Impending Failures Research Articles

Expected First Occurrence Time of Uncertain Future Events in One-Dimensional Linear Systems

The rapid advancement of machine learning algorithms has significantly enhanced tools for monitoring system health, making data-driven approaches predominant in Prognostics and Health Management (PHM). In contrast, model-based approaches have seen modest progress, as they are often constrained by the need for prior knowledge of specific governing equations, limiting their applicability to a wide range of problems. Recently, rigorous theoretical foundations have been established to extend dynamical systems theory by incorporating prognosis of uncertain events. This article leverages this formal framework to introduce and demonstrate a fundamental mathematical result for one-dimensional linear dynamical systems. The presented theorem offers an exact expression for calculating the expected time at which an event will first occur in the future. Unlike typical thresholds, this event is triggered by a hazard zone, defined as an uncertain event likelihood function over the system's state space. Applications of this theorem can be found in implementing real-time prognostic frameworks, where it is crucial to quickly estimate the magnitude of impending failures. Emphasis is placed on minimizing computational burden to facilitate prognostic decision-making.

Implementation of machine learning methods in predicting failures in electrical submersible pump machines

The Electric Submersible Pump (ESP) is one of the artificial lift tools widely used in oil and gas wells. Approximately 25% of oil and gas wells have been equipped with ESP. The ESP unit excels in lifting fluid rates much larger than most other types of artificial lift. The purpose of this research is to develop a machine learning model that can predict the likelihood of ESP installation failure. By making predictions at the early stage, necessary actions can be prepared to prevent and anticipate these issues. Thus, the reduction in the amount of produced fluid can be minimized. The model in this study uses 8 input parameters, namely, motor Ampere, Frequency, Pump Intake Pressure, Temperature Motor, Output Volt, Pump Discharge Pressure, Input Voltage, and Motor Horse Power (HP). There are also 8 failure classifications for which the model will be created. The initial stage involves collecting and cleaning raw data and processing it to detect anomalies. Once the data is processed, it proceeds to the model formation and model validation stage. The method used for the model is based on the comparison of two methods, namely, decision tree and k-nearest neighbor. The output of the research is a dashboard that can display visualizations of data and the prediction results from the created model. Millions of data points are used to create a database consisting of 77 wells over the past 2 years. The model developed generates several types of failures that can be read. These failures have their own characteristic parameters. The model obtained has an accuracy rate above 90%. The output of this model includes the most likely failure prediction based on the latest input data, thus effectively addressing well problems, warning of impending failures, reducing failures, and assisting in scheduling ESP repairs and maintenance.

Strategies for overcoming data scarcity, imbalance, and feature selection challenges in machine learning models for predictive maintenance

Predictive maintenance harnesses statistical analysis to preemptively identify equipment and system faults, facilitating cost- effective preventive measures. Machine learning algorithms enable comprehensive analysis of historical data, revealing emerging patterns and accurate predictions of impending system failures. Common hurdles in applying ML algorithms to PdM include data scarcity, data imbalance due to few failure instances, and the temporal dependence nature of PdM data. This study proposes an ML-based approach that adapts to these hurdles through the generation of synthetic data, temporal feature extraction, and the creation of failure horizons. The approach employs Generative Adversarial Networks to generate synthetic data and LSTM layers to extract temporal features. ML algorithms trained on the generated data achieved high accuracies: ANN (88.98%), Random Forest (74.15%), Decision Tree (73.82%), KNN (74.02%), and XGBoost (73.93%).

Implementation of Health Monitoring System of Induction Machine Using Computational Intelligence

The induction machine is a critical component in various industrial applications, and its reliable operation is paramount for uninterrupted processes. To ensure the continuous performance and prevent unexpected failures, the implementation of an effective health monitoring system is imperative. In this study, we propose a novel approach to monitor the health of induction machines using computational intelligence techniques. The proposed system integrates advanced computational intelligence algorithms, including machine learning and signal processing methods, to analyze the operational data and identify potential faults or anomalies in the induction machine. Specifically, we employ techniques such as artificial neural networks, support vector machines, and genetic algorithms to process the sensor data acquired from the induction machine. The health monitoring system operates in real-time, continuously analyzing various parameters such as current, voltage, temperature, and vibration patterns to detect deviations from normal operating conditions. By establishing baseline performance profiles and employing pattern recognition algorithms, the system can identify early signs of degradation or impending failures, allowing for proactive maintenance interventions. Moreover, the proposed system offers flexibility and scalability, enabling adaptation to different types of induction machines and operating environments. It facilitates remote monitoring capabilities, enabling maintenance personnel to access real-time health status and diagnostic information from anywhere, facilitating timely decision-making and proactive maintenance strategies.

