Big data simulation for effective thermal conductivity modeling of thermosyphon

Thermosyphon or thermosiphon has been continuously studied for various thermal engineering applications in terms of large heat delivery and removal. Their thermal resistances have been investigated as they are the key to infer the effective thermal conductivity in thermal systems. This study aims to provide an essential basis or modeling equation to design a thermal system to estimate the effective thermal conductivity with thermal resistance. Collecting big data using computational fluid dynamics is newly introduced to examine the thermosyphon, which cannot be tested in an experiment. More than 36,000 simulation cases are conducted by changing total length, radius, heat sink area, heater area, heat transfer coefficient at heat sink, and power of the heater. Selected cases cover the actual ranges by referring to existing experimental or numerical results in other literature. As a result of numerical analysis, a correlation between thermosyphon morphology, thermal resistance, and effective thermal conductivity is derived. The Deep Neural Network (DNN) is employed to validate the simulation data. It is confirmed that the results are consistent and verifiable. These results open a way as we can see that the effective thermal conductivity is a function of geometry and complex function of heat power or heat transfer coefficient. This correlation model is expected to play an essential role in designing thermal systems.

Automated summary

This study aims to design a thermal system to estimate the effective thermal conductivity in thermosyphon by analyzing the relationship between thermal resistance and effective thermal conductivity. Computational fluid dynamics is used to simulate the thermosyphon under different parameters, and a correlation model is derived. The simulation data is validated using a Deep Neural Network, confirming the consistency and reliability of the results.