Failure Prediction Research Articles

Predicting business bankruptcy: A comparative analysis with machine learning models

Business failure prediction has become crucially important for today's firms, enabling them to reduce financial risks and make informed decisions. This study uses a dataset of 6819 companies and 96 financial and macroeconomic variables to present a comparative analysis of machine learning (ML) models for predicting corporate bankruptcies. Behind this research is to improve the accuracy of bankruptcy prediction, which can help companies make more informed decisions and reduce financial risks. This study aims to evaluate the effectiveness of 16 ML algorithms in terms of accuracy, sensitivity, and other relevant metrics. The work uses methodologies that include data collection and cleaning, exploratory data analysis, model preprocessing and training, and model performance evaluation. Data preprocessing and hyperparameter optimization techniques were used to improve model performance. The evaluated algorithms include Classifiers such as Stacking Classifier (SCC), Randomized Search Classifier (RCV), Historical Gradient Boosting Classifier (HGBC), MLP Classifier (MLPC), K-Neighbors Classifier (KNC), Decision Tree Classifier (DTC), XGBRF Classifier (XGBRFC), Support Vector Classifier (SVC), Logistic Regression Classifier (LR), Linear SVC Classifier (LSVC). With an accuracy of 97.63 %, recall of 97.63 %, and F1-score of 97.63 %, the results show that the SCC algorithm was the best. Other models, such as RCV and DTC, also showed good results, with accuracies above 97 %. However, models such as PAC and BNB had lower performance and accuracy below 90 %. Finally, this study compares the results of ML models in predicting business failures and highlights their effectiveness. The SCC algorithm is considered the most suitable model for this task, as it suggests that it can help economic actors make more informed decisions and reduce financial risks in the context of firms.

Intraoperative Prediction of Coronary Graft Failure Based on Transit Time Flow Measurement: A PRELIMINARY STUDY.

Myocardial revascularization has been known to not affect the prognosis in some patients. Coronary artery bypass graft (CABG) failure may develop one year after CABG surgery. This is accompanied by a high risk of developing myocardial infarction after complete myocardial revascularization in obstructive coronary artery disease (CAD) due to microvascular dysfunction. The study of microvascular dysfunction using intraoperative stress tests with adenosine triphosphate (ATP) allows for the assessment of the coronary bypass flow reserve (CBFR) and the risk of graft failure one year after surgery. The study included 79 CAD patients (238 grafts) who underwent dynamic single-photon emission computed tomography (SPECT) before CABG and dynamic transit time flow measurement (TTFM) during CABG at rest and at stress. The CBFR was calculated by the ratio of the mean graft flow (MGF) at stress to the MGF at rest. A multivariate regression model showed that the MGF at rest (p = 0.043), the MGF at stress (p = 0.026) and the CBFR (p = 0.0001) were significant independent predictors of graft failure. As a result of ROC analysis, the threshold CBFR < 1.67 units correlated with graft failure more closely (sensitivity 82%, specificity 90%) The CBFR is a significant independent predictor of graft failure for up to 16 months.

Design methodology for fused filament fabrication with failure theory: framework, database, design rule, methodology and study of case

PurposeAdditive manufacturing (AM) is growing economically because of its cost-effective design flexibility. However, it faces challenges such as interlaminar weaknesses and reduced strength because of product anisotropy. Therefore, the purpose of this study is to develop a methodology that integrates design for additive manufacturing (AM) principles with fused filament fabrication (FFF) to address these challenges, thereby enhancing product reliability and strength.Design/methodology/approachDeveloped through case analysis and literature review, this methodology focuses on design methodology for AM (DFAM) principles applied to FFF for high mechanical performance applications. A DFAM database is constructed to identify common requirements and establish design rules, validated through a case study.FindingsExisting DFAM approaches often lack failure theory integration, especially in FFF, emphasizing mechanical characterizations over predictive failure analysis in functional parts. This methodology addresses this gap by enhancing product reliability through failure prediction in high-performance FFF applications.Originality/valueWhile some DFAM methods exist for high-performance FFF, they are often specific cases. Existing DFAM methodologies typically apply broadly across AM processes without a specific focus on failure theories in functional parts. This methodology integrates FFF with a failure theory approach to strengthen product reliability in high-performance applications.

