Showing papers by "Yaguo Lei published in 2020"

A Hybrid Prognostics Approach for Estimating Remaining Useful Life of Rolling Element Bearings

Abstract: Remaining useful life (RUL) prediction of rolling element bearings plays a pivotal role in reducing costly unplanned maintenance and increasing the reliability, availability, and safety of machines. This paper proposes a hybrid prognostics approach for RUL prediction of rolling element bearings. First, degradation data of bearings are sparsely represented using relevance vector machine regressions with different kernel parameters. Then, exponential degradation models coupled with the Frechet distance are employed to estimate the RUL adaptively. The proposed approach is evaluated using the vibration data from accelerated degradation tests of rolling element bearings and the public PRONOSTIA bearing datasets. Experimental results demonstrate the effectiveness of the proposed approach in improving the accuracy and convergence of RUL prediction of rolling element bearings.

A Polynomial Kernel Induced Distance Metric to Improve Deep Transfer Learning for Fault Diagnosis of Machines

Abstract: Deep transfer-learning-based diagnosis models are promising to apply diagnosis knowledge across related machines, but from which the collected data follow different distribution. To reduce the distribution discrepancy, Gaussian kernel induced maximum mean discrepancy (GK-MMD) is a widely used distance metric to impose constraints on the training of diagnosis models. However, the models using GK-MMD have three weaknesses: 1) GK-MMD may not accurately estimate distribution discrepancy because it ignores the high-order moment distances of data; 2) the time complexity of GK-MMD is high to require much computation cost; 3) the transfer performance of GK-MMD-based diagnosis models is sensitive to the selected kernel parameters. In order to overcome the weaknesses, a distance metric named polynomial kernel induced MMD (PK-MMD) is proposed in this article. Combined with PK-MMD, a diagnosis model is constructed to reuse diagnosis knowledge from one machine to the other. The proposed methods are verified by two transfer learning cases, in which the health states of locomotive bearings are identified with the help of data respectively from motor bearings and gearbox bearings in laboratories. The results show that PK-MMD enables to improve the inefficient computation of GK-MMD, and the PK-MMD-based diagnosis model presents better transfer results than other methods.

An Incorrect Data Detection Method for Big Data Cleaning of Machinery Condition Monitoring

Abstract: The presence of incorrect data leads to the decrease of condition-monitoring big data quality. As a result, unreliable or misleading results are probably obtained by analyzing these poor-quality data. In this paper, to improve the data quality, an incorrect data detection method based on an improved local outlier factor (LOF) is proposed for data cleaning. First, a sliding window technique is used to divide data into different segments. These segments are considered as different objects and their attributes consist of time-domain statistical features extracted from each segment, such as mean, maximum and peak-to-peak value. Second, a kernel-based LOF (KLOF) is calculated using these attributes to evaluate the degree of each segment being incorrect data. Third, according to these KLOF values and a threshold value, incorrect data are detected. Finally, a simulation of vibration data generated by a defective rolling element bearing and three real cases concerning a fixed-axle gearbox, a wind turbine, and a planetary gearbox are used to verify the effectiveness of the proposed method, respectively. The results demonstrate that the proposed method is able to detect both missing segments and abnormal segments, which are two typical incorrect data, effectively, and thus is helpful for big data cleaning of machinery condition monitoring.