Implement Condition Monitoring Research Articles

Enhancing Friction Stir Welding: Quality Machine Learning Based Friction Stir Welding Tool Condition Monitoring

Ensuring the quality and optimizing the tool in Friction Stir Welding (FSW) process is quite complex and the solution relies on implementing Condition Monitoring. The major impact of this process yields good quality welds and cuts down the non-operational timing and cost. Condition Monitoring is the key to find a solution to the challenging problem of ensuring quality and optimizing the tool in the FSW process. The creation of a graphical user interface (GUI) and the development and comparison of several models, including Decision Tree (DT), Random Forest (RF), Light Gradient Boosted Machine (LGBM), and Extreme Gradient Boosting (XGBoost), are the main objectives of this study. By offering an uniform interface for tracking and evaluating tool condition data, GUI can make it easier for operators and the maintenance crew to collaborate. Vibration analysis is the first step in tool condition monitoring. Al5083 and AZ31B are used as the workpiece and H13 as the tool in this investigation. The signals are obtained from the experimental setup via DAQ, and LabView processes them. A Python script converts the raw signals into statistical data. Following that, the data was loaded into ML models and optimized using Optuna. TKinter has been used to create the GUI. For prediction, the best models were included in the GUI. By the deployed models, LGBM generates 96% for 1000 rpm, 96.55% for 1200 rpm, and 95.90% for 1400 rpm for Al5083 93.22% for 1000 rpm, 99.29% for 1200 rpm, and 91.50% for 1000 rpm for AZ31B. For real-time prediction, these models are thus connected to a graphical user interface. In each case, the LGBM classifier topped the others. This work served as an initial basis for the creation of a semi-onboard diagnostic approach that requires minimal human input.

Implementing Condition Monitoring for Medium Voltage Switchgear for the Distribution Network in Singapore

Implementing Condition Monitoring for Medium Voltage Switchgear for the Distribution Network in Singapore

Deep Learning Approach to Within-Bank Fault Detection and Diagnostics of Fine Motion Control Rod Drives

Control rod motion is one of the primary means of regulating the rate of fission in a nuclear reactor core to ensure safe and stable operation. Reactor power distribution and thermal power output can be fine-tuned by adjusting the control rod position. For high-precision control of rod movements, Fine Motion Control Rod Drives (FMCRDs) are often used. The operation of FMCRDs provides a unique opportunity to implement condition monitoring related to the intermittency of motion and the use of control rod banks. This research sets out to detect three types of faults in an electrically driven FMCRD. In addition to detecting faults, this work will attempt to determine both the type of fault and the source of each fault, completing the fault detection and diagnostics (FDD) pipeline on a scarcely researched system. The three types of faults to be investigated are short-circuit faults, ball screw wear faults, and ball screw jam faults. This is a potential advancement to the within-bank FDD of this specific drive system intended for deployment in an advanced nuclear reactor plant. Using encoder-decoder structured convolutional neural networks and autoencoders, the three tested faults were confidently detected and isolated as well as reasonably diagnosed by monitoring the FMCRD servomotor torque.

Physics verification and validation for transferring data between bearings

Bearings are critical components for transferring load and motion between subsystems with reduced friction for rotational equipment. Manufacturers implement condition monitoring technology to prevent failures of bearings by using different data acquisition methods, such as vibration, acoustics, temperature, motor current, and ultrasonic sensing, to monitor and predict changes that could indicate early equipment degradation. However, the data quality and availability varies depending on the application, and low data availability from physical environments can lead to poorly trained models. This paper explores how to transfer data and information about failure propensity between bearings of different sizes using a combined physics and data-driven approach. Though the approach is exemplified with roller bearings, the method is extensible to other types of failures in different-sized equipment. Data are generated using an experimental test stand measuring failure of one component, with the intent to scale findings to represent a real-world system; findings are verified with a physics-based model. Data from three different bearing sizes, i.e. 6205, 6206, and 6207, are used to train the algorithms to identify similarity among the sets. Classifiers trained with the raw data provide over 90% accuracy, leading to the conclusion that the data classes are separable based on bearing size. 6205 and 6207 data were scaled to simulate 6206 and tested to see if a classifier could differentiate between the true 6206 bearing data and the simulated bearing data derived from the 6205 and 6207 data. In the simulated data, the classifier accuracy for each algorithm dropped below 90% to as low as 50% (Naive Bayes case). The lower accuracy implied a greater overlap in the vibration features, increasing the data similarity between the different bearing sizes. Future work will further investigate defining dimensionless numbers as scaling parameters for bearing data.

