Bounded Rationality and Materials Selection
- PDF / 1,397,763 Bytes
- 5 Pages / 604.8 x 806.4 pts Page_size
- 43 Downloads / 260 Views
MRS BULLETIN/OCTOBER 1999
data. This may enable the design of more appropriate validation procedures required in proving a material for an artifact.
new materials are used, the new testing regimes will ignore some of the more subtle differences between materials that can have a significant impact on performance. In this article, we present a framework for characterizing and representing the engineering problem posed when an artifact is moved from a situation where both the material and design have been p r o v e n in the laboratory a n d in the field—materials-selection problems of this kind are well understood and are treated, for example, by Ashby 1 —to a situation where one or both of them have been unproven in the field. Our objective is to make this framework as computational and decision-rule-oriented as possible, given the context that materials selection is based on materials databases and can be linked to materials properties.
The Framework Choosing Materials A materials-selection decision is often disguised as a materials substitution decision. That is, a new material is selected to replace a material for one reason or another^ When this substitution takes place, certain salient properties of the material are seen to change. These changes in a material's most notable properties motivate certain changes in the design, fabrication, testing, and finally the performance of the new artifact. The difficulty, however, is that it is not only these properties that affect a material's eventual performance. Many other, more subtle properties change as well, and the effects of these changes may cause failures and, in some cases, even catastrophic damage. We will argue that at least part of the error lies in the decision-making method used to prove and validate the new material. The old set of proving trials (tests) are used to generate a new set of tests. In other words, the new testing regime is developed by using the differences between old and new materials to generate a set of changes to the old testing regime. Obviously, if only the salient differences between the salient properties of old and
S o m e t i m e s , of course, the substitution is metaphorical rather than physical. When we design a new device or system, there is no material that is being replaced. Instead, the substitution is based on experience by comparing the new device with an old device that has many similar operating conditions. The point is not that there is nothing new, but rather that even new artifacts must, in the n a t u r e of things, share many attributes with older, more familiar artifacts.
O u r framework is a coarse, t w o dimensional characterization of an artifact. We characterize an artifact as having two components: material and design. Furthermore, each dimension has just two values: proven or unproven. T h u s our characterization consists of four cells arranged in a conventional 2 X 2 matrix. An artifact has either proven or unproven materials (Pm or Um) and the design itself is again either proven or unproven (Pd or Ud).
Data Loading...
