the database, which, in some cases may not re�ectthe actual situation of a speci�c country. With that said, it is remarked that theprocesses were chosen in order to be as most consistent as possible and the data referto one single country (in this case Germany, ecoinvent code: "DE") rather than anaveraged value over Europe or world production.

The third group, the additional processes, represents all the processes whichare needed in order to guarantee the operation of the biore�nery systems; they areneeded to take into account the upstream activities of the biore�nery process andthe respective emissions.

For what concerns the �nal demand vector, f, some clari�cations should be made.Most of the values assumed at this point were taken from on-line available databasesand some of the quantities needed to be converted into the appropriate units ofmeasurement, which in this case are kg/year. In particular, the amount of the petrolproduction was converted from barrels/day through the value of 119, 07 kg/barrel[1]; the natural gas density was supposed to be 0, 8 kg/m3 [ecoinvent 3.2 ]; forthe gypsum, a density value of 2, 3 g/cm3 [2] was needed for the conversion fromvolume to mass annual production. The amount of lactic acid was the trickiest toevaluate due to the lack of data available. The problem was tackled by makingsome hypotheses on the production volumes of bio-plastics and the shares of thisproduction according to Europe and material type [24]. Speci�cally, a world totalbio-plastics production of 2, 03million tonnes was divided into European share (13%)and subsequently into the share related to PLA (11%). Taking also into account thefact that the percentage of lactic acid utilized for bio-plastics is about 30% (whilethe remaining part is used in the food and pharmaceutical industries) and that theratio of lactic acid (after the fermentation) to PLA (after the polymerization) isnearly 1, 25, the �nal value for the lactic acid demand was estimated.


The characterization matrix, Q, in this case is represented by a single columnreporting the values of the CO2 equivalent associated with each chemical species.

6.3 Mixed Integer Linear Programming (MILP)

It has been implemented a mixed integer linear program to assess multiple newtechnology options with regard to their environmental advantages. Thereby, thenew technology options compete with currently available processes to produce theircorresponding products. The target of the program is to �nd an optimal scalingvector s that is able to minimize the impact indicator result h:

Minimize h = QBs (7)

Subject to:

As =

{≥ fi = ϕ, if i = r,

= fi = 0, otherwise

and:

lb ≤ s ≤ ub

The scaling factors of the new technologies shall be integers, due to the factthat planned technology concepts are usually not linearly upscaled or downscaled.The �rst constraint is synonymous to the LCI problem (1). To deal with the multi-functionality of processes, the equations can be relaxed into inequations. Equationconstraints persist for �ows that are only used as pre-products ("Additional pro-cesses") and hence do not contribute to the �nal demand. Vectors of lower boundslb and upper bounds ub have to be identi�ed. These bounds set the limits of theprocesses in which the optimization is carried out: in particular they concern thewood availability in a certain region as shown in Tab. 6.2 and hence the amount ofpossible biore�nery plants that could be built in a certain area.

Table 6.1: Characteristics of the regions under study

Region Wood availability[Mio. t](Scenario 1)

Wood availability[Mio. t](Scenario 2)

Averagedistance[km]

Maximum number ofbiore�neries(Scenario 1)

Maximum number ofbiore�neries(Scenario 2)

Region Nordhessen(CA1)

1,146 0,452 350 1,4 0,6

Region LändereckBayern/Hessen/Thüringen(CA2)

1,344 0,541 250 1,7 0,7

RegionBaden-Württemberg(CA3)

0,887 0,283 150 1,1 0,4

Total 3,377 1,276

Adapted from [43]

6.3. Mixed Integer Linear Programming (MILP) 79

Figure

6.2:Process

Matrixfortheoptimizationmodel

Leftside


Figure

6.3:Process

Matrixfortheoptimizationmodel

Rightside


The upper bounds were evaluated considering the wood availability, the amountof wood which is needed as feedstock to the biore�nery each year (400000 tonnes,dry; w = 50%); calculating the ratio between these two quantities, approximatingto the nearest integer, it was possible to estimate the upper limits for the biore�neryinstallations in each area.

Two di�erent scenarios were considered (Scenario 1, Scenario 2): the �rst onerefers to the total availability of wood made up of forest wood, wood deriving fromindustrial processes and sawing products; in the second case the amount of forestwood was reduced to half of its amount.

The problem 7, with the data considered in Fig.6.2 and Fig.6.3 has been imple-mented in a Matlab R© code and the the results yielded are presented in the followingtable:

Table 6.2: Results from the optimization model: GWP savings

Scenario 1 Scenario 2

Total biore�neriesbuilt

7 2

h current[kg CO2 eq]

4,97E+12 4,97E+12

h consequential[kg CO2 eq]

4,53E+12 4,83E+12

h di�erence[kg CO2 eq]

4,47E+11 1,48E+11

GWP saving[Mton CO2 eq]

447 148

So the results indicate that it would be possible to built seven and two biore�ner-ies for the scenario 1 and 2 respectively. The di�erence lies in the wood availabilitywhich is not the same for the di�erent scenarios.

As a consequence, a proportional amount of products belonging to the �naldemand vector will be displaced by the alternative processes (biore�neries) leadingto a reduction in the overall emissions when compared to the base case consideringonly current processes for the productions of materials.

More speci�cally, the current production processes to meet the �nal demandof the considered products entails the emissions of about 4, 98 · 1012 kg of carbondioxide equivalent (this value, of course, is the same for both the scenarios). Thisvalue is reduced to 4, 53 · 1012 and 4, 83 · 1012 kg of carbon dioxide equivalent for thescenario 1 and 2 respectively, due to the e�ects of the new processes. This results ina di�erence in the emissions (avoided emissions or saving) equal to about 447 and148 Mton of carbon dioxide equivalent, respectively.