Siamese multiscale residual feature fusion network for aero-engine bearing fault diagnosis under small-sample condition

Implementing condition monitoring and fault diagnosis of aero-engine bearings is crucial to ensure that aircraft operate safely and reliably. In engineering practice, the fault data for aero-engine bearings are extremely limited. However, the traditional fault diagnosis methods have two shortcomings under extremely small sample conditions: (1) they have limited diagnostic performance and generalization ability, and (2) they do not mine fault information sufficiently or efficiently. This article proposes a Siamese multiscale residual feature fusion network (SMSRFFN) for aero-engine bearing fault diagnosis under small-sample conditions to overcome the weaknesses above. In the proposed SMSRFFN, the training samples are first paired according to the matching rules to realize the expansion of the sample size. Second, a multiscale residual feature extraction network (MSRFEN) is constructed to excavate the fault features of different scales and speed up the convergence speed of the network. Then, a multiscale attention mechanism feature fusion module (MSAMFFM) is designed to achieve efficient fusion of fault features at different scales. Finally, the distance of the input sample is measured based on the fused deep feature representation to identify the fault state of the aero-engine bearing. The proposed SMSRFFN is evaluated using three bearing fault data and also compared with some state-of-the-art small-sample diagnostic methods. The experimental results demonstrate the effectiveness and superiority of the proposed SMSRFFN in mining fault information and improving diagnosis accuracy under extremely small sample conditions.

Condition Monitoring of Mechanical Components Based on MEMED-NLOPE under Multiscale Features

An increasing popularity of researches focuses on the vibration signal with the characteristics of nonstationary, nonlinear, and strong noise interference. A nonlinear dimension and feature reduction method called multiple empirical mode entropy decomposition-nonlocal orthogonal preserving embedding (MEMED-NLOPE) is proposed to implement condition monitoring in this paper. Different from multiple empirical mode decomposition (MEMD), MEMED adopts maximum entropy method, which can directly output the subsignal with the maximum correlation and realize nonlinear dimensionality reduction. Besides, multiscale feature extraction method is used during preprocessing nonlinear data process, which realizes feature reduction. Finally, nonlocal orthogonal preserving embedding algorithm-exponentially weighted moving average (NLOPE-EWMA) realizes the automatic detection of the fault. Taking the laboratory rolling bearing test and naval gun pendulum mechanism test as cases, the effectiveness of MEMED-NLOPE is verified.

Combining Design Thinking and Agile to Implement Condition Monitoring System: A Case Study on Paper Press Bearings

Nowadays, technologies and strategies to bring processes toward the Industry 4.0 paradigm go beyond simple paper proposals and focus on their implementation. This article presents a case study of University-Business collaboration focused on technology transfer as part of a viable solution to real problems in the company. The objective of the collaborative project was the transfer of knowledge and use of technologies by the research group and, together with the knowledge based on the experience of the company's maintenance department, to solve a real problem in the industry. Therefore, we set out with the industry the objective of implementing a system for early detection of failures in press bearings to ensure that maintenance personnel performs maintenance actions on time. We combined agile methodologies and design thinking for the implementation because agile methodologies allow controlling the development progressively, and design thinking focuses on the user's needs. The implementation process was executed in six phases: empathize, ideate, define, test, prototype, and implement. The project was successfully implemented in six months and met industry requirements. Combining agile methodologies and design thinking was fundamental for the implementation, as it was possible to get feedback from industry personnel and to have materialized progress in each meeting.

Life Monitoring and Evaluation Strategy of Traction Transformer on Freight Line

The traction transformer is regarded as the core power equipment in the traction power supply system of railway freight-dedicated line. As the time goes on, lots of traction transformers are under the long-term service status in the domestic electrified railway. Nevertheless, based on the present time-based preventive test code for electric power equipment, it is difficult to conduct the decision-makings on theirs running status. Therefore, it is necessary to study and implement condition monitoring and diagnosis (CMD) technologies. In this paper, some typical characteristics of aging conditions are described, such as tensile strength (TS), polymerization degree (DP), CO/CO2 content, furfural concentration, water content. Some analysis methods with development prospect are analyzed based on dielectric response theories, such as recovery voltage (RV) method, polarization-depolarization current (PDC) method and frequency domain spectrum (FDS) method of the oil-immersed insulation traction transformer. However, the insulation aging of traction transformer is a complex kinetic process of chemical reaction, and it is definitely necessary to combine all the effective methods together and synthetically analyze the experimental data and parameters for obtaining a comprehensive conclusion.

