Early Failure Prediction Research Articles

Progressive degradation behavior and mechanism of lithium-ion batteries subjected to minor deformation damage

Minor deformation damage poses a concealed threat to battery performance and safety. This study delves into the progressive degradation behavior and mechanisms of lithium-ion batteries under minor deformation damage induced by out-of-plane compression. The effects of varying initial state of charge and loading speed on battery degradation are also analyzed. It has been observed that a deformation damage degree as low as 3.1 % can cause a significant transient capacity reduction up to 6.3 %. More transient capacity reduction can occur at higher initial state of charge and loading speed. Additionally, the investigation reveals that the capacity decay rate between the normal cell and cells experiencing deformation damage degree of <4.7 % is almost similar, with a maximum difference of merely 1.887 mAh/cycle and a minimum difference of 0.001 mAh/cycle, both observed over 3000 cycles. Non-destructive and destructive analysis methods are used to investigate degradation behaviors of the cells experiencing minor deformation dagame. It is found that such progressive degradation is primarily driven by the loss of active lithium ions, followed by the loss of cathode and anode active materials. It is anticipated that the findings will offer theoretical underpinnings for advancements in battery damage assessment, early failure prediction, and facilitation of loss adjustment and compensation by insurance agencies.

Industrial IoT Sensor Data Federation in an Enterprise for an Early Failure Prediction

Industrial IoT Sensor Data Federation in an Enterprise for an Early Failure Prediction

Quantitative prediction of wear failure based on dissipative methodology under dry friction

The evaluation and prediction on failure of friction pair are important for improving equipment reliability and life span. In engineering practice, researchers often determine the severe-wear stage by monitoring changes in wear and the friction coefficient. However, existing criteria for evaluating the wear process are nebulous, leading to the ambiguous relationship between the measured parameters and the wear process which cannot be precisely quantified. In this study, based on dissipation theory, an exploration into the system entropy production rate variations throughout the wear process has been conducted. This investigation reveals that entropy production rate variations at different wear stages, providing a basis for segmenting the wear process. Through experimental methods, a clear transitional stage with substantial entropy production rate variations between the mild-wear and severe-wear stages has been discovered. These findings gave a novel method for predicting wear failure. The efficacy of this proposed approach has been validated by experiment, thus laying a theoretical groundwork for early failure prediction of critical components principally prone to wear failure. Furthermore, this work provides the foundation for the construction of visual data system in friction-wear digital twin systems.

Bayesian change point prediction for downhole drilling pressures with hidden Markov models

AbstractIn the drilling of oil wells, the need to accurately detect downhole formation pressure transitions has long been established as critical for safety and economics. In this article, we examine the application of Hidden Markov Models (HMMs) to oilwell drilling processes with a focus on the real time evolution of downhole formation pressures in its partially observed state. The downhole drilling pressure system can be viewed as a nonlinear, non‐degrading stochastic process whose optimum performance is in a region in its warning state prior to random failure in time. The differential pressure system is modeled as a hidden 3 state continuous time Markov process. States 0 and 1 are not observable and represent the normally pressured (initiating ) and abnormally pressured or warning (reducing ) states respectively. State 2 is the observable failure state (from negative and loss of well control). The signal process of the evolution of differential pressure is identified in the changes in the observable rate of penetration (ROP) encoded in drilling performance data. The state and observation parameters of the HMM are estimated using the Expectation Maximization (EM) algorithm and we show, for a univariate system with a depth dependent time relationship, that the model parameter updates of the EM algorithm equation have explicit solutions. A Bayesian inference model, to determine the safety threshold of the system and early failure prediction at each sampling epoch, is thereafter proposed. The application of our stochastic model of the dynamic evolution of downhole pressures in operational time is illustrated with a hindcast case example. The analysis showed strong early indication of probable failure in real time and was validated in the field post drilling system failure that resulted in significant recovery costs. The potential to predict the differential pressure state transitions ahead of the bit represents a capability not currently available in the industry. This opens up significant opportunity for value creation from optimizing drilling operations to deliver substantial savings in well construction costs.

MRI radiomics may predict early tumor recurrence in patients with sinonasal squamous cell carcinoma.

