A Fuzzy Performance Evaluation Model for a Gearbox System Using Hidden Markov Model

摘要: In order to track and grasp the operation situations of gearboxes, vertical vibration signals three different gear fault states, normal, worn broken teeth, are collected via a gearbox experiment. An online diagnosis performance evaluation model with hidden Markov (HMM) fuzzy comprehensive is proposed. To address limitation maximum membership principle in case equal or very close each other, closeness strategy proposed by defining likelihood ratio HMM as similarity selecting an combined function semi-trapezoidal intermediate-ridge distribution. Results show that has achieved good strategy. Compared principle, more accurately distinguished from results teeth state state, especially for membership.