考虑爬坡特性的短期风电功率概率预测  

Short-term probabilistic wind power forecast considering ramp characteristics

作　　者：甘迪[1] 柯德平[1] 孙元章[1] 崔明建[1] 

机构地区：[1]武汉大学电气工程学院,湖北武汉430072

出　　处：《电力自动化设备》2016年第4期145-150,共6页Electric Power Automation Equipment

基　　金：国家重点基础研究发展计划(973计划)资助项目(2012CB215101)~~

摘　　要：短期风电功率概率预测有助于调度部门提前安排发电计划,提高风电的消纳能力。提出一种考虑爬坡特性的风电功率概率预测方法,首先通过分析不同风电爬坡定义的特点,阐述互补组合预测的思路;然后采用小波神经网络建立风电功率确定性预测模型,并在其基础上建立不同功率分区内风电爬坡率和风电功率预测误差的二维核密度估计概率预测模型;最后由二者的联合概率分布求取后者的条件概率分布,得到风电功率概率预测结果。仿真结果表明,所提模型具有很高的短期风电功率概率预测精度。Short-term probabilistic wind power forecast is conducive to electric power scheduling and wind power accommodation. A method of probabilistic wind power forecasting with the consideration of ramp characteristics is proposed. The characteristics of different wind power ramp definitions are analyzed and the concept of complementarily combined forecasting is elaborated. The deterministic wind power forecasting model is established based on the wavelet neural network and a probabilistic forecasting model with the two-dimensional kernel density estimation of wind power ramp rate and wind power forecast error is established for different wind power sections. The conditional probability distribution of the latter is calculated by the joint probability distribution for obtaining the probabilistic wind power forecast. The simulative results show that,the proposed model has excellent accuracy of short-term probabilistic wind power forecast.

关 键 词：风电功率 概率预测 风电爬坡事件 小波神经网络 二维核密度估计 

分 类 号：TM614[电气工程—电力系统及自动化]