To have a more detailed look into the model, the scaling (s) vector for both thescenarios is reported below :

s =

BR1(CA1)BR2(CA1)BR3(CA1)BR1(CA2)BR2(CA2)BR3(CA2)BR1(CA3)BR2(CA3)BR3(CA3)

Petrol productionEthylene productionLactic Acid productionNatural gas production

Lignite productionPhenol production

Gypsum(stucco) productionHeat from CHPWood productionWood productionWood production

Transportation of woodNutrients (P, S,Mg) production

Y east productionEnzymes production

SodiumHydroxide productionLime production

Sulfuric Acid productionWaste Water Treatment

Refrigerant R134− a productionEthanol productionMethanol productionAmmonia production

=

Scenario 1110201110

1, 81E + 093, 36E + 090, 00E + 001, 50E + 103, 58E + 098, 85E + 084, 39E + 091, 82E + 108, 00E + 081, 20E + 098, 00E + 081, 40E + 097, 49E + 061, 62E + 077, 77E + 063, 02E + 065, 06E + 071, 60E + 082, 15E + 071, 48E + 052, 16E + 071, 13E + 064, 00E + 05

;

Scenario 2100001000

2, 02E + 093, 44E + 091, 49E − 081, 52E + 104, 49E + 091, 20E + 094, 39E + 094, 96E + 094, 00E + 084, 00E + 080, 00E + 004, 80E + 082, 14E + 064, 64E + 062, 22E + 065, 04E + 055, 04E + 071, 41E + 086, 34E + 064, 94E + 042, 16E + 071, 13E + 064, 00E + 05

Furthermore, the discrepancy vector (d) was evaluated, which represents the dif-ference between the �nal supply vector of the optimized system and the intended�nal demand vector:


d =

Process heat [MJ ]Wood form area 1 [tons]Wood form area 2 [tons]Wood form area 3 [tons]Transport of wood [tkm]Nutrients (P, S,Mg) [kg]

Y east [kg]Enzymes [kg]

SodiumHydroxide [kg]Lime [kg]

Sulfuric Acid [kg]Water [m3]

Refrigerant R134− a [kg]Ethanol [kg]Methanol [kg]Ammonia [kg]

Ethanol / Petrol [kg]Ethylene [kg]Lactic Acid [kg]

Biomethane / Natural gas [kg]Lignin(H) / Lignite [kg]Lignin(O)/ Phenol [kg]

Calcium Sulfate / Gypsum [kg]

=

Scenario 100000000000000000000000

;

Scenario 200000000000000000000000

It can be noticed that the �nal demand is met without neither an overproduction

nor a shortage.The case study demonstrates the possibility of determining optimal production

mixes of newly developed technologies. The mixed integer linear program can solvethe multifunctional problem without using the partitioning method. Solving themultifunctional problem for new technologies can be interpreted as a substitutionapproach where the substitution refers not to products but rather to processes,which can be multi-functional. When minimizing the total environmental impactsof the system under study larger amounts of supplied products are allowed comparedwith those in the �nal demand. It is crucial to take this into consideration, sincemore products and hence more functions are possible in the �nal supply vector ofthe optimized system. On the other hand, if products of current technologies arenot provided by the new technologies, the program would ensure that at least theamounts of the current system are generated.

The de�nition of the current system with processes and its corresponding produc-tion volumes is crucial. Various LCI databases provide data supporting to generatethe A and B matrices while the amounts of the �nal demand vector can be collectedfrom statistical agencies.

Appropriate current processes with its corresponding product volumes must bedetermined to de�ne the �nal demand vector. However, the choice for multifunc-


tional processes that can be substituted by new technologies seems closer to realitythan choosing processes and declaring it as mono-functional.

As in any study, the research question de�nes the system under study. Theprogram addresses the question of whether new technology options are more advan-tageously to provide products than current processes.

The result of the program can be used to support decision-making by identifyingthe system's highest potential for optimization from a set of prede�ned alternatives.

Chapter 7

Conclusions and outlook

Is it useful to review what was done during the analysis of the case study and tosummarize the results obtained:

Table 7.1: Summary of the results of the LCA calculations

BR1 BR2 BR3Ethanol Ethylene Lactic Acid

Impact vector, h [-/year]

climate change [kg CO2 eq] 86067510 40758594 27023558agricultural land occupation [m2*yr] 1343703529 1312863162 1354785705water depletion [m3] 1255472771 1557489641 1312447914abiotic depletion, fossil fuels [MJ] 4371991227 3989364167 5023075748

Impact vector, h [-/kg wood]

climate change [kg CO2 eq] 0,21 0,1 0,07agricultural land occupation [m2*yr] 3,36 3,28 3,4water depletion [m3] 3,14 3,89 3,28abiotic depletion, fossil fuels [MJ] 10,93 9,97 12,6

Normalized impacts, h* [-/year]

climate change [kg CO2 eq] -11710629804 -10534045600 -28892429854agricultural land occupation [m2*yr] 1069718222 1117562532 1036707109water depletion [m3] 627130341,6 958360562 -344635821abiotic depletion, fossil fuels [MJ] -8409225481 -9137506335 -13685575032

Reference impact vector, h' [-/kg wood]

climate change [kg CO2 eq] -29,49 -26,43 -72,29agricultural land occupation [m2*yr] -0,68 -0,48 -0,79water depletion [m3] -1,57 -1,49 -4,14abiotic depletion, fossil fuels [MJ] -31,95 -32,81 -46,77

Normalized impacts, h* [-/kg wood]

climate change [kg CO2 eq] -29,27 -26,33 -72,23agricultural land occupation [m2*yr] 2,67 2,79 2,59water depletion [m3] 1,57 2,39 -0,86abiotic depletion, fossil fuels [MJ] -21,02 -22,84 -34,21

In the �rst place, an attributional LCA was carried out, in order to �nd out whatare the impacts associated with the biore�neries, seen as a unit process treating acommon input: the same amount of feedstock, beech wood (Fig. 7.1).

85

86 Conclusions and outlook

Figure 7.1: Impacts deriving from the treatment of 1 kg of beech wood

This approach is mostly indicated when a single process is under study and thefocus is more on the output of the process itself; for example if it was to analyze justone biore�nery concept and we wanted to evaluate what are the emissions related toa certain amount of main output (e.g. 1 kg of bioethanol). In our case, this resultsare hard to compare due to the di�erence in the products portfolio of each concept,therefore a further step was undertaken.

A consequential LCA was then performed, considering the so-called avoided bur-dens for each concept, to being able to compare the results between the di�erentbiore�neries (Fig. 7.2). The net results are reported in Fig.7.3 and it is possibleto see that the biore�nery concept 3 (BR3) is the one that guarantees the greatestvalue of savings for all four the impact categories.

