<p>Spindle bearing is one of the machine elements in the spindle that is mostly vulnerable to failure. Its failure may result into total machine tool breakdown and other associated catastrophic consequences. An early identification of the failure is emphasized for reducing extreme damages of the machine tools and there has been an increasing demand for the development of effective and efficient machine health condition monitoring techniques. Efforts to address early detection and diagnosis of spindle bearing localized faults have resulted into the development of the zero resonator frequency filter (ZRFF) and discrete wavelet packet transform (DWPT) techniques. However, the application of these techniques individually has not managed to accurately identify the intended fault in a machine. This is due to the fact that, at the early stage of bearing failure, the fault impulses are weak and have low Signal to Noise Ratio (SNR) due to being overwhelmed by strong machine vibrations from other machine elements and the surrounding environment. This study aims to develop a novel hybrid algorithm combining the ZRFF and DWPT for early spindle bearing fault detection and diagnosis. The integrated algorithm uses the ZRFF as the first level of de-noising the non-stationary bearing vibration signals and the DWPT for clear extraction of crucial periodic impulse features that were not easily visible from the first de-noising. The obtained frequency spectrum gives a dominant peak line which corresponds the fault frequency of interest. To minimize the algorithm processing time and achieve efficient algorithm, an optimum wavelet decomposition level selection is of upmost important. The optimum decomposition level was determined using the minimum Shannon entropy criteria. The experimental datasets from Case Western Reserve University (CWRU) and simulated signal were used to test the validity of the proposed algorithm. The vibration data of normal bearing, inner race fault, and outer race fault from CWRU were considered for analysis. The simulations were performed using an Intel Core i5-4200U laptop computer with a 2.3 GHz speed running on MATLAB R2017a 64-bit software. The proposed algorithm had superior performance in terms of computational efficiency (45s) and high classification accuracy (99.87%) of the bearings faults when compared with other methods. The integrated algorithm of ZRFF and DWPT demonstrated its effectiveness and efficiency by accurately extracting the Bearings Characteristics Frequencies (BCF’s) of interest from the bearing raw vibration signals masked by strong noise. It is also simple, easy interpretation, requires no computational complexity and can be used for other machine elements faults detection and diagnosis.</p>

Spindle Bearings Fault Diagnosis Technique Based on Integration of Zero Resonator Frequency Filter and Discrete Wavelet Packet Transform

Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2021/volume/85543

ASME 2021 International Mechanical Engineering Congress and Exposition (IMECE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.