Power system emulation for rapid exploration of water-energy policies

Abstract Resource systems like energy and water enable human development and economic growth. The strong linkages between river basins, which include irrigated agriculture and freshwater ecology, and electricity grids motivate adopting an integrated multiobjective approach to their planning and management to increase service levels, seek economic efficiency, and promote wider societal goals. Connecting multi-sector infrastructure system computer simulators to multi-objective search algorithms offers a practical tool for identifying synergies in these interdependent systems. However, this approach is constrained by the computational resources required to run joint energy-water simulation models many times to explore the multi-objective search space. The proposed framework links a fast power system model emulator to a river system simulator, then connects that integrated model to a multi-objective design process. The emulator in our case is an artificial neural network, the search algorithm is a multi-objective evolutionary optimizer. We compare optimized designs that use emulation with those from a conventional physical simulator to check the emulators' ability to guide a computer-aided multi-objective design process. The framework is applied to the Sudanese power system and Eastern Nile River Basin; the design problem is to search for operational strategies of the Grand Ethiopian Renaissance hydropower Dam (GERD) considering multi-sector and multi-country performance objectives. Using the power system emulator in the design process leads to results similar to ones generated using conventional simulation. The proposed emulation method provides a promising new approach to explore sensitivities, synergies, and trade-offs in multi-sector resource systems.