LIFE CYCLE ASSESSMENT OF HIGHWAY RECONSTRUCTION: A …docs.trb.org/prp/17-02163.pdf · 1 LIFE CYCLE...

LIFE CYCLE ASSESSMENT OF HIGHWAY RECONSTRUCTION: A CASE STUDY 1 2 3 4 Eleanor Bloom, M.S., EIT 5 Geological Engineering Program, University of Wisconsin-Madison 6 2243 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706 7 Tel: (219) 771-0409; Email: [email protected] 8 9 Aaron Canton 10 Civil and Environmental Engineering Department, University of Wisconsin-Madison 11 2221 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706 12 Tel: (781) 698-5275; Email: [email protected] 13 14 Angela Pakes Ahlman, P.E., LEED AP, Corresponding Author 15 Geological Engineering Program and Recycled Materials Resource Center, University of 16 Wisconsin-Madison 17 2204 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706 18 Tel: (608) 890-4966; Email: [email protected] 19 20 Tuncer Edil, Ph.D., P.E., D.GE, Dist. M. ASCE 21 Geological Engineering Program and Recycled Materials Resource Center, University of 22 Wisconsin-Madison 23 2258 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706 24 Tel: (608) 262-3225; Email: [email protected] 25 26 27 Total Words: 5,975 words text + 6 figures/tables x 250 words (each) = 7,475 words 28 29 30 31 32 33 34 Submission Date: November 14, 2016 35

Bloom, Canton, Ahlman, and Edil 2

ABSTRACT 1 The use of recycled materials in the reconstruction of an urban highway was quantitatively 2 analyzed to assess the environmental and economic benefits of sustainable road construction. 3 Recycled materials in the reconstruction included fly ash, recycled concrete aggregate, recycled 4 asphalt shingles, and recycled asphalt pavement. In addition to assessing the benefits of recycled 5 material use, this case study was used to evaluate two methodologies for gathering input data for 6 life cycle assessments (LCAs). In past studies on road LCAs conducted post or prior to 7 construction, significant assumptions were made to determine the input data. This project offered 8 a unique opportunity to compare the post-construction, estimated input data with data explicitly 9 collected and tracking while the road construction was on-going. Furthermore, multiple LCA 10 tools were used to conduct, compare, and verify the LCA results. The actual design including 11 recycled materials was compared to a reference design using only virgin materials to determine 12 the reduction in environmental impacts. Results show that use of the recycled materials does 13 reduce impacts in most environmental criteria, including energy consumption (15-25% 14 reduction) and CO2 emissions (17-24% reduction). Although impact results did differ between 15 LCA tools, the data collection method comparison revealed significant variances in LCA inputs 16 regardless of which tool was used for the impact assessment. Therefore, it is recommended that 17 Departments of Transportation should focus future efforts on material tracking for LCAs when 18 these issues are critical. Overall, the study’s results suggest that incorporating recycled materials 19 improved the sustainability of road construction. 20 21 22 Keywords: Environmental impact, Life cycle assessment, Recycled material, Sustainability 23


INTRODUCTION 1 There is interest in determining and validating the environmental benefits of incorporating 2 recycled materials into road construction using life cycle assessments (LCA) tools. However, the 3 process of collecting the necessary data for LCAs from departments of transportations (DOTs) 4 and road construction contractors is not well defined. In previous case studies, LCA data was 5 estimated from planned design and average mix specifications gathered after the road 6 construction was completed. Post construction data led to issues such as over-generalization of 7 mix designs and sourcing and inability for real-time data collection. For this study of a typical 8 urban highway, the Recycled Materials Resource Center (RMRC) was able to work with local 9 engineers and contractors to explicitly track and quantify the material used in construction, 10 identify material sources, and determine transportation distances. This project provided a study 11 of real-time data collection to compare with the results of estimated post-construction LCA data. 12 The goal of this comparison is to determine a data collection precedent for environmental 13 analyses of future projects. Additionally, two prominent LCA tools were used in conducting the 14 LCA and the results were compared. 15

16 Background 17 Sustainable roadway construction has become an increasingly popular topic because of the 18 impacts on climate change and rising costs of virgin, or non-recycled materials in road 19 construction. Buildings and infrastructure utilize 40% of all materials extracted in the U.S. (1), 20 and the construction industry emits approximately 6% of total U.S. industry-related greenhouse 21 gasses (GHGs) (2). To be sustainable, environmental impacts of highways must be reduced 22 through thoughtful planning, design and construction. This includes reducing the use of virgin 23 materials (3). Production of common road construction materials, such as crushed rock aggregate 24 or cement, consume significant energy, generate GHG emissions, are increasingly limited in 25 supply, and often incur high transportation costs (4), (5). After demolition, previously used 26 concrete or asphalt pavement is either recycled or sent to a landfill to remains unused (6)–(8). 27 However, sustainable road construction incorporates as much existing material on site as 28 possible. Recycled coal by-products such as fly ash and bottom ash are proven useful alternatives 29 to using virgin material (9), (10). The RMRC studies the viability of fly ash as an alternative 30 binder both in the surface concrete mix and as stabilization in the base course. 31 For this case study, recycled materials in the reconstruction and expansion of a 2.4 km 32 (1.5-mi) stretch of an urban highway was quantitatively analyzed. Preliminary findings were 33 presented at the 4th international conference in Sustainable Construction Material and 34 Technologies (SCMT4) and Geo-Chicago 2016 (11), (12). Recycled materials used in this 35 reconstruction included: fly ash, blast furnace slag, recycled asphalt shingles (RAS), recycled 36 asphalt pavement (RAP), and recycled concrete aggregate (RCA). Fly ash and blast furnace slag 37 can be used as a partial replacement of Portland cement in the ready-mix concrete (9), (13). RAP 38 was used in both the hot mix asphalt (HMA) pavement as well as a base course material. 39 Properly processed RAP consists of high-quality, well-graded aggregates coated by asphalt 40 cement (14). The aggregated RAP often undergoes a specific gradation process such that it 41 becomes fractionated recycled asphalt pavement (FRAP). In HMA, the asphalt content of RAP 42 can be used as binder, while the aggregate portion can replace virgin aggregates (15). Similarly, 43 RAS was substituted for binder and aggregate material in some HMA mix designs. For this 44 highway, approximately 20% of RAS and 4% of RAP was allowed as binder replacement in the 45 HMA design. RAP was also used with RCA as base aggregate and fill material. 46


