Failure Rate Prediction Research Articles

Prediction of Pipe Failure Rate in Heating Networks Using Machine Learning Methods

The correct prediction of heating network pipeline failure rates can increase the reliability of the heat supply to consumers in the cold season. However, due to the large number of factors affecting the corrosion of underground steel pipelines, it is difficult to achieve high prediction accuracy. The purpose of this study is to identify connections between the failure rate of heating network pipelines and factors not taken into account in traditional methods, such as residual pipeline wall thickness, soil corrosion activity, previous incidents on the pipeline section, flooding (traces of flooding) of the channel, and intersections with communications. To achieve this goal, the following machine learning algorithms were used: random forest, gradient boosting, support vector machines, and artificial neural networks (multilayer perceptron). The data were collected on incidents related to the breakdown of heating network pipelines in the cities of Kazan and Ulyanovsk. Based on these data, four intelligent models have been developed. The accuracy of the models was compared. The best result was obtained for the gradient boosting regression tree, as follows: MSE = 0.00719, MAE = 0.0682, and MAPE = 0.06069. The feature «Previous incidents on the pipeline section» was excluded from the training set as the least significant.

An ensemble learning model for forecasting water-pipe leakage

Based on the benefits of different ensemble methods, such as bagging and boosting, which have been studied and adopted extensively in research and practice, where bagging and boosting focus more on reducing variance and bias, this paper presented an optimization ensemble learning-based model for a large pipe failure dataset of water pipe leakage forecasting, something that was not previously considered by others. It is known that tuning the hyperparameters of each base learned inside the ensemble weight optimization process can produce better-performing ensembles, so it effectively improves the accuracy of water pipe leakage forecasting based on the pipeline failure rate. To evaluate the proposed model, the results are compared with the results of the bagging ensemble and boosting ensemble models using the root-mean-square error (RMSE), the mean square error (MSE), the mean absolute error (MAE), and the coefficient of determination (R2) of the bagging ensemble technique, the boosting ensemble technique and optimizable ensemble technique are higher than other models. The experimental result shows that the optimizable ensemble model has better prediction accuracy. The optimizable ensemble model has achieved the best prediction of water pipe failure rate at the 14th iteration, with the least RMSE = 0.00231 and MAE = 0.00071513 when building the model that predicts water pipe leakage forecasting via pipeline failure rate.

A sensor-driven operations and maintenance planning approach for large-scale leased manufacturing systems

The rise of mass customisation drives manufacturers to adopt leasing agreements for their machinery rather than outright their ownership. This industrial trend results in large-scale leased manufacturing systems, where the asset conditions and maintenance requirements for a fleet of machines and/or production lines from several manufacturers are continuously monitored and managed by a lessor. In this paper, we propose an operations and maintenance planning model that explicitly models (i) dynamic real-time failure rate predictions for machines based on sensor-driven degradation data, (ii) optimal routes for maintenance teams across a number of geographically-distributed manufacturing sites, and (iii) operational outcomes. The model can handle complex maintenance and operational outcome interdependencies between multiple machines, and incorporate demand for different products, heterogeneous machine capacities and capabilities, factory topology, and the resulting production flow capacities of products. To demonstrate the model’s practicality, it is applied to three experimental case studies with various numbers of machines, operating schedules, product types, and flow capacities. The results show that the proposed model significantly outperforms the traditional reliability-based maintenance model in key metrics such as the number of unexpected failures, the cost of preventive maintenance actions, machine downtime, the percentage of unsatisfied demand, and total cost.

High Overload Failure Rate Prediction Method for Packaged Devices Based on Stress–Strength Interference Modeling

High Overload Failure Rate Prediction Method for Packaged Devices Based on Stress–Strength Interference Modeling

Reliability evaluation of non-isolated high gain interleaved DC-DC converter

A high gain DC-DC converter is the crucial part in renewable energy systems (RES) and in electric vehicular systems. The reliability of those high-gain converters needs to be assessed for the long-term operation of renewable energy systems. This article presents the reliability analysis of non-isolated high gain interleaved DC-DC converter. The analysis primarily relies on calculating the mean time between failures (MTBF). Based on military handbook (MIL-HDBK-217) criteria, the reliability calculation is performed. Stress factors and predicted failure rate for each component of presented converter is evaluated and tabulated. Reliability evaluation is performed for 1.5 kW hardware prototype. Based on reliability evaluation results, a reliable converter with better operating life time has been introduced.