Conclusions and outlook 87

Figure 7.2: Impacts deriving from the biore�nery systems (blue) compared to the avoidedimpacts (red))

AD: abiotic depletion, fossil fuels [MJ]; WD: water depletion [m3] ;ALO:agricultural landoccupation [m2 ∗ yr]; CC:climate change [kg CO2 eq]

Figure 7.3: Normalized impacts considering the avoided emissions deriving from the dis-placement of �nal products equivalent

88 Conclusions and outlook

An alternative way to perform the analysis was used then to integrate both anLCA methodology with a decision-making process based on the criterion of minimumenvironmental impacts.

This thesis work showed that mixed integer linear programming (MILP) can beused to extend standard LCA. Additional research questions can be addressed suchas the determination of the optimal number and location of new production plants.Thereby, the optimal decision is determined in terms of minimizing future potentialenvironmental impacts for the total system under study.

Further improvements to the model are possible. In principle it is possibleto modify the program according to other assumptions. For instance, maximumamounts for products that are additionally provided by the new technologies can beintroduced. Or, the �nal demand for products that are not generated by the newtechnologies can be reduced.

To take into account additional impact categories the Q matrix of the targetfunction can be easily modi�ed by corresponding characterization factors. Further-more, normalization and weighting can be introduced within the target function.Thus optimization can be carried out in terms of a single score, which takes intoaccount various normalized and weighted impact indicator results.

Appendix A

Results for the Bioethylene concept

Figure A.1: Comparison between a biore�nery and BAU production (Business As Usual)

89

90 Appendix A. Results for the Bioethylene concept

Table

A.1:Process


Matrix

A:bluepart,Matrix

B:violetpart

CHP(Naturalgas)

Industrialfurnace(Naturalgas,>100kW)

MagnesiumSulphate


Sulfur

Nutrients(Mix)


Woodchips


Yeast(fodder)

Transportofyeast

Enzymes(Cellulase)

Transportofenzymes

SodiumHydroxide

Transportofsodiumhydroxide

Lime(milled)[1kg]

Transportoflime

RefrigerantR134-a

TransportofrefrigerantR134-a


SulfuricAcid


Wastewatertreatment

Transportation

BioEthylene

Biomethane

Lignin(H)

Lignin(O)

Energy

from

NaturalGas

[MJ]

10

00

00

00

00

00

00

00

00

00

00

00

-1E+05

-41141

-24221,25

-96800

Heatfrom

NaturalGas

[MJ]

01

00

00

00

00

00

00

00

00

00

00

00

-4167

-1264,3

-744,3281

-2974,7

Magnesium

Sulphate[kg]

00

10

0-39,666

00

00

00

00

00

00

00

00

00

00

00

Phosphorus[kg]

00

01

0-39,666

00

00

00

00

00

00

00

00

00

00

00

Sulfur[kg]

00

00

1-39,666

00

00

00

00

00

00

00

00

00

00

00

Nutrients

(P,S,M

g)[kg]

00

00

0119

-10

00

00

00

00

00

00

00

00

00

00


[kg]

00

00

00

10

00

00

00

00

00

00

00

00

-60,93

-18,488

-10,8843

-43,499

WoodChips[kg]

00

00

00

01

-10

00

00

00

00

00

00

00

00

00


00

00

00

00

10

00

00

00

00

00

00

00

-22769

-6908,7

-4067,377

-16255

Yeast

[kg]

00

00

00

00

01

-10

00

00

00

00

00

00

00

00


00

00

00

00

00

10

00

00

00

00

00

00

-131,9

-40,029

-23,56638

-94,182

Enzymes

[kg]

00

00

00

00

00

01

-10

00

00

00

00

00

00

00

Transportedenzymes

[kg]

00

00

00

00

00

00

10

00

00

00

00

00

-63,3

-19,206

-11,30731

-45,189

Sodium

Hydroxide[kg]

00

00

00

00

00

00

01

-10

00

00

00

00

00

00

Transportedsodium

hydroxide[kg]

00

00

00

00

00

00

00

10

00

00

00

00

-28,69

-8,705

-5,124895

-20,482

Lim

e[kg]

00

00

00

00

00

00

00

01

-10

00

00

00

00

00

Transportedlime[kg]

00

00

00

00

00

00

00

00

10

00

00

00

-2,186

-0,6632

-0,390468

-1,5605

Refrigerant,R134-a[kg]

00

00

00

00

00

00

00

00

01

-10

00

00

00

00

Transportedrefrigerant,R134-a[kg]

00

00

00

00

00

00

00

00

00

10

00

00

-2,812

-0,8533

-0,502337

-2,0076

Building[pcs]

00

00

00

00

00

00

00

00

00

01

00

00

-2E-06

-6E-07

-3,39E

-07

-1E-06

SulfuricAcid[kg]

00

00

00

00

00

00

00

00

00

00

1-1

00

00

00


[kg]

00

00

00

00

00

00

00

00

00

00

01

00

-214

-64,942

-38,23334

-152,8

Treated

waste

water

[m3]

00

00

00

00

00

00

00

00

00

00

00

10

-161,8

-49,091

-28,90115

-115,5

Transportation

[tkm]

00

00

00

-0,15

0-0,15

0-0,15

0-0,15

0-0,15

0-0,15

0-0,15

00

-0,15

01

00

00

Bioethylene[kg]

00

00

00

00

00

00

00

00

00

00

00

00

4469,9

00

0

Biomethane[kg]

00

00

00

00

00

00

00

00

00

00

00

00

08722,6

00

Lignin

(Hydrolysis)

[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

22684,2

0

Lignin

(Organosolv)[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

07770,6

CO2[kg]

0,0632237

0,0678823

0,269419

8,848

0,241833286

00

-1,753548

0-15,0167

05

01,2281

00,00686

06,68346

012836657

0,0979

075,019

0,1636

7058,3

2141,7

1260,8869

5039,1

N2O

[kg]

9,18E-07

5,646E

-07

1,08E-05

3E-04

4,92021E

-06

00

1,97E-06

00,01337

00

05E

-05

09,6E

-07

00,00018

0331,02051

3E-06

00,0003

3E-06

00

00

CH4[kg]