The reconstruction involved expanding from two to three lanes in each direction. The 1 existing HMA surface pavement was replaced with concrete pavement, at times over an asphaltic 2 base. Six ramps were updated and four were added. The reconstruction design chosen for the 2.4 3 km (1.5-mi) stretch is generally comprised of 28-cm (11-in) Portland cement concrete (PCC) 4 pavement over a 15-cm (6-in) base aggregate over a 41-cm (16-in) subbase of select crushed 5 material (SCM). An 8-cm (3-in) asphaltic base layer appears between the concrete pavement and 6 base course in certain locations. 7 8 Life Cycle Assessment 9 LCAs were conducted to quantify the environmental benefits gained from using recycled 10 materials in reconstruction. LCA quantifies environmental impacts over the lifetime of a product 11 by using a meticulous evaluation methodology (16). According to the FHWA, LCAs provide a 12 comprehensive approach to evaluating the total environmental burden of a particular product or a 13 more complex system of products and processes (17). In the case of pavements, LCAs often 14 evaluate the impacts from the materials and processes used to construct the highway, the 15 transportation of those material, and any rehabilitation processes. Rehabilitation and maintenance 16 can be considered the end-of-life practices for roads. Impacts during the use phase are often not 17 considered, as those impacts are mainly caused by the vehicles on the road rather than the road 18 itself. In this study, two LCA computer tools are used to evaluate the environmental benefits of 19 recycled material use, as discussed in the LCA Methods section. 20 21 DATA COLLECTION AND LCA METHODS 22 23 Data Collection 24 Two data collection methodologies were employed to gather the necessary inputs for LCA of the 25 reconstruction: 1) material quantities estimated from designs and specifications as planned prior 26 to construction (referred as Planned), and 2) material quantities explicitly tracked and collected 27 while construction was on-going (referred as Constructed). 28 29 Collecting Planned Data 30 Plans for the reconstruction project were analyzed to estimate LCA input material quantities. 31 Pavement volumes were calculated from cross-sectional drawings using the provided 32 dimensions. The road layers were divided into component materials based on average mix 33 designs and estimated ratios. Also included in the design plans were total material volumes for 34 the concrete bridges and structures. Mix designs for concrete and HMA, as well as estimated 35 recycled-to-virgin aggregate ratios in the base and subbase, were provided by DOT engineers 36 and the project contractors. For this project, all RAP and some RCA in the base was recycled on-37 site from existing pavements. Some RCA was imported from a near-by aggregate supplier. 38

To perform a life cycle analysis, materials for future maintenance were required. The 39 DOT provided a maintenance schedule over the 50-year lifetime of the road, including material 40 quantities for the rehabilitation processes (18). Maintenance materials included concrete for two 41 repair and grind rehabilitations and HMA for a 10-cm (4-in) overlay. The mix designs and 42 sourcing for the maintenance materials are estimated based on initial pavement specifications, 43 although they may vary in time. 44

45


Collecting Constructed Data 1 DOT engineers and project managers helped facilitate the collection process while construction 2 was on-going. Contractors and sub-contractors clarified quantities and procedures. Key, site-3 specific DOT and sub-contractor files, including Item Record Account (IRA) spreadsheets, 4 Quality Management Plan (QMP) specifications, concrete and hot mixed asphalt (HMA) 5 pavement mix designs, weigh tickets, site plans, and bid item lists, were accessed for information 6 on materials. QMP plans kept particularly detailed records of the type and amount of material 7 being used, as well as their sources. QMP specifications are used to verify product acceptance 8 based on contractor’s quality control testing (19). Quality testing must be verified per a certain 9 amount of product used in the road construction. Therefore, QMP plans tracked the quantity of 10 materials, supplier, and date of placement in great detail. Weigh tickets were also critical in data 11 collection because they specified the material, its origin, and its quantity. Weigh tickets were 12 used to track the quantity and supplying quarries for subbase SCM. 13

Omitted from weigh tickets were pavements recycled in situ, such as RAP and RCA used 14 as base course or fill. To account for these un-weighed materials, the site plans were used to 15 calculate the tonnage of RAP and RCA recycled from the existing road. This estimate was 16 deemed valid because the contractor stated that in the construction, almost all RAP and RCA was 17 immediately used for the new highway as either base or fill. 18

