Bubble Behaviors And Characteristics Of Flow Boiling In Microchannel With ?-grooves

The microchannel flow boiling for heat dissipation has the advantages of small volume and high heat transfer coefficient,and it is considered as one of the effective methods for solving the cooling problem of microelectronic devices with high heat flux.The artificial nucleation sites such as making grooves on the heating wall of the microchannel can effectively reduce flow boiling instability and enhance the heat transfer performance in the microchannel.The flow boiling heat transfer characteristics in the microchannel with grooves is related to the bubble behaviors and the flow pattern evolution of the vapor-liquid two phase flow.Therefore,exploring the bubble behaviors and flow pattern evolution during the process of flow boiling in microchannel with grooves will improve heat transfer performance,and the heat transfer mechanism of flow boiling can be well understood.In this paper,?-grooves are artificially fabricated on the heating wall in the straight microchannel.Based on the VOF model of CFD-Fluent software,the user-defined function is programmed and compiled,and the flow boiling model in microchannel for numerical simulation is established.The study is focused on the bubble behaviors,flow pattern evolution of vapor-liquid two phase flow,and the characteristics of flow boiling heat transfer in the microchannel with ?-grooves.Based on the numerical calculation results and Matlab's artificial neural network toolbox,BP and RBF neural networks are applied to identify the two phase flow pattern in microchannel with ?-grooves.The main work and conclusions are as follows:(1)The bubble behaviors in microchannel with ?-grooves are investigated.Compared with the straight microchannel,the bubble behaviors in microchannel with?-grooves are significantly different,such as the behaviors of bubble nucleation,detachment and coalescence.In microchannel with ?-grooves,liquid films are found between the merging bubbles and the heating wall,which can reduce the temperature of the liquid region in microchannel and improve the flow boiling instability.(2)The flow pattern evolutions of the vapor-liquid two phase flow in microchannel with ?-grooves are explored.On the basis of the numerical calculation results,It is observed that the two phase flow patterns in microchannel with ?-grooves are isolated bubble flow,confined bubble flow,elongated bubble flow and annular flow.Investigating the effect of flow velocity and groove's structure on the flow pattern evolution,it is found that with the increase of flow velocity,the vapor-liquid flow pattern gradually changes from annular flow to elongated bubble flow,and shows a tendency to develop to the isolated bubble flow.The different groove structure takes a different effect on the flow pattern evolution.(3)The boiling heat transfer coefficient and pressure drop in the microchannels with different ?-grooves are studied.The results show that changing the groove structure can influence the effect of flow boiling heat transfer enhancement.In order to improve the flow boiling instability,it is necessary to choose a certain groove depth and cavity's diameter.However,too deep depth will cause partially dry-out in the groove,and too large diameter will increase the bubbles' volume fraction in the microchannel,which will weaken the effect of heat transfer enhancement.The pressure drops in microchannel with ?-grooves are higher than those of the straight microchannel,and the cavity diameter plays a dominant role on pressure drop in microchannel.(4)BP and RBF neural networks are applied to identify the vapor-liquid two phase flow pattern in microchannel with ?-grooves.According to the numerical calculation results,the related data is picked up and processed.Then,the training samples are input into the network for training,and the overall performance of the network is tested.The results show that the recognition success rates of BP and RBF neural networks are 94.4% and 88.9% respectively.The flow pattern recognition of both neural networks has a good accuracy,which provides a new idea for flow pattern recognition inmicrochannel when flow boiling occurs.