Impact of condition monitoring on reciprocating compressor efficiency

PurposeReciprocating compressors offer an efficient method of compressing almost any gas composition in a wide range of pressures and have numerous applications. Condition monitoring of critical rotating machinery is widely accepted by operators of centrifugal compressors. However, condition monitoring of reciprocating machinery has not received the same degree of acceptance. An earlier study (Townsend et al., 2016) was conducted on temperature monitoring. The purpose of this paper is to examine the impact of continuous pressure monitoring on electric-driven compressors.Design/methodology/approachThis research analyzes the impact of continuous pressure monitoring on a fleet of 14 compressors transporting CO2 for enhanced oil recovery. The reliability and efficiency data on 14 reciprocating compressors over a three-year period were analyzed for failures detectable by the condition monitoring technology. The engineering economic analysis is presented to determine the impact this technology will have on the productivity of the compressors.FindingsThe study considers utilizing condition monitoring technology to analyze the pressure of the swept volume of the compressor cylinders. The results of the study indicate that continuous pressure monitoring technology has a strong impact on the productivity of the compressor fleet. The internal rate of return not only exceeds the operators hurdle rate, but the payback period is also dramatic. Pressure monitoring was found to be economically better than temperature monitoring.Originality/valueThe study reveals the economic benefits of implementing condition monitoring in the form of continuous pressure monitoring on reciprocating compressors.

An Information Processing Framework Facilitating the Implementation of Condition Monitoring in Cyber-Physical Systems

Implementing condition monitoring functionality in production machinery often proves to be a difficult task. Device-and process-specific algorithms must be created while inhomogeneous industrial communication networks hinder the integration of control signals and process variables. Further challenges arise from the advance of flexible Cyber-Physical Systems (CPS) and the Industrial Internet of Things (IIoT). They demand a service-oriented condition monitoring architecture, which seamlessly adapts to quickly changing production topologies. Existing condition monitoring systems (CMS) and reference architectures for CMS however do not possess the capabilities to meet the requirements originating from CPS and the IIoT. This paper presents a software module serving as a generic framework to ease the implementation of decentralized condition monitoring functionalities. A decentralized component, the monitoring module, constitutes a part of a holistic condition monitoring architecture managed by a central server deployed on an edge or cloud server. Therefore, the monitoring module offers an interface through which its data processing flow and algorithms are entirely remotely configurable during operation. Algorithms are encapsulated in function blocks, which can easily be setup and interconnected. Scalability is ensured by the use of web technology optimized for efficient data handling and parallel data-processing. The examined use cases show the potential of the developed software module for its deployment in the use case of monitoring turn-mill-centers.

A novel approach for data-driven process and condition monitoring systems on the example of mill-turn centers

Implementing condition monitoring functionality in production machinery often proves to be a difficult task. Device- and process-specific algorithms must be created while inhomogeneous industrial communication networks hinder the aggregation of control signals and process variables. Further challenges arise from the advance of flexible cyber-physical systems (CPS) and the industrial internet of things (IIoT). They demand a service-oriented condition monitoring architecture, which seamlessly adapts to quickly changing production topologies. In this context, data-driven systems which are capable of unsupervised learning are promising approaches. The aim is the autonomous identification of significant process variables and patterns. This paper describes a machine learning approach for a condition and process monitoring system on the basis of pattern recognition within structure-borne noise of rotating cutting machinery. Process states are defined under application of non-negative matrix factorization (NMF). A production model is learned and deployed on the basis of Gaussian mixture models (GMM) and hidden Markov models (HMM) in a two stage process. Additionally a generic framework to ease the implementation of decentralized condition monitoring functionalities is given. A decentralized component, the monitoring module, constitutes a part of a holistic condition monitoring architecture managed by a central server. The approach is evaluated on the example of mill-turn centers.

Condition Monitoring of Sensors in a NPP Using Optimized PCA

An optimized principal component analysis (PCA) framework is proposed to implement condition monitoring for sensors in a nuclear power plant (NPP) in this paper. Compared with the common PCA method in previous research, the PCA method in this paper is optimized at different modeling procedures, including data preprocessing stage, modeling parameter selection stage, and fault detection and isolation stage. Then, the model’s performance is greatly improved through these optimizations. Finally, sensor measurements from a real NPP are used to train the optimized PCA model in order to guarantee the credibility and reliability of the simulation results. Meanwhile, artificial faults are sequentially imposed to sensor measurements to estimate the fault detection and isolation ability of the proposed PCA model. Simulation results show that the optimized PCA model is capable of detecting and isolating the sensors regardless of whether they exhibit major or small failures. Meanwhile, the quantitative evaluation results also indicate that better performance can be obtained in the optimized PCA method compared with the common PCA method.

