Modeling for control of laser melting processes

Laser processing has become an important technology in manufacturing due to the ability to precisely control the position, duration, and intensity of the applied energy, but the quality of laser processed parts can vary significantly over the course of processing parts. Laser melting processes are highly sensitive to variations in material properties and geometry and these process disturbances can perturb the desired melt depth and temperature distribution. Poor reproducibility and variable part geometries lead to the conclusion feedback control is necessary to obtain acceptable processing quality. To implement the control, a sufficiently accurate model of the process is required. Current models of laser melting are too complex for implementation of feedback control schemes (FEM models), fail to accurately predict process output parameters (simplified 1-D models), or lack flexibility in application (system identification models). This work addresses the challenge of deriving a laser melting model that balances the need for accurate melt depth and temperature distribution prediction with the requirements of feedback control compatibility (i.e. a linear model) and flexibility in application. Experimental studies of two different laser melting applications, i.e. selective laser remelting and laser polishing, are conducted to gain insights into the physical phenomena involved. From those insights, a nonlinear laser melting model is derived using energy based techniques consisting of three components: the isotherm shape factor, the preheating model, and the melting model. A linearized laser melting model is then established and the system response is evaluated. Finally, the controllability and observability are evaluated for to ensure compatibility with linear control techniques