Article pubs.acs.org/EF
Slagging and Fouling during Coal and Biomass Cofiring: Chemical Equilibrium Model Applied to FBC Paula Teixeira,*,† Helena Lopes,† Ibrahim Gulyurtlu,† Nuno Lapa,‡ and Pedro Abelha† †
LNEG, Estrada do Paço do Lumiar, 22, Ed. J, 1649-038 Lisboa, Portugal UNL-FCT, Quinta da Torre, 2829-516 Caparica, Portugal
‡
S Supporting Information *
ABSTRACT: A thermodynamic model was applied to foresee the occurrence of fouling, slagging, and bed agglomeration phenomena during fluidized bed monocombustion of three different types of biomass, namely straw pellets, olive cake, and wood pellets. The cocombustion effect in reducing the occurrence of deposits and agglomerates of blends of 5, 15, and 25% (wt.) biomass with coal was also assessed. Chemical fractionation was applied to evaluate the reactive and nonreactive fraction of elements in the fuels, which was used to estimate their partition between the freeboard and bottom zone of the boiler. Qualitative and semiquantitative analytical techniques, namely, X-ray diffraction and scanning electronic microscopy − energy dispersive spectroscopy were used to compare the results from the simulation with the mineralogical and morphological composition of ash and deposits formed during combustion. The thermodynamic modeling revealed to be a powerful tool in foreseeing the formation of melt and liquids salts, depending on the temperature and chemical composition of fuels. The main discrepancies observed between the experimental and simulated data were due to particularities of the combustion process, which are not incorporated in the software, namely, kinetic limitations of the reactions, possible occurrence of secondary reactions in the ashes, and elutriation effects of ash and silica sand particles.
1. INTRODUCTION Biomass thermochemical conversion through combustion (and cocombustion) has socioeconomic and environmental benefits; however, problems associated with its use are also documented.1 Ash related problems like deposit formation in the convective heat transfer zones (fouling) and refractory zones of the boiler (slagging), agglomeration (in case of the fluidized bed combustion), and corrosion are usually the principal causes for the malfunctioning of the combustion systems. According to the actual knowledge, the ash problems associated with biomass combustion are mainly related to the quantity, reactivity, and interaction of Si, K, Ca, Cl, and S. Na and Mg have similar behavior to K and Ca, respectively. Nevertheless these elements exist usually in minor quantities in biomass. K in the presence of Cl, S, and Si undergoes many undesirable reactions during combustion, e.g. alkali reactions with Si forming alkali silicates (which melt or soften at low temperatures), and reactions of alkali with S and Cl forming alkali sulfates and chlorides, which deposit inside the combustor and heat exchanger surfaces.2 Ca may affect the equilibrium reactions binding to Si as calcium silicates, decreasing potassium silicates formation and increasing potassium salts.3,4 The physical properties and the inorganic composition of the different types of biomass could vary significantly, turning difficult the prevision of biomass ash behavior during combustion. Usually, herbaceous biomass ash has high Si and K content and woody biomass ash has high Ca content, which may justify the different behavior of the ashes formed during combustion. Soil contamination is another problem5 contributing to the ash formation. Possible operational measures to decrease problems related to biomass ashes produced in fluidized beds combustors (FBC) © 2013 American Chemical Society
are as follows: 1) decrease process temperatures and avoid hot spots; 2) fuel refinement (e.g., through leaching, thus extracting Na and K); 3) cofiring with fuels having less problematic ashes; 4) modify the composition of bed material and their size distribution and 5) use of additives to modify ash behavior.3 The cofiring of biomass with coals allows diluting problematic elements and modifies ash composition. Besides dilution effect, the reactivity of inorganic elements has also an important role in the compounds formation. Some reactions are favored relatively to others, influencing the mineralogical composition of ashes and the salts formation and consequently the ash melting temperatures. The interactions between Al, Si, and S from coal and the alkali elements from biomass may interfere with the properties of fly ash and ash deposits.6,7 During cofiring of coal with biomass the formation of alkali-aluminasilicates is thermodynamically favored, which promotes the alkali sequestration in bottom ashes and its decrease in the gaseous stream. The alkali-alumina-silicates remaining on the bottom ashes have relatively high melting temperatures, decreasing slag, and sintering of the ash and bed material. The ash deposition in the convective zones of the boilers and surface corrosion may also be reduced because less alkali chlorides are volatized.8 The reaction of S from coals with alkali species from biomass may also have an important role in preventing fouling and corrosion, because K2SO4 is less corrosive than KCl.6 Most of the interactions between the elements during coal and biomass cofiring have been identified. Nevertheless, the Received: September 9, 2013 Revised: December 2, 2013 Published: December 5, 2013 697
dx.doi.org/10.1021/ef4018114 | Energy Fuels 2014, 28, 697−713
