A Renewable Energy Aware Scheduling Approach for Grid Pumped Storage

Pumped storage technology is essential for the power grid industry, and effective scheduling of its operations is a paramount concern, as it can reduce the gap between peak and off-peak power generation. To address the uncertainties from renewable energy sources, this work first develops a rainfall prediction model based on a temporal convolutional network to improve forecasting accuracy. Subsequently, an Integer Linear Programming (ILP) based scheduling approach is proposed for pumped storage operations, incorporating the predicted renewable energy inputs. Case study results demonstrate that the proposed methodology achieves an economic benefit of up to 5.06 million yuan, which is 1.53 times greater than state-of-the-art pumped storage optimization techniques.