Failure Statistics Research Articles

On the matter of evaluation of the variation coefficient of the time to failure based on low-level quantiles

In the context of various tasks related to dependability estimation of systems by probabilistic physical methods the most important a priori information that ensures effective solutions is the information on the variation coefficient of the time to failure. Given the low failure statistics, the estimation of the variation coefficient of the time to failure is complicated due to significant sample censoring. In these cases, methods of variation coefficient evaluation with additional a priori information and the method of quantiles are used. The solution of a number of dependability-related tasks that require taking into consideration various failure distributions is significantly simplified if the functions of such distributions are tabulated in the relative operation time and variation coefficient parameters. An effective solution of dependability-related tasks with the use of tables of DN distribution function was first proposed for the parametrization of distribution in parameters x and v, where x is the scale parameter, relative operation time x = at; v is the shape parameter, variation coefficient v = V; a is the average degradation rate. That allowed performing tabulation out of real time, simplifying function tabulation and its use in a number of dependability-related tasks by method of quantiles. The paper analyzed the effectiveness of the method of quantiles in the estimation of the variation coefficient of the time to failure, that is at the same time the shape parameter of the DN distribution, under scarce failure statistics and based on it proposes a new, more effective, method. The method of estimation of the variation coefficient using low and ultralow-level quantiles is based on the behaviour analysis of function ai = f(t) obtained using the method of quantiles. It is considered that the best choice of the a priori value of v is a choice under which the dependence graph ai = f(t) is most accurately described by a straight horizontal line, which is in complete compliance with the hypothesis of constant degradation rate accepted in the context of DN distribution formalization. In cases when the dependence graph ai = f(t) does not easily allow concluding on the best choice of the a priori value v (it is especially difficult to make a choice based on the statistics of first failures), the following formal criterion can be used: the most acceptable a priori value of the shape parameter v lies within the range of values, where the sign of the trend of the average degradation rate (h) in graph ai = f(t) changes. Studies have established that the most significant errors in the estimation of the variation coefficient are associated with first failures. When processing the results of dependability tests it is assumed the first failures in a sample have the lowest information weight, as their occurrence is due to serious defects not detected by final quality inspection of products. The first failures normally “fall out” of the overall statistical pattern, and it is recommended to omit them from further analysis. The proposed method of estimation of the variation coefficient of the time to failure based on ultralow-level quantiles enables – in the context of limited failure statistics, when other methods are inefficient – for sufficiently accurate identification of not only the variation coefficient of the time to failure and DN distribution parameters, but also make conclusions regarding the feasibility and legitimacy of equalization (description) of the considered sample using this diffusion distribution, i.e. it can be used as a kind of criterion of compliance of the empirical failure distribution under consideration with the chosen theoretical dependability model. The described process of finding the truest values of the variation coefficient of the time to failure using the formal criterion can be computerized.

Energy-Efficient Neural Network Acceleration in the Presence of Bit-Level Memory Errors

As a result of the increasing demand for deep neural network (DNN)-based services, efforts to develop hardware accelerators for DNNs are growing rapidly. However, while highly efficient accelerators on convolutional DNNs (ConvDNNs) have been developed, less progress has been made with regards to fully-connected DNNs. Based on analysis of bit-level SRAM errors, we propose memory adaptive training with in-situ canaries (MATIC), a methodology that enables aggressive voltage scaling of accelerator weight memories to improve the energyefficiency of DNN accelerators. To enable accurate operation with voltage overscaling, MATIC combines characteristics of SRAM bit failures with the error resilience of neural networks in a memory-adaptive training (MAT) process. Furthermore, PVT-related voltage margins are eliminated using bit-cells from synaptic weights as in-situ canaries to track runtime environmental variation. Demonstrated on a low-power DNN accelerator fabricated in 65 nm CMOS, MATIC enables up to 3.3× energy reduction versus the nominal voltage, or 18.6× application error reduction. We also perform a simulation study that extends MAT to Conv-DNNs, and characterize the accuracy impact of bit failure statistics. Finally, we develop a weight refinement algorithm to improve the performance of MAT, and show that it improves absolute accuracy by 0.8-1.3% or reduces training time by 5-10×.

