基于自适应线性逻辑网络的风电功率预测方法性能评估与分析  

Performance assessment and analysis of wind power forecasting method based on adaptive linear logic network

作　　者：王媛媛[1] 秦政[1] 张超 包德梅[1] 赖晓路[1] 岳以洋[1] 

机构地区：[1]国电南京自动化股份有限公司,江苏南京210003 [2]中国中材国际工程股份有限公司自动化电气设计研究所,江苏南京211100

出　　处：《可再生能源》2013年第6期61-65,共5页Renewable Energy Resources

基　　金：中国华电集团科技项目(CHFKJ12-01-22)

摘　　要：介绍了一种基于自适应线性逻辑网络(ALLN)的风电功率预测方法,描述了ALLN算法的特点、基本结构、建模方法以及预测与学习方法;采用该方法进行了风速预测试验,对试验结果进行了对比分析,详细分析了影响试验结果的因素。首次提出了模型适应性的概念,并给出了计算公式。从模型训练输入数据和预测输入数据的角度,分析了影响ALLN算法模型适应性与预测结果的主要因素。A wind power forecasting method based on adaptive linear logic network （ALLN） is introduced in this paper. The characteristics, basic mechanism and modeling methods, prediction and studying methods of ALLN algorithm are given in details. Wind speed forecasting experiments are carried out and then experimental results are compared, analyzed and evaluated to obtain factors which effect on testing results. The concept of model adaptability and its definition are brought forward for the first time. In the end, the main factors which have effects on ALLN forecasting results are analyzed viewing from input data used for model training and input data used for wind speed forecasting.

关 键 词：风电功率预测 短期预测 自适应线性逻辑网络 预测性能评估 预测模型 

分 类 号：TK83[动力工程及工程热物理—流体机械及工程] TP302[自动化与计算机技术—计算机系统结构]