Pipe Failure Prediction in the Water Distribution System Using a Deep Graph Convolutional Network and Temporal Failure Series

Pipe Failure Prediction in the Water Distribution System Using a Deep Graph Convolutional Network and Temporal Failure Series

Finite element mesh transition for local–global modeling of composite structures

This study presents an automatic mesh generation algorithm designed to address computational challenges in simulating small-scale defects within large composite structures. The algorithm seamlessly transitions from a coarse mesh, corresponding to the global structure, to a highly refined mesh in targeted local regions of interest. The transition element number and shape can be adjusted by the specified parameters. Tailored to complement this method for non-homogeneous composite models, which include multiple materials such as cohesive layers representing interlayer properties, a volume fraction calculator is integrated to automatically assign the mixture material property in each transition element. Entire processes are fully automated using a MATLAB script, eliminating the need to open the FEA software interface. The validation studies of the reconstructed two-dimensional models, assembled with the wrinkle-defect model, demonstrate their feasibility. The performance of the model is examined in terms of strain and displacement at the connecting boundaries, load–displacement curve, and interlayer failure prediction. The mesh transition model achieves agreeable results compared to a fully fine mesh model, and a 92% reduction in computational time in stress analysis, showing the efficiency of the mesh transition for local–global modeling of composite structures.

Polymer gear failure prediction: A regression-Based approach using FEA and photoelasticity technique

The research aims to develop a model to predict the maximum contact stress depth, which will be correlated with polymer gear failures such as wear and rapid thermal deformation. A thorough 3D Finite Element Analysis (FEA) was performed for different gear parameters (varying module, number of teeth, and face width) in different combinations (Polymer – Polymer, Steel – Polymer, and Polymer – Steel). The simulation results showed that the contact stress depth increases with respect to the module and number of teeth; meanwhile, very negligible variations were found along the face width. Also, the findings indicate that the contact stress depth varied in the following descending combinations of gear pair: Polymer – Polymer, Steel – Polymer, and Polymer – Steel. The 3D FEA contact stress patterns were validated experimentally by utilizing the in-house developed photoelasticity test rig. The data obtained from the simulation results were used to develop regression models (Linear regression and Random Forest regressor) for estimating the maximum contact stress depth. A feature importance study was done using the mean reduction of the Gini index or impurities in random forest. It was found that torque applied and contact width of the gear pair were significantly influenced by the contact stress depth. Finally, an equation was developed correlating torque and contact width to predict the depth occurrence of maximum contact stress through which the failure mode of the polymer gear could be predicted.

Mathematical Framework for Enhancing Machine Failure Prediction in Aviation and Beyond: Leveraging Deep Convolutional Neural Networks and Visual Analytic

Mathematical Framework for Enhancing Machine Failure Prediction in Aviation and Beyond: Leveraging Deep Convolutional Neural Networks and Visual Analytic

Temperature-based energy changes in gold nanowire sensors for flow rate detection and failure prediction

This study delves into the dynamic operational states of nanowire sensors within microfluidic channels for flow velocity detection, classifying these states into decline, steady, and failure states. It conducts a comparative analysis of the steady-state operational characteristics of nanowire sensors across various voltages and inlet flow rates, alongside examining the interplay between nanowire resistivity and temperature. This investigation identifies an optimal working condition for the sensor at a 1 V operating voltage and a nanowire dimension of 150 × 150 nm. Employing both theoretical and experimental approaches, the research elucidates the convective heat transfer mechanism that underpins the functionality of the nanowire sensor for flow rate detection. It unveils the pattern of nanowire temperature changes during inlet flow rate measurement within a range from 10 to 50 μL/min range and establishes standard values of temperature and calorimetric differences for various flow velocities. Furthermore, the study achieves the prediction of the time to reach the steady and failure states under given operational conditions for the nanowire-based sensor. This advancement will enhance the reading precision of the signal output from the nanowire sensor, mitigate the risk of sensor failure, and contribute to prolonging the operational lifespan of equipment.

Full-field experiment-aided virtual modelling framework for inverse-based stochastic prediction of structures with elastoplasticity

Forecasting of stochastic ductile failures in fabrication and service stages are challenging tasks of advanced structures in practical engineering. Due to the prohibitive costs of repetitive experimental tests to quantify optimal failure-related parameters, numerous studies have turned to simulation-based uncertainty quantification. However, the credibility of these approaches is frequently doubted by the function-generated distribution of random data involved. Thus, an accurate assessment of the material parameters ascertains the appropriate estimation of uncertain parameters, which is essential for the risk/safety assessment of structures in different stages. In this paper, a full-field experiment-aided virtual modelling framework for inverse-based prediction of stochastic elastoplastic failure (EVMF-ISP) is proposed to deliver precise insights regarding the effective distribution range of mechanical parameters. To this end, all available experiment observations serve as the reference for assessing the prior and posterior probability density of the unknown parameters through the real-time inverse uncertainty quantification (UQ) module. The framework can be divided into three parts, where initially a pre-virtual model (PRVM) is formulated, and Bayesian inference is implemented to propagate the experiment observations backwards to ascertain the uncertain parameters. Then, an advanced multidimensional slice sampling method is developed to deal with the derived complex posterior probability density function (PDF) of mechanical parameters. In the end, a reliable stochastic elastoplastic analysis can be conducted with the revised uncertain samples and finalised with post-virtual models (POVMs) for the concerned structures. Such that, accurate and efficient determination of nonlinear response of structures can be directly predicted. The EVMF-ISP framework is logically presented and thoroughly illustrated with practical applications.

