摘要
提出一种基于随机过程自相关性的风速预测模型,在传统概率模型分析的基础上引入了随机过程的概念,将每个时刻的风速均看作是一个随机变量,利用随机过程多维分布函数的统计特性描述风速过程。为了表征风速在时间上的自相关特性,引入连续马尔科夫模型,依据前述随机过程模型,求解马尔科夫模型的状态转移函数,从而表征风速相邻时刻间的演化规律,并从理论上证明了该模型具有较好保持自相关特性的能力。仿真表明,利用该模型能够更好地模拟风速分布,预测风速大小,并且具有良好的置信度。
It is of increasing importance to predict wind speed in times of wind energy widely utilized. This paper proposed a new wind speed prediction model based on random process, taking autocorrelation of raw data into consideration. Concept of random process was introduced in this model in parallel with traditional probability model. This model took wind speed at every moment as a random variable and utilized multidimensional distribution function of random process to describe wind changing process. In order to feature wind＇s time autocorrelation, this paper introduced continuous Markov chain model and solved Markov chain state transfer function based on random process model to characterize evolution law of wind speeds between adjacent times. Simulation shows that the model can better simulate wind speed distribution and predict wind speed with higher confidence level.
出处
《电网技术》
EI
CSCD
北大核心
2017年第2期529-535,共7页
Power System Technology
基金
国家科技支撑计划项目(2015BAA01B04)
国家电网公司科技项目(522727160002)~~
关键词
随机过程
正态过程模型
连续马尔科夫链模型
状态转移函数
random process
normal distribution process model
continuous Markov chain model
state transfer function
