关键词: 短期电力负荷, 负荷预测, 深度信念网络, 混合蜂群优化, 多变量耦合

Abstract:

The traditional load forecasting algorithm is difficult to effectively deal with the nonlinearity in the power load data.A novel multivariable coupling forecasting algorithm for short-term power load is designed.The deep belief network is used to automatically learn the complex coupling relationship between multiple variables and the deep feature representation,by which the effective extraction of nonlinear features and deep features is realized.At the same time,the hyper parameter adaptive selection strategy is introduced into the traditional bee colony optimization algorithm to form a hybrid bee colony optimization algorithm,which realizes the global optimization of key parameters,and effectively improves the generalization ability and prediction accuracy of the algorithm.The results show that the prediction accuracy and prediction efficiency of the proposed algorithm are exceed 99%,which realizes the effective and accurate capture of the overall trend of load change,and lays an important foundation for the real-time scheduling and optimization of the power system.

Key words: short-term power load, load forecasting, deep belief network, hybrid bee colony optimization, multivariable coupling