Abstract
The refrigerant-direct convective-radiant cooling (RCC) system is attracting widespread concern due to its advantages of good thermal comfort, high energy efficiency and simple structure. However, researches on thermal and economic optimization of this system are rare. In this study, a novel heuristic approach is proposed to optimize the aluminum column-wing type refrigerant-direct convective-radiant cooling (ACT-RCC) system, which adopts artificial neural network (ANN) integrated with multi-objective genetic algorithm (MOGA). The numerical and economic models of the ACT-RCC terminal are developed and the numerical model is validated by the experimental data. Besides, the ANN model is adopted to accelerate the prediction of the thermal and economic performances of this system. Results show that the training values of the ANN model are fitted well with simulated results and the ANN model can greatly improve the runtime in comparison with original numerical and economic models. Based on the heuristic optimization approach, the optimal structure of the ACT-RCC terminal is the copper pipe diameter with 8.7 mm, the copper pipe spacing with 25.5 mm and the rib height with 30.3 mm. Compared with the original structure, the cooling capacity of the improved ACT-RCC system is enhanced by 16.0% and the initial cost is reduced by 10.0%. The appearance area equals to the direct product of the length and width, and results show that the appearance area of the improved ACT-RCC terminal is decreased from 1.04 m2 to 0.78 m2. Therefore, the proposed heuristic approach provides guidance for improving the thermal and economic performances of the RCC systems.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have