Generation and use of a statistical as-built/as-is model

Abstract

Using the principles of as-built/as-is engineering and decision analysis, it is possible to generate a highly accurate as-built model of a manufactured product, and an as-is model of an aged and/or used product from a relatively small amount of measurement information. Using manufacturing data from multiple product instantiations, one can determine probabilistic distributions for design parameters, establish influence relationships between variables, and identify the key characteristics which represent the entire group of instantiations in a single model. As-built/as-is engineering emphasizes the importance of knowing the actual state of a manufactured, aged, and/or used product, as opposed to the nominal design or the state of the product at an earlier time. Decision analysis is a commonly used procedure for determining the optimal course of action in the presence of uncertainty. Combining the principles of as-built/as-is engineering with decision analysis can lead to the creation of a highly accurate and relatively inexpensive numerical model, which can be used to enhance analysis, simulation, or decision making. The merits of as-built/as-is engineering are discussed, as well as the application of decision theory to this topic. The methodology for creating and verifying the as-built model provides the core of this research, and examples are given to display its effectiveness.