An enhanced multipoint optimal minimum entropy deconvolution approach for bearing fault detection of spur gearbox

Previous research has shown that minimum entropy deconvolution (MED) is an effective technique for detecting impulse-like signals, such as the bearing fault and gear fault signals. However, some problems still exist in this technique. With the aim of overcoming these limitations, in this paper, an enhanced MED called multipoint optimal minimum entropy deconvolution adjusted (MOMEDA) is proposed. MOMEDA can succeed in detecting multiple impulses. Unfortunately, according to some simulations and real tests in this work, the results of applying this technique to the fault signals directly were grudgingly acceptable but not very satisfactory, especially under a harsh working condition. This means that MOMEDA is a little sensitive to intensive background noise and vibration interference. To overcome this drawback, a novel mode decomposition method, named time-varying filtering for empirical mode decomposition (TVFEMD), is applied to adaptively eliminate background noise and vibration interference prior to using MOMEDA. According to this proposed method, the weak bearing fault features can be identified clearly. The proposed approach is utilized in bearing fault detection of a spur gearbox and the results show its superiority and effectiveness.