基于遗传程序设计的回转支承寿命状态识别

Life state recognition of slewing bearing based on genetic programming

针对回转支承低转速、故障信号微弱的特点,提出了一种遗传程序(GP)设计的方法对其寿命状态进行准确的识别。为保证回转支承运转信息的完整性,该方法从不同领域提取了多个特征指标组成特征向量矩阵。以模型的性能和复杂度为衡量指标,从遗传程序设计建立的模型中选择出最佳模型,再将测试样本输入模型实现对回转支承寿命状态的识别。利用自主研发的回转支承综合性能实验台对某型号的回转支承进行了全寿命疲劳实验,实验结果表明,所提出的方法能够准确地识别出回转支承的寿命状态,为实时维修奠定了基础。

Genetic programming（GP） was used to deal with the weak fault feature of low-speed slewing bearing.The feature indexes composed of feature matrix were extracted from different domains to guarantee the integrity of information.The performance and complexity were chosen as indexes for the best model selection from the genetic programming.The life state of slewing bearing was identified by the testing data with the best model.Based on the test rig,experiments on the full life test of slewing bearing were also conducted.Results showed that the proposed method could be used to recognize the life test of slewing bearing accurately,which was the foundation of real-time maintenance.