Calculation methods for the gas pipeline failure rate

Determination of gas pipeline failure rates is essential for quantitative risk assessment and integrity management of gas pipelines. To this end, the functional relationship between failure rate and failure probability is established by reliability theory, considering that the two terms are often confused in the literature. Then, existing methodologies for the estimation of failure rates are divided into three categories, namely methods based on historical failure data statistics, methods based on expert knowledge and experience, and methods based on structural reliability analysis. Finally, characteristics and applicability of the three methods are compared. We find three essential aspects to consider in determination of the choice of method in practical engineering: (1) operation and management level of gas suppliers, (2) features or merits of detection technologies of pipelines, and (3) integrity of pipeline failure databases. Considering that external interference as a cause of failure belongs to a type of extreme failure load, it would be appropriate to adopt methods based on historical failure statistics to determine failure rates. On the contrary, when historical data are relatively scarce or the pipeline detection technology is advanced, it is advisable to use methods based on structural reliability analysis to obtain the failure rates. In this study, the calculation model and procedures for the method based on historical failure data statistics are established. Furthermore, the so-called bathtub curve of pipeline failure rates is obtained by methods based on structural reliability analysis.

Designing Optimal Middlebox Recovery Schemes With Performance Guarantees

Enabling functionality in a modern network is achieved through the use of middleboxes. Middleboxes suffer from temporal unavailability due to various reasons, such as hardware faults. We design a backup scheme that takes advantage of network function virtualization, an emerging paradigm of implementing network functions in software, deployed on commodity servers. We utilize the agility of software-based systems, and the gap between the resource utilization of active and standby components, in order to design an optimal limited-resource backup scheme. We focus on the case where a small number of middleboxes fail simultaneously, and study the backup resources required for guaranteeing full recovery from any set of failures, of up to some limited size. Via a novel graph-based presentation, we develop a provably optimal construction of such backup schemes. Since full recovery is guaranteed, our construction does not rely on failure statistics, which are typically hard to obtain. Simulation results show that our proposed approach is applicable even for the case of larger numbers of failures.

Comparison of dam failure frequencies and failure probabilities

AbstractRecently the author has compared observed collapse frequencies and computed failure probabilities of bridges. In contrast, in this paper the comparison of observed failure frequencies and the failure probabilities of large dams is carried out.In the paper the failure frequencies are estimated based on various publications related to the failure statistics of dams for different regions and different times. The various publications are considered in terms of a meta‐analysis to provide a robust estimation of the failure frequencies and to identify trends of the failure frequencies. The failure trend analysis considers both; the failure frequency per year and the failure frequency related to the accumulated experience in dam years. The latter enables the comparison to the generalized learning curve. The failure frequencies are compared to the computed probabilities of failure for dams. The comparison is evaluated with regard to the accuracy of the computed values to the observed ones.

EXPERIMENTAL STUDY OF NATURAL FORMS AND THE VIBRATION FREQUENCIES OF THE COMPRESSOR BLADES OF AIRCRAFT AUXULARY POWER UNIT

The article presents the results of the blades vibration studies, where the blades are rigidly fixed to the disk of the high-pressure compressor of the auxiliary power unit (APU). The task of the study is set because of the destruction of the blades cascade, which led to the APU failure. The designated APU resource was not provided. Destruction of compressor blades occurred since the introduction of gas turbine engines, but it was not always possible to comprehensively investigate the problem and completely prevent the destruction of compressor blades. Analysis of fatigue failure statistics led to the conclusion that all the blades, regardless of their size and material are destroyed in certain areas on the back, edges and tip. The relative coordinates of the crack location are approximately the same. This indicates the availability of common patterns of the blades dynamic loading and the resulting tension in them with resonance vibrations. Moreover, the availability of operational damage on the blades does not affect these patterns. However, these patterns are not fully applicable to the blades represented in this study (wide-chord, with variable thickness from the leading to the trailing edge), but they have some similarities. The study of the forms spectrum and the blade frequency vibration is the first step in solving the problem of the blades destruction. Every detail has an infinite number of frequency characteristics that form the spectrum. Each frequency of natural oscillations corresponds to its own form of vibration. The form of the vibrations is determined by the pattern of alternating displacements or deflections and the position of the nodal lines. In nodal lines the displacements are infinitely small and considered null. The displacements of the vibrating sections on different sides of the nodal lines are in antiphase. The combination of the vibration shapes forms a variety of shapes. In this study, a sand method for determining the spectrum of vibration frequencies was used. This experiment will allow to define the sectors of defects occurrence with the subsequent development of fatigue cracks and to define real dynamic loading of the given blades design.

