Condition Monitoring Approach Research Articles

Abnormal Pavement Condition Detection with Vehicle Posture Data Considering Speed Variations.

Pavement condition monitoring is an important task in road asset management and efficient abnormal pavement condition detection is critical for timely conservation management decisions. The present work introduces a mobile pavement condition monitoring approach utilizing low-cost sensor technology and machine-learning-based methodologies. Specifically, an on-board unit (OBU) embedded with an inertial measurement unit (IMU) and global positioning system (GPS) is applied to collect vehicle posture data in real time. Through a comprehensive analysis of both time domain and frequency domain data features for both normal and abnormal pavement conditions, feature engineering is conducted to identify how the most important features affect abnormal pavement condition recognition. Six machine learning models are then developed to identify different types of pavement conditions. The performance of different algorithms and the significance of different features are then analyzed. Moreover, the influence of vehicle speed on pavement condition assessment is further examined and classification models for different speed intervals are developed. The results indicate that the random forest (RF) model that considers vehicle speed achieves the best performance in pavement condition monitoring. The outcomes of the present work would contribute to cost-effective pavement condition monitoring and provide an important reference for pavement maintenance sectors.

Electrical impedance based condition monitoring of machine elements–a systematic review

Condition monitoring of machine elements becomes more important over the last years. Different approaches to detect failures in mechanical components have been developed. All these methods are located at a distance from the point of interest, the observed machine element. This leads to uncertainties in the data, which influences the data quality negatively. Using the electrical impedance for condition monitoring enables in situ measurement with reduced uncertainties and higher data quality. In the last years, research considering this topic was done, but a systematic overview is missing. In this article, a systematic literature research according to the PRISMA approach is fulfilled. The main questions are, what application fields for electrical impedance-based condition monitoring approaches exists and which research gaps are not addressed yet. At the end, 21 articles are categorized in their application fields. Analyzing their content, research questions are identified which have to be addressed in further investigations.

Development of a condition monitoring of heat exchangers for corrosion progress and coolant condition

The transport sector, and thus also rail transport, needs to be completely decarbonised by 2050 at the latest. Considering the long service life of rail vehicles, this goal means enormous innovation and growth potential for the vehicle and supplier industry of railway technology in the coming years. This contribution presents the first results of an R&D project to reduce the total cost of ownership of passenger and freight rail vehicles. The reduced life cycle costs can be achieved over the operating life, among other things, with condition monitoring approaches for critical components of cooling systems to enable diagnosis-based maintenance. Based on guided elastic waves, an integrated system is being developed to analyse the depth and position of corrosion damage to heat exchangers in rail vehicles. This method is to be supplemented by a refractive index sensor integrated in the coolant circuit which can be used to derive influences of the coolant condition on corrosion. First, an acoustic measuring technique was put into operation. Suitable piezoelectric transducers were selected and test measurements, including the introduction of artificial damages, along different paths on a representative heat exchanger were carried out and evaluated. Measurement signals can be received over different distances, and a monitoring of larger areas of the heat exchanger seems possible. For the development of the monitoring of the coolant, plasmonically active sensor substrates with a gold nanostructure were produced and installed in a measuring flow cell. In laboratory tests, concentrations of water-coolant mixtures could be distinguished with the refractive index sensor. In the future, the refractive index sensing will be carried out by means of bypass flow measurement. Since the coolant is pumped in a circuit, the sensor system can be attached to one of the inlet/outlet lines.

An end-to-end deep learning approach for tool wear condition monitoring

An end-to-end deep learning approach for tool wear condition monitoring

CNN-Based Machine Tool Monitoring with STFT Image Analysis

Abstract: In the realm of machine tool operation, effective condition monitoring holds predominant importance to ensure operational reliability and safety. Leveraging deep learning methodologies, particularly convolutional neural networks (CNN), for defect identification has gained significant attention. However, inherent challenges persist, including the extraction of salient features and potential information loss while extracting the features from raw vibration signals. In response, this study proposes an intelligent approach for condition monitoring in machine tools, integrating short-time Fourier transform and convolutional neural networks (STFT-CNN). The process entails using STFT to convert one-dimensional vibration signals into time-frequency pictures, which are then fed into the STFT-CNN model to acquire and identify fault features. Furthermore, the study explores optimizing STFT parameters, such as window type, window width, and translation overlap width, across various typical window functions to improve the effectiveness of the transformation process. Within the STFT-CNN model, the utilization of stacked double convolutional layers aims to augment the model's nonlinear expression capacity, thereby facilitating robust fault diagnosis capabilities in machine tool condition monitoring applications

