基于搜索策略的通用态势评估方法的研究

在态势状态转换图表示的基础上，提出了一种通用的态势评估搜索策略。该策略将态势评估的整个过程看作为对当前态势路径的识别和搜索过程。在算法中以Bayes分类器识别当前态势发展可能的目标状态，以parzen窗算法对态势各类别的类条件概率密度进行修正，以改进的深度优先算法实现态势图的搜索。在实验中，将其应用于丰满水电仿真系统水轮发电机组的状态评估系统中，经数据对比分析，该算法优于传统的发电机诊断分析方法。

Study on General Situation Assessment Method Based on Search Strategy

A general situation assessment search strategy, based on the representation of situation transfer graph, is proposed, and the process of situation assessment is looked upon that of identification and search of situation path in the graph. In the proposed algorithm, Bayesian path classifier is used to classify target situation that is probably reached, Parzen window algorithm is used to modify the conditional probability density of each situation path, and the search in situation graph is realized via the improved depth-first search algorithm. In the experiment, the method had been successfully applied in the running state assessment system of Hydro-generators of Hydroelectric Simulation System of Jilin Fengman. The performance takes precedence over that of the traditional method.