Elbistan linyitinin oluşturduğu kirleticilerin akışkan yataklı bir sistemde kontrolü

Abstract

Akışkan yatak teknolojisi, dünyada uzun yıllardan beri bilinen, ülkemizde de çeşitli çapta uygulamalara başlanılan temiz ve verimli bir yakma sistemi olarak kabul edilmektedir. Ülkemizin en büyük rezervine sahip Elbistan linyitinin düşük ısıl değer, yüksek kül ve nem içeriğine karşın akışkan yataklı sistemlerde değerlendirilebileceği düşünülmektedir. Bu çalışmanın amacı ortalama 4600 kj/kg ısıl değere sahip, Elbistan linyitinin 0.2 m çapındaki bir akışkan yatakta yakılarak hava kirletici yayınımların sürekli olarak izlenmesi ve optimum Ca/S mol oranının saptanarak önceden belirlenen sabit sıcaklık, hava hızı ve linyit besleme debisinde SO2 ve NOx yayınımlarının azaltılmasıdır. Yatak malzemesi olarak 3xl0~4-6xl0~4 m Ve 6x1 0~4- 10"-* m tane boyutlarındaki kuartz kum kullanılarak 10-3-2xl0-3, 2xl0"3-3.33xl0-3 ve 4xl0~3- 5x1 0"3 m tane boyutlarındaki linyit numuneleri yakılmış ve yatak içi sıcaklık dağılımları incelenmiştir. Beklenenin aksine her üç fraksiyonun da 1123 K sabit sıcaklıkta oldukça kararlı yanma gösterdikleri görülmüştür. Yapılan bir dizi karakterizasyon çalışması ile mevcut olan kireçtaşı, iki farklı dolomit ve trona numunelerinin yoğunluk ve poroziteleri incelenmiş ve termogravimetrik analizler ile de sorbentlerin oksitlere dönüşüm sıcaklıkları saptanmıştır. Tane boyutuna bağlı olmakla birlikte akışkan yatakta kalma sürelerinin de etkin olduğu kireçtaşı ve dolomitlerin SO2 sıyırma etkinliklerinin benzer olduğu gözlenmiştir. Bu durum sorbentlerin yatak içi etkileşim ve ısıl şok sonucu ortaya çıkan ufalanma ile de açıklanabilmektedir. Elbistan linyitinin bünyesindeki CaO içeriğinin yüksek olması SO2 sıyırma açısından olumlu sonuçlar vermiştir. Stokiometrik Ca/S mol oranının SO2 sıyırma üzerindeki etkisi düşük mol oranlan için az etkin, yüksek oranlarda ise daha kararlı olmuştur. Trona kullanımının kireçtaşı ile birlikte Na/Ca oranlarının ayarlanarak yapılması durumunda verimli olduğu, buna karşın tek başına trona kullanımının aglomerasyon riskini ortaya çıkardığı görülmüştür. Çalışmalar sonucunda etkin bir sıyırma için kararlı yanma şartlan ile birlikte sorbentin fiziksel ve kimyasal özelliklerinin de etkin olduğu saptanmıştır.