As the rehabilitation construction will not begin for a number of years, the material used 19 for these processes could not be collected. Therefore, this maintenance data was estimated based 20 upon the requirements for rehabilitation strategies anticipated by the DOT. 21

22 Data Input Assumptions 23 The assumptions made while performing the LCAs are as follows: 24

• Existing road dimensions were used to calculate volume of RAP and RCA. Average 25 widths were used when ranges were provided. 26

• Transportation distances were determined from the material origin to either the plant 27 locations or the reconstruction site as appropriate. 28

• Unless otherwise stated, the assumed transportation vehicles were dump trucks, with the 29 exception of cement trucks for cement/fly ash. Cement was also shipped from the plant 30 via barge. 31

• To analyze a comparable reference design with no recycled material, conventional virgin 32 material was substituted ton-for-ton for all recycled material used in the project (i.e. 33 cement for fly ash, virgin aggregate for RCA, etc.). In reality, different dimensions or 34 quantities of virgin material may be required to construct the virgin road to meet the 35 structural support requirement. Additionally, the all-virgin design may also result in a 36 different service life or rehabilitation procedures. All quantities, service lives, and 37 rehabilitation schedules were considered equal for the reference, virgin design. 38

39 Life Cycle Assessment Tools 40 Two prominent tools were used to conduct the LCA with the objective of validating LCA result. 41 Both tools provide individual impact assessment parameters, as well mutual impact categories 42 that could be used for a comparison. 43 44


PaLATE 1 The Pavement Life-cycle Assessment Tool for Environmental and Economic Effects (PaLATE) 2 is an open-source LCA program commissioned by the RMRC and designed by the Consortium 3 on Green Design and Manufacturing (20). It is an LCA tool in spread sheet format specifically 4 developed for highway construction and available in the public domain. Users input the initial 5 construction and maintenance materials and transportation details. PaLATE calculates 6 environmental impacts from material production, materials transportation, and construction 7 processes. Environmental outputs include: 8

• Energy consumption (GJ) 9 • Water consumption (kg) 10 • Carbon dioxide (CO2) emissions (Mg) 11 • Nitrous oxide (NOx) emissions (kg) 12 • Particulate matter 10 (PM10) emissions (kg) 13 • Sulfur dioxide (SO2) emissions (kg) 14 • Carbon monoxide (CO) emissions (kg) 15 • Mercury (g) 16 • Lead (g) 17 • Resource Conservation Recovery Act hazardous waste (Haz Waste) generation (kg) 18

When comparing the LCAs of two or more products, a relative ranking of alternatives 19 was analyzed as well as the absolute impacts, or the impact magnitudes in a certain category. For 20 this study, the design of the actual roadway that incorporated recycled material (referred to as 21 Recycled) was compared to a reference, hypothetical design comprised of no recycled material 22 (referred to as Virgin). This method demonstrates the impact reduction from the use of recycled 23 material. A Recycled and Virgin design was analyzed for both the Planned and Constructed data. 24

One challenge of LCAs is comparing absolute impacts of two or more designs across 25 different environmental categories, as each category differs in units. This can be aided by 26 normalization (21). Raw LCA scores are normalized per category and per product as: 27

𝐿𝐶𝐴!! =!"#!"#!

!"#!"#! 28

where 𝐿𝐶𝐴!! is the normalized impact per category x per design, 𝐿𝐶𝐴!"#! is the raw impact per 29 category x per design, and 𝐿𝐶𝐴!"#! is the maximum value across all designs for category x. The 30 results of normalization provide impacts on a scale of 0 to 1, with 1 being the maximum impact 31 magnitude across the designs. 32 33 SimaPro 34 SimaPro is one of the leading commercial software programs for LCA studies and is employed 35 worldwide (22), (23). It is a professional LCA software used to collect, analyze, and monitor the 36 sustainability performance data of products and services. Unlike PaLATE, SimaPro LCAs are 37 not specific to road construction projects. SimaPro follows the traditional four-step LCA method 38 as described by ISO standard 14040 (16), including inventory analysis and impact assessment 39 procedures. 40

The inventory analysis includes a compilation, tabulation, and preliminary analysis of all 41 environmental exchanges of the materials and processes of the final product, in this case, the 42 highway (24). The inputs (raw material, energy, etc.) and outputs (waste, emissions, etc.) for 43 some common road construction processes such as concrete material, asphaltic material, rock 44 crushing, stone quarrying, and transportation are readily available in SimaPro. However, some 45


recycled materials specific to roads are not included (e.g. (F)RAP and bottom ash). To simulate 1 impacts for grinding, milling, and crushing (F)RAP and RAS, the impact for the hypothetical 2 amount of diesel fuel used in these processes was found. This is a similar assessment 3 methodology used in other LCAs, namely PaLATE (20). Concrete recycling was present in 4 SimaPro’s inventory and included the impact from concrete demolition. An additional process of 5 crushing was added to the RCA inventory to simulate grading the demolished concrete into 6 desired aggregate sizes. Also included in SimaPro’s inventory was cement with fly ash and slag 7 replacement. 8

Many road construction processes were not included in SimaPro’s built-in inventory. 9 However, based on previous LCAs conducted by the RMRC, construction processes had a 10 relatively insignificant (less than 10%) environmental impact as compared to the material 11 production and transportation (25). Therefore, the impacts from these processes were ignored in 12 the SimaPro analysis. 13