Probabilistic assessment of complex corrosion in pipelines considering River-Bottom Profile information

One of the important causes of failure and incidents in pipelines is corrosion. Due to the severe consequences and impact in several social, economic, and environmental areas, pipelines must be regularly inspected. The possibilities for pipeline assessment are standard equations and numerical simulations. While the first is simple and fast, it tends to be conservative and the second is expensive. The finite element method (FEM) has been successfully used to predict the failure pressure of pipelines under different conditions. Commonly assessment methods overlook the uncertainties for failure predictions, often relying on safety factors. However, it is imperative to consider uncertainties to ascertain the reliability and risk of corroded pipes. While the reliability analysis of pipelines with idealized defects has been extensively explored in the literature, studies considering complex corrosion profiles, are still demanding. This paper addresses this gap considering the traditional Monte Carlo (MC) and first-order reliability (FORM) methods. The assessment is based on the Effective Area Method (EAM) within the failure function. As well know, MC can be expensive due to the high number of function evaluations. To mitigate this, FORM is employed as it requires fewer function evaluations. However, function gradients are required as it relies on an iterative optimization algorithm. Particularly challenges arise when computing the gradient using EAM for burst pressure estimation. The EAM is an iterative procedure, and small perturbations can yield incorrect sensitivity values, causing problems for FORM convergence. To address this, we propose a novel procedure for efficiently and less conservatively obtaining the gradients required by FORM using EAM as a failure function. The modified FORM method achieved convergence within 15 iterations for all test cases. This alternative approach offers a faster, more precise and less conservative method for Level-2 probabilistic failure analysis of complex-shaped corrosion defects compared to traditional Level-1 methods.

A HEART FAILURE PREDICTION ALGORITHM BASED ON IMPROVED REINFORCEMENT LEARNING FRAMEWORK

Reducing the occurrence of diseases has become an important research direction in today’s social medicine. Accurate prediction of heart failure can greatly reduce the complication rate, allowing patients to know the condition in advance and receive treatment. To deal with heart failure, this paper proposes a heart failure prediction algorithm based on a reinforcement learning framework. This paper combines and improves the reinforcement learning algorithm and the swarm intelligence (SI) optimization algorithm, aiming to use the reinforcement learning algorithm to improve the global optimization capability of the SI optimization algorithm. In order to better improve the global search ability of the SI optimization algorithm, the penalty function will be replaced by a combination mode of dynamic and static reward. The improved reinforcement learning algorithm framework is significantly better than the previous algorithm framework and has been used to predict heart failure. The experimental results have proven the effectiveness of the algorithm and its accuracy in prediction of heart failure.

Dynamic risk assessment of natural gas transmission pipelines with LSTM networks and historical failure data

In the realm of energy infrastructure, ensuring the security of gas transmission pipelines is critical. This research introduces an advanced dynamic risk assessment framework that leverages the predictive capabilities of LSTM networks, presenting an improvement over conventional failure prediction models. Unlike traditional approaches that rely on averaging historical failure records, this framework dynamically processes historical pipeline failure incidents into sequential time series analysis, facilitating and improving the accuracy of the current failure rate. The model refines the failure rate estimation for individual pipelines by incorporating unique characteristics and modification factors, resulting in a highly precise failure likelihood estimation. Additionally, this study introduces a quantifiable linkage between mortality risk and the fatality probit value across various accident scenarios, enhancing consequence evaluation. A sensitivity analysis is then performed to assess the impact of various input parameters on the model's performance. The practical application of the model on a U.S. pipeline confirms its effectiveness. This proposed methodology substantially enhances the understanding of incident causation in gas pipeline systems, paving the way for superior safety management strategies. It is instrumental in enhancing pipeline safety, refining infrastructure planning, and optimizing safety resource allocation. The methodology offers benefits for pipeline operators, industry professionals, and regulatory agencies, contributing to improved operational safety and resource management in the pipeline industry.

