In the single phase AC Resistance spot welding process, consistent weld quality cannot be guaranteed due to its highly coupled electrical-thermal-mechanical nature. As one of the most critical factors for weld quality, methods to control the welding current have been studied by researchers in both industry and academia for many years. However, real-time control of welding current is difficult because the controlled process is typically nonlinear, uncertain, time-varying and fast-response.

For a dynamic system which is uncertain, repetitive and fast-response, adaptive learning control is a good solution. Iterative learning control has been developed by many researchers and proves to be effective for tracking control to certain extent. However, most of them are designed in the time domain, where the system time delay is a big problem for these controllers.

In this thesis, a controller using Fourier series based learning control scheme is developed to control the current of RSW process. The control law is updated on-line in the frequency domain, and the stability of the closed-loop system is theoretically analyzed based on the Lyapunov theory. It has been shown that the asymptotic stability and good performance of the closed-loop system can be achieved. The effectiveness of the proposed controller is experimentally verified on a 63KVA single phase AC RSW machine.

The work presented in this thesis shows that the Fourier learning control method is suitable and effective for controlling the RSW process. We believe that this control method can also be applied to other welding processes, such as arc welding and MFDC welding.