Neural Net Software for Optimization of Ceramic Grinding

The purpose of this Phase 1 project is to demonstrate the ability to develop a neural network model of the ceramic grinding process with fusion of the sensor/process control parameter data. A model will be developed from existing NIST data using artificial neural systems (ANS) with algorithms that have previously been specifically designed to provide good results for manufacturing applications, for the primary customer base of N.A. Technologies Company. The model will provide the ability to do model based simulation of the process in an automated interface capable of an output weighted optimized search with cost functions for the input parameters. The system will be developed in the 32 bit Windows environment making it compatible with all other certified Windows software and capable of being embedded into an off-line planning (OLP) and automated concurrent engineering (ACE) system. The system will include full Windows Dynamic Data Exchange (DDE), Object Linking and Embedding (OLE), and Open Database Connectivity (ODBC) data communication linkages. The system will also include pseudo-three dimensional response surface graphics dynamically linked to the neural network model with both visual and sound pitch human interface allowing multiple parameter "visualization" using combined visual and auditory feedback. Finally, the system will be designed for easy integration into a 32 bit Windows based, (multi-platform), integrated concurrent engineering, Off-Line Planning (OLP) system.