Effect of Self-Heat Circulation on VOCs Decomposition in Regenerative Thermal Oxidizer

Transcription

1 Effect of Self- Circulation on VOCs Decomposition in Regenerative Thermal Oxidizer Shinsuke Iijima 1, Katsuya Nakayama 1, Koichi Ushiroebisu 1 Mitsuhiro Kubota 2 and Hitoki Matsuda 2 1. Engineering Division, Sinto Kogio, Ltd, Nagoya, Japan 2. Department of Energy Engineering and Science, Nagoya University, Nagoya, Japan Abstract: The oxidative decomposition behavior of volatile organic compounds (VOCs) was studied in a pilot-scale unit of regenerative thermal oxidizer (RTO). As a result, it was found that thermal efficiency of RTO was increased from 92.8 to 93.4% by recovering heat of treated gas decreasing the switching time between two regenerative heat exchangers from 120 to 60 sec at a temperature of 800 C and at a contact time of 0.23 sec, which resulted in the reduction of fuel consumption by 9.2%. It was also found that thermal efficiency was increased from 92.1 to 94.8% saving more than 27.7% of fuel consumption when the contact time was increased from 0.20 to 0.46 sec, at 800 C and at a switching time of 120 sec. It was estimated from both the fuel consumption and toluene s exothermic heat that the critical concentration of VOC for self-combustion was g-toluene/nm 3, without additional heating of the RTO employed. Keywords: Regenerative, Self-Combustion, Toluene, Volatile Organic Compounds 1. INTRODUCTION VOCs are used mainly as a solvent in several industrial processes such as painting, printing and laminating. Consequently, as an example, an amount of 132,419 tons of toluene was discharged into the atmosphere from a total number of 34,820 facilities in FY 2001 in Japan, according to the 2003 PRTR data report published by the Ministry of the Environment and the Ministry of Economy [1]. Moreover, VOCs are responsible for the generation of photochemical oxidants and suspended particulate matter (SPM) which is harmful to human health, agricultural production and/or a forest [2]. From this viewpoint, many regulations such as the Council Directive 1999/13/EC [3] and the Clean Air Act Amendments of 1990 [4] were established for the limitation of VOCs emission since 1990 in Europe and USA, but only 22 odorous compounds such as toluene, ethyl acetate as well as other VOCs such as benzene, trichloroethylene and tetrachloroethylene have so far been regulated in Japan [5]. Hence, air pollutant abatement regulation has been revised more rigorously in May This new air pollutant abatement regulation targets the controls of exhaust gases from driers, paint booth, industrial cleaning and VOCs storage tank. For example, total hydrocarbon concentration permitted in a drier exhaust air for a gravure printing is 700 ppm, and these new limits have to be met in existing application by March 2010[6]. For cleaning of VOCs containing emission gases from printing dryer and laminator, etc., catalytic oxidizer or wet scrubber using some chemicals has been employed [7]. In addition, regenerative thermal oxidizer (RTO) has recently attracted attention, owing to its advantages in terms of economic efficiency, no-influence on catalysis caused by sulfur- and chloric-elements and the applicability to a wide concentration range up to 1/4 of explosion limit. To increase thermal efficiency and optimize the RTO operation as well as its design, in terms of VOCs decomposition efficiency, the parameters such as gas flow rate, contact time in the heat exchanger and switching s for cyclic operation should be considered. In the present work, the effect of these parameters on the performance of VOC decomposition was investigated by using our pilot-scale RTO system. 