Cloud-Based CFD Optimization of Magnetic Drive Pumps
Presentation of the problem and objective of the experiment
The experiment will investigate and optimize magnetic-drive centrifugal pumps using Cloud-based HPC with the objective of improving performance and developing new products. Mag-drive chemical-process pumps eliminate the need for shaft sealing, reducing costs. Axial thrust balancing is critical for the design: CFD allows this to be predicted, but the optimization of the pumps to maximize efficiency, minimize thrust and avoid cavitation requires very large models. Thus, advanced HPC infrastructure and CFD engineering tools are essential.
Short description of the experiment
CDR Pompe will provide the necessary specifications to create a model to study the operating conditions of its pumps, aiming at improving their performance and at creating a product that better fits the market needs. As these goals require High-Performance Computing (HPC) infrastructure and engineering tools, the simulation will involve high fidelity CFD transient analyses, on ANSYS software. The optimization process will be geometry-related and based on the Design of Experiment (DoE) and Response Surface methodology. In this regard, EnginSoft has committed to supplying the engineering know-how in terms of CFD modelling, while CINECA will provide the hardware and support needed to conduct CPU-intensive parallel computations through the cloud.
Partner CINECA is part of the NCC Italy.