Elbistan lignite with 3 billion tons of visible reserves is the largest open mine coal sector in Turkey. Its low calorific value, high moisture and ash content made it quite difficult to obtain beneficial results in the utilisation side. The aim of this study was to investigate the steady-state combustion and pollution characteristics of Elbistan lignite in a fluidised bed combustor. Elbistan lignite referred as El was provided by the Turkish Coal Board, South Eastern Regional Office in K.Maraş, Elbistan. 8 tons of low quality lignite sample was collected from the mixed coal which was used by Turkish Electricity Board (TEK) power plant. The main reason to carry out a series of experiments was to monitor continously the emissions which occurred during the combustion of this lignite. It was also aimed to make continous addition of different sorbents to the bed in order to reduce SO2 and NOx emissions. The fluidised bed combustor used in this work was made of stainless steel tube of 0.2 m diameter and 1 m. high. A 310 stainless steel distributor plate was designed to maintain proper fluidisation and mixing of particles. The experimental rig used in this study consisted of a fluidised bed, a preheating system, lignite and sorbent feeding systems including inverters, pressure transducer, a draft fan and two cyclones. All of these components were fully controlled by means of a data acquisition and control system communicated with a computer. A computer programme was developed in order to monitor and store the data on coal feeding rate, sorbent feeding rate, air feeding rate, bed and distributor plate pressure drops, sulphur dioxide, nitrogen oxide, carbon monoxide emissions and oxygen percentage. These parameters were continously visualised from the monitor, printed out and even plotted. A continous emission monitoring system was used to determine the pollutants. The gas samples were continously taken from the exit on the vertical line above the second cyclone. vu The regression analysis results indicated that the development of a model for the sorption of SO2 emissions using different sorbents is quite difficult. This is mainly due to the variety of sorbents, lignite and sorbent particle sizes and Ca/S molar ratio. The lignite particle size is an important factor since its ignition and combustion time in the bed differred. The combustion of small lignite particle sizes takes place instantaneously whereas that of the bigger particles combustion proceeds longer. Consequently, the smaller the lignite particle size, the more SO2 emits. XI The effect of particle size, bed material and temperature profiles were determined in order to establish steady-state combustion conditions. The lignite samples used in this study, prior to experiments, were crushed and sieved to 10_3-2xl0-3, 2xl0"3-3. 33x10'^ and 4xl0-3-5xl0"3 m fractions and designated with code numbers Nl, N2 and N3 respectively. The proximate analysis of Nl, N2 and N3 fractions is given in Table- 1. The calorific value of a representative sample is around 4600 kJ/kg. Quartz sand used as bed material and sorbents such as limestone, dolomites ( two types referred as Dl and D2) and a nahcolite (trona mineral) were used as sorbents to capture SO2. These materials were collected from different regions and were crushed and sieved to 6xl0"4-10-3 and 10~3-1.25xl0~3 m fractions. Table- 1- The proximate analysis of three different particle sizes of El lignite. A series of characterisation work was carried out in order to characterize lignite, sorbents and bed material samples. A mercury porosimeter, a thermal analyser and a scanning electron microscope with EDX facility were used for the characterisation work. The porosimeter results showed that the original limestone and dolomites had small surface areas before calcination at 1123 K (850 °C). Therefore, after calcination, total surface area of limestone sample slightly increased whereas total surface area of dolomites showed considerable augmentation. Active soda, obtained by the calcination of trona at 523 K (250 °C), had a low surface area after the operation temperature at 1073 K (800 °C) due to pore closure. The maximum pore size was obtained between 423 and 523 K (150-250 °C). It was therefore difficult to obtain promising results by using active soda as a sorbent in a fluidised bed combustor. The addition of limestone or high CaO content of lignites could be a relief for trona mineral in order to prevent the fluidised bed from agglomeration phenomena. The Na content of the lignite sample is accepted to be a key factor for agglomeration. Hence, the high percentage of Na in trona mineral would have a hindering effect on the behaviour of the bed towards agglomeration. Thermogravimetric analyses were carried out in a Rigaku thermal analyser. Limestone sample gave a peak at 1099 K (826 °C) due to heat loss with CO2 release. Dl, dolomite originally different from D2, showed two peaks, one at 1043 K (770 °C) for the conversion of MgC03 to MgO and the other at 1101 K (828 °C) for the conversion of CaCÛ3 to CaO. A third peak occurred at 1 122 K (849 °C). At similar conditions, D2 dolomite which was originally soft compare to Dl dolomite presented V1I1 rigid. 4x1 0"3 -5x1 0"3 m, N3 fraction was the best performed lignite sample as far as the SO2 emissions were concerned. For Dl dolomite, Ca/S molar ratio of 3.5 was sufficient in order to obtain a SO2 removal efficiency over 80 % while a Ca/S ratio of 5 and greater was needed for D2 dolomite. NOx emissions increased with the decrease in SO2 emissions. This conclusion is compatible with the findings of some researchers. Despite its low calorific value, high moisture, ash and comparably high sulphur content, Elbistan lignite can be burned quite satisfactorily in a fluidised bed combustor. The high CaO content reduces the SO2 emission and consequently reduces the amount of sorbent needed to meet the standards. The tendency of the fluidised be towards agglomeration due to an increase in Na content of the bed hindered by trona's chemical composition followed by a breaking-up in hydrodynamic behaviour of the bed was easily visualised from temperature profiles. The Na/Ca molar ratio was adjusted and the presence of Ca behave as agglomeration retardant factor for the whole bed. However, the formation of soda-lime-glass was quite clear from the sample taken from the bed after the occurrence of agglomeration phenomena. This was observed from the samples collected from the bed after agglomeration. The reason for the occurrence of soda- lime-glass was explained by the high Na content of active soda as well as the calcium content of lignite and addition of CaC03 which was needed for the proper Ca/S and Ca/Na ratios. There was also another observation on the crucial role of active soda on the activation of sorbents. Therefore, the SEM micrograph of activated soda- limestone mixture after calcination presented open pores compare to raw limestone. A multiple regression analysis was carried out in order to assess the experimental and theoretical results. The correlation coefficient were found around 0.8. The effect of lignite particle size was observed whereas there was no correlation between SO2 and sorbent particle size unlike our expectations. Ca/S molar ratio has a correlation with SO2 emission. A general equation S02= K+AX+BX2+CX3 ( where K,A,B,C were chosen as coefficients and X as Ca/S molar ratio) was developed according to experimental results. Ca/S molar ratio was chosen between 0 and 5. Calculated and experimental data were found compatible even though there was differences in the physical and chemical properties of sorbents. The calcination and sulphation behaviour of limestone, Dl and D2 dolomites also differred due to their particle sizes. Two different particle size for sorbents were used during the experiments. The correlation coefficients were calculated between 0.95 and 0.99. Another regression analysis were accomplished in order to obtain a correlation between Ca/S molar ratios which were chosen above 2.5. The correlation coefficients were calculated as 0.52 for limestone and 0.36 for both dolomites. As could be seen from the regression coefficients, these results are not acceptable. The reason for these results could be explained by the diversity of the parameters and the complexity of the reactions which occurred during the sorption of SO2 in a fluidised bed combustor, even in steady-state experimental conditions. The regression analysis results indicated that the development of a model for the sorption of SO2 emissions using different sorbents is quite difficult. This is mainly due to the variety of sorbents, lignite and sorbent particle sizes and Ca/S molar ratio. The lignite particle size is an important factor since its ignition and combustion time in the bed differred. The combustion of small lignite particle sizes takes place instantaneously whereas that of the bigger particles combustion proceeds longer. Consequently, the smaller the lignite particle size, the more SO2 emits.