Content uploaded by Gennadii Manko
Author content
All content in this area was uploaded by Gennadii Manko on Mar 05, 2021
Content may be subject to copyright.
A preview of the PDF is not available
Теорія і практика ідентифікації об'єктів управління : Монографія
ResearchGate has not been able to resolve any citations for this publication.
Model verification and validation (V&V) is an enabling methodology for the development of computational models that can be used to make engineering predictions with quantified confidence. Model V&V procedures are needed by government and industry to reduce the time, cost, and risk associated with full-scale testing of products, materials, and weapon systems. Quantifying the confidence and predictive accuracy of model calculations provides the decision-maker with the information necessary for making high-consequence decisions. The development of guidelines and procedures for conducting a model V&V program are currently being defined by a broad spectrum of researchers. This report reviews the concepts involved in such a program. Model V&V is a current topic of great interest to both government and industry. In response to a ban on the production of new strategic weapons and nuclear testing, the Department of Energy (DOE) initiated the Science-Based Stockpile Stewardship Program (SSP). An objective of the SSP is to maintain a high level of confidence in the safety, reliability, and performance of the existing nuclear weapons stockpile in the absence of nuclear testing. This objective has challenged the national laboratories to develop high-confidence tools and methods that can be used to provide credible models needed for stockpile certification via numerical simulation. There has been a significant increase in activity recently to define V&V methods and procedures. The U.S. Department of Defense (DoD) Modeling and Simulation Office (DMSO) is working to develop fundamental concepts and terminology for V&V applied to high-level systems such as ballistic missile defense and battle management simulations. The American Society of Mechanical Engineers (ASME) has recently formed a Standards Committee for the development of V&V procedures for computational solid mechanics models. The Defense Nuclear Facilities Safety Board (DNFSB) has been a proponent of model V&V for all safety-related nuclear facility design, analyses, and operations. In fact, DNFSB 2002-1 recommends to the DOE and National Nuclear Security Administration (NNSA) that a V&V process be performed for all safety related software and analysis. Model verification and validation are the primary processes for quantifying and building credibility in numerical models. Verification is the process of determining that a model implementation accurately represents the developer's conceptual description of the model and its solution. Validation is the process of determining the degree to which a model is an accurate representation of the real world from the perspective of the intended uses of the model. Both verification and validation are processes that accumulate evidence of a model's correctness or accuracy for a specific scenario; thus, V&V cannot prove that a model is correct and accurate for all possible scenarios, but, rather, it can provide evidence that the model is sufficiently accurate for its intended use. Model V&V is fundamentally different from software V&V. Code developers developing computer programs perform software V&V to ensure code correctness, reliability, and robustness. In model V&V, the end product is a predictive model based on fundamental physics of the problem being solved. In all applications of practical interest, the calculations involved in obtaining solutions with the model require a computer code, e.g., finite element or finite difference analysis. Therefore, engineers seeking to develop credible predictive models critically need model V&V guidelines and procedures. The expected outcome of the model V&V process is the quantified level of agreement between experimental data and model prediction, as well as the predictive accuracy of the model. This report attempts to describe the general philosophy, definitions, concepts, and processes for conducting a successful V&V program. This objective is motivated by the need for highly accurate numerical models for making predictions to support the SSP, and also by the lack of guidelines, standards and procedures for performing V&V for complex numerical models.
Рассматривается применение критерия хи-квадрат Пирсона для проверки гипотезы о принадлежности выборки нормальному закону. Построены таблицы процентных точек и модели предельных распределений статистики. Получены оценки мощности критериев хи-квадрат Пирсона и Никулина-Рао-Робсона относительно ряда конкурирующих гипотез. Приводятся результаты сравнительного анализа мощности множества критериев нормальности.
Computer simulation of a chemical process (e.g., distillation) usually reduces to solving a system of nonlinear equations with many unknowns using a database of physicochemical properties of gases and liquids. Here, Broyden’s method is applied to systems of nonlinear equations involved in multicomponent distillation calculations, with the convergence of the algorithm ensured.
In this paper, a seemingly contradictory experimental result concerning Fisher criterion is first exhibited. Then, we analyze this result from the statistical correlation point of view and give a reasonable explanation. More importantly, we emphasize that Fisher criterion is not an absolute criterion, and, it should be associated with the statistical correlation together to assess the descrimination of a set of discriminant vectors. © 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.
Two Pearson-type goodness-of-fit test statistics for parametric families are considered for randomly right-censored data. Asymptotic distribution theory for the test statistics is based on the result that the product-limit process with MLE for nuisance parameters converges weakly to a Gaussian process. The Chernoff-Lehmann (1954) result extends to a generalized Pearson statistic. A modified Pearson statistic is shown to have a limiting chi-square null distribution.