A Comprehensive Analysis of the Reliability of Electric Vehicle Motor Systems: Exploring the Intricacies of Performance and Durability

Electric vehicles (EVs) have garnered significant attention in recent years due to their distinctive zero carbon emission capability. While their environmental advantages are widely acknowledged, concerns about the reliability of key components, particularly the drive motors and associated controllers, continue to linger. Previous studies have made strides in evaluating the reliability of individual drive motors in EVs. However, given the integrated nature of drive motors and motor controllers in EVs, a comprehensive assessment that considers both components as a unified system is imperative for more precise reliability forecasting. Furthermore, the multifaceted nature of drive motors and motor controllers, encompassing various components with diverse structures, types, and characteristics, can significantly influence the overall reliability of the motor system. Unfortunately, these nuances have often been overlooked in prior research, creating a knowledge gap that needs to be addressed. To address this gap, our study offers a holistic investigation into the reliability of the entire motor system in pure electric vans, which includes both the drive motors and motor controllers. We initiate our research by employing theoretical models to predict the failure rates of individual subassemblies and components within the drive motor and motor controller. Building upon these predictions, we conduct an in-depth assessment of the overall motor system reliability. Our findings reveal the vulnerabilities associated with specific subassemblies and components within the motor system, offering critical insights that challenge some of the prevailing assumptions in existing reliability research. These insights are instrumental in guiding future advancements in reliability design and maintenance practices tailored for pure electric vans. The adoption of EVs continues to grow, driven by their environmental benefits and expanding market share, it becomes increasingly crucial to address concerns related to the reliability of essential components like drive motors and control systems. While previous research has largely focused on component-level reliability, our study adopts a more comprehensive approach, evaluating the entire motor system. Through a combination of theoretical modeling and empirical analysis, we aim to provide a robust framework for predicting failure rates and enhancing the overall reliability of EV motor systems. By uncovering vulnerabilities and potential areas for improvement, this research offers valuable insights to advance the design, development, and maintenance of reliable electric vehicles.

Reliability analysis of wind turbines considering seasonal weather effects

The failure rate of wind turbines shows obvious fluctuation due to seasonal environmental factors. However, few efforts have been devoted to modeling the seasonal failure rate. This paper develops a novel model that consists of a baseline failure rate function, seasonal indices, and a residual term to describe the monthly failure rate of wind turbines. A two-stage procedure is developed to estimate the 16 unknown parameters in the model. The first stage explores the relationship between the parameters in the baseline function and the monthly coefficients by maximum likelihood estimation and then integrates the properties into the genetic algorithm to estimate the main parameters. In the second stage, the variance of the residual term is estimated based on the analysis of the differences between the observed and predicted failure rates. The failure history of 48 months has been used to illustrate the proposed approach. The results show that the monthly failure rate function can well fit the real failure history of wind turbines, and it outperforms the traditional reliability model.

Adopting New Machine Learning Approaches on Cox’s Partial Likelihood Parameter Estimation for Predictive Maintenance Decisions

The Proportional Hazards Model (PHM) under a Condition-Based Maintenance (CBM) policy is used by asset-intensive industries to predict failure rate, reliability function, and maintenance decisions based on vital covariates data. Cox’s partial likelihood optimization is a method to assess the weight of time and conditions into the hazard rate; however, parameter estimation with diverse covariates problem could have multiple and feasible solutions. Therefore, the boundary assessment and the initial value strategy are critical matters to consider. This paper analyzes innovative non/semi-parametric approaches to address this problem. Specifically, we incorporate IPCRidge for defining boundaries and use Gradient Boosting and Random Forest for estimating seed values for covariates weighting. When applied to a real case study, the integration of data scaling streamlines the handling of condition data with diverse orders of magnitude and units. This enhancement simplifies the modeling process and ensures a more comprehensive and accurate underlying data analysis. Finally, the proposed method shows an innovative path for assessing condition weights and Weibull parameters with data-driven approaches and advanced algorithms, increasing the robustness of non-convex log-likelihood optimization, and strengthening the PHM model with multiple covariates by easing its interpretation for predictive maintenance purposes.

