- To review state-of-the-art numerical modelling and process monitoring techniques to identify the most suitable selections for the proposed concept
- To develop numerical models for welding process parameter optimisation along with its integration with the robot control system, using robotic Cold Metal Transfer (CMT) welding as the demonstrator process
- To develop an automated process monitoring system (laser line scanning) along with its integration with robot control and numerical models
- To integrate the developed modelling and monitoring systems with the robotic welding system and demonstrate the automated process optimisation capability
The concept of this proposed project is an automated parameter optimisation system based on numerical process modelling and real-time process monitoring for robotic arc welding, as shown in Figure 2. By inputting the welding requirements (e.g. weld size, penetration depth, residual stress / distortion limits, etc.) into the validate numerical (finite element) models, the optimised process parameters and welding sequence will be automatically calculated and translated into robot and welding programmes for the robotic system to subsequently perform the test welding. The test weld will be measured / inspected by the automatic process monitoring units. If the unsatisfactory results are gained from the automatic inspection, the data will be fed into the numerical models to correct, re-calculate and re-optimise the process parameters. The concept aims to provide a highly automated intelligent process parameters development and optimisation solution for robotic arc welding.
Relevant Industry Sectors
Mechanised programmable industrial robots are widely implemented to automate arc welding processes, in order to increase productivity and improve product quality and repeatability. While robotic arc welding has played an important role in automotive industry, it recently has been rapidly developed for aerospace, power generation and oil & gas industrial applications.
Due to the complex physics behind fusion welding, when developing an automated arc welding procedure for a new product or for improvement of an existing product (in terms of quality and productivity), the majority of research and development (R & D) time may be spent on welding parameters determination and optimisation. The robotic welding parameters not only dominate the weld formation, welding quality, mechanical properties and other key aspects of the welded components, but also directly influence the productivity. Most arc welding processes involve a series of welding parameters, which results in a large number of possible combinations. High cost and long lead time can be caused by the difficulty on selecting correct parameters from the large number of combinations.
With current robotic arc welding processes (Figure 1), identification of the correct welding parameters to deliver the required products is mainly based on trial-and-error experiments incorporating the welding engineers’ knowledge and experience. As a result, the complicated combinations of welding parameters can generate a large number of welding trials which are costly and time consuming. This negatively affect the implementation and development of robotic arc welding in high value (e.g. aerospace, power and oil & gas) and high volume (e.g. automotive) manufacturing industries.
Some basic mathematical (analytical) modelling tools and parametric envelopes (based on previous experiences) have been applied to reduce the difficulty but due to their lack of accuracy for changing applications they frequently could not offer effective help. Besides that, the current analytical modelling technique and experience based data cannot produce reasonable prediction for the weld’s mechanical performance which is key of importance for critical components.
There is therefore a need for an effective intelligent process parameter optimisation tool to be part of a robotic arc welding system. The concept can also be adapted by other robotic fusion welding and material processing processes, e.g. laser beam welding and arc additive manufacturing (AM).
This project aims to produce a novel method for optimising welding parameters for robotic welding, to significantly reduce the number of trials that need to be undertaken using traditional techniques.