Condition Monitoring Of Equipment Research Articles

Evaluation of the Fast Synchrophasors Estimation Algorithm Based on Physical Signals

The goal of this study is to evaluate the performance of the fast algorithm for synchrophasor estimation proposed on the basis of a physical system. The test system is represented by a physical model of a power system with four synchronous generators (15 and 5 kVA). Three synchronous machines represent steam turbine generators, while the fourth machine represents a hydro generator. The proposed method of accuracy assessment is based on comparison of the original and the recovered signals, using values of amplitude and phase angle. The experiments conducted in the study include three-phase faults, two-phase faults and single-phase faults at various buses of the test model. Functional dependencies of initial signal standard deviation from the recovered signal are obtained, as well as those for sampling rate and window width. Based on the results, the following requirements for measurement system and window width are formulated: sampling rate of analog-to-digital converter should be 10 kHz; and window width should start from 5 ms. In addition, the fast algorithm of synchrophasor estimation was tested on event recorder signals. The sampling rate of these signals was 2 kHz. Acceptable window width for event recorder signals is 8 ms. The algorithm was implemented using programming language Python 3 for the testing purposes. The proposed fast algorithm of synchrophasor estimation can be applied in methods for emergency control and equipment state monitoring with short time response.

ПРОГНОЗНЫЙ МОНИТОРИНГ ТЕХНИЧЕСКОГО СОСТОЯНИЯ ПУТЕМ АНАЛИЗА СОСТАВЛЯЮЩИХ СИГНАЛА ВИБРАЦИИ ПОДШИПНИКОВ ВРАЩАЮЩИХСЯ ЭЛЕКТРИЧЕСКИХ МАШИН

Relevance In industry, any failures and unplanned downtime of machines or systems can degrade or interrupt a company's core business, potentially leading to significant penalties and irreparable loss of reputation. Existing traditional maintenance approaches (failure or routine maintenance) suffer from some assumptions and limitations, such as high prevention or repair costs and inadequate or inaccurate mathematical degradation processes. Due to the trend of intelligent manufacturing, data mining, and artificial intelligence, predictive maintenance control is proposed as a new type of maintenance paradigm only after analytical models predict certain failures or degradations. Equipment condition monitoring can reduce the likelihood of failures occurring. Successful technical diagnostics and monitoring of complex electromechanical subsystems and assemblies provides an increase in service life and reliability of the systems as a whole. Aim of research A general overview of maintenance aims and objectives, which mainly include detection and prevention of failures and malfunctions, maintaining performance within established limits, condition prediction in order to improve machine reliability and full use of its resource, cost minimization, maximization of availability/reliability and multi-criteria optimization. Object of research The object of the study is the technical condition of rolling bearings of rotating electric machines. The main defects that are characteristic of this object are analyzed. In addition, a practical and effective sequential preventive maintenance method for rotating electrical machines is proposed, consisting of four steps: feature generation and fault diagnosis selection, data dimensionality reduction and feature fusion, fault detection and diagnosis, and estimation of the remaining useful life of rotating electrical machine bearings. Vibration signal analysis (from the time domain waveform and spectrum of the original or derivative signal, characteristics can be calculated and their effectiveness in diagnosing a fault can be assessed). Research methods To verify the performance of the proposed method, bearing failure is simulated in the Simulink MATLAB module. The bearing failure is created by feeding the vibration perturbation into the shaft system. Results Based on the results of the simulation, a conclusion on the operability of the proposed method for investigating the technical condition of rotating electric machines is drawn.

A novel multiscale hybrid neural network for intelligent fine-grained fault diagnosis

<abstract> <p>Various intelligent methods for condition monitoring and fault diagnosis of mechanical equipment have been developed over the past few years. However, most of the existing deep learning (DL)-based fault diagnosis models perform well only when applied to deal with limited types of general failures, and these models fail to accurately distinguish fine-grained faults under multiple working conditions. To address these challenges, we propose a novel multiscale hybrid model (MSHM), which takes the raw vibration signal as input and progressively learns representative features containing both spatial and temporal information to effectively classify fine-grained faults in an end-to-end way. To simulate fine-grained failure scenarios in practice, more than 100 classes of faults under different working conditions are constructed based on two benchmark datasets, and the experimental results demonstrate that our proposed MSHM has advantages over state-of-the-art methods in terms of accuracy in identifying fine-grained faults, generality in handling fault classes of different granularity, and learning ability with limited data.</p> </abstract>