Exploring predicted failure rates in long-term hepatitis B vaccination: An analysis of study data

Exploring predicted failure rates in long-term hepatitis B vaccination: An analysis of study data

Prediction of Remaining Useful Life (RUL) of Electronic Components in the POSAFE-Q PLC Platform under NPP Dynamic Stress Conditions

In the Korean domestic nuclear industry, to analyze the reliability of instrumentation and control (I&C) systems, the failure rates of the electronic components constituting the I&C systems are predicted based on the MIL-HDBK-217F standard titled ‘Reliability Prediction of Electronic Equipment’. Based on these predicted failure rates, the mean time to failure of the I&C systems is calculated to determine the replacement period of the I&C systems. However, this conventional approach to the prediction of electronic component failure rates assumes that factors affecting the failure rates such as ambient temperature and operating voltage are static constants. In this regard, the objective of this study is to propose a prediction method for the remaining useful life (RUL) of electronic components considering mean time to failure calculations reflecting dynamic environments, such as changes in ambient temperature and operating voltage. Results of this study show that the RUL of electronic components can be estimated depending on time-varying temperature and electrical stress, implying that the RUL of electronic components can be predicted under dynamic stress conditions.

Sliding Window-Based Machine Learning for Environmental Inspection Resource Allocation.

Environmental regulation is pivotal in mitigating environmental risks and promoting sustainable development, yet regulators frequently encounter resource constraints when inspecting enterprises. To address this limitation, we employed four sliding window-based machine learning techniques to enhance effective environmental inspections. Utilizing feature-engineered time-series data of characteristic and compliance records from 16,777 chemical enterprises in Jiangsu between 2010 and 2021, the four models were used to establish the predictive models that link enterprise characteristics to the likelihood of inspection failure. The results indicated that the models were comparable with the widely utilized deep learning sequence model, the Long Short-Term Memory model, achieving areas under the ROC (receiver operating characteristic) curves exceeding 0.83. Past violation significantly influences future violations, with a more recent violation history exerting stronger impacts. Besides, using predicted failure rates, we proposed seven resource allocation scenarios, considering different regulation intensities to target high-risk enterprises. Notably, the provincial-level risk-based method yielded a significant inspection failure rate increase (detecting violation), surpassing 8-fold the baseline. Overall, our study offers regulators an optimized inspection resource allocation method, thereby enhancing the regulatory effectiveness. Moreover, this study demonstrates the potential of window-based machine learning techniques in environmental regulation and highlights the importance of data-driven decision-making for promoting sustainable development.

Introduction of an Auxiliary Breaker into the Generator-Transformer Block for Energy Saving in Open Switchgear Circuits of Power Plants

A need in finding of new ways of energy saving at open switchgears of power plants is substantiated. In order to increase energy saving efficiency, an auxiliary breaker is suggested to be inserted between a transformer of a block and its two high-voltage circuit breakers. The reasonability of such an insertion is proved on the basis of comparing of calculations resultsof under-discharge of electricity (UE) by the tabular-logical method (Yu. B. Guk) of the obtained schemes and of the traditional ones. For the calculations, the conditions that arose due to the change in the main circuit of a power plant are studied. Also, equations are given for calculation of a decrease in UE, damage due to it during reconstruction, and costs for the construction of a power plant (the costs are assumed to be the same in all options). Russian statistical data and the predicted failure rate λEV of a 750 kV SF6 circuit breaker are used. An option of the introduction of an SF6 circuit breaker with and without replacement of other circuit breakers with SF6 circuit breakers is considered. The results of calculations of UE, damage, and costs for the introduction suggested are tabulated, where changes in them due to the introduction of the circuit breaker are estimated for 18 ring circuits and 17 “3/2” and “4/3” circuits of 330–750 kV switchgears at condensation, nuclear, and hydroelectric power plants. It is demonstrated that the presence of a generator breaker in the blocks makes it possible to reduce these energy-saving efficiency indicators several times. A technique for determining the failure rate of a hypothetical circuit breaker, which, in the case of traditional replacement, is capable of producing the same effect as an SF6 circuit breaker inserted is proposed. An example of determining this frequency is given. Results of the calculated reduction of UE, damage and costs for the case of an air circuit breaker having been substituted to an SF6 circuit breaker are presented.

Uncertainty quantification of a deep learning model for failure rate prediction of water distribution networks

Predicting the time-dependent pipe failure rate of the water distribution networks (WDNs) is important for planning its renewal budget but also challenging due to the complex factors involved. The recent development of machine learning techniques provides a novel approach for accurate failure rate prediction based on historical data. However, the inherited randomness and uncertainty of water pipe failures and machine learning algorithms are often ignored in the training and prediction process. This article develops an uncertainty quantification framework for deep learning-based WDN system-wide failure rate prediction. The framework integrates a probabilistic long short-term memory (LSTM) model with a Monte Carlo method. The historical climate data and WDN pipe maintenance data over the past 35 years for Cuyahoga County, USA, are used to illustrate this deep-learning model-based uncertainty quantification framework in this study. A statistical time series regression model, ARIMAX, is used as a comparison benchmark. The results show that the LSTM model outperforms the ARIMAX model in prediction accuracy in most years by considering the uncertainties. Besides, the uncertainty range of the LSTM model prediction is 50% of that of the ARIMAX model. The results also identified the major contributing factors to the uncertainties of LSTM machine learning model prediction. The proposed uncertainty framework features excellent extensibility and can be adapted to quantify uncertainties with other types of machine learning models.

