Model Predictive Controller For Water Treatment Process

  • Beatrice C. Langat
  • Stanley I. Kamau
  • Peterson K. Hinga


Water treatment processes incur relatively long  transport delays, simply because the process variable is controlled by the addition of chemicals where a reaction time is allowed before the process variable can be sampled. This process dead time makes it difficult to control water treatment plants using standard feedback techniques mainly because the control action takes some time to affect the controlled variable. The feedforward control when used exhibits low performance and instability when the flow rates vary rapidly and when there are large changes in other water quality variables. During the rainy seasons, raw water quality changes frequently and widely posing a challenge in the water coagulation process where the application of optimum amount of chemicals is required in  order to meet the laid down standards. The process being continuous and with no feedback, the consumers may receive water that does not meet the set down quality standards. This paper focuses on the application of Model Predictive Control (MPC) to control of turbidity in the water treatment plants. The system monitors the incoming water quality and prescribes an optimized coagulation chemicals dosage for the process before large changes in turbidity values can be seen in the outlet.
Oct 7, 2019
How to Cite
LANGAT, Beatrice C.; KAMAU, Stanley I.; HINGA, Peterson K.. Model Predictive Controller For Water Treatment Process. Proceedings of Sustainable Research and Innovation Conference, [S.l.], p. 25-30, oct. 2019. ISSN 2079-6226. Available at: <>. Date accessed: 05 july 2020.