The interaction of experimental and numerical methods to optimize solid mechanics models

In the 1950’s the finite element method was “discovered” by structural engineers when it was observed that solutions of multibar structures converged to continuum solutions. The rest of course is history. Today the finite element method has become a universal tool that is used not only by structural engineers, but also by all kind of engineers, physicist and of course mathematicians have become involved with it. Since the very beginning of Renaissance the solution of practical engineering problems was direct experimentation and Leonardo da Vinci was one of the first recorded persons to test materials to fracture and come with some laws, for example that the strength of a wire was a function of the length. This conclusion may have some nineteen century engineers smile but we know today that he was correct. To his name we can add that of Galileo investigation of the strength of a cantilever beam. These two illustrious historical figures can be considered as the forefathers of Experimental Mechanics. From Renaissance we can jump to the end of the 19 century and the beginning of the 20th century France, to Photoelasticity and to the study of the structure of a bridge over the river Seine by Mesnager. Then we have all the developments of the 20th century with the large variety of methods and instrumentation for Experimental Mechanics. Traditionally Experimental Mechanics methods were used to get the solution of stress analysis problems too difficult or impossible to solve with the available tools provided by Continuum Mechanics methods. Experimental Mechanics was called then Experimental Stress Analysis and the different methods developed within this discipline could be looked upon as analog procedures to solve complex systems of partial differential equations, reflection of the lack of universal tools to handle this type of equations. The development of finite element first, then of boundary elements and the different refinements of these techniques filled the gap. Hence we had two competing approaches to go around the lack of universal tools for the solution of systems of partial differential equations, to use analog methods utilizing Experimental Mechanics or to resort to numerical techniques. With the enormous progress of computer technology helping numerical techniques, in the opinion of many professionals the dices were loaded, the final result predictable, in the long run the demise of Experimental Mechanics was inevitable. Many of our colleagues were sold on this idea and supported their opinion on the true fact that a lot of the industrial clients were evaporating and did not come any more with many of their problems to be rescued from their predicaments by the experimental mechanician. Where are we today? Are we witnessing the end of the road of a fading discipline? Many Engineering schools seem to have adopted this point of view and have discontinued the teaching of Experimental Mechanics. Many engineers think that we are moving to a world were the development of very complex algorithms will be possible and these algorithms will solve the most complex problems that we may face with total accuracy. Let us think some more about this dilemma and let us see if we can arrive to some conclusions.