The paper, titled “Hybrid physics-machine learning approach for predicting manifold distribution non-uniformity effects on heat exchanger performance with discrete channel network models,” has been prepared on behalf of Friterm by Burhan Yörük and Mustafa Zabun. The study focuses on the effects of flow distribution non-uniformity on system performance in fin-and-tube heat exchangers. By combining physics-based discrete channel network modeling with machine learning methods, the research aims to provide a fast and reliable way to predict performance losses caused by manifold-related flow maldistribution.

Supported by experimental validation, this hybrid approach offers rapid evaluation capabilities at a significantly lower computational cost compared to conventional CFD and iterative network solution methods. As a result, preliminary analysis, sensitivity assessment, and optimization processes in heat exchanger design can be carried out more efficiently, while also enabling engineering decisions to be made through a more data-driven approach.

Friterm’s study not only demonstrates the company’s engineering expertise in heat exchanger design, experimental validation, and numerical modeling on an international platform, but also contributes to the industry’s transformation toward more efficient, faster, and more predictable design processes.