ABSTRACT

Low-pressure refrigerants such as R245fa allow compact design of IT cooling systems intended for harsh environments. Nucleate boiling of R245fa on an electrically heated flat horizontal surface was experimentally studied. Heat fluxes $q_w$ were 30–65 $\mathrm{k}\mathrm{W}/{\mathrm{m}}^2$ and saturation temperatures 40–70${ }^{\circ }$C. Results show that the heat-transfer coefficient HTC increases with both heat flux and temperature, and that higher heat fluxes/temperatures lessen the HTC dependence on temperature/heat flux. For instance, increasing temperature from 40 to 70${ }^{\circ }$C translates into change of the exponent $n$ in the $\mathrm{H}\mathrm{T}\mathrm{C}\propto q_w^n$ from 0.60 to 0.50. Therefore, pool-boiling correlations employing $n=const$ do not accurately reflect the $\mathrm{H}\mathrm{T}\mathrm{C}\propto q_w^n$ dependence. Comparing the experimental data with nucleate boiling correlations showed that the models of Kruzhilin and Mostinski were the most accurate with the Mean Absolute Percentage Error MAPE equal to 16.6% and 32.9%, respectively. The data were then used to optimize the model of Rohsenow and increase its accuracy. This model was chosen for optimization as it is the only model among the analyzed in this paper which incorporates a description of both surface roughness and contact angle. These properties are addressed using an empirical parameter $C_{sf}$. Optimizing this parameter allowed the characterization of the interface between R245fa and heating surface used in this study. The resulting MAPE is 5.9%.

摘要

R245fa 等低压制冷剂允许用于恶劣环境的 IT 冷却系统的紧凑设计。实验研究了 R245fa 在电加热的平坦水平表面上的核沸腾。热通量$q_{瓦}$ 30–65 $\mathrm{克}\mathrm{宽}/{\mathrm{米}}^2$ 和饱和温度 40–70${ }^{\circ }$C. 结果表明传热系数 HTC 随着热通量和温度的增加而增加，并且较高的热通量/温度降低了 HTC 对温度/热通量的依赖性。例如，将温度从 40 升高到 70${ }^{\circ }$C 转化为指数的变化 $n$ 在里面 $\mathrm{H}\mathrm{吨}\mathrm{C}\propto q_{瓦}^n$从 0.60 到 0.50。因此，池沸腾相关性采用$n=C○n秒吨$ 没有准确反映 $\mathrm{H}\mathrm{吨}\mathrm{C}\propto q_{瓦}^n$依赖性。将实验数据与核沸腾相关性进行比较表明，Kruzhilin 和 Mostinski 的模型最准确，平均绝对百分比误差 MAPE 分别等于 16.6% 和 32.9%。然后使用这些数据来优化 Rohsenow 的模型并提高其准确性。选择该模型进行优化，因为它是本文分析的唯一模型，其中包含对表面粗糙度和接触角的描述。这些属性使用经验参数解决$C_{秒F}$. 优化此参数允许表征本研究中使用的 R245fa 和加热表面之间的界面。所得 MAPE 为 5.9%。