Roger W. Logan

Verification & Validation: Process and Levels Leading to Qualitative or Quantitative Validation Statements

The concepts of Verification and Validation (V&V) can be oversimplified in a succinct manner by saying that verification is doing things right and validation is doing the right thing. In the world of the Finite Element Method (FEM) and computational analysis, it is sometimes said verification means solving the equations right and validation means solving the right equations. In other words, if one intends to give an answer to the equation 2+2=, then one must run the resulting code to assure that the answer 4 results. However, if the nature of the physics or engineering problem being addressed with this code is multiplicative rather than additive, then even though Verification may succeed (2+2=4 etc), Validation will fail because the equations coded are not those needed to address the real world (multiplicative) problem. When this simple explanation of V&V is extended to the multidimensional world of nonlinear FEM with multiple application scenarios, the V&V process becomes complicated very quickly. It is essentially impossible to fully verify a code or fully validate a model. The appropriate Level of V&V is a function of the time available to do the V&V evaluations, and this should in turn be a function of the Risk that will be incurred if the V&V is not done, or the risk that will be mitigated if a given level of V&V is done. We will describe a process for V&V based on Levels (with a fractional rating system from 0 to 1). V&V Levels can provide a necessary first step beyond yes or no in answering the question of whether a capability has been verified or validated. We then discuss a 4-step quantitative implementation for V&V once a given Level has been chosen. Next, we provide short examples from metal forming, crashworthiness, and engine performance of the different process Levels for V&V, and the qualitative and quantitative statements that can be credibly made as a function of the V&V Level assessed. We suggest that one of the key end products of V&V is to provide the information needed for predictive adequacy for the intended application, and that adequacy is a balance between rigor and expediency obtained from risk management.

Validation, Uncertainty, and Quantitative Reliability at Confidence (QRC)

This paper represents a summary of our methodology for Verification and Validation and Uncertainty Quantification. A graded scale methodology is presented and related to other concepts in the literature. We describe the critical nature of quantified Verification and Validation with Uncertainty Quantification at specified Confidence levels in evaluating system certification status. Only after Verification and Validation has contributed to Uncertainty Quantification at specified confidence can rational tradeoffs of various scenarios be made. Verification and Validation methods for various scenarios and issues are applied in assessments of Quantified Reliability at Confidence and we summarize briefly how this can lead to a Value Engineering methodology for investment strategy.

Uncertainty Quantification During Integral Validation: Estimating Parametric, Model Form, and Solution Contributions

One of the final steps in building a numerical model of a physical, mechanical, thermal, or chemical process is to assess its accuracy as well as its sensitivity to input parameters and modeling technique. In this work, we demonstrate one simple process to take a top-down or integral view of the model, one which can implicitly reflect any couplings between parameters, to assess the importance of each aspect of modeling technique. We illustrate with an example of a comparison of a finite element model with data for violent reaction of explosives in accident scenarios. We show the relative importance of each of the main parametric inputs, and the contributions of model form and grid convergence. These can be directly related to the importance factors for the system being analyzed as a whole, and help determine which factors need more attention in future analyses and tests.

Risk Reduction as the Product of Model Assessed Reliability, Confidence, and Consequence

JDMS, Volume 2, Issue 4, October 2005 Pages 191–207

Use of non-quadratic yield surfaces in design of optimal deep-draw blank geometry

Planar anisotropy in the deep-drawing of sheet can lead to the formation of ears in cylindrical cups and to undesirable metal flow in the blankholder in the general case. For design analysis purposes in non-linear finite-element codes, this anisotropy is characterized by the use of an appropriate yield surface which is then implemented into codes such as DYNA3D . The quadratic Hill yield surface offers a relatively straightforward implementation and can be formulated to be invariant to the coordinate system. Non-quadratic yield surfaces can provide more realistic strength or strain increment ratios, but they may not provide invariance and thus demand certain approximations. Forms due to Hosford and Badat et al. have been shown to more accurately address the earning phenomenon. in this work, use is made of these non-quadratic yield surfaces in order to determine the optimal blank shape for cups and other shapes using ferrous and other metal blank materials with planar anisotropy. The analyses are compared to previous experimental studies on non-uniform blank motion due to anisotropy and asymmetric geometry.

Energy Absorption in Aluminum Extrusions for a Spaceframe Chassis

This work describes the design, finite-element analysis, and verifications performed by LLNL and Kaiser Aluminum for the prototype design of the CALSTART Running Chassis purpose-built electric vehicle. Component level studies, along with our previous experimental and finite-element works, provided the confidence to study the crashworthiness of a complete aluminum spaceframe. Effects of rail geometry, size, and thickness were studied in order to achieve a controlled crush of the front end structure. These included the performance of the spaceframe itself, and the additive effects of the powertrain cradle and powertrain (motor/controller in this case) as well as suspension. Various design iterations for frontal impact at moderate and high speed are explored.

The Relative Sensitivity of Formability to Anisotropy

This work compares the relative importance of material anisotropy in sheet forming as compared to other material and process variables. The comparison is made quantitative by the use of normalized dependencies of depth to failure (forming limit is reached) on various measures of anisotropy, as well as strain and rate sensitivity, friction, and tooling. Comparisons are made for a variety of forming processes examined previously in the literature as well as two examples of complex stampings in this work. 7 The examples rover a range from nearly pure draw to nearly pure stretch situations, and show that for materials following a quadratic yield criterion, anisotropy is among the most sensitive parameters influencing formability. For materials following higher-exponent yield criteria, the dependency is milder but is still of the order of most other process parameters. However, depending on the particular forming operation, it is shown that in some cases anisotropy may be ignored, whereas in others its consideration is crucial to a good quality analysis.