A full-size model
by Bernard McGrath, Inspection Validation Centre

Monty Python was not everybody's taste in comedy. Well, it certainly wasn't my Mum and Dad's! I don't think I ever got to see a full, or even a partial, episode when it was first broadcast. Yet, despite only having seen clips of the series since then, I am familiar with lots of the sketches; mainly because I shared night shifts with someone who could, and did, recite them word for word. Two members of the Monty Python team, Michael Palin and Terry Jones, also wrote a parody of Tom Brown's Schooldays called Tomkinson's Schooldays. I didn't get to see this when it was broadcast either! But, like the Monty Python sketches, bits have wormed their way into my memory, never to leave. Who can forget Tomkinson scoring a try and then keeping on running in order to escape, only to be caught some miles away by the school leopard?

One anecdote, of which I am reminded frequently, is when Tomkinson builds a full-scale model icebreaker in the model boat club. He is told to melt it down and informed (as I remember being told) that a full-scale model is not a model – it is the real thing! I think of this every time I see the test specimens used by the IVC for the validation of Sizewell inspections. These are not really test-pieces, they are real components or part components, albeit defective ones! I try to explain my view of these impressive specimens, in the context of the Tomkinson quotation, to visitors that I show around. Unfortunately, I am usually the only one that laughs.

Of course, building duplicate plant components and inserting defects in them, so they can be used as test-pieces, is not very economic. Neither is it very economic to build a sufficient number of test-pieces to cover a statistically valid number of the possible real defects, in order to prove the capability of an inspection by practical trials alone. Luckily, there is a complementary substitute – models. The increase in computing power over the years has made possible the development of theoretical models for a number of NDT techniques. These models allow many different scenarios to be assessed easier, quicker and potentially cheaper than a similar assessment through a practical exercise, if an exercise of this sort is actually possible.

As such models become more widely available, there is a danger that an over-reliance is placed upon them. Recent history has given us a salutary lesson in what can happen when too much reliance is placed on economic models. A Sunday Times review of a biography of Maynard Keynes quotes from one of his letters: "With a free hand to choose coefficients and time lags, one can with enough industry, always cook a formula to fit moderately well a limited range of past facts." Keynes goes on to say that such quantitative modelling "often cannot support one tenth of the burden that is placed upon it." The review goes on to point out that a particular company's computer-modelled risk assessments underwrote the entire system of credit swaps.

NDT theoretical models are different from economic ones because they are based, at least in part, on physical principles. But they can become not too dissimilar if they incorporate risk, probability and human activity. So the advice given in another Sunday Times article, written by John Kay, is worth heeding: "Models are not right or wrong, but only more or less useful. What is useful depends upon the problem and the context." Kay goes on to restate the lesson that Tomkinson learnt back in 1913 – models are not the real thing! They are simplified representations of reality. Even when they are complex, they are still not reality.

It is easy to be seduced by the fancy graphics and the virtual reality of computer-based models. But some really useful models are the simpler ones which pre-date such technology: the Pollitt model of radiography was first described in a paper in the British Journal of NDT in 1962*. And models do not have to be theoretical calculations. Test-pieces can be models. Many years ago, in the CEGB, there was a mechanical model which showed effectively how a beam propagated through austenitic material.
Skill is required to apply a model, the output of which provides insight and allows judgements to be made about the real world in which we all live and work. Unless, of course, you work nights with someone like my colleague and find yourself in an alternative universe of dead parrots and spam!

*C G Pollitt, 'Radiographic sensitivity', British Journal of NDT, Sept 1962.

Please note that the views expressed in this column are the author’s own personal ramblings for the purpose of encouraging discussion within the NDT Newspaper. They do not represent the views of the IVC, Serco Assurance or the HSE who funded the PANI projects.

Letters can be mailed to The Editor, NDT News, Newton Building, St George's Avenue, Northampton NN2 6JB. Fax: 01604 89 3861; Email: ndtnews@bindt.org or email Bernard McGrath direct at Bernard.McGrath@sercoassurance.com