We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Novel tools for coating developers: a solution to the inverse problem in formulation.
- Authors
Vitela, Javier E.; Nahmad-Achar, Eduardo; Nakamura, Rene; Gama, Armando; Rodríguez, Javier
- Abstract
The inverse problem in coating formulation, i.e., that of finding a formulation that will satisfy a given set of performance properties, is much more difficult than the direct problem of finding the properties of a given formula. While the latter case may be labor intensive and time-consuming, we have at hand a wealth of test methods that make the solution straight-forward. By contrast, the former case calls for experienced formulators plus intensive trial-and-testing work. In this work we describe a way to use intelligent systems in the form of artificial neural networks and genetic algorithms to tackle the more difficult inverse problem. It is applied to a set of five different families of alkyd enamels. In the direct mode, we are able to predict performance of a formulation with 90% efficiency. Through a sensitivity analysis, we can also obtain those variables to which a given property responds more strongly. In the inverse mode, we can obtain enamel formulations that will satisfy a given set of performance requirements. On the whole, it constitutes a powerful lab tool for coating developers. Computer-proposed formulations with some prescribed characteristics and their laboratory tests are discussed.
- Subjects
COATING processes; FINISHES &; finishing; SURFACES (Technology); GENETIC algorithms; ENAMEL &; enameling; TEST methods
- Publication
Journal of Coatings Technology & Research, 2007, Vol 4, Issue 3, p231
- ISSN
1945-9645
- Publication type
Article
- DOI
10.1007/s11998-007-9034-z