Sinonasal squamous cell carcinoma (SCC) follows a poor prognosis with high tendency for local recurrence. We aimed to evaluate whether MRI radiomics can predict early local failure in sinonasal SCC. Sixty-eight consecutive patients with node-negative sinonasal SCC (January 2005-December 2020) were enrolled, allocated to the training (n = 47) and test sets (n = 21). Early local failure, which occurred within 12 months of completion of initial treatment, was the primary endpoint. For clinical features (age, location, treatment modality, and clinical T stage), binary logistic regression analysis was performed. For 186 extracted radiomic features, different feature selections and classifiers were combined to create two prediction models: (1) a pure radiomics model; and (2) a combined model with clinical features and radiomics. The areas under the receiver operating characteristic curves (AUCs) were calculated and compared using DeLong's method. Early local failure occurred in 38.3% (18/47) and 23.8% (5/21) in the training and test sets, respectively. We identified several radiomic features which were strongly associated with early local failure. In the test set, both the best-performing radiomics model and the combined model (clinical + radiomic features) yielded higher AUCs compared to the clinical model (AUC, 0.838 vs. 0.438, p= 0.020; 0.850 vs. 0.438, p= 0.016, respectively). The performances of the best-performing radiomics model and the combined model did not differ significantly (AUC, 0.838 vs. 0.850, p= 0.904). MRI radiomics integrated with a machine learning classifier may predict early local failure in patients with sinonasal SCC. MRI radiomics intergrated with machine learning classifiers may predict early local failure in sinonasal squamous cell carcinomas more accurately than the clinical model. • A subset of radiomic features which showed significant association with early local failure in patients with sinonasal squamous cell carcinomas was identified. • MRI radiomics integrated with machine learning classifiers can predict early local failure with high accuracy, which was validated in the test set (area under the curve = 0.838). • The combined clinical and radiomics model yielded superior performance for early local failure prediction compared to that of the radiomics (area under the curve 0.850 vs. 0.838 in the test set), without a statistically significant difference.

Predictive Maintenance in Rotating Machinery Using Deep Learning Techniques

The integration of Artificial Intelligence (AI) in mechanical systems has revolutionized the way we perceive maintenance and failure detection. This research paper introduces a novel methodology to predict and assess machinery failures in rotating equipment, such as turbines and motors, using Deep Learning techniques while traditional methods often rely on rule-based algorithms and basic statistical models, our approach leverages Convolutional Neural Networks (CNN) to process vibration and acoustic signals, and Long Short-Term Memory networks (LSTM) to analyze time-series data. By assimilating these techniques, we aim to achieve early and accurate failure prediction. A dataset consisting of 10,000 hours of machinery operation was collected from an industrial partner. Features like vibration amplitude, operating temperature, and acoustic signatures were captured in varying operational conditions. Data augmentation techniques were also employed to amplify rare but crucial failure signatures. Our model achieved a prediction accuracy of 98.5%, outperforming traditional methods by 15%. Furthermore, the implemented model showed high robustness against noisy data, indicating its potential for real-world applications. This research not only presents a powerful tool for mechanical engineers but also underscores the immense potential of integrating AI in traditionally non-digital domains. The methodology can be expanded to other mechanical systems, paving the way for a new era of predictive maintenance.

Self-optimised cost-sensitive classifiers for early field failure prediction in storage systems

Data storage systems such as disk arrays go through rigorous testing in the production phase, however, a few of these DAs fail in the field and are returned back to the manufacturer. Although the failure appears in relatively a small percentage of the manufactured DAs, it results in a significant loss of data, time and money. This paper is motivated by the hypothesis that many of these failures could be predicted at the testing stage through data mining and machine learning. Field failure is modelled as a classification problem, however, as in many real-world problems, the problem suffers from significant class imbalance. Several approaches have been proposed that attempt to improve the performance of imbalanced classification by either modifying the dataset (resampling), or assigning classification costs to the classes’ cost matrix. These methods have been shown to improve performance, but they come with many parameters that need to be set, something that usually requires a lengthy exhaustive search, especially on problems with several classes. This paper presents a new scalable genetic algorithm approach for automating the design of the cost matrix CM along with the algorithm parameters. The proposed algorithms are tested on a real-world manufacturing dataset from Seagate disk arrays; the target is to predict from the devices’ testing data those that are likely to fail in the field. To demonstrate its performance, the proposed approach evaluated on a number of standard datasets and compared with other state-of-the-art methods.

