Survey on Deep Learning applied to predictive maintenance (original) (raw)
Related papers
Engineering Applications of Artificial Intelligence, 2020
Deep learning applications have been thriving over the last decade in many different domains, including computer vision and natural language understanding. The drivers for the vibrant development of deep learning have been the availability of abundant data, breakthroughs of algorithms and the advancements in hardware. Despite the fact that complex industrial assets have been extensively monitored and large amounts of condition monitoring signals have been collected, the application of deep learning approaches for detecting, diagnosing and predicting faults of complex industrial assets has been limited. The current paper provides a thorough evaluation of the current developments, drivers, challenges, potential solutions and future research needs in the field of deep learning applied to Prognostics and Health Management (PHM) applications.
Prognostics and Health Management of Industrial Assets: Current Progress and Road Ahead
Frontiers in Artificial Intelligence, 2020
Prognostic and Health Management (PHM) systems are some of the main protagonists of the Industry 4.0 revolution. Efficiently detecting whether an industrial component has deviated from its normal operating condition or predicting when a fault will occur are the main challenges these systems aim at addressing. Efficient PHM methods promise to decrease the probability of extreme failure events, thus improving the safety level of industrial machines. Furthermore, they could potentially drastically reduce the often conspicuous costs associated with scheduled maintenance operations. The increasing availability of data and the stunning progress of Machine Learning (ML) and Deep Learning (DL) techniques over the last decade represent two strong motivating factors for the development of datadriven PHM systems. On the other hand, the black-box nature of DL models significantly hinders their level of interpretability, de facto limiting their application to real-world scenarios. In this work, we explore the intersection of Artificial Intelligence (AI) methods and PHM applications. We present a thorough review of existing works both in the contexts of fault diagnosis and fault prognosis, highlighting the benefits and the drawbacks introduced by the adoption of AI techniques. Our goal is to highlight potentially fruitful research directions along with characterizing the main challenges that need to be addressed in order to realize the promises of AI-based PHM systems.
A review on the application of deep learning in system health management
Mechanical Systems and Signal Processing, 2018
Given the advancements in modern technological capabilities, having an integrated health management and diagnostic strategy becomes an important part of a system's operational life-cycle. This is because it can be used to detect anomalies, analyse failures and predict the future state based on up-to-date information. By utilising condition data and on-site feedback, data models can be trained using machine learning and statistical concepts. Once trained, the logic for data processing can be embedded on on-board controllers whilst enabling real-time health assessment and analysis. However, this integration inevitably faces several difficulties and challenges for the community; indicating the need for novel approaches to address this vexing issue. Deep learning has gained increasing attention due to its potential advantages with data classification and feature extraction problems. It is an evolving research area with diverse application domains and hence its use for system health management applications must been researched if it can be used to increase overall system resilience or potential cost benefits for maintenance, repair, and overhaul activities. This article presents a systematic review of artificial intelligence based system health management with an emphasis on recent trends of deep learning within the field. Various architectures and related theories are discussed to clarify its potential. Based on the reviewed work, deep learning demonstrates plausible benefits for fault diagnosis and prognostics. However, there are a number of limitations that hinder its widespread adoption and require further development. Attention is paid to overcoming these challenges, with future opportunities being enumerated.
Predictive maintenance using cox proportional hazard deep learning
Advanced Engineering Informatics, 2020
Predictive maintenance (PdM) has become prevalent in the industry in order to reduce maintenance cost and to achieve sustainable operational management. The core of PdM is to predict the next failure so corresponding maintenance can be scheduled before it happens. The purpose of this study is to establish a Time-Between-Failure (TBF) prediction model through a data-driven approach. For PdM, data sparsity is regarded as a critical issue which can jeopardize algorithm performance for the modelling based on maintenance data. Meanwhile, data censoring has imposed another challenge for handling maintenance data because the censored data is only partially labelled. Furthermore, data sparsity may affect algorithm performance of existing approaches when addressing the data censoring issue. In this study, a new approach called Cox proportional hazard deep learning (CoxPHDL) is proposed to tackle the aforementioned issues of data sparsity and data censoring that are common in the analysis of operational maintenance data. The idea is to offer an integrated solution by taking advantage of deep learning and reliability analysis. To start with, an autoencoder is adopted to convert the nominal data into a robust representation. Secondly, a Cox PHM is researched to estimate the TBF of the censored data. A long-short-term memory (LSTM) network is then established to train the TBF prediction model based on the pre-processed maintenance data. Experimental studies using a sizable real-world fleet maintenance data set provided by a UK fleet company have demonstrated the merits of the proposed approach where the algorithm performance based on the proposed LSTM network has been improved respectively in terms of MCC and RMSE.
