Cloud-Based CFD- and Data-Driven Prediction and Optimization of Fluid Flow Equipment

The CFDPredict project aims to develop innovative methods for predicting and optimizing fluid flow equipment by combining computational fluid dynamics (CFD) simulations with data-driven approaches in the cloud. This technology is essential for companies in the HVAC, transportation, process engineering, and pharmaceutical industries that are seeking improved solutions for their fluid flow applications.

Rising energy costs, stricter regulatory requirements, and increasing market pressure are key reasons why companies in the HVAC, transportation, process engineering, and pharmaceutical industries have a growing need to improve fluid flow equipment and processes. CFD is particularly well suited to this task because it provides detailed insights into complex physical phenomena and enables straightforward “digital experiments.”

The CFDPredict project develops data-driven methods for predicting and optimizing fluid flow equipment based on a combination of CFD simulations and machine learning algorithms. The objective is to reduce the computational time and cost associated with fluid flow optimization by several orders of magnitude through reduced-order modeling. The algorithms are demonstrated using the following prototype devices and configurations:

• Rotary heat exchangers
• Baffles in ducts

• Swirl-nozzle atomizers

All developments are available on the project website: https://github.com/AndreWeiner/CFDPredict

Cooperation	DHCAE GmbH
Funding	“Central Innovation Programme for small and medium-sized enterprises (SMEs)” (ZIM) - Cooperation part KF (Funding reference 16KN119021)
Contact	M. Sc. Janis Geise Dr.-Ing. Andre Weiner