Deep learning-based cross-sensor domain adaptation for fault diagnosis of electro-mechanical actuators
- PDF / 1,230,815 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 27 Downloads / 186 Views
Deep learning-based cross-sensor domain adaptation for fault diagnosis of electro-mechanical actuators Shahin Siahpour1 · Xiang Li1
· Jay Lee1
Received: 5 June 2020 / Revised: 10 July 2020 / Accepted: 13 July 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Recently, the development of intelligent data-driven machinery fault diagnosis methods have received significant attention. In most studies, the training and testing data are assumed to be collected from the same sensor. However, in real practice, due to the mounting limitation and sensor malfunctioning, it cannot be generally guaranteed to obtain the data from the same sensor location at all times. The testing and training data can be possibly from different sensor locations. Consequently, different data distributions exist, which remarkably deteriorates the data-driven model performance in different scenarios. In order to address this issue, this paper proposes a deep learning-based cross-sensor domain adaptation approach for machinery fault diagnosis. The maximum mean discrepancy is deployed as a distance metric to realize marginal domain fusion. The unlabeled parallel data is further exploited to achieve conditional domain alignment with respect to different machine health conditions. An electro-mechanical actuator dataset is used as a case study for the validation of the proposed method. Different tasks are designed to simulate different cross-sensor domain adaptation problems in fault diagnosis. The experimental results suggest the proposed method achieves higher than 95% testing accuracies in most tasks, and it offers a promising approach for cross-sensor fault diagnosis problems. Keywords Deep learning · Domain adaptation · Electro-mechanical actuator · Transfer learning · Fault diagnosis
1 Introduction Recently, due to the higher reliability, lower overall weight, better maintainability, etc., fly-by-wire control actuators are gaining more attention in comparison to the conventional hydraulic control actuators in aerospace applications [1]. The Electro-Mechanical Actuator (EMA) plays a crucial role as an important fly-by-wire actuation in the aerospace industries like robotic spacecrafts and civilian airliners [2–4]. The low utilization time of EMAs in aerospace applications prohibits the accumulation of reliable fault statistics. On the other hand, the EMAs are mostly deployed in high-risk conditions [5]. These two factors highlight the importance of
B
Xiang Li [email protected] Shahin Siahpour [email protected] Jay Lee [email protected]
1
Department of Mechanical Engineering, University of Cincinnati, Cincinnati 45221, USA
developing a fault detection method to prevent the occurrence of catastrophic consequences [1]. In the current literature, numbers of studies methods have been proposed to develop a prognostic health management (PHM) system for the EMAs. These methods mostly can be categorized into two different approaches: model-based methods [6,7], and data-driven methods [8]. Due to the complexity of th
Data Loading...