Meta-learning-based approach for tool condition monitoring in multi-condition small sample scenarios

Tool Condition Monitoring (TCM) technology in machining is crucial for maintaining safety and optimizing costs. However, its practical application faces two significant challenges: difficulties in data collection and a decline in generalization performance across different monitoring tasks. To this end, a hybrid feature boundary-enhanced meta-learning network with adaptive gradients (HFBEAML) is proposed. This method combines cutting process parameters with time-series signal features, employing a multimodal one-dimensional Convolutional Neural Network (1D-CNN) based on the DenseNet architecture and a multi-head self-attention mechanism (MHSA) to mine multi-dimensional features sensitive to tool wear. To further enhance the model’s feature discrimination capability, a multi-loss joint optimization strategy is introduced, combining task-level loss with an improved triplet loss. This approach incorporates strategies for adaptive boundaries and handling the hardest positive and negative sample sets, thereby enhancing the robustness of feature metrics. Additionally, this research innovatively proposes an adaptive meta-level gradient update mechanism, dynamically adjusting gradient weights according to the characteristics of multiple tasks, aiming to improve the model’s generalization ability in multi-task learning environments. The effectiveness of the proposed method is demonstrated through experiments in two different scenarios, comparing its results with five other models showcasing its significant advantages in multi-task, small-sample data monitoring environments.

Offshore field experimentation for novel hybrid condition monitoring approaches

This study details the development of a fully automated pipeline for the condition monitoring of wind turbine drive trains. Vibration data is collected using hardware designed and manufactured in-house and used directly to monitor the condition of the drive trains. The complex nature of wind turbine vibration signals, due to the large number of components and highly variable operating conditions, makes drive train condition monitoring a challenging task. This paper details the full data measurement and analysis flow from sensor to insights and proposes a hybrid automated pipeline with signal processing and data-driven techniques to address the complexity of dealing with wind turbine vibration data. The vibration signals are directly employed to estimate the wind turbine’s instantaneous angular speed to compensate for any rotation speed fluctuations. Pre-processing is performed on the speed-independent signals to evaluate condition indicators in both the time and spectral domain for the vibration signals and their envelopes. Machine learning is then employed to distinguish the healthy state of the machine from a faulty one using the computed condition indicators. Besides the scalar indicators, also two-dimensional vibration decompositions such as the cyclic spectral correlation maps are used as inputs to the machine learning pipeline. This comprehensive and automated approach ensures both an early and reliable fault detection. Experimental results demonstrate that the fully automated hybrid pipeline can effectively be used for fleet-based health tracking of offshore wind turbine drivetrains.

On-line condition monitoring for rotor systems based on nonlinear data-driven modelling and model frequency analysis

This paper proposes a novel on-line rotor system condition monitoring approach using nonlinear data-driven modelling and model frequency analysis. First, the dynamic process model of the vibration transmission path between the vibration measurement points of two fulcrum structures is established by utilizing nonlinear data-driven modelling. Then, the unique frequency properties are extracted from the established model to reveal, in real time, the health condition of the rotor system. Finally, using the frequency properties as features, the unsupervised learning technology is applied to the on-line monitoring of the rotor system. Compared to conventional condition monitoring methods, the proposed approach can output an early warning 26 min before a shaft fracture occurs, without generating false alarms. Consequently, this approach can greatly enhance diagnostic accuracy, demonstrating its potential to contribute to the advancement of rotor system condition monitoring techniques.