Impact of the gate driver voltage on temperature sensitive electrical parameters for condition monitoring of SiC power MOSFETs

Condition monitoring using temperature sensitive electrical parameters (TSEPs) is widely recognized as an enabler for health management of power modules. The on-state resistance/forward voltage of MOSFETs, IGBTs and diodes has already been identified as TSEPs by several researchers. However, for SiC MOSFETs, the temperature sensitivity of on-state voltage/resistance varies depending on the device and is generally not as high as in silicon devices. Recently the turn-on current switching rate has been identified as a TSEP in SiC MOSFETs, but its temperature sensitivity was shown to be significantly affected by the gate resistance. Hence, an important consideration regarding the use of TSEPs for health monitoring is how the gate driver can be used for improving the temperature sensitivity of determined electrical parameters and implementing more effective condition monitoring strategies. This paper characterizes the impact of the gate driver voltage on the temperature sensitivity of the on-state resistance and current switching rate of SiC power MOSFETs. It is shown that the temperature sensitivity of the switching rate in SiC MOSFETs increases if the devices are driven at lower gate voltages. It is also shown, that depending on the SiC MOSFET technology, reducing the gate drive voltage can increase the temperature sensitivity of the on-state resistance. Hence, using an intelligent gate driver with the capability of customizing occasional switching pulses for junction temperature sensing using TSEPs, it would be possible to implement condition monitoring more effectively for SiC power devices.

Fault diagnosis of compressor based on decision tree and fuzzy inference system

There is a strong need to use different methods for maintaining high product quality and safety production line, given the rapid technological progress. Thus, the condition monitoring is widely used as an efficient method in various industries. A variety of methods have been used so far in order to implement condition monitoring system, the most common one has been seismic waves analysis in the time-frequency domain. The current paper proposes a Fuzzy Inference System for monitoring the status of the compressor based on Daubechies wavelet transform and decision trees. The J48 algorithm was used as a tool for classification design and selection of effective features on troubleshooting. The J48 algorithm output results showed that signal processing technique by Daubechies wavelet mother has the highest accuracy for the implementation of the FIS system. The J48 algorithm output is a decision tree used on production of if-then rules and fuzzy set membership functions of the system. Finally, the combination of WT-J48-FIS overall accuracy for classifying compressor defects of 93.33 % was obtained.

Effect of Condition Monitoring on Risk Mitigation for Steam Turbines in the Forest Products Industry

Industry has been implementing condition monitoring (CM) for turbines to minimize losses and to improve productivity. Deficient conditions can be identified before losses occur by monitoring the equipment parameters. For any loss scenario, the effectiveness of monitoring depends on the stage of the loss scenario when the deficient condition is detected. A scenario-based semi-empirical methodology was developed to assess various types of condition monitoring techniques, by considering their effect on the risk associated with mechanical breakdown of steam turbines in the forest products (FPs) industry. A list of typical turbine loss scenarios was first generated by reviewing loss data and leveraging expert domain knowledge. Subsequently, condition monitoring techniques that can mitigate the risk associated with each loss scenario were identified. For each loss scenario, an event tree analysis (ETA) was used to quantitatively assess the variations in the outcomes due to condition monitoring, and resultant changes in the risk associated with turbine mechanical breakdown. An application was developed following the methodology to evaluate the effect of condition monitoring on turbine risk mitigation.

Measurement Science for Prognostics and Health Management for Smart Manufacturing Systems: Key Findings from a Roadmapping Workshop.

The National Institute of Standards and Technology (NIST) hosted the Roadmapping Workshop - Measurement Science for Prognostics and Health Management for Smart Manufacturing Systems (PHM4SMS) in Fall 2014 to discuss the needs and priorities of stakeholders in the PHM4SMS technology area. The workshop brought together over 70 members of the PHM community. The attendees included representatives from small, medium, and large manufacturers; technology developers and integrators; academic researchers; government organizations; trade associations; and standards bodies. The attendees discussed the current and anticipated measurement science challenges to advance PHM methods and techniques for smart manufacturing systems; the associated research and development needed to implement condition monitoring, diagnostic, and prognostic technologies within manufacturing environments; and the priorities to meet the needs of PHM in manufacturing. This paper will summarize the key findings of this workshop, and present some of the critical measurement science challenges and corresponding roadmaps, i.e., suggested courses of action, to advance PHM for manufacturing. Milestones and targeted capabilities will be presented for each roadmap across three areas: PHM Manufacturing Process Techniques; PHM Performance Assessment; and PHM Infrastructure - Hardware, Software, and Integration. An analysis of these roadmaps and crosscutting themes seen across the breakout sessions is also discussed.

