Lack Of Failure Data Research Articles

The Reliability Modeling and Evaluation of a Cusped Field Thruster When Undertaking a Gravitational Wave Detection Mission

The propulsion system, particularly electric propulsion, holds immense significance in the context of gravitational wave detection missions. One of the key factors of a deep space exploration mission is the lifetime of the electric propulsion. Ensuring the high reliability of the propulsion system is of paramount importance; however, achieving this is challenging in the absence of adequate failure data. Conducting ground tests for a thruster tends to encounter two limitations: a lack of failure data and time constraints. To address these challenges, we propose a semi-physics sputtering method that combines a physical erosion model with empirical processes. In this study, we focus on evaluating the lifespan of a cusped field thruster (CFT) for potential application in gravitational wave detection missions. Our analysis revolves around modeling non-conservative forces in a space environment and examining their impact on a thruster’s longevity. The results indicate that, in gravitational wave missions, the survival rate of a thruster’s lifespan at 8000 h is 0.75. At a constant voltage of 500 V, the maximum corrosion depth after 5000 h is 3.1 mm, while the minimum is 0.49 mm.

Delamination Detection Framework for the Imbalanced Dataset in Laminated Composite Using Wasserstein Generative Adversarial Network-Based Data Augmentation

As laminated composites are applied more commonly, Prognostics and Health Management (PHM) techniques for the maintenance of composite systems are also attracting attention. However, applying PHM techniques to a composite system is challenging due to the data imbalance problem from the lack of failure data and unpredictable failure cases. Despite numerous studies conducted to address this limitation, including techniques like data augmentation and transfer learning, significant challenges remain. In this study, the Wasserstein Generative Adversarial Network (WGAN) model using a time-series data augmentation technique is proposed as a solution to the data imbalance problem. To ensure the performance of the WGAN model, time-series data augmentation of experimental data is executed with a frequency analysis. After that, a One-Dimensional Convolutional Neural Network (1D CNN) is used for fault diagnosis in laminated composites, validating the performance improvement after data augmentation. The proposed data augmentation significantly elevated the performance of the 1D CNN classification model compared to its non-augmented counterpart. Specifically, the accuracy increased from 89.20% to 91.96%. The precision improved remarkably from 29.76% to 74.10%, and its sensitivity rose from 33.33% to 94.39%. Collectively, these enhancements highlight the vital role of data augmentation in improving fault diagnosis performance.

Availability analysis of safety-critical systems of nuclear power plant using ordinary differential equations and reachability graph

With the growing advancement and demand for technology, computer-based systems (CBS) have become an integral part of human life. Numerous CBSs are safety and mission-critical in nature, and are employed in various industries, including nuclear power plant (NPP), locomotive control, avionics, medical automation, etc. Safety-critical systems (SCS) are one such CBS that are vital to control and maintain the infrastructure of NPP. This paper presents an inventive technique to assess the performance metric availability of SCSs. A Petri net (PN) models such systems, which feature multiple processing nodes interacting with one another. PN models are useful for generating the reachability graph. This article uses reachability graph to derive a collection of ordinary differential equations (ODEs), whose solution can be applied for assessing the availability of the system. Conventional methods like Markovian chains, Reliability Block diagrams (RBDs), Fault Tree Analyses (FTAs), and Flow Networks fail to cover the systems' behaviors and structures properties, as well as the lack of failure data, and the diversity of possible failures. The suggested technique has been applied to Digital Feed Water Control System (DFWCS) of NPP, which consider the failure, maintenance and repairment of the main-steam safety valves. We achieved 99.20% accuracy of availability measurement, proving the efficacy of methodology.

Predictive maintenance using digital twins: A systematic literature review

Predictive maintenance is a technique for creating a more sustainable, safe, and profitable industry. One of the key challenges for creating predictive maintenance systems is the lack of failure data, as the machine is frequently repaired before failure. Digital Twins provide a real-time representation of the physical machine and generate data, such as asset degradation, which the predictive maintenance algorithm can use. Since 2018, scientific literature on the utilization of Digital Twins for predictive maintenance has accelerated, indicating the need for a thorough review. This research aims to gather and synthesize the studies that focus on predictive maintenance using Digital Twins to pave the way for further research. A systematic literature review (SLR) using an active learning tool is conducted on published primary studies on predictive maintenance using Digital Twins, in which 42 primary studies have been analyzed. This SLR identifies several aspects of predictive maintenance using Digital Twins, including the objectives, application domains, Digital Twin platforms, Digital Twin representation types, approaches, abstraction levels, design patterns, communication protocols, twinning parameters, and challenges and solution directions. These results contribute to a Software Engineering approach for developing predictive maintenance using Digital Twins in academics and the industry. This study is the first SLR in predictive maintenance using Digital Twins. We answer key questions for designing a successful predictive maintenance model leveraging Digital Twins. We found that to this day, computational burden, data variety, and complexity of models, assets, or components are the key challenges in designing these models.

