论文部分内容阅读
针对基于机匣测点信号的航空发动机滚动轴承早期故障诊断问题,提出了一种基于正则化多核判别分析的融合诊断方法.该方法首先提取多种类型的滚动轴承故障特征;然后采用相同的一组核参数为不同类型的特征分别构造一组核矩阵,并将所有核矩阵组合在一起;最后通过求解一个半无限线性规划得到该组核矩阵关于正则化核判别分析的目标函数的最优线性组合系数,进一步采用该系数计算所有核矩阵的线性组合,从而实现多种类型特征信息的融合.实验结果表明:该方法诊断正确率与采用单一类型特征诊断的最高正确率相比提高了9.25%,同时可以避免核矩阵需要人工选择的问题,从而进一步提高了故障诊断的自动化水平.
In order to solve the early fault diagnosis of aeroengine bearing based on receiver signal, a new fusion diagnosis method based on regularized multi-core discriminant analysis is proposed. Firstly, the fault features of many types of rolling bearings are extracted. Then, Parameters for different types of features, respectively, to construct a set of nuclear matrix, and all the nuclear matrix together; finally by solving a semi-infinite linear programming to obtain the nuclear matrix of regularized nuclear discriminant analysis of the objective function of the optimal linear combination of coefficients , The linear combination of all the kernel matrices is further used to achieve the fusion of various types of feature information.The experimental results show that the diagnostic accuracy of this method is improved by 9.25% compared with the highest correct rate of single type feature diagnosis, The problem of manual selection of the nuclear matrix can be avoided, thereby further improving the automation of fault diagnosis.