Experimental investigation and multi-objective optimization of a novel multi-fluid heat exchanger performances using response surface methodology and genetic algorithm
Belal Almasri,Taraprasad Mohapatra,Hillol Joardar,Sudhansu S. Mishra
DOI: https://doi.org/10.1007/s40430-024-04987-8
IF: 2.361
2024-05-29
Journal of the Brazilian Society of Mechanical Sciences and Engineering
Abstract:This experimental investigation focuses on the heat transfer performance of a novel multi-fluid heat exchanger (NMFHE) in regard to variation in the flow rate of heat transfer fluids HF 1 (hot water), HF 2 (normal water), and inlet temperature of HF 1 (hot water). A brazed helix tube with inside and outside surface waviness is fitted inside NMFHE prepared from helical coil tube is novel. The co-relations between input parameters and output responses are predicted and statistically modeled using response surface methodology and multi objective genetic algorithm (MOGA)-artificial neural network (ANN) and validated with experimental data reasonably. The Nusselt number and entropy generation number are considered as the performance measure for this study. HF 2 flow rate is identified as the most critical input parameter for the Nusselt number for HF 1 , HF 2 , and entropy generation number with an increment of 24.34%, 24.14%, and 24.41%, respectively, whereas HF 1 inlet temperature critically affects the Nusselt number of HF 3 with a decrement of 14.86%. For HF 1 flow rate of 190 LPH, HF 2 flow rate of 200 LPH, and HF 1 inlet temperature of 75 °C, the maximum Nusselt number for HF 1 , HF 2 , and HF 3 , and the minimum entropy generation number, are predicted as 42.84, 52.5, 52.1, and 0.0191, respectively, from the present optimization analysis with a composite desirability (D) of 0.939. The optimal outcomes from MOGA-ANN are experimentally verified and it is found that the data variation within ± 4% in the MOGA-ANN hybrid technique.
engineering, mechanical