Product Code: ICA13_1704

Clad Height Control in Laser Cladding Using a Nonlinear Optimal Output Tracking Controller
Authors:
Mohammad Farshidianfar, University of Waterloo; Waterloo USA
Amir Khajepour, University of Waterloo; Waterloo ON Canada
Saeid Khosravani, University of Waterloo; Waterloo ON Canada
Adrian Gerlich, University of Waterloo; Waterloo ON Canada
Presented at ICALEO 2013

Laser cladding also known as laser powder deposition has proven to be one of the most efficient methods of metallic deposition in the industry. Due to its advantages, it is now being used in applications such as rapid prototyping, metallic coating and parts repair. Despite all the benefits provided by the laser cladding process its applications are limited mainly because of the shortcomings in its automated control. Several parameters such as scanning speed, laser power and powder federate are influential on the clad mechanical, metallurgical and geometrical properties. Although, obtaining an accurate model for the process is difficult to achieve but, system identification techniques can provide a good prediction of the process dynamics. Variation in system parameters as well as disturbances in the working environment will produce unexpected changes in the clad height. Thus a closed-loop control system for the clad height will improve the production quality and manufacturing speed.
In this paper a nonlinear optimal output tracking controller is designed to control the clad height. The substrates moving speed is used as the control action. An approximate nonlinear model obtained by previous researchers is linearized in the neighborhood of the desired clad height. The linearized model is found to be in the form of a bilinear system. A novel control technique is proposed to develop an optimal output tracking controller for the bilinear system. First an optimal output regulator algorithm is developed. The optimal output tracking problem is solved by two separate methods in order to obtain a sub optimal output tracking controller. The method is then applied to the laser cladding model. The response of the designed controller is verified using various input signals. Results are obtained and presented to corroborate the effectiveness of the proposed method.