Anomaly Detection in Binary Time Series Data: An unsupervised Machine Learning Approach for Condition Monitoring

A key element of smart manufacturing is condition monitoring and heath controlling of production machines. In today's rapidly evolving landscape of industrial machinery and equipment, optimizing the operation of production lines is critical to ensure high productivity and product quality. Timely detection and prevention of faults in the production process plays a crucial role in minimizing downtime, reducing costs, and ensuring optimal performance. The scientific challenge here is that the increasing number of sensors and actuators with digital input and output signals in the production machines creates different patterns, which are difficult to evaluate using conventional statistical methods. Another difficulty is identifying the cause of the failure to be able to intervene rapidly and in a focused manner in the event of irregularities. For this reason, this research study presented a comprehensive analysis of anomaly detection in binary time series data using various machine learning models. The study included preprocessing of the dataset, normalizing the data, and evaluation of the anomaly detection performance of the different models. The accuracy, detection rate, and F1-score are used as evaluation measures. The execution time of each model is also analyzed. In addition, the identification of sensors that cause anomalies is investigated and the impact of false detections is discussed. Experimental results show the strengths and weaknesses of each model and provide valuable insights for selecting the appropriate anomaly detection approach. The Isolation Forest, Local Outlier Factor, DBSCAN, and kMeans models show high precision and detection, while the Autoencoder and Variational Autoencoder models show high precision but lower detection. The one-class Support Vector Machine model achieves balanced performance. AutoML shows excellent results in recognition rate but is not real-time capable. The results highlight the trade-offs between performance and computational efficiency and point to further research potential in real-time implementation.

Online Condition Monitoring of Industrial Loads Using AutoGMM and Decision Trees

Condition monitoring and fault management approaches can help with timely maintenance planning, assure industry-wide continuous production, and enhance both performance and safety in complex industrial operations. At the moment, data-driven approaches for condition monitoring and fault detection are the most attractive being conceived, developed, and applied with less of a need for sophisticated expertise and detailed knowledge of the addressed plant. Among them, Gaussian mixture model (GMM) methods can offer some advantages. However, conventional GMM solutions need the number of Gaussian components to be defined in advance and suffer from the inability to detect new types of faults and identify new operating modes. To address these issues, this paper presents a novel data-driven method, based on automated GMM (AutoGMM) and decision trees (DTree), for the online condition monitoring of electrical industrial loads. By leveraging the benefits of the AutoGMM and the DTree, after the training phase, the proposed approach allows the clustering and time allocation of nominal operating conditions, the identification of both already-classified and new anomalous conditions, and the acknowledgment of new operating modes of the monitored industrial asset. The proposed method, implemented on a commercial cloud-computing platform, is validated on a real industrial plant with electrical loads, characterized by a daily periodic working cycle, by using active power consumption data.

A novel fault detection and diagnostic Petri net methodology for dynamic systems

Faults can have significant, negative impacts on the operation and performance of simple and complex dynamic systems. Based on the integration of Bayesian network diagnostic features with Petri net formalism, the existing Bayesian-supported Petri net tool has demonstrated the flexibility of using the Petri net approach for diagnosing failure scenario of a dynamic system. However, studies on using the proposed hybrid Petri net approach for condition monitoring and early detection and diagnosis of single and multiple failures in a dynamic system with feedback control loops are yet to be investigated. Thus, this paper presents a methodology to address this research gap using the operation of a water tank level control system as a case study. The method combines the constructed Generalised Stochastic Petri Net (GSPN) model of the system operation with its corresponding fault diagnostic Petri net model, created using the proposed modified Bayesian Stochastic Petri Net (mBSPN) formalism. The GSPN model establishes the causal relationships between the system’s components and/or subsystems. It further identifies deviations in the sensor measurements of the observable process variables characterising the system operation. The information provided by the sensors in the system model are then inputted into the mBSPN model to diagnose the root cause of the observed deviations. The obtained results demonstrated the capability of using the proposed integrated Petri net methodology for system condition monitoring, early fault detection and diagnosis of single and multiple failures in a dynamic system with feedback control loops.

