Feasibility of using an electronic nose for quality assessment of air in poultry slaughterhouse rendering plants

Introduction:
Industrial activities growth and increasing emissions getting attention to environmental problems in the world. Unpleasant smell (malodor) can be classified as pollutants that had negative impact on quality of life. About one third to half of animal products that are unusable for humans are used as raw materials in the processing and recycling industry. One of the most important features of slaughterhouse rendering plants is the release of volatile compounds and unpleasant odors, which disturb the working people and the adjacent inhabitants of this industry. The purpose of this study was to evaluate an electronic nose system as a tool for displaying vapors from a slaughterhouse rendering plant, select an array of appropriate sensors and determine the fingerprint for this pollutant unit.
 Material and methods:
The batch cooking was carried out at temperature 140 ° C under pressure 3 bar for 4 hours and uniformly stirred at a speed of 20 rpm (ISIRI 2389, 1995). A total of 30 liquid samples, each one with volume of 10 ml, were obtained from condensed vapors of batch cooker. Samples were transferred to the laboratory located at the Agricultural Machinery Department of the University of Tehran, and an electronic nose system was used for analysis and investigation. The electronic nose system was based on metal oxide semiconductor (MOS) sensors and included an array of six gas sensors. The principal component analysis (PCA) method was used to evaluate the sensors and select appropriate array of sensors. 
Results and discussion:
Observation of sensor’s response until steady state with aim of reducing consumption of time and oxygen, showed that 30 seconds to reach sensors to their baseline (baseline time), 100 seconds to contact the vapors with sensors (injection and response time) and 40 seconds to clear the sensors (purging time) were suitable . Correlation analysis of the sensors showed that some of the sensors had a correlation higher than 0.85. The analysis of variance distribution of two primary principal components showed that these two components comprises for more than 99% of the total variance, and the primary component (PC-1) comprise over than 98% of the total variance. The highest effect on the first component was related to the MQ135 sensor, after that MQ136, MQ9, MQ131, MQ5 and finally MQ3 sensors were effective. The analysis of the principal components showed that MQ135 and MQ136 sensors with high sensitivity to ammonia, hydrogen sulfide and benzene were the most suitable sensors for detecting the unpleasant smell released from the poultry slaughterhouse waste batch cooking (rendering plant). Selection of MQ135 and MQ136 sensors as an array can be used to distinguish between pollutants from poultry rendering plant and pollutants released from other industries, and these sensors can be used to determine the fingerprint of the odor emitted from poultry rendering plant. 
Conclusion:
According to the results of this research, the electronic nose system with array of MQ135 and MQ136 can be used to monitor and control the quality of pollutants from poultry rendering plant. Also, with referring to this fact that unpleasant smell in air quality control is not defined standardly in Iran, an electronic nose system could be used to control the air quality in terms of the presence of unpleasant odors in the environment.