Houston, USA
Register for the event and come see our session: Machine Learning-Based Surrogate Model for Computational Fluid Dynamics in Centrifugal Pump Design
Dassault Systèmes is thrilled to be a Bronze Sponsor for the Annual Symposium: AI Accelerated Physics Based Modelling and its Role in Energy Industry, hosted by Computational Fluid Dynamics.
Come see our speaking session at the event, Machine Learning-Based Surrogate Model for Computational Fluid Dynamics in Centrifugal Pump Design.
Centrifugal pumps are used in a large range of applications in the Oil & Gas industries during the production process including extraction, transportation, and refining phases. These pumps typically have a set of vanes, also known as diffuser, surrounding the rotating impeller to enhance efficiency by increasing pressure and reducing flow velocity as the fluid passes through. Therefore, a well-designed diffuser is critical to ensure optimal pump performance.
We propose to use a Design of Experiment through Computational Fluid Dynamics (CFD) in order to train a Machine Learning model allowing to quickly predict the effect of a diffuser’s design on the pump’s performance. Once trained, this surrogate model approach allows to get information on the main performance indicators, notably the total pressure/head rise by inferring the Machine Learning model, which is much faster than running additional full CFD simulations.