Buildability analysis on effect of structural design in 3D concrete printing (3DCP): An experimental and numerical study

The importance of numerical modeling and simulation approaches increases with increased scale as it allows for early results predictions. This study evaluated the buildability of novel construction and demolition waste (CDW)-based geopolymer materials for 3-dimensional printing (3DP) structures for the built environment. The feasibility of the developed numerical model for 3DP of built environment structures was also evaluated. This study reported the fresh-state properties of geopolymer-based materials, and based on these results, numerical modeling for 3DP of geopolymer material was performed. Three cylindrical structures with diameters of 300 mm, 450 mm, and 600 mm were printed, with the same printing conditions, resulting in buildability of 352.5 mm, 322.5 mm, and 277.5 mm (total height constructed before the collapse), respectively. In the next stage of the study, a comparative analysis was performed to assess the compatibility of numerical models for CDW-based geopolymer materials. In the numerical results, three structures showed early failure prediction from experimental results at 31.91%, 29.10%, and 29.73%, respectively. The geopolymer materials developed were suitable for 3DP, and close agreement in results and repeatability were observed. The numerical model provided a reliable failure prediction for different structures. Hence, the model can be utilized to evaluate the buildability of the structures before the actual 3DP, which can significantly reduce the time and cost of the project.

Predictive Maintenance in IoT: Early Fault Detection and Failure Prediction in Industrial Equipment

The Industrial Internet of Things (IoT) has ushered in a new era of predictive maintenance, revolutionizing the way industries manage and maintain their critical equipment. This paper presents a comprehensive exploration of predictive maintenance strategies, with a primary focus on early fault detection and classification in industrial equipment. We introduce the Triplet Fault Injection Algorithm, capable of injecting three distinct fault types—spike, bias, and stuck—into sensor data for realistic and rigorous testing. Leveraging this algorithm, we employ the powerful Extreme Gradient Boosting (XGBoost) machine learning approach to detect and classify these faults. Our experimental results showcase the superiority of XGBoost over baseline machine learning methods, across various data types commonly found in industrial equipment. The consistent higher accuracy and F1 scores obtained with XGBoost underscore its effectiveness in minimizing false alarms and enhancing the reliability of early fault detection. Moreover, we discuss the transformative role of IoT in predictive maintenance, highlighting its potential to optimize equipment performance and reduce downtime in the industry 4.0 landscape. This paper contributes valuable insights and empirical evidence to the domain of predictive maintenance in IoT-enabled industries, emphasizing the significance of early fault detection for efficient and cost-effective maintenance practices.

Clinical and Radiomics-Based Deep Learning Predictive Model for Early Treatment Failure after Neoadjuvant Radiochemotherapy for Rectal Cancer