SimaPro’s Tool for the Reduction and Assessment of Chemical and other environmental 14 Impacts (TRACI) impact assessment method was selected to calculate the impacts because it was 15 developed by the U.S. EPA specifically for North America using input parameters consistent 16 with U.S. locations (26). TRACI’s impact categories were constructed to represent potential 17 effects in the U.S. and include: 18

• Ozone depletion (kg chlorofluorocarbon (CFC)-11 equivalents (eq)) 19 • Global warming (kg CO2 eq) 20 • Smog (kg ozone (O3) eq) 21 • Acidification (kg SO2 eq) 22 • Eutrophication (kg nitrogen (N) eq) 23 • Carcinogenic (CTUh, a comparative toxic unit for human toxicity impacts) 24 • Non-carcinogenic (CTUh) 25 • Respiratory effects (kg in particulate matter 2.5 (PM2.5) eq) 26 • Ecotoxicity (CTUe, a comparative toxic unit for aquatic ecotoxicity impacts) 27 • Fossil fuel depletion (surplus MJ) 28

Impacts predicted by PaLATE versus SimaPro were compared to validate the LCA 29 results. However, none of the TRACI impacts can be directly compared to PaLATE’s. 30 Additional SimaPro analyses were conducted for single LCA factors similar to PaLATE’s, 31 including energy, CO2, NOx, SO2, CO, and lead. Because construction processes were ignored in 32 the SimaPro analysis, they are also removed from the PaLATE results when comparing the two 33 LCA tools’ impact predictions. 34 35 RESULTS 36 A summary of the Planned and Constructed materials for the highway is shown in Table 1. In 37 general, the Constructed data predicts slightly greater material use as compared to the Planned 38 data. When explicitly tracking the material during construction, a more thorough collection of all 39 of the materials and constructed features was recognized. For example, the designs specified the 40 dimensions of concrete for the road surface alone. However, the Constructed data also include 41 ancillary concrete quantities, which is any concrete item not explicitly part of the pavement such 42 as curbs, gutters, and dividers. 43

Similarly, only the HMA used in the asphaltic base pavement was included in the 44 designs. More HMA of varying mixes was used in the actual construction as driveways, 45 temporary pavements, shoulders, and tie-ins to existing pavement on the construction ends. 46


Additionally, a lesser amount of base and subbase materials was specified in the designs as were 1 actually purchased during construction. Likely, some of the recycled pavements were used in 2 embankment fill as well as base courses, thus were not evident from design plans. 3 4 TABLE 1. Summary of Initial Construction Material Quantities Found from Planned and 5 Constructed Data Collection Methodologies 6

Material Planned Volumes (m3)

Constructed Volumes (m3)

Cement 1,880 2,380 Fly ash 715 567

Slag 0.00 17.2 PCC aggregate 15,800 20,100 PCC mix water 2,720 3,500 Concrete total 21,100 26,900

RAP binder 3.15 10.6 RAS binder 15.4 30.7

Asphalt binder 55.6 227 FRAP 206 395 RAS 44.1 86.9

HMA aggregate 1,210 1,690 HMA total 1,550 2,440

On-site RAP 2,510 7,630 On-site RCA 2,480 7,550

Imported RCA 4,230 7,100 Imported virgin aggregate 1,220 3,010

SCM subbase 6,980 9,040 RCA subbase 13,700 12,700

7 LCA Results 8 9 PaLATE Results 10 The PaLATE results of the LCA are represented by Figure 1. In Figure 1a, the percent reduction 11 was calculated by the difference in impact of the Recycled and Virgin divided by the Virgin’s 12 impact. Across both data collection methods, reductions were seen in most PaLATE categories. 13 Most categories also predicted slightly greater impact reductions for the Planned data as 14 compared to the Constructed data. However, the trends between the two data sets are similar. 15 Many transportation sectors focus on reducing their impacts of CO2 emissions, energy, and water 16 in particular. The reductions in these three categories range from 15-17% (Planned) and 12-13% 17 (Constructed). These impacts largely stem from the materials’ production. Mining and grading 18 virgin aggregate is more resource intensive than milling and grinding existing pavement. 19 Similarly, milling asphalt pavement and grinding recycled shingles to use in HMA is a less 20 environmentally intensive than producing their virgin counterparts. Because fly ash is a by-21 product, it is considered to have no impact. In the production of concrete, substituting fly ash for 22 cement results in significant impact reductions. 23


The results are shown in Figure 1. In Figure 1a, the reductions from the Planned data 1 appear greater than the Constructed data. However, the normalized results in Figure 1b reveal 2 that the absolute impacts predicted from the Planned data are less than the Constructed data. The 3 absolute impact results are related to the total quantity of material inputs. The data estimated 4 from design plans and contractors suggested a smaller quantity of materials than those actually 5 used during construction (see Table 1). 6

7

FIGURE 1. Results of the PaLATE analysis of both the Constructed and Planned data sets 8 represented by (a) percent reduction and (b) normalized absolute impacts 9 10 SimaPro Results 11

The TRACI assessment results for the two data sets are portrayed in Figure 2. There are 12 reductions in most TRACI impact categories for both data sets (Figure 2a). The greatest 13

