Process Control Development for the Spray Forming Process
Abstract
Spray forming is an alternate alloy production technique to both conventional and powder metallurgy methods. In an effort to develop process control of this process, relationships must be established between process parameters and product quality parameters. Because mathematical modeling of the spray forming process has not yet been able to determine well-define relationships between process parameters and product quality, neural networks were employed to more clearly define this relationship. It was the goal of initial neural network development to prove the feasibility of neural network use in spray forming control. Because this initial work was successful, the focus of subsequent development was on determining and improving the accuracy of these neural predictions. Not only can neural networks successfully predict trends in quality data, but they are as accurate as an experienced operator in predicting quality outputs. Finally, process control development resulted in Windows compatible software program that puts neural network predictions within easy access for the spray forming plant operator
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1994
- Accession Number
- ADA280898
Entities
People
- Rochelle D. Payne
Organizations
- Naval Surface Warfare Center Carderock Division