|ROMERO-FLORES, A. - University Of Maryland
|MCCONNELL, L.L. - University Of Maryland
|RAMERIZ, M. - Collaborator
|TORRENTS, A. - University Of Maryland
Submitted to: Sensors and Actuators B: Chemical
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/26/2017
Publication Date: 7/27/2017
Citation: Romero-Flores, A., McConnell, L., Hapeman, C.J., Rameriz, M., Torrents, A. 2017. Evaluation of an electronic nose for improved biosolids alkaline-stabilization treatment and odor management. Sensors and Actuators B: Chemical. 186:151-159. https://doi.org/10.1016/j.chemosphere.2017.07.135.
Interpretive Summary: Improved gas monitoring tools have been used in food industry and more recently in disease detection and agricultural and environmental engineering applications. The electronic nose sensors have evolved over the last decade and are a promising approach for applications which require the rapid and automated detection of specific gases or patterns of gases. They can also be used to discriminate between air samples containing complex mixtures of compounds. No information is available, however, on the application of sensor arrays to monitor process performance of biosolids treatment processes. Treated biosolids are now used as an alternative fertilizer source. The objective of this work, therefore, was to examine the feasibility of a series of electronic nose sensors to discriminate between treatment of biosolids stabilized with various amounts of lime and to explore its performance to detect key odors versus sophisticated analytical instruments. The electronic nose sensors were able to discriminate between lime dosages for alkaline stabilized biosolids, but its usefulness in recognizing patterns was somewhat limited. This was most likely due to the variability in biosolids treatment processes. The sensors also showed close agreement with sophisticated analytical instrumentation, however, the instruments could detect the gases at much lower concentrations. These results are expected to provide important information on the use of commercially-available electronic nose technology to increase the effectiveness of biosolids treatment programs and to assist in reducing odor complaints when biosolids are applied to agricultural fields.
Technical Abstract: Electronic nose sensors are designed to detect differences in complex air sample matrices. For example, they have been used in the food industry to monitor process performance and quality control. However, no information is available on the application of sensor arrays to monitor process performance of biosolids treatment processes. The objective of this work, therefore, was to examine the feasibility of an electronic nose to discriminate between treatment conditions of alkaline stabilized biosolids and to explore its performance by comparison with quantitative analysis of key odorants. Seven quick lime treatment rates from 0 to 30% (w/w) were prepared, and the off gas was monitored by with an electronic nose with ten metal oxide sensors. Pattern recognition models were created from the electronic nose data using using linear discriminant analysis (LDA) and principal component analysis (PCA)LDA and PCA. LDA performed better than PCA and showed clear discrimination when single tests were evaluated, . However,although when the full data set was included in the pattern creation, the discrimination of classes (lime dosages) was not as clear as with single test data. The Fsrequency of recognition was tested by using three algorithms with Euclidan and Mahalonbi performing at 81.1% accuracy while discriminant Function Analysis DFA at 69.870%. These limitations are found shown to be due to the intrinsic variability of the wastewater treatment process and to the low specificity of the sensor evaluated in this study. The electronic nose was able to discriminate from 0%, 15% and 30% lime and, to some extent 25% lime. These data suggest that under-dosed and over-dosed classes may be created to support the alkaline stabilization process of biosolids, assuming that 15% lime is the dosage needed to safely achieve a pH of 12 required for 99% pathogen reduction.