基于SVDD与信息融合技术的设备性能退化评估

Equipment performance degradation assessment based on SVDD and information fusion technology

为了能够准确地对大型设备的性能退化过程进行描述,提出了一种基于支持向量数据描述(SVDD)和信息融合技术的评估方法。通过SVDD算法分别评估来自单个传感器的数据,然后运用D-S证据理论对来自多传感器的局部评估结果进行信息融合,最终给出设备的整体性能评估结果。实验分析表明,SVDD算法能够真实地反映设备局部性能退化状态的变化,而利用D-S证据理论得出的整体设备状态评估结果符合实际情况,同时有效地消除局部信息之间的矛盾,提高了设备整体评估的可靠性。

To improve the accuracy of description of equipment performance degradation process,a novel method for performance degradation assessment was proposed,it was based on two techniques: support vector data description (SVDD) and dempeter-shafer(D-S) theory. SVDD was used to assess the performance according to the data from a single sensor,and the performance assessment results based on multi-sensor were sent to D-S theory as input to obtain the assessment result of the whole system. As shown in experiments,SVDD could reflect the performance degradation of equipment parts,the results using D-S theory accorded with the practical situation,and the algorithms could deal with the evident conflict problem effectively,it improved the accuracy and reliability of the assessment results.