摘要
结合隐马尔可夫模型(HMM)所需训练样本少及可解释的优点,提出了基于HMM的矿井提升机故障诊断方法。利用多个加速度传感器在提升机运行的不同转速阶段采集数据,通过快速傅里叶变换(FFT)从提升机振动信号中进行特征抽取后,再由劳埃德算法(Lloyd)进行标量量化,根据HMM建模理论,训练HMM诊断库,再利用训练好的HMM对提升机进行状态监测和故障诊断。
Because of the advantage of Hidden Markov Model （HMM） which requires less training samples and explained easily, puts forward a method for the fault diagnosis based on HMM. Using multiple acceleration sensors in different rotating speed of the elevator operation stage to collect data, second, through the FFT method, the feature extraction is carried out from the vibration signal of the elevator, then, using the method of Lloyd algorithm to scalar quantization. Based on the HMM modeling theory, training the HMM diagnostic library, trained HMM is used to analysis the hoisting machine condition monitoring and fault diagnosis.
出处
《煤炭技术》
CAS
北大核心
2017年第2期254-256,共3页
Coal Technology
基金
国家自然科学基金项目(51475318)
