{"title":"神经变压器:受大脑启发的噪声下轻量级机械故障诊断方法","authors":"","doi":"10.1016/j.ress.2024.110409","DOIUrl":null,"url":null,"abstract":"<div><p>Recently, as a representative of deep learning methods, Transformers have shown great prowess in intelligent fault diagnosis, offering powerful feature extraction and modeling. However, their high computational demand and low robustness limit industrial application. Therefore, this paper proposes an innovative Neural-Transformer to realize high-precision robust fault diagnosis with low computational cost. First, a two-dimensional representation method, the frequency-slice wavelet transform (FSWT), is introduced to reflect the dynamic characteristics and frequency component variations of signals, enhancing the fault identifiability of vibration signals. Second, a separable multiscale spiking tokenizer (SMST) is developed to project time-frequency input of multiple scales to spike features with a fixed patch, ensuring consistency in feature extraction and improving the recognizability of specific frequencies in mechanical faults. Subsequently, a multi-head spatiotemporal spiking self-attention (MHSSSA) mechanism is constructed, which abandons the cumbersome multiplication operations with high computational costs and can also focus on key fine-grained time-frequency features in a global range. Experimental cases validate the advantages of the Neural-Transformer in comparison to baseline methods and state-of-art methods on one public dataset and two real-world datasets. In particular, the proposed method only consumes 0.65mJ of energy to achieve an optimal diagnostic accuracy of 93.14% on real-world dataset.</p></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":null,"pages":null},"PeriodicalIF":9.4000,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Neural-transformer: A brain-inspired lightweight mechanical fault diagnosis method under noise\",\"authors\":\"\",\"doi\":\"10.1016/j.ress.2024.110409\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Recently, as a representative of deep learning methods, Transformers have shown great prowess in intelligent fault diagnosis, offering powerful feature extraction and modeling. However, their high computational demand and low robustness limit industrial application. Therefore, this paper proposes an innovative Neural-Transformer to realize high-precision robust fault diagnosis with low computational cost. First, a two-dimensional representation method, the frequency-slice wavelet transform (FSWT), is introduced to reflect the dynamic characteristics and frequency component variations of signals, enhancing the fault identifiability of vibration signals. Second, a separable multiscale spiking tokenizer (SMST) is developed to project time-frequency input of multiple scales to spike features with a fixed patch, ensuring consistency in feature extraction and improving the recognizability of specific frequencies in mechanical faults. Subsequently, a multi-head spatiotemporal spiking self-attention (MHSSSA) mechanism is constructed, which abandons the cumbersome multiplication operations with high computational costs and can also focus on key fine-grained time-frequency features in a global range. Experimental cases validate the advantages of the Neural-Transformer in comparison to baseline methods and state-of-art methods on one public dataset and two real-world datasets. In particular, the proposed method only consumes 0.65mJ of energy to achieve an optimal diagnostic accuracy of 93.14% on real-world dataset.</p></div>\",\"PeriodicalId\":54500,\"journal\":{\"name\":\"Reliability Engineering & System Safety\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":9.4000,\"publicationDate\":\"2024-08-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Reliability Engineering & System Safety\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0951832024004812\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, INDUSTRIAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Reliability Engineering & System Safety","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0951832024004812","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, INDUSTRIAL","Score":null,"Total":0}
Neural-transformer: A brain-inspired lightweight mechanical fault diagnosis method under noise
Recently, as a representative of deep learning methods, Transformers have shown great prowess in intelligent fault diagnosis, offering powerful feature extraction and modeling. However, their high computational demand and low robustness limit industrial application. Therefore, this paper proposes an innovative Neural-Transformer to realize high-precision robust fault diagnosis with low computational cost. First, a two-dimensional representation method, the frequency-slice wavelet transform (FSWT), is introduced to reflect the dynamic characteristics and frequency component variations of signals, enhancing the fault identifiability of vibration signals. Second, a separable multiscale spiking tokenizer (SMST) is developed to project time-frequency input of multiple scales to spike features with a fixed patch, ensuring consistency in feature extraction and improving the recognizability of specific frequencies in mechanical faults. Subsequently, a multi-head spatiotemporal spiking self-attention (MHSSSA) mechanism is constructed, which abandons the cumbersome multiplication operations with high computational costs and can also focus on key fine-grained time-frequency features in a global range. Experimental cases validate the advantages of the Neural-Transformer in comparison to baseline methods and state-of-art methods on one public dataset and two real-world datasets. In particular, the proposed method only consumes 0.65mJ of energy to achieve an optimal diagnostic accuracy of 93.14% on real-world dataset.
期刊介绍:
Elsevier publishes Reliability Engineering & System Safety in association with the European Safety and Reliability Association and the Safety Engineering and Risk Analysis Division. The international journal is devoted to developing and applying methods to enhance the safety and reliability of complex technological systems, like nuclear power plants, chemical plants, hazardous waste facilities, space systems, offshore and maritime systems, transportation systems, constructed infrastructure, and manufacturing plants. The journal normally publishes only articles that involve the analysis of substantive problems related to the reliability of complex systems or present techniques and/or theoretical results that have a discernable relationship to the solution of such problems. An important aim is to balance academic material and practical applications.