Machine Health Research Articles

VR-based digital twin for remote monitoring of mining equipment: Architecture and a case study

BackgroundTraditional methods for monitoring mining equipment rely primarily on visual inspections, which are time-consuming, inefficient, and hazardous. This article introduces a novel approach to monitoring mission-critical systems and services in the mining industry by integrating virtual reality (VR) and digital twin (DT) technologies. VR-based DTs enable remote equipment monitoring, advanced analysis of machine health, enhanced visualization, and improved decision making. MethodsThis article presents an architecture for VR-based DT development, including the developmental stages, activities, and stakeholders involved. A case study on the condition monitoring of a conveyor belt using real-time synthetic vibration sensor data was conducted using the proposed methodology. The study demonstrated the application of the methodology in remote monitoring and identified the need for further development for implementation in active mining operations. The article also discusses interdisciplinarity, choice of tools, computational resources, time and cost, human involvement, user acceptance, frequency of inspection, multiuser environment, potential risks, and applications beyond the mining industry. ResultsThe findings of this study provide a foundation for future research in the domain of VR-based DTs for remote equipment monitoring and a novel application area for VR in mining.

Attention-based ConvNeXt with a parallel multiscale dilated convolution residual module for fault diagnosis of rotating machinery

Convolutional Neural Networks have promoted development of the fault diagnosis in the machine prognostics and health management. However, the existing methods have limited applicability under strong noisy conditions. We propose an attention-based ConvNeXt with a parallel multiscale dilated convolution residual module for the rotor fault diagnosis. The parallel multiscale dilated convolution residual module is firstly introduced to filter noise and extract multiscale discriminative features intelligently. The multi-head attention module is then embedded in ConvNeXt for global discriminative features adaptively. Focal Loss is finally implemented for identifying hard-to-classify samples to further improve the diagnostic accuracy. The application results demonstrate the superiority and robustness of our method in two case studies.

Energy-driven TBM health status estimation with a hybrid deep learning approach

This study provides a hybrid deep learning approach that enables the researchers to estimate the Tunnel Boring Machine (TBM) health status during excavation process using energy consumption data. By analyzing the energy forms and paths in the tunneling process, the key energy factors' consumption performance behavior affecting TBM safety is identified. The Ensemble Empirical Mode Decomposition (EEMD) method and a Convolution neural network-Long short term memory-Depth Neural Network (CNN-LSTM-DNN, CLDNN) model are integrated for Health Performance Parameter (HPP) prediction. A classification method is constructed by the Generative Adversarial Imputation Networks (GAIN) model for health status estimation. The SHapley Additive exPlanations (SHAP) method is performed to quantify the relationships between energy consumption and system health status. A tunneling case in Singapore is used to test the effectiveness of the developed methodology. It is found that (1) the proposed method provides TBM cutter wear and cooling oil prediction with a R-square (R2) of 0.8591 and 0.8727, respectively, and (2) classification results indicate that the estimation F1-score of 0.9863 can be achieved based on health status evaluation indicator (HSEI). The novelty of the proposed approach lies in its capabilities of (1) fully extracting and utilizing the energy consumption information in every regard to predict TBM health performance; (2) providing a good one-step-ahead health status estimation for the TBM excavation process. Therefore, this study proposes a feasible direction for the application of energy consumption data in TBM health status estimation.

An on-line imaging sensor based on magnetic deposition and flowing dispersion for wear debris feature monitoring

Sensors that can real-time monitor the wear particles suspended in lubricating oil are of great value in evaluating the health state of various machines (e.g., engines, transmissions). Here we report an on-line imaging sensor with a new structure, operating with magnetic deposition and flowing dispersion methods. A highly divergent magnetic field is generated by the cage electromagnet to collect ferromagnetic particles which will then disperse via the flow field in the disc-like channel. Wear particles in the chain- and separated-form are imaged respectively by the imaging module to acquire four debris features (concentration, size distribution, shape, and colour). We demonstrated that the developed sensor offers a low size detection limit (1–10 μm), a broad concentration-sensing range (2–200 ppm), and a high throughput (∼100 ml/min). Beyond that, the type and material of wear debris can also be identified in the test, utilising the shape and colour features getting by the sensor. This sensor shows promise for use in monitoring the machine health condition and studying the wear mechanism of surfaces.