2. EXPERIMENTAL 2.1. Operation Principle of RTO Figure 1 illustrates the operation principle of a RTO unit. The RTO consists of two regenerative heat exchangers charged with ceramic honeycombs. The raw gas and the treated gas are made flow into these two heat exchangers alternately by regulating the gas flow direction into the heat exchanger through the main valves V01 to V04. At first, the inlet valve V01 and outlet valve V04 is opened (the inlet valve V02 and outlet valve V03 remained closed), the raw gas containing VOCs is allowed to enter the RTO through heat exchanger A. The VOCs undergo oxidative decomposition releasing the combustion heat in the combustion chamber. Then, the treated gas free from VOCs flows into the honeycombs placed in the heat exchanger B and leaves the RTO through valve V04. In this process, the heat of treated gas is stored in the heat exchanger B. In the next step, the direction of gas flow in the RTO is reversed by opening the valves V02, and V03 and closing the valves V01 and V04. Consequently, the raw gas enters the RTO through the heat exchanger B and the heat stored in the heat exchanger B was utilized to preheat the raw gas. Then, the preheated gas enters the combustion chamber, and the VOCs are thermally decomposed. After VOCs combustion, the treated gas free of VOCs leaves the RTO through heat exchanger A, and similarly to the former step, the heat of treated gas is stored in the heat exchanger A. Finally, the Corresponding author: S. Iijima, s-iijima@sinto.co.jp 976

2 direction of gas flow in the RTO is reversed again in order to operate the RTO in cyclic manner. The time between each change of gas flow direction, hereafter referred to as switching time, is kept constant for a given experiment. 1st step 2nd step V01 Combustion chamber A B V03 V02 V04 Main valves Fig.1 Schematic diagram of RTO operation 2.2. Experimental apparatus As shown in Fig. 3, the experimental apparatus used in this study was a pilot-scale RTO unit. In the combustion chamber, a gas burner (YOKOIKIKAI KOSAKUSHO, SGL-1S) was installed and a thermo-controller (YAMATAKE, SDC31) was used to maintain the temperature at a given value from 500 to 900 C. The gas volume of LPG consumed was measured by a gas flow meter (YAMATAKE, CMG250P). Ceramic honeycombs of total area of (A) x 1200 height(h) were loaded in two heat exchanger towers, of which the specification is shown in Table 1. The combustion chamber and the heat exchanger towers were covered by insulating material(250 in thickness) to reduce the heat loss. Four gas flow valves controlled by a sequencer (OMRON, C200) were connected to the inlet and the outlet duct in the two heat exchangers. The temperature of the raw gas and the purified gas was measured by K-type thermocouples. The raw gas was sucked into the RTO using exhaust gas fan (SUNADA SEISAKUSHO, 18A56) controlled by frequency controller (FUJI ELECTRIC, FVR) and the flow rate was adjusted using a flow controller (YAMATAKE, SDC31). LPG Burner Burner Controller C Figure 2 shows the temperature gradient along heat exchangers charged with ceramic honeycombs, when the RTO was operated at T c =800 C, gas flow rate of 30m 3 /min and gas flow direction from heat exchanger A to B. Each point in this figure represents a temperature measured at a given height of heat exchanger just before the end of switching time. insulation Combustion Chamber 600 x 600 A T ca B T cb Φ200 T ha T hb Gas temperature ( C) T hb in heat exchanger B T ha exchanger height () T cb raw gas treated gas T ca in heat exchanger A Fig.2 Temperature gradient in the honeycombs In Fig. 2, the raw gas temperature was observed to increase from T ha to T ca by receiving the heat stored in heat exchanger A. The T ca value was found to be slightly lower than the combustion temperature of T c =800 C, which was attributed to heat transfer efficiency lower than 1. Then, the heat required to achieved T c =800 C was supplied by a burner or by VOCs combustion. Conversely, the temperature of treated gas decreased from T cb to T hb, during the downward flow of the treated gas through heat exchanger B, owing to the heat storage process in heat exchanger B. Inlet Duct Φ200 P T i Outlet Duct Fig.3 Schematic diagram of RTO unit Table 1 Specification of honeycombs 150 x x Cross Section Height Channel Size Wall Thickness CSI Geometric Surface Open Space (O s ) Weight Material Raw Density Specific Capacity Thermal Conductivity Cordierite T o Exhaust FAN M VVVF cells/inch 2 m 2 /m 3 % kg g cm -3 J kg -1 K -1 W m -1 K Experimental procedure At the beginning of RTO operation, the ambient air was introduced into RTO to start the preheating period of 2 hours. During this period, the gas flow valves were continuously switched at given switching s, and the combustion chamber as well as heat exchangers was preheated to a given value of T c. Then, the raw gas containing toluene was introduced into the RTO. During this operation, the consumption of LPG and the gas temperatures in the inlet (T i ) and the outlet duct (T o ) were meas- 977

