One of the challenges to applying machine learning in non-destructive testing (NDT) is the limited amount of suitable data to train the algorithms. The medical world has always been ahead of the NDT community, not least because of the higher level of investment but also because the subject of their diagnostic tools is the human body and there are so many of us. This presents an opportunity for the NDT profession to take advantage of the learning from medicine in the development of equipment and techniques, which it certainly has done.
A recent review of a book by Marty Makary, titled: Blind Spots: When Medicine Gets It Wrong and What It Means for Our Health, certainly provoked a number of thoughts, one of which was “what else can the NDT profession take from the medical profession?” First, a disclaimer: I am not saying that the points I highlight from the book apply to the NDT profession, I will leave it up to you to consider what is applicable. But even those things that are not applicable may be issues that we need to be on our guard to avoid in the future.
In the following list, bearing in mind the disclaimer above, I have translated from medical speak to NDT speak to make it easier to see if we can learn lessons:
Data from research and development accumulate and change, so NDT professionals need to change in response. Evidence in the medical world suggests that we do not change quickly enough.
It takes far too long for the best science to reach the workplace.
Bias and groupthink are prevalent.
Reputational protection is too often put before truth.
There is a reluctance to own up to mistakes.
Reviewing the book, The Times quoted the author’s reasons for detailing various medical groupthink errors:
To broaden peoples’ perspective about the range of scientific study on topics.
To encourage civil discourse.
To encourage an open-minded approach to scientific dialogue so that scientific method can more universally govern the medical profession.
I know that item 2 is never going to happen for NDT, nor do I think it is necessarily needed, but 1 and 3 are particularly relevant.
With respect to item 1, I have often pondered on the amount of NDT information generated over the years and published in various journals and presented at all the different conferences. How much gets implemented outside of the authors’ affiliations? Do people know what actually exists? I can hear people shouting that artificial intelligence (AI) will solve this problem, but I think that is a long way off.
There are still ‘old’ pieces of work that have stood the test of time. Taking a personal example and blowing my own trumpet, I recently read a paper in the e-Journal of Nondestructive Testing (ISSN 1435-4934, www.ndt.net) by E and R Ginzel regarding the effect of the couplant gap under wedges when performing scanning of phased arrays. This is an important effect for people to understand when conducting critical inspections. However, in 1989, at the BINDT conference in Sheffield, I co-authored a paper on the same topic for conventional ultrasonic probes (see: www.sciencedirect.com/science/article/abs/pii/B9780080401911500126). Knowledge transfer is an important subject that needs addressing.
Over the years, there has been evidence of some groupthink in NDT, stemming from the desire to find a universal solution. In my career, I have observed time-of-flight diffraction (TOFD) being widely promoted before phased array knocked it off the pedestal, to be hotly followed by full matrix capture (FMC) and the total focusing method (TFM). The phrase ‘horses for courses’ aptly describes the selection of techniques for any particular application, but this is often dismissed in the rush to utilise advanced techniques. I have seen the inability of a common technique to detect defects in a test-piece used as a justification to apply advanced techniques. The advanced techniques may have been the optimum techniques to use but the justification was flawed: the common technique would never have found the defects and an experiment was not needed.
Hopefully, the above may stimulate the discourse and dialogue that these articles are designed to encourage.
Please note that the views expressed in this column are the author’s own personal ramblings for the purpose of encouraging discussion within NDT News. They do not represent the views of Amentum or BINDT.
Letters can be mailed to The Editor, NDT News, Midsummer House, Riverside Way, Bedford Road, Northampton NN1 5NX, UK. Email: ndtnews@bindt.org or email Bernard McGrath direct at bernard.mcgrath1@global.amentum.com
Comments by members
This forum post has no comments, be the first to leave a comment.