A multi-function integrated PVDF transducer for fault detection and speed measurement of cylindrical roller bearings

Monitoring the condition of bearings in rotating machinery is crucial to ensure reliability, safety, and efficiency. Recently, integrated sensors have become more popular than traditional accelerometers in bearing condition monitoring because of their compact design, cost-effectiveness, and improved efficiency in fault detection through a short transmission path. This study presents a new approach to bearing condition monitoring by developing a bearing with an integrated Polyvinylidene Fluoride transducer. This polymer-based transducer, known as PVDF, has a piezoelectric effect which generates electric charge in response to mechanical excitation. The PVDF transducer was selected for its high sensitivity, flexibility, and thin structure. The developed bearing with the integrated PVDF transducer can detect faults under constant and variable speeds, and the damage size can be estimated with an accuracy of 10%. It outperformed a commercial accelerometer in noisy environments with artificial impacts from an air motor, making it a highly reliable option for bearing condition monitoring in noisy industrial settings. The PVDF transducer can also be used as a low-cost speed sensor with an accuracy of 5.2%, offering a versatile solution for machine monitoring. By simultaneously recording rotational speed and vibration, the developed integrated transducer provides a comprehensive view of machine health and performance as a multi-function transducer, enabling more accurate and efficient maintenance scheduling.

Cardiovascular Health Management in Diabetic Patients with Machine-Learning-Driven Predictions and Interventions

The advancement of machine learning in healthcare offers significant potential for enhancing disease prediction and management. This study harnesses the PyCaret library—a Python-based machine learning toolkit—to construct and refine predictive models for diagnosing diabetes mellitus and forecasting hospital readmission rates. By analyzing a rich dataset featuring a variety of clinical and demographic variables, we endeavored to identify patients at heightened risk for diabetes complications leading to readmissions. Our methodology incorporates an evaluation of numerous machine learning algorithms, emphasizing their predictive accuracy and generalizability to improve patient care. We scrutinized the predictive strength of each model concerning crucial metrics like accuracy, precision, recall, and the area under the curve, underlining the imperative to eliminate false diagnostics in the field. Special attention is given to the use of the light gradient boosting machine classifier among other advanced modeling techniques, which emerge as particularly effective in terms of the Kappa statistic and Matthews correlation coefficient, suggesting robustness in prediction. The paper discusses the implications of diabetes management, underscoring interventions like lifestyle changes and pharmacological treatments to avert long-term complications. Through exploring the intersection of machine learning and health informatics, the study reveals pivotal insights into algorithmic predictions of diabetes readmission. It also emphasizes the necessity for further research and development to fully incorporate machine learning into modern diabetes care to prompt timely interventions and achieve better overall health outcomes. The outcome of this research is a testament to the transformative impact of automated machine learning in the realm of healthcare analytics.

A multimodal dynamic parameterized bilinear factorized framework for remaining useful life prediction under variational data

Remaining useful life (RUL) prediction, as the cornerstone of prognostics and health management, transforms machine maintenance from a passive to a predictive workflow. With the growing availability of sensor types and installations, monitoring data are becoming progressively multimodal, providing comprehensive perspectives on machine health status. In this paper, we propose a dynamic parameterized bilinear factorized framework to tackle some challenges encountered in exploiting multimodal data. To overcome blending and interference among different modalities, a parallel strategy equipped with distillation blocks is adopted to independently process each modality, extracting specialized sensitive features. Subsequently, high-dimensional features are explicitly fused by a bilinear factorized attention module based on their degree of contained degradation-related information. Notably, we design and validate a new loss function called RUL loss, tailored for industrial requirements, considering the asymmetric risk of different prediction errors. Additionally, a dynamic structure distinct from the regular network paradigm is developed, allowing weight parameters to adapt quickly to data variations rather than being fixed during the reasoning process. The superiority of the proposed method is demonstrated using gear life-cycle experimental data. Experimental results show the effectiveness of each component and the overall framework for RUL prediction, showcasing improvements in accuracy and robustness.

Intelligent process control system for predicting operating conditions of slag grinding machines: a data mining approach for improved efficiency and energy savings

This study develops an intelligent data-driven approach for optimising slag grinding systems. Slag grinding exhibits complex nonlinear dynamics that challenge control. The proposed system monitors key operating parameters to assess machine health and automate control adjustments. Operating data are collected, and features linked to health status are identified using data mining techniques. Cluster analysis categorises historical data into healthy/unhealthy modes to build a condition library. Real-time data are then evaluated against this library. A predictive model forecasts future trend. The system was implemented in an industrial slag mill. Results demonstrated reduced vibration and energy use versus manual control. Validation confirmed improved accuracy in predicting mill responses. Significant energy savings were achieved annually through optimised control. The system enhances safety by automating adjustments while minimising costs and environmental impacts. Data-driven strategies overcome the limitations of traditional methods, representing an advance for intelligent management of industrial processes. Benefits were confirmed through rigorous factory implementation and performance monitoring.

