Search NASA⌕ Search

SEARCH · Search NASA

Results for “reliability modeling”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Reliability Growth Modeling and Testing

Reliability growth has been modelled as an exponential decline in the cumulative failure rate that continues indefinitely as long as testing continues. Contrary to this, most reliability growth data show a brief high initial failure rate due to infant mortality followed by a long period of constant low failure rate. A two part failure rate model with an initial exponential decline followed by a constant failure rate usually fits the data and provides a more realistic description of reliability growth. The reliability growth process consists of testing, experiencing failures, finding the failure causes, and redesigning the system to remove them. The cost of reliability growth increases with the number of inherent failure modes and the time needed for them to occur and be removed. The failure modes with the lower failure rates will tend to occur later, as their Mean Time Before Failure (MTBF) is the inverse of the failure rate. Reliability growth testing has diminishing returns, since it takes longer to find and remove the less probable failures.This paper first discusses the reliability bathtub curve and then explains that reliability growth is produced by testing, identifying failure causes, and designing to remove them. A simple model of reliability growth is introduced, with a brief group of early failures followed by a constant failure rate. The cumulative failure rate n(t)/t can decline as rapidly as1/t or t-1butdeclines more slowly if additiona lfailures occur. The 56-failure Crow data seti s used to demonstrate the two-phase model of reliability growth followed by a constant failure rate. 13 additional data sets are modeled, with 9 of the 14 data sets showing reliability growth approximately as n(t)/t =1/t or t-1and substantial final failure rates. The model fits most of the data sets, but 4of the 14 show no reliability growth. The reliability growth period typically includes six failures and extends one-quarter or half the total test time. As reliability growth testing continues, the cumulative failure rate should be tracked to estimate the reliability growth exponent and the final failure rate.

reliability growth modeling↗
NASA NTRS1990

Surrogate oracles, generalized dependency and simpler models

Software reliability models require the sequence of interfailure times from the debugging process as input. It was previously illustrated that using data from replicated debugging could greatly improve reliability predictions. However, inexpensive replication of the debugging process requires the existence of a cheap, fast error detector. Laboratory experiments can be designed around a gold version which is used as an oracle or around an n-version error detector. Unfortunately, software developers can not be expected to have an oracle or to bear the expense of n-versions. A generic technique is being investigated for approximating replicated data by using the partially debugged software as a difference detector. It is believed that the failure rate of each fault has significant dependence on the presence or absence of other faults. Thus, in order to discuss a failure rate for a known fault, the presence or absence of each of the other known faults needs to be specified. Also, in simpler models which use shorter input sequences without sacrificing accuracy are of interest. In fact, a possible gain in performance is conjectured. To investigate these propositions, NASA computers running LIC (RTI) versions are used to generate data. This data will be used to label the debugging graph associated with each version. These labeled graphs will be used to test the utility of a surrogate oracle, to analyze the dependent nature of fault failure rates and to explore the feasibility of reliability models which use the data of only the most recent failures.

Wilson, Larry↗
DOE OSTI2016

Photovoltaic Reliability Performance Model v 2.0

PV-RPM is intended to address more “real world” situations by coupling a photovoltaic system performance model with a reliability model so that inverters, modules, combiner boxes, etc. can experience failures and be repaired (or left unrepaired). The model can also include other effects, such as module output degradation over time or disruptions such as electrical grid outages. In addition, PV-RPM is a dynamic probabilistic model that can be used to run many realizations (i.e., possible future outcomes) of a system’s performance using probability distributions to represent uncertain parameter inputs.