17% 15%

17%

5%

21%

1% 4%

9% 11%

13% 13% 12% 12%

2%

24%

1% 2% 5%

8% 8%

0%

5%

10%

15%

20%

25%

30%

Energy Water CO2 NOx PM10 SO2 CO Hg Pb Haz Waste Planned Constructed

0.0

0.2

0.4

0.6

0.8

1.0

Energy Water CO2 NOx PM10 SO2 CO Hg Pb Haz Waste

Planned, Recycled Constructed, Recycled Planned, Virgin Constructed, Virgin

(a)

(b)


reductions are seen in the TRACI results for carcinogens, eutrophication, ecotoxicity, and non-1 carcinogens. The impact associated with quarrying and crushing the virgin base and subbase 2 aggregate in the Virgin design is contributing to much of the reductions. SimaPro does predict an 3 increase in respiratory effects from the Recycled design, stemming from recycling concrete as 4 compared to the production of virgin aggregate from a quarry. The absolute impacts were also 5 normalized (Figure 2b). SimaPro predicts greater impacts for the Constructed data set as 6 compared to the Planned data for both the Virgin and Recycled designs. This is also consistent 7 with the trends seen in the PaLATE normalized impacts, further suggesting that the greater 8 quantities found with the Contracted data collection method result in greater impacts. 9

10 11


1 FIGURE 2. Results of SimaPro’s TRACI assessment of both the Constructed and Planned 2 data sets represented by (a) percent reduction and (b) normalized absolute impacts 3

4 In addition, those SimaPro categories comparable to PaLATE impacts were included in 5

the impact assessment. The comparable results from both tools are shown in Table 2. 6 7

10%

24% 15%

21%

34% 35% 32%

-37%

36%

11% 14%

30% 23%

32%

49% 50% 43%

-15%

48%

16%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

50%

60%

Planned Constructed

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Planned, Recycled Constructed, Recycled

Planned, Virgin Constructed, Virgin

(a)

(b)


TABLE 2. Comparable Planned and Constructed Data Results for PaLATE and SimaPro 1 2

Data Set, LCA Tool, Design Energy (GJ)

CO2 (Mg)

NOx (kg)

SO2 (kg)

CO (kg)

Lead (g)

Planned, PaLATE

Recycled 76,700 5,380 41,300 16,400 10,300 3,780

Virgin 99,200 6,980 44,600 20,600 11,500 4,700

Savings 22,500 1,600 3,300 4,200 1,200 920

Reduction 23% 23% 7% 20% 10% 20%

Planned , SimaPro

Recycled 72,000 4,470 14,200 8,370 8,800 1,190

Virgin 85,700 5,840 16,800 11,300 12,400 1,890

Savings 13,700 1,370 2,600 2,930 3,600 700

Reduction 16% 23% 15% 26% 29% 37%

Constructed, PaLATE

Recycled 97,100 6,930 51,400 20,500 13,100 4,760

Virgin 118,000 8,300 53,700 24,300 14,000 5,560

Savings 20,900 1,370 2,300 3,800 900 800

Reduction 18% 17% 4% 16% 6% 14%

Constructed, SimaPro

Recycled 98,500 6,000 22,620 10,600 14,600 1,480

Virgin 115,000 7,510 22,600 16,800 19,000 2,760

Savings 16,500 1,510 -20 6,200 4,400 1,280

Reduction 14% 20% -0.1% 37% 23% 46% 3 4 5


DISCUSSION 1 2 Planned vs. Constructed Data 3 One goal of this project was to evaluate the two data collection methodologies and their LCA 4 results. PaLATE’s reduction comparison is made in Figure 1a. In the data collection process, the 5 Planned materials often over-generalize the actual materials used in construction. For example, 6 the average concrete pavement mix contained 3% fly ash. However, detailed data collection 7 during construction revealed that some sections of the road were paved with mixes with no fly 8 ash. Because cement in concrete has a large influence in the PaLATE analysis, the ratio of fly 9 ash to cement is significant in recognizing impact reductions. The Planned data predicts a larger 10 ratio of fly ash to cement. Additionally, only the asphaltic base was included in the design plans. 11 From the construction data, it was found that HMA was used in numerous other places such as 12 medians and shoulders. However, there was a larger ratio of RAP substitution in the binder for 13 the Planned data’s HMA as compared to the Constructed data. Therefore, the Planned data’s 14 higher RAP to virgin binder ratio leads to greater reductions. Due to similar recycling ratios, the 15 base materials have little contribution to the percent reduction differences from the two data sets. 16 Unlike PaLATE, SimaPro predicted greater reductions from the Constructed data (Figure 17 2a). There are a number of reasons for this inconsistency. The TRACI impact categories are 18 different than PaLATE’s. The same trends we see in PaLATE’s impacts may not be the same in 19 a TRACI analysis. Additionally, SimaPro’s material inventory differs from that included in 20 PaLATE. A more detailed comparison of PaLATE and SimaPro is included later in this 21 discussion. 22 The normalized results in Figures 1b and 2b also aide in the comparison. All analyses 23 predict greater absolute environmental impacts for the Recycled and Virgin designs from the 24 Constructed data. The greater impacts are due to the greater quantity of materials identified 25 during construction as compared to the quantities estimated from the designs. Calculating 26 materials from the designs plans caused some details found from the Constructed data to be 27 excluded. When collecting data during construction, it was found that more material quantities 28 were used than depicted in the design plans. For the HMA pavement, this is because the plans 29 only specify the asphaltic base and no other smaller HMA pavement work. For concrete, the 30 difference is likely caused from changes from the plans in the actual construction. For example, 31 ranges of width are provided for certain lanes in the plans, and average widths were used in the 32 design quantities calculation. 33