Remaining Useful Life Prediction With Partial Sensor Malfunctions Using Deep Adversarial Networks
IEEE/CAA Journal of Automatica Sinica, 2023
In recent years, intelligent data-driven prognostic methods have been successfully developed, and good machinery health assessment performance has been achieved through explorations of data from multiple sensors. However, existing data-fusion prognostic approaches generally rely on the data availability of all sensors, and are vulnerable to potential sensor malfunctions, which are likely to occur in real industries especially for machines in harsh operating environments. In this paper, a deep learning-based remaining useful life (RUL) prediction method is proposed to address the sensor malfunction problem. A global feature extraction scheme is adopted to fully exploit information of different sensors. Adversarial learning is further introduced to extract generalized sensor-invariant features. Through explorations of both global and shared features, promising and robust RUL prediction performance can be achieved by the proposed method in the testing scenarios with sensor malfunctions. The experimental results suggest the proposed approach is well suited for real industrial applications.
Physics Based Deep Learning Technique for Prognostics
Annual Conference of the PHM Society
Remaining useful life (RUL) estimation is one of the most important aspects of prognostics and health management (PHM). Various deep learning (DL) based techniques have been developed and applied for the purposes of RUL estimation. One limitation of DL is the lack of physical interpretations as they are purely data driven models. Another limitation is the need for an exceedingly large amount of data to arrive at an acceptable pattern recognition performance for the purposes of RUL estimation. This research is aimed to overcome these limitations by developing physics based DL techniques for RUL prediction and validate the method with real run-to-failure datasets. The contribution of the research relies on creating hybrid DL based techniques as well as combining physics based approaches with DL techniques for effective RUL prediction.
Automobile Predictive Maintenance using Deep Learning
International Journal of Artificial Intelligence and Machine Learning
There are three types of maintenance management policy Run-tofailure (R2F), Preventive Maintenance (PvM) and Predictive Maintenance (PdM). In both R2F and PdM we have the data related to the maintenance cycle. In case of Preventive Maintenance (PvM) complete information about maintenance cycle is not available. Among these three maintenance policies, predictive Maintenance (PdM) is becoming a very important strategy as it can help us to minimize the repair time and the associated cost with it. In this paper we have proposed PdM, which allows the dynamic decision rules for the maintenance management. PdM is achieved by training the machine learning model with the datasets. It also helps in planning of maintenance schedules. We specially focused on two models that are Binary Classification and Recurrent Neural Network. In Binary Classification we classify whether our data belongs to the failure class or the non failure class. In Binary Classification the number of cycles is entered an...
Predictive maintenance of rotational machinery using deep learning
International Journal of Electrical and Computer Engineering (IJECE), 2024
This paper describes an implementation of a deep learning-based predictive maintenance (PdM) system for industrial rotational machinery, built upon the foundation of a long short-term memory (LSTM) autoencoder and regression analysis. The autoencoder identifies anomalous patterns, while the latter, based on the autoencoder's output, estimates the machine's remaining useful life (RUL). Unlike prior PdM systems dependent on labelled historical data, the developed system doesn't require it as it's based on an unsupervised deep learning model, enhancing its adaptability. The paper also explores a robust condition monitoring system that collects machine operational data, including vibration and current parameters, and transmits them to a database via a Bluetooth low energy (BLE) network. Additionally, the study demonstrates the integration of this PdM system within a web-based framework, promoting its adoption across various industrial settings. Tests confirm the system's ability to accurately identify faults, highlighting its potential to reduce unexpected downtime and enhance machinery reliability.