3 ured. In addition, radiation heat loss (Q r ) from the surface of the combustion chamber and the heat exchangers were estimated. The experimental conditions were as follows; the temperature of the combustion chamber was set between 500 and 900 C and the switching time s of gas flow valves were set from 60 to 180sec, and the gas flow rate was in the range of m 3 min RESULTS AND DISCUSSION 3.1. Effect of operating conditions on heat exchanging efficiency A typical time change of the temperatures of the raw gas and the exhaust gas in the steady state condition at T c =800 C, the switching time of 120sec and the gas flow rate of 30 m 3 /min, is shown in Fig.4. Gas temperature ( C) gas flow A B B A time (s) Fig.4 Time change of the temperatures of the raw gas and the exhaust gas To Ti In this figure, it was found that the exhaust gas temperature T o, regularly fluctuated in the range of 50 to 100 C, and the temperature peaks corresponded to the end of switching time. At first, as mentioned in section 2.1, the treated gas discharged from the combustion chamber at 800 C was continuously introduced into the honeycomb-type heat exchanger, where the heat was transferred from the exhaust gas to a heat exchanger. As a result, the temperature of the exhaust gas dropped from 800 C in the combustion chamber to an outlet value of T o. To explain the fluctuation of T o shown in Fig. 4, only gas flow direction from heat exchanger A to B in one cycle was considered. Additionally, it should be noticed that the heat exchanger B was cooled down by the raw gas (T i =19 C) in the previous cycle and the temperature measured at the bottom of heat exchanger B (T hb ) was about 22 C at the end of switching time. Then, it can be seen that the value of T o was about 100 C at the beginning of A B cycle and the value quickly dropped to about 50 C. Such a behavior resulted from the fact that the exhaust gas from the combustion chamber (T c =800 C) was introduced to relatively cold heat exchanger B and was quickly cooled down to a temperature of 50 C. However, as the temperature of heat exchanger B started to rise owing to the heat transferred from the exhaust gas, the heat exchange efficiency of heat exchanger B was reduced, which ultimately led to an increase in outlet temperature to about 100 C at the end of A B cycle. Finally, when the gas flow was reversed, the same behavior was observed in the next B A cycle. When the treated gas flowed into heat exchanger B, the treated gas temperature was decreased from T c to T hb. Therefore the heat calorific value transferred from the treated gas to the honeycombs was calculated by the following equation (1), on the assumption that T hb was equal to T o. Q co = W (T c C c -T o C o ) (1) where, Q co is the regenerated calorific value of the honeycombs, W is the mass flow rate of the gas [kgh -1 ], C c, and C o is the specific heat capacity of the gas at T c and T o [kjkg -1 K -1 ]. The efficiency of heat exchange (ή) of the honeycombs was calculated by the equation (2), while the heat calorific value for heating the raw gas to combustion temperature was calculated by the equation (3). ή [%] = Q co / Q ci 100 (2) Q ci = W (T c C c -T i C i ) (3) where, Q ci is the calorific value for heating the raw gas from T i to T c, C i is the specific heat capacity at T i [kjkg -1 K -1 ]. When the value of C o and C i was assumed to be equal in equation (2) and equation (3), equation (1) was replaced by equation (4). ή [%] = (T c -T o )/ (T c -T i ) 100 (4) Equation (4) represents the temperature efficiency, which was derived on the assumption that the heat loss could be neglected [8,9]. However, radiation heat loss (Q r ) from the surface to atmosphere can not be neglected in the pilot scale apparatus. In the present work, Q r was estimated about 28 MJ h -1 at T c =800 C. Consequently, the heat balance of the combustion chamber was considered as below. Q i + Q x + Q L + Q v = Q co + Q o + Q r (5) where, the left side of equation is the input heat calorific value, the right side of equation is the output heat calorific value, Q i is the heat calorific value of the raw gas at T i [kj h -1 ], Q x is the heat calorific value transferred from honeycombs to the raw gas [kj h -1 ], Q L is the combustion heat calorific value of the LPG [kj h -1 ], Q v is the combustion heat calorific value of the VOCs [kj h -1 ], Q o is the heat calorific value of the treated gas at T o [kj h -1 ]. Equation (5) was replaced by equation (6), since the value of Q x was equal to Q co in equation (5) when the RTO was operated in the steady-state condition. Q L + Q v = ( Q o - Q i ) + Q r (6) Therefore the thermal efficiency of the RTO was cal- 978

4 culated by the following equation (7) using the value of Q L and Q v. ή [%]= (1-( Q L + Q v ) / Q ci ) 100 (7) As shown in Fig. 5, when the gas flow vales were switched to adjust the switching time between 60 to 180sec for the cyclic operation of the RTO at T c =800 C and the gas flow rate of 30 m 3 /min, the thermal efficiency of the RTO was changed with the switching time. Thermal efficiency 95% 94% 93% 92% 91% 90% Time of cyclic operation [s] Fig.5 The effect of the switching time In this figure, it was found that the thermal efficiency was almost in inverse proportion to switching time in the range of 60 to 180sec and increased from 92.4 to 93.4% when the switching time was decreased from 120 to 60 sec at T c =800 C, which further resulted in the fuel consumption reduction by 9.2%. Further, when the gas flow rate was changed, the thermal efficiency of RTO at T c =800 C and the switching time of 120sec was changed as shown in Fig.6. The contact time, t in Fig.6 was calculated by the following equation (8). t [s] = V / (F / 60) (8) V = A O s H (9) where, V is the volume of ceramic honeycombs channel [m 3 ], F is the gas flow rate at the average temperature in the ceramic honeycombs[m 3 min -1 ]. Thermal efficiency 95% 94% 93% 92% 91% 90% Contact time to heat exchanger [s] Fig.6 Effect of contact time It is seen in Fig. 6 that the thermal efficiency is increased from 92.1% at a contact time of 0.2 sec to 94.8% at a contact time 0.46 sec, under the condition of combustion chamber temperature, T c =800 C and switching time of 120 sec. By the increase in the thermal efficiency from 92.1 to 94.8%, the consumption of LPG was saved by more than 27.7%. Therefore it was also found that the effect of contact time on the fuel consumption saving was approximately 3 times higher than the effect of the switching time. When VOCs are involved in a raw gas to be treated, LPG consumption can be decreased, since VOCs are usually combustible which may supplement the combustion heat of LPG. In the present study, the critical concentration of VOCs in the raw gas was defined as the minimum VOCs content in 1 Nm 3 of raw gas which can keep steady-state combustion condition without LPG burning. According to the calculation from both the fuel consumption and toluene s exothermic heat, the critical VOC concentration for self combustion in the RTO employed was 1.4 g-toluene/nm 3 at a contact time 0.46 sec and 2.2 g-toluene/nm 3 at a contact time 0.2 sec Ratio of heat exchange calorific value to stored heat calorific value The gas temperature in the honeycombs was measured at 300 s from T ha to T ca along the height of heat exchanger tower in order to investigate the stored calorific value. From Fig. 2, it was found that the temperature difference of the gas was about 130 C at 300 height measuring point, while it was about 65 C at 900 height. It was therefore considered that the contribution of the lower part of the honeycombs was higher than the upper part to the heat exchange. The average value of heat stored in the honeycomb was calculated about 150MJ by the following equation. Q a = w C a T a (10) where, Q a is the average stored heat calorific value of the honeycombs in one heat exchanger, w is the weight of the honeycombs in one heat exchanger [kg], C a is the specific heat capacity of honeycombs [kj kg -1 K -1 ], T a is the time and height averaged temperature of honeycombs [ C]. On the other hand, the calorific