Energy & Fuels
Article
most promissory software is Factsage, which is a tool that can perform a wide range of thermochemical calculations. It is a combination of two linked models in thermochemical, namely FACT-Win and ChemSage that makes calculations based upon the concept of Gibbs free energy minimization. Assuming that chemical equilibrium is attained, it provides information about the phases formed, their proportions and compositions, the activities of individual chemical components, and the thermodynamic properties for a wide range of pressures and temperatures.6,7,30,31This tool may be useful to predict the fouling tendency based on the condensed phases and the melt salts formed along the convective zone of the combustion systems, when the gaseous compounds cools down. It can be equally useful to predict the slagging and agglomeration tendency based on oxide slag formed in the radiant zone of combustion systems, as a function of temperature, pressure, and elemental composition. 1.2.1. Limitations. Adequate reproduction of the installation behavior during combustion processes depends of the input data and the selection of databases. Critical decisions have to be made before the thermodynamic calculations, and a detailed knowledge about the chemical system is required for a better employment of this tool. The main limitations or risks during the use of equilibrium thermochemical tools include the following:3,5,6,28,31−35 1) During combustion not all the reactants reach equilibrium at the same time, the chemical equilibrium depends on reactants mixture degree and time of residence. If the calculations include the formation of compounds/phases thermodynamically very stable, which could be kinetically prohibited within the relevant time scale of the considered process, this will inevitably lead to misleading results. 2) Variations in the composition of the fuel and in the temperature gradients, which occur inevitably in the boiler, are not considered. This effect is special critical in the case of biomass due to its high quantity of volatile matter. 3) Only chemical reactions are considered. Physical processes, such as particle nucleation, agglomeration, and adsorption in the gas are not taken into consideration. 4) Many mechanisms of deposit formation are also highly specie-specific, resulting in deposit compositions that are not easily related to fuel ash composition. 5) If the selected databases do not contain all possible species for given conditions like, pressure, temperature, and input species, the software inevitably leads to misleading results. 6) Elutriation problems often associated with FBC are not considered; nevertheless they are frequently responsible for ash particles transport, affecting particles distribution, and the type of possible reactions. 1.2.2. Modeling Approaches. Several chemical equilibrium models have been developed recently, which evidence the efforts employed to minimize the inherent limitations of this tool. Based on inorganic composition and on fuel compounds reactivity, the developed models group the elements in a way that favors gas phase reactions or gas−solid reactions. However, due to the involved assumptions they still present some weaknesses. In 1998, based on quantitative elemental analysis of laboratory ashes, Skrifvars et al.22 recalculated the components using stoichiometric assumptions and thermodynamic considerations. The calculations assumed that all phosphor was present as calcium phosphate, analyzed carbonate was present as both potassium and calcium carbonate, the rest of calcium as oxide, all chlorine as potassium
effect of coal replacement by different types and quantities of biomass in ash and gas composition is not well established. The effect of biomass composition and reactivity of ash forming elements also needs a better evaluation. Moreover, the thermal behavior of inorganic elements during combustion depends also on other factors like temperature, air/ fuel mixing degree, and residence time. Due to the limited residence times of gases inside the boilers, usually lower than 3−6 s, the reactivity of elements is extremely important and may dictate the occurrence of the favorable reactions. The reactivity of inorganic elements in fuels is influenced by their chemical associations in the fuel matrix.7,9,10 Based on this, the chemical fractionation is an analytical methodology that allows distinguishing the elements into more or less reactive forms. The knowledge of the elements reactivity in fuels helps on predicting, in a more realistic way, the reactions that may occur during the combustion. Due to the elements reactivity dependence on temperature, which is influenced by the fuel volatile matter content of fuels, some considerations must be performed about the temperature profile in the boiler during biomass combustion. Ideally, during combustion on FBC systems a uniform temperature should be maintained on the bed zone, most of the volatiles should be burned in the bed zone and hot spots in the bed should be avoided. However, during combustion of high volatile content of fuels, as in the case of biomass, some difficulties related to the homogeneity of temperatures may occur, especially if the installation is not designed for biomass combustion.11,12 Even at optimum combustion conditions, not all the volatiles will burn out in the bed zone, and some of them will be burned in the freeboard, which means that in case of