Polymer Gear Fault Classification Using EMD-DWT Analysis Based on Combination of Entropy and Hjorth Features

Polymer gears have proven to be an adequate replacement for traditional metal gears in various applications. They are lighter, have less inertia, and are much quieter than their metal counterparts. Polymer gears, however, are rarely employed because there is a lack of failure data. Hence, there is tremendous scope for fault detection of polymer gears. In this paper, a novel technique of polymer gear fault detection is proposed following the double decomposition of vibration signals. The experimentally acquired vibration signals are processed through two steps of decomposition, i.e., empirical mode decomposition and discrete wavelet transform based Time-Frequency decomposition. Subsequently, entropy features (EF), Hjorth parameter (HP), and a combination of EF and HP are extracted. A combination of these feature sets is used to train the classifier: support vector machine (SVM), ensemble learning, and decision tree. Among all classification methods, the ensemble learning classifier reached the maximum classification accuracy of 99.2 % using a combination of EF and HP features. Furthermore, EMD and DWT are compared with the proposed double decomposition method (EMD-DWT) for accuracy validation. The experiments demonstrated that the proposed EMD-DWT method is efficient and yields promising results for classifying polymer gear faults.

A Weakly Supervised Learning-Based Oversampling Framework for Class-Imbalanced Fault Diagnosis

With the lack of failure data, class imbalance has become a common challenge in the fault diagnosis of industrial systems. The oversampling methods can tackle the class-imbalanced problem by generating the minority samples to balance the training set. However, one of the main challenges of the existing oversampling methods is how to generate high-quality minority samples. Traditional oversampling methods regard all synthetic samples as minority ones to be added to the training set without filtering. The low-quality synthetic samples would distort the distribution of the dataset and worsen the classification performance. In this article, we propose a weakly supervised oversampling method that treats all synthetic samples as unlabeled samples and develops a graph semisupervised learning algorithm to select high-quality synthetic samples, adding into the final training set as minority samples. To improve the quality of synthetic samples, we propose a cost-sensitive neighborhood component analysis dimensionality reduction method to enhance domain information validity in high-dimensional datasets. Finally, combining a boosting-based ensemble framework, we propose a new imbalanced learning framework suitable for high dimensionality and highly imbalanced fault diagnosis in industrial systems. The experimental validation is performed on five real-world wind turbine blade cracking failure datasets and compared to 15 benchmark methods. The experimental results show that average performances and robustness of the proposed framework are significantly better than those of the benchmark methods.

A comparative study of different algorithms using contrived failure data to detect robot anomalies

Unexpected downtime remains a costly, preventable burden in manufacturing. To mitigate and eliminate unexpected downtime, manufacturers have incorporated machine learning to diagnose equipment faults and determine the equipment remaining useful life. However, such models suffer from a lack of failure data and context knowledge surrounding data gathered from production (i.e. rich labeled sets). The purpose of this paper is to conduct an algorithm comparison study over a previously collected contrived anomaly data set from an industrial robot. The goal of the study is to measure the effectiveness of algorithms to eliminate bias and variance from the classification results. The tested system is a 6-DOF collaborative robot from Universal Robots, on which an anomalous condition is artificially induced on a robot to simulate robot overload. The different algorithms are assessed based on their accuracy in determining the overloading case on the robot. From the analysis of data-driven machine learning and deep learning models, a deep learning regression was determined as the best model from the assessment of the data both qualitatively (low overfitting and bias) and quantitatively (98% overall accuracy). As part of the future research direction, different failure cases should be created based on wear applied to specific robot joint components and an adaptability assessment carried out for the model to other robot paths.

