Condition Monitoring System Research Articles

Low-Cost Microcontroller-Based System for Condition Monitoring of Permanent-Magnet Synchronous Motor Stator Windings

Low-Cost Microcontroller-Based System for Condition Monitoring of Permanent-Magnet Synchronous Motor Stator Windings

Acoustic signals analysis from an innovative UAV inspection system for wind turbines

Wind energy is emerging as a leading renewable energy source, with the deployment of large and more innovative wind turbines (WTs). This expansion requires new condition monitoring systems (CMS) and diagnostic techniques to reach competitiveness, improve reliability and availability, and minimize maintenance costs associated with WT operations. This research proposes a novel CMS based on acoustic analysis of WTs, combined with advanced analytics for pattern recognition. An acoustic CMS embedded in an unmanned aerial vehicle is developed to capture, send, and process the sound emitted in the nacelle to an acoustic receiver by a ground station for further analysis to assess the viability of the methodology. The article presents initial results from the laboratory using the fast Fourier transform algorithm, studying the signals in the time-frequency domain aspect and measuring the energy. An advanced signal processing method is presented to filter and define patterns that identify the state and condition of different proposed scenarios. The methodology is tested in a working WT, and the results demonstrate that the acoustic analysis is suitable for maintenance management in WTs.

Combination of generic novelty detection and supervised classification pipelines for industrial condition monitoring

Abstract Machine learning in industrial condition monitoring is currently a rapidly developing field of research, to improve the efficiency and reliability of industrial processes. Many of the used algorithms are supervised methods, which can learn and recognize hidden patterns in the data. However, training data is required to learn these patterns, which can only be generated to a limited extent in an industrial environment due to the high costs involved. Furthermore, it is impossible to represent all possible events in the training data. In contrast, unsupervised or semi-supervised methods can be used to detect new conditions or events. However, these usually do not allow diagnosis or quantification of a fault condition, which is why their usefulness for modern maintenance strategies is limited. Consequently, a robust condition monitoring system should combine the functionality of both approaches. This paper presents a methodology for the combination of supervised classification and semi-supervised novelty detection to build an expandable and adaptable condition monitoring by transferring recurring novelties as new conditions to the supervised classification. A superordinate algorithm is proposed to achieve a stepwise extension of the supervised model based on new conditions detected by novelty detection. With this approach, a condition monitoring system can at first be based on “normal” data of a new machine or process by adding failures or novel conditions step-by-step. Furthermore, the supervised methods can be used to help the corresponding staff identify unknown conditions by analyzing the features selected by the supervised classification. The general workflow is demonstrated for condition monitoring of the pneumatic drive system of a welding gun.

A wideband vibration sensor with piecewise nonlinear and up-frequency coupling strategy for mechanical fault monitoring

Vibration analysis using vibration sensors is an effective method of monitoring mechanical faults. However, most current vibration sensors have a narrow working frequency range and a low-frequency vibration response. Therefore, broadening the frequency band to detect high-frequency vibration signals is imperative. This paper proposes a self-powered vibration sensor based on a broadband hybrid generator (VS-BHG) that incorporates a piecewise nonlinear and up-frequency coupling strategy through collisions between the suspension and fixed part, thereby effectively expanding the operational frequency range. By implementing this strategic approach, the VS-BHG can effectively capture the vibration energy across a broad frequency spectrum ranging from 5 to 38 Hz and exhibit a robust linear response to acceleration within the frequency range of 20–1600 Hz. A machine running state monitoring system based on the VS-BHG is established, enabling the recognition of 16 different running states and fault types in industrial fans with an accuracy rate of 98.38 % using a deep-learning model. This paper presents an efficient and viable method for monitoring broadband mechanical fault.