A Flexible Monitoring Framework via Dynamic-Multilayer Graph Convolution Network

Due to the continuous technological innovation in industrial processes, many deep learning based methods have shown powerful capability in handing equipment status monitoring, but most of them ignore the temporal features and the dynamic changes of diverse spatial structure of the raw data. Meanwhile, these methods usually focus on handling a single downstream task, but rarely consider different tasks simultaneously. To solve these issues, this paper proposes a more flexible monitoring framework based on dynamic-multilayer graph convolution network, which can be adapted to different downstream tasks simultaneously by agile combinations of the modules according to the different industrial scenarios. First, the time series of fault samples are segmented and constructed into a feature matrix to extract the temporal information by the temporal module. At the same time, in order to fully characterize the dynamic change of different spatial structures among samples, the dynamic graph of each moment is expanded into a multilayer graph through various composition indexes. Then, the spatial information is extracted by the intra-layer and inter-layer convolution operations in the spatial module, and the fused features are applied to different downstream tasks. Finally, the experiments are performed in two different downstream tasks, namely three datasets for fault diagnosis and one dataset for remaining useful life prediction. The results of both diagnosis and prediction are better than those of the comparison algorithms, so the effectiveness and flexibility of the proposed framework are proved.

PD Signal Detection Model Based on Multiscale Simulation of EM Generation and Transmission

Nowadays, the digital twin method has been gradually applied to power equipment condition monitoring. Since a main fundamental part of the digital twin is a simulation process with better fitting for the real one, this paper is devoted to building up a PD detection model to simulate the generation and transmission of electromagnetic (EM) wave signals caused by PDs, especially those inside the void defects. This model combines the micro discharge process with the transmission process inside certain geometry structures. Specifically, two commonly used methods are applied to simulate the processes correspondingly. The dielectric barrier discharge (DBD) model is used for the PD simulation inside void defects, while the finite-difference time-domain (FDTD) method is for the EM signal transmission process. The pulsed discharge current can be obtained during the simulation of DBD, which can function as the excitation source of the transmission process solved by FDTD. According to the PD detection model, the influences on the simulation results are discussed, including the impact on the discharge current caused by coefficient variation and the impact on the transmission process caused by the power equipment geometry structure and materials. As mentioned above, this paper is aimed at building a direct relationship between the observed EM signal and the micro discharge process, which reveals the physical process of PD signal measurement in virtual space that can provide more features for PD pattern recognition or noise elimination, and further the mechanism and theoretical analysis support for digital twin application.

Thermal Fault Diagnosis of Electrical Equipment in Substations Using Lightweight Convolutional Neural Network

Real-time equipment condition monitoring is crucial to ensure the regular operation of the electric system. However, the deployed heavy convolutional neural networks (CNNs) are defective for edge computation and offline diagnosis. The purpose of this article is to present a system for detecting overheating faults of substation equipment using infrared photos and U-Net deep learning techniques. First, a stepwise encoder employs a lightweight CNN (LCNN) based on inverted residuals with depthwise separable convolution. The fault location is then decoded in the decoder using stepwise upsampling and nearest-neighbor interpolation. We additionally incorporate low-level detail feature information from the encoder to include additional fault data. Finally, testing findings on our dataset demonstrated the proposed method’s superior reliability and efficiency under a variety of evaluation metrics. Ultimately, we reached a lighter structure and leading estimation with a small training dataset, so our method is well-suited for deployment on mobile devices.

A Review on Data-Driven Condition Monitoring of Industrial Equipment

This paper presents an up-to-date review of data-driven condition monitoring of industrial equipment with the focus on three commonly used equipment: motors, pumps, and bearings. Firstly, the general framework of data-driven condition monitoring is discussed and the utilized mathematical and statistical approaches are introduced. The utilized techniques in recent literature are discussed. Then, fault detection, diagnosis, and prognosis on the three types of equipment are highlighted using a variety of popular shallow and deep learning models. Applications of these techniques in recent literature are summarized. Finally, some potential future challenges and research directions are presented.

