基于最优测试点的小波包-隐马尔科夫模型故障诊断  

Fault Diagnosis System of Wavelet Packet and Hidden Markov Model Based on Optimal Test Point

作　　者：鞠建波[1] 胡胜林[1] 祝超[1] JU Jian-bo HU Sheng-lin ZHU Chao(Navy Aeronautics and Astronautics University, Yantai 264001, China)

机构地区：[1]海军航空工程学院,山东烟台264001

出　　处：《兵器装备工程学报》2017年第4期91-94,共4页Journal of Ordnance Equipment Engineering

基　　金：国家自然科学基金资助项目(60874112);军队科研专项资助项目(41512322)

摘　　要：采用小波包变换和隐马尔科夫模型结合的方法,提出了机载通信电台的故障诊断模型。通过评价函数的建立,得到最优测试点集合;利用小波包变换提取测试点的能量特征值作为模型的输入,利用历史数据对隐马尔科夫模型进行训练,得到各个状态下的HMM参数,通过模式识别,实现电台的故障诊断;试验研究结果验证了所提出的方法的可行性和有效性。By using wavelet packet transform and hidden Markov model,the fault diagnosis of airborne radio communication model was put forward. Through the establishment of the evaluation function,the optimal test set point was obtained. Using wavelet packet transform,we extracted the test point of energy eigenvalues for the input of the model,and the hidden markov model was trained using the history data,and we got the HMM parameters of each state,so as to realize the fault diagnosis of the radio by pattern recognition; Experimental results verified the feasibility and effectiveness of the proposed method.

关 键 词：故障诊断 最优测试点 小波包分析 隐马尔科夫模型 模式识别 

分 类 号：TB114[理学—概率论与数理统计]