Sparse random projection-based hyperdisk classifier for bevel gearbox fault diagnosis. (August 2022)

• Record Type: Journal Article
• Authors: Zhu, Zuanyu; Yang, Yu; Hu, Niaoqing; Cheng, Zhe; Cheng, Junsheng
• Is Part Of: Advanced engineering informatics. Volume 53(2022)
• Publication Date: 2022-08
• Subjects: Sparse random projection-based hyperdisk classifier -- Fast algorithm -- Bevel gearbox -- Fault diagnosis -- Class imbalance problem; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2022.101713
• Languages: English
• ISSNs: 1474-0346

Abstract

Highlights: A novel hyperdisk classifier called SRPHD is proposed in this paper. SRPHD can screen out the core samples that affect the decision boundary. SRPHD develops a strategy to deal with imbalanced data. SRPHD can greatly reduce the training time while guaranteeing a high accuracy. SRPHD has better performance and efficiency in imbalanced fault data. Abstract: The fault diagnosis of bevel gearbox is of great significance. At present, the commonly used methods are based on pattern recognition, such as support vector machine, convex hull classifier and hyperdisk classifier. However, the number of elements in the kernel matrix of these kernel function-based classification methods increases squarely with the data size, resulting in intolerable training time. Based on this, a sparse random projection-based hyperdisk classifier model is proposed. The proposed method has the following novelties: First, based on sparse random projection and the geometrical characteristics of the hyperdisk model, a method is designed to efficiently screen out the core samples, and these samples are given different weights in this process. Second, the proposed method introduces slack variables and the dynamic penalty parameter to obtain a hyperdisk model with more reasonable boundary. Last, a strategy is developed to minimize the adverse effects of imbalanced training data. The effectiveness and applicability of the proposed method are verified on bevel gearbox fault data. The experimental results show that compared with other classifiers, the proposed method can greatly reduce the training time while guaranteeing a high classification accuracy. What's more, it has better performance and efficiency in fault diagnosis with imbalanced training data.