Equipment Operational Reliability Evaluation Method Based on RVM and PCA-Fused Features

Reliability assessment is of great significance in ensuring the safety and reducing maintenance cost of equipment. The traditional statistical method is widely used to estimate the reliability of mass equipment; however, it cannot efficiently predict the overall reliability of single or small batch equipment due to lack of failure data. This paper introduced the operational reliability concept to describe the running condition of single or small batch equipment and proposed a method based on the combination of Relevance Vector Machines (RVMs) and Principal Component Analysis (PCA) to evaluate the operational reliability. Some representative characteristic indexes of operating equipment were firstly selected, and PCA was applied to obtain a hybrid index of the equipment’s running condition. Then, a RVM prediction model was trained to predict the development of the hybrid index and corresponding probability density function (PDF). Based on this, the operational reliability of the equipment was calculated by the interval integral defined by the failure threshold and the predicted value of the hybrid index. The approach was validated using the experimental test conducted on the aero-engine rotor bearings. The results show a good agreement in the evaluations of the failure time between the proposed method and the experimental test.

An expert-based method for the risk analysis of functional failures in the fracturing system of unconventional natural gas

This paper proposes an expert-based method to analyze functional failure risk in the fracturing system of unconventional natural gas. The newly proposed Fuzzy-FFIP method integrates functional failure identification and propagation (FFIP) with fuzzy logic. Fuzzy-FFIP is sufficient for dealing with the complexity and correlation of functional failures and the lack of failure data, as well as capturing the fuzziness of functional states of a critical component and entire system. From a configuration-behavior-function aspect, FFIP is a common framework used to represent system-wide logical relationships and study functional failure propagation paths. This framework is an expert system where knowledge base and inference engine are based on behavioral rules (BRs) and functional failure logic (FFL), respectively. Using FFL as a support, fuzzy logic is specifically applied to analyze the fuzzification and defuzzification of functional states – operating, degraded, and lost. The proposed method can clearly provide functional failure modes, effectively reveal failure propagation paths, and quantitatively assess functional failure risk levels in the fracturing system. To illustrate its validity, an on-site pump lubricant subsystem of fracturing unit is selected as a test case. Results show that Fuzzy-FFIP is more detailed and accurate, and contributes to system safety during the whole fracturing period.

A two-phase model to predict the remaining useful life of corroded reinforced concrete beams

A two-phase modeling and estimation framework to structural health monitoring of infrastructure components is presented in this paper. The main contribution of this paper is to present a mathematical and implementation framework to detect failures and to predict the remaining useful life in long-life infrastructure components (e.g., bridge components), where actual failure data may be unavailable. A hybrid approach where laboratory data are fused with in-situ measurements is proposed to overcome the difficultly of data unavailability. Taking into account practical constraints related to field measurements, surrogate measures are used as an indirect means to model the mechanism of degradation in the component(s) of interest. The model proposed for degradation is a two-phase gamma process which can handle cases where an event causes a change in the rate of degradation (e.g., stiffness loss). A Bayesian approach is used to estimate the model parameters and the residual useful life. The key step of determining the priors is accomplished using the laboratory test data, which is subsequently integrated with data from a monitored specimen for posterior inference. This hybridization overcomes some of the difficulties associated with the lack of failure data in long-life components, typical of civil infrastructure. The methodology is demonstrated using measurement data acquired from near full-scale laboratory corroded reinforced concrete beam specimens. For the sake of completeness, the results are compared with the single-phase degradation modeling approach, and the practical advantages of using a two-phase model are underscored.

Bearing Procurement Analysis Method by Total Cost of Ownership Analysis and Reliability Prediction

In making bearing procurement analysis, price and its reliability must be considered as decision criteria, since price determines the direct cost as acquisition cost and reliability of bearing determine the indirect cost such as maintenance cost. Despite the indirect cost is hard to identify and measured, it has high contribution to overall cost that will be incurred. So, the indirect cost of reliability must be considered when making bearing procurement analysis. This paper tries to explain bearing evaluation method with the total cost of ownership analysis to consider price and maintenance cost as decision criteria. Furthermore, since there is a lack of failure data when bearing evaluation phase is conducted, reliability prediction method is used to predict bearing reliability from its dynamic load rating parameter. With this method, bearing with a higher price but has higher reliability is preferable for long-term planning. But for short-term planning the cheaper one but has lower reliability is preferable. This contextuality can give rise to conflict between stakeholders. Thus, the planning horizon needs to be agreed by all stakeholder before making a procurement decision.

