基于CFOA-MKHSVM的滚动轴承健康状态评估方法

Health state assessment of a rolling bearing based on CFOA-MKHSVM method

为了更有效评估滚动轴承性能退化程度,提出一种混沌优化果蝇算法(CFOA)与多核超球体支持向量机(MKHSVM)相结合的滚动轴承健康状态定量评估方法。该方法针对滚动轴承各状态数据分布不均匀、单一核函数分类存在局限性的问题,提出利用多核核函数的凸组合来优化超球体支持向量机。为消除人为选择分类器多参数的盲目性、避免果蝇优化算法陷入局部最优,将果蝇算法与混沌理论相结合,对多参数进行寻优。同时构建混沌优化果蝇算法-多核超球体支持向量机(CFOAMKHSVM)模型,并提出归一化差别系数评估指标。通过实验研究,与支持向量数据描述(SVDD)算法评估指标进行对比,验证了所提指标的有效性,实现了滚动轴承健康状态的定量评估。

In order to effectively assess performance degradation of the rolling bearing, a quantitative assessment method of rolling bearing health state is proposed combining chaos fruit fly optimization algorithm （CFOA） with multi-kernel hypersphere support vector machine （MKHSVM）. Aiming at the uneven distribution characteristics of rolling bearing each state data and the classification limitation problem of single kernel function, the hypersphere support vector machine is optimized by multi-kernel function convex combination method. In order to eliminate the unreasonableness of artificially selecting parameters of the classifier and avoid locally optimal solution for fruit fly optimization algorithm, the parameters are optimized by using the fruit fly optimization algorithm combined with chaos theory. At the same time, the CFOA-MKHSVM model is constructed, and an assessment index is proposed based on normalized difference coefficient. Comparing with the assessment index of support vector data description （SVDD） algorithm, the effectiveness of the proposed index is verified, and the health state quantitative assessment of a rolling bearing is achieved.