基于相似性修正的风电功率爬坡事件预测方法

发布时间：2018-05-16 21:20

本文选题：风电功率爬坡预测 + 多变量模型　；参考：《中国电机工程学报》2017年02期

【摘要】：为了满足爬坡预测所需的长期高精度风电功率预测要求,提出基于相似性修正的风电功率爬坡预测方法。该方法首先根据Granger因果检测法提取数值天气预报(numerical weather prediction,NWP)中对风电功率预测有效的气象变量,保证基于气象变量和统计模型的混合预测模型的可实现性,并以支持向量回归模型作为基本预测模型。其次,结合历史数据分析气象背景相似性与爬坡事件相似性,给出相似爬坡事件的选取机制。考虑到较高精度的风电功率预测可提高爬坡预测的性能,为此,结合相似爬坡的功率变化修正风电功率的预测结果,并由误差指标分析验证修正模型的优越性。最后,对实际算例进行仿真分析,验证基于相似性修正的风电功率预测模型的可行性。
[Abstract]:In order to meet the requirements of long-term high-precision wind power prediction for slope climbing prediction, a method based on similarity correction is proposed for predicting wind power climbing. In this method, the effective meteorological variables for wind power prediction are extracted from numerical weather prediction NWPs according to Granger causality test method, and the realizability of mixed forecasting models based on meteorological variables and statistical models is guaranteed. The support vector regression model is used as the basic prediction model. Secondly, the selection mechanism of similar climbing events is given by analyzing the similarity of meteorological background and climbing events with historical data. Considering that the prediction of wind power with high accuracy can improve the performance of slope climbing prediction, the prediction result of wind power is corrected by combining the power variation of similar climbing slope, and the superiority of the modified model is verified by error index analysis. Finally, the feasibility of wind power prediction model based on similarity correction is verified by simulation and analysis of practical examples.
【作者单位】：武汉大学电气工程学院;
【基金】：国家重点基础研究发展计划项目(973计划)(2012CB215101)~~
【分类号】：TM614

【相似文献】