Evaluation of a Condition Monitoring Algorithm for Early Bearing Fault Detection.

Roller bearings are critical components in various mechanical systems, and the timely detection of potential failures is essential for preventing costly downtimes and avoiding substantial machinery breakdown. This research focuses on finding and verifying a robust method that can detect failures early, without creating false positive failure states. Therefore, this paper introduces a novel algorithm for the early detection of roller bearing failures, particularly tailored to high-precision bearings and automotive test bed systems. The featured method (AFI-Advanced Failure Indicator) utilizes the Fast Fourier Transform (FFT) of wideband accelerometers to calculate the spectral content of vibration signals emitted by roller bearings. By calculating the frequency bands and tracking the movement of these bands within the spectra, the method provides an indicator of the machinery's health, mainly focusing on the early stages of bearing failure. The calculated channel can be used as a trend indicator, enabling the method to identify subtle deviations associated with impending failures. The AFI algorithm incorporates a non-static limit through moving average calculations and volatility analysis methods to determine critical changes in the signal. This thresholding mechanism ensures the algorithm's responsiveness to variations in operating conditions and environmental factors, contributing to its robustness in diverse industrial settings. Further refinement was achieved through an outlier detection filter, which reduces false positives and enhances the algorithm's accuracy in identifying genuine deviations from the normal operational state. To benchmark the developed algorithm, it was compared with three industry-standard algorithms: VRMS calculations per ISO 10813-3, Mean Absolute Value of Extremums (MAVE), and Envelope Frequency Band (EFB). This comparative analysis aimed to evaluate the efficacy of the novel algorithm against the established methods in the field, providing valuable insights into its potential advantages and limitations. In summary, this paper presents an innovative algorithm for the early detection of roller bearing failures, leveraging FFT-based spectral analysis, trend monitoring, adaptive thresholding, and outlier detection. Its ability to confirm the first failure state underscores the algorithm's effectiveness.

A machine learning approach for particle accelerator errant beam prediction using spatial phase deviation

Particle accelerators are extremely complex systems that are expected to operate on high availability. Predicting impending failures only by utilizing data collected from diagnostic equipment already on board can help operators to avoid installing expensive sensors, unscheduled downtime and associated costs. For this purpose we explore the predictive power of Machine Learning algorithms to detect faulty beams prior to the failure. In this study, we propose a Machine Learning approach to model mapping from a pair of sensors located across the accelerator. While the model is trained to represent normal operation, we evaluate the predictive performance on known faulty beam pulses. We also investigate the model performance on unseen data through k-fold cross-validation. Then we recap the analysis with a neural architecture search and hyperparameter optimization study to fine tune our initial model. This paper will also introduce a sustainable framework that can standardize Machine Learning workflow applied to particle accelerators.

Reliable Ship Emergency Power Source: A Monte Carlo Simulation Approach to Optimize Remaining Capacity Measurement Frequency for Lead-Acid Battery Maintenance

<div>The development of predictive maintenance has become one of the most important drivers of innovation, not only in the maritime industry. The proliferation of on-board and remote sensing and diagnostic systems is creating many new opportunities to reduce maintenance costs and increase operational stability. By predicting impending system faults and failures, proactive maintenance can be initiated to prevent loss of seaworthiness or operability. The motivation of this study is to optimize predictive maintenance in the maritime industry by determining the minimum useful remaining lead-acid battery capacity measurement frequency required to achieve cost-efficiency and desired prognostic performance in a remaining battery capacity indication system. The research seeks to balance operational stability and cost-effectiveness, providing valuable insight into the practical considerations and potential benefits of predictive maintenance. The methodology employed in this study includes outlining the theoretical development of a fully automated condition monitoring system and describing data cleansing steps to account for environmental effects on system performance. A Monte Carlo simulation is used to evaluate the sensitivity of the remaining useful life prediction to varying measurement frequencies, prediction models, and parameter settings, leading to an estimate of the optimal measurement frequency for the system. The results show that a certain minimum measurement frequency is required to achieve the target prediction accuracy while balancing cost-efficiency and operational stability. Reliable failure prediction with negligible changes in prognostic accuracy can be achieved by performing useful remaining lead-acid battery capacity measurements twice a day or every 5 ship voyage cycles with the underlying utilization.</div>