Analysis of the Degradation Tendency of NPP Cables Using a Wireless Cable Indenting Robot

In nuclear power plants, there are many cables that perform safety-related functions. These cables should implement condition monitoring during the operation period in the nuclear power plant, in order to assess the remaining qualified life and extend the qualified life. In this study we focused on the indenting method which can measure the hardness of the cable jacket. This method is selected because it is non-destructive and requires short testing time and small sized equipment. In order to address the problems with the existing indenting test equipment, we developed new indenting test equipment, which could automatically move on the surface of the object cable. The newly developed equipment is designed for a small-sized and light-weight robot using wireless communication in order to implement condition monitoring in a harsh environment or locations that are inaccessible to the tester. The developed wireless cable indenting robot is composed of three parts, which are mechanical and electrical hardware parts and remote-control part. In order to analyze the degradation tendency of the cable, we prepared four thermally aged specimens and one un-aged specimen. Using the developed robot, we measured the modulus of the cable jacket of each specimen. The test data showed that the modulus of the cable jacket increased linearly as the accelerated aging time increased. From these results, we can analyze the degradation trends pertaining to cables installed in nuclear power plant according to the operation period.

Data Mining and Visualization of Diagnostic Messages for Condition Monitoring

Complex technical systems consist of a huge amount of electromechanical subsystems and components with more or less interdependencies. During the operational phase these systems are subject to wear and tear and other degradation mechanisms. Therefore condition monitoring is a major challenge for operators of such systems. A well-established practice of implementing condition monitoring is to equip the system, or its functionally relevant components with appropriate sensors and measurement technology. Selection of sensors and diagnostic algorithms requires a deep understanding of physical correlations which is usually a core competence of the system provider. Moreover, subsequent installation of a condition monitoring system using additional sensors in most cases is attended by an interference with the system and may affect its functionality if constructional changes of the system for sensor mounting become necessary. A more simple way to access data for technical diagnosis is the use of an on board diagnosis system that collects the system messages that rise while the system is in operation. To control and monitor functionality and interaction between subsystems, each of them generates internal state variables and communicates a subset of these as system messages via task specific bus systems. Here the challenge is to find a suitable way of analyzing the messages. The present paper discusses how to handle system messages for condition monitoring purposes by using data mining and visualization.

Research on Condition Monitoring of Bearing Health Using Vibration Data

A reliable condition monitoring system is very useful in a wide range of industries to detect the occurrence of incipient defects so as to prevent machinery performance degradation, malfunction and sudden failure. Among the rotating machinery, many mechanical problems are attributed due to bearing failures. So implementing condition monitoring for bearing is critically needed. Considering that most research for condition monitoring only focus on detecting the existing fault, this paper add degradation tendency prognostics into the condition monitoring process. The kernel of bearing condition monitoring method presented in this paper is related to condition features extraction and remaining useful life prediction. The former is realized by the comprehensive vibration analysis for specific fault frequencies. The latter is achieved by adaptive neuron-fuzzy inference system based on extracted degradation signal. For illustration purpose, a bearing case from NASA data repository is used to validate the feasibility of the proposed method. The result indicates that the performance degradation of bearing can be effectively monitored and the predicted remaining useful life with 5.6% relative error can be the important reference for maintenance decision making.

Accuracy improvement of indenting test results by using wireless cable indenting robot

In nuclear power plants, there are many cables that perform safety-related functions. These cables should implement condition monitoring during the operation period in the NPP, in order to assess the remaining qualified life and extend the qualified life. In this study we focused on the indenting method, which can measure the hardness of the cable jacket. This method is selected because it is nondestructive and requires short testing time and small sized equipment. In order to address the problems of the existing indenting test equipment, we developed new indenting test equipment, which could automatically move on the surface of the object cable. The newly developed equipment is designed for a small-sized and lightweight robot using wireless communication in order to implement condition monitoring in a harsh environment or locations that are inaccessible to the tester. The developed wireless cable indenting robot is composed of three parts, mechanical and electrical hardware parts and remote-control part. In order to verify accuracy improvement of indenting test data, an indenting test for cable specimens using both existing indenting test equipment and wireless cable indenting robot was performed. Analysis of the indenting test results shows that the accuracy of the results obtained using the wireless cable indenting robot is improved remarkably.