Wind Turbine Predictive Fault Diagnostics Based on a Novel Long Short-Term Memory Model

The operation and maintenance (O&M) issues of offshore wind turbines (WTs) are more challenging because of the harsh operational environment and hard accessibility. As sudden component failures within WTs bring about durable downtimes and significant revenue losses, condition monitoring and predictive fault diagnostic approaches must be developed to detect faults before they occur, thus preventing durable downtimes and costly unplanned maintenance. Based primarily on supervisory control and data acquisition (SCADA) data, thirty-three weighty features from operational data are extracted, and eight specific faults are categorised for fault predictions from status information. By providing a model-agnostic vector representation for time, Time2Vec (T2V), into Long Short-Term Memory (LSTM), this paper develops a novel deep-learning neural network model, T2V-LSTM, conducting multi-level fault predictions. The classification steps allow fault diagnosis from 10 to 210 min prior to faults. The results show that T2V-LSTM can successfully predict over 84.97% of faults and outperform LSTM and other counterparts in both overall and individual fault predictions due to its topmost recall scores in most multistep-ahead cases performed. Thus, the proposed T2V-LSTM can correctly diagnose more faults and upgrade the predictive performances based on vanilla LSTM in terms of accuracy, recall scores, and F-scores.

Using acoustic emission for condition monitoring of the main shaft bearings in 4-point suspension wind turbine drivetrains

ABSTRACT The continuous growth of wind power technology makes condition monitoring of wind turbine components crucially important for their operational efficiency. The main shaft bearings in wind turbines have been identified as one of the most critical components in the system, especially with the ongoing increase in rotor size and weight. This increase made the 4-point suspension drivetrain more preferable. In this study, we present a novel approach for condition monitoring of the main shaft bearings in a 2 Megawatt wind turbine with 4-point suspension drivetrain using primarily acoustic emission (AE). The focus was on the analysis of time and frequency domains of the AE signal, where the dominant frequency of each AE hit was identified and plotted back in the time domain to create the so-called dominant frequency map in specific time intervals for each bearing. A comparison between the two dominant frequency maps of the two bearings gives valuable insights into the condition of the two bearings. The distinctive nature of the dominant frequency bands in the dominant frequency maps presented promising potential for this method. The presented method is straightforward and can be automated and then integrated into a planned predictive maintenance programme for this wind turbine.

A New Approach to Study the Effect of Complexity on an External Gear Pump Model to Generate Data Source for AI-Based Condition Monitoring Application

The external gear pump, like any other hydraulic component, is vulnerable to failure, which may lead to downtime as well as the failure of other components linked to it, thereby causing production loss. Therefore, establishing a condition monitoring system is crucial in identifying failure at an early stage. Traditional condition monitoring approaches rely on experimental data that are collected by means of sensors. However, the sensors utilized in the experiments may have calibration issues, which lead to inaccurate measurements. The availability of experimental data is also limited as it is difficult and expensive to create and detect a fault in a component. Hence, it is essential to develop a simulation model that mimics the performance of the actual system. The data generated from the model can be utilized to create the data source required for automated condition monitoring. A new methodology based on a detailed geometric model for simulating the External Gear Pump is described and compared to two models analyzed in the authors’ previous work, namely Schlosser’s loss model and simple geometric model. In this paper, the three models are compared with experimental data and the method utilized for fault injection. Schlosser’s loss model, as well as the detailed geometric model, are found to be suitable in terms of validation; however, the latter is a better candidate in terms of fault injection. Hence, the detailed geometric model can be implemented as a tool to generate the data source for condition monitoring applications.

Uncovering sensor faults in wind turbines: An improved multivariate statistical approach for condition monitoring using SCADA data

Uncovering sensor faults in wind turbines: An improved multivariate statistical approach for condition monitoring using SCADA data

Road Condition Monitoring Using Vehicle Built-in Cameras and GPS Sensors: A Deep Learning Approach