An efficient method with tunable accuracy for estimating expected interruption cost of distribution systems

The paper presents a novel efficient method with tunable accuracy for estimating expected interruption cost (ECOST) of distribution systems. ECOST index quantifies the reliability of a distribution system in monetary basis. The performance of ECOST estimation could be influenced by various factors such as time-varying load and cost models, computational limitation and random nature of component failure and repair time. Generally, the interruption cost is estimated based on an analytical method which does not consider the input and parameter uncertainties that are represented as random variables. The simulation method based on Monte Carlo (MC) simulation could provide a more accurate approximation of ECOST due to consideration of stochastic factors. However, one basic challenge related to the MC method is the high computational cost in order to run a large number of iterations for a specified high accuracy. Speed up the accurate estimation process using fast computation method could be an important feature in distribution systems management software. An advancement of the MC method with controllable accuracy is the Multilevel Monte Carlo (MLMC) estimator which is proposed to estimate the system ECOST. The proposed method could reduce the huge computational cost needed for accurately estimating the index. To illustrate the performance of this method, five different size distribution systems of Roy Billinton Test System are utilized. The impacts of the network topology, customer load type, time-varying load and cost models, failure and repair statistics on the MLMC based system ECOST assessment performance are also investigated.

Investigation Into Modeling Australian Power Transformer Failure and Retirement Statistics

Analyzing the probability of power transformer failure is challenging because to observe a sufficient number of failures requires a fleet size larger than what is often operated by one utility. Parametric distributions have been applied to small datasets to model failure. However, an assumption is that the failure follows the chosen distribution. Reliability data from 97% of the 6057 utility-owned power transformers operating in mainland Australia and Tasmania were collected and analyzed to develop a more accurate understanding of failure statistics. Information on 564 failures and retirements occurring between 2000 and 2015 was collected, stratified into distribution, subtransmission, and transmission voltage class units, and then analyzed. The useful life of the three different populations was estimated, and the relationship between age and different failure types was investigated.

SALT CONDENSATE рН EFFECT ON PITTING CORROSION OF ALUMINUM ALLOYS

Background The late stages of oil fields development include the technologies increasing oil recovery, and therefore oil production is carried out in conditions of its high water cut, which leads to increased equipment corrosion at production, preparation, storage, transport and processing facilities oil. Technological fluids high aggressiveness is associated with the presence of dissolved gases, mechanical impurities, salt ions, water-soluble naphthenic and mineral acids and bases, as well as microorganisms whose life-activity products cause active biocorrosion of the metal. According to statistics of oil and gas equipment failures and destruction, it can be noted that corrosion of metal and the formation of its products, contributing to the transported raw materials contamination, represent the main complications that significantly reduce its resource. Aims and Objectives Impact assessment of corrosive environment pH on the pitting corrosion development in formation water. Results Studies have shown that an increase in the pH of a water-salt solution saturated with chlorine ions promotes the substantial shift of the aluminum alloy surface stationary electrode potential (up to 900 mV) in the negative direction. This contributes to the cathodic corrosion reactions enhancement and the breakdown of passive films. Passive film breakdown causes the pitting corrosion development. According to the forecast, if the pH is increased to 14, the corrosion rate can be increased to 17 mm/year. Characteristics and composition of alkaline salt condensate require constant monitoring. With significant local increase in the condensate pH, it is necessary to protect the aluminum alloy surface with paint and varnish. Taking into account the study results and following the recommendations made will significantly increase the life of pontoons, reservoirs and other structures during their operation in the oil and gas industry.

Stability in a fiber bundle model: Existence of strong links and the effect of disorder.

The present paper deals with a fiber bundle model which consists of a fraction α of infinitely strong fibers. The inclusion of such an unbreakable fraction has been proven to affect the failure process in early studies, especially around a critical value α_{c}. The present work has a twofold purpose: (i) a study of failure abruptness, mainly the brittle to quasibrittle transition point with varying α and (ii) variation of α_{c} as we change the strength of disorder introduced in the model. The brittle to quasibrittle transition is confirmed from the failure abruptness. On the other hand, the α_{c} is obtained from the knowledge of failure abruptness as well as the statistics of avalanches. It is observed that the brittle to quasibrittle transition point scales to lower values, suggesting more quasi-brittle-like continuous failure when α is increased. At the same time, the bundle becomes stronger as there are larger numbers of strong links to support the external stress. High α in a highly disordered bundle leads to an ideal situation where the bundle strength, as well as the predictability in failure process is very high. Also, the critical fraction α_{c}, required to make the model deviate from the conventional results, increases with decreasing strength of disorder. The analytical expression for α_{c} shows good agreement with the numerical results. Finally, the findings in the paper are compared with previous results and real-life applications of composite materials.