The largest difference in material quantities is found in the base course. There is almost a 34 60% decrease from the Constructed to the Planned base course material predictions. While the 35 ratios of RCA and RAP to virgin aggregate are relatively uniform, the total volume of base 36 calculated from the plans differs significantly from those collected during construction. The 37 Constructed base was gathered mainly from two sources: (1) calculated volumes of existing 38 pavement from design plans and (2) QMP testing of imported material and recycled pavement 39 used in base. All base material, both imported and recycled on site, were tested for their quality 40 and therefore explicitly tracked by DOT personnel. However, a certain amount of recycled 41 pavement was used for embankment fill and not tested or tracked. This is likely the cause for the 42 greater quantity of base material in the Constructed data set, and thus greater overall 43 environmental impact. 44

Overall, the Planned and Constructed data produced similar results. In particularly 45 important PaLATE categories such as energy and CO2 emissions, the two data sets differed by 7-46


8%. In similar TRACI categories such as global warming and fossil fuel depletion, the 1 Constructed data predicted a 5-6% difference in impacts reductions. Although the normalized 2 results did show greater impacts from the Constructed data, both data sets’ results are similar, 3 with no difference between like-design results greater than 30% (PaLATE) to 40% (SimaPro). It 4 may be concluded that to gain the most accurate understanding of environmental benefits from 5 road construction, detailed recycled material tracking is necessary. However, should the DOT be 6 unable to explicit track recycled material use and application, an evaluation of quantities based 7 on design plans’ estimated quantities and typical mix designs would provide a reasonable 8 estimate of the benefits. 9 10 SimaPro vs. PaLATE 11

Comparing results across multiple LCA tools can be challenging. There has not yet been 12 an internationally accepted data format for LCAs (27). Different formats lead to different 13 boundaries within material inventories, i.e. the inputs and outputs of the same material might not 14 be consistent. Most material inventories are an average of the inputs and outputs of processes and 15 products. For example, because PaLATE was specifically designed for roads, rock crushing 16 impacts are calculated for processes included in crushing rock for aggregate in roads. Because 17 SimaPro is a general LCA tool, its rock crushing process averages impacts of rock crushing for 18 multiple purposes. Additionally, the location and temporal range of the data within the software 19 can vary. Most of SimaPro’s inventory is more recent than PaLATE’s. PaLATE’s inventory was 20 created prior to 2004. SimaPro’s database pulls from multiple LCA inventories including Agri-21 footprint 2.0 (28), ecoinvent v3.1 (29), and the USA Input Output, or CEDA (30). SimaPro is an 22 international software, and therefore has inventory data from multiple nations. While some of the 23 materials, process, and assessment methods are specifically for the U.S., many inventories are 24 aggregated from global or developing world averages. PaLATE was created with U.S.-specific 25 data and designed for state DOTs. 26

To further evaluate the differences between PaLATE and SimaPro’s LCAs, a course 27 sensitivity analysis for energy consumption was conducted. For the Recycled designs based on 28 both the Constructed and Planned data quantities, SimaPro and PaLATE analyses were 29 conducted while varying broad material quantities. The inputs were varied by: 1) no change in 30 inputs, 2) doubling base and subbase quantities, 3) doubling surface pavement quantities, i.e., 31 concrete and HMA materials, and 4) doubling binder quantities, i.e., cement, fly ash, asphalt, etc. 32 Doubling the base led to small increases (8-9%) in overall impacts for both PaLATE and 33 SimaPro. However, doubling the surface material, particularly the binder, led to greater increases 34 in energy, particularly for PaLATE. Doubling the surface material led to a 74% (Planned) and 35 75% (Constructed) increase in energy consumption in the PaLATE analysis. From the SimaPro 36 results, the energy impact increased 34% (Planned) and 54% (Constructed) due to doubling the 37 surface material. Similar increases are felt when doubling the binder material alone. This trend 38 indicates that both analysis tools are most sensitive to changes in binder material inputs, but 39 PaLATE may be more sensitive to the ratio of recycled to virgin binder material as compared to 40 SimaPro. 41

With these differences in mind, six common environmental impacts were evaluated 42 between the two softwares (see Table 2). Figure 3 shows the percent reductions predicted for 43 both the Constructed and Planned data from PaLATE and SimaPro. The reductions in the energy 44 and CO2 impact categories for all analyses are within 10%. There is more variability in the 45 predictions for NOx, SO2, CO, and lead due to greater differences in each LCAs’ inventories and 46


assessment methods. For instance, PaLATE calculates larger savings in NOx, SO2, CO, and lead 1 from fly ash replacement, whereas SimaPro predicts larger savings from recycled pavement 2 replacement in base. The results are related to each software’s estimation of impact reduction per 3 unit of material production. In PaLATE, estimates are based on many EPA emissions standards. 4 Because SimaPro is a proprietary software, the calculation methods and inventory are not readily 5 visible. Therefore, it is not clear how the software estimates its impact reductions. 6