Road authorities worldwide can leverage the advances in vehicle technology by continuously monitoring their roads’ conditions to minimize road maintenance costs. The existing methods for carrying out road condition surveys involve manual observations using standard survey forms, performed by qualified personnel. These methods are expensive, time-consuming, infrequent, and can hardly provide real-time information. Some automated approaches also exist but are very expensive since they require special vehicles equipped with computing devices and sensors for data collection and processing. This research aims to leverage the advances in vehicle technology in providing a cheap and real-time approach to carry out road condition monitoring (RCM). This study developed a deep learning model using the You Only Look Once, Version 5 (YOLOv5) algorithm that was trained to capture and categorize flexible pavement distresses (FPD) and reached 95% precision, 93.4% recall, and 97.2% mean Average Precision. Using vehicle built-in cameras and GPS sensors, these distresses were detected, images were captured, and locations were recorded. This was validated on campus roads and parking lots using a car featured with a built-in camera and GPS. The vehicles’ built-in technologies provided a more cost-effective and efficient road condition monitoring approach that could also provide real-time road conditions.

Dynamic CUSUM Chart With an Integrated Indicator for Bearing Condition Monitoring

The operation status of bearings plays a critical role in the mechanical systems. Current models for condition monitoring of bearings tend to overlook the complex properties of the process, including its non-Gaussian, nonlinear, and dynamic characteristics. To address these issues, this paper proposes a novel approach for condition monitoring based on a novel dynamic cumulative sum (CUSUM) chart with an integrated indicator. The integrated indicator accounts for the non-Gaussian, nonlinearity, and dynamic properties of the process by utilizing optimal components that are decomposed via independent component analysis (ICA) and its extended versions. The developed dynamic CUSUM chart can adapt to the variation of the process. The proposed condition monitoring strategy is validated using the simulation and experiment data of bearings. Results demonstrate that the proposed method yields a significant improvement in hit rates and a reduction in false alarms compared to conventional methods.

A condition monitoring approach for the detection of defects in the primary suspension of railway vehicles

This work presents the main problems in the primary suspension of metro vehicles and a methodology to detect those using wayside systems. The primary suspension is usually based on rubber elements, which are space saving and have damping properties. However, rubber characteristics change with time, often in a different way in each element. Differences between elements can cause irregular transmission of loads to the wheels, and thus, a reduction or loss of guidance forces. The ageing of rubber can also lead to a general stiffening of the primary suspension, which increases the risk of derailment. This paper presents a wayside system that uses a few strain gauge bridges to detect those problems independently of the vehicle load and speed. Firstly, the system is developed by means of a multibody model. Secondly, a demonstrator is installed in an operating metro exploitation. Finally, the accuracy of the predictions is proven by testing some of the elements in the laboratory.

A machine learning approach for condition monitoring of high voltage insulators in polluted environments

A machine learning approach for condition monitoring of high voltage insulators in polluted environments

3상 DC-AC 컨버터용 필름 커패시터 내부 상태 추정 기법 및 전해 커패시터와 비교 연구

Metalized polymer-film capacitors have acquired a distinctive position among a variety of capacitor types due to their self-healing ability. Therefore, these devices are appropriate for critical power applications that demand reliability and durability. Nevertheless, as converters are increasingly being used for transmissions in networks, it is essential to improve stability to ensure the safety of system operations. Therefore, it is necessary to have a monitoring process that enables predictive maintenance to evaluate the health status and ensure the stability of electrical systems. However, the research in this field concentrates on electrolytic capacitors; and the characteristics of electrolytic and film capacitors differ. Hence, the need for further investigation into film capacitors is evident. This research proposes a condition monitoring approach that employs frequency signal analysis to assess the health status of capacitors in a three-phase AC-DC converter. The capacitor current is subjected to discrete wavelet transform and normalized by various indices, which serve as the input for learning algorithms. In addition, for comparison, capacitor voltage, output current, and output voltage are investigated using the discrete wavelet transform and fast Fourier transform. In this study, various indexes including root-mean-squared value, variance, average, and median, are utilized as inputs for artificial intelligent models to investigate factors affecting film capacitors. Eight learning algorithms are implemented to monitor the health status of film capacitors. The results show that utilizing the discrete wavelet transform combined with indexes for capacitor current yields a high accuracy of approximately 99.85%. These findings offer valuable insights into monitoring film capacitors using advanced techniques, and are anticipated to be informative for practical applications of film capacitor monitoring.