Rapid BOD assessment with a microbial array coupled to a neural machine learning system

WATER RESEARCH

卷:166

Article

Jouanneau, Sulivan¹  Grange, Emilie¹  Durand, Marie-Jose¹  Thouand, Gerald¹ 

[1] Univ Nantes, UMR CNRS GEPEA 6144, 18 Blvd Gaston Defferre, F-85035 La Roche Sur Yon, France

关键词: BOD5;    Multivariate approach;    Bacterial biosensor;    Neural network;    Wastewater;   

DOI  :  10.1016/j.watres.2019.115079

来源: Elsevier

【 摘 要 】

The domestic usage of water generates approximately 310 km(3) of wastewater worldwide (2015, AQUASTAT, Food and Agriculture Organization of United Nations). This sewage contains an important organic load due to the use of this water; this organic load is characterized using a standard method, namely, the biological oxygen demand measurement (BOD5). The BOD5 provides information about the biodegradable organic load (standard ISO 5815). However, this measurement protocol is very time-consuming (5 days) and may produce variability in approximately 20% of results mainly due to variation in the environmental inocula. To remedy these limitations, this work proposes an innovative concept relying on the implementation of a set of rigorously selected bacterial strains. This publication depicts the different steps used in this study, from bio-indicator selection to validation with real wastewater samples. The results obtained in the final step show a strong correlation between the developed approach and the reference method (ISO 5815) with a correlation rate of approximately 0.9. In addition, the optimization of the experimental conditions and the use of controlled strains (8 selected strains) allow significant reduction in the duration of the BOD5 analysis, with only 3 h required for the proposed method versus 5 days for the reference method. This technological breakthrough should simplify the monitoring of wastewater treatment plants and provide quicker, easier and more coherent control in terms of the treatment time. (C) 2019 Elsevier Ltd. All rights reserved.

【 授权许可】

Free