论文部分内容阅读
在故障诊断中,从不完备数据中获取规则要比从完备数据中获取规则困难。利用给出的分辨矩阵基元的定义,提出了一种直接从不完备数据中获取最优广义诊断决策规则的粗糙集方法。该方法以极大相容块为单位构造了不完备故障决策表的分辨矩阵中的列元素,实现了不完备故障诊断决策表中面向对象的约简计算和最优广义故障诊断规则的获取。该方法不需要改变原始不完备故障诊断决策表的规模,且具有更高的约简计算效率。结合电力系统操作点的安全状态诊断实例给出了所提出的方法在工程实践中的应用步骤,并证明了该方法的有效性。
In troubleshooting, getting rules from incomplete data is more difficult than getting rules from complete data. Using the definitions of the discernable matrix elements, a rough set method is proposed to obtain the optimal generalized diagnostic decision rules directly from the incomplete data. The method constructs the column elements in the discernibility matrix of incomplete decision table in the unit of maximally compatible block, and realizes object-oriented reduction calculation and optimal generalized fault diagnosis rules in incomplete decision table. This method does not need to change the scale of the original incomplete fault diagnosis decision table and has higher reduction computational efficiency. The practical application of the proposed method in engineering practice is given with examples of the safety diagnosis of the power system operating point. The effectiveness of the proposed method is proved.