<h3>Purpose/Objective(s)</h3> As shown in RAPIDO and PRODIGE-23 trials, total neoadjuvant therapy for rectal cancer increases the complete response rate, however, the overall survival benefit remains uncertain. Although postponing surgery assures more time for tumor downstaging, it also increases surgical complications and can worsen prognosis in non-responding tumors. We aimed to develop a clinical and radiomics-based model for early treatment failure (ETF) prediction after neoadjuvant radiochemotherapy (RCHTx) for rectal cancer. <h3>Materials/Methods</h3> We conducted a prospective cohort analysis of 92 consecutive and eligible patients with rectal adenocarcinoma treated with neoadjuvant radiochemotherapy in the years 2018-2020, and followed until February 2022. We defined ETF as postoperative poor pathological response and no MRI-confirmed downstaging; or unresectability after neoadjuvant treatment; or cancer-related or treatment-related death within 18 months since treatment initiation. Collected clinical data included physical examination, grading, staging (before and after RCHTx), biomarker levels, RCHTx prescription details, and pre-operative MRI tumor and nodal characteristics. Radiomic features were extracted from GTV delineated on axial MRI T2 TSE sequences. To develop a predictive model, we followed OmicSelector-based (https://biostat.umed.pl/OmicSelector) feature selection, logistic regression (LR), and artificial deep neural network model development procedure. After splitting into training (60%), testing (20%), and validation (20%) datasets, the final feature set was selected based on the highest average logistic regression accuracy on all sets, while all patients with MRI performed outside our institution were included in the testing set (compensating possible batch effect). <h3>Results</h3> ETF was experienced by 24% of patients and was associated with significantly shorter overall survival (median 12.1 months [95%CI:8.5-16.5] vs. not reached, p<0.0001). None of the clinical factors was significantly associated with ETF (adjusted p value>0.05; AUC ROC range 0.39-0.64). LR model based on clinical factors presented overfitting (accuracy 82% on the training set, 56.2% on testing and 68.8% on validation set). The best feature set included 14 selected radiomic features, patient's weight, T-stage, grading, mesorectal fascia and nodal involvement. LR model based on those features achieved perfect accuracy on training and testing sets and 81.3% accuracy on the validation set. Deep neural network model utilizing selected clinical and radiomic features achieved 87.5% accuracy (80% sensitivity, 91% specificity) on testing and 93.8% accuracy (100% sensitivity, 91% specificity) on validation set. <h3>Conclusion</h3> Combined clinical and radiomics-based deep learning model could predict ETF after RCHTx for rectal cancer with at least 80% sensitivity and 90% specificity. Prediction of ETF after RCHTx could select high-risk patients who may require non-standard treatment.

Early Detection of Main Bearing Damage in Wind Turbines

According to the European Wind Energy Academy (EAWE), the wind industry has recognized that main bearing failures are a major concern in order to increase the reliability and availability of wind turbines. This is due to the high replacement cost of major repairs and the long downtime associated with main bearing failures. As a result, predicting main bearing failure has become an economically important problem as well as a technological difficulty. This paper presents a data-driven technique based on a closed recurrent unit (GRU) neural network for early failure prediction (months in advance). The main contributions of this work are: (i) The prediction is made exclusively using SCADA (Supervision Control and Data Acquisition) data already present in all industrial wind turbines. Therefore, there is no need to add additional sensors intended for a specific use. (ii) Since the proposed approach only requires healthy data, it can be used in any wind farm even if it has not recorded faulty data. (iii) The suggested algorithm operates under a variety of operational and environmental circumstances. (iv) The methodology is validated in two real inproduction wind turbines. production.

A model for early failure prediction of blood pressure measurement devices in a stepped validation approach.

Blood pressure monitoring (BPM) devices have to be validated according to strict international validation protocols. Each protocol requests a specific number of participants to be included. All protocols use vast amounts of resources, as three people have to be present for every measurement, making trials costly, especially when the manufacturer has no intention to execute a validation study, reflected in the low share of validated in the commercially available BPM devices. The aim of our study was to develop criteria, which could detect low accuracy devices that could not pass a validation protocol early in the course of the validation process. The 2010 European Society of Hypertension International Protocol (ESH‐IP) and the Universal Standard for Validation of BPM devices (AAMI/ESH/ISO) were scrutinized for criteria which can be used for preclusion of passing. Based on this, we developed a fail model. We found that a BPM device cannot pass the ESH‐IP protocol, if there are ≥27, 13, or 4 single measurements differing more than 5, 10, or 15 mmHg, respectively, from the reference. For the AAMI/ESH/ISO protocol, we developed a model, which calculates best‐case standard deviations (SDs) to detect SDs which would prevent the passing of the protocol before its completion, making a stepwise validation process possible. In conclusion, we found that our model is able to predict failure of low‐accuracy BPM devices early during a validation protocol if used in a stepwise‐approach. This can be useful to keep costs of validation studies low and to enable investigator‐initiated trials.

