Proceedings

Introduction Increasing levels of instrumentation across wastewater treatment works have produced a wider range of information on process operation. There are also increased incentives for process optimisation, for example due to the rising cost of energy used, and the reduction in the size of the operational staff workforce. This changes the manner in which these works are operated, with a more focussed objective of real-time dynamic control and pushing towards energy, cost and operational expenditure targets. Whilst traditionally, the wastewater industry has operated plants in a conservative manner, minimising risk of consent failure by over-treating, there is now a move towards a more balanced approach to optimisation of the plant with respect to multiple objectives. While the reigning priority is still quality objectives and meeting regulatory consents, the multivariable optimisation approach provides more scope for operational savings, exploiting the wealth of information that is now more readily available and facilitating a more coherent ‘risk based’ approach to final effluent quality.

Keywords Control, Optimisation, Model Predictive Control, Wastewater, Efficiency Data Based Strategies

This paper details the application of a robust data-based control and optimisation strategy that has been successfully implemented on a wastewater treatment process. A key consideration when utilising online process data is the reliability of the information available. The level of robustness of the control system depends heavily on the availability of reliable data, or – in the case where the data quality is questionable – the ability to ‘reconstruct’ that information from other correlated process variables. The important objective here is to determine when the process is behaving abnormally, as opposed to an unusual process value being due to an instrumentation problem. For operators to ‘believe’ in closed loop control of their process, they must be satisfied that the new control system is capable of operating safely throughout a range of process conditions, and does not need constant maintenance to perform its function correctly.