Hydropower Scheduling with Power Load Prediction: Optimizing Energy Efficiency and Navigation Performance

Abstract Hydropower stations integrated into the grid system often suffer from imbalanced scheduling of power generation and shipping benefits due to untimely acquisition of load data. This paper proposes a joint scheduling model for hydropower stations based on power load forecasting and applies it practically to the Shatuo Hydropower Station in Guizhou Province of China. The model uses forecasting algorithms to predict power load data and optimization algorithms to schedule the discharge flow for each time period. Experiments show that the CNN-GRU model (i.e., the Convolutional Neural Network - Gate Recurrent Unit) is superior to other models in load balancing data for water stations. Using predicted power load data and the GA-NSGA-II algorithm (i.e., the improved elite non-dominated sorting genetic algorithm), the scheduling results are optimized, making the Shatuo Hydropower Station achieve both power generation and navigation benefits. This approach solves the problems of untimely load data acquisition and insufficient scheduling, providing an effective way for hydropower generation navigation scheduling.