Miller, StephenP.↗
NASA NTRS1990

Reliability growth modeling analysis of the space shuttle main engines based upon the Weibull process

The Weibull process, identified as the inhomogeneous Poisson process with the Weibull intensity function, is used to model the reliability growth assessment of the space shuttle main engine test and flight failure data. Additional tables of percentage-point probabilities for several different values of the confidence coefficient have been generated for setting (1-alpha)100-percent two sided confidence interval estimates on the mean time between failures. The tabled data pertain to two cases: (1) time-terminated testing, and (2) failure-terminated testing. The critical values of the three test statistics, namely Cramer-von Mises, Kolmogorov-Smirnov, and chi-square, were calculated and tabled for use in the goodness of fit tests for the engine reliability data. Numerical results are presented for five different groupings of the engine data that reflect the actual response to the failures.

Wheeler, J. T.↗
DOE OSTI2023

Dynamic Model Agnostic Reliability Evaluation of Machine-Learning Models Integrated in Instrumentation & Control Systems

In recent years, the field of data-driven neural network-based machine learning (ML) algorithms has grown significantly and spurred research in its applicability to instrumentation and control systems. While they are promising in operational contexts, the trustworthiness of such algorithms is not adequately assessed. Failures of ML-integrated systems are poorly understood; the lack of comprehensive risk modeling can degrade the trustworthiness of these systems. In recent reports by the National Institute for Standards and Technology, trustworthiness in ML is a critical barrier to adoption and will play a vital role in intelligent systems' safe and accountable operation. Thus, in this work, we demonstrate a real-time model-agnostic method to evaluate the relative reliability of ML predictions by incorporating out-of-distribution detection on the training dataset. It is well documented that ML algorithms excel at interpolation (or near-interpolation) tasks but significantly degrade at extrapolation. This occurs when new samples are "far" from training samples. The method, referred to as the Laplacian distributed decay for reliability (LADDR), determines the difference between the operational and training datasets, which is used to calculate a prediction's relative reliability. LADDR is demonstrated on a feedforward neural network-based model used to predict safety significant factors during different loss-of-flow transients. LADDR is intended as a "data supervisor" and determines the appropriateness of well-trained ML models in the context of operational conditions. Ultimately, LADDR illustrates how training data can be used as evidence to support the trustworthiness of ML predictions when utilized for conventional interpolation tasks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Dynamic Model Agnostic Reliability Evaluation of Machine-Learning Models Integrated in Instrumentation & Control Systems

In recent years, the field of machine learning (ML), specifically neural networks, has grown significantly and has spurred research in its applicability to digital instrumentation and control systems (DI&C). While ML models have shown promise in operational contexts, the trustworthiness of using such algorithms has not been adequately assessed. Failures of ML integrated systems are not well understood, and the lack of comprehensive risk modeling can degrade the trustworthiness in these systems. In recent reports by the National Institute for Standards and Technology (NIST) [1] and the Nuclear Regulatory Commission (NRC) [2], they indicate that trustworthiness in ML is a critical barrier and will play a vital role in the safe, accountable, and secure operation of intelligent systems. Thus, in this work, we demonstrate a dynamic model-agnostic method to quantify the relative reliability of AI/ML predictions by incorporating out-of-distribution (OOD) detection on the training dataset. It is well documented that most ML algorithms excel at interpolation (or near-interpolation) tasks but experience significant performance degradation at extrapolation. The method, referenced as the Laplacian distributed decay for reliability (LADDR), determines the difference between the operational and training datasets which can used to the relative reliability of AI/ML predictions. LADDR is then demonstrated on a feedforward neural network based digital twin used for the prediction of safety significant factors during a loss-of-flow transient. LADDR is used to demonstrate how training data can be used as evidence to support the relative reliability of ML/AI predictions enhancing the overall trustworthiness of the system.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2017

Development of a bayesian belief network model for the software reliability assessment of nuclear digital I and C safety systems - 206

