| With higher environmental quality requirement and more stringent environmental legislation,tobacco industry is also increasing concern about off-gas treatment in China.Biofiltration has developed into a cost-effective and environmentally benign technology for gaseous contaminants in many industries. However,only limited information is available about the application of an open-bed, full-scale biofilter for odor abatement in tobacco industry.Compared with flue-cured tobacco processing,burley tobacco processing is known to produce the highest and most odorous exhaust gases.A modular scrubber with full-scale biofilter was installed to treat emissions from a burley tobacco processing plant,and the main pollutants were total particulate matters(TPM),odors(primarily ammonia and nicotine) and total volatile organic compounds(TVOC).In this work,a nicotine-biodegradation strain was isolated and identified,operation conditions of the industrial-scale biofilter were optimized,the performance of the modular biofiltration system was evaluated and a model to simulate biofiltration process was proposed.The major research contents and results are as follows.A high nicotine-degrading bacterium was isolated and named as ZUT5 from the full-scale biofilter.On the basis of morphological,physiological characters,Biolog GN,16S rDNA sequence and phylogenetic characteristics,it was identified as a member of Pseudomonas sp.The characteristics of nicotine degradation by strain ZUT5 were investigated.The optimal temperature and initial pH value for nicotine degradation by strain Pseudomonas sp.ZUT5 was 30℃and pH 7.5,respectively.The results also showed that Pseudomonas sp.ZUT5 had an ability to degrade nicotine in high concentration.The addition of NH4CL and glucose in the medium could improve the nicotine degradation.Four new intermediate metabolites such as 2,3'-bipyidine, 3-(2,3,4-3H-pyrrol-5-yl)-pyridine,cotinine and 3-pyridinecarboxylic acid could be detected in the medium by GC-MS.The optimized operating conditions of the biofiltration system were studied.The effects of inlet concentration,gas flow rate,empty bed residence time,medium initial moisture contents,pH,off-gas temperature and shut-down periods on full-scale biofilter performance were investigated.The results indicated that the optimum range of the system gas flow rates was about 21000-25000 m~3/h,and corresponding to empty bed residence time of 38~45 s.The results also showed that the optimum medium initial moisture contents were about 40-55%(wet weight basis),off-gas inlet temperature in biofilter should not be higher than 40℃,and the pH value of filter bed bottom-medium should be lower than pH8.Under various gas flow rates,the exhaust gases had been sampled and analyzed for evaluating the performance of modular scrubber with full-scale biofilter.The modular scrubber efficiencies for total particulate matters removal ranged from 93.39 %to 96.30%.Owing to high removal efficiency,the particulate matters removal process can protect biofilter from possible clogging for off-gas containing high particle concentration.In addition,nicotine,ammonia and TVOC were treated in the modular scrubber unit by physical-chemical methods.Because more than half of the ammonia and nicotine had been removed in the modular scrubber,the full-scale biofilter had been protected from basification.The elimination capacities of the full-scale biofilter for nicotine,ammonia and TVOC increased with inlet load to a maximum of 81.57,1032.90 and 33791.10 mg/m~3.h,respectively.As a result,the overall efficiencies of the whole system for the removal of nicotine,ammonia and TVOC were in the range of 85.76-99.30%,73.33~91.71%,and 94.10-100%, respectively.A kinetic model was developed that described basic transport and biological processes in biofilter.Model parameter was determined by a partial least-square regression for two odorous contaminants.The expression also indicated that a linearly proportional relationship existed between elimination capacity and the inlet concentration.This theoretical model was fitted with experimental data.The results showed that the model predictions were basically closer to the experimental data. |