Risk Factors for Rotator Cuff Repair Failure and Reliability of the Rotator Cuff Healing Index (RoHI) in Thai Patients: Comparison of the RoHI With a Modified Scoring System

Background: The success rate of surgical treatment for rotator cuff (RC) tear ranges from 16% to 94%. The Rotator Cuff Healing Index (RoHI) is a system for predicting failure after RC repair and is based on a combined score of factors, including age, anteroposterior (AP) tear size, tendon retraction, fatty infiltration of the infraspinatus muscle, bone mineral density (BMD), and level of work activity. Purpose: To determine the factors leading to RC repair failure in a Thai population, to test the reliability of the RoHI in this population, and to compare the RoHI with a modified RoHI (m-RoHI) based on the factors for repair failure as determined. Study Design: Case-control study; Level of evidence, 3. Methods: This study included 133 Thai patients who underwent arthroscopic RC repair between February 2012 and February 2021. Postoperative magnetic resonance imaging was performed at 6 to 24 months to evaluate RC healing. Variables that might affect failure rates were evaluated, including demographic characteristics, AP tear size and retraction, radiographic measurements, and magnetic resonance imaging findings. The m-RoHI was created using factors that significantly predicted repair failure on multivariate analysis. The area under the receiver operating characteristic curve was calculated to determine the reliability of the RoHI and to compare the reliability of the RoHI and m-RoHI to predict failure rates. Results: Multivariate logistic regression analysis revealed that body mass index ≥23 (adjusted odds ratio [OR], 9.02; P = .034), high work activity (adjusted OR, 19.53; P = .008), AP tear size ≥2.5 cm (adjusted OR, 19.04; P = .001), and a retraction size of 2 to <3 cm (adjusted OR, 20.36; P = .013) were the independent factors that predicted repair failure in our population. BMD was not independently predictive of repair failure. We used these 4 significant independent factors to generate the m-RoHI. The area under the curve of the final adjusted m-RoHI was slightly improved as compared with the original RoHI, but this difference was not significant (0.827 [95% CI, 0.741-0.913] vs 0.780 [95% CI, 0.686-0.875], respectively; P = .447). Conclusion: The m-RoHI had a similar predictive value for repair failure to the original RoHI in our study population, but it did not require obtaining BMD. The m-RoHI may be useful in populations where BMD is not routinely obtained.

A robust clustering-based multi-objective model for optimal instruction of pipes replacement in urban WDN based on machine learning approaches

ABSTRACT Water distribution networks (WDNs) face serious management challenges due to the high investment necessity for pipe maintenance and high performance as well as the uncertainties of input variables. To address these challenges, this study aims to prepare and implement the optimal instructions for pipe replacement with maximum hydraulic performance, minimum cost, and minimum uncertainty. Herein, a robust clustering multi-objective (RCMO) approach is developed by combining five models, including hydraulic simulation, multi-objective optimization, pipe failure rate prediction, non-linear interval programming, and multi-criteria decision-making. In this procedure, a clustering method is implemented to reduce the uncertain scenarios of the multi-objective optimization. The new approach is applied to a WDN in Gorgan, Iran. Implementing the optimal instruction increases the network’s physical and hydraulic performance by 56% and 35%, respectively, and decreases the annual deficit of nodes’ demand between 69% and 93%. Also, the proposed methodology reduces the optimization run time by about 99%.

Predictive deep learning for pitting corrosion modeling in buried transmission pipelines

Despite significant efforts and investments in the renewable energy sector, fossil fuels continue to provide the majority of the world's energy supply. Transmission pipelines, which are extensively used in the oil and gas industry, are vulnerable to various failure mechanisms, such as corrosion. Among these, pitting corrosion in offshore pipelines is the most prevalent type of external corrosion. This study explores the potential of deep learning models (Generalization and Generalization-Memorization models) to predict the maximum depth of pitting corrosion in oil and gas pipelines. The models are trained considering various characteristics of the soil where the pipe is buried and different types of the protective coating of the pipes. The application of deep neural networks resulted in a mean squared error of prediction of 0.0055 in training data and 0.0037 in test data. These results demonstrate that deep learning models outperform all empirical and hybrid models applied in previous studies on the same dataset. The proposed model in this study has the potential to predict failure rates of the pipelines due to external corrosion and enhance the safety and reliability of these facilities.