7 8 FIGURE 3. Percent reductions from PaLATE and SimaPro analyses of Planned and 9 Constructed data sets 10 11 Figure 4 shows the normalized impacts from the Recycled and Virgin designs from both 12 data sets analyzed by the two LCA tools. Again, the Constructed data consistently predicts 13 greater impacts than the Planned data. For the same design (i.e. Recycled) from the same data set 14 (i.e. Constructed), SimaPro and PaLATE predict similar results, particularly in energy, CO2 15 emissions, and carbon monoxide. For energy and CO2 impacts, PaLATE found slightly greater 16 impacts as compared to SimaPro. Overall, these trends indicate that the data collection methods 17 and resulting LCA inputs have a greater influence in environmental impact predictions as 18 compared to analysis tools, particularly for the relevant categories of energy and CO2 emissions. 19

23% 23%

7%

20%

10%

19% 18% 17%

4%

15%

6%

14% 16%

24%

15%

26% 29%

37%

15%

20%

36%

23%

47%

0% 5%

10% 15% 20% 25% 30% 35% 40% 45% 50%

Energy CO2 Nox SO2 CO Pb

Planned, PaLATE Constructed, PaLATE

Planned, SimaPro Constructed, SimaPro

0.1%


1

FIGURE 4. Absolute impacts from PaLATE (Pa) and SimaPro (S) analyses of Recycled (R) 2 and Virgin (V) designs from Planned (Pl) and Constructed (C) data sets. In the legend, the 3 labels should be read as the initials for: LCA tool (Pa vs. S), Data set (Pl vs. C), Design (R 4 vs. V) 5 6 CONCLUSION 7 This case study indicated that fairly typical recycled material use in highway construction 8 resulted in significant positive environmental impact reductions over the service life of the road. 9 The study evaluated two methods of data collection for the purpose of LCA. Based on the LCAs 10 results, the two data sets provided similar impacts and reductions. However, greater absolute 11 impacts were consistently predicted from the Constructed data (actual data collected during 12 construction). This is directly related to the quantity of materials found by the data collection. 13 The Constructed data was able to capture more applications of material, as well as a greater 14 variety of material types and mix designs. Although the in-depth tracking of material may have 15 resulted in more accurate environmental impact predictions, the Planned data (material quantities 16 estimated from plans and specifications) provided similar enough results to suggest that it could 17 be an acceptable method for estimating impacts should the DOT be unable to explicitly track 18 recycled material use and application. Additionally, results from two separate LCA tools 19 indicated that the environmental impacts for energy consumption and CO2 emissions were within 20 7% and 1%, respectively. Although PaLATE and SimaPro LCA results did differ, the data 21 collection method comparison revealed significant variances in LCA inputs regardless of which 22 tool was used for the impact assessment. Therefore, it is recommended that DOTs should focus 23 future efforts on material tracking for the purpose of LCAs when environmental impact issues 24 are critical. Overall, the study’s results suggest that incorporating recycled materials improved 25 the sustainability of road construction. 26 27 28

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Energy CO2 Nox SO2 CO Pb Pa,Pl,R Pa,Pl,V Pa,C,R Pa,C,V S,Pl,R S,Pl,V S,C,R S,C,V


1 ACKNOWLEDGEMENT 2 3 Funding for this project was provided by the Wisconsin Department of Transportation 4 (WisDOT) through the Construction Materials Support Center (CMSC) at the University of 5 Wisconsin-Madison. The authors thank Steve Krebs and Chris Fredrick from WisDOT and Gary 6 Whited at CMSC. Additional acknowledgements are to contractors on the project including Joe 7 Jirsa, Mary Gherke, Greg Courter, and Kevin Kaufman. Final thanks to Greg Horstmeier for 8 assistance with the initial data collection and analysis. 9


REFERENCES 1 2

1. C. J. Kibert. Policy instruments for a sustainable built environment. J. Land Use Environ. 3 Law, Vol. 17, No. 2, 2002, pp. 379–394. 4

2. P. Truitt. Potential for reducing greenhouse gas emissions in the construction sector. 5 Publication U.S. Environmental Protection Agency, Washington, DC, 2009. 6

3. J. A. Gambatese and S. Rajendran. Sustainable roadway construction: Energy 7 consumption and material waste generation of roadways. Construction Research 8 Congress, Vol. 183, 2005. 9

4. AASHTO. Primer on Transportation and Climate Change. Publication PCRT-1, 10 American Association of State Highway and Transportation Officials, Apr. 2008. 11

5. J. Lee, T. Edil, J. Tinjum, and C. Benson. Quantitative Assessment of Environmental and 12 Economic Benefits of Recycled Materials in Highway Construction. Transp. Res. Rec. J. 13 Transp. Res. Board, Vol. 2158, Dec. 2010, pp. 138–142. 14

6. T. B. Edil. Characterization of recycled materials for sustainable construction. Proc 18th 15 ICSMGE, 2013. 16

7. FHWA User Guideline for ByProducts and Secondary Use Materials in Pavement 17 Construction. Federal Highway Administration, VA, U.S.A., FHWA-RD-97-148, 2008. 18

8. W. Guthrie, D. Cooley, and D. Eggett. Effects of Reclaimed Asphalt Pavement on 19 Mechanical Properties of Base Materials. Transp. Res. Rec. J. Transp. Res. Board, Vol. 20 2005, Jan. 2007, pp. 44–52. 21

9. RMRC. Coal Fly Ash. Recycled Material Resource Center, 2010. 22 http://rmrc.wisc.edu/coal-fly-ash/. Accessed: June 17, 2015. 23