Ultrasonic detection of white etching defect based on convolution neural network

&lt;sec&gt;Unlike classical defects formed by rolling contact fatigue, white etching defect (WED) including white etching area and white etching crack will cause surface to spall in the early stage and the service life to shorten seriously. Located in the subsurface of bearings, the tiny size WED is difficult to detect by conventional ultrasonic methods. The root cause of WED generation remains unclear. It is time consuming and expensive to prepare samples during the evolution of such defects. For characterizing the WED at early stage, five evolving states concerning the existing microscopic information are established in this paper. The immersion ultrasonic inspection process is simulated based on &lt;i&gt;k&lt;/i&gt;-space pseudo spectrum method.&lt;/sec&gt;&lt;sec&gt;For the later evolutionary stage with crack, the bearing can be simplified into a homogeneous three-layer model by ignoring the internal grain structure. The crack depth is obtained by using the ultrasonic reflection coefficient amplitude spectrum (URCAS), with an error of 1.5%. For other states without crack, the spectrum characteristic is no longer evident with slight acoustic impedance difference between layers. The polycrystalline structure on a microscale is thus realized based on Voronoi diagram, from which the grain induced backscattering can be used to amplify the microstructure variations at different stages. The backscattering signal is influenced by the grain size and detection frequency from the simulation. Since a direct comparison of backscattering information among evolutionary stages is difficult, the five different evolutionary stages of WED are recognized with the help of deep learning. The received waveform is transformed into a time-frequency map by short-time Fourier transform. Based on RESNET network structure, the results show that the train accuracy and validation accuracy reach 92% and 97% respectively. This study provides a sound way to characterize WED, which is conducive to early failure prediction and residual life evaluation.&lt;/sec&gt;

Control of damage‐sensitive features for early failure prediction of wind turbine blades

Control of damage‐sensitive features for early failure prediction of wind turbine blades

Data-driven method for real-time prediction and uncertainty quantification of fatigue failure under stochastic loading using artificial neural networks and Gaussian process regression

Various engineering systems such as naval and aerial vehicles, offshore structures, and mechanical components of motorized systems, are exposed to fatigue failures due to stochastic loadings. Methods for early failure prediction are essential for engineering, military, and civil applications. In addition to the prediction of time to failure (TtF), uncertainty quantification (UQ) is of major importance for real-time decision-making purposes. Usually, time domain or frequency domain methods are used for fatigue prediction, such as rainflow counting and Miner’s rule or Dirlik’s method. However, those methods suffer from over-simplistic modeling and inaccurate failure predictions under stochastic loadings. During the last years, several data-driven models were suggested for offline fatigue failure. However, most of them are not capable of both accurate real-time fatigue prediction and UQ. In the current work, a probabilistic data-driven model is introduced. A hybrid architecture of a fully connected artificial neural network (FC-ANN) and Gaussian process regression (GPR) is proposed to ensure enhanced predictive abilities and simultaneous UQ of the predicted TtF. The real-time prediction and UQ performances of the suggested model are validated using both synthetic and experimental data. This novel hybrid method is fully data-driven and extends the forecasting capabilities of existing time-domain and machine learning-based methods for fatigue prediction. It paves the way towards the development of a preventive system that provides real-time safety and operational instructions and insights for structural health monitoring (SHM) purposes, allowing prevention of environmental damage, and loss of human lives.

Tracking and Visualizing Signs of Degradation for Early Failure Prediction of Rolling Bearings

Predictive maintenance, which means detection of failure ahead of time, is one of the pillars of Industry 4.0. An effective method for this technique is to track early signs of degradation before failure occurs. This paper presents an innovative failure predictive scheme for machines. The proposed scheme combines the use of the full spectrum of vibration data from the machines and a data visualization technology. This scheme requires no training data and can be started quickly after installation. First, we proposed to use the full spectrum (as high-dimensional data vectors) with no cropping and no complex feature extraction and to visualize the data behavior by mapping the high-dimensional vectors into a two-dimensional (2D) map. This ensures simplicity of the process and less possibility of overlooking important information as well as provide a human-friendly and human-understandable output. Second, we developed a real-time data tracker that can predict failure at an appropriate time with sufficient allowance for maintenance by plotting real-time frequency spectrum data of the target machine on a 2D map created from normal data. Finally, we verified our proposal using vibration data of bearings from real-world test-to-failure measurements obtained from the IMS dataset.