0,0001871

0,0002009

0,000681

0,037

0,00105325

00

5,939E

-05

00,23964

00

00,0036

01,2E

-05

00,02185

046193,637

0,0004

00,2015

0,0002

482,52

146,41

86,196181

344,48

Coal,hard[kg]

0,0012974

0,001393

0,062996

3,798

0,019037171

00

0,0021735

01,34957

00

00,3877

00,00078

01,41805

04765163,7

0,0167

00,7514

0,0075

00

00


0,0008481

0,0009106

0,072482

1,909

0,003854215

00

0,0003815

00,23453

00

00,0837

00,00049

00,47419

0696650,01

0,0052

00,8828

0,001

00

00

NaturalGas

[Nm3]

0,0295338

0,03171

0,033762

1,085

0,126829752

00

0,0012321

00,86697

00

00,1221

00,00043

01,64497

0715970,59

0,0317

00,6098

0,0046

00

00

Oil,crude[kg]

0,0002684

0,0002882

0,007141

0,448

0,52421293

00

0,0124978

01,21578

00

00,0611

00,00087

01,15435

0676045,52

0,1147

00,2211

0,0518

00

00

LandUse

[m2*yr]

0,0791712

0,00021

0,017016

0,916

0,00658292

00

2,6875387

018,167

02

00,0807

00,00611

00,31526

05785028,3

0,0146

01,2623

0,0113

00

00

Water

Use

[m3]

0,0722614

5,863E

-05

2,117797

10,299263652

00

0,0366644

00,78777

00

00,1667

00,87739

00,04146

085396,526

0,4817

00

0,0932

77414

23490

13829,082

55268

Thevalues

are

expressed

onanhourlybasis

91

Table A.2: Partitioning method




Bioethylene 1,33 6857,14 45,53

Biomethane 0,46 2080,66 13,82

Lignin (Hydrolysis) 0,05 1224,95 8,13


Total 15058,23 100

Table A.3: Avoided burdens for the biore�nery concept 2

Ethilene producion, average [1 kg] natural gas, high pressure [1 m3] pulverised lignite production [1 MJ] phenol production [1 kg]CO2 [kg] 1,68 0,41 0,3 4,19N2O [kg] 0,002 0,001 0,007 0,44CH4 [kg] 0,345 0,47 0,76 0,36Coal, hard [kg] 0,04 0,004 0,004 1,23Coal, soft, lignite [kg] 0,0001 0,003 0,12 0,21Natural Gas [Nm3] 0,64 1,08 0,6 0,95Oil, crude [kg] 0,92 0,001 0,0002 1,15Land Use [m2*yr] 1,003 0,61 1,82 1,63Water Use [m3] 0,11 0,7 0,71 8,68



Table A.4: Final demand vector, f and scaling vector s

fFinal

dem

andvector

sScalingvector


Heat from Natural Gas [MJ] 0 9149,976



Sulfur [kg] 0 44,59925





Yeast [kg] 0 289,7


Enzymes [kg] 0 139


Sodium Hydroxide [kg] 0 63

Transported sodium hydroxide [kg] 0 63

Lime [kg] 0 4,8


Refrigerant, R134-a [kg] 0 6,1751955

Transported refrigerant, R134-a [kg] 0 6,1751955




Treated waste water [m3] 0 355,28


Bioethylene [kg] 4469,92195 1


Lignin (Hydrolysis) [kg] 22684,2 1

Lignin (Organosolv) [kg] 7770,6 1

Table A.5: Environmental interventions vector, g


CO2 [kg] -27844

N2O [kg] 4,42

CH4 [kg] 1264

Coal, hard [kg] 1456

Coal, soft, lignite [kg] 771

Natural Gas [Nm3] 9740

Oil, crude [kg] 1646

Land Use [m2*yr] 164107


93

Table A.6: Impact vector h






Table A.7: Avoided impacts h'



Reference impactvector, h' [-/kg wood]


1321850 10574804194 26,43


24412 195300629 0,48

water depletion [m3] 74891 599129077 1,49abiotic depletion,fossil fuels [MJ]

1640858 13126870501 32,81

Table A.8: Normalized impacts h*


h*=h-h'

-10534045599 -26,33

1117562532 2,79

958360562 2,39

-9137506334 -22,8


(a)

(b)

(c)

(d)

Figure

A.2:Analysisof

thecontribution

ofeach

process


hour)

95

Figure

A.3:Relativecontribution

ofeach

process

tothetotalenvironmentalinterventions


(a)

(b)

(c)

(d)

Figure

A.4:Analysisof

thecontribution

ofeach

process

totheim

pactcategories

g(per

hour)

97

Figure

A.5:Relativecontribution

ofeach

process

tothetotalim

pactcategories

Appendix B

Results for the Lactic Acid concept

Figure B.1: Comparison between a biore�nery and BAU production (Business As Usual)

99

100 Appendix B. Results for the Lactic Acid conceptTable

B.1:Process


Matrix

A:bluepart,Matrix

B:violetpart

-

CHP(Naturalgas)

MagnesiumSulphate


Sulfur

Nutrients(Mix)


Woodchips


Yeast(fodder)

Transportofyeast

Enzymes(Cellulase)

Transportofenzymes

Lime(milled)

Transportoflime

RefrigerantR134-a

TransportofrefrigerantR134-a


SulfuricAcid


Wastewatertreatment

Ethanol

Transportofethanol

Ammonia

Transportofammonia

Methanol

Transportofmethanol

Transportation

LacticAcid

Biomethane

Lignin(H)

Lignin(O)

CalciumSulfate

Energy

from

NaturalGas

[MJ]

10

00

00

00

00

00

00

00

00

00

00

00

00

0-2E+05

-23692

-12382,7

-49488

-19501

Magnesium

Sulphate[kg]

01

00

-39,7

00

00

00

00

00

00

00

00

00

00

00

00

00

0

Phosphorus[kg]

00

10

-39,7

00

00

00

00

00

00

00

00

00

00

00

00

00

0

Sulfur[kg]

00

01

-39,7

00

00

00

00

00

00

00

00

00

00

00

00

00

0

Nutrients

(P,S,M

g)[kg]

00

00

119

-10

00

00

00

00

00

00

00

00

00

00

00

00

0


[kg]

00

00

01

00

00

00

00

00

00

00

00

00

00

0-82,41

-11,589

-6,05697

-24,207

-9,539

WoodChips[kg]

