Short-term load forecasting based on fuzzy gray correlation clustering and AMPSO-BP neural network

Accurate selection of similar days was the key to accurately predict short-term load, and a fuzzy gray clustering method was proposed to select similar days. The neural network could well predict short-term load for prediction due to its strong nonlinear approximation ability and need not to establish mathematical models. However, the commonly used neural network also had the disadvantages of slow learning speed and easy to fall into local minimum values. A Particle Swarm Optimizition with Adaptive Mutation (AMPSO) algorithm was proposed to optimize BP neural network parameters in order to improve the shortcomings of traditional neural networks and improve the prediction accuracy, Finally, using Tangshan power grid data to simulate through matlab, the experimental results show that the proposed load forecasting method has better prediction accuracy and stability.