流体热力学性质准确、可靠的建模是动力、航天、石油、化工、制冷、制药等工业过程的重要基础。状态方程是流体热力学性质的唯象模型,以特定的函数形式趋近于流体的真实热力学性质。分子热力学理论为建立状态方程和消除已有模型的非物理表现提供了途径。针对常用的立方型状态方程应用于临界温度以上时呈非物理表现的问题,基于经典统计力学理论,由正则系综配分函数和亥姆霍兹自由能的关系,不指定势能函数而在低密度下进行推导,获得了α函数(立方型状态方程唯一的温度相关参数)随温度变化的普适规律。给出了α函数的热力学一致性判据,指明了可正确描述热力学性质定性规律的α函数及其各阶导数应满足的一般准则。以常用的Soave和Twu α函数作为亚临界分支,构建了三阶连续的分段函数(不引入额外的可调参数),将α函数外推至临界温度以上,避免了原模型在高温下计算维里系数和比热容的非物理表现。针对在中低温热能利用和制冷行业中有广泛应用的氟代烃,采用基于热力学微扰理论的立方缔合状态方程计入其弱氢键作用,并将模型应用于15种纯流体和40种二元混合体系。提出了求解纯流体和二元混合体系缔合分数的简化表达式,显著提高了求解速度和稳健性。针对氟代烃弱氢键的特征和典型应用对气液相精度均有较高要求的特点,改进了参数优化目标函数,在保持缔合模型对液相性质改进效果的同时,避免了气相性质计算精度的降低。新模型同时适用于气相和液相、低温和高温、饱和和高压计算,总体精度优于以往同类模型。通过与水的方程缔合项贡献的比较,印证了本文氟代烃缔合模型对弱氢键的描述符合物理规律。针对经典状态方程在近临界区失效的问题,探索了建立跨接多参数状态方程的一般方法(以二氧化碳为例)。以基于重整化群理论的Kiselev跨接方法替代原模型中导致非物理行为的非解析项,避免了原模型热力学性质计算中近临界等温线的非物理振荡。提出高阶跨接函数,实现了远离临界点时的快速收敛,适应了多参数状态方程的特性。提出孪生高斯项,补偿了由所去除的非解析项与原模型解析部分相耦合导致的简单替代下的精度下降。新方法以较小的精度损失为代价,采用近临界区实验数据拟合跨接参数和孪生高斯项的参数,避免了对多参数表达式中大量参数的复杂优化,实现了对临界渐近奇异性和幂定律的复现。
Accurate and reliable modeling of fluid thermodynamic properties is essential for the power generation, aerospace, petroleum, chemical, refrigeration, and pharmaceutical sectors. Equation of state (EoS) is a phenomenological approach of modeling fluid thermodynamic properties, aiming to formulate and reproduce experimental property data and physical behavior of real fluids. Molecular thermodynamics theory provides insights into establishing new EoSs and eliminating non-physical behavior in available EoSs.The widely used cubic EoSs present non-physical behavior above the critical temperature (Tc). The temperature dependence of the α function (the only temperature-dependent parameter in the cubic EoS) is rigorously derived based on the relationship between the canonical ensemble partition function and the Helmholtz free energy determined by the classical theory of statistical mechanics. In order to obtain universal rules, the derivation is performed at low densities without specifying the intermolecular interaction potential. Furthermore, a set of thermodynamic consistency conditions is proposed for the behavior of the α function and its derivatives. Accordingly, a 3rd-order continuous piecewise function is proposed using the widely used Soave and Twu α functions as the subcritical branch without introducing extra adjustable parameters as extending to above Tc. The revised α function avoids nonphysical behavior of virial coefficient and heat capacity calculations at high temperatures.Fluorinated hydrocarbons are widely used as working fluids in medium-and-low temperature heat utilization and refrigeration systems. The weak hydrogen bond between their molecules is accounted for using the cubic-plus-association (CPA) EoS, which is based on the thermodynamic perturbation theory. The model is applied on 15 pure fluids and 40 binary mixtures. Simplified expressions are proposed for association fraction calculations of pure fluids and binary mixtures, significantly improving the calculation speed and robustness of the model. Because that the hydrogen bond is weaker and that the typical applications demand for high accuracy in both the liquid and vapor phases, the parameter fitting objective function is improved, retaining the accuracy improvements for liquid phase properties, in the meantime avoiding the loss of accuracy for vapor phase properties. The present model is suitable for the calculation in both vapor and liquid phases, at both low and high temperatures, and for both saturated and compressed liquids, and is in general superior to previous similar models. Furthermore, according to a comparison with the association term in the CPA EoS for water, it is confirmed that the present association model for fluorinated hydrocarbons is consistent with the physical behavior of the weak hydrogen bond between fluorinated hydrocarbon moleculesClassical EoSs fail near the critical point. The general procedure of developing crossover multiparameter EoS is investigated, using carbon dioxide as a demonstration. In order to avoid the non-physical oscillations on the near-critical property isotherms, the fallacious non-analytical terms are replaced using the Kiselev crossover method, which is based on the renormalization group theory. In order to adapt to the characteristics of the multiparameter equation of state, a high-order crossover function is proposed which forces the critical effects to diminish rapidly as departing from the critical point. In order to compensate for the loss of accuracy upon the removal of the non-analytical terms due to their coupling with the analytical part of the original formulation, the twin Gaussian terms is proposed. With a small loss of accuracy, the present model fits the crossover and twin Gaussian parameters to near-critical experimental data, avoiding the complex optimization of the large number of parameters in the multiparameter expression, in the mean time reproducing the asymptotic singular behavior and exponential laws at the critical point.