摘要
提出一种基于高斯混合模型(Gaussian mixture model,GMM)的配电网负荷量测权重优化算法,包括对GMM参数的优化和权重确定。首先采用引力搜索算法(gravitational search algorithm,GSA)对数据的最佳聚类个数进行判断,利用K-means算法获取数据的初始聚类中心、方差和混合权重;然后通过组合马尔科夫链蒙特卡洛期望最大化(Markov chain Monte Carlo-expectation maximum,MCMC-EM)算法对GMM的参数进行估计;最后根据优化的GMM,提出负荷伪量测权重优化方法,确定负荷伪量测的权重。以改进IEEE-12节点系统对所提方法进行验证,结果表明其合理、有效。
A kind of optimization algorithm for load measurement weight of power distribution network based on Gaussian mixture model（GMM）is presented which includes optimization and weight confirmation for GMM parameters.It firstly uses gravitational search algorithm（GSA）to judge optimal clusteringnumbers of data and K-means algorithm to obtain an initial clustering center,variance and mixed weight of data.Then it uses Markov chain Monte Carlo-expectation maximum（MCMC-EM）algorithm to estimate GMM parameters.Finally it presents the optimization method for pseudo measurement weight of load according to optimized GMM so as to confirm pseudo measurement weight of load.Improved IEEE-12 node system is taken for an example to verify the proposed method and the result indicates its reasonability and effectiveness.
出处
《广东电力》
2016年第5期86-91,123,共7页
Guangdong Electric Power
关键词
配电网
状态估计
伪量测权重
高斯混合模型
组合马尔科夫链蒙特卡洛期望最大化算法
power distribution network
state estimation
pseudo measurement weight
Gaussian mixture model（GMM）
Markov chain Monte Carlo-expectation maximum（MCMC-EM）