Enhanced Condition Monitoring of Gearboxes using Fuzzy Comprehensive Evaluation and Group Decision-making: A Case Study

In this study, we introduce a novel methodology for the condition monitoring of gearboxes by integrating fuzzy comprehensive evaluation and group decision-making. Traditional risk assessment methods often suffer from subjective biases and uncertainties inherent in expert evaluations. Our approach mitigates these limitations by aggregating expert opinions using fuzzy membership functions and assigning weights based on the similarity of individual evaluations to the group consensus. This converts qualitative judgments into quantitative measures, resulting in more precise and objective Risk Priority Numbers (RPNs). We validate the efficacy of our methodology through a case study involving a gearbox. The primary failure modes identified include gear tooth wear, misalignment, bearing failure, lubrication failure, and thermal overload. Our results indicate a significant improvement in condition monitoring accuracy, with calculated fuzzy RPN values closely aligning with historical data and expert feedback. Comparative analysis highlights the advantages of our methodology over conventional RPN calculations, particularly in reducing subjective biases and enhancing the reliability of risk assessments. Our findings demonstrate that this methodology can be effectively applied in various industrial settings, establishing a robust framework for mechanical system condition monitoring. Future research should explore integrating advanced data analytics and machine learning techniques to enhance the methodology's accuracy and efficiency.

Research on Key Technology of Wind Turbine Drive Train Fault Diagnosis System Based on Digital Twin

Fault diagnosis of wind turbines has always been a challenging problem due to their complexity and harsh working conditions. Although data-mining-based fault diagnosis methods can accurately and efficiently diagnose potential faults, the visibility is extremely poor. In this paper, digital twin technology is introduced into the fault diagnosis of wind turbine drive train systems, and a wind turbine drive train fault diagnosis method based on digital twin technology is proposed, which monitors and simulates the actual operating condition in real-time by establishing a digital twin model of the wind turbine drive train. In addition, an improved variational modal decomposition combined with particle swarm optimization least squares support vector machine (IVMD-PSO-LSSVM) fault diagnosis method is proposed, which not only improves the accuracy of fault diagnosis but also effectively shortens the diagnosis time and strengthens the response speed of the system. Finally, a digital twin system for condition monitoring and fault diagnosis of wind turbine drive trains is developed based on the Unity 3D platform. Experiments show that the proposed IVMD-PSO-LSSVM can accurately identify fault types with an accuracy rate of 99.1%, which is an improvement of 3.4% compared to before. The proposed digital twin model can be used for real-time monitoring of wind turbine vibration data and provide a more intuitive real-time simulation of the wind turbine’s operating status. This facilitates quick fault location and enables more accurate and efficient maintenance.

Differences in driver takeover performance and physiological responses in conditionally automated driving: Links to emotional instability

When the automated driving system reaches its technical limit, drivers will have to take over control from the vehicle within a limited time. The drivers’ emotional instability may affect whether they can safely control the vehicle during takeover transitions. Accordingly, this paper explores the association between emotional instability and physiological responses and driver takeover performance. 42 drivers engaged in negative emotion induced by movie clips were involved in a sequence of takeover events in a driving simulator. Personality questionnaires were used to assess the driver’s level of emotional instability. Their physiological data before the takeover request (TOR) and operational behavior data fragments for the takeover process were extracted. The results demonstrated that integration of subjective evaluations with electroencephalogram (EEG) and electrocardiogram (ECG) signals allows for more objective measurement of drivers’ level of emotional instability. The statistical analyses indicated that emotional instability has a significant effect on drivers’ heartbeat interval, heart rate, the frequency bands of θ(4–8 Hz), α(8–12 Hz), β(13–30 Hz), γ (30–50 Hz) and lateral takeover performance. Those findings of differences in driver physiological responses and takeover performance under emotional instability may provide additional support for the design of driver state monitoring and adaptive warning systems.

