Fault Detection in Wastewater Treatment Plants: Application of Autoencoders Models with Streaming Data

• Type: Article in International Conference Proceedings Book
• Date: 2023

Authors

Salles, R (Author) Other; Mendes, J (Author) Other; Rita Ribeiro (Author) FCUP; João Gama (Author) FEP

Conference proceedings International

Title: Machine Learning and Principles and Practice of Knowledge Discovery in Databases - International Workshops of ECML PKDD 2022, Grenoble, France, September 19-23, 2022, Proceedings, Part I

Pages: 55-70

Indexing

ISI Web of Knowledge - 1 Citation

Scopus - 3 Citations

Other information

Authenticus ID: P-00X-T5N

DOI: 10.1007/978-3-031-23618-1_4

Abstract (EN): Water is a fundamental human resource and its scarcity is reflected in social, economic and environmental problems. Water used in human activities must be treated before reusing or returning to nature. This treatment takes place in wastewater treatment plants (WWTPs), which need to perform their functions with high quality, low cost, and reduced environmental impact. This paper aims to identify failures in real-time, using streaming data to provide the necessary preventive actions to minimize damage to WWTPs, heavy fines and, ultimately, environmental hazards. Convolutional and Long short-term memory (LSTM) autoencoders (AEs) were used to identify failures in the functioning of the dissolved oxygen sensor used in WWTPs. Five faults were considered (drift, bias, precision degradation, spike and stuck) in three different scenarios with variations in the appearance order, intensity and duration of the faults. The best performance, considering different model configurations, was achieved by Convolutional-AE.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 16