Numerical analysis of size reduction of municipal solid waste particles on the traveling grate of a waste-to-energy combustion chamber

Transcription

1 Numerical analysis of size reduction of municipal solid waste particles on the traveling grate of a waste-to-energy combustion chamber Masato Nakamura, Marco J. Castaldi, and Nickolas J. Themelis Earth Engineering Center and epartment of Earth and Environmental Engineering, Columbia University, 5 West 12 th St., 918 Mudd, New York, NY 127, USA Tel: , Fax: mn228@columbia.edu Abstract The size reduction of municipal solid waste () particles on the reverse acting traveling grate of a waste-to-energy (WTE) combustion chamber was estimated by means of a numerical model combining the particle size distributions (PS) of and combustion residues and the Shrinking Core Model (SCM). This new integrated model was used to simulate the particle behavior on the grate. uring their travel on the moving grate, the sizes of the particles are reduced by combustion, breakage, and compaction. This study shows the calculation of the particle size change using this model and comparison of the numerically derived PSs of and ash particles with experimental data. There is good agreement between calculated and measured values. 1. Introduction The size and volume of municipal solid waste () particles is reduced during the combustion process in the waste-to-energy (WTE) combustion chamber. As they travel over the length of the grate particles are subjected to drying, volatilization, and char oxidation and finally turn into ash (Figure 1). The Martin reverse-acting traveling grate depicted in Figure 1 controls the rate of flow and also enhances mixing. The grate operation is necessary because is much more heterogeneous than coal and other fuels. Major factors in the size reduction of particles are the physical and chemical transformations occurring on the grate, including drying, volatilization, and combustion. In addition, because of the motion of the moving bars of the Martin grate, some particles break up and their size is reduced. This breakage process can be numerically expressed as a breakage matrix based on experimental data. For example, Campbell and Webb [1] analyzed roller milling performance and developed a breakage equation. Before the enters into the inlet of the combustion chamber, the particles are accumulated in the feed hopper from the bottom of which they are pushed into the combustion chamber by the ram feeder. The resulting compaction process can be estimated by means of a ram pressure-density curve, because varies greatly in both densities and compressibility. After the drying, volatilization, and combustion processes in the chamber, particles become ash consisting mostly of noncombustible residues. The particles that are on the surface on the bed, where the temperature rises above 11 o C, are subjected to some fusion and agglomeration, and they form larger particles (clinker). ue to the motion of the traveling grate, clinker particles of ash can break again and their size is finally reduced to the PS of ash at the outlet of the combustion chamber (Figure 1). In order to simulate the size reduction of particles, we modeled the PS of particles and ash particles and applied the shrinking core model (SCM) to these distributions. Also, Image Analysis was used to determine the PS of and ash samples. Finally, thermogravimetric analysis (TGA) was applied on particles of the main constituents of in order to examine the effect of thermal decomposition on particle size.

2 particle 1. rying (moisture evaporation) particle size distribution 2. evolatilization 3. Combustion (char oxidization) particle particle size distribution Figure 1: Combustion and transport phenomena of one particle in a bed on the traveling grate 2. Modeling Particle Size istributions (PSs): PSs of and ash were described in this study by using the Gamma function distribution that is expressed as follows: f 1 α 1 β p ( ) = a e (Eq. 1) α β Γ( α) diffusion through ash layer, or chemical reaction control. g(, t) = for < < t and g(, t) = 1 - X for t < < max Therefore, the overall conversion for all particle sizes (X) is where a is constant, is the particle size, and α, β are positive parameters, respectively. Figure 2 shows the gamma distributions for different values of α but constant β. X = 1 max t 1 g(, t) a α β Γ( α) α 1 e β (Eq. 3) d Combustion of particles: The shrinking core model (SCM) is used to simulate the thermal decomposition of particles. The SCM was originally developed by Yagi and Kunii [2] for describing solid and fluid reactions for non-porous particles. Gbor and Jia developed a SCM combined with PSs [3]. The unreacted mass fraction in the core of a particle can be defined as follows: Unreacted Fraction = (, t) ) ( f ( ) ) g p ( d (Eq. 2) where g(, t) = 1 - X is the unreacted fraction and X depends on the type of the controlling regime, such as diffusion through gas film, fp() Gamma istribution α= 1 β= 5 α= 2 α= 3 α= 4 α= Figure 2: Particle size density function (Gamma distribution) for different values of α (=1, 2, 3, 4, 5) and constant β (= 5)