IEEE Access
In the era of industry 4.0, safety, efficiency and reliability of industrial machinery is an elementary concern in trade sectors. The accurate remaining useful life (RUL) prediction of an equipment in due time allows us to effectively plan the maintenance operation and mitigate the downtime to raise the revenue of business. In the past decade, data driven based RUL prognostic methods had gained a lot of interest among the researchers. There exist various deep learning-based techniques which have been used for accurate RUL estimation. One of the widely used technique in this regard is the long short-term memory (LSTM) networks. To further improve the prediction accuracy of LSTM networks, this paper proposes a model in which effective pre-processing steps are combined with LSTM network. C-MAPSS turbofan engine degradation dataset released by NASA is used to validate the performance of the proposed model. One important factor in RUL predictions is to determine the starting point of the engine degradation. This work proposes an improved piecewise linear degradation model to determine the starting point of deterioration and assign the RUL target labels. The sensors data is pre-processed using the correlation analysis to choose only those sensors measurement which have a monotonous behavior with RUL, which is then filtered through a moving median filter. The updated RUL labels from the degradation model together with the pre-processed data are used to train a deep LSTM network. The deep neural network when combined with dimensionality reduction and piece-wise linear RUL function algorithms achieves improved performance on aircraft turbofan engine sensor dataset. We have tested our proposed model on all four sub-datasets in C-MAPSS and the results are then compared with the existing methods which utilizes the same dataset in their experimental work. It is concluded that our model yields improvement in RUL prediction and attains minimum root mean squared error and score function values. INDEX TERMS Deep learning, long short-term memory networks, remaining useful life, turbofan engine. This technique focuses on forecasting the error by model-39 ing the degradation trends between input sensors and time-40 to-failure duration of the machine. So, the benefits of this 41 maintenance strategy is that we can eliminate unplanned 42 downtime, reduced maintenance costs and maximize the 43 machine lifetime for safety critical circumstances. One such 44 example is aircraft engine which requires continuous mon-45 itoring of the engine performance. The fault diagnostics 46 and prognostics of aircraft engine has gained great attention 47 over the last few decades [4], [5], [6], [7]. One important 48 component in aircraft engine maintenance is to accurately 49 determine its remaining useful life (RUL) for reducing the 50 maintenance costs while attaining the reliability [8], [9]. RUL 51 prediction model is developed based upon the degradation 52 trends among the various condition monitoring sensors. This 53 model helps in development of maintenance strategy in a 54 targeted manner to eliminate unplanned downtime and maxi-55 mize machine lifetime for safety critical circumstances. Early 56 anomaly detection and timely warning of a failure is vital for 57 maximum utilization of the system. There are basically three 58 types of prognostics techniques used for estimating RUL, 59 physical model-based approaches [10], [11], data-driven 60 approaches [12], [13] and hybrid approaches [14]. 61 Model based approach initially required a comprehensive 62 understanding of the physical architecture of the machine and 63 then applying the laws of physics to obtain the mathematical 64 model of the machine for RUL estimation [15]. Mathemat-65 ical models often take some simplifying assumptions with 66 uncertainty management for a complex industrial machinery, 67 which can impose serious limitations on these techniques and 68 hence degrade the RUL prediction accuracy [16]. 69 Data-driven based prognosis approaches use various sta-70 tistical and machine leaning (ML) algorithms to discover the 71 trends or patterns in the underlying sensor data to estimate 72 RUL of the system. These techniques are suitable for com-73 plex industrial machinery and further, it does not require a 74 thorough understanding of a complete engine or the process. 75 Hybrid method combines both the physics and data-driven 76 based model techniques [17]. 77 In the past decade, data-driven based prognostics meth-78 ods have been exploited by many researchers. These mod-79 els estimate the RUL by analyzing the degradation trend 80 and target trajectory of sensor data. Deep learning methods 81 like autoencoder, convolutional neural networks (CNN), long 82 short-term memory (LSTM) networks and their varianta and 83 combinations have achieved a massive success in the fields 84 of computer vision, speech recognition, video segmentation 85 and predictive maintenance [18]. The major drawback of deep 86 learning algorithm is that it requires a large volume of data 87 for offline training and in the field of prognostics, it is very 88 157 techniques usually employ algorithms like Kalman filter 158 (KF), extended Kalman filter (EKF) and particles filters to 159 come up with mathematical formulation of machine based on 160 multi sensor time series sequence data [26], [27], [28]. Clas-161 sical degradation method such as Eyring model or Weibull 162 distribution was implemented in [29]. Salahshoor et al. [30] 163 used a unified framework of EKF based design for sen-164 sor data fusion algorithm to further enhanced the detec-165 tion and diagnosis of degradation trends and system faults. 166 Ordonez et al. [31] implemented the auto-regressive inte-167 grated moving average (ARIMA) model and support vector 168 machine (SVR) methods collectively to estimate the RUL. 169 The desired features can be created by analyzing prior learn-170 ing about the degradation models as presented in [32]. In [33], 171 it is suggested that failure thresholds or degradation state esti-172 mation is no longer required in learning-oriented approach. 173 Khelif et al. [33] presented machine learning based support 174 vector regression (SVR) model to project the direct associa-175 tion between multivariate sensor data or health index and the 176 aircraft turbofan engine RUL. 177 Across all these techniques for turbofan engine RUL pre-178 diction, deep neural network-based methods have gained vast 179 popularity. Zhang et al. [34] introduced a multi-objective 180 evolutionary algorithm to expand and organized the deep 181 belief network into multiple parallel networks simultane-182 ously to accomplish the two convicting objectives i.e. diver-183 sity and accuracy. These networks attained a fine RUL 184 prediction accuracy especially in case of complicated oper-185 ations and in the presence of noise in input data [35], [36]. 186 Saeidi et al. [37], proposed a naive Bayesian classification 187 algorithm to measure the health index for turbofan engine.