10. RMRC. Coal Bottom Ash/Boiler Slag. Recycled Material Resource Center, 2010. 24 http://rmrc.wisc.edu/ug-mat-coal-bottom-ashboiler-slag/. Accessed: June 17, 2015. 25

11. Bloom, E., Horstmeier, G., Pakes Ahlman, A., Edil, T. Urban Highway Life Cycle 26 Assessment and Data Collection Methodology. Presented at the 4th international 27 conference in Sustainable Construction Materials and Technology (SCMT4), Las Vegas, 28 NV, 2016. 29

12. Bloom, E., Horstmeier, G., Pakes Ahlman, A., Edil, T., Whited, G. Assessing the Life 30 Cycle Benefits of Recycled Material in Road Construction. Presented at Geo-Chicago 31 2016, Chicago, IL, 2016. 32

13. RMRC. Blast Furnace Slag. Recycled Material Resource Center, 2010. 33 http://rmrc.wisc.edu/ug-mat-blast-furnace-slag/. Accessed: April 22, 2016. 34

14. RMRC. Reclaimed Asphalt Pavement. Recycled Material Resource Center, 2010. 35 http://rmrc.wisc.edu/ug-mat-reclaimed-asphalt-pavement/. Accessed: April 22, 2016. 36

15. O. Reyes-Ortiz, E. Berardinelli, A. E. Alvarez, J. S. Carvajal-Muñoz, and L. G. Fuentes. 37 Evaluation of Hot Mix Asphalt Mixtures with Replacement of Aggregates by Reclaimed 38 Asphalt Pavement (RAP) Material. Procedia - Soc. Behav. Sci., Vol. 53, Oct. 2012, pp. 39 379–388. 40

16. ISO. Environmental Management - Life cycle assessment - Principles and framework. 41 Publication 14040:2006. International Organization for Standardization, 2006. 42

17. FHWA. Life Cycle Assessment of Pavements. Publication FHWA-HIF-15-001. Federal 43 Highway Administration, U.S. Department of Transportation, 2014. 44


18. WisDOT. Chapter 14: Pavement, Facilities Development Manual, 2015. 1 http://wisconsindot.gov/Pages/doing-bus/eng-consultants/cnslt-rsrces/rdwy/fdm.aspx. 2 Accessed December 9, 2015. 3

19. WisDOT. Quality management program, 2015. http://wisconsindot.gov/Pages/doing-4 bus/eng-consultants/cnslt-rsrces/qmp/default.aspx. Accessed May 10, 2016. 5

20. A. Horvath. PaLATE - Pavement Life-cycle Tool. Consortium on Green Design and 6 Manufacturing, June 5, 2007. http://www.ce.berkeley.edu/~horvath/palate.html. 7 Accessed: March 24, 2015. 8

21. M. P. Tsang, M. E. Bates, M. Madison, and I. Linkov. Benefits and Risks of Emerging 9 Technologies: Integrating Life Cycle Assessment and Decision Analysis to Assess 10 Lumber Treatment Alternatives. Environ. Sci. Technol., Vol. 48, No. 19, Oct. 2014. pp. 11 11543–11550. 12

22. I. T. Herrmann and A. Moltesen. Does it matter which Life Cycle Assessment (LCA) tool 13 you choose? – a comparative assessment of SimaPro and GaBi. J. Clean. Prod., Vol. 86, 14 Jan. 2015, pp. 163–169. 15

23. PRe Sustainability. SimaPro, the world’s leading LCA software. SimaPro, 2016. 16 https://simapro.com/. Accessed: May 10, 2016. 17

24. G. Rebitzer, T. Ekvall, R. Frischknecht, D. Hunkeler, G. Norris, T. Rydberg, W.-P. 18 Schmidt, S. Suh, B. P. Weidema, and D. W. Pennington. Life cycle assessment. Environ. 19 Int., Vol. 30, No. 5, Jul. 2004, pp. 701–720. 20

25. Bloom, E., Del Ponte, K., Madras Natarajan, B., Pakes Ahlman, A., Edil, T., Whited, G. 21 State DOT Life Cycle Benefits of Recycled Material in Road Construction. Presented at 22 Geo-Chicago 2016, Chicago, IL, 2016. 23

26. EarthShift. TRACI 2 Impact Assessment Method | EarthShift. EarthShift: Building 24 pathways to sustainability, 2016. http://www.earthshift.com/software/simapro/traci2. 25 Accessed: May 11, 2016. 26

27. M. Goedkoop, M. Oele, J. Leijting, T. Ponsioen, and E. Meijer. Introduction to LCA with 27 SimaPro. PRe Consulting, 2016. 28

28. Agri-footprint. Agri-footprint 2.0: Description of data. 2015. 29 29. E. Moreno Ruiz, T. Levova, G. Bourgault, and G. Wenet. Documentation of changes 30

implemented in ecoinvent version 3.1. ecoinvent, Zurich, 2014. 31 30. S. Suh. CEDA 4.0 User’s Guide. IERS, LLC, 2010. 32

33

LIFE CYCLE ASSESSMENT OF HIGHWAY RECONSTRUCTION: A …docs.trb.org/prp/17-02163.pdf · 1 LIFE CYCLE...

Documents

Transcript of LIFE CYCLE ASSESSMENT OF HIGHWAY RECONSTRUCTION: A …docs.trb.org/prp/17-02163.pdf · 1 LIFE CYCLE...