Validation of the fully rationalized Tsai-Wu failure criterion for unidirectional laminates under multiaxial stress states through a ring-on-ring test

Validating a failure criterion under complex stress states is important because composites often encounter multiaxial stress in practical applications, but it also helps to expose any deficiencies of the criterion in failure prediction in the presence of stress interactions. This study examines the recently formulated fully rationalized Tsai-Wu criterion for unidirectional laminates under multiaxial stress states created by ring-on-ring loading. Distinct stress distributions under this loading condition are exhibited and quantified by numerical modeling, featuring critical biaxial tension on the bottom surface and triaxial compression on the top surface. The fully rationalized Tsai-Wu criterion and the original Tsai-Wu criterion are used to predict failure, with the results indicating good agreement between the former and experimental data. In comparison with the latter, quadric failure surfaces of the two criteria are constructed to analyze differences in prediction. The findings demonstrate the validity of the rationalization work for the Tsai-Wu criterion and its potential in predicting failure under multiaxial stress states.

Structural Failure Prediction of RCC Structure Under Earthquake Using Machine Learning

Earthquake is a most dangerous and catastrophic disaster in the world. It is essential to introduce earthquake resistance features in construction. An earthquake -prone region must deal with the more challenging problem of earthquakes. And there is no more hundred percent earthquake proof building. The best way to avoid the building damages is to learn from the failures that can occur during a seismic event. This paper presents the failure prediction of RCC building using k-Nearest Neighbors (KNN) algorithm which can be tuned to gives up to 90% accuracy of result. A past- earthquake assessment of building data is used to train and create a model for this project to predict the necessary pattern for a future construction project. Thereby, can be predicted whether the proposed building will be collapsed or facing a minimum damage. Key Words: Earthquake Assessment, Machine Learning, Python, K-Neighbour Algorithm, RCC Building, Failure Prediction, Site issues, Soil and Foundation Condition

Experimental study on overtopping failure of concrete face rockfill dam

With the frequent occurrence of natural disasters, the problem of dam failure with low probability and high risk has gradually attracted people's attention. This paper uses flume model tests to systematically analyze the overtopping failure mechanisms of concrete face rockfill dam (CFRD) and identify its failure modes. The tests reveal that the longitudinal erosion of the CFRD breach progress through stages of soil erosion, panel failures, and water flow stabilization. Meanwhile, the cross-section breach process involves the evolution of breach size in rockfill materials, including traceable erosion, lateral broadening, and breach morphology stabilization. The fracture characteristics of the water-blocking panel are primarily evident in the flow-time curve. By analyzing the breach morphology evolution processes in longitudinal and cross sections, the flow-time curve can be subdivided into stages of burst flow formation, breach expansion with flow increase, rapid increase of breach flow discharge due to panel failures, and stabilization of breach flow and size. The primary damage process of the CFRD occurs in a cyclical stage of breach expansion, flow increase, panel failure, and rapid discharge. The rigid face plate and granular body structure contribute to partial dam failure, showing a tendency for gradual expansion of the breach. The longitudinal section illustrates dam failure resulting from panel fracture and rockfill erosion interaction, while downstream slopes exhibit failure due to lateral intrusion of rockfill and cyclic instability. These research results can serve as a reference for constructing a concrete CFRD failure prediction model and conducting disaster risk assessments.

On an efficient global/local stochastic methodology for accurate stress analysis, failure prediction and damage tolerance of laminated composites

The quantification of uncertainties in the mechanical response of composite structures can be a computationally demanding task. This is due both to the number of uncertain parameters in a real study case and the complexity of the model to be analysed. In this paper, an efficient global/local approach to estimate the uncertainties of the quantities of interest in specific regions of interest with limited computational effort is proposed. This is achieved by refining only locally the model taking advantage of Refined Structural Theories. At the same time, since the variance of the uncertain parameters is usually relatively small, the stochastic analysis is dealt with a sensitivity study carried out both in the global and in the local model. In this way, it is possible to assess the influence of global and local uncertain parameters in the same submodeling analysis. The methodology presented is applied to several study cases of interest. The results focus on obtaining probabilistic distributions of the stress field that can be later used in failure criteria to evaluate the subsequent distribution of the failure index. Furthermore, a damage tolerance study case is investigated, showing good correlation with the reference Monte Carlo simulations.

