Yadav Swati, Angira Rakesh
Process Systems Engineering Laboratory, University School of Chemical Technology, Guru Gobind Singh Indraprastha University, Sector 16-C, Dwarka, Delhi, 110078, India.
Environ Sci Pollut Res Int. 2024 Oct 29. doi: 10.1007/s11356-024-35257-8.
Biodiesel stands out as the most favorable alternatives to fossil-derived diesel, offering a multitude of environmental benefits. At various stages of biodiesel production, the separation processes necessitate thermodynamic models with the capability to correlate and predict phase equilibria of mixtures. In this study, application of the classical differential evolution (DE) and its new enhanced version, the OPDE algorithm, for modeling vapor-liquid equilibrium (VLE) associated with components related to biodiesel production is presented. The algorithms were analysed and contrasted in terms of their performance, in estimating parameters for Wilson, NRTL and UNIFAC models. Additionally, classical least-squares (LS) and error-in-variable (EIV) approaches were examined and compared. Also, VLE datasets for a specific system have been merged, and parameters have been estimated. The findings suggest that parameters obtained through LS approach align with those reported in literature, indicating faster convergence in all problems. In contrast, the EIV approach achieved a higher objective function value compared to the LS approach, exhibiting low deviation. OPDE outperformed DE in terms of performance. The enhancement in RMSTD value has been found within range 91%-99% for EIV approach. Further, novel findings derived from some of the studied VLE datasets are presented.
生物柴油是化石衍生柴油最有利的替代品之一，具有诸多环境效益。在生物柴油生产的各个阶段，分离过程需要能够关联和预测混合物相平衡的热力学模型。本研究介绍了经典差分进化（DE）及其新的增强版本OPDE算法在与生物柴油生产相关组分的气液平衡（VLE）建模中的应用。对这些算法在估计Wilson、NRTL和UNIFAC模型参数方面的性能进行了分析和对比。此外，还研究并比较了经典最小二乘法（LS）和变量误差（EIV）方法。同时，合并了特定系统的VLE数据集并估计了参数。结果表明，通过LS方法获得的参数与文献报道的参数一致，表明在所有问题中收敛速度更快。相比之下，EIV方法与LS方法相比目标函数值更高，偏差较小。OPDE在性能方面优于DE。对于EIV方法，RMSTD值的提高幅度在91%-99%范围内。此外，还展示了一些研究的VLE数据集得出的新发现。