Machine Learning and Health Science Research: Tutorial.

Machine learning (ML) has seen impressive growth in health science research due to its capacity for handling complex data to perform a range of tasks, including unsupervised learning, supervised learning, and reinforcement learning. To aid health science researchers in understanding the strengths and limitations of ML and to facilitate its integration into their studies, we present here a guideline for integrating ML into an analysis through a structured framework, covering steps from framing a research question to study design and analysis techniques for specialized data types.

Bayesian Hyper-parameter Optimization in One-dimensional Convolutional Autoencoder for Monitoring Bearing Health Status

This study introduces a framework that integrates hyperparameter optimization techniques for feature representation and abnormality detection in machine health diagnostics. The framework leverages one-dimensional convolutional autoencoders (1D-CAEs) and multilayer perceptron (MLP) as the key components. We propose a systematic approach that begins with the development of the 1DCAE model to encode characteristics into a latent vector. Subsequently, we employ an MLP to effectively detect and identify abnormalities within the encoded data. Our study presents a novel contribution by integrating an adjusted hybrid stopping criterion into the Bayesian optimization framework to improve the efficiency and robustness of anomaly detection. The accuracy and training time of these approaches were examined and evaluated using the bearing degradation data under real-life scenarios. The findings of our study provide strong evidence that the utilization of the Bayesian optimization method, along with the suggested stopping criterion, surpasses conventional tuning techniques in both training time and accuracy. This achievement sets a novel standard for optimizing hyperparameters in the prognostics and health management field. The superiority observed in our study extends beyond the fidelity of the derived models, encompassing the accelerated identification of ideal hyperparameters and holding practical value for real-world applications for early detection of abnormality.

Designing a Deep Autoencoder Neural Network for Detecting Sound Anomalies in Smart Factories Using Unsupervised Learning

Modern world technologies such as the integration of technologies such as the Internet of Things (IoT), cloud computing, and machine learning (ML) enhance the challenges of smart industrial management. Detecting anomalies in predictive maintenance within smart factories, and monitoring machine health to prevent unexpected breakdowns. This research presents an advanced model for designing automatic encoders capable of distinguishing between sounds emitted by machines in industrial environments and identifying faults. The MIMII dataset and advanced feature extraction techniques, such as MFCCs, are adopted as key factors in making the proposed model. The four evaluation measures: accuracy, recall, recall, and F1 score, in addition to the confusion matrix, were also adopted. To evaluate the model's performance. The results confirm the effectiveness and robustness of the proposed deep neural network model designed for autoencoders in the field of artificial audio classification. With a commendable accuracy rate of 93.95% and F1 score of 95.31%,

Intelligent Machinery Fault Diagnosis With Event-Based Camera

Event-based cameras are the emerging bio-inspired technology in vision sensing. Different from the traditional standard cameras, the event-based cameras asynchronously record the brightness change per pixel, and have the great merits of high temporal resolution, low energy consumption, high dynamic range, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">etc</i> . While the event-based cameras have been initially exploited in several common vision-based tasks in the recent years, the investigation on machine condition monitoring problem is quite limited. This paper offers the first attempt in the current literature on exploring the contactless event vision data for machine fault diagnosis. A vibration event representation is proposed to transform the event records into typical data samples, and a deep convolutional neural network model is used for processing the event information. To enhance the model robustness against environmental noisy vision events, an event data augmentation method is proposed to introduce variations of the event patterns. A deep representation clustering method is further proposed to improve the pattern recognition performance with respect to different machine health conditions. Experiments on the event vision-based rotating machine fault diagnosis problem are carried out. It is extensively validated that high fault diagnosis accuracies can be obtained using the vision data from the event-based cameras, which are competitive with the popular accelerometer data. Considering the properties of flexibility, portability and data recognizability, the event-based cameras thus provide a promising new tool for contactless machine health condition monitoring and fault diagnosis.

