一种多链MCMC电力系统光伏出力预测模型及其算例分析

为了提高电力系统光伏出力预测精度,论文建立了一种应用多链马尔科夫-蒙特卡洛(MCMC)算法的预测模型,并展开算例分析。研究结果表明:通过多链算法生成的序列概率密度度达到了和历史序列非常接近程度,有效反映了对历史序列实施统计的情况。在光伏电站达到较高的相关性时,采用多链算法计算出的预测序列均值和标准差比历史序列误差更小,因此能够保持明显的历史序列统计特征。利用二种算法计算得到的预测序列相关系数,采用多链算法可以得到比历史序列更优的自相关曲线,有效保留了原序列的时间相关性。由此可见利用论文算法分析多光伏电站出力时间序列是可行的。

A Multi-chain MCMC Power System Photovoltaic Output Prediction Model and Its Example Analysis

In order to improve the accuracy of photovoltaic output prediction of power system,a prediction model based on multi-chain Markov chain Monte Carlo(MCMC)algorithm is established,and an example analysis is carried out.The results show that the probability density of the sequence generated by the multi-chain algorithm is very close to the historical sequence,which ef?fectively reflects the implementation of statistics on the historical sequence.When the photovoltaic power station reaches a high cor?relation,the error of the mean and standard deviation of the prediction series calculated by the multi-chain algorithm is smaller than that of the historical series,so the statistical characteristics of the historical series can be maintained obviously.The correlation coef?ficients of prediction sequences are calculated by using the two algorithms,and the multi-chain algorithm can get a better autocorre?lation curve than the historical sequence,can effectively retain the time correlation of the original sequence.Therefore,it is feasible to analyze the output time series of multi-photovoltaic power stations by using this algorithm.