In this work, first, the phase equilibrium of the CO₂/water system has been conducted by using two equations of state, including Two-State and perturbed chain statistical association fluid theory equations of state. The experimental value of cross-association energy (exists in the previous studies) is used for this study to model the equilibrium composition CO₂ in the aqueous phase. Two strategies are applied to carbon dioxide. In the first strategy, carbon dioxide can only have cross-association with water, so it has no self-association. The second strategy considers CO₂ as both self and cross-associating fluid. Also, an additional equation is used for the cross-association section which reduces the number of adjustable parameters. The results of this study show that the first strategy is successful for all cases, and it is accurate while the second strategy is unsuccessful for Two-State equation of state. Moreover, the application of the first strategy and experimental cross-association energy makes the model independent of simplicity of the model and number of adjustable parameters. Then with the sufficient amount of datasets, machine learning techniques were applied to predict the solubility of CO₂ in the water with high accuracy. The results are in good agreement with the experimental data with the correlation coefficient (R) of 0.999 and mean-square root of 4.45e−6 for multilayer perceptron network, which means that the network can predict the solubility for the wide range of temperature and pressure.

在这项工作中，首先，利用两个状态方程进行了CO _{2 /水体系的相平衡，包括二态和扰动链统计关联流体理论状态方程。}本研究使用交叉缔合能的实验值（存在于之前的研究中）来模拟水相中的平衡组成CO _{2 。}两种策略适用于二氧化碳。在第一种策略中，二氧化碳只能与水发生交叉缔合，因此它没有自缔合。第二种策略将CO ₂视​​为自缔合流体和交叉缔合流体。此外，交叉关联部分还使用了一个附加方程，这减少了可调节参数的数量。研究结果表明，第一种策略在所有情况下都是成功的，并且是准确的，而第二种策略对于双态状态方程则不成功。此外，第一种策略和实验交叉关联能的应用使得模型独立于模型的简单性和可调参数的数量。然后，利用足够数量的数据集，应用机器学习技术来高精度预测CO ₂在水中的溶解度。结果与实验数据吻合较好，多层感知器网络的相关系数（R）为0.999，均方根为4.45e−6，这意味着该网络可以预测较宽温度和范围的溶解度。压力。