基于小波神经网络的中长期电力负荷预测

电力系统负荷预测是1项复杂的系统工程,其不仅涉及的领域广泛,而且不确定性的因素较多。文中在传统BP神经网络算法、改进型BP神经网络算法基础上,将BP神经网络与小波分析相结合,构建了小波神经网络模型,然后分别应用BP神经网络、改进型BP神经网络和小波神经网络对宁夏石嘴山地区电力负荷进行了中长期预测。通过对比分析表明,采用小波神经网络获得的预测数据比前2种方法获得的预测数据误差均要小。这说明了小波神经网络的预测结果更加准确,即采用BP神经网络与小波分析相结合的方法比单纯地采用BP神经网络算法进行电网负荷预测的效果更佳

Mid-long-term Power Load Forecasting Based on Wavelet Neural Network

Load foreeasting is a complex and tedious work, It involves many fields widely anti has so much uncertain filetors. Based on the traditinnal BP neural network algorithm and improved BP neural network algurithm,the paper adopts tile method of wavelet neural nelwork which combined BP neural network with wavelet analysis, Then it forecasts respectively the mid-long-term load in the region of Shizuishan with Ihe three methods. Through comparative analysis, it is showed thai Ihe errors of using wavelet neural network In obtain data are excessively smaller than lhem of using the previous two methods. The wavelet neural network prediction is more accurate. Using BP neural network algorithm combined with wavelet analysis is better than simply using the BP neural nelwork algorilhm in forecasting power load.