Signal processing of acoustic reflectometry in estimation of paint thickness measurement over composites using neural networks

An accurate method of measuring the thickness of paint on composite aircraft parts is important from a weight standpoint, but more importantly, due to damage that can occur from lightning strikes on composite parts with excessively thick paint coatings. When the lightning strikes the aircraft, the embedded conductive wires (which are made up of aluminum) evenly spread the current over the surface of the aircraft. The current generated due to the lightning reaches the different layers of the aircraft skin depending on the paint thickness or, in general, coating thickness. If the paint is thick, the current generated by lightning cannot spread and burns a hole in the aircraft. This project was charged with a task of looking at a method to improve the accuracy and ease of operation of a commercial paint thickness measurement system that analyses the reflection pattern of an acoustic pulse echo (time domain reflectometry).