Optimizing Predictive Maintenance Decisions: Use of Non-Arbitrary Multi-Covariate Bands in a Novel Condition Assessment under a Machine Learning Approach

Jointing Condition-Based Maintenance (CBM) with the Proportional Hazards Model (PHM), asset-intensive industries often monitor vital covariates to predict failure rate, the reliability function, and maintenance decisions. This analysis requires defining the transition probabilities of asset conditions evolving among states over time. When only one covariate is assessed, the model’s parameters are commonly obtained from expert opinions to provide state bands directly. However, the challenge lies within multiple covariate problems, where arbitrary judgment can be difficult and debatable, since the composite measurement does not represent any physical magnitude. In addition, selecting covariates lacks procedures to prioritize the most relevant ones. Therefore, the present work aimed to determine multiple covariate bands for the transition probability matrix via supervised classification and unsupervised clustering. We used Machine Learning (ML) to strengthen the PHM model and to complement expert knowledge. This paper allows obtaining the number of covariate bands and the optimal limits of each one when dealing with predictive maintenance decisions. This novel proposal of an ML condition assessment is a robust alternative to the expert criterion to provide accurate results, increasing the expectation of the remaining useful life for critical assets. Finally, this research has built an enriched bridge between the decision areas of predictive maintenance and Data Science.

Maintenance cost prediction for the vehicle based on maintenance data

With the fierce competition in the automobile market, the focus of competition in the automobile industry had gradually turned to the automobile extended warranty service. The prediction of maintenance cost was a very important premise for the formulation of automobile extended warranty service. Combining the failure frequency data of vehicles in multiple sales batches, the mixed Weibull model was used to fit the failure process, and the single vehicle failure rate prediction model was obtained; At the same time, combined with the maintenance cost data of the same batch of vehicles, the prediction model of single vehicle maintenance cost was obtained by iterative solution. Using the existing maintenance data, the maintenance cost prediction model based on user group had been verified. The results showed that the model was real, effective and had strong engineering application value.

Fault Rate Estimation of Electric Energy Meter Based on WDPOD-NHBM

To solve the problems of the small amount of sample information in the original failure rate data of smart energy meters and the difficulty of mining under typical environmental stress, a smart energy meter failure rate prediction and reliability evaluation model based on double precision combined with nonlinear hierarchical Bayes is proposed. First, the KNN algorithm is used to test the outliers in the original data, and a double-precision outlier detection algorithm is further established to assign the weight to outliers. Then, a nonlinear hierarchical Bayesian model is built based on Weibull distribution, and couples the failure rate data of smart energy meters with various typical environmental stresses. Through multi-dimensional analysis of the data distribution rule, the most prior distribution is selected. Finally, the Markov Chain Monte Carlo method is used to sample and solve the posterior condition distribution of the model parameters. The posterior parameters and confidence interval estimation are obtained, and the reliability of the smart energy meters is calculated as 95%. The experimental results show that the method can accurately predict the trend of the failure rate of smart energy meters over time in a typical environment.

A novel reliability prediction with input transients for an LLC converter

Utility applications demand highly reliable power converters to match market quality needs. The classical reliability prediction methods do not account for the sudden transients involved with the power converter. This work envisages a new reliability prediction procedure for LLC converter which accounts for the input transients’ impact on the failure rate. Experiments are conducted to collect the actual stress values at input transient and fault conditions, which aids to predict the failure rate with more accuracy. The reliability prediction has been performed using the collected experimental data from the component level to the system level at different mission profiles with transient operating conditions. The impact of various faults and transients on the converter failure rate prediction has been clearly projected from the quantitative analysis presented in this paper. To have a clear picture, the effect of reliability with respect to other stress factors like temperature stress, environmental stress, electrical voltage, current, and power stress on failure rates are also compiled and tabulated. Failure rate and Mean Time Between Failures (MTBF) have been calculated for an LLC converter using the experimental data. The proposed reliability model can be used in the design phase to have an optimal design, planning, and operation of a power electronic converter in the field. This enables to reach out power converters with better reliability profile to cater the industrial needs for real time applications. From the results, it is evident that the reliability prediction is more realistic when the input transients are considered using the experimental data.