Mechanism Analysis And Experimental Investigation Of Flow Boiling Heat Transfer In Micro/Mini-channels

The emergence of heat transfer devices with high heat flux stimulates the development of research on the flow boiling in micro/mini-channel. However, it is well known that there are some difference between the heat transfer in micro/mini-channel and that in conventional scale channel. The flow characteristics in micro/mini-channel are more complex, and the heat transfer mechanism is still not clear. Therefore, to carry out a study focusing on the flow boiling in micro/mini-channel and to clarify the flow and heat transfer mechanism are essential to guide the research and design of the compact high heat flux heat transfer components.A rectangular mini-channel flow boiling experiment rig with self-designed and processed test section is established in this study. Influence caused by different channel widths, pressure, mass flow rates, inlet subcooling and heating power on the flow boiling in mini-channel are experimentally investigated for channel width from 0.5mm to 2.0mm under the pressure ranging from 0.12MPa to 0.15MPa. A visualization study is also conducted to define and explore the observed flow patterns, and to analyze the relationship between flow patterns and system pressure drop and heat transfer coefficient. A flow pattern map is obtained based on the experimental data.The nucleate boiling theory of Van Stralen is amended in the present study, and a method of computing the total heat flux of flow boiling is proposed from theoretical analysis. The flow boiling heat transfer area is divided into bubble covered area and liquid covered area, and the weighted parameters of the bubble number and the bubble occupied wall area are employed to amend the formula of total heat flux. The analysis of flow boiling mechanism in mini-channel reveals the relationship between the boiling heat transfer coefficient and channel size. The results show that there is no absolute heat transfer enhancement when using mini-channel, but depending on the exponent related with the bubble departure diameter. Only when the exponent n<1, the flow boiling heat transfer coefficient will increase along with the decrease of mini-channel size.Subcooled boiling and saturated boiling are two important components of the flow boiling phenomenon, and also the study focus of boiling heat transfer mechanism. By summarizing the experimental data, factors affect the onset of subcooled boiling are analyzed, diagrams showing the influence exerted by inlet subcooling, mass flow, width of the mini-channel, system pressure. The proportion taken by each influential factor is compared, and then the formula to calculate the onset heat flux of subcooled boiling is fitted. The saturated boiling is also discussed in this thesis, and the transfer of latent heat dominates at such area. The heat transfer process is closely related with the bubble formation frequency, departure diameter and the mini-channel width, etc, and an empirical formula is proposed. According to the experimental results, from heating surface to the adiabatic surface, three regions are defined as bubble formation area, bubble growth region and bubble annihilation area. The boundary area of these regions can be affected by heat flux, resulting in different movement patterns of bubble.The problem of pressure drop has been the focus of the study about vapor-liquid two phase flow. This thesis work makes a penetrating analysis on the relationship between inlet subcooling, mass flow rate, the width of mini-channel, the system operation pressure and the flow pressure drop. Phase flow model is employed to deal with the problem of pressure drop during the flow boiling process. Two basic assumptions are made as vapor and liquid phases have different flow velocities and these two phases are in a thermodynamic equilibrium state. Based on such assumptions, the expression formula of two-phase total pressure drop gradient is derived, which amends the traditional expressions, fitting better to the flow boiling process in vertical mini-channel.On the base of experimental study, this thesis establishes the flow boiling model in vertical rectangular mini-channels with different sizes and cross-section shapes. Finite volume method is used to discrete the control equations and simulation area, and the SIMPLER method is chosen to solve the pressure field and velocity field. The properties of the working fluid are treated as variable during the simulation and the standard k-e model is selected to simulate the fully developed turbulence area, while wall function is employed for region neat the wall. In addition, the UDF function of the CFD software is used to enable the numerical computation of homogeneous and non-homogeneous boiling. The motion characteristics of bubble generation, growth up and departure are obtained together with the vapor-liquid two-phase pressure profile, velocity profile and temperature profile. The relationship between bubble motion and the heat transfer coefficient is analyzed, and a fairly good agreement is achieved from the comparison of experimental results and simulation output, indicating that the established model can be used to simulate the flow boiling in mini-channel. Taking into account the relatively stringent requirement by the experimental study, many experiments under extreme conditions are difficult to carry out, so the numerical simulation in this thesis can contribute to the research data in a wide scope of parameter values, promoting the development of study on the flow boiling phenomenon in mini-channel.It is well known that the flow patterns of two-phase flow have significant impact on both system pressure drop and heat transfer coefficient. Based on the experimental results and simulation output, this thesis introduces artificial neural network to identify the two-phase flow patterns. Three groups of parameters, i.e. pressure, temperature, void fraction, are served as the input feature vectors, which will be inputted into the BP neural network and Elman neural network after normalization for the identification of flow patterns. It is found that both of the two neural networks possess relatively high reliability and accuracy, providing support to the improvement and development of the online intelligent identification system.