Machine Learning Implementation for Curing Carbon Fibers using Microwave Technology

Sammanfattning: Curing of carbon fibers using microwave technologies was shown to improve ove rprocesses based on thermal heating, both by reducing time and energy cost. However, realization of the full potential of the approach is hampered by the non-even distribution of the energy during curing. This challenge can be addressed by integrating hardware and software in an actively controlled microwave system. This project investigates feasibility of the approach by implementing and testing an embedded system actively controlling the microwave generation. This system is based on feedback generated by an infrared camera. The core of the solution is an automatic control solution based on the device of the so-called beambook. The presence of a complex feedback mechanism relates this directly to a method of artificial intelligence. First experimental evidence is given of feasibility and efficacy of the approach. It is shown that curing using this system results in sufficiently even distribution of curing of carbon fiber reinforced polymer samples.