Abstract: Among various types of carbon fibers, PAN-based carbon fibers have become the most significