Abstract
In modern industrial engineered systems, variant working conditions disturb the distributions of machines’ operational data, which results in different feature distributions (DFD) problems for fault prognostics. Domain adaptation (DA) have been proved good at dealing DFD problems, and several deep DA-based methods have been also proposed in fault prognostics filed. However, existing methods refer to the DA tasks from one working condition to another, without considerations of transferring between datasets under complex working conditions. The prior distribution of working conditions will influence the distributions of machines’ operational data, and few studies take prior distribution of working conditions into consideration of DA for fault prognostics. Thus, in this paper, a working-condition-based deep domain adaptation network (Deep wcDAN) is proposed to overcome the DFD problems caused by variant complex working conditions. In the proposed method, CNNs combines LSTMs with domain adaptive transfer technique to minimize the distribution discrepancy between training and testing datasets. Furthermore, a working-condition-based MMD (wcMMD) is proposed to optimize the DA process based on the prior distribution of each working condition. The performance of proposed model is evaluated and the negative transfer effects have been analyzed based on C-MAPSS datasets. The results show that the proposed method performs better than baseline methods on predicting remaining useful life (RUL) with DFD problems.
Highlights
Prognostics and Health Management (PHM) has a significant importance for modern industrial systems
WORK In this paper, we focus on establishing an advanced deep Domain adaptation (DA) model by fully utilizing the prior distributions of working conditions, to optimize the different feature distributions (DFD) problems on fault prognostics in industrial systems
Based on an assumption that similar working condition generates similar feature distribution of samples and vice versa, we propose a novel Deep wcDAN to minimizing the distribution discrepancy of samples among different working conditions
Summary
Prognostics and Health Management (PHM) has a significant importance for modern industrial systems. In real industries, environmental noise and variation of operating condition inevitably make the feature distributions of training and testing data different from each other, it is difficult for the well-trained neural network to generalize the learned pattern knowledge from the labeled training data, denoted as source domain, to the new unlabeled testing data, denoted as target domain. In this paper, this challenge is referred as different feature distributions (DFD) problem in industrial system, which is caused by variant working conditions. Such challenge have been pointed out in [4], [5] for fault prognostics
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have