00

00

00

1-1

00

00

00

00

00

00

00

00

00

00

00

00


00

00

00

01

00

00

00

00

00

00

00

00

00

0-30795

-4330,7

-2263,44

-9045,9

-3565

Yeast

[kg]

00

00

00

00

1-1

00

00

00

00

00

00

00

00

00

00

00


00

00

00

00

01

00

00

00

00

00

00

00

00

0-178,4

-25,092

-13,1144

-52,412

-20,65

Enzymes

[kg]

00

00

00

00

00

1-1

00

00

00

00

00

00

00

00

00

00

Transportedenzymes

[kg]

00

00

00

00

00

01

00

00

00

00

00

00

00

0-85,61

-12,039

-6,29237

-25,148

-9,909

Lim

e[kg]

00

00

00

00

00

00

1-1

00

00

00

00

00

00

00

00

00

Transportedlime[kg]

00

00

00

00

00

00

01

00

00

00

00

00

00

0-3883

-546,12

-285,434

-1140,7

-449,5

Refrigerant,R134-a[kg]

00

00

00

00

00

00

00

1-1

00

00

00

00

00

00

00

00

Transportedrefrigerant,R134-a[kg]

00

00

00

00

00

00

00

01

00

00

00

00

00

0-3,803

-0,5349

-0,27954

-1,1172

-0,44

Building[pcs]

00

00

00

00

00

00

00

00

10

00

00

00

00

0-3E-06

-4E-07

-1,9E-07

-8E-07

-3E-07

SulfuricAcid[kg]

00

00

00

00

00

00

00

00

01

-10

00

00

00

00

00

00


[kg]

00

00

00

00

00

00

00

00

00

10

00

00

00

0-10563

-1485,4

-776,36

-3102,8

-1223

Treated

waste

water

[m3]

00

00

00

00

00

00

00

00

00

01

00

00

00

0-245,1

-34,469

-18,0152

-71,998

-28,37

Ethanol

[kg]

00

00

00

00

00

00

00

00

00

00

1-1

00

00

00

00

00

Transportedethanol

[kg]

00

00

00

00

00

00

00

00

00

00

01

00

00

0-1662

-233,68

-122,135

-488,12

-192,3

Ammonia

[kg]

00

00

00

00

00

00

00

00

00

00

00

1-1

00

00

00

00

Transportedam

monia

[kg]

00

00

00

00

00

00

00

00

00

00

00

01

00

0-30,8

-4,3307

-2,26344

-9,0459

-3,565

Methanol

[kg]

00

00

00

00

00

00

00

00

00

00

00

00

1-1

00

00

00

Transportedmethanol

[kg]

00

00

00

00

00

00

00

00

00

00

00

00

01

0-86,84

-12,213

-6,3829

-25,51

-10,05

Transportation

[tkm]

00

00

0-0,15

0-0,15

0-0,15

0-0,2

0-0,15

0-0,15

00

-0,15

00

-0,15

0-0,15

0-0,15

10

00

00

LacticAcid[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

015151

00

00

Biomethane[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

00

5042,1

00

0

Lignin(H

ydrolysis)

[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

00

022683,8

00

Lignin

(organosolv)[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

00

00

7770,6

0

Calcium

Sulfate[kg]

00

00

00

00

00

00

00

00

00

00

00

00

00

00

00

012860

CO2[kg]

0,063

0,269

8,848

0,2418

00

-1,75

0-15,02

05

00,007

06,68

01,28E+07

0,098

075,02

-0,5

01,46

00

00,164

4477,7

629,68

329,104

1315,3

518,28

N2O

[kg]

9E-07

1E-05

3E-04

5E-06

00

00

0,013

00

01E

-06

00

03,31E+02

3E-06

03E

-04

00

00

00

3E-06

00

00

0CH4[kg]

2E-04

7E-04

0,037

0,0011

00

00

0,24

00

01E

-05

00,02

04,62E+04

4E-04

00,202

00

00

00

2E-04

734,15

103,24

53,9597

215,65

84,977

Coal,hard[kg]

0,001

0,063

3,798

0,019

00

00

1,35

00

08E

-04

01,42

04,77E+06

0,017

00,751

0,16

00

00

00,007

00

00

0Coal,soft,lignite[kg]

8E-04

0,072

1,909

0,0039

00

00

0,235

00

05E

-04

00,47

06,97E+05

0,005

00,883

0,02

00

00

00,001

00

00

0NaturalGas

[Nm3]

0,03

0,034

1,085

0,1268

00

00

0,867

00

04E

-04

01,64

07,16E+05

0,032

00,61

0,19

00,6

00.65179

00,005

00

00

0Oil,crude[kg]

3E-04

0,007

0,448

0,5242

00

0,01

01,216

00

09E

-04

01,15

06,76E+05

0,115

00,221

0,09

00,2

00

00,052

00

00

0LandUse

[m2*yr]

0,079

0,017

0,916

0,0066

00

2,69

018,17

02

00,006

00,32

05,79E+06

0,015

01,262

2,51

00

00

00,011

00

00

0Water

Use

[m3]

0,072

2,118

10,2993

00

0,04

00,788

00

00,877

00,04

08,54E+04

0,482

00

0,37

00,14

00,01

00,093

77730

10931

5713,07

22833

8997,1

Thevalues

are

expressed

onanhourlybasis

101

Table B.2: Partitioning method




Lactic Acid 1,10 16665 61,59

Biomethane 0,46 2343,68 8,66

Lignin (Hydrolysis) 0,05 1224,93 4,53


Calcium Sulphate 0,15 1929,05 7,13

Total 27059,02 100

Table B.3: Avoided burdens for the biore�nery concept 3lacticacid

production[1

kg]

naturalgas,highpressure

[1m3]

pulverised

ligniteproduction[1

MJ]

phenol

production[1

kg]

gypsum

(stucco)

production[1

kg]

CO2 [kg] 3,79 0,41 0,3 4,18 1,69N2O [kg] 0,43 0,008 0,007 0,43 0,008CH4 [kg] 0,37 0,47 0,76 0,36 0,59Coal, hard [kg] 0,34 0,003 0,004 1,23 0,002Coal, soft, lignite [kg] 0,20225 0,003 0,12099 0,21 0,001Natural Gas [Nm3] 0,85 1,08 0,60 0,94 0,26Oil, crude [kg] 0,71 0,001 0,0002 1,15 0,004Land Use [m2*yr] 1,45 0,6 1,82 1,63 1,8Water Use [m3] 8,93 0,7 0,71 8,68 1,14


