Improved BP neural network algorithm to wind power forecast

• Published in: Journal of engineering (Stevenage, England) Vol. 2017; no. 13; pp. 940 - 943
• Main Authors: Wang, Zheng, Wang, Bo, Liu, Chun, Wang, Wei-sheng
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 2017
• Subjects: backpropagation; big noise; forecasting power; improved BP neural network algorithm; maximum correntropy criteria algorithm; maximum entropy methods; MCC algorithm; mean square error methods; neural nets; nonGaussian error; power engineering computing; power generation dispatch; power grid dispatching uncertainty reduction; real power; root MSE reduction; wind power; wind power forecast; wind power plants; wind power prediction; real power; wind power forecast; power engineering computing; power generation dispatch; power grid dispatching uncertainty reduction; wind power; root MSE reduction; wind power prediction; improved BP neural network algorithm; big noise; backpropagation; nonGaussian error; maximum correntropy criteria algorithm; maximum entropy methods; forecasting power; MCC algorithm; mean square error methods; wind power plants; neural nets
• ISSN: 2051-3305; 2051-3305
• DOI: 10.1049/joe.2017.0469

To constantly enhance the accuracy of wind power prediction and furthermore reduce the uncertainty of power grid dispatching, this study proposes an improved back propagation (BP) neural network algorithm. The original prediction method of BP neural network algorithm has been improved, and the traditional minimum square error (SE) perform function is abandoned. Maximum correntropy criteria (MCC) algorithm which is more conducive to deal with non-Gaussian error and big noise is introduced, and a new perform function is created. Through the analysis of examples, the feasibility of MCC algorithm is verified. Comparing to the traditional mean SE (MSE) perform function, MCC perform function could drop the limit error of prediction, reduce root MSE and increase the correlation between forecasting power and real power. The most important is that the prediction accuracy is enhanced.