Deep-Learning-Enabled Predictive Maintenance in Industrial Internet of Things: Methods, Applications, and Challenges

With the great evolution of human society from the information era to the smart automation era, intelligent production and maintenance have become the core orientation of Industry 4.0. Data-driven intelligent predictive maintenance (IPdM) offers potential solutions to satisfy the urgent demand of predicting impending failures and mitigating unplanned downtime of industrial machinery, and is envisioned as a critical enabler for future industrial Internet of Things. In this article, we provide a comprehensive survey of deep learning (DL)-based IPdM for the researchers and practitioners who focused on the promotion of fault diagnosis/prognosis. First of all, four mainstream algorithms of DL are reviewed around the specific IPdM applications of recent five years. Some prominent discussions and insights on the DL-based IPdM are given here in terms of data characteristics and model performance. Particularly, as an important contribution, a new industrial dataset of dual-bearings rotating machinery is published for opening research to address one of the main challenges for the development of DL-based IPdM, i.e., the shortage of high quality industrial datasets. Compared with current datasets, our dataset is multisensory, which is closer to reality and more suitable for building advanced DL-based IPdM models. We also give a DL-based case study to demonstrate how to build a predictive model for maintenance decisions based on this dataset. Finally, potential research trends and challenges are also discussed for enhancing the intelligence of IPdM in the industrial scenarios.

A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data

Predictive Maintenance (PdM) methods aim to facilitate the scheduling of maintenance work before equipment failure. In this context, detecting early faults in automated teller machines (ATMs) has become increasingly important since these machines are susceptible to various types of unpredictable failures. ATMs track execution status by generating massive event-log data that collect system messages unrelated to the failure event. Predicting machine failure based on event logs poses additional challenges, mainly in extracting features that might represent sequences of events indicating impending failures. Accordingly, feature learning approaches are currently being used in PdM, where informative features are learned automatically from minimally processed sensor data. However, a gap remains to be seen on how these approaches can be exploited for deriving relevant features from event-log-based data. To fill this gap, we present a predictive model based on a convolutional kernel (MiniROCKET and HYDRA) to extract features from the original event-log data and a linear classifier to classify the sample based on the learned features. The proposed methodology is applied to a significant real-world collected dataset. Experimental results demonstrated how one of the proposed convolutional kernels (i.e. HYDRA) exhibited the best classification performance (accuracy of 0.759 and AUC of 0.693). In addition, statistical analysis revealed that the HYDRA and MiniROCKET models significantly overcome one of the established state-of-the-art approaches in time series classification (InceptionTime), and three non-temporal ML methods from the literature. The predictive model was integrated into a container-based decision support system to support operators in the timely maintenance of ATMs.