adiabatic furnace walls the temperature of combustion above the bed zone may exceed the bed temperature by up to 200 °C.12−14 1.1. Ash Related Problems Prediction. Along time several methodologies were developed to predict bed material agglomeration and deposits formation during combustion. Indices based on the chemical composition or ash fusion tests,15−17 indices based on the chemical composition of biomass,18,19 utilization of ternary phase diagrams,2 or the empiric evaluation of some elements known for its problematic behavior during combustion20 are some of the methodologies often applied to predict the ash related problems. Thermomechanical analysis,21,22 simultaneous thermal analysis,23 and the controlled fluidized bed agglomeration method2,24−26 are other approaches described in the literature. In the past decade a significant increase was observed in the use of chemical equilibrium tools16,27,28 to predict the formation of problematic compounds during combustion. The use of tools to predict the ash related problems during combustion are useful in many aspects; however, due to the diversity of factors that may influence the fuel behavior during combustion an accurate prediction of these problems is always limited. The physical-chemical characteristics of the fuels, the technology, and the operational parameters are some of the factors that may influence the fuel behavior during the combustion, and it is not possible to include all of them on a single tool. 1.2. Chemical Equilibrium Tool. The progress in chemical thermodynamic and viscosity models of oxide systems, the development of computational methods, and the upgrade of computer software and hardware improved the prediction, with an increasing accuracy, of the equilibrium phases in complex multicomponent systems.30 In thermochemistry, probably the 698
dx.doi.org/10.1021/ef4018114 | Energy Fuels 2014, 28, 697−713
Energy & Fuels
Article
bottom ashes agglomerates together with the components of a typical combustion atmosphere, as input for a qualitative interpretation of data,39 or the use of the amount and composition of the predicted melt phase as an input into other software to predict viscosity of the melt phase.40 1.2.3. Output Interpretation. The output data are directly dependent on the species, temperature, pressure, and database initially selected. The products of reactions may include compounds with different physical characteristics, gaseous compounds, pure solids, pure liquids, and solution species (ex. solid solution, slag-liq, and salts-liq). Relatively to ash related problems, the output results may usually be evaluated in terms of problematic compounds quantity, physical state, and composition. It is common to evaluate the amount of melt present in the ash formed by condensation, i.e., in ashes formed when the combustion gases cools down and condense, forming solid or melt compounds. The formation of deposits on a surface implies that the condensed ash particles should contain a certain amount of liquid melt. Usually, the fraction of melt for each temperature is represented by Tx, where T is the temperature and x is the melt percentage. Supported in experimental data with black liqueur,41 it was postulated that all the ash with a melt fraction between 15% (T15) and 70% (T70) are sticky and thus may be accumulated on the heat transfer surfaces, contributing to deposit formation. These limits are only valid for pure alkali melts and only partly valid for alkali earth melts, i.e., they may be applied for simple ionic salts, but for deposits containing silicon, leading to viscous melts, another criterion is required. The viscosity of a melt could be a criterion for stickiness.35,41 For melt fractions higher than 70% the opinions differ. Backman cited by Zevenhoven et al.41consider that in this case the ash will flow from the surface without causing additional slagging or fouling. Plaza et al.29 state that when more than 70% of the deposit is liquid all new particles transported to the molten surface flow off and cause severe slagging. Usually, due to the elevated temperatures in the flame zone, the agglomeration and slagging problems are not related to the liquid salts that evaporate but with silicates presence in the bottom ashes. When silica is available in ashes, especially as amorphous silica from biomass, highly viscous melts may be formed. However, the melting temperature of silicates is influenced mainly by Si, K, and Ca proportions. The phase diagram of ternary system K2O-CaO-Si2O evidenced that an increase of K content relatively to Ca implies a formation of eutectic compounds with lower melting temperatures.42 The aim of this paper is to apply and validate a chemical equilibrium modeling to predict the tendency for deposit formation in a pilot bubbling fluidized bed combustor. The elements present in fuel were distributed between the refractory section of the boiler and the convective section according to the chemical fractionation method. In a previous paper, the suitability of chemical fractionation to classify the reactivity of elements and to predict its distribution in the fluidized bed had been evaluated.10
