Development of a Hybrid Model and Intelligent Control System for the Water Ozonation Process

This article addresses the problem of improving the efficiency of water treatment by ozonation under conditions of unstable raw water composition and limited observability of process parameters. The continuous ozonation process in a perfectly mixed contact reactor is treated as a complex control object, and a hybrid mathematical model is developed that combines physicochemical reaction kinetics, mass-transfer processes between the gas and liquid phases, and artificial intelligence methods. The model describes the joint dynamics of dissolved ozone, total pollutant load, organic matter, dissolved iron and manganese, microbial concentration, pH, and temperature, while treating raw-water characteristics and hydrodynamic conditions as disturbance inputs. A software module was created for predicting treatment efficiency, optimizing ozone dosage, and controlling the process based on the OWEN PLC۲۱۰ programmable logic controller. An animation model for real-time visualization of the technological process was implemented as the basis of a digital twin. A control algorithm based on model predictive control (MPC) and the extended Kalman filter (EKF) was proposed. Numerical simulations of the hybrid model demonstrate adaptive ozone dosing and a monotonic improvement in treatment efficiency; a separate test case of the standalone MPC/EKF loop, however, revealed a parameterization under which the controller fails to reach the target pollutant concentration, indicating a model–parameter mismatch that requires further calibration before field deployment. The obtained results, together with the identified limitation, informed the development of digital twins for water treatment systems.

کلیدواژه ها:

Ozonation ، Water Treatment ، intelligent control system ، Model predictive Control ، Extended Kalman Filter ، Digital Twin