The Customer is often interested on tools delivering information about an entire process system or an entire physical phenomenon in real time, with data that are quickly understandable and usable. In the pre-design stage this approach is powerful and cost-effective, but the same can not be said during the design stage or when the product optimization is investigated. In the latters a deep insight of the involved phenomena often simplifies designers work and reduces the product time-to-market.
The required simulation tools are two and they look diametrically opposed: on one side fast and rapid tools (0D/1D), on the other side tools able to deeply investigate phenomena otherwise unintelligible (3D). The link is represented by SURROGATE MODELS, that is, simple algebraic relations able to replace detailed 3D computations. Once inserted in 0D/1D structures, they behave as the other standard blocks precisely, allowing to solve quickly even complex problems. Surrogate models are built over a wide amount of data concerning specific phenomena:
- most influencing parameters are identified which mostly alterate the phenomena behaviour;
- a database is built made up of a myriad of 3D (both numerical, CFD and FEM, and experimental) evaluations performed varying the previously identified influencing parameters;
- a mathematical correlation is found and calibrated which is able to reproduce the database information.
The accuracy and the cohesion degree to the real physics depend on the size of the computational database and on the range the model is applied over.