Advancing multimodal diagnostics: Integrating industrial textual data and domain knowledge with large language models

The rapid advancement and application of large language models (LLMs) in various domains prompt an investigation into their potential in the field of prognostics and health management (PHM), particularly for enhancing data-driven model capabilities. This study explores the integration of domain knowledge accumulated in unstructured text data such as technical documents and maintenance logs into diagnostics models using LLMs. The study demonstrates the new possibilities to exploit data that are traditionally underutilized due to their complexity and the presence of domain-specific jargon. By leveraging LLMs for contextual understanding and information extraction from such texts, this study proposes a novel approach that combines textual data with existing condition monitoring systems to improve the accuracy of diagnostics models. A case study on hydrogenerators illustrates the feasibility and value of integrating LLMs into PHM systems. The findings suggest that the incorporation of LLMs can lead to more informed, accurate diagnostics, ultimately enhancing operational efficiency and safety within industrial environments.

A Novel Scalable Digital Data Acquisition System for Industrial Condition Monitoring

A Novel Scalable Digital Data Acquisition System for Industrial Condition Monitoring

Innovative machine learning based approach for reliable and accurate measurement of guide roll alignment in continuous casting plants

Continuous casting (strand casting) is the leading-edge technology in modern steel production. Even slight improvements of the casting process offer a range of benefits, such as improved quality, higher productivity, energy savings and a reduced environmental impact. The precise alignment of a plants guiding rolls is a crucial factor for its performance. Strand condition monitoring systems usually incorporate modules to gauge the alignment of the guiding rolls. These modules measure the angle formed by two successive rolls by using metal rulers fitted with inclination sensors. However, the angle of the ruler only represents a valid measurement where it maintains firm contact with two successive rolls. Most of these systems rely on inductive switches that trigger a measurement at these locations. Yet this approach provides unreliable results due to the trigger’s insufficient accuracy. Furthermore, this method stores only one data point per roll, which is problematic due to significant measurement noise. Apparently, a reliable and accurate alignment measurement remains a significant challenge. As part of an ongoing research project, this paper presents the development of a novel alignment measurement module. The module will provide continuous readings from the inclination sensors without the need for a trigger. In addition, an innovative machine learning based approach to distinguish between valid and invalid measurements is developed. This method enables the averaging of multiple measurements per roll, drastically reducing the impact of measurement noise while providing reliable alignment angles. To overcome the limited availability of real-world data, which is insufficient for model training, the measurement module is simulated to generate artificial training data. Our machine learning model achieves excellent performance on this artificially generated dataset. The paper concludes by outlining how the new approach will be applied to real-world data generated by the new module

Enhanced prediction of reflected spectrum for FBG sensors with metallic coating embedded in CFRP composites: Unveiling the impact of process-induced residual stress and coating thickness

The occurrence of peak-split or distortion in the reflected light of fiber Bragg grating (FBG) sensors with metallic coatings embedded in composites is inevitable during the curing process, regardless of protection layers. In this study, we present a comprehensive methodology to numerically predict the reflected spectrum of metallic-coated FBG sensors, considering the process-induced residual stress in carbon fiber/epoxy composites. The finite element analysis was utilized to simulate the residual stress, which primarily arises from mechanical, thermal, and chemical cure mechanisms of the composites, including the thermosetting resin. Subsequently, the reflected spectra were calculated using the coupled mode theory. Contrary to common expectations, our findings indicate that the coating thickness has minimal influence on the reflected spectrum, while the residual stress and embedding position significantly impact it. By employing this proposed methodology, the number of experimental trials can be reduced, enabling the development of robust structural and state monitoring systems for composites using metallic-coated FBG sensors.

