Numerical Simulation, An Art of Prediction 1

2Simulation Practice

In her novel Birth of a Bridge, the French writer Maylis de Kerangal recounts a human adventure: that of building a bridge in an imaginary California. Sneaking into the psychology of the characters with subtlety, she paints a portrait of women and men accustomed to the mud of the construction sites and the adventures, technical or human, of a civil engineering project. She described Georges Diderot, the engineer in charge of the construction project, in these terms:

[He was] not a supersonic brain lubricated with force diagrams, functions with multiple variables, derivatives, strength-of-material analyses, Euclidean spaces and Fourier series. […] What filled him with joy was operating the lifesized fulfilment of thousands of hours of calculations… [KER 14]

Before carrying out this “lifesized fulfilment”, engineers build confidence in calculation tools and define the rules of practice for numerical simulation. This chapter aims to present them briefly.

When developing a computational model, engineers generally examine the different options that allow them to represent the system under study. A good knowledge of the models is generally sufficient to make the most appropriate choice of equations for the situation concerned. Their practice of simulation, oriented towards the modeling of real systems, often confronts them with the limits of model validity, for example when:

• – the equations on which they base their calculations do not reflect all the physical ...