Semi-universal geo-crack detection by machine learning

Introduction: Cracks are a key feature that determines the structural integrity of rocks, and their angular distribution can be used to determine the local or regional stress patterns. The temporal growth of cracks can be monitored in order to predict impending failures of materials or structures such as a weakened dam. Thus, cracks and their spatial-temporal distributions should be automatically monitored for assessing their structural integrity, the associated stress patterns and their potential for failure.Method: We show that the U-Net convolutional neural network, semantic segmentation and transfer learning can be used to accurately detect cracks in drone photos of sedimentary massifs. In this case, the crack distributions are used to assess the safest areas for tunnel excavation. Compared to the coarse performance of ridge detection, the U-Net accuracy in identifying cracks in images can be as high as 98% when evaluated against human identification, which is sufficient for assessing the general crack properties of the rock faces for the engineering project.Result: Based on approximately 100 h of manual cracks labeling in 127 drone photos and 20 h of network training, the U-Net was able to successfully detect cracks in 23,845 high-resolution photographs in less than 22 h using two Nvidia V100 GPUs. Meanwhile, the network was able to detect more than 80% of the observable cracks of a volcanic outcrop in Idaho without additional training. With a modest amount of extra labeling on photos of the volcanic outcrop and transfer training, we found that the accuracy significantly improved. The surprising outcome of this research is that the U-Net crack detector laboriously trained on photos of sedimentary rocks can also be effectively applied to photos of volcanic rock faces. This can be important for real-time assessment of geological hazards and lithology information for dam inspection and planetary exploration by autonomous vehicles. For another application, we accurately detected fractures and faults with a scale of tens of kilometers from Martian photographs.Conclusions: In summary, our methodology of using CNN with transfer training suggests that it can be used as a semi-universal detector of cracks in across a range of diverse geological settings.

Successful emergency separation of premature omphalopagus conjoined twins: an anaesthetic experience

Anaesthesia for early emergency separation of premature conjoined twins is extremely rare as surgery generally done electively between 2 to 4 months of age. However, urgent separation may be needed due to life-threatening complications. We report a case of successful early separation of premature omphalopagus twins at 36 weeks of gestational age with a combined weight of 2.7 kg. To the best of our knowledge, this was the lowest weight yet reported of successful surgical separation in Malaysia. Early separation was indicated as extrauterine twin-to-twin transfusion with unbalanced blood shunting through the porto-systemic anastomoses within the shared liver parenchyma potentially risked impending life-threatening organ failures. Anaesthesia for the separation of premature conjoined twins in the emergency setting requires extensive multidisciplinary discussion and planning. Factors predicting difficult anaesthesia in this case were the twins’ size and age as well as duration of anaesthesia. Two separate anaesthetic teams were required with all team members well versed in the operative workflow. Simultaneous airway management, prevention of hypothermia, and vigilant haemodynamic monitoring are key to successful anaesthesia in premature conjoined twin separation.

Use of Data-Driven Approaches for Defect Classification in Stator Winding Insulation

Partial discharges (PD) in the high voltage insulation systems are both a symptom and cause of terminal and impending failures. The use of data-driven methods based on PD measurements will enable predictive strategies to replace traditional maintenance strategies. This paper employs machine learningbased classification models to identify and characterizePD signals originating from lab-made artificial defects in epoxy-mica material samples. Three different PD sources were studied: surface discharges in air, corona discharges, and discharges caused by internal cavities/delaminations. To generate high-quality datasets for the training, validation, and testing of classification models, Phase-Resolved PD (PRPD) data for each test object was obtained at room temperature under 50 Hz AC excitation at 10 % above the PD inception voltage (PDIV) of each sample. Relevant statistical and deterministic features were extracted for each observation and were labeled based on the defect type (supervised learning). Finally, the trained and validated ML models were used to identify PD sources in the service-aged stator winding insulation. Support vector machines (SVM), ensemble, and k-nearest neighbor (kNN) algorithms achieved significantly high accuracy (≥ 95 %) of defect identification.

Bi-level bayesian control scheme for fault detection under partial observations

Due to limitations of sensor techniques and system structures, the internal states of most complex devices are unobservable. Only partial observations can be monitored at discrete time epochs, which poses great challenges for fault detection and prediction. To facilitate the development of advanced approaches, this paper proposes a bi-level Bayesian control scheme for detecting faults of complex devices under partial observations. The deterioration of an operational system is modeled as a hidden three-state multivariate Markov process, in which the parameter estimates are obtained by an expectation maximization (EM) algorithm. Based on the multivariate Markov process, a bi-level Bayesian control scheme is presented to monitor the Bayesian posterior probability of the system in a warning state and uses two levels of sampling frequency to dynamically detect impending failures. The decision variables of the bi-level fault detection scheme are optimized and solved in a semi-Markov decision process (SMDP) framework. The particularity of this work is that two phases of system deterioration are considered into detection scheme design to avoid unnecessary false alarms. Using partially observed multivariate data from discrete monitoring, the proposed approach is validated to be capable of detecting the early fault occurrences. By comparing with other advanced methods, the superiority of the proposed model is demonstrated.

