Study On High Efficiency Heat Transfer Characteristics Of Low Temperature Phase Change Thermal Storage Device

In 2020,China aimed to reach "peak carbon emissions" and "carbon neutrality".As renewable energy gains traction,natural gas,a clean and low-carbon fossil fuel,becomes vital.However,foreign dominance in liquefied natural gas processes,notably in spiral-tube heat exchangers,hampers local production.Understanding condensation flow in these exchangers is crucial for localization.During natural gas liquefaction,spiral tubes contain non-condensable hydrocarbons,involving intricate heat transfer processes.This study blends theory and numerical simulations to explore these dynamics,facilitating technology localization.Firstly,to establish a numerical model for the condensation flow and heat transfer of non-azeotropic mixtures within spiral tubes,a three-zone physical structure conducive to both comprehensive development and backflow prevention is initially employed.The Reynolds Stress Model,capable of capturing anisotropy,is selected to accurately simulate flow conditions.The influence of wall grid partitioning on liquid film capture is analyzed,yielding a rational grid arrangement.By considering condensation heat transfer mechanisms and the impact of mixing,a theoretical formula for calculating the relaxation coefficient of mixed refrigerants during condensation is formulated.Comparative analysis is conducted on the effects of entrainment,interface velocity gradient correction,and turbulence Prandtl number correction on simulation results.A numerical model for condensation flow and heat transfer of mixed refrigerants is established,simultaneously considering mixing effects,relaxation coefficient values,entrainment,interface velocity gradient correction,and turbulence Prandtl number correction.This model is validated for accuracy and reliability using experimental data for ethane/propane mixed refrigerants from classical literature.The validation results indicate good agreement between simulation and experimental data.Specifically,the maximum deviation between simulated and experimental heat transfer coefficients is within ±15%,within ±10% for pressure drop,and within ±7% for void fraction compared to Steiner’s correlation results.Secondly,utilizing our established numerical model,we conduct simulations to explore the impact of various operational and structural parameters on condensation flow and heat transfer in the liquefaction process.Our analysis under steady-state conditions reveals insightful correlations between different parameters and condensation heat transfer coefficients.Specifically,we find that mass flow rate and vapor quality exhibit a positive correlation with heat transfer coefficients,while saturation pressure and pipe diameter display a negative correlation.Within the scope of our investigation,the sensitivity of heat transfer coefficients to winding angle and diameter is relatively modest.Similar trends are observed in pressure drop compared to condensation heat transfer coefficients.Additionally,we extend our simulations to explore the influence of rolling conditions by incorporating the rolling equation into dynamic coordinates.Our findings demonstrate that varying rolling periods and amplitudes have similar effects on the heat transfer process.Hourly heat transfer coefficients exhibit periodic variations under different rolling conditions,indicating a simultaneous enhancement and reduction of heat transfer.We delve deeper into the impact of rolling by analyzing the spatiotemporal distribution characteristics of liquid film turbulent energy and thickness.The thickening of the liquid film results in increased thermal resistance,weakening heat transfer,while thinning enhances heat transfer.The combined effects of these mechanisms manifest differently under diverse conditions,leading to either enhanced or reduced time-averaged heat transfer coefficients.Within the simulated scope of our study,the influence of period on heat transfer coefficients falls within a range of ±20%,while the influence of amplitude remains within ±10%.These insights contribute to a comprehensive understanding of the intricate dynamics governing heat transfer in liquefaction processes.Finally,a comparative analysis of several commonly used classic heat transfer and pressure drop correlations is conducted.The results indicate poor universality of existing correlations,with prediction results strongly dependent on flow type,necessitating the use of correlations distinguishing flow types for prediction.To improve the prediction accuracy and universality of correlations,a half-empirical correlation based on flow type for heat transfer and pressure drop is developed,ensuring continuity between different flow type transitions through the critical angle of drying,and the accuracy of the developed correlation is evaluated based on various alkane literature experimental data as well as the simulation results of this study.In terms of pressure drop,90.38% of alkane predictions have deviations within ±20%,and in terms of heat transfer coefficients,91.06% of alkane predictions have deviations within ±20%.Additionally,a prediction model based on backpropagation neural networks is developed.For the BP neural network prediction model,when there is sufficient data,prediction results are not influenced by flow type,with 95.34% of pressure drop verification data predictions within ±10% deviation and 97.67% of heat transfer coefficient verification data predictions within ±10% deviation.Within the training range,the neural network prediction model can provide high prediction accuracy and can be used to broaden the application range in combination with theoretical correlations,enabling a better understanding of flow and heat transfer mechanisms.