Reliability Assessment for Telecommunication Channel of Electric Power System Using AHP

The reliability assessment is a key function that should be performed in advance of any security deployment. Since the reliability of electric power communication channel is influenced by many factors, modification factor set for telecommunication channel of electric power system is needed. According to past research, the confidence for equipment reliability would decline caused by the lack of failure data. Therefore, the paper modified each initial index calculated only by the failure data to obtain the reliability. Comprehensive modification factor is obtained according to the weight of each index calculated with Hierarchy Process whose computing standard is the importance of the modification factor in the actual operation. Finally, an instance communication network of power control system is modeling to test the validity of reliability assessment. The result supports the usefulness of the reliability assessment.

Data Requisites for Transformer Statistical Lifetime Modelling—Part I: Aging-Related Failures

Statistical lifetime models are regarded as an important part of the replacement management of power transformers. The development of transformer lifetime models, however, is hindered by the lack of failure data since most of the transformer fleets have not yet completed their first lifecycle. As researchers realized the importance of survival data, lots of lifetime models are developed based on failure data together with survival data. This paper analyzes the effect of survival data on the accuracy of lifetime models through a series of Monte Carlo simulations. It has been proved that the accuracy of lifetime models can be improved by taking the survival data into account. However, the degree of improvement is greatly confined by the censoring rate and the sample size of the collected lifetime data. Practical implications of the simulation results and suggestions on measures to further improve the accuracy of lifetime models are subsequently provided.

Key Subsystem Identification of the CNC Machine Tools

In order to find the key subsystems which affect the reliability of the CNC machine tools is an important and yet harder problem due to the lack of failure data and statistical analysis. A novel method based on posterior probability function for identifying the key subsystems of the CNC machine tools was proposed. Firstly, subsystem reliability of the CNC machine tools was estimated using the John estimator, since the failure data were censored. Secondly, parametric fitting was conduct with Weibull distribution using the maximum likelihood estimation (MLE) or the least square estimation (LSE). Thirdly, the probability function of each subsystems were attained by the reliability modes of the systems and the whole system. Finally, a case study was given, which proved that the proposed method could find out the key subsystems rapidly and accurately.

Research opportunities in maintenance of office building services systems

PurposeThe purpose of this study is to highlight special characteristics of building services systems and investigate how practitioners view reliability and maintenance. These characteristics include energy‐hungry services systems, operating modes, maintenance types, the relationship between procurement costs and maintenance costs.Design/methodology/approachThe practitioners' viewpoints on reliability and maintenance are explored through a workshop. The authors wish to draw the attention of researchers in the reliability and maintenance community and furthermore emphasise the difference between building services systems and systems in industries other than construction.FindingsIt is shown that a lack of failure data and maintenance data is the main problem from both academic researchers' and industrial practitioner's points of view. The paper suggests that there exists no fixed cost ratio available to apply to building services systems; the analysis of RAMS (Reliability, Availability, Maintainability and Safety) should include duty cycles and the environment; and clients of the construction industry would benefit from mandating a LCC to be applied to the build.Practical implicationsThe gap between academia and practitioners should be bridged through better understanding each other's needs. Accurately estimating the ratio between procurement and maintenance costs is needed from a whole life costing perspective.Originality/valueThis paper is a good reference for building designers, facility managers and maintenance staff of building services systems. It also offers reliability researchers references on special characteristics of building services systems.

An update of the probabilistic safety assessment of the Radiation Monitor Calibration Laboratory of a Brazilian nuclear power station

In 2001, a probabilistic safety assessment of the Radiation Monitor Calibration Laboratory of a Brazilian nuclear power station was performed to estimate its radiological risk. Owing to the lack of data, many conservative assumptions were considered. The purpose of this paper is to obtain the updated laboratory radiological risk. A human reliability analysis of the procedure for calibrating the monitors was performed by applying the Technique for Human Error Rate Prediction (THERP) methodology. For facing the lack of failure data, the Bayesian approach was used for the calculation of the irradiator failure probability. The exponential distribution was used to model the failure times of the electronic dosimeters. The radiological risk was obtained through an event tree, in which all relevant scenarios were considered. Although more accident scenarios were considered in the analysis, like external events, the approach was by no means conservative, and a radiological risk about one order of magnitude smaller was obtained.