针对AspenPlus在乙烯淤浆聚合过程的回流工况中无法收敛的难题,提出建立乙烯淤浆聚合的联立方程模型。基于多活性位的Ziegler-Natta催化剂催化乙烯聚合的机理,依据矩方法理论建立乙烯聚合的过程模型。通过瞬时分子量分布方法,模拟计算了聚乙烯的分子量分布。针对PC-SAFT中的序贯求解模式,提出了基于数据驱动的物性计算方法,通过Kriging估计函数,建立了适用多个工况的、可联立求解的物性计算模型,非聚合物体系误差低于03%,聚合物体系误差控制在2%左右。联立方程模型的求解打破了传统的物性计算与流程计算的嵌套式求解模式,实现了真正全联立模拟求解。针对5个不同牌号工况,联立方程模型可以顺利收敛,模拟结果与Aspen计算结果进行了对比。乙烯淤浆聚合模型的全联立求解提高了模型的收敛性,各工况的模拟验证了Kriging函数的PC-SAFT模型的准确性和稳健性。
Ethylene slurry polymerization process is an important process for the production of high-density polyethylene (HDPE). In this paper, an equation-oriented model is developed with simultaneous solving approach for the process simulation. By using the method of moments, a mathematical model was developed for ethylene polymerization with multiple-site-type Ziegler-Natta catalyst. The molecular weight distribution (MWD) of polyethylene was calculated by using the method of instantaneous distributions. To facilitate the process simulation with equation-oriented approach, a data-driven model was presented to substitute the complex PC-SAFT property calculation model. Kriging estimation was used to establish the property calculation model, which was verified to be suitable for multi-grades. A completely simultaneous method was applied to simulate the process by breaking the traditionally nested thermodynamic and process loops, Comparisons with commercial software were also conducted, Simultaneous simulation approach shows the major advantage in convergence, and the simulation results show the preciseness and robustness of Kriging model.