Abstract

A Capability Scale for Engineering Simulation Keith Meintjes, CIMdata k.meintjes@cimdata.com Bill Webster, Fidelis FEA Bill.Webster@fidelisfea.com NAFEMS World Congress, 2001 Session: Simulation Governance and Management Overview: We present a capability scale for engineering (digital, physics-based) simulation. This scale is used as a planning tool to subjectively assess simulation applied in product design and development, particularly in relation to physical testing. Introduction: Following Moore?s Law, the capability for technical computing has doubled every eighteen months for nearly sixty years. His has completely changed how products are designed and developed. All of this means that the product engineering process must be continually reassessed and changed to leverage increasing simulation capability. A Scale for Simulation Capability: A scale for simulation capability has been used for over two decades to provide a somewhat objective planning framework. This scale is in the spirit of the CMMI (ref) and other progressions proposed with various goals in mind. We note: a. The assessment is subjective and represents an organization?s self-assessment. b. The assessment is primarily of technical ability. c. It does not so much measure maturity, or process capability. d. It seeks to balance simulation and physical test, not to make a binary choice between them. We look at every product system and subsystem requirement, ask the question: Can this be addressed by simulation? If yes, we then rank the simulation capability on the following scale: 1. Simulation is not useful; it cannot address this requirement. 2. Simulation can be used to rank design alternatives; A is better than B. 3. Simulation can assess whether a design meets the requirement, but testing is required to calibrate the simulation model. 4. Simulation can assess whether a design meets the requirement, and no development testing is required. Final product validation testing is required. 5. Simulation can assess whether a design meets the requirement, and no testing is required. Comments and Examples: If a requirement is not ranked on this scale, that means it has not been evaluated. Level 1: There are many requirements that are assessed by expert opinion: Neither test nor simulation are needed. Other requirements may not have a basis in physics, and simulation simply cannot be applied. Level 2: At a first pass, most simulation procedures will be in this category. But this capability is particularly important: It can support early decision-making in ranking alternatives, and it can guide experimental evaluations. Level 3: As an organization develops its simulation capability, most simulations will be ranked at level 3. Engineers will use fractional rankings that asymptote to Level 4: Giving up that confirmation test is a difficult leap to make. Level 4: Imagine. Products are engineered without development testing and, by implication, are expected to pass validation with no major issues Level 5: Feasible in many cases of simple physics, for example, linear strength and deflection. Discussion: We immediately see this is a very fine-grained assessment, at the level of each individual load case. For a complex product like an automobile, there are thousands of items to consider. This leads to a process of reaching a consensus on how the organization does its work. We asked for agreement between all stakeholders. Test vs. Simulation Comment on issue of test and simulation being in competition for resources. ?Simulation is Better Than Test? Here we will discuss applications (stochastics, robust design, etc.) that are infeasible to accomplish by physical test alone. Closure This process has been very successfully applied over two decades to help create an organizational culture that excels, and a culture that is quick to change and adapt.