Imminent threat of rock-ice avalanches in High Mountain Asia

Upsurge of glacier-related hazards in High Mountain Asia (HMA) has been evident in recent years due to global warming. While many glacial-related hazards are instantaneous, some large landslides were preceded by slow gravitational deformation, which can be predicted to evade catastrophes. Here, we present robust evidence of historical deformation in 2021 Chamoli rock-ice avalanche of Himalaya using space imaging techniques. Multi-temporal satellite data provide evidence of a precursor event in 2016 and expansion of a linear fracture along joint planes, indicating 2021 rock-ice avalanche is a retrogressive wedge failure. The deformation history shows that the fracture propagated at a velocity of ~0.07 m day−1 until September 2020, and with an accelerated velocity (~0.14 m day−1 on average) lately. Analysis of recent similar cases in HMA supported our inference on global warming-induced glacier retreat and thermomechanical effects in enhancing the weakening of fractured rock masses in tectonically active mountain belts. Recent advances in Earth observation and seismic monitoring system can offer clues to the location and timing of impending catastrophic failures in high mountain regions.

Pinpointing Early Signs of Impending Slope Failures From Space

AbstractA promised potential of spaceborne interferometric synthetic aperture radars (InSAR) is a capability for regularly monitoring ground deformation with millimeter accuracy, for timely forecasting of impending natural hazards such as landslides. The main limitation in InSAR being actually capable of unleashing this potential for hazard prediction is that key precursory ground displacements are, in the majority of cases, a very small subset of the entire big data set provided by the method over large regions. Consequently, pinpointing a single impending failure may become very difficult or impossible. We develop a data‐driven framework that can handle such imbalanced spatiotemporal data based on the concept of outlying aspects mining, to find a subset of features out of a collection of potential features, which best distinguishes the landslide source area from the others. We show that the identified feature subspace can be used to find anomalous areas across multiple spatial scales, such that Sentinel‐1 satellite monitoring points which persistently lie in these areas can accurately detect the location of the Xinmo landslide (China) almost 1 year in advance—without false alarms. In a second case study, we identify the area affected by rockfalls on Stromboli volcano, a task that is generally infeasible with traditional methods applied to InSAR data. With continuing improvements in the spatial and temporal resolution from the new generation of satellites, such as Sentinel‐1, this approach opens the door to reliable and early prediction of failure over a broad range of slope instabilities.

Adaptive monitoring scheme of stochastically failing systems under hidden degradation processes

Limited by the installation of sensors and the structure of systems, only external observation information can generate an association with the internal health status of a system. In addition, modern industrial systems are often accompanied by hidden degradation processes, which present great challenges when monitoring and alerting impending failures. To address this issue, this paper proposes an adaptive monitoring scheme for predicting faults of systems with hidden degradation processes. A multivariate, continuous-time hidden Markov model with three states (hidden healthy state, hidden unhealthy state, and observable failure state) is established to describe the degradation process of the system. An expectation maximization (EM) algorithm is presented to estimate the parameters of the stochastic model. On the basis of the hidden degradation model, an adaptive Bayesian control scheme is developed for monitoring the potential risk of the system with two switchable sampling intervals. We calculate optimal decision variables of the control scheme to achieve the minimum expected average cost by a policy iteration algorithm under a semi-Markov decision process (SMDP). The performance and characteristics of the proposed scheme are demonstrated via real oil measurement data obtained from mechanical generators. A comparison case illustrates the superiority of the prediction performance of our monitoring scheme.

Nonintrusive Measurements for Detecting Progressive Equipment Faults

The gradual environmental degradation of materials can lead to corrosion and mechanical failures of electrical loads. Vibration and corrosion both contribute to the failure of electrical connections and conductors, and electrical arcs destroy materials in ways that enhance the opportunities for further damage. These failures lead to loss of system capability and the possibility of electrical fires. The signatures of these impending electrical failures, which may occur over weeks or months, can be detected in many systems through nonintrusive electrical monitoring. This paper demonstrates techniques for decoding the type of electrical signatures that lead to progressive failure. Example applications of these techniques are illustrated with observations of the degradation and failure of copper-sheathed jacket water heaters aboard a United States Coast Guard vessel. Electrical, material, and vibration analysis are presented to demonstrate the basis for diagnostic electrical signatures.