Representing Time Series Data in Intelligent Training Systems

Many of the most popular intelligent training systems, including driving and flight simulators, generate user time series data. This paper presents a comparison of representation options for two different student modeling problems: 1) early failure prediction and 2) classifying student activities. Data for this analysis was gathered from pilots executing simple tasks in a virtual reality flight simulator. We demonstrate that our proposed embedding which uses a combination of dynamic time warping (DTW) and multidimensional scaling (MDS) is valuable for both student modeling tasks. However, Euclidean distance + MDS was found to be a superior embedding for predicting student failure, since DTW can obscure important agility differences between successful and unsuccessful pilots.

Hybrid Particle Filter Trained Neural Network for Prognosis of Lithium-Ion Batteries

Prognostics and Health Management (PHM) plays a key role in Industry 4.0 revolution by providing smart predictive maintenance solutions. Early failure detection and prediction of remaining useful life (RUL) of critical industrial machines/components are the main challenges addressed by PHM methodologies. In literature, model-based and data-driven methods are widely used for RUL estimation. Model-based methods rely on empirical/phenomenological degradation models for RUL prediction using Bayesian formulations. In many cases, the lack of accurate physics-based models emphasizes the need to resort to machine learning based prognostic algorithms. However, data-driven methods require extensive machine failure data incorporating all possible operating conditions along with all possible failure modes pertaining to that particular machine/component, which are seldom available in their entirety. In this work, we propose a three-stage hybrid prognostic algorithm (HyA) combining model-based (Particle Filters-PF) and data-driven (Neural Networks-NN) methods in a unique way. The proposed method aims to overcome the need for accurate degradation modeling or extensive failure data sets. In the first stage, a feedforward neural network is used to formulate lithium-ion battery’s degradation trends and the corresponding NN model parameters are used to define the initial prior distribution of PF algorithm. In the second stage, the PF algorithm optimizes the model parameters and the posterior model parameter distributions are utilized to ‘ warm-start ’ the neural network used for prognosis and the third/final stages focuses on prognosis and RUL estimation using the trained NN model leveraging on the posterior distributions of the PF fine-tuned weights and biases. The proposed method is demonstrated on CALCE and NASA lithium-ion battery capacity degradation datasets. The efficacy of the proposed hybrid algorithm is evaluated using root mean square error (RMSE) values and alpha-lambda prognostic metrics. Also, the impact of the NN architecture on the prediction accuracy and computational load are analyzed.

Changepoint-Based Anomaly Detection for Prognostic Diagnosis in a Core Router System

Prognostic diagnosis is desirable for commercial core router systems to ensure early failure prediction and fast error recovery. The effectiveness of prognostic diagnosis depends on whether anomalies can be accurately detected before a failure occurs. However, traditional anomaly detection techniques fail to detect “outliers” when the statistical properties of the monitored data change significantly as time proceeds. We describe the design of a changepoint (CP)-based anomaly detector that first detects CPs from collected time-series data, and then utilizes these CPs to detect anomalies. Different CP detection approaches are implemented to detect various types of CPs. A clustering method is then developed to identify normal/abnormal patterns from CP windows. Data collected from a set of commercial core router systems are used to validate the proposed anomaly detector. Experimental results show that our CP-based anomaly detector achieves better performance than traditional methods in terms of two metrics, namely success ratio and nonfalse-alarm ratio.

Prediction of Online Students Performance by Means of Genetic Programming

ABSTRACT Problem: Online higher education (OHE) failure rates reach 40% worldwide. Prediction of student performance at early stages of the course calendar has been proposed as strategy to prevent student failure. Objective: To investigate the application of genetic programming (GP) to predict the final grades (FGs) of online students using grades from an early stage of the course as the independent variable Method: Data were obtained from the learning management system; we performed statistical analyses over FGs as dependent variable and 11 independent variables; two statistical and one GP models were generated; the prediction accuracies of the models were compared by means of a statistical test. Results: GP model was better than statistical models with confidence levels of 90% and 99% for the training testing data sets respectively. These results suggest that GP could be implemented for supporting decision making process in OHE for early student failure prediction.