新能源组合出力的置信区间估计

Estimation on Confidence Intervals of Combined Output of Renewable Energy

随着新能源的不断发展,电力系统中常涉及新能源组合出力的分布问题。文中综述了几种常见新能源出力的分布特性,基于其分布函数,给出了多种新能源组合出力的累积分布函数和概率密度函数,提出了求解新能源组合出力置信区间的2种算法:二分查找—数值积分法和基于拉丁超立方抽样的蒙特卡洛法。通过一个三组合出力的具体算例的分析可知,利用推导的公式能准确给出其组合出力的概率分布曲线。对于置信区间的计算,二分查找—数值积分法平均花费228.25s,但具有较高的计算精度,而抽样类方法计算时间均在1s以内。基于拉丁超立方抽样的蒙特卡洛法抽样1 000次的效果即可达到随机抽样10 000次的效果,且其所用的计算时间较少,平均误差低,是一种较好的计算方法。

With the development of renewable energy, the probability distribution of combined output of renewable energy generation becomes an outstanding issue in the power system. This paper summarizes the distribution characteristics of several common renewable energy, and proposes the cumulative distribution function and probability density function of the combined output. In addition, two methods are proposed to get the confidence intervals： dichotomy search numerical integration method and Latin hypercube sampling. Monte Carlo method. Numerical tests are conducted with a microgrid system. Testing results indicate that the proposed formula is accurate. Additionally, dichotomy search numerical integration method has higher accuracy hut has a time consumption of 228.25 s, while sampling methods need less than 1 s. Compared to random sampling, Latin hypercube sampling Monte Carlo method has higher sampling efficiency, higher accuracy and lower time consumption.