Sensor-fault tolerance in a wastewater treatment plant by means of ANFIS-based soft sensor and control reconfiguration

• Published in: Neural computing & applications Vol. 30; no. 10; pp. 3265 - 3276
• Main Authors: Belchior, Carlos Alberto C., Araújo, Rui Alexandre M., Souza, Francisco Alexandre A., Landeck, Jorge Afonso C.
• Format: Journal Article
• Language: English
• Published: London Springer London 01.11.2018; Springer Nature B.V
• Subjects: Activated sludge process; Adaptive systems; Artificial Intelligence; Artificial neural networks; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Computer simulation; Correlation analysis; Data Mining and Knowledge Discovery; Fault detection; Fault tolerance; Fuzzy logic; Fuzzy systems; Image Processing and Computer Vision; Original Article; Probability and Statistics in Computer Science; Reconfiguration; Sensors; Wastewater treatment; Data-driven soft sensor; Decentralized dissolved oxygen control; Activated sludge process; ANFIS; Sensor-fault-tolerant control
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-017-2901-3

This work presents a sensor-fault-tolerant design applied to a decentralized dissolved oxygen control in an activated sludge process subject to sensor faults such as bias and slow drifts. The core idea is to use a data-driven soft sensor based on adaptive neuro-fuzzy inference system to act as a backup of the joint sensor and controller block, and to exploit the data/analytical correlations existing in the system. After fault detection and isolation, a control reconfiguration technique takes action in order to surmount/counteract the effect of the fault until the faulty sensor is repaired. The approach presented here was applied to the Benchmark Simulation Model n.1 and was able to demonstrate the improvements on the control system dependability.