Advanced Numerical Simulations - 5EUS5SNA

Goal(s)

Learning outcomes (LO)

CFD part for energy system design

• Evaluate the quality of a CFD simulation using verification and validation methods.
• Explain the principles of turbulence modeling and the limitations of standard approaches.
• Describe the numerical models used to simulate multiphase flows.
• Implement a numerical simulation to analyze head losses in a complex hydraulic system (hydroelectric power plant diffuser).
• Determine the optimum geometric parameters of a hydraulic component using the results of a CFD simulation.

Part Optimization for energy system design

With regard to sensitivity analysis, students will be able to :

• explain the principles of global sensitivity analysis methods, in particular the Morris and Sobol methods.
• use sensitivity analysis to identify influential parameters in an expensive numerical simulation.
• interpret sensitivity indices to guide dimension reduction or variable prioritization.

With regard to non-linear optimization algorithms, students will be able to :

• implement and compare optimization algorithms (e.g. quasi-Newton, heuristics, metaheuristics) in the context of energy modeling.
• identify the advantages and limitations of different approaches depending on the type of problem (noisy, costly, multimodal, etc.).

With regard to indirect optimization, students will be able to :

• explain the fundamental concepts of Bayesian optimization and its interest for objective functions that are costly to evaluate.
• build a metamodel (e.g. Gaussian Process/Kriging) from a simulated dataset.
• formulate and solve an optimization problem using a metamodel and an active acquisition strategy (e.g. Expected Improvement).

Content(s)

• CFD for energy system design (G. Balarac): turbulence, two-phase flows, verification and numerical validation; project on the head loss of a hydroelectric diffuser.
• Optimization for energy system design (B. Delinchant): algorithms (quasi-Newton, metaheuristics), multi-objective optimization, sensitivity analysis, metamodeling; optimal building design project.

Basic knowledge of fluid mechanics, heat transfer, applied mathematics (optimization, numerical analysis), and programming (Python).