On-line hybrid soft sensor for water quality COD based on synchronous clustering

摘要

Aiming at the problem that the wastewater treatment process is affected by many uncertain factors leading to frequent fluctuations of operating conditions, single water quality model cannot guarantee the accuracy of soft sensor, an online hybrid soft sensor of effluent COD based synchronous clustering is presented in this paper. The soft sensor consists of simplified mechanism model and error compensation model, of which the simplified model based ASMI is used to describe the basic characteristics of dynamic mechanism in the wastewater treatment process; the error compensation model based integrated linear model is used to compensate mode errors under different operating conditions. Synchronization clustering algorithms can achieve clustering of online data in the case of unknown number of integrated model, it improves the real-time performance of soft sensor, and linear model in error compensation model reduces the computational cost. The simulation results show that COD online hybrid soft sensor has better prediction accuracy under multiple working conditions.