A Novel Predictive Maintenance Method Based on Deep Adversarial Learning in the Intelligent Manufacturing System

Abstract

Along with the number and the functional complexity of machines increase in the intelligent manufacturing system, the probability of faults will increase, which may lead to huge economic losses. Traditional passive or regular maintenance methods of solving the faults have the problems of low efficiency and huge resource consumption. Besides, traditional maintenance methods mostly contain single model, so all the prognostics and maintenance tasks of the intelligent manufacturing system can hardly be addressed at the same time. Therefore, this paper proposes a novel predictive maintenance (PDM) method based on the improved deep adversarial learning (LSTM-GAN). The long-short-term memory (LSTM) network can solve the disadvantage of vanishing gradients and the mode collapse from the generative adversarial network (GAN). The method can not only avoid the mode collapse of GAN but also realize the self-detection of abnormal data. Meanwhile, the predictive maintenance model includes two prediction models and a maintenance decision model. The prediction models can predict the state of the machine and the fault of the machine in advance. Then the maintenance decision model will arrange maintenance personnel and offer a plan of maintenance. Finally, a case study about predictive maintenance using LSTM-GAN in the intelligent manufacturing system is presented. The fault prediction accuracy of LTSM-GAN is as high as 99.68%. With the comparison between LSTM-GAN and other traditional methods, LSTM-GAN shows priority both in accuracy and efficiency. Moreover, the proposed PDM can reduce maintenance costs and downtime so that the life of machines in the intelligent manufacturing system will extend.

Machine learning techniques applied to mechanical fault diagnosis and fault prognosis in the context of real industrial manufacturing use-cases: a systematic literature review

“…As can be seen in Fig. 6, 53.3% of publications employ supervised learning techniques, 28.9% use unsupervised learning techniques, 15.6% make use of both supervised and unsupervised techniques and 2.2% [56] conference Advances in Manufacturing [57] journal Applied Sciences [58] journal Business & Information Systems Engineering [59] journal Complexity [60] journal Computers & Industrial Engineering [61] journal Electronics [62] journal Engineering Applications of Artificial Intelligence [63] journal Expert Systems with Applications [64] journal IEEE Transactions on Industrial Electronics [31] journal IEEE Transactions on Industrial Informatics [65] journal Journal of Manufacturing Systems [66] journal Simulation Modelling Practice and Theory [67] journal Studies in Informatics and Control [68] journal 2019 31st International Conference on Advanced Information Systems Engineering (CAiSE) [69,70] conference CIRP Annals [71,72] journal Sensors [73,74] journal The International Journal of Advanced Manufacturing Technology [75,76] journal IEEE Access [77][78][79][80][81] journal Fig. 4 Proportion of publications in conferences and journals combine semi-supervised, unsupervised, and supervised techniques.…”

“…Instance-based algorithms K-NN [39] supervised Latent Variable Models PCA [65] unsupervised GMM [47] unsupervised K-Means [54] unsupervised PLSR [64] supervised K-SVD [60] unsupervised K-MDTSC [62] unsupervised Artificial Neural Networks ANN [57] supervised BPNN [40] supervised CNN [78] supervised DNN [77] supervised LSTM [70] supervised MLP [56] supervised SSAE + BPNN [31] unsupervised + supervised SSAE + Softmax Classifier [81] unsupervised + supervised LSTM Autoencoder [73] supervised LSTM -GAN [79] supervised RNN [55] supervised Conditional Variational Autoencoder [66] unsupervised Rule-based models R4RE ("Rules 4 Rare Events" based on QARMA) [49] supervised XCSR [51] supervised consists in the principal components obtained from the application of DPCA, which do not represent any physical properties or measurements of the system. The study presented in [67] used an ensemble method as well due to its efficiency in terms of computation time and memory when handling large amounts of data.…”

“…Proposed approach: using multiple ANNs can be a difficult task; might not be able to identify neighbour states; requires labelled data. [79] Proposed approach: capable of generating large volumes of fault data; avoids mode collapse; improved accuracy and efficiency.…”

“…Moreover, to handle scenarios where labels might be delayed or inexistent, SERMON includes a mapping unit that employs an autoencoder to suggests possible data labels. An LSTM-GAN was used in [79] as part of a PdM methodology that can not only monitor the health state of machines and predict faults before they occur, but also provides the factory's maintenance staff with maintenance plans that are appropriate to deal with the issues detected by the state prediction and fault prediction models. The GAN was used to generate a large volume of synthetic fault data to improve the accuracy of the model.…”

Deep Neural Network Structure Design for Equipment Failure Prediction in Smart Factory

“…To demonstrate the feasibility of the approach, an experiment based on simulation methodology has been conducted, and the results show that their achieves. Changchun [15] propose a facilities maintenance scheme based on a mixed deep neural network that has both long-short term memory (LSTM) and generative adversarial network (GAN). In the scheme, GAN was used to provide a reliable fault dataset to dramatically enhance prediction performance.…”