chloride, sulfur as both sodium and potassium sulfate, the rest was potassium as oxide, all silicon as oxide, the rest of the analyzed elements assumed as their oxides. The ash components were divided into three groups, simple alkali melts, silicates, and rest, presenting different melting behavior. The amount of melt present at different temperatures was estimated for silicates and simple alkali melts.22 Two years later Zevenhoven-Onderwater et al.16 used a methodology which allowed the evaluation of the chemical association of the elements that constitutes the fuel. The chemical fractionation consists in the consecutive use of three increasingly strong solvents which allows to distinguish the reactive elements (ionic salts and bounded to the organic matrix) from the elements that are not in a so reactive form (included and excluded elements). The reactive and less reactive elements are assumed to constitute the fine and the coarse ash fractions, respectively. Assuming chemical equilibrium, the interaction of the two ash fractions with the combustion gases formed could be modeled.16,35−37 This model minimizes the first limitation mentioned before. On the other hand, it does not consider the interaction between the reactive and the less reactive fractions. The assumption that all reactive ash forming elements leave the bed could lead to an overestimation of the amount of deposit formation due to the neglected interaction of reactive species within the bed zone. Nutalapati et al.28 and Plaza et al.29 used chemical fractionation to distinguish the inorganic part going into reactive and less reactive fractions. Because this methodology does not take into consideration the reactions between the two fractions, it was assumed that a percentage (values between 5 to 25% were evaluated) of the nonreactive fraction reacts with the reactive fraction. This second approach was supported by laboratory investigations during coal/biomass combustion, where it was evidenced that the alkali ash compounds, which vaporize during combustion, could interact with the surface of nonreactive silica particles. This gives rise to alkali silicates formation which melts at low temperature and contributes to the melt phase that occurs in boiler. The model developed by Nutalapati et al.28 assume that all the ash particles are spherical and have the same diameter (10 μm); the thickness of the reacting layer was about 0.3 μm; the same proportion of particles react at all temperatures; the particle temperature is the same as the gas temperature; there are no differences between the included and excluded minerals. The authors concluded that if a percentage of nonreactive elements was considered jointly with the reactive fraction, relevant differences would be evidenced in potassium sequestration by bottom ashes and condensed phase. An increase of nonreactive fraction implies a higher retention of potassium on bottom ashes. Korbee et al.38 suggested a more effective technique to predict the chemical association between the fuel elements, which was based in a combination of pH extraction tests. The maximum leached quantity was used as an input into the chemical equilibrium model to determine the most probable mineral elements speciation for the biomass used. For example, for elements presenting constant leaching behavior at all pH values, it was most likely that their speciation would mostly consist of salts or free ions with a small amount of dissolved organics. On the other hand, for elements that had optimum leaching range at the acidic pH regions such as Ca, Mg, Al, and Si, it was likely that a larger speciation of either minerals or solid organics exists.7 Other possible applications of chemical equilibrium included the use of the main elements constituting
2. EXPERIMENTAL SECTION The evaluation of the effect of partial substitution of coal by biomass was performed by addition of increasing proportions of biomass, namely, 5%, 15%, and 25% (wt.). For comparative reasons and for identification of problematic biomass types, tests with biomass alone were also performed. Three different types of biomass were used: straw pellets (SP), olive cake (OC), and wood pellets (WP). The coals used were from Europe and South America, namely, Polish coal (PC) 699
dx.doi.org/10.1021/ef4018114 | Energy Fuels 2014, 28, 697−713
Energy & Fuels
Article
Table 1. Proximate Analysis, Heating Values, Ash Fusibility of Fuels, and Respective Relative Uncertainty (95% Confidence) Polish coal
Colombian coal
Ù coal (%) (K = 2)
straw pellets
olive cake
wood pellets
Ù biomass (%) (K = 2)
2.1 6.2 32.2 61.6 28397
9.3 9.2 37.5 53.3 27028
5 0.5 1 6 5
10.6 5.8 76.6 17.6 16590
7.9 4.9 76.7 18.4 18857
8.4 0.4 86.2 13.4 18808
11 4 1 12 6
1223 1233 1251 1284
1202 1358 1397 1443
21 5 5 7
819 1014 1167 1238
751 830 1367 1386
1238 1265 1282 1291
7 3 0.4 3
moisture (a.r., wt %) ash (d.b., wt %)a volatile matter (d.b., wt %) fixed carbon (d.b., wt %) low heating value (d.b., KJ/kg) Ash Fusibility (oxidant atmosphere) initial deformation temp (°C) softening temp (°C) hemispherical temp (°C) fluid temp (°C) a
For biomass samples a temperature of 550 °C (CEN/TS 14774) was used. a.r.: as received; d.b.: dry base.
Table 2. Elemental Analysis of Fuels and Respective Relative Uncertainty (95% Confidence)
a
elemental analysis (d.b., wt %)
Polish coal
Colombian coal
Ù coal (%) (K = 2)
straw pellets
olive cake
wood pellets
Ù biomass (%) (K = 2)
C H N S Cl Al Fe Ca K Na Mg Si P Ti
71.0 4.9 1.2 0.51 0.255 0.694 0.307 0.513 0.104 0.045 0.222 1.011 0.012 0.028
66.2 5.9 1.4 0.65 0.070 1.108 0.453 0.171 0.208 0.011 0.132 2.674 0.006 0.036
1 1 21 3 13 14 12 14 12 15 15 20 18 3
46.7 7.0 0.7 0.14 0.270 0.012 0.010 0.359 1.306 0.029 0.077 1.258 0.080
