Time-series change detection has been studied in several fields. From sensor data, engineering systems, medical diagnosis, and financial markets to user actions on a network, huge amounts of temporal data are generated. There is a need for a clear separation between normal and abnormal behaviour of the system in order to investigate causes or forecast change. Characteristics include irregularities, deviations, anomalies, outliers, novelties or surprising patterns. The efficient detection of such patterns is challenging, especially when constraints need to be taken into account, such as the data velocity, volume, limited time for reacting to events, and the details of the temporal sequence. This paper reviews the main techniques for time series change point detection, focusing on online methods. Performance criteria including complexity, time granularity, and robustness is used to compare techniques, followed by a discussion about current challenges and open issues.
Fault detection based on deep learning has been intensively investigated in the recent decade due to increasing availability of data and its ability to engineer features with deep neural network architectures. Despite much attention to its application, the major challenge is the lack of available labelled datasets to build the models since maintenance is usually conducted regularly to avoid significant defects. This paper aims to propose a successful real-time fault detection framework based on unsupervised deep learning using only healthy normal data. The approach is based on autoencoder architecture and a one-class support vector machine as a classifier. As a case study, large real-world datasets acquired from railway door systems have been employed. The five different types of deep learning models and a one-class classifier are trained and comprehensively validated based on performance metrics and sensitivity analysis. In addition, two experiments have been carried out to verify the model's adaptability and robustness to variational timeseries data. The result shows a typical autoencoder is the least sensitive to a decision boundary set by the one-class classifier. However, the two experiments show that the fault detection accuracy for a bidirectional long short-term memory-based autoencoder is considerably higher than other autoencoder-based models at 0.970 and 0.966 as F1 score, meaning only this model is adaptable and robust to variational data. The experimental result allows us to obtain the understandability of the deep learning models. Furthermore, the regions of anomalies are localised with unsupervised models, which enables diagnosing the cause of failure. 17 18 Fault detection serves an important role in PHM and has 42 been investigated in recent decades. Researchers in such 43 diverse disciplines as medicine, engineering and sciences 44 have been developing methodologies to detect fault or 45 anomaly conditions, pinpoint or isolate which component or 46 object in a system or process is faulty, and decide on the 47 potential impact of a failing or failed component on the health 48 of the system [4]. 49 In this area of study, the methodologies usually centre 50 on model-based or data-driven approaches. Model-based 51 approaches incorporate a physical understanding of the sys-52 tems through mathematical representations and include sys-53 tem modelling. The output of the model is then compared 54 with the actual output measurement throughout the residual 55 analysis [5], [6]. However, the mechanical system contains 56 many components interconnected with various uncertainties, 57 which makes the modelling approach of limited value. On the 58 other hand, data-driven approaches use statistical pattern 59 recognition and machine learning to detect changes [5]. Data-60 driven approaches do not require mathematical modelling 61 of the systems and have gained much attention with the 62 increasing availability of data. 63 The data-driven approaches include traditional machine 64 learning (ML) and deep learning ...
This paper focuses on real-time techniques for fault detection in railway assets through large real-world datasets. It aims to investigate data mining methods to detect faulty behaviour in time series data. A fault detection on railway door systems is carried out using motor current and encoder signal. The door data highlighted start-stop characteristics, with discontinuities in the data. This paper presents a successful fault detection technique, which is a feature-based machine learning method that requires several steps for time-series data processing, such as signal segmentation and the extraction of features. Principal Component Analysis (PCA) is applied to reduce the dimensionality of the extracted feature set and generate condition indicators. Then, the k-means algorithm is employed to separate normal and abnormal behaviour. This is followed by an evaluation of the proposed method and discussion about current challenges and prognosis possibility.TABLE OF CONTENTS 1. INTRODUCTION .
Fault detection for railway door systems based on data-driven approaches has been investigated in recent years due to the massive amount of available monitoring data. Despite much attention to its application, the major challenge is the lack of available faulty datasets to build a reliable model since railway maintenance is usually conducted regularly to avoid significant defects from economic and safety points of view. We aimed to tackle the issue by employing transfer learning. Firstly, we built a long-short term memory-based deep learning model using linear actuator experimental datasets. Then, we employed a transfer learning technique to adjust the deep learning model to be available to real-world railway door systems using a small amount of faulty data. As a result, high fault detection accuracy can be obtained at 0.979 as F1 score. The result reveals that an accurate fault detection model can be built even though a large amount of labelled datasets is unavailable. In addition, the proposed method is applicable to other door systems or electro-mechanical actuators since the method is unspecific to physical mechanisms and fault modes, and the only motor current signal is used in this research. The signal is primarily available from the controller or motor drive without additional sensors.
Prognosis is a challenging technology that aims to accurately predict and estimate the remaining useful life of a component or system in order to enhance its reliability and performance. Although prognosis research for predictive maintenance is a well-researched topic, practical examples of successful prognostic applications remain scarce. This is due to the lack of available run-to-failure data to build the prediction model as maintenance is usually conducted regularly to avoid significant defects. This paper proposes a novel prognosis method that can be applied to real-world railway maintenance planning without employing run-to-failure data. The key idea is that the fault severity assessment and approximate remaining time prediction are often all that is needed in order to plan maintenance. Firstly, using motor current signals, a degradation indicator on railway door systems is generated based on the dynamic time warping method to measure similarity between typical normal and faulty behaviour. Then, the K-means algorithm is applied to assess fault severity, followed by the representative time estimation for each level of fault severity. This estimation thus allows the remaining time prediction until reaching the critical fault severity level without using run-to-failure data. As a result, the proposed method enables predictive maintenance planning for railway door systems. In addition, the fault severity threshold can be updated by additional operational data, enabling the remaining time prediction to be more reliable. Furthermore, the proposed method can be applied to conventional railway assets and other electro-mechanical actuators as motor current signals are primarily available from the controller or motor drive without additional sensors.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.