Using Model Predictive Control in a Water Infrastructure Planning Model for the Zambezi River Basin

Abstract

Hydroeconomic optimization models considering the inter-relations in the water-energy-food nexus are potential tools to evaluate water infrastructure and policy development that will contribute to multiple Sustainable Development Goals. However, most of these models are deterministic and assume perfect foresight. Thus, the optimal management decisions are found with perfect knowledge of future conditions, which might bias the economic evaluation of infrastructure investments. We show how the Model Predictive Control (MPC) framework can be used to overcome the perfect foresight assumption. By implementing the MPC framework in WHAT-IF, a perfect foresight hydroeconomic optimization model, we show how MPC leads to more realistic simulated reservoir operations and consequently to a more realistic valuation of investments. To evaluate the impact of the perfect foresight assumption, we evaluate infrastructure investments in the Zambezi river basin with and without the MPC framework. We find significant differences (up to 12%) between the perfect foresight and MPC frameworks when estimating the value of hydropower and irrigation investments. By carrying out the analysis for four different climate change scenarios, we find that the impact of the perfect foresight assumption is particularly important in a water scarce context.

Optimizing the Layout of Run-of-River Powerplants Using Cubic Hermite Splines and Genetic Algorithms

“…Moreover, ref. [12] presents the most important features of the turbine model from a control point of view, and, in [13], the use of model predictive control to deal with prediction errors in the available water in the reservoir and is proposed.…”