基于概率盒理论的电网规划方案中指标不确定性建模

Index Uncertainty Modeling in Grid Planning Based on Probability Box Theory

随着电网规划要求的日益精化,获取科学可靠的电网评估指标十分迫切.但是,目前研究成果主要侧重于获取指标后的决策指标体系的优化问题,大多忽略了指标的不确定性问题的表达,这将影响整个电网规划的可靠性和经济性.针对指标的不确定性,首先,介绍概率盒处理不确定性信息的优越性;然后,对不同类型指标分别采用不同的方法进行概率盒的建模;最后,以指标负荷的不确定性获取为例,将样本数据分别用概率盒和传统的方法进行处理,并采用支持向量机模式识别的方法进行对比.实验结果表明,基于概率盒理论的建模方法有效可行,并提高了决策方案的准确性.

As grid planning requirements are increasingly refined,the establishment of scientific and reliable grid evaluation indicators has become a matter of urgency. Previous research has mainly focused on the optimization problem of decision indicator systems after the indicators have been obtained,and has ignored the uncertainty of the indicators,which affect the reliability and economy of the entire grid plan. To consider the uncertainty of the indicators,we introduce the superiority of the probability box in dealing with uncertain information,and then use the probability box to model different types of indicators. Finally,using the uncertainty associated with obtaining load indicators as an example,we process sample data by constructing a probability box and by the traditional method,and compare our results with those of the support-vector-machine modeling method. The result shows that the modeling method based on probability theory has great significance for improving decision-making planning.