102 Appendix B. Results for the Lactic Acid concept

Table B.4: Final demand vector, f and scaling vector s

fFinal

dem

andvector

sScalingvector




Sulfur [kg] 0 44,5992504





Yeast [kg] 0 289,7


Enzymes [kg] 0 139


Lime [kg] 0 6305,3


Refrigerant, R134-a [kg] 0 6,1751955

Transported refrigerant, R134-a [kg] 0 6,1751955




Treated waste water [m3] 0 397,96

Ethanol [kg] 0 2698

Transported ethanol [kg] 0 2698

Ammonia [kg] 0 50

Transported ammonia [kg] 0 50

Methanol [kg] 0 141

Transported methanol [kg] 0 141


Lactic Acid [kg] 15150,8 1


Lignin(Hydrolysis) [kg] 22683,8 1

Lignin (organosolv) [kg] 7770,6 1

Calcium Sulfate [kg] 12860,3 1

Table B.5: Environmental interventions vector, g


CO2 [kg] -34040

N2O [kg] 6,7

CH4 [kg] 1416

Coal, hard [kg] 2159

Coal, soft, lignite [kg] 929

Natural Gas [Nm3] 9945

Oil, crude [kg] 4020

Land Use [m2*yr] 169348


103

Table B.6: Impact vector h






Table B.7: Avoided impacts h'



Reference impactvector, h' [-/kg wood]


3614931 28919453412 72,29


39759 318078596 0,79

water depletion [m3] 207135 1657083735 4,14abiotic depletion,fossil fuels [MJ]

2338581 18708650780 46,77

Table B.8: Normalized impacts h*


h*=h-h'

-28892429854 -72,23

1036707108 2,59

-344635821 -0,86

-13685575032 -34,21


(a)

(b)

(c)

(d)

Figure

B.2:Analysisof

thecontribution

ofeach

process


hour)

105

Figure

B.3:Relativecontribution

ofeach

process

tothetotalenvironmentalinterventions


(a)

(b)

(c)

(d)

Figure

B.4:Analysisof

thecontribution

ofeach

process

totheim

pactcategories

g(per

hour)

107

Figure

B.5:Relativecontribution

ofeach

process

tothetotalim

pactcategories

Appendix C

A focus on the Organosolv process

The key components of lignocellulosic biomass, i.e., cellulose, hemicelluloses andlignin, are closely associated with each other at the plant cell level. This close as-sociation, together with the partly crystalline nature of cellulose, reduces cellulosereactivity towards acid and enzymatic hydrolysis in native biomass. Thus, organo-solv pretreatment is necessary to render the carbohydrate fraction to acid, enzymaticand microbial action.

From structure study to pulping and currently to energy usage organosolv frac-tionation has a long history. The earliest study applying organic solvents to treatlignocellulosic material was back in 1893, when Klason used ethanol and hydrochlo-ric acid to separate wood into its components to study the structure of lignin andcarbohydrates. Nowadays organosolv pretreatment has been used for lignin andother potentially valuable co-products (e.g. acetone, butanol, biogas) production

Organosolv is based on the retatment of biomass with an (aqueous) organicsolvent at elevated temperatures. Commonly used solvents are ethanol, methanol,acetone and organic acids like acetic acid and formic acid or combinations thereof.Organosolv processes delignify lignocellulose, with the organic solvent functioning aslignin extractant, while the hemicellulose is depolymerized through acid-catalysedhydrolysis. In general, organosolv processes aim to fractionate the lignocellulosicbiomass as much as possible into its individual major fractions in contrast to otherpre-treatment technologies such as steam explosion and dilute acid hydrolysis. Thesetechnologies merely make the cellulose fraction suitable for further processing with-out recovery of a puri�ed lignin fraction.

Organosolv lignin has a high purity (limited amounts of residual carbohydratesand minerals) due to its isolation process. Consequently, its application spectrumis broader compared to the more impure lignin-containing residues derived fromconventional pre-treatments which are targeted primarily towards the productionof cellulose for paper or second generation bioethanol. The latter are a complexmixture of unconverted carbohydrates, lignin, minerals and process chemicals ormicrobial residues. Hardly any applications for such complex byproducts have beenidenti�ed other than combustion for combined heat and power (CHP). Organosolvlignins also have a relatively low molecular weight with a narrow distribution and avery low sulphur content.

109

110 Appendix C. A focus on the Organosolv process

Figure C.1: Organosolv-based lignocellulosic biore�nery

Organosolv lignins can be used as a functional high-quality additive in inks, var-nishes and paints. Other examples are the use in blends with polyethylene oxide,as radical scavenger (anti-oxidant) and as matrix material in biobased composites.Organosolv lignin is also a candidate for high-value applications such as carbon �-bres and aromatic (specialty) chemicals. For the latter application the lignin shouldin general be depolymerized by appropriate technology such as chemocatalytic de-polymerization, partial oxidation or pyrolysis. Finally, the high quality organosolvlignin is a preferable candidate for phenolic resins and polyurethane (PU) foams.

Bibliography

[1] http://www.aqua-calc.com/calculate/volume-to-weight.

[2] https://www.mindat.org/min-1784.html.

[3] http//www.gabi-software.com/international/news/news-detail/article/a-brief-history-of-life-cycle-assessment-lca.

[4] http://www.indexmundi.com/energy/?country=de&product=oil&graph=production.

[5] http://www.indexmundi.com/germany/natural_gas_consumption.html.

[6] https://minerals.usgs.gov/minerals/pubs/commodity/gypsum/mcs-2015-gypsu.pdf.

[7] R. Clift A. Azapagic. Life cycle assessment and multiobjective optimisation.Journal of Cleaner Production, 7:9, 1998.

[8] Allan Astrup Jensen, Leif Ho�man, Birgitte T. Møller, Anders Schmidt, KimChristiansen, John Elkington, Franceska van Dijk. Life cycle assessment_ aguide to approaches, experiences and information sources.