A novel adaptive kernel-guided multi-condition abnormal data detection method

Data acquisition devices and environmental factors may lead to abnormal data and misjudgment of mechanical health status in mechanical work. A new method for abnormal data detection in mechanical equipment health monitoring is proposed in this paper. Firstly, an adaptive kernel bandwidth local outlier factor (AKLOF) detection method is proposed. Furthermore, multiple time-domain statistical features are obtained from sliding window segmentation data, and the anomaly degree of each data segment is calculated. Then, the relationship between the sliding window length and the maximum AKLOF function is studied, and the optimal sliding window length is obtained. Subsequently, the AKLOF value and threshold are calculated according to the optimal sliding window length to detect the final abnormal data. Four groups of experiments result show that the method can accurately identify all the missing data, and can also effectively detect other types of abnormal data, which provide guarantee for equipment condition monitoring.

A Deep-Learning-Based Multi-Modal Sensor Fusion Approach for Detection of Equipment Faults

Condition monitoring is a part of the predictive maintenance approach applied to detect and prevent unexpected equipment failures by monitoring machine conditions. Early detection of equipment failures in industrial systems can greatly reduce scrap and financial losses. Developed sensor data acquisition technologies allow for digitally generating and storing many types of sensor data. Data-driven computational models allow the extraction of information about the machine’s state from acquired sensor data. The outstanding generalization capabilities of deep learning models have enabled them to play a significant role as a data-driven computational fault model in equipment condition monitoring. A challenge of fault detection applications is that single-sensor data can be insufficient in performance to detect equipment anomalies. Furthermore, data in different domains can reveal more prominent features depending on the fault type, but may not always be obvious. To address this issue, this paper proposes a multi-modal sensor fusion-based deep learning model to detect equipment faults by fusing information not only from different sensors but also from different signal domains. The effectiveness of the model’s fault detection capability is shown by utilizing the most commonly encountered equipment types in the industry, such as electric motors. Two different sensor types’ raw time domain and frequency domain data are utilized. The raw data from the vibration and current sensors are transformed into time-frequency images using short-time Fourier transform (STFT). Then, time-frequency images and raw time series data were supplied to the designed deep learning model to detect failures. The results showed that the fusion of multi-modal sensor data using the proposed model can be advantageous in equipment fault detection.

Marine-Hydraulic-Oil-Particle Contaminant Identification Study Based on OpenCV

Particulate pollutants mixed in hydraulic oil will lead to the failure of the marine hydraulic system. Nowadays, the current identification methods of particulate pollutants in oil make it challenging to obtain the specific parameters of pollutants. For this reason, this paper proposes a recognition method of marine-hydraulic-oil-particle pollutants based on OpenCV. The image of particles in the marine hydraulic oil was preprocessed by OpenCV software and using the Canny operator edge detection algorithm to extract the contour of particle pollutants to obtain their area and perimeter. The recognition accuracy reached 95%. Using the Douglas–Peucker algorithm for fit polygons, then image moments to obtain the angle-distance waveform of particulate pollutants, the shape of marine-hydraulic-oil particulate pollutants was successfully identified. The designed method has the advantages of fast calculation efficiency, high accuracy, and real-time detection of various parameters of particulate pollutants in marine hydraulic oil. It has great significance for the fault diagnosis of hydraulic systems and prolonging the working life of hydraulic equipment. This research provides a new idea for the condition monitoring and fault diagnosis of ships and offshore engineering equipment.

Sparse Random Reconstruction of Data Loss With Low Redundancy in Wireless Sensor Networks for Mechanical Vibration Monitoring

Wireless sensor networks (WSNs) for condition monitoring of mechanical equipment have been shown effective for ensuring operational safety and reducing breakdown losses. For battery-supply sensor nodes, energy efficiency for data transmission of mass vibration signals has become a significant challenge. In this article, an efficient low redundant data method is designed and implemented in the WSN nodes for reliable transmission. The proposed method includes two stages. Stage 1 involves the delta-discrete cosine transform (delta-DCT) that reduces the temporal and frequency redundancy of the original data so that the transferred data can be downsized via range coding. In stage 2, the sparse random reconstruction is injected to increase the redundancy of the compressed data to guarantee reliable transmission. Experimental results demonstrate that the original data can be downsized to 50%–55%, and the compression data can be recovered via sparse reconstruction without any data loss.

