Balancing Operator's Risk Averseness in Model Predictive Control of a Reservoir System

Ja-Ho Koo,Edo Abraham,Andreja Jonoski,Dimitri P. Solomatine

2024-07-05

Abstract:Model Predictive Control (MPC) is an optimal control strategy suited for flood control of water resources infrastructure. Despite many studies on reservoir flood control and their theoretical contribution, optimisation methodologies have not been widely applied in real-time operation due to disparities between research assumptions and practical requirements. First, tacit objectives such as minimising the magnitude and frequency of changes in the existing outflow schedule are considered important in practice, but these are nonlinear and challenging to formulate to suit all conditions. Incorporating these objectives transforms the problem into a multi-objective nonlinear optimisation problem that is difficult to solve online. Second, it is reasonable to assume that the weights and parameters are not stationary because the preference varies depending on the state of the system. To overcome these limitations, we propose a framework that converts the original intractable problem into parameterized linear MPC problems with dynamic optimisation of weights and parameters. This is done by introducing a model-based learning concept under the assumption of the dynamic nature of the operator's preference. We refer to this framework as Parameterised Dynamic MPC (PD-MPC). The effectiveness of this framework is demonstrated through a numerical experiment for the Daecheong multipurpose reservoir in South Korea. We find that PD-MPC outperforms `standard' MPC-based designs without a dynamic optimisation process under the same uncertain inflows.

Systems and Control

What problem does this paper attempt to address?

This paper primarily discusses how to balance the risk aversion of operators when using Model Predictive Control (MPC) in reservoir management systems. Currently, although MPC has theoretical advantages in flood control, it is not widely used in practical applications due to the gap between assumptions in research and actual requirements. Operators typically consider implicit objectives such as reducing the frequency and magnitude of outflow changes, but these objectives are difficult to accurately express and solve.

The paper proposes a parameterized dynamic MPC (PD-MPC) framework to address the following issues: first, transforming nonlinear implicit objectives into manageable linear MPC problems by dynamically optimizing weights and parameters to adapt to changes in system states; second, considering the short-term prediction window and the uncertain length of flood events, operators cannot always rely on long-term rainfall or inflow predictions.

Specifically, the paper proposes several practical flood control objectives, including minimizing the peak and total outflow through the spillway, limiting the frequency of outflow changes, maintaining the reservoir water level close to the target range, avoiding the reservoir water level exceeding a specific range, and protecting the continuous operation conditions of the spillway. In addition, the paper considers the dynamic changes of operator preferences with hydrological conditions and solves the multi-objective nonlinear optimization problem by parameterized dynamic MPC framework combined with Genetic Algorithm (GA) to dynamically optimize weights and parameters at each time step.

Through numerical experiments, the paper demonstrates that PD-MPC performs better than standard MPC design in the face of uncertain inflows, and can more effectively cope with complex flood situations.