[9] A. Azapagic and R. Clift. The application of life cycle assessment to processoptimisation. Computers & Chemical Engineering, 23(10):1509�1526, 1999.

[10] Adisa Azapagic. Life cycle assessment and its application to process selection,design and optimisation. Chemical Engineering Journal, 73:21, 1999.

[11] Harro von Blottnitz and Mary Ann Curran. A review of assessments conductedon bio-ethanol as a transportation fuel from a net energy, greenhouse gas, andenvironmental life cycle perspective. Journal of Cleaner Production, 15(7):607�619, 2007.

[12] Maik Budzinski and Roy Nitzsche. Comparative economic and environmentalassessment of four beech wood based biore�nery concepts. Bioresource technol-ogy, 216:613�621, 2016.

[13] Francesco Cherubini. The biore�nery concept: Using biomass instead of oilfor producing energy and chemicals. Energy Conversion and Management,51(7):1412�1421, 2010.

111

http://www.aqua-calc.com/calculate/volume-to-weight

https://www.mindat.org/min-1784.html

http//www.gabi-software.com/international/news/news-detail/article/a-brief-history-of-life-cycle-assessment-lca

http//www.gabi-software.com/international/news/news-detail/article/a-brief-history-of-life-cycle-assessment-lca

http://www.indexmundi.com/energy/?country=de&product=oil&graph=production

http://www.indexmundi.com/energy/?country=de&product=oil&graph=production

http://www.indexmundi.com/germany/natural_gas_consumption.html

https://minerals.usgs.gov/minerals/pubs/commodity/gypsum/mcs-2015-gypsu.pdf

https://minerals.usgs.gov/minerals/pubs/commodity/gypsum/mcs-2015-gypsu.pdf

112 BIBLIOGRAPHY

[14] Francesco Cherubini and Gerfried Jungmeier. Lca of a biore�nery conceptproducing bioethanol, bioenergy, and chemicals from switchgrass. The Inter-national Journal of Life Cycle Assessment, 15(1):53�66, 2010.

[15] Francesco Cherubini and Sergio Ulgiati. Crop residues as raw materials forbiore�nery systems � a lca case study. Applied Energy, 87(1):47�57, 2010.

[16] Jorge Cristóbal, Cristina T. Matos, Jean-Philippe Aurambout, Simone Man-fredi, and Boyan Kavalov. Environmental sustainability assessment of bioecon-omy value chains. Biomass and Bioenergy, 89:159�171, 2016.

[17] S. C. Davis, K. J. Anderson-Teixeira, and E. H. Delucia. Life-cycle analysisand the ecology of biofuels. Trends Plant Sci, 14(3):140�146, 2009.

[18] A. Demirbas. Progress and recent trends in biofuels. Progress in Energy andCombustion Science, 33(1):1�18, 2007.

[19] Ayhan Demirbas. Biofuels sources, biofuel policy, biofuel economy and globalbiofuel projections. Energy Conversion and Management, 49(8):2106�2116,2008.

[20] Ayhan Demirbas. Biore�neries: Current activities and future developments.Energy Conversion and Management, 50(11):2782�2801, 2009.

[21] Ayhan Demirbas. Political, economic and environmental impacts of biofuels: Areview. Applied Energy, 86:S108�S117, 2009.

[22] Dimitrios K. Sidiras and Ioanna S. Salapa. Organosolv pretreatment as a majorstep of lignocellulosic biomass re�ning.

[23] US DoE. Lignocellulosic biomass for advanced biofuels and bioproducts: Work-shop report. US Department of Energy�O�ce of Science. http:// genomic-science. energy. gov/ biofuels/ lignocellulose. Accessed Jun, 201, 2015.

[24] HJ Endres, A Siebert-Raths, H Behnsen, and C Schulz. Biopolymers facts andstatistics. IfBB�Institute for Bioplastics and Biocomposites, Hanover, 2014.

[25] European Commission - Joint Research Centre - Institute for Environmentand Sustainability. General guide for life cycle assessment - detailed guidance.International Reference Life Cycle Data System (ILCD) Handbook, 2010.

[26] European Commission - Joint Research Centre - Institute for Environmentand Sustainability. Review schemes for life cycle assessment. InternationalReference Life Cycle Data System (ILCD) Handbook, 2010.

[27] European Commission - Joint Research Centre - Institute for Environment andSustainability. Speci�c guide for life cycle inventory data sets. InternationalReference Life Cycle Data System (ILCD) Handbook, 2010.

BIBLIOGRAPHY 113

[28] European Commission - Joint Research Centre - Institute for Environmentand Sustainability. Recommendations for life cycle impact assessment in theeuropean context. International Reference Life Cycle Data System (ILCD)Handbook, 2011.

[29] M. Fatih Demirbas. Biore�neries for biofuel upgrading: A critical review. Ap-plied Energy, 86:S151�S161, 2009.

[30] M. FitzPatrick, P. Champagne, M. F. Cunningham, and R. A. Whitney. Abiore�nery processing perspective: treatment of lignocellulosic materials forthe production of value-added products. Bioresour Technol, 101(23):8915�8922,2010.

[31] E. Gnansounou, P. Vaskan, and E. R. Pachon. Comparative techno-economicassessment and lca of selected integrated sugarcane-based biore�neries. Biore-sour Technol, 196:364�375, 2015.

[32] Bärbel Hahn-Hägerdal, Mats Galbe, Marie-Francoise Gorwa-Grauslund, Gun-nar Lidén, and Guido Zacchi. Bio-ethanol�the fuel of tomorrow from theresidues of today. Trends in biotechnology, 24(12):549�556, 2006.

[33] Carlo N. Hamelinck, Geertje van Hooijdonk, and André P. C. Faaij. Ethanolfrom lignocellulosic biomass: techno-economic performance in short-, middle-and long-term. Biomass and Bioenergy, 28(4):384�410, 2005.

[34] Anders Rasmuson Hans Björk. A method for life cycle assessment environmen-tal optimisation of a dynamic process exempli�ed by an analysis of an energysystem with a superheated steam dryer integrated in a local district heat andpower. Chemical Engineering Journal, 87:14, 2001.

[35] John Houghton, Sharlene Weatherwax, and John Ferrell. Breaking the biolog-ical barriers to cellulosic ethanol: a joint research agenda. Technical report,EERE Publication and Product Library, 2006.

