Automated Data Acquisition for In-Situ Material Process Modeling

This Phase I SBIR proposal plans to first identify modeling methodologies that are applicable to real-time data acquisition. Information agents will be developed to discover and isolate discrete and continuous features within materials process data. Process discovery implemented through the information agents will allow the continuous features of a process to be identified and grouped into episodes via the discrete features that segment them. Encoding real-time data into a given modeling methodology will be the primary consideration used in determining the suitability for a modeling method. Then the InfoScribe data acquisition system will be used to demonstrate sensor-actuator connectivity on a target process. Following the application of material process data acquisition, preliminary identification testes will be performed to characterize and distinguish process noise. Noise characterization will be implemented as the inverse of process characterization that is utilized by many process control tuning methodologies. Information agents will observe process actuation parameters for categorized features and correlate them to dependent process parameters to characterize the noise in a given materials process system. The proposed in-situ material process modeling capability will be designed, implemented and evaluated in a Phase II follow-on SBIR program.