On power-law tail distribution of strength statistics of brittle and quasibrittle structures

Understanding the tail behavior of the probability distribution of structural strength is of paramount importance for reliability-based design of engineering structures. For brittle and quasibrittle structures, the tail distribution not only determines the design strength at a low failure probability, but also governs the functional form of the strength distribution of large-size structures. There exists clear evidence that the Weibull distribution is applicable to brittle structures. Based on the theory of extreme value statistics, the applicability of Weibull distribution suggests that the tail distribution must follow a power law. The justification of the power-law tail distribution was first proposed by Freudenthal based on an assumed type of flaw statistics and linear elastic fracture mechanics of non-interacting flaws. A series of recent studies suggested that the power-law tail can be explained by the transition rate theory governing the statistics of material failure at the nano-scale. In this study, we investigate the tail distribution by considering the randomness in both material strength and applied stress field. The present analysis adopts a nonlocal strength-based failure criterion, in which the spatial variability of the material strength is represented by an autocorrelated random field. The failure statistics of the structure is calculated as a first passage probability. We analyze this problem in 1D, 2D and 3D settings, and the results indicate that in all cases the tail distribution of the nominal structural strength follows a power law. It is shown that the power-law tail behavior of strength distribution stems from the tail distribution of material strength. The flaw statistics introduces additional randomness to the nominal structural strength, but does not dictate the power-law form of its tail distribution.

Thermal Ageing of XLPE Cable Insulation under Operational Temperatures – Does It Exist?

Thermal ageing of power cables may cause premature breakdown if the cables are operated at elevated temperatures, a relation usually described by Arrhenius’ law. The purpose of the present study is to investigate if any significant ageing takes place under normal operational temperatures. The investigation shows that thermal degradation at these comparably low temperatures is hardly reported and seems not to be of practical concern to transmission system operators. This may be the result of a conservative mind-set during the dimensioning of power systems, here in the case of power cables, which means that the lines are generally lightly loaded. Based on the literature found or, to be more precise, on the literature not found it seems to be very difficult to determine that thermal ageing should be problematic for operational power cables at all, in particular as compared to laboratory investigations at very high temperatures, where thermal ageing clearly takes place. To this conclusion contributes the fact that failure statistics do not correlate failures with severity of loading, and that the laboratory studies (which mosly conclude thermal ageing to be problematic) exposes the materials to unrealistic conditions.

Influence of failures of overhead lines on reliability of autonomous power supply system

We research influence of failures of overhead lines on reliability of autonomous power supply system with distributed generation in the given article. Simulation of failures of overhead lines are done on the base of long-term data of exploitation distributing electricity networks. Considered autonomous power supply system consists of wind and solar power stations. Diesel generators are emergency sources of electricity in the system. For adequacy estimation of electricity supply we many times simulated and analyzed of random states of autonomous power supply system. Results of experimental research confirm that detailed accounting of statistics of real failures of overhead lines increases of quality of reliability analysis.

Approach to detecting latent defects of the printed circuit boards thin-filmed coatings

One of the main reasons for the space vehicles (SV) power supply systems failure is a secondary arc. In the statistics of fatal failures that can lead to the SV loss, the failure rate of the power supply systems is 9%, but despite this, there are no standard methods for diagnosing the on-board equipment for the arc stability, due to the relative novelty of the problem. The problem we encountered is associated with the need to switch power supply systems to the voltage levels that are much higher than the arc voltage in the space vacuum. From the point of view of the defects downing that can lead to arcing, the most vulnerable are the insulation coatings. Therefore, a significant part of the problem solution is concentrated in the development of the specific methods for these coatings diagnostics. In space technology, such diagnostics significance is increased due to the inability to provide necessary repairs or services in the outer space.

ПОВЫШЕНИЕ ЭФФЕКТИВНОСТИ ПРОЦЕССОВ В ДИЛЕРСКО-СЕРВИСНЫХ ЦЕНТРАХ С ПОМОЩЬЮ ТЕХНОЛОГИИ DIGITAL TWIN

High competition among car manufacturers requires modifying business models and digitalising not only production processes, but also service processes. The use of modern digital and hardware means allows to monitor the technical condition of vehicle parts and systems, to collect, store and analyze vehicle failure statistics, data on vehicle operation conditions, mileage, as well as data on the performed maintenance and repair works. This will allow to plan the loading of service stations in advance, predict the need for spare parts, as well as notify owners of vehicles about the need to visit dealers and service centres. To achieve this, we suggest an Intelligent System of Failure Prediction and Planning of Vehicle Visits to the Dealer and Service Centres based on the Digital Twin technology.

