Accurate Analysis of Boiler Efficiency

100 VGB PowerTech 12/2008

Accuracy Improvement Analysis of the Standard Indirect Method for Determining a Steam Boilers EfficiencyAndrej Senegacnik, Igor Kutrin and Mihael Sekavcnik

Authors

Dr. Andrej Senegacnik, Assistant Professor

Dr. Igor Kutrin, Teaching Assistant

Dr. Mihael Sekavcnik, Assistant Professor

University of Ljubljana Faculty of Mechanical Engineering Ljubljana/Slovenia.

Kurzfassung

Verbesserte Methoden zur Bestimmung des

Keselwirkungsgrades

Der Kesselwirkungsgrad von Kohlekraftwerken wird gewhnlich nach standardisierten Verfah-ren bestimmt. Bei Einsatz eines Festbrenn-stoffs, wird der Massenstrom indirekt nach dem in DIN 12952-15 beschriebenen Verfah-ren berechnet. In dem vorliegenden Beitrag werden bestimmte Vernderungen an der standardisierten Methode DIN 12952-15 hin-sichtlich der Konsistenz von Berechnungen und die Genauigkeit der Ergebnisse mit der Sensitivittsanalyse vorgeschlagen, prsentiert und bewertet,. Die wesentlichen Berechnungs-verfahren bleiben identisch mit DIN 12952-15. Es werden sechs Vernderungen bercksich-tigt. Vier davon beeinflussen die Bestimmung der entsprechenden Rauchgasmasse, nm-lich: nicht Vernachlssigung der partiellen Oxi-dation des Kohlenstoffes zu CO, nicht Ver-nachlssigung von Unverbranntem in der Asche und Schlacke bei der Bestimmung des Luftberschusses mit zwei Modellen und nicht Vernachlssigung von Restwasser in der Rauchgasmessprobe. Die fnfte Vernderung behandelt den Einfluss der kalten Falschluft und die sechste Vernderung bercksichtigt die Verwendung der Berechnungsmethode der Enthalpie fr Rauchgas und Verbrennungs-luft nach VDI 4670. Der Einfluss der vorge-schlagenen Vernderungen auf den Kesselwir-kungsgrad wurde mit zwei realen Datenreihen fr Stein- und Braunkohle bewertet.

Introduction

The efficiency of coal-fired power plants is usually determined according to standardised procedures. Currently, the Guidelines VDI 3986 and standard DIN EN 12952-15 [1] are valid. If solid fuel is used, its mass flow is cal-culated indirectly according to the procedures stated in DIN 12952-15. The indirect method requires the determination of boiler losses. Losses can be measured directly except for radiation and convection losses which are es-timated according to the standard procedure stated in DIN 12952-15. The article at hand proposes, presents and evaluates some modi-fications of the standardised indirect method DIN 12952-15 regarding the consistency of computations and accuracy of the results. Most of the modifications were introduced re-cently in reference [2] where they were treated generally, i.e. from a stoichiometric calcula-tion point of view. Additional research pre-sented in this paper is focused on incorporat-ing additional calculations into the standard procedure [3] and to evaluate their impacts using sensitivity analysis. The main flow of computations remains identical to DIN 12952-15. Six modifications are considered. Four of them affect the referred flue gas mass deter-mination, i.e.: partial oxidation of carbon to CO, including unburned matter in ash and slag while determining the excess air ratio with two models and including residual water vapour content in the flue gas sample. The fifth modification studies the impact of cold uncontrolled leakage air. The sixth modifica-tion is applying the VDI 4670 procedure [4] for flue gas and combustion air enthalpies cal-culation. The impacts of modifications were studied for two real data sets for hard coal and lignite.

Air Ratio Determination

Par t i a l Ox ida t ion o f Carbon t o CO ( I )

If some of the carbon from fuel burns to CO instead to CO2 the actual oxygen consump-tion is lower than the calculated consumption according to DIN 12952-15. To evaluate the

CO impact, the carbon from fuel should be divided into quantities of carbon combusted to CO2 and to CO, respectively.

Let the factor CO2 be a quantity of carbon forming CO2. The factor CO2 can be calculat-ed exactly with the next derivative formulae:

CO2d1,8650 C VAod CO2Ad COdCO2 = (1)

CO2dC 1,8534 + 1,865 COdFactor CO2 is dependent on , COd and CO2d (if the O2d is measured in the flue gas, the CO2d value needs to be calculated with the formulae (6) and (7)).