[36] Ed de Jong and Gerfried Jungmeier. Biore�nery concepts in comparison topetrochemical re�neries. pages 3�33, 2015.

[37] Susan G. Karp, Adriana H. Igashiyama, Paula F. Siqueira, Julio C. Carvalho,Luciana P. S. Vandenberghe, Vanete Thomaz-Soccol, Je�erson Coral, Jean-LucTholozan, Ashok Pandey, and Carlos R. Soccol. Application of the biore�neryconcept to produce l-lactic acid from the soybean vinasse at laboratory andpilot scale. Bioresource technology, 102(2):1765�1772, 2011.

[38] Eric D. Larson. A review of life-cycle analysis studies on liquid biofuel systemsfor the transport sector. Energy for Sustainable Development, 10(2):109�126,2006.

114 BIBLIOGRAPHY

[39] Alya Limayem and Steven C. Ricke. Lignocellulosic biomass for bioethanol pro-duction: Current perspectives, potential issues and future prospects. Progressin Energy and Combustion Science, 38(4):449�467, 2012.

[40] Lin Luo, Ester van der Voet, and Gjalt Huppes. Biore�ning of lignocellulosicfeedstock�technical, economic and environmental considerations. Bioresourcetechnology, 101(13):5023�5032, 2010.

[41] Fabio Andres Castillo Martinez, Eduardo Marcos Balciunas, José Manuel Sal-gado, José Manuel Domínguez González, Attilio Converti, and Ricardo Pinheirode Souza Oliveira. Lactic acid properties, applications and production: a re-view. Trends in food science & technology, 30(1):70�83, 2013.

[42] Vishnu Menon and Mala Rao. Trends in bioconversion of lignocellulose: Bio-fuels, platform chemicals & biore�nery concept. Progress in Energy and Com-bustion Science, 38(4):522�550, 2012.

[43] Jochen Michels. Pilotprojekt: Lignocellulose-biora�nerie. page 239, 2009.

[44] S. N. Naik, Vaibhav V. Goud, Prasant K. Rout, and Ajay K. Dalai. Productionof �rst and second generation biofuels: A comprehensive review. Renewableand Sustainable Energy Reviews, 14(2):578�597, 2010.

[45] Roy Nitzsche, Maik Budzinski, and Arne Grongroft. Techno-economic assess-ment of a wood-based biore�nery concept for the production of polymer-gradeethylene, organosolv lignin and fuel. Bioresource technology, 200:928�939, 2016.

[46] Stephane Octave and Daniel Thomas. Biore�nery: Toward an industrialmetabolism. Biochimie, 91(6):659�664, 2009.

[47] C. Oliveira, D. Coelho, and C. H. Antunes. Coupling input�output analysis withmultiobjective linear programming models for the study of economy�energy�environment�social (e3s) trade-o�s: A review. Annals of Operations Research,247(2):471�502, 2016.

[48] Rajeev M. Pandia. The phenol-acetone value chain: prospects and opportuni-ties.

[49] P.J. de Wild, W.J.J. Huijgen, R. van der Linden, H. den Uil, J. Snelders, B.Benjelloun-Mlayah. Organosolv fractionation of lignocellulosic biomass for anintegrated biore�nery. ECN, 14(010):14, 2015.

[50] S. Suh R. Heijungs. The computational structure of life cyce assessment. Eco-E�ciency in Industry and Science, 11:235, 2002.

[51] L. MeyerA. R.K. Pachauri. Climate change 2014: Synthesis report. contributionof working groups i, ii and iii to the �fth assessment report of the intergovern-mental panel on climate change. IPCC, page 151, 2014.

BIBLIOGRAPHY 115

[52] Per Sassner, Mats Galbe, and Guido Zacchi. Techno-economic evaluation ofbioethanol production from three di�erent lignocellulosic materials. Biomassand Bioenergy, 32(5):422�430, 2008.

[53] Anke Siebert, Alberto Bezama, Sinéad O'Kee�e, and Daniela Thrän. Social lifecycle assessment: in pursuit of a framework for assessing wood-based productsfrom bioeconomy regions in germany. The International Journal of Life CycleAssessment, 2016.

[54] Ralph E. H. Sims, Warren Mabee, Jack N. Saddler, and Michael Taylor. Anoverview of second generation biofuel technologies. Bioresource technology,101(6):1570�1580, 2010.

[55] T. Werpy: Paci�c Northwest National Laboratory and G. Petersen: NationalRenewable Energy Laboratory. Top value added chemicals from biomass: Vol-ume i�results of screening for potential candidates from sugars and synthesisgas.

[56] Gail Taylor. Biofuels and the biore�nery concept. Energy Policy, 36(12):4406�4409, 2008.

[57] R Taylor, L Nattrass, G Alberts, P Robson, C Chudziak, A Bauen, IM Li-belli, G Lotti, M Prussi, R Nistri, et al. From the sugar platform to biofuelsand biochemicals: Final report for the european commission directorate-generalenergy. Technical report, E4tech/Re-CORD/Wageningen UR, 2015.

[58] Suh S. Huppes G. Udo de Haes H., Heijungs R. Three strategies to overcomethe limitations of life-cycle assessment. Journal of Industrial Ecology, 8:19�32,2004.

[59] Andreas Uihlein and Liselotte Schebek. Environmental impacts of a lignocel-lulose feedstock biore�nery system: An assessment. Biomass and Bioenergy,33(5):793�802, 2009.

[60] U.S. Energy Information Administration. International energy outlook 2016.

[61] V.A. Lignite in germany 2015 facts and �gures. Bundesverband Braunkohle.

[62] Konstantin M. Zech, Kathleen Meisel, André Brosowski, Lars Villadsgaard Toft,and Franziska Müller-Langer. Environmental and economic assessment of theinbicon lignocellulosic ethanol technology. Applied Energy, 171:347�356, 2016.

Dichiarazione

Piacenza, April 2017

Mattia Sisca

POLITECNICO DI MILANO Mattia... · luogo, di mettere in pratica le conoscenze acquisite migliorando...

Documents

Transcript of POLITECNICO DI MILANO Mattia... · luogo, di mettere in pratica le conoscenze acquisite migliorando...