3 A large part of the size reduction of is due to thermal decomposition that occurs at temperatures below 4 o C. and can be expressed as follows: moisture + volatile matters + char (Eq. 4) Camera For this reaction, k (, ) = 1 = 1 rnt g t X (Eq. 5) 3 where k rm =k r, t = k rn t, and k r = k 1 C Ag / ρr. R is radius of particles (R = /2). Other possible size reduction phenomena that will be included in this model: (1) Breakage of particles: As the particles travel on the reverse acting grate, some of them break due to movement of the traveling grate and mechanical interaction with neighboring particles. For example, a glass bottle fed into the chamber may break into hundreds of small pieces. Sample Figure 3: The image analyzer system, outlining the shapes of particles regarding length, width, perimeter, area (2) Compaction of particles: The size of particles can be reduced due to compaction. When particles are fed into the chamber, the feeder compresses them and they become smaller than the particles in black bags as collected. 15 cm 5 cm 3. Experimental work Measuring and ash samples by Image Analysis: Particles of and of ash samples were measured in terms of height, width, area and perimeter. The size and surface area of each particle were measured using an image analyzer (Figure 3). There are two steps in image analysis: a) capturing particle images from the camera with color inversion and b) defining the shape of each particle with respect to length, width, perimeter, and surface area [4]. In order to measure particle sizes and shape factors from photographic images, the Able Image Analyzer version 2.1 (Mu Labs 24), one of the powerful image analysis software packages, was used. Thermogravimetric Analysis (TGA) of particles: TG Analysis was carried out paper samples. The effects of heating rate and sample size on the thermal decomposition process are shown in Figure 5 [5]. These effects depend on heat transfer through the gas film surrounding the particle into the particle. In addition to the heating rate and particle size, the shape of particle is also an important factor in thermal decomposition. 15 cm Figure 4: igital camera images of samples (left) and ash samples (right) TG % mg Temp (C) 18 mg Figure 5: Sample Size effect of pelletized Baker Cellulose, 5.8mg and 18 mg [5] 5 cm

4 a) Size istributions b) Sphericity istributions Sphericity 16 c) Aspect Aspect Ratio istributions d) Roundness istributions Aspect Ratio Roundness Figure 6: istributions by particle numbers of residential and combined ashes in NYC a) Particle size distributions (PS), b) Sphericity distributions, c) Aspect Raito distributions (AR), d) Roundness distributions 4. Results and discussion Figure 5 shows the size effect of palletized baker cellulose between a 5.8mg and 18mg. Both samples start reducing their mass around 38 o C, but the mass loss rate of the 18 mg sample is slower than that of the 5.8 mg sample. Figure 6 shows the particle size distributions (PS), sphericity distributions, aspect ratio distributions (AR) and roundness distributions of and ash. In figure 6-a, the peak of PS is around 1cm and shifts to the peak of an ash PS of around 2cm during the combustion process. Also, asymptotic tail of the PS is larger than ash PS. In Sphericity istributions (Fig 6-b) two peaks of and ash particles are located between 1.5 and 2. The two peaks range from 1.2 to 1.4 in Aspect Ratio istributions (Figure 6-c) and from 1.3 to 1.4 in Roundness istributions (Figure 6-d). As seen in Figure 6, the ash distribution has the higher peak compared to the distribution. That means the ash particles are more homogeneous in size, sphericity, aspect ratio, and roundness. Figure 7 shows simulation results of the gamma distribution. ots are experimental data and lines are the predicted distribution. The PS of has an a constant value of 2.5 x 1, an α parameter of 4, a β parameter of 3.2, a mean µ of 12.8, a standard deviation, σ, of 6.4, and a covariance, CV, of.5. The PS of ash resulted in having a constant a value of.3 x 1, an α parameter of 12, a βparameter of.17, a mean μof 2.4, a standard deviation, σ, of.5889, and a covariance, CV, of.289 (Table 1). Compared with the measured PSs shown in Fig 6-a, predicted PSs have a similar shape except for the asymptotic tail of the PS (3-4 cm). This tail group of particles contains a lot of inorganic waste that is not combustible, which is currently not well simulated in this combustion model. That is primarily because some of physical phenomena such as breakage and compaction have not been included in this combustion model yet.

5 Particle Size istribution Particle Size (cm) Figure 7: Particle size distributions (PS) by particle numbers of residential and combined ashes in NYC: (lines: estimated gamma distributions, dots: experimental data) particle size distribution particle size distribution Parameter Predicted Experimental Predicted Experimental Mean µ Standard dev. σ Covariance CV a.3 X X 1 - α β Table 1: Particle size parameters obtained from residential and combined ashes in NYC 5. Conclusions Size reduction model using a shrinking core model (SCM) and particle size distributions (PS) is developed for particles traveling WTE combustion chamber. PS and distributions of several shape factors of were measured using image analysis. Thermogravimetric analysis (TGA) is applied for decomposition of solid waste. The size and shape change of particles were compared with those of ash particles. In the future work using the size and shape information a combustion model integrated with a mixing submodel will be developed to analyze heat transfer and gas-particle reactions..

6 Acknowledgement The authors gratefully acknowledge the support of the Waste-to-Energy Research and Technology Council to graduate student Masato Nakamura. The support of this research and Mr. S, H. Lee by Covanta Energy, Montenay Power, Wheelabrator Technologies, and IWSA is gratefully acknowledged. References [1] Campbell, G., M., Webb, C. (21). On predicting roller milling performance Part I: the Breakage equation. Powder Technology 115 pp [2] Yagi, S., Kunii,. (1955). Studies on combustion of carbon particles in flames and fluidized beds. In: Fifth Symposium (International) on Combustion Reinhold, New York, pp [3] Gbor, P. K. Jia, C. Q. (24). Critical evaluation of coupling particle size distribution with the shrinking core model, Chemical Engineering Science 59 pp [4] Nakamura, M., Castaldi, M.J., and Themelis, N.J. (25). Measurement of Particle Size and Shape of New York City Municipal Solid Waste and Combustion Residues Using Image Analysis," Proc. 16th Japan Society of Waste Management Experts (JSMWE) Fall Conference, pp. 1-3, Sendai, Japan [5] Gaur, S., and Reed, T. B. (1998). Thermal ata for Natural and Synthetic Fuels, Marcel eeker, Inc.