Study on the Failure Mechanism of the Polymorphic Mixture for Remanufactured Machinery Parts

The polymorphic mixture failure mode for multiple heterogeneity of remanufactured (RM) machinery parts makes it difficult to assess their lifetime. The Weibull distribution failure model of RM parts (substrate, coating layer, bonding surface and sudden failure) is constructed with failure time statistics of the parts in service, the latter is used to characterize the failure patterns of RM parts. In view of the multiple heterogeneity of RM parts, the Kaplan–Meier type decoupling method is used to analyze four sets of failure statistics, and each state of the Weibull failure function of the above parts is solved. It reveals the in-service failure mechanism of polymorphic mixtures for multiple heterogeneity of RM machinery parts. The validity and feasibility of the model are verified by the case study. Research results provide the theoretical basis for the design and preparation of a RM alloy powder and the improvement of RM technology. Moreover, the method for lifetime prediction and failure time evaluation of RM parts is proposed and validated.

Bayesian information fusion for probabilistic health index of power transformer

This study proposes a Bayesian information fusion approach for determining the probabilistic health index of power transformer. The proposed approach integrates a variety of data obtained from transformer measurements, maintenance records, and failure statistics. By making use of these data, an inference model is constructed using Bayesian belief network (BBN). In the inference model, the significance of each individual measurement on the corresponding component in the BBN is decided by both principal component analysis and expert's experience, and subsequently quantified with a score-probability transform. Finally, the inference model is used to derive a probabilistic health index. Case studies are provided to demonstrate the applicability of the proposed approach for evaluating transformer condition.

IMPROVING THE RELIABILITY OF POWER TRANSFORMERS OLTC DEVICES

It is established that one of the effective ways to improve the reliability of the power transformer as a whole is to increase the reliability of its on-load tap-changer (OLTC). As a result of the analysis of statistical material obtained under real operating conditions (the number of transformers is 1000 pieces, the observation period is 4 years), the failure statistics of its individual elements are given. It is shown that the most not reliable elements of the OLTC are the drive and the contactor. To improve the reliability of the OLTC, the possibility of its development on the basis of an engine with a rolling rotor is considered. A critical analysis of modern OLTC and the proposed technical solution is carried out. The technical and economic feasibility of developing and implementing a new device has been established. On the basis of the general theory of the electromagnetic field, a technique and an algorithm for calculating new sizes of an engine with a rolling rotor for OLTC of power transformers has been developed. The analysis of operational reliability of OLTC based on statistical materials is carried out; the main reliability indices of its elements are calculated. It is established that the introduction of a new technical solution will increase the reliability of the OLTC by 25% and the power transformer by 2-4%. The perspective of the development of a new OLTC device is presented.

Reliability assessment platform for the power semiconductor devices – Study case on 3-phase grid-connected inverter application

Because of the high cost of failure, the reliability performance of power semiconductor devices is becoming a more and more important and stringent factor in many energy conversion applications. Thus, the need for appropriate reliability analysis of the power electronics emerges. Due to its conventional approach, mainly based on failure statistics from the field, the reliability evaluation of the power devices is still a challenging task. In order to address the given problem, a reliability assessment platform is proposed in this paper. An advanced reliability design tool software, which can provide valuable reliability information based on given mission profiles and system specification is first developed and its main concept is presented. In order to facilitate the test and access to the loading and lifetime information of the power devices, a novel mission profile based stress emulator experimental setup is proposed and designed. The link between the stress emulator setup and the reliability tool software is highlighted. Finally, the reliability assessment platform is demonstrated on a 3-phase grid-connected inverter application study case.

HUMAN FACTOR INFLUENCE ON TRAIN BRAKE EQUIPMENT RELIABILITY

[For the English abstract and full text of the article please see the attached PDF-File (English version follows Russian version)].ABSTRACT The article considers the human factor influence on trouble-free operation of brake equipment of trains. The study was carried out by an analytical method, based on the statistics of equipment failures, assessment of implementation of the current rules for railway rolling stock maintenance. The problem of the lack of an effective device for diagnosing the brake network of a train, which would provide control over its integrity and density in the course of traffic and during stops, is revealed. At the same time, it is also necessary to reduce time for measuring density of the brake network, the train to automate this process. Keywords: railway, traffic safety, braking equipment, train brake network density, reliability, human factor.