Since the digital instrumentation and control systems are expected to play an important role for the safety systems in nuclear power plants (NPPs), the need has emerged to not only establish a basis for incorporating software behavior into digital I and C system reliability models, but also to quantify the failure probability of the software used in NPP digital protection systems. In this study, a Bayesian belief network (BBN) model is developed to quantitatively assess software reliability by estimating the number of faults in a software program considering its software development life cycle (SDLC). The model structure and parameters are established based on the information applicable to NPP safety-related systems and the evidence used to construct and quantify the BBN model was collected from three stages of expert elicitation. The software failure probability is estimated from the number of residual defects in a software program at the end of SDLC phase. As a case study, the BBN model was applied to quantify the software reliability of a typical digital protection software having the size of 50 function points and having the Medium development and validation and verification (V and V) qualities. The developed model can be applied to estimate the failure probability for both developing and deployed safety-related NPP software, and such results can be used to evaluate the quality of the digital I and C systems in addition to estimating potential reactor risk due to software failure. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS1983

A Bayesian modification to the Jelinski-Moranda software reliability growth model

The Jelinski-Moranda (JM) model for software reliability was examined. It is suggested that a major reason for the poor results given by this model is the poor performance of the maximum likelihood method (ML) of parameter estimation. A reparameterization and Bayesian analysis, involving a slight modelling change, are proposed. It is shown that this new Bayesian-Jelinski-Moranda model (BJM) is mathematically quite tractable, and several metrics of interest to practitioners are obtained. The BJM and JM models are compared by using several sets of real software failure data collected and in all cases the BJM model gives superior reliability predictions. A change in the assumption which underlay both models to present the debugging process more accurately is discussed.

Littlewood, B.↗
NASA NTRS1989

Software reliability growth models dominated by randomness

The Jelinski-Moranda and Geometric models for software reliability failed the consistency test which was proposed. These models were challenged to take data which comes from a process which they have correctly modeled and to make predictions about the reliability of that process. It was found that either model, given data precisely from a process it correctly models, will usually fail to make good predictions. These problems are attributed to randomness in the data used as input to the models and a remedy is indicated for this lack of robustness, namely replication of data.

Shen, Wenhui↗
NASA NTRS1995

Integrating Reliability Analysis with a Performance Tool

A large number of commercial simulation tools support performance oriented studies of complex computer and communication systems. Reliability of these systems, when desired, must be obtained by remodeling the system in a different tool. This has obvious drawbacks: (1) substantial extra effort is required to create the reliability model; (2) through modeling error the reliability model may not reflect precisely the same system as the performance model; (3) as the performance model evolves one must continuously reevaluate the validity of assumptions made in that model. In this paper we describe an approach, and a tool that implements this approach, for integrating a reliability analysis engine into a production quality simulation based performance modeling tool, and for modeling within such an integrated tool. The integrated tool allows one to use the same modeling formalisms to conduct both performance and reliability studies. We describe how the reliability analysis engine is integrated into the performance tool, describe the extensions made to the performance tool to support the reliability analysis, and consider the tool's performance.

Nicol, David M.↗
NASA NTRS1982

A Markov chain model for reliability growth and decay

A mathematical model is developed to describe a complex system undergoing a sequence of trials in which there is interaction between the internal states of the system and the outcomes of the trials. For example, the model might describe a system undergoing testing that is redesigned after each failure. The basic assumptions for the model are that the state of the system after a trial depends probabilistically only on the state before the trial and on the outcome of the trial and that the outcome of a trial depends probabilistically only on the state of the system before the trial. It is shown that under these basic assumptions, the successive states form a Markov chain and the successive states and outcomes jointly form a Markov chain. General results are obtained for the transition probabilities, steady-state distributions, etc. A special case studied in detail describes a system that has two possible state ('repaired' and 'unrepaired') undergoing trials that have three possible outcomes ('inherent failure', 'assignable-cause' 'failure' and 'success'). For this model, the reliability function is computed explicitly and an optimal repair policy is obtained.