Partial discharge localization using time reversal: Application to gas insulated switchgear

• Single sensor partial discharge (PD) localization in gas insulated switchgear. • Localization of PD using electromagnetic the time reversal (EMTR) method. • Investigation of the effect of noise, PD length, location, and orientation. • Single sensor localization of multiple, simultaneous PD in gas insulated switchgear. Gas-insulated switchgear (GIS) are some of the most reliable and most expensive equipment that are nowadays used in power systems. Condition monitoring of the power system equipment can help operators detect defects like partial discharges (PDs) early, prevent severe damage to the insulation of the apparatus and the related costs. Different techniques such as time difference of arrival (TDoA) have already been implemented to localize PDs. The TDoA technique requires four synchronized sensors to find the 3D location, it does not yield accurate results in a realistic situation when having multiple PDs, noise, scattering or obstacles which mask the line of sight between the sensor(s) and the PD. To have a powerful and accurate technique that can localize single or multiple PDs in power transformers in the presence of noise, scattering, obstacles, etc., using a single sensor, the time reversal (TR) technique has recently been proposed in both the acoustic and electromagnetic regimes. In this paper, the electromagnetic time reversal (EMTR) technique is applied to localize PDs in GIS, which have a more complicated structure than a transformer. The performance of the method is demonstrated considering several realistic case studies involving two pipe-shape and T-shape parts of a GIS.

A Remote Monitoring and Maintenance System for Industrial Robots

Technologies such as external sensors, data acquisition, trend analysis, and cloud services make remote operation and maintenance of equipment status monitoring more convenient and lower maintenance costs. In this paper, an industrial robot operation and maintenance system is designed. First, the online monitoring of the robot is realized by collecting the control parameters, state parameters, fault information and external sensor monitoring information of the industrial robot. Then, the health baseline of the industrial robot is established based on the historical operation data. The online data is compared with the baseline data to realize the online assessment of the robot’s health status. Finally, a visualization platform for remote operation and maintenance of the robot is built to realize the online display of parameters, health trend tracking and fault status early warning, which is convenient for maintenance personnel to take timely countermeasures.

Research and application of equipment intelligent operation and maintenance platform

The intellectualization of equipment operation and maintenance is one of the main development directions under the background of intelligent manufacturing. Based on lean theory and an intelligent algorithm, this paper establishes the equipment hierarchical management model and equipment intelligent operation and maintenance framework, introduces the lean management strategy and equipment operation state prediction algorithm in the operation and maintenance framework in detail, and finally realizes the design and development of equipment intelligent operation and maintenance platform. The platform has been proved in a steel structure manufacturing enterprise that it can effectively realize equipment condition monitoring, fault diagnosis, and performance prediction, and has a significant effect in reducing accidental shutdown loss and improving machine utilization.

An anomaly detection method based on random convolutional kernel and isolation forest for equipment state monitoring

Anomaly detection plays an essential role in health monitoring and reliability assurance of complex system. However, previous researches suffer from distraction by outliers in training and extensively relying on empiric-based feature engineering, leading to many limitations in the practical application of detection methods. In this paper, we propose an unsupervised anomaly detection method that combines random convolution kernels with isolation forest to tackle the above problems in equipment state monitoring. The random convolution kernels are applied to generate cross-dimensional and multi-scale features for multi-dimensional time series, with combining the time series decomposing method to select abnormally sensitive features for automatic feature extraction. Then, anomaly detection is performed on the obtained features using isolation forests with low requirements for purity of training sample. The verification and comparison on different types of datasets show the performance of the proposed method surpass the traditional methods in accuracy and applicability.

Data-Driven Approach toward Long-Term Equipment Condition Assessment in Sterile Drug Product Manufacturing.

A two-stage data-driven methodology for long-term equipment condition assessment in drug product manufacturing is presented with a case study for a commercially operating aseptic filling line. The methodology leverages process monitoring data. Sensor measurements are partitioned using process information and maintenance schedules that are available on different databases. Data is processed to tackle heterogeneity in sources and formats. The data is cleaned to remove the effects of short-term variabilities and to enhance underlying long-term trends. Two approaches are presented for data analysis: first, anomaly detection using independent component analysis (ICA), where clusters of outliers are identified. The frequency and timing of such outliers yield important insights regarding maintenance schedules and actions. The second approach enables condition monitoring using principal component analysis (PCA). Long-term operational baselines are identified and shifts therein are linked with different process and equipment faults. This approach highlights the impact of equipment deterioration on shifting operational data baselines and shows the potential for the combined application of ICA and PCA for equipment condition monitoring. It can be applied within predictive maintenance applications where the installation of new specialized sensors is difficult, like in the pharmaceutical industry.

