The interaction model parameters are the key thermodynamic parameters in activity coefficient models such as Wilson and non-random two-liquid (NRTL) models. Their accuracy directly determines the reliability of the results of vapor – liquid phase equilibrium and separation process calculations. The core bottlenecks of traditional parameter acquisition methods relying on experimental data regression and universal functional activity coefficient (UNIFAC) model estimation include limited data coverage, unstable solution of nonlinear equations and limited scope of application. To address these issues, we constructed a dual-branch graph neural network (GNN) model based on molecular structure characterization enabling efficient and direct prediction of the parameters of the Wilson interaction model. This model takes molecular diagrams as input, separately extracts the molecular features of solutes and solvents, and, through the message-passing mechanism, deeply learns the potential structural correlations between molecules. At the same time, we introduce a parallel residual strategy to reduce the prediction error. Using 4068 binary system interaction model parameter data for training and testing, our proposed parallel residual GNN model significantly outperforms the UNIFAC method across multiple error indicators, and compared with the single GNN model, the prediction accuracy is further improved. The predicted parameters Aij, Aji, Bij, Bji of the Wilson interaction model were substituted into the Wilson equation to calculate the infinite dilution activity coefficient γi∞. The results were highly consistent with the calculated value in the literature, verifying the thermodynamic rationality and engineering applicability of our model.
- Article type
- Year
Open Access
Issue
Open Access
Issue
Although UV flash is currently the most rigorous flash algorithm in the field of dynamic simulation, the algorithm suffers from problems such as poor stability and significant solution errors. By fully considering the influence of composition changes on various factors in the separation process, the traditional UV flash algorithm has been improved by simplifying the iterative equations and optimizing the calculation order, and applied to the dynamic simulation of distillation columns. A dynamic simulation model of distillation columns based on the improved UV flash algorithm was proposed. The improved UV flash algorithm was verified using four flash cases under different operating conditions and one actual industrial distillation column case. The results show that the improved UV flash algorithm outperforms the traditional UV flash algorithm in terms of computational efficiency, accuracy and robustness. Compared with the conventional UV flash algorithm, the calculation results for the distillation column model based on the improved UV flash algorithm are closer to the simulation results of the HYSYS distillation column dynamic model. The calculation time is reduced from 1.331 s to 0.310 s, and the relative errors of the simulation results of the intensity variable and the final product composition are less than 1% and 1.5%, respectively. The results show that the improved UV flash algorithm is more suitable for calculations of the dynamic simulation of a distillation column, and can provide a reliable flash algorithm for the establishment of a digital twin model of the dynamic process in a distillation column.
京公网安备11010802044758号