Background and Objectives:
The aim of my Engineering Doctorate project, which is sponsored by Peak NDT, is to build upon the work currently being undertaken by the University of Strathclyde’s Centre for Ultrasonic Engineering (CUE) in the area of in-process Phased Array Ultrasonic Testing (PAUT) of fusion welded components. My work will look to incorporate process control at the point of manufacture by utilising traditional ultrasonic weld testing techniques and Peak NDT’s instrumentation to image the molten weld pool in real-time. The information gathered through this in-process inspection can then be used to inform a real-time process control algorithm.
Traditionally, welding and quality control inspection practices are distinctly separate within manufacturing process timeline. This limits productivity, throughput and ultimately can increase rework times and associated costs. There are many defects which occur at the point of welding which, if detected early, can be rectified quickly resulting in reduced rework and material costs. Furthermore, by controlling and optimising the welding process in real-time, it may be possible to prevent some of these defects from occurring in the first place.
I began by simplifying the use-case down and utilising conventional, single-element transducers in order to fully understand the wave propagation for the specific case of a liquid metal weld pool within solid metal parent material. By using one of Peak NDT’s conventional UT controllers, the MicroPulse LT, along with some developed LabVIEW software, I have been able to experiment easily with different combinations of single element transduction through various media, i.e. Longitudinal Pulse-Echo, Longitudinal Through-Transmission and Shear Pulse Echo. Initial experiments were performed on replica steel-blocks which were designed and machined with a groove to simulate the geometrical properties of a weld pool when filled with liquid media. This allowed different ultrasonic inspection configurations to be tested at room temperature without the need for live-arc welding to be performed. These initial studies have shown promising results, indicating that there is useful information present within the signals which can be used to infer the geometrical and positional properties of the weld pool.
The effects that the elevated working temperatures and thermal gradients present during welding have on the amplitude and position of received signals is not trivial. Part of my work has been directed towards understanding and predicting these gradients through the use of FEA simulation and experimental validation in order to help develop crucial temperature compensation strategies.