Similarity-Based Echo State Network for Remaining Useful Life Prediction

The health state of aero engines will affect economic benefits, availability, and safety of aircrafts. Fault prognosis is considered to be one of useful methods to evaluate the health state of aero engines and avoid unscheduled down time. Prognosis is the ability to forecast the evolution of engine deterioration. The two major objectives of prognosis are: forecasting the impending failures and predicting the engines’ remaining useful life (RUL). This paper proposes a method that combines echo state network (ESN) with similarity-based methods for RUL prediction. First, principal component analysis (PCA) is used to pre-process data to obtain degradation trajectories, which handles the fusion problem of multiple data. Second, in order to use historical information accurately, the Euclidean distance is applied to calculate the similarity among different degradation trajectories. Third, the trajectory with the highest similarity is used in the ESN training process. Then, a numerical simulation on the NASA C-MAPSS dataset verifies the accuracy of the method. Finally, the comparison with other methods shows the superiority of the method.

Theoretical research of Stand Error of Mean Coupling Markov chain models for early warning of landslide displacement rates: A case study of human-induced landslide in Inner Mongolia (Northern China)

Slope failures are always frequent and important issues in surface mining, where a single calamity can disarrange mining schemes and endanger workers’ lives. Simultaneously, the design and verification of “landslide early warning systems” in open-pit mines is a pivotal tool for mitigating risks. Notwithstanding, a prevalent model for heralding slope impending failures has not been available yet. The purpose of this article is to present a novel approach based on mean standard deviation (MSD) coupling weighted Markov chain (WMC) model to evaluate and perform the evolution that can detect onset-of-acceleration of landslides analysis procedure based on the indicator of real-time ground-based radar low-frequency measurements (e.g. displacement rate). The proposed method has been applied and tested in a open-pit coal mine situated in Inner Mongolia, China. An explanatory example of back analysis of a 4-month continuous surface monitoring dataset were regarded as a random process, and a binary classifier (i.e. steady-state and unsteady-state) was constructed based on the characteristics of mean standard deviation (MSD). Afterwards, properties of none aftereffect and bouncing allocation on the efficiency of Markov chains are studied. In order to modify and optimize the model, three indexes are integrated to verify the accuracy of the model and proof the predicting results: (1) precaution sensitiveness; (2) proper rate; (3) consensus rate. The consequences show that when the size of training samples was 20 days, the early warning accuracy reached 93%. When the model was performed seven days before the landslide occurrence, the true positive rate of the model was 84%, thus indicating the early warning for the landslide was timely. The integration of the proposed model and on-site monitoring data provides a useful tool for early warning of landslide displacement rates, and opens new perspectives on predicting slope instabilities.

Age and Condition-Based Preventive Replacement Timing for Periodic Aircraft Maintenance Checks

By anticipating impending failures and addressing those with preventive replacements, Condition-Based Maintenance (CBM) can provide several economic benefits to aircraft operators. While Prognostics and Health Monitoring (PHM) methods are widely available, scheduling of tasks originating from those methods is a relatively new challenge. To avoid incurring extra cost and ground-time, these maintenance tasks are typically scheduled during already existing (conventional) maintenance slots such as periodic checks. Following this strategy, an aircraft component would have to be replaced if a failure precursor is detected that is expected to result in failure between the next two checks. The decision following from that detection depends on the chosen decision threshold of the prognostic or diagnostic model, which can be represented by a point on the Receiver Operating Characteristic curve. Selecting the optimal operating point for each maintenance check is challenging, as it depends on the age-dependent reliability of the component, the performance of the prognostic (diagnostic) model, the interval of the periodic check and the cost of (corrective and preventive) maintenance. This paper presents an innovative method for selecting optimal operating points for all periodic checks throughout the lifetime of an aircraft component. This is done by means of a numerical optimization model that finds operating points that minimize the component’s total maintenance cost per flight hour. A case study on a compressor of a wide-body aircraft is presented, which shows that by using this method, additional economic value from existing PHM can be realized without the need for additional investments.