Hybrid CFD-ANN approach for evaluation of bio-inspired dolphins dorsal fin turbulators of heat exchanger in turbulent flow

The central aim of the present investigation is to evaluate the effects of bio-inspired Dolphin's dorsal fin tur-bulators (DFTs) with a hybrid CFD-ANN approach in a counter flow double pipe heat exchanger. The DFT's physical model is adapted from the infrared thermal image of a bottlenose dolphin dorsal fin. The effects of different influential parameters, such as the length scale (D) of DFTs including 4, 8, 12 mm and the number of turbulators (N) including 6, 10, and 14 are numerically examined at different Reynolds numbers. Also, using CFD data, the neural network is trained with great accuracy of R2 above 0.999. Important parameters such as Nu, f, 6, NTU can be predicted in different conditions by the ANN method. This method greatly reduces the costs of CFD solutions, which is advantageous. The results indicate that the DFT's streamlined geometry could significantly reduce friction loss by streamlining the fluid flow. Also, colliding the fluid stream with the DFTs improves heat transfer by boundary layer destruction. The UA Ratio (UAR) of the Dolphin's dorsal fin turbulator heat exchanger (DFTHE) to the simple heat exchanger (SHE) is between 1.02 and 1.74. Also, the ratio of Nu, 6, NTU for the DFTHE to the SHE is equal to 2.48, 1.63, and 1.74, respectively.