Automatic damping estimation via bootstrap technique and Bayesian analysis for mechanical system condition monitoring

Monitoring mechanical systems that are exposed to dynamic loads during operation requires a technical indicator that provides reliable information on degradation. In this paper, the damping ratio is used as an indicator and can be obtained from monitoring data such as acceleration. Its change from the initial value highly correlates with the system change that can lead to failure. The damping ratio can be relatively easily estimated from the free response data. Such events occur frequently for systems that are subjected to environmental loads such as wind gusts on structures, or sudden stops and changes in the operational cycle.The problem addressed by this research is the fully autonomous detection of these free-response events and the subsequent estimation of the damping ratio, including uncertainty. Therefore, a novel monitoring approach is proposed, utilising the Bootstrap technique and Bayesian analysis. Bootstrap provides automatic free response detection without monitoring the excitation force. Once the event is detected as a valid free response, Bayesian analysis in the form of a gamma-exponential conjugate prior is applied to estimate the damping ratio. As damping estimation is in the form of distribution, uncertainty is also quantified, which is a required feature when monitoring the degradation process and remaining useful life prediction.The monitoring approach is demonstrated and verified numerically and experimentally on a single degree of freedom air-bearing testbed as well as on multiple degrees of freedom beam system. As shown, the combination of the Bootstrap technique and Bayesian analysis provided satisfactory results, demonstrating that the degradation process can be reliably monitored by the proposed approach.

Analysis of a raise boring reamer in operation based on acceleration measurements

AbstractThe understanding of the complex dynamic phenomena occurring during the reaming operations of the raise boring process is limited. Any operational damage of the reamer is very likely to be caused by component failure due to fatigue effects from the dynamic processes. Therefore, an innovative measurement setup consisting of acceleration sensors in three axes was installed on a reamer to capture its movements and vibrations during operation. The capture of the reaming process by acceleration sensors was successful, however no operational damage on the reamer was observed during the reaming of three vertical shafts. The collected vibration data was implemented into a dynamic finite element assessment, which is intended to serve for lifetime predictions of components and predictive maintenance. A further step is investigating the potential for the implementation of the measurement data into a condition monitoring system and for optimizing the reaming operation.

Edge conditional node update graph neural network for multivariate time series anomaly detection

With the rapid advancement in cyber-physical systems, the increasing number of sensors has significantly complicated manual monitoring of system states. Consequently, graph-based time series anomaly detection methods have gained attention due to their ability to explicitly represent relationships between sensors. However, these methods often apply a uniform source node representation across all connected target nodes, even when updating different target node representations. Moreover, the graph attention mechanism, commonly used to infer unknown graph structures, could constrain the diversity of source node representations. In this paper, we introduce the Edge Conditional Node Update Graph Neural Network (ECNU-GNN). Our model, equipped with an edge conditional node update module, dynamically transforms source node representations based on connected edges to represent target nodes aptly. We validate performance on three real-world datasets: SWaT, WADI, and PSM. Our model demonstrates 5.4%, 12.4%, and 6.0% higher performance, respectively, compared to baseline models with best F1 scores. Our code is available at https://github.com/hayoung-jo/ECNU-GNN.

Enhancing interpretability in data-driven modeling of photovoltaic inverter systems through digital twin approach

The utilization of data-driven modeling techniques has been extensively employed in the simulation analysis, power prediction, and condition monitoring of photovoltaic power generation systems. However, the absence of interpretability regarding the intrinsic mechanisms in the modeling process has resulted in numerous constraints in practical implementation and subsequent promotion. To this end, we propose a novel digital twin modeling approach that eliminates the need for injecting additional signals or sensors, estimating unknown parameters in the mechanism model solely by using operational data from physical systems. A time synchronization filter was added to address the frequency mismatch between the actual sampling frequency and the solution step size. The results of numerical research indicate that the proposed digital twin model has the ability to accurately simulate the dynamic characteristics of photovoltaic grid connected inverters. The digital twin model of photovoltaic inverters has achieved good results in the cross experiment of device degradation trend monitoring, indicating that the proposed method is expected to make significant contributions to the simulation, power prediction, and degradation monitoring of grid connected photovoltaic systems.

