利用熱力學模型研究水合物之相平衡行為

Abstract

本研究建立一理論計算模型包含 van der Waals and Platteuw model (vdWP) 和 Peng-Robinson-Stryjek-Vera Equation of State (PRSV EoS)，並將之應用於計算單一氣體水合物(甲烷、乙烷、丙烷、二氧化碳及氮氣)之相平衡行為，並可進一步預測單一氣體水合物添加抑制劑(甲醇、乙醇、乙二醇、二甘醇及丙三醇)及混合氣體水合物系統之相平衡行為。在此模型中，PRSV EoS搭配van der Waals (vdW)及modified Huron-Vidal (MHV1) 混合規則來求得狀態方程式中之特性參數。MHV1 混合規則中之活性係數(activity coefficient)則藉由UNIQUAC和COSMO-SAC 液態模型來求得。 COSMO-SAC 模型不包含兩物質之間之交互作用力迴歸參數，為一具有預測性之液態模型。而在本研究中利用COSMO-SAC 模型所求得之相平衡條件結果，與具有交互作用力迴歸參數之 vdW 混合規則及UNIQUAC模型所得之結果比較，並不會有太大之偏差。在乙烷及丙烷氣體水合物系統中，利用MHV1＋COSMO-SAC所求得之壓力平均誤差為1.93 % 和 2.61 %，利用vdW所求之壓力平均誤差為1.96 % 和 2.63 %，利用MHV1+UNIQUAC得之平均誤差為1.94 % 和2.62 %。此理論模型中所需之參數可經由計算單一氣體水合物之相平衡條件而得，並可進一步應用於添加抑制劑以及混合氣體水合物之系統，預測其相平衡條件與相平衡狀態下系統組成。

In this study, the simple hydrate thermodynamic model including van der Waals and Platteuw model (vdWP) model and Peng-Robinson-Stryjek-Vera Equation of State (PRSV EoS) is developed to calculate the phase equilibrium of pure gas hydrates (CH4, C2H6, C3H8, CO2, N2). Moreover, it can be extended to predict the phase behaviors of gas hydrate systems in the presence of inhibitors e.g. methanol, ethanol, ethylene glycol, diethylene glycol and glycerol, and mixed hydrates systems. The PRSV EoS is incorporated with the van der Waals (vdW) mixing rules and the modified Huron-Vidal (MHV1) mixing rules, and the UNIQUAC and the COSMO-SAC models are used to determine the activity coefficient in the MHV1 mixing rules. The COSMO-SAC model is the completely predictive model without any fitting binary interaction parameters, and the results obtained by using the COSMO-SAC model are comparable to those derived from vdW mixing rules and the UNIQUAC model containing fitting parameters in this work. For the C2H6 + H2O and C3H8 + H2O hydrate systems, the AADP (%) are 1.93 % and 2.61 % (MHV1+COSMO-SAC), 1.96 % and 2.63 % (vdW), and 1.94% and 2.62 % (MHV1+UNIQUAC). The model parameters are fitted to experimental data for pure gas hydrates, and the phase equilibria for the hydrate systems with inhibitors and mixed hydrate systems can be predicted successfully without any additional parameters.