On the applications of neural ordinary differential equations in medical image analysis

Medical image analysis tasks are characterized by high-noise, volumetric, and multi-modality, posing challenges for the model that attempts to learn robust features from the input images. Over the last decade, deep neural networks (DNNs) have achieved enormous success in medical image analysis tasks, which can be attributed to their powerful feature representation capability. Despite the promising results reported in numerous literature, DNNs are also criticized for several pivotal limits, with one of the limitations is lack of safety. Safety plays an important role in the applications of DNNs during clinical practice, helping the model defend against potential attacks and preventing the model from silent failure prediction. The recently proposed neural ordinary differential equation (NODE), a continuous model bridging the gap between DNNs and ODE, provides a significant advantage in ensuring the model’s safety. Among the variants of NODE, the neural memory ordinary differential equation (nmODE) owns the global attractor theoretically, exhibiting superiority in prompting the model’s performance and robustness during applications. While NODE and its variants have been widely used in medical image analysis tasks, there is a lack of a comprehensive review of their applications, hindering the in-depth understanding of NODE’s working principle and its potential applications. To mitigate this limitation, this paper thoroughly reviews the literature on the applications of NODE in medical image analysis from the following five aspects: segmentation, reconstruction, registration, disease prediction, and data generation. We also summarize both the strengths and downsides of the applications of NODE, followed by the possible research directions. To the best of our knowledge, this is the first review regards the applications of NODE in the field of medical image analysis. We hope this review can draw the researchers’ attention to the great potential of NODE and its variants in medical image analysis.

Customised-sampling approach for pipe failure prediction in water distribution networks

This paper presents a new methodology for addressing imbalanced class data for failure prediction in Water Distribution Networks (WDNs). The proposed methodology relies on existing approaches including under-sampling, over-sampling, and class weighting as primary strategies. These techniques aim to treat the imbalanced datasets by adjusting the representation of minority and majority classes. Under-sampling reduces data in the majority class, over-sampling adds data to the minority class, and class weighting assigns unequal weights based on class counts to balance the influence of each class during machine learning (ML) model training. In this paper, the mentioned approaches were used at levels other than “balance point” to construct pipe failure prediction models for a WDN with highly imbalanced data. F1-score, and AUC–ROC, were selected to evaluate model performance. Results revealed that under-sampling above the balance point yields the highest F1-score, while over-sampling below the balance point achieves optimal results. Employing class weights during training and prediction emphasises the efficacy of lower weights than the balance. Combining under-sampling and over-sampling to the same ratio for both majority and minority classes showed limited improvement. However, a more effective predictive model emerged when over-sampling the minority class and under-sampling the majority class to different ratios, followed by applying class weights to balance data.

Bilinear constitutive model for the anisotropic failure prediction of PLA specimens produced by material extrusion additive manufacturing

Bilinear constitutive model for the anisotropic failure prediction of PLA specimens produced by material extrusion additive manufacturing

Model for the Failure Prediction Mechanism of In-Service Pipelines Based on IoT Technology

With the rapid increase in pipeline mileage in China, the accurate prediction of corrosion issues in in-service pipelines has become crucial for ensuring safe pipeline operation. Traditional pipeline leakage monitoring methods are significantly limited by human factors and equipment precision, making it challenging to predict and identify leakage points accurately. Therefore, aligned with the trend of intelligent pipeline development, this study aims to construct a failure pressure prediction mechanism model for corroded pipelines based on IoT technology. This model leverages intelligent sensing and prediction to assess the safety status of corroded pipeline sections. Ultrasonic phased array technology detects specific corrosion points and detailed defect parameters within pipeline sections. The parameters are then utilized in the Simdroid domestic finite element analysis model to simulate the ultimate burst pressure of the pipeline. A single-variable approach is employed to analyze the sensitivity of different parameters to the pipeline’s ultimate burst pressure, with the minimum burst pressure point of multi-point corroded sections selected as the overall segment failure pressure. Finite element simulation data are integrated into a neural network database to predict the pipeline failure pressure. The real-time operational data of the pipeline are monitored using negative-pressure wave sensing. The operational pressure of the corroded points is compared with the algorithm-predicted failure pressure; if the values approach a critical threshold, an alarm is triggered. Moreover, the remote control terminals evaluate the pipeline’s self-rescue time, providing a buffer for pipeline leakage self-rescue. The failure prediction mechanism model for in-service pipelines was applied to the Fujian–Guangdong branch of the West–East Gas Pipeline III to verify its accuracy and feasibility. The research results offer technical support for the maintenance and emergency repair of pipeline leakage scenarios, leveraging intelligent pipeline technology to reduce costs and increase the efficiency of pipeline operations, thereby supporting the sustainable development of China’s oil and gas pipelines with theoretical and technical backing.