Ori-inspired bistable piezoelectric energy harvester for scavenging human shaking energy: Design, modeling, and experiments

Origami structure has nonlinear characteristics, which can broaden the vibration frequency band and improve the dynamic responses. By combining the origami structure with the piezoelectric theory, we can obtain a simple, portable, and bistable energy harvester that can be used to scavenge low-frequency vibration, such as human shaking. Therefore, an ori-inspired bistable piezoelectric energy harvester is proposed. The mathematical model of the harvester is established based on the energy method and Lagrange equation. Dynamic responses and harvesting characteristics are explored by theoretical and experimental methods respectively. Comparing the theoretical and experimental results, the rationality of the established mathematical model is verified. The output performance of the harvester is enhanced by increasing the external energy input and reducing the damping coefficient. On the other hand, by increasing the length of the connecting rod, the harvester can obtain higher output performance at the cost of discarding the bandwidth of interwell vibration. In the experiment, when the excitation acceleration is 1.2 g and the frequency is 6.2 Hz, the ori-inspired harvester could have the root-mean-square voltage of 25.7 V and output power of 2.2 mW at a resistance of 0.3 MΩ, which has great potential to provide continuous power for condition or biological monitoring equipment.

A Review of Fault Diagnosis Methods for Rotating Machinery Using Infrared Thermography.

At present, rotating machinery is widely used in all walks of life and has become the key equipment in many production processes. It is of great significance to strengthen the condition monitoring of rotating machinery, timely diagnose and eliminate faults to ensure the safe and efficient operation of rotating machinery and improve the economic benefits of enterprises. When the state of a rotating machine deteriorates, the thermal energy that is much more than its normal operation will be generated due to the increase in the friction between the components or other factors. Therefore, using the infrared thermal camera to collect the infrared thermal images of rotating machinery and judge the health status of rotating machinery by observing the temperature distribution in the thermal images is often more rapid and effective than other technologies. Nevertheless, after decades of development, the research achievements of infrared thermography (IRT) and its application in various industrial fields are numerous and complex, and there is a lack of systematic sorting and summary of the achievements in this field. Accordingly, this paper summarizes the development and application of IRT as a non-contact and non-invasive tool for equipment condition monitoring and fault diagnosis, and introduces the basic theory of IRT, image processing technology and fault diagnosis methods of rotating machinery in detail. Finally, the review is summarized and some future potential topics are proposed, which will make the subject easier for beginners and non-experts to understand.

Intelligent machine fault diagnosis with effective denoising using EEMD-ICA- FuzzyEn and CNN

ABSTRACT With the advances in smart sensing and data mining technologies of Industry 4.0, condition monitoring of key equipment in manufacturing has brought transformations in production and maintenance management. However, in practical applications, noise from both the working environment and the sensing devices is inevitable, which causes the low performance of data-driven fault diagnosis. To address this challenge, the paper develops a robust two-stage joint denoising method by integrating ensemble empirical mode decomposition (EEMD) and independent component analysis (ICA), with fuzzy entropy discriminant as a threshold. The developed method can filter noisy components from decomposed modal components and reconstruct a new signal with denoised independent components. Moreover, an improved convolutional neural network (CNN) model based on the VGG structure has been constructed as a classifier to achieve end-to-end fault diagnosis. The experimental results demonstrate the high accuracy and superior anti-interference capability of the proposed method for rolling bearing fault diagnosis under various noise levels. Compared with state-of-the-art denoising methods and fault diagnosis methods, the proposed method achieves higher accuracy and robustness under variable noise interference. The proposed method can be applied to broader fault diagnosis tasks of production equipment in complex practical environments.

A Sensor Fusion Method Using Transfer Learning Models for Equipment Condition Monitoring.

Sensor fusion is becoming increasingly popular in condition monitoring. Many studies rely on a fusion-level strategy to enable the most effective decision-making and improve classification accuracy. Most studies rely on feature-level fusion with a custom-built deep learning architecture. However, this may limit the ability to use the widely available pre-trained deep learning architectures available to users today. This study proposes a new method for sensor fusion based on concepts inspired by image fusion. The method enables the fusion of multiple and heterogeneous sensors in the time-frequency domain by fusing spectrogram images. The method’s effectiveness is tested with transfer learning (TL) techniques on four different pre-trained convolutional neural network (CNN) based model architectures using an original test environment and data acquisition system. The results show that the proposed sensor fusion technique effectively classifies device faults and the pre-trained TL models enrich the model training capabilities.