5th November, 2014
Today we would like to thank Mike Lowe, Mechanical Engineering Professor at the Imperial College of London, who accepted to share with us his universitarian point of view about NDT simulation.
Pr. Lowe, you started to study Mechanical Engineering in 1987, and finally specialized in NDT. Why did you choose this field?
In fact we can go back earlier than that when I started as a Civil Engineer! My first job was with an engineering consultancy company, WS Atkins, which I began in 1979. At that time I worked in stress analysis and structural integrity, including a lot of work on writing and using Finite Element programs. Many of my projects addressed the structural integrity of nuclear power stations and the safe design of offshore oil and gas structures. It was a natural step from the topic of structural integrity to the topic of NDT, because NDT is vital to assess and assure structural integrity. It was also a natural step for me, with my interests in modelling, to move from the modelling of stress analysis and vibration to my particular NDT interests in the modelling of guided waves, wave propagation in solids, and the reflection of waves from defects. That change brought me from a Civil Engineering discipline to Mechanical Engineering, but that was not a big change, because the specific topics of my interest exist within both of these branches of engineering. My work in engineering consultancy in the first 10 years of my career gave me a strong motivation towards solving problems for industry, and this has given me the same motivation to address real industrial problems in my research since I moved to work in the university.
The imperial College is a CIVA user (UT, ET and RT modules). How often do you personally use CIVA?
Yes, we use CIVA in my NDT research group. Unfortunately I do not get time to use it with my own hands. I have always liked "getting my hands dirty" by doing my own research on the computer and in the laboratory, but these days I don't get the time to do that, so my work with CIVA, and many other areas of interest, is through the hands of the researchers in my group.
What is your personal point of view on NDT simulation?
I have no doubt at all: NDT simulations are vital to the future of NDT. We can already see that simulation tools have brought many capabilities to NDT. Simulations enable us to identify the sensitivity of proposed NDT inspections, which allows us to decide whether or not a proposed inspection plan or new idea is worth pursuing. It also enables us to optimise the parameters to get the best performance out of an inspection. And all of this can be done, at least initially, just on the computer. If such simulations can be validated and formally approved, then the follow-on benefit in the reduction of the need for experimental qualification, using expensive test pieces, can be saved, and at the same time the confidence in safe deployment of NDT can be increased. Simulations can also be used for the investigation and understanding of measured data, for the training of inspectors, and in many other contexts in NDT.
If I am not mistaken, you conducted research on analytical and numerical modeling. In your opinion, what are the current limits that people should be trying to improve/work on? What are the main difficulties of modeling?
Yes, analytical and numerical modelling is a particular interest for me. I think the most important challenge we face is to develop reliable and accurate simulations that can be calculated in sensible amounts of time. For my main interest in ultrasonic NDT, the problem is that the simulation of wave propagation and scattering is hugely demanding on computers, so we are limited in what we can achieve. This has led to approximations that have been necessary in order to be able to perform the simulations reasonably quickly. Very fast progress has been happening in recent years, thanks to the revolution in computing power, so that the approximations are being eroded and at the same time the solutions are being calculated faster. This is an important area for reasearchers to work on. Another area that is very important is the qualification of simulations and simulation tools. After developing powerful new capabalities, we then need to demonstrate their validity and prove them such that the people who use these tools are able to trust them.
You are taking part in the SIMPOSIUM project, aiming to integrate in a unique platform interoperable NDE simulation tools, to make possible the virtual testing of parts at the early stages of manufacturing and design. What is the role of the imperial College in this project?
We have developed an interface with CIVA so that separate Finite Element programs can be linked to CIVA to calculate the scattering of ultrasound from complex defects. The idea is that users can run CIVA to simulate an inspection, but a part of the inspection, the reflection of the ultrasound from the defect, can be calculated externally by the Finite Element program instead of within CIVA. The advantage of this is that the reflection from any kind of complex defect, such as rough cracks, inclusions, or branched cracks, can be modelled accurately by the Finite Element program, while the user still has all the advantages of the CIVA program. The possibility that this link may be to any external Finite Element program is important for industrial users who already use particular Finite Element programs, so they may already have the necessary qualifications for that program in place, as well as familiarity with how to set up the model of the defect.
EXTENDE recently released some tools dedicated to universities, including the eNDT application. Did you have the occasion to try any of them? If so, do you find it useful?
I don't think we have used such tools.