Fusion of Multi-Layer Attention Mechanisms and CNN-LSTM for Fault Prediction in Marine Diesel Engines

Timely and effective maintenance is imperative to minimize operational disruptions and ensure the reliability of marine vessels. However, given the low early warning rates and poor adaptability under complex conditions of previous data-driven fault prediction methods, this paper presents a hybrid deep learning model based on multi-layer attention mechanisms for predicting faults in a marine diesel engine. Specifically, this hybrid model first introduces a Convolutional Neural Network (CNN) and self-attention to extract local features from multi-feature input sequences. Then, we utilize Long Short-Term Memory (LSTM) and multi-head attention to capture global correlations across time steps. Finally, the hybrid deep learning model is integrated with the Exponential Weighted Moving Average (EWMA) to monitor the operational status and predict potential faults in the marine diesel engine. We conducted extensive evaluations using real datasets under three operating conditions. The experimental results indicate that the proposed method outperforms the current state-of-the-art methods. Moreover, ablation studies and visualizations highlight the importance of fusing multi-layer attention, and the results under various operating conditions and application scenarios demonstrate that this method possesses predictive accuracy and broad applicability. Hence, this approach can provide decision support for condition monitoring and predictive maintenance of marine mechanical systems.

Inter-turn Short Circuit Fault Diagnosis and Severity Estimation for Wind Turbine Generators

While preventive maintenance is crucial in wind turbine operation, conventional condition monitoring systems face limitations in terms of cost and complexity when compared to innovative signal processing techniques and artificial intelligence. In this paper, a cascading deep learning framework is proposed for the monitoring of generator winding conditions, specifically to promptly detect and identify inter-turn short circuit faults and estimate their severity in real time. This framework encompasses the processing of high-resolution current signal samples, coupled with the extraction of current signal features in both time and frequency domains, achieved through discrete wavelet transform. By leveraging long short-term memory recurrent neural networks, our aim is to establish a cost-efficient and reliable condition monitoring system for wind turbine generators. Numeral experiments show an over 97% accuracy for fault diagnosis and severity estimation. More specifically, with the intrinsic feature provided by wavelet transform, the faults can be 100% identified by the diagnosis model.

Smart EV Charging Station Monitoring and User Feedback System

The efficiency and reliability of electric vehicle (EV) charging stations are pivotal for user satisfaction and broader EV adoption. This research introduces a system that monitors essential parameters of EV charging stations, such as temperature, voltage, and current, using advanced sensors and the ESP32 microcontroller. The collected data is transmitted through the Bylnk platform, for real-time monitoring and analysis. To enhance the user experience, a mobile application has been developed, allowing users to review these critical details and receive notifications. The app provides a detailed overview of each charging station's performance, helping users identify the best stations based on historical data and user reviews. By offering insights into the operational status and efficiency of charging stations, the system aids users in making informed decisions about where to charge their vehicles. This integration of real-time monitoring with user-friendly mobile access improves user convenience and optimizes the EV charging infrastructure management.

Multi-criterion analysis-based artificial intelligence system for condition monitoring of electrical transformers

Multi-criterion analysis-based artificial intelligence system for condition monitoring of electrical transformers

Detection of Signal Integrity Issues in Vibration Monitoring Using One-Class Support Vector Machine

AbstractThis paper presents an analysis of the common signal integrity issues in vibration monitoring caused by sensor saturation and signal distortion, or sensor loosening and detachment, and the development of a method of detecting the occurrence of vibration signal integrity issues using a one-class support vector machine. For this, vibration signals with distortions due to sensor saturation and/or sensor detachment are analysed to determine parameters sensitive to common integrity issues. These features are then extracted from training data of good quality vibration signals. Principal Component Analysis is used to dimensionally reduce the parameters which are then used in the training of the one-class support vector machine. The proposed method was validated on trained and untrained health conditions. Additionally, model sensitivity was also tested on trained health conditions with varying defect sizes. The method successfully detected all signal integrity issues tested proving to be an effective pre-processing step in condition monitoring systems for increased reliability and accuracy.