| Abstract Scope |
Laser cladding has found intense use in the manufacturing industry, thanks to its capabilities of rebuilding pieces with tailored and even enhanced properties. This has led to an increased necessity of correctly controlling one or multiple of the involved parameters in this process such as dimension of the pool, layer height, dilution, laser power, powder flow, glas flow, etc.
The area of control that deals with this type of processes is process control . Obtaining the best result translates into an optimization problem. Today, this problem is often being tackled with rules of thumb procedures, which limits production rates, less flexibility and restrictions in hardware and materials. Only a small number of automation procedures have found their way into daily production for laser cladding processes.
In this document we tackle the process of analyzing and building a control methodology making use of already developed prediction models and process control techniques specifically for dilution. The main objective for this work is to provide a methodology and platform to successfully automate the use of the machinery involved in the process. This methodology reduces the input from the on site engineers for manually tuning design parameters during dry run and the actual process execution, and increases engineering efficiency and practicality.
The correct use and implementation of an SCADA system will correctly make use of the predictive equations (models), tests (empirical verification) and the correct use of all the functional levels of the Distributed Control system that has been put together at the Canadian Centre for Welding and Joining. The aim is to have a system that receives production scheduling as input with the client’s requirements and outputs the finalized product meeting the quality standards.
For this purpose we have put together an automated cell in which individual control loops will be implemented and tested in order to achieve the desired characteristics. The goal is to determine the correct values from predictive equations, verify the results with experimental analysis and finetune the implemented control loops.
The challenge of this system and work is to correctly bridge the gap between the theoretical work and the actual setup in the production line. This means to automatically map unknown and complex relationships between variable representations from equations to the real hardware parameters that may include robot speed, powder feeder speed, laser power, etc. This setup is unique in the sense that no programming of the robot is intended and the system should reconfigure itself according to the operator inputs by using the embedded models and empirical verification in order to automatically set the correct parameters to achieve the desired results. |