Multi-Node Feature Learning Network Based on Maximum Spectral Harmonics-to-Noise Ratio Deconvolution for Machine Condition Monitoring

Since the cyclostationarity in vibration signals is the key to judge the rotating machine health state, spectral harmonics-to-interference ratio (SHIR) has been used to construct single node feature learning network (SHIR-based blind deconvolution, BD-SHIR) to realize condition monitoring of machine. However, BD-SHIR still has several obvious limitations, including the need for prior fault information and the tendency to fall into local optimal solutions which will affect its feature learning and state monitoring performance. Therefore, this paper proposes a multi-node feature learning network based on BDSHIR, which does not rely on prior information, multi-node adaptive BDSHIR (MABD-SHIR). It uses each filter in the 1/3 binary tree filter bank to design each initial node (filter) in the learning network correspondingly, and introduces an adaptive fault characteristic frequency (FCF) detection process. Since this initialization can give some filters close to the fault resonant band, iterations starting from these nodes are more likely to detect the precise FCF and converge to the optimal solution. Finally, MABD-SHIR can obtain multiple local optimal solutions, among which the one with the largest SHIR is generally closer to the global optimal solution than the one obtained by BD-SHIR. Simulation and experimental data of defective rotating machinery verify the effectiveness of MABD-SHIR in machinery condition monitoring and fault feature learning. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This paper is motivated by the problems that the rotating machine condition monitoring based on BDSHIR is easy to prematurely fall into local optimal solution and requires prior fault information. These problems can be effectively solved by extending BDSHIR from single node structure to multi-node structure and introducing adaptive period detection technique (A-DPT). The core idea is to start iterations from multiple initial filters and to detect the FCF adaptively during each iteration. Multiple features conforming to filtered signals’ SHIR maximization can be finally learned. Therefore, in this paper, MABD-SHIR is proposed, which is equivalent to a multi-node feature learning network. Since the objective function of BD is usually non-convex, MABD-SHIR can obtain multiple local optimal solutions without requiring prior FCFs. This is more likely to achieve a global optimal solution than BDSHIR, which starts iteration from only one initial filter and ends up with one local optimal solution. In addition, the objective function of MABD-SHIR has lower mathematical complexity, which reduces the difficulty of optimization.

Machine Health in a Click: A Website for Real-Time Machine Condition Monitoring

Machine Health in a Click: A Website for Real-Time Machine Condition Monitoring

An optimized Laplacian of Gaussian filter using improved sparrow search algorithm for bearing fault extraction

Precise detection of fault characteristics in rolling bearings is imperative for machine health management. However, due to the presence of interfering components including noise and periodic components caused by vibration sources, the extraction of weak fault-related information cannot be achieved precisely. In this study, we propose an optimized Laplacian of Gaussian (LoG) filtering technique to handle this issue. The proposed algorithm utilizes the envelope entropy and Gini of square envelope as an objection function to optimize two important parameters, namely standard deviation and filter order of the LoG filter, through an improved sparrow search algorithm (SSA) named adaptive spiral flying SSA. Afterward, the LoG filtering method with the optimal parameters is employed to filter the raw vibration data. Finally, the filtered signal undergoes envelope analysis for fault feature detection. A simulated test and two case studies demonstrate the effectiveness and superiority of the LoG technique.

Vibration-based rotation speed estimation for Industrial IoT

The Industrial Internet of Things (IIoT) uses wireless vibration sensors to predict machine health. In many cases, communication bandwidth and sensor processing resources are limited, and machine kinematics is not available at the sensor level. Consequently, only basic condition indicators, such as overall signal energy or peak-to-peak values, are computed at the sensor and transmitted to the cloud for further analysis. The behavior of these indicators is strongly influenced by machine speed, which is not available in many cases. This paper presents a novel and straightforward method for estimating machine relative speed based on vibrations, enhancing the interpretation of changes in basic condition indicators. The method does not require any prior knowledge about the machine's kinematics or expected speed changes. It was evaluated on simulated signals and wind turbine vibrations, providing a general solution for estimating varying rotating speed when no supplementary information is available. The method estimates rotating speed up to a constant factor in a set of signals (or signal segments) by optimally utilizing peak frequency statistics in the spectrum. The method demonstrates high accuracy under different operating conditions and noise and holds the potential to be used in a variety of IIoT applications.

Next-generation predictive maintenance: leveraging blockchain and dynamic deep learning in a domain-independent system.

