Share

Data-driven modeling of microchannel heat exchangers utilizing dimensionless numbers for enhanced prediction.

Number: 2496

Author(s) : ZOU J., CHEN Y., ZHENG C., HUANG L.

Summary

Microchannel heat exchangers are widely used in the automotive industry and are becoming increasingly popular in the air-conditioning and refrigeration applications. The heat transfer performance perdition of microchannel heat exchangers is of great interest. However, the complexity of the heat transfer processes in these exchangers presents a significant challenge in accurately predicting their performance. Traditional methods often lead to excessive computational resource consumption due to their computational complexity. Machine learning models as a data-driven method can handle complex, non-linear data, and present a promising solution. In this study, we compared various machine learning algorithms to predict heat exchanger performance. We employed Artificial Neural Networks (ANN), Gaussian Process Regression (GPR), and Random Forest (RF) as prediction models. The training data are generated from Computational Fluid Dynamics (CFD) models. The inputs for these algorithms are the mass flow rates and inlet conditions of the fluids, while the output variables include the outlet conditions, capacity and pressure drops. Furthermore, we evaluated different fin structures and incorporated dimensionless parameters such as Reynolds number (Re), and Nusselt number (Nu) into our machine learning models. This method integrates dimensionless numbers into machine learning algorithms, enhancing the interpretability of the prediction results. The validation and evaluation of these algorithms have demonstrated that machine learning models which incorporate dimensionless parameters as inputs possess higher predictive accuracy. This approach significantly enhances the capability to forecast the thermohydraulic efficiency of heat exchangers across various fin geometries, proving their utility in advanced thermal system design.

Available documents

Format PDF

Pages: 10 p.

Available

Free

Download

Details

  • Original title: Data-driven modeling of microchannel heat exchangers utilizing dimensionless numbers for enhanced prediction.
  • Record ID : 30033046
  • Languages: English
  • Subject: Technology
  • Source: 2024 Purdue Conferences. 20th International Refrigeration and Air-Conditioning Conference at Purdue.
  • Publication date: 2024/07/17

Links


See other articles from the proceedings (187)
See the conference proceedings

Indexing