heat value exchanged from the treated gas for switching time of 120sec calculated by equation (1) was about 62MJ at T c =800 C and and the gas flow rate of 30 m 3 /min. Hence, the average calorific heat value stored in ceramic honeycombs was 2.4 times as large as the calorific heat value recovered from the treated gas. In the present work, only one type of ceramic honeycomb was used for evaluating thermal efficiency of heat exchanger in the RTO. Further investigation on the effect of other parameters such as the geometric surface, the specific heat capacity and the thermal conductivity is needed, in order to obtain the optimal conditions for VOC decomposition and to predict the thermal efficiency in this type of RTO. 979

5 4. CONCLUSIONS The main findings obtained in this paper are suarized as follows. 1) It was found that thermal efficiency increased from 92.1 to 94.8% and more than 27.7% of fuel consumption decreased with increasing contact time from 0.20 to 0.46 sec at T c =800 C and switching time of 120sec. 2) The critical VOC concentration for self combustion in the RTO without additional heating was 1.4 g-toluene/nm 3 at a contact time 0.46 sec and 2.2 g-toluene/nm 3 at a contact time 0.2 sec. 3) The average calorific heat value stored in ceramic honeycombs was 2.4 times as large as the calorific heat value recovered from the treated gas. 5. Nomenclature A = Total area of ceramic honeycombs [m 2 ] C a = Specific heat capacity of honeycombs [kj kg -1 K -1 ] = Specific heat capacity at gas temperature C c C i T c [kj kg -1 K -1 ] = Specific heat capacity at gas temperature T i [kj kg -1 K -1 ] C O = Specific heat capacity at gas temperature T o [kj kg -1 K -1 ] F = Gas flow rate at the average temperature in the ceramic honeycombs [m 3 min -1 ] H = Height of ceramic honeycombs [m] O = Open space of ceramic honeycombs [%] Q a = Average calorific value stored in the honeycombs [kj h -1 ] Q r = Radiation heat loss from surface of the combustion chamber and the heat exchangers [kj h -1 ] Q ci = Calorific value for heating the raw gas from T i to T c [kj h -1 ] Q co = Regenerative calorific value of ceramic honeycombs [kj h -1 ] Q L = Combustion heat calorific value of LPG [kj h -1 ] Q v = Combustion heat calorific value of VOCs [kj h -1 ] Q x = Transferred heat calorific value from the honeycombs to the raw gas [kj h -1 ] Q y = Transferred heat calorific value from the treated gas to honeycombs [kj h -1 ] T a = Time and height averaged temperature of honeycombs [ C] T c = Set point of the gas temperature in combustion chamber [ C] T ca = Gas temperature at the upper part of heat exchanger A [ C] T cb = Gas temperature at the upper part of heat exchanger B [ C] T ha = Gas temperature at the lower part of heat exchanger A [ C] T hb = Gas temperature at the lower part of heat exchanger B [ C] T i = Raw gas temperature in the inlet duct [ C] T o = Treated gas temperature in the outlet duct [ C] V = Volume of ceramic honeycombs channel [m 3 ] W = Mass flow rate of the gas [kg h -1 ] w = Honeycombs weight in one heat exchanger [kg] ή = Efficiency of heat exchange [%] 6. REFERENCES 1. Ministry of the Environment Government of Japan, Law Concerning Reporting, etc. of Releases to the Environment of Specific Chemical Substances and Promoting Improvements in Their Management, (2003) (in Japanese) 2. M. Izumo, Seminar Report for Organic Solvent Regulation Trend in 1995, The Japan Society of Industrial Machinery Manufacturers (1995), pp.1-16 (in Japanese) 3. The Council of the European Union, Council Directive 1999/13/EC, en pdf (1999) 4. U.S. Environmental Protection Agency, Clean Air Act, (1990) 5. Ministry of the Environment Government of Japan, Offensive Odor Control Law, (1971) (in Japanese) 6. The Japan Society of Industrial Machinery Manufacturers, Seminar Report of VOC treatment technology, The Japan Society of Industrial Machinery Manufacturers (1995), pp (in Japanese) 7. Ministry of the Environment Government of Japan, Air Pollution Control Law, (1968) (in Japanese) 8. D. Kunii, Thermal Unit Operations, Maruzen (1976), pp (in Japanese) 9. A. Tsutida, S. Yamazaki and M. Akiyama, Problems and Solutions in Engineering Transfer, Gakken (1965), pp (in Japanese) 980