The fourth industrial revolution, often referred to as Industry 4.0, has revolutionized the manufacturing sector by integrating emerging technologies such as artificial intelligence (AI), machine and deep learning, Industrial Internet of Things (IIoT), cloud computing, cyber physical systems (CPSs) and cognitive computing, throughout the production life cycle. Predictive maintenance (PdM) emerges as a critical component, utilizing data analytic to track machine health and proactively detect machinery failures. Deep learning (DL), is pivotal in this context, offering superior accuracy in prediction through neural networks' data processing capabilities. However, DL adoption in PdM faces challenges, including continuous model updates and domain dependence. Meanwhile, centralized DL models, prevalent in PdM, pose security risks such as central points of failure and unauthorized access. To address these issues, this study presents an innovative decentralized PdM system integrating DL, blockchain, and decentralized storage based on the InterPlanetary File System (IPFS) for accurately predicting Remaining Useful Lifetime (RUL). DL handles predictive tasks, while blockchain secures data orchestration. Decentralized storage safeguards model metadata and training data for dynamic models. The system features synchronized two DL pipelines for time series data, encompassing prediction and training mechanisms. The detailed material and methods of this research shed light on the system's development and validation processes. Rigorous validation confirms the system's accuracy, performance, and security through an experimental testbed. The results demonstrate the system's dynamic updating and domain independence. Prediction model surpass state-of-the-art models in terms of the root mean squared error (RMSE) score. Blockchain-based scalability performance was tested based on smart contract gas usage, and the analysis shows efficient performance across varying input and output data scales. A comprehensive CIA analysis highlights the system's robust security features, addressing confidentiality, integrity, and availability aspects. The proposed decentralized predictive maintenance (PdM) system, which incorporates deep learning (DL), blockchain technology, and decentralized storage, has the potential to improve predictive accuracy and overcome significant security and scalability obstacles. Consequently, this system holds promising implications for the advancement of predictive maintenance in the context of Industry 4.0.

Comparison-embedded evidence-CNN model for fuzzy assessment of wear severity using multi-dimensional surface images

AbstractWear topography is a significant indicator of tribological behavior for the inspection of machine health conditions. An intelligent in-suit wear assessment method for random topography is here proposed. Three-dimension (3D) topography is employed to address the uncertainties in wear evaluation. Initially, 3D topography reconstruction from a worn surface is accomplished with photometric stereo vision (PSV). Then, the wear features are identified by a contrastive learning-based extraction network (WSFE-Net) including the relative and temporal prior knowledge of wear mechanisms. Furthermore, the typical wear degrees including mild, moderate, and severe are evaluated by a wear severity assessment network (WSA-Net) for the probability and its associated uncertainty based on subjective logic. By integrating the evidence information from 2D and 3D-damage surfaces with Dempster–Shafer (D–S) evidence, the uncertainty of severity assessment results is further reduced. The proposed model could constrain the uncertainty below 0.066 in the wear degree evaluation of a continuous wear experiment, which reflects the high credibility of the evaluation result.

Recent Advances in Flexible Multifunctional Sensors.

Wearable electronics have received extensive attention in human-machine interactions, robotics, and health monitoring. The use of multifunctional sensors that are capable of measuring a variety of mechanical or environmental stimuli can provide new functionalities for wearable electronics. Advancements in material science and system integration technologies have contributed to the development of high-performance flexible multifunctional sensors. This review presents the main approaches, based on functional materials and device structures, to improve sensing parameters, including linearity, detection range, and sensitivity to various stimuli. The details of electrical, biocompatible, and mechanical properties of self-powered sensors and wearable wireless systems are systematically elaborated. Finally, the current challenges and future developmental directions are discussed to offer a guide to fabricate advanced multifunctional sensors.

Developing a hypertension visualization risk prediction system utilizing machine learning and health check-up data

As an important risk factor for many cardiovascular diseases, hypertension requires convenient and reliable methods for prevention and intervention. This study designed a visualization risk prediction system based on Machine Learning and SHAP as an auxiliary tool for personalized health management of hypertension. We used ten Machine Learning algorithms such as random forests and 1617 anonymized health check data to build ten hypertension risk prediction models. The model performance was evaluated through indicators such as accuracy, F1-score, and ROC curve. We used the best-performing model combined with the SHAP algorithm for feature importance analysis and built a visualization risk prediction system on the web page. The LightGMB model exhibited the best predictive performance, and age, alkaline phosphatase, and triglycerides were important features for predicting the risk of hypertension. Users can obtain their risk probability of hypertension and determine the focus of intervention through the visualization system built on the web page. Our research helps doctors and patients to develop personalized prevention and intervention programs for hypertension based on health check data, which has significant clinical and public health significance.