Towards the automation of TIG welding sensors and control

Below has a brief summary of each of the chapters content and conclusions.

Chapter 1: This chapter presents an introduction to the topic of TIG welding and sensors. The research aims and the objective of the work are also outlined. An outline of the novel work within this thesis is contained within section 1.2 with some background information for the different aspects of the thesis in section 1.3.

Chapter 2: This chapter provides research into the existing methodologies for tackling the different aspects of the project. The chapter focus is on parametric studies, the use of thermography within autonomous welding, laser scanner technologies and the improvements which could be introduced to existing vision systems for the purposes of creating a device for path planning and bead geometry measurement before during and after the welding process

Chapter 3: Experiments were undertaken to understand the influences of Key Performance Variables affect the quality and strength of the welded joint of thin plate stainless steel 316L. Feedback from informal interviews with human welders determined the initial set points for the variables to be tested. A computer designed; D-optimal, experimental setup was chosen to find the optimal parameters to limit the number of experiments undertaken whilst still enabling the statistical difference between each of the findings. The results of these experiments present the current input to the joint and the travel speed of the welding torch over the surface had the greatest effect on the strength and geometry of the joint, more so than the angle of the electrode to the work surface and the distance of the electrode from the welding surface. With the current input and travel speed combining to form the largest collective influence on the weld. These results are then taken for the basis of experimentation for the weld setting in the rest of this work. Additional studies were undertaken to determine if the geometry of the welding electrode was analysed as a means to understand if the welding voltage could be used to control the electrode standoff from the work surface. Further the tests show how the welding arc voltage varies as the tip deteriorates.

Chapter 4: An initial study to investigate the effect of using a thermal imaging based closed-loop control system to address HAZ fluctuations observed during the TIG welding process. A study to assess the optimum robotic TIG welding parameters and temperature range that will give optimal welding performance was undertaken on stainless steel 316L. The temperature range was chosen with 1420°C as the centroid temperature with a boundary of ± 20°C. This allows for the steel to change phase into a liquid but to reduce the chance of material evaporating from the joint due to excessive heat/energy being present in the joint. The experimental setup comprised of a robotic TIG welding system, single wavelength IR thermal camera, a data acquisition system and an NI (National Instruments) based DAQ controller. The results show that the online monitoring and closed loop control of the TIG welding process can help to minimise defects and the proposed system can be readily adapted to any thermal based manufacturing techniques.

Chapter 5: Modern manufacturing has identified a need for the use of a seam following system for use in a constrained environment during the welding process. This chapter presents the design and evaluation of the performance of a small for factor and low cost laser scanner for the TIG welding process. At the time of the creation of this chapter most commercial laser scanners are designed as bulky general-purpose scanners, with stand offs in excess of 200mm from the measurement surface. A benchmark system was developed and compared to a commercial scanner to ensure that a laser line scanner of equal resolution and accuracy could be developed. This developed the tools which could then be used to form the basis of the calibration processes before the miniaturisation process could take place. After these various combinations of sensors and laser modules were tested to understand if a macro sized scanner could fit within the specifications and operational requirements of the scanning system. The results present comparison of commercial systems vs the benchmark system and the macro system. With the benchmark system offering comparable performance to the commercial system and the macro system allowing for acceptable operation in confined spaces.

Chapter 6: The use of move exotic and processed materials has highlighted limitations in longer wavelength scanning systems. A systematic investigation was carried out to compare and quantify how the wavelength of a laser module affects the quality of a laser line on metal of differing surface roughness. The wavelengths utilised for these experiments were; red (635 nm), green (520 nm), blue (450 nm) and violet (405nm). Definitions of missing, saturated and good data points were defined and used as parameters to assess the scanner performance. The setup used a fixed laser module and measurement surface with the camera module attached to a robotic arm and moved to achieve the varying triangulation angles to analyse the number of good data points observable. At each of the triangulation angles the effect of exposure time on the number of good data points was investigated for samples of varying surface roughness. Comparison of the red, green, blue and violet laser scanners clearly shows that as the wavelength of the laser shortens the amount of good data coming from the measurement system increases and therefore yields the highest measurement range for use within laser scanning systems.

Chapter 7: Conclusions of the work presented in the thesis and further work arising from the thesis, discussions around the results and the significance of the results and how they relate to the research questions are also included in the primary section of this chapter. In the second half of the chapter follow-on work is presented, with suggestions on how the follow up work may integrate with the existing work to continue the development of some of the control approaches and developed sensor technology within this thesis.

Appendix 1: The appendix identifies how the system integration was undertaken and how the sensor components operate. Specifications of each of the components is detailed as well as the methods of integration used.