基于MEA优化BP神经网络的天然气短期负荷预测  

Short-Term Gas Load Forecasting Based on MEA Optimized BP Neural Network

作　　者：张以帅[1] 赖惠鸽[1] 李勇[1] 唐光耀[1] 张晨艺 

机构地区：[1]宁夏大学机械工程学院,银川750021 [2]山东省产品质量检验研究院,济南250102

出　　处：《自动化与仪表》2016年第5期15-19,共5页Automation & Instrumentation

摘　　要：天然气负荷中包含大量非线性因素,单一的神经网络很难达到理想的预测精度,为了提高预测效果,提出了一种思维进化算法(MEA)优化BP神经网络智能预测模型。利用MEA的全局搜索性对BP神经网络的权值和阈值进行优化,避免了单一BP网络的局部最优和过拟合等缺点,然后建立最优预测模型。将这种组合模型应用于银川某县的天然气负荷预测,结果表明该组合模型具有更优的非线性映射能力和更高的预测精度。Natural gas load contains a great number of nonlinear factors which make a single neural network forecasting system difficult to achieve anticipated accuracy. So in order to improve the forecasting accuracy,a BP neural network intelligent prediction model based on the mind evolutionary algorithm（ MEA） has been proposed. By taking advantage of the MEA＇s overall-searching competence to optimize the weights as well as the thresholds of the BP neural network,in this way to avoid the defects as partial-optimization and over-fitting. After that,a optimal predictive model was founded. The combination model was put into effect on the forecasting of gas load in a certain county in Yinchuan,and the results showed that it has better nonlinear mapping competence and higher prediction accuracy.

关 键 词：天然气 BP神经网络 思维进化算法 负荷预测 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]