Lambda , is calculated from the basic pa-rameters which are given in DIN 12952-15 and are complemented with the carbon mon-oxide (CO) coefficients which are calculated according to [5]. The complemented equa-tions are:

Aod = 11,5122 CO2 C + 5,7561 (1 CO2) C + 34,2974 H + 4,3129 (2) S 4,3129 O

God = 12,5122 CO2 C + 6,7561 (1 CO2) C + 26,3604 H + 5,3129 (3) S 3,3129 O + N

VGod = 8,8930 CO2 C + 5,3847 (1 CO2) C + 20,9724 H + 3,3190 (4) S 2,6424 O + 0,7997 N

CO2 = 3,6699 CO2 C + 0,0029 (1 CO2) C + 0,0173 H + 0,0022 (S O) (5)

If CO2d is measured is calculated according to: VGod CO2d CO2d = +1 (6) VAod CO2d CO2Ad if yO2d is measured is calculated according to: VGod O2d = +1 (7) VAod O2Ad O2d The upper formulae for are given in [5].Iteration is needed since factors CO2 and are correlated. The CO impact on referred flue-gas mass G is always negative. This means that in presence of CO in flue gas, actual flue gas mass flow is lower than

Accurate Analysis of Boiler Efficiency

VGB PowerTech 12/2008 101

flue gas mass flow calculated according to DIN 12952-15. F i g u r e 1 shows the impact of CO content in dry flue gas on referred flue gas mass G. Actual operating points are also marked. The CO impact on G is negative. If the CO content would rise to 1000 ppm, the difference of referred flue gas mass G would be in the case of hard coal 0.28 % and in the case of lignite 0.22 %. The gradient of CO impact is slightly higher in the case of hard coal due to the larger mass fraction of carbon in coal.

Unbur ned Combus t ibl e s i n Ash and S lag

The air ratio is calculated on the basis of the measured flue gas composition at the boilers outlet and the fuel composition. Standard pro-cedure for the flue gas mass flow calculation assumes that all the carbon from the burned fuel is burned completely and that the compo-sition of the burned fuel matches the com-position of the supplied fuel i.e. fuel sample. In practice, some of the larger coal dust parti-cles do not burn completely; they leave the boiler with ash as coke. Therefore, the un-burned carbon does not participate in CO2 formation. Actual oxygen consumption is therefore lower. In order to determine the ex-cess air ratio more accurately, unburned car-bon should be subtracted from the carbon mass fraction in the fuel and added to the ash mass fraction. The air ratio calculation should then be repeated using the new apparent fuel composition. Two explanatory models are in-troduced in the following sections.

In the following paragraphs annotations coalX means composition of the fuel sam-ple while X means apparent fuel composi-tion (i.e. composition that actually burned).

Model 1 Unburned Combustible Matter in Ash and Slag is Pure Carbon (II)

Model 1: The assumption is that the combus-tible matter in ash and slag is pure carbon. New apparent fuel composition, which ac-tually burns, is therefore:

C = coalC lu (1 coalAsh coalH2O) (8)

H = coalH (9)

S = coalS (10)

O = coalO (11)

N = coalN (12)

Ash = coalAsh + lu (1 coalAsh coalH2O) (13)

H2O = coalH2O (14)

Model 2 Composition of Unburned Combustible Matter in Ash and Slag Matches the Combustible Component of Fuel Linear Model (III)

Model 2: The assumption is that the combus-tible matter in ash and slag has the same com-

position as the combustible component of the fuel. New apparent fuel composition is therefore:

C = (1 lu) coalC (15)

H = (1 lu) coalH (16)

S = (1 lu) coalS (17)

O = (1 lu) coalO (18)

N = (1 lu) coalN (19)

Ash = coalAsh + lu (coalC + coalH + coalS + coalO + coalN) (20)

Sensitivity Analysis

F i g u r e 2 shows the impact of unburned combustibles in ash and slag to referred flue gas mass G. The x-axis of Figure 2 represents the amount of unburned combustible in ash because there were no unburned combustibles in slag. To get a more objective impression about the impact of unburned combustibles in ash and slag, the total amount of unburned combustible can be also expressed as a mass fraction of carbon in the supplied fuel [6]. In-dex unburned carbon is calculated by expres-sion (21).unburned carbon

mSL SL + mFA FA= 100 % (21) mFo CThe impact of unburned combustibles in ash and slag on referred flue gas mass G is al-ways negative. This means that in a presence of unburned combustibles in ash and slag, ac-tual flue gas mass flow is always lower than flue gas mass flow calculated according to DIN 12952-15.

The results of Model 1 and Model 2 are pre-sented in Figure 2. The proposed models are

extreme cases. The actual impact lies some-where in between. From the results we can conclude that the average impact gradient of unburned combustibles to referred flue gas mass is

hard coal: 0.87 % of referred flue-gas mass per 1 % of unburned carbon

lignite: 0.71 % of referred flue-gas mass per 1 % of unburned carbon

The impact is larger for hard coal due to a higher mass fraction of carbon in the fuel. In the case of lignite, when the ratio C/H is lower, the gap between pure carbon and the linear model becomes wider.

Res idua l Vapour Con ten t i n Gas Sample ( IV)

The standard procedure for the flue gas mass flow calculation requires a dry