QGA-VPMCD智能诊断模型研究  

QGA-VPMCD intelligent diagnosis model

作　　者：杨宇[1] 李紫珠 何知义 程军圣[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《振动与冲击》2015年第13期31-35,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金(51175158;51375152);湖南省自然科学基金(11JJ2026)

摘　　要：针对多变量预测模型模式识别(Variable Predictive Model-based Class Discriminate,VPMCD)分类方法中只选择了某单一模型的缺陷,提出一种基于量子遗传算法优化的多变量智能诊断模型(Quantum Genetic Algorithm-Variable Predictive Model-Based Class Discriminate,QGA-VPMCD)。该模型采用最优权值矩阵来综合考虑各诊断模型对分类结果的影响。即首先通过样本训练来建立多个SVPM(Subordinate Variable Predictive Model,SVPM);然后采用量子遗传优化算法求出各SVPM的权值,从而得到最优权值矩阵;最后用最优权值矩阵加权融合测试样本的SVPM特征变量预测值,得到最佳特征变量预测值,并以预测误差平方和最小为判别函数来识别故障的类型。滚动轴承振动信号的分析结果表明了该模型的有效性。Aiming at the defect that only a single model is selected in the variable predictive model-based class discriminate （VPMCD）classification method,an intelligent diagnosis model called quantum genetic algorithm -variable predictive model-based class discriminate （QGA-VPMCD） was presented.The optimal weight matrix was used to comprehensively consider the effect of each diagnosis model on classification results with this new model.Firstly,multiple subordinate variable predictive models （SVPMs）were established through samples-training.Secondly,the intelligent quantum genetic algorithm was used to acquire the weight value of each SVPM and the optimal weight matrix was obtained.Finally,the optimal weight matrix was used to get the optimal feature variable predictions through weighted fusing feature variable predictions of SVPMs of test samples.Fault types were identified according to the minimum error square sum taken as the discrimination function.The analysis results of vibration signals of rolling bearings verified the effectiveness of the proposed model.

关 键 词：多变量预测模型 量子遗传算法 最优权值矩阵 智能诊断模型 

分 类 号：TH113[机械工程—机械设计及理论]