作者: C. Amaris , B.C. Miranda , M. Balbis-Morejón
DOI: 10.1016/J.TSEP.2020.100684
关键词:
摘要: Abstract In this paper, the performance of a gas/oil heat recovery unit is assessed experimentally and by development an Aspen model artificial neural networks. The cross-flow exchanger used to recover residual exhaust gases coming from microturbine drive absorption chiller. test facility consists mainly microturbine, unit, air-cooled experiments were conducted at partial power loads different thermal oil mass flows. Regarding models, depends on inlet conditions, mechanical description fluids thermophysical properties, whereas ANN 3 trained neurons, 4 inputs (inlet flows temperatures), 2 outputs (thermal load overall transfer coefficient). experimental tests show that recovers 18.8 kW 8.1 kW when output varied 23 kWe kWe. Results also coefficient ranges between 243 W.m−2.K−1 89 W.m−2.K−1, while they evidence resistance controlled resistance. Furthermore, simulation results predicts with average relative differences 0.93% 11.27%, respectively, experiments. evidences 0.51% 3.48% for coefficient, respectively.