Eventos Anais de eventos
ENCIT 2020
18th Brazilian Congress of Thermal Sciences and Engineering
CORRELATION OF TWO-PHASE PRESSURE DROP OF THE R1234yf IN SMOOTH HORIZONTAL TUBES: AN ARTIFICIAL INTELLIGENCE APPROACH
Submission Author:
Juan Jose Garcia Pabon , MG
Co-Authors:
Ali Khosravi, Juan Jose Garcia Pabon, Marcos Silva, Mamdouh El Haj Assad, Mohammad Malekan
Presenter: Juan Jose Garcia Pabon
doi://10.26678/ABCM.ENCIT2020.CIT20-0712
Abstract
High global warming potential (GWP) refrigerants that have been used as working fluids in industrial and residential applications are alleged to pose a threat to the environmental security. R134a refrigerant is the most widely used refrigerant medium temperature with GWP of 1430, which therefore must be replaced. One substitute is the R1234yf refrigerant with a GWP lower than 1. In this study, pressure drop during evaporation of R1234yf refrigerant in smooth horizontal tubes is investigated by experimental study and developing intelligent methods. Calculation of the pressure drop for two-phase flow in evaporation and condensation processes needs many experimental tests. Therefore, in this investigation, machine learning algorithms (MLAs) are employed to predict the pressure drop during evaporation of R1234yf refrigerant. Three methods of MLAs are developed to predict the target. These methods are adaptive neuro-fuzzy inference system (ANFIS), ANFIS optimized with particle swarm optimization (ANFIS-PSO), and ANFIS optimized with genetic algorithm (ANFIS-GA). The intelligent models are constructed based on tube diameter (3.2, 4.8, 6.4, and 8 mm); saturation pressure, mass velocity and quality of vapor as input variables and the pressure drop is selected to be the target. The results demonstrated that increasing the mass velocity increases the pressure gradient. Also, for the larger value of tube diameter, the pressure drop has lower values. The behavior of two-phase pressure drop was accurately predicted by the ANFIS-PSO model. Moreover, the PSO algorithm for increasing the performance prediction of the ANFIS model performs better than the GA.
Keywords
pressure drop, R1234yf refrigerant, Adaptive neuro-fuzzy inference system, Particle Swarm Optimization, genetic algorithm
DOWNLOAD PDF VIEW PRESENTATION
