BOTTOM ASH CHARACTERIZATION OF MUNICUPAL SOLID WASTE AND CONTAINED GLASS RECOVERY Hugo Tiago Antunes Jardim ABSTRACT Instituto Superior Técnico, Universidade de Lisboa, Lisboa, 2015 The MSW incineration is a reality in Portugal since the twenty-first century. Currently, two large industrial facilities are operating in Portugal mainland and a smaller in Madeira. The volume reduction of 90% and mass reduction of 70% turns this waste management model to be preferred for developed countries, but some disadvantages are brought. Bottom ash production is a disadvantage of this process, however its recovery and valorization might present a great environmental and economic potential. This dissertation deals with the glass recovery contained in the incineration bottom ash from the installation of Valorsul, fueled by MSW in Lisbon s region. The study began with a bibliographic search, then by the characterization of bottom ash in particle size and composition. The RecGlass equipment, a dry process used to separate particles with different shape in the size range 4-12 mm, approximately, was used in processing the samples taken, in order to obtain a final product rich in glass and with low content in stones. First, a preliminary study was undertaken on the study of the effects of the operational variables in the materials recovery in the glass product. A statistical design experimental program was then performed to identify the model that relates the glass and stones recovery with the operational variables and the use of Software Design- Expert allowed to obtain the optimal solution. The levels of the variables of the optimal solution were 20.3º angle inclination and drop height 16 cm. The optimization solution for the validation test was experimentally validated by carrying out experimental tests which lead to 56% in glass and 30% recovery of stones recovery in the glass product. Because this product has not a satisfactory quality an image study was performed to evaluate the aptitude for reprocessing. The study allowed to find out that significant percentage of glass particles were rejected when one only treatment is carried out, so, the experimental testing of reprocessing of the rejected product was re-processed and an upgrading of 7% was achieved. 1
1. INTRODUCTION In 2013, the Municipal Solid Waste (MSW) produced in Portugal was 4.362 million tons. These were subjected to the following management operations: 43% were sent to landfill, 22% to energy recovery through incineration and 17% followed to Mechanical and Biological treatment. Comparing to 2012, there was a significant decrease in the MSW landfilled, 55% to 43% (APA, 2014). The policy of the European Community (EC) as well as national legislation transposed from the European directives, consider to exist value in recycling packaging waste, as there is an increase over the demand to comply with the waste management hierarchy goals. With regard to the unsorted municipal solid waste and to achieve the established recycling targets, Portugal has several tools, which are largely the most common and used in most countries, such as Incineration, a controlled burning of waste, allowing the volume reduction of municipal wastes, the removal of pathogens and production of energy. Incinerators are included in the strategies to minimize the quantity of waste landfilled, being optimized today to maximize material and energy recovery. However, as it is fed with Mixed Municipal Solid Waste (MMSW), it produces residual streams one of which is composed mainly by inert materials that are currently being landfilled. In Portugal (continental) there are 2 incinerators located in the two largest urban areas (Lisbon and Porto). These areas cover about one fourth of the Portuguese population. In the present paper it is presented the characterization, in terms of composition and particle size distribution, of the bottom ash produced by Valorsul s incinerator which is located in the largest Portuguese urban area. Valorsul has as maximum capacity of 662,000 t / year (90% availability), and produce about 200 kg of bottom ash / ton incinerated MMSW (Valorsul, 2015). 2. EXPERIMENTAL 2.1 Sample Preparation The sample of the bottom ash from the municipal incinerator of Valorsul was collected in April and May 2015. This sample was a composite constituted by gathering 24 increments (6 per day) of the output of the bottom ash treatment plant (see figure 1a). Each increment was collected with 10 minutes interval and weighed about 240 kg. Then the composite sample was divided in a riffles divider 4 times (see figure 1b). 2
