| In this research, a portable intelligent electronic nose prototype, which can be used both in the laboratory and for field studies, was developed to analyze and predict the strength of livestock farm odour. The prototype consists of 14 gas sensors that were especially selected for measuring odorous components from livestock farms. An expert system called "Odour Expert" was developed to support researchers' and farmers' decision making about odour control strategies for livestock operations. "Odour Expert" incorporates both conventional analysis methods and artificial intelligent methods, and it can not only take odour measurement from livestock farms, but also give expert advice on what control actions to take in order to efficiently reduce odour. Downwind odour measurements were conducted at 14 Ontario livestock farms in 2004. The data gathered by the electronic nose were correlated with the readings from a Nasal Ranger Field Olfactometer. The analysis of the data shows that using the electronic nose to predict livestock farm odour yields more accurate results.; A neural network based multi-component, multi-factor analytical odour model was proposed to adequately model livestock farm odours. A structural learning with forgetting algorithm was proposed for contribution analysis of chemical components and factors to the perception of odour. Contribution analysis would allow identification of significant odour components and major contributing factors, thus improving the efficiency in developing odour measurement and reduction technologies. The effectiveness of the proposed approach is demonstrated by using a pork farm odour database. |