Siegrist, K.↗
NASA NTRS1991

Reliability growth models for NASA applications

The objective of any reliability growth study is prediction of reliability at some future instant. Another objective is statistical inference, estimation of reliability for reliability demonstration. A cause of concern for the development engineer and management is that reliability demands an excessive number of tests for reliability demonstration. For example, the Space Transportation Main Engine (STME) program requirements call for .99 reliability at 90 pct. confidence for demonstration. This requires running 230 tests with zero failure if a classical binomial model is used. It is therefore also an objective to explore the reliability growth models for reliability demonstration and tracking and their applicability to NASA programs. A reliability growth model is an analytical tool used to monitor the reliability progress during the development program and to establish a test plan to demonstrate an acceptable system reliability.

Taneja, Vidya S.↗
NASA NTRS2024

The abcd Reliability Growth Model

This paper presents a modification of the well-known Duane-Crow reliability growth model. In the abcd reliability growth model, the initial period of exponential decline of the failure rate in the Duane-Crow model may be followed by a period of constant failure rate. Data often show that an exponential decline in failures is followed by a constant failure rate. If a growth model including only the initial period of exponential decline is applied to increasingly longer failure rate data sets, the data will include longer periods of constant failure rate, and the estimated reliability growth rate will decline from an initially high value down toward zero. Using the Duane-Crow model without extending it to include a possible period of constant failure rate may create the mistaken impression that the initial reliability growth continues forever, but at an ever decreasing rate.

Reliability growth↗
NASA NTRS2024

The abcd Reliability Growth Model

This paper presents a modification of the well-known Duane-Crow reliability growth model. In the abcd reliability growth model, the initial period of exponential decline of the failure rate in the Duane-Crow model may be followed by a period of constant failure rate. Data often show that an exponential decline in failures is followed by a constant failure rate. If a growth model including only the initial period of exponential decline is applied to increasingly longer failure rate data sets, the data will include longer periods of constant failure rate, and the estimated reliability growth rate will decline from an initially high value down toward zero. Using the Duane-Crow model without extending it to include a possible period of constant failure rate may create the mistaken impression that the initial reliability growth continues forever, but at an ever decreasing rate.

Reliability growth↗
DOE OSTI2015

Microgrid Design Toolkit (MDT) Technical Documentation and Component Summaries

The Microgrid Design Toolkit (MDT) is a decision support software tool for microgrid designers to use during the microgrid design process. The models that support the two main capabilities in MDT are described. The first capability, the Microgrid Sizing Capability (MSC), is used to determine the size and composition of a new microgrid in the early stages of the design process. MSC is a mixed-integer linear program that is focused on developing a microgrid that is economically viable when connected to the grid. The second capability is focused on refining a microgrid design for operation in islanded mode. This second capability relies on two models: the Technology Management Optimization (TMO) model and Performance Reliability Model (PRM). TMO uses a genetic algorithm to create and refine a collection of candidate microgrid designs. It uses PRM, a simulation based reliability model, to assess the performance of these designs. TMO produces a collection of microgrid designs that perform well with respect to one or more performance metrics.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

ON THE LANGUAGE OF RELIABILITY: A SYSTEM ENGINEER PERSPECTIVE

In its classical definition, risk is defined by three elements: what can go wrong, what are its consequences and how likely is it to occur. While this definition makes sense in a regulatory based framework to estimate risk associated to power plants (in terms of core damage frequency and large early release frequency), this approach does not provide a useful snapshot of the health of the plant. A possible alternate path can start by redefining the word “risk” to a broader meaning that better reflects the needs of a system health and asset management decision making process. Rather than asking how likely an event can occur (in probabilistic terms), we can ask how far this event is from occurring. We will show how, given the data available from plant equipment reliability and monitoring/diagnostic/prognostic centers, a margin can be described and determined for all type of maintenance approaches (e.g., corrective or predictive maintenance). We will show how to link SSC margin-based reliability models to system reliability models (i.e., fault trees) in order to assess system/plant health and how to perform margin-based system calculations. These calculations are not solved using classical probabilistic calculations applied to sets (as performed by any PRA code) but, instead, through metric spaces operations (i.e., distance/margin based approach).

97 - MATHEMATICS AND COMPUTING↗