基于三维依赖矩阵的故障预测策略设计

Fault Prognostic Strategy Design Based on Three-dimensional Dependency Matrix

针对不同基于状态监测的故障预测方法难以在一个平台下集成和组合的问题,设计了基于故障、测试、衰退信息三维依赖矩阵的故障预测策略设计框架,在框架内给出了故障预测需求设计和预测用测试选择的基本方法,并在灰色预测和支持向量回归预测的基础上,通过融合粒子群和声算法给出了一种灰色支持向量回归智能组合衰退趋势预测分析方法,结合故障预测实例验证了所给方法的有效性和准确性。

In order to solve the problem of being difficult to integrate different status monitoring based prognostic algorithms in a plat- form, architecture for design for prognostic strategy based on fault, test and degradation information three--dimensional dependency matrix is given out. Methods of prognostie requirement design and test selection for prognosis are given out. On basis of gray prognosis and support vector regression, a gray support vector regression prognostic method is given out by fusing particle swarm harmony search algorithms. At last, an example of fault prognosis is given out to validate the effectiveness and correctness of the method.