Figure 1 Bottom ash s sample collection (a) and division (b) For the characterization process, four sub-samples of BA, with near 3-4 kg each were used and special care was taken, during sample preparation and analyses, to avoid particle fragmentation during manipulation. 2.2. Particle size analysis and composition The particle size distribution was performed using a Fritsch Analysette 3 Spartan apparatus. The DIN 4188 sieving series was used with 22.4, 16.0, 11.2, 8, 5.6, 4, 2.8, 2, 1.6 and 0.710 mm mesh. The detailed characterization procedure used can be found in Dias et al (2014a). The composition was determined by manual sorting of the different materials occurring in each particle size fraction, weighing and calculation of the relative percentage of each material in each particle size fraction. In fractions with particle size bellow 4 mm, the materials could not be clearly distinguished visually. So, below this particle size the characterization was not made. The global composition of the sub-samples was calculated from the composition of the particle size fractions, considering that the 4 mm fraction was constituted by ashes. The following materials were identified: ashes, glass, stones and mineral type particles, ceramics, metals, plastics and paper and cardboard particles. A very low percentage of glass and mineral type particles were found to be meltdown leading to difficult to classify particles. These were separately classified. A very low percentage of particles could not be identified by material. 3
2.3. Recovery of glass by RecGlass device Preliminary study A preliminary study was carried out to determine a set of parameter ranges for yield optimization. The preliminary study was conducted in the RecGlass s equipment (Carvalho et al, 2015). The inclination angle range of 18-25º, drop height range 16-23 cm, belt velocity range 1.73-4.97 cm/s and horizontal position range -7-7 cm.. Experimental design and data analysis Software package, Design-Expert 9.0.3, Stat-Ease, Inc., Minneapolis, USA, was used for experimental design, data analysis and model building to study the relationship between glass and stone recovery and three operating parameters. Variance s analysis (ANOVA) was used to estimate the statistical parameters. F-test was used to estimate the significance of all terms in the polynomial equation within 95% confidence interval. Table 1 shows the 2 3 full factorial design matrix. The three selected operating parameters, inclination angle, drop height and horizontal displacement were defined as Inc., Alt. and Desl., respectively. Each parameter was coded at two levels, -1 (minimum), and +1 (maximum) (table 1). Glass recovery response was coded as ηvid and stone recovery response as ηped. Running the full complement of all possible factor combinations means that all the main and interaction effects, as well as the optimum solution for the response system can be estimated (table 2). Table 1 - Plan of experiments domain defined by the level of the parameters Manipulation Interval Factor -1 0 +1 Drop height (cm) 16 19 23 Inclination angle (⁰) 18 22 25 Horizontal Displacement (cm) -7 0 7 4
% acumulative passive Table 2 - Full factorial 23 design matrix and experiments results Run Alt. (cm) Incl. (º) Desl. (cm) ηvid (%) ηped (%) 1-1 -1-1 66 30 2-1 -1-1 65 35 3-1 1-1 17 0 4-1 1-1 11 0 5 1-1 -1 64 20 6 1-1 -1 55 16 7 1 1-1 11 0 8 1 1-1 10 0 9-1 -1 1 66 39 10-1 -1 1 64 20 11-1 1 1 20 0 12-1 1 1 23 0 13 1-1 1 53 14 14 1-1 1 55 15 15 1 1 1 10 0 16 1 1 1 7 0 17 0 0 0 46 20 18 0 0 0 42 13 19 0 0 0 44 12 3. RESULTS AND DISCUSSION 3.1. Characterization of bottom ash Figure 2 shows the particle size distribution of the subsamples. 100,00 80,00 60,00 40,00 20,00 sample 1 sample 2 sample 3 sample 4 0,00 1,00 10,00 Size(mm) Figure 2 Particle size distributions of the samples collected 5
% acumulative passive Approximately 50% of the material has a particle size higher than 4 mm. The particle size distribution in this sample is quite smooth. Figure 3 presents the composition of one of the sub-samples characterized. The other subsamples are similar. It appears that bottom ash, as expected, is a product composed mainly by ash but with a high percentage of glass (Jardim, 2015). 48% 1% 0% 0% 4% 4% 24% 15% 4% Glass (g) Ferrous Metals (g) Stone (g) Mixed glass (g) Ceramics (g) Light plastics (g) Paper and cardboard (g) Others (g) Ash (g) Figure 3 - Composition of the Valorsul s bottom ash Figure 4 shows the distribution of the main components contained in the fractions + 4 mm in the BA sample. While the glass and stone are concentrated in the fractions -8+5.6 and -11.2+8 mm the ceramics are mostly concentrated in the fractions higher + 11.2 mm. The remaining components present a more homogenously distribution. 60,00 50,00 40,00 30,00 20,00 10,00 0,00 light plastics (%) paper & cardboard (%) others (%) ferrous metals (%) ceramics (%) stones (%) glass (%) Size (mm) Figure 4 - Distribution by particle size fraction of the main components of the BA in the fractions + 4 mm 6
Figure 5 describes the distribution of the glass and stones in + 4 mm fractions per sub-sample. It is shown that -8.0, +5.6 and -11.2, + 8.0 mm fractions concentrate over 70% of glass and stones, respectively. Figure 5 - Glass and Stone distribution per fraction higher than + 4mm 3.2. Experiments results and discussion According to preliminary results, glass recovery ranged between 7 to 64%. Inclination angle, drop height and horizontal displacement were considered to be the most influent operational parameters. Table 2 shows the experiment results. In table 3 and 4 the ANOVA statistical analysis for both glass recovery response and stones recovery response, respectively, are presented. The factors angle inclination and drop height are significant factors for glass recovery. Though, for stone recovery only inclination angle appears to be significant. The values in table 3 and 4 indicate the will fitting of the experimental results to the polynomial model equations for both responses and hence accuracy of these models. ηvid (%) = 206.34 6.70Inc 1.20Alt (1) ηped (%) = 84. 55246 3. 31585Inc (2) Table 3 - ANOVA for glass recovery response Source Sum of Degree of squares freedom Mean square F-value p-value Model 9115.68 2 4557.84 170.89 <0.0001 Inc 8835.12 1 8835.12 331.26 <0.0002 Alt 280.56 1 280.56 10.52 0.0051 Residual 426.74 16 57.71 Lack of fit 346.24 6 8.05 7.17 0.0036 Pure error 80.5 10 Correction Total 9542.42 18 7
Table 4 - ANOVA for stone recovery response Source Sum of Degree of squares freedom Mean square F-value p-value Model 2161.93 1 2161.93 38.93 <0.0001 Alt 2161.93 1 2161.93 38.93 <0.0001 Residual 944.07 17 55.53 Lack of fit 669.90 7 95.7 3.49 0.0366 Pure error 274.17 10 27.42 Correction Total 3106.00 18 3.3. Optimization of the experimental parameters Table 6 shows the optimum parameters that lead to maximum recovery of glass with minimum recovery of stone (the optimization criteria). The experimental results of applying these parameter settings (Table 5) are shown in table 6 and 7. It is shown that is possible to recover 56% glass and 30% stones in the glass concentrate. Table 5 - Optimization parameters and responses for RecGlass separation process Inc (º) Alt (cm) ηvid (%) ηped(%) 20.3086 16 cm 51.0714 17.2124 Table 6 Glass recovery response of RecGlass equipment at optimum parameters settings, in 2 tests ηvid (1) (%) ηvid (2) (%) η Vid (%) σ2 60 52 56 16 Table 7 - Stone recovery response of RecGlass equipment at optimum parameters, in 2 tests ηped (1) (%) ηped (2) (%) η Ped (%) σ2 37 23 30 49 8
4. CONCLUSIONS In Portugal, the bottom ash of municipal incinerators are sent to municipal sanitary landfills. In this study, first a characterization of the BA was undertaken showing that the bottom ash is mainly composed by fine particles with size below 22 mm (mostly ash, glass and stones). Glass is the most frequent recyclable material constituting approximately 70% of the + 4 mm fraction of the sample. Separation between glass and stone particles using RecGlass equipment separation was studied. A statistical design using a full factorial 2 3 design of experiments was conducted to study the effect of main combination parameters. Two mathematical models for calculation of glass and stone recovery in the separation were suggested according to the statistical experimental design. The best combination for the RecGlass is obtained with inclination angle 20.3º, drop height 16 cm. The optimization solution for the validation experimental test had resulted in 56% glass and 30% recovery of stones recovery in the glass product. Using image analysis. It was observed that a significant percentage of glass particles in the reject product had low circularity, so a re-treatment of this product was experimentally performed. ACKNOWLEDGMENTS The authors acknowledge the financial support by Valorsul, for the samples supply. REFERENCE APA, 2013. Resíduos Urbanos - Relatório Anual, 2011, s.l.: s.n. BELO, N. 2013. Recuperação de vidro contindo no rejeitado pesado de instalações de TMB, dissertação para obtenção do grau de mestre em Engenharia do Ambiente, Lisboa, Instituto Superior Técnico - Universidade Técnica de Lisboa CARVALHO, M. T., DIAS, N., BROGUEIRA, P. 2013. Separation of glass from stones using the difference in the shape of particles - the RecGlass device. Submitted to Waste Management, aceite para publicação em 2015 CORTEZ, L. & DURÃO, F., 1982. Análises densitárias. Lisboa: IST- Laboratório de mineralurgia e planeamento. COUTINHO, M. & MATA, P. 2003. Monitorização Ambiental de uma Unidade de Incineração de Resíduos, Aveiro: Instituto do Ambiento e Desenvolvimento [Online] Disponível em: https://www.ua.pt/idad/readobject.aspx?obj=9473. Acedido em Setembro de 2015 DIAS, N., 2011. Caracterização e recuperação do vidro de embalagem contido no rejeitado de Tratamento Mecânico Biológico, dissertação para obtenção do grau de mestre em Engenharia do Ambiente, Lisboa, Instituto Superior Técnico - Universidade Técnica de Lisboa 9
FERREIRA, H. M., 2014. Recuperação de Metais Provenientes de Resíduos, dissertação para obtenção do grau de mestre em Enge FERREIRA, H. M., 2014. Recuperação de Metais Provenientes de Resíduos, dissertação para obtenção do grau de mestre em Engenharia do Geológica e de Minas, Lisboa, Instituto Superior Técnico - Universidade Técnica de Lisboa VALORSUL, 2014 Relatório de Contas 2014, [Online] Disponível em: http://www.valorsul.pt/pt/mediateca/planos-e-relat%c3%b3rios/relat%c3%b3rios/2014.aspx. Acedido em Setembro de 2015 VALORSUL. Disponível em: http://www.valorsul.pt. Acedido entre Junho e Outubro de 2015 10