The aim of our research activity is the development of a system, based on the use of electronic noses, for the continuous monitoring of environmental odours at specific receptors, i.e. directly where the presence of odours is lamented. This paper describes an experimental monitoring conducted at an Italian composting plant, and it focuses especially on the training procedures and on the principles followed for the data processing. The training consisted in the analysis of different gas samples of known olfactory quality at different odour concentration values, in order to teach the instrument to recognize odours from the qualitative and quantitative point of view. The first step of data processing was exploratory data analysis, which allowed to identify and to exclude two outliers from the training data. Specific qualitative and quantitative recognition tests enabled the selection of the optimal features and of the best pattern recognition algorithms. The optimized electronic nose presented a qualitative classification accuracy of 96.4% and the correlation index relevant to the odour concentration determination was R2 = 0.90172. Two instruments were installed inside the composting plant and at two specific receptors, respectively, with the function of analyzing the ambient air every 15 min. The results of these analyses allowed the determination of the odour impact of the monitored plant on both receptors and the identification of the plant principal odour source, which turned out to be represented by the plant area dedicated to the storage of the refined compost heaps. Finally, the reliability of the results obtained by the air analyses with electronic noses was validated by comparing the instrumental responses with the meteorological data (wind speed and wind direction) relevant to the monitoring period.
Development of a system for the continuous monitoring of odours from a composting plant: Focus on training, data processing and results validation methods
SIRONI, SELENA;CAPELLI, LAURA MARIA TERESA;CENTOLA, PAOLO;DEL ROSSO, RENATO
2007-01-01
Abstract
The aim of our research activity is the development of a system, based on the use of electronic noses, for the continuous monitoring of environmental odours at specific receptors, i.e. directly where the presence of odours is lamented. This paper describes an experimental monitoring conducted at an Italian composting plant, and it focuses especially on the training procedures and on the principles followed for the data processing. The training consisted in the analysis of different gas samples of known olfactory quality at different odour concentration values, in order to teach the instrument to recognize odours from the qualitative and quantitative point of view. The first step of data processing was exploratory data analysis, which allowed to identify and to exclude two outliers from the training data. Specific qualitative and quantitative recognition tests enabled the selection of the optimal features and of the best pattern recognition algorithms. The optimized electronic nose presented a qualitative classification accuracy of 96.4% and the correlation index relevant to the odour concentration determination was R2 = 0.90172. Two instruments were installed inside the composting plant and at two specific receptors, respectively, with the function of analyzing the ambient air every 15 min. The results of these analyses allowed the determination of the odour impact of the monitored plant on both receptors and the identification of the plant principal odour source, which turned out to be represented by the plant area dedicated to the storage of the refined compost heaps. Finally, the reliability of the results obtained by the air analyses with electronic noses was validated by comparing the instrumental responses with the meteorological data (wind speed and wind direction) relevant to the monitoring period.File | Dimensione | Formato | |
---|---|---|---|
RDR160.pdf
Accesso riservato
:
Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione
1.05 MB
Formato
Adobe PDF
|
1.05 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.