ANN based ternary diagrams for thermal performance of a Ranque Hilsch vortex tube with different working fluids

In this study, an artificial neural network-based ternary diagram was used to predict temperature separation in a counter-flow Ranque-Hilsch vortex tube. The working fluid and nozzle materials were selected as the effect parameters, and the temperature difference between the hot and cold outlets was used as the performance indicator. In the multiple regression and neural network analysis programs, some values obtained from the experimental set were used as input parameters, and statistical evaluations were performed. Different algorithms combinations have been attempted to obtain the best estimates. Finally, new equations were developed to estimate the temperature difference in the vortex tube using the values measured in the experimental set. In addition, a ternary diagram was developed for oxygen gas and air using the experimental conditions to evaluate the temperature differences.

Automated summary

An artificial neural network-based ternary diagram was used to predict temperature separation in a counter-flow Ranque-Hilsch vortex tube. The working fluid and nozzle materials were chosen as effect parameters, and the temperature difference between the hot and cold outlets was used as the performance indicator. Different algorithms combinations were attempted to obtain the best estimates, and new equations were developed based on the values measured in the experimental set to estimate the temperature difference in the vortex tube. Additionally, a ternary diagram was developed to evaluate the temperature differences using experimental conditions for oxygen gas and air.