基于广义回归神经网络的风力发电场设备温度自适应预测方法  

Adaptive Temperature Prediction Method for Wind Farm Equipment Based on Generalized Regression Neural Network

作　　者：张二辉 徐兴朝 郑卫剑 贾政 ZHANG Erhui;XU Xingchao;ZHENG Weijian;JIA Zheng(Hebei Xintianke Innovation Energy Technology Co.,Ltd.,Zhangjiakou 075000,China)

机构地区：[1]河北新天科创新能源技术有限公司,张家口075000

出　　处：《自动化与仪表》2024年第10期72-75,共4页Automation & Instrumentation

摘　　要：传统预测方法很难有效处理风力发电场设备温度各种影响因素之间的非线性关系,从而导致预测结果的不准确。针对上述问题,研究一种基于广义回归神经网络的风力发电场设备温度自适应预测方法。分析风力发电场设备温度影响因素并收集这些因素对应的数据,组成样本,对样本实施离群值处理和归一化处理。利用广义回归神经网络自适应预测设备温度并利用鸽群优化算法(PIO算法)自适应调整广义回归神经网络预测模型参数——平滑因子σ,提高其自适应能力。结果表明,所研究方法的预测偏度最高误差仅为0.3℃,说明该方法在预测温度时具有良好的准确性,预测值接近实际值。Traditional prediction methods are difficult to effectively handle the nonlinear relationship between various influencing factors of wind farm equipment temperature,resulting in inaccurate prediction results.To address the above issues,a temperature adaptive prediction method for wind farm equipment based on generalized regression neural network is studied.Analyze the factors affecting the temperature of wind farm equipment and collect data corresponding to these factors to form a sample,and perform outlier and normalization processing on the sample.Using PIO algorithm to adaptively adjust the parameters of the generalized regression neural network prediction model-smoothing factorσ,improve its adaptive ability.The results indicate that the prediction bias of the studied method is small,the maximum error is only 0.3℃,indicating that the method has higher accuracy in predicting temperature and the predicted values are closer to the actual values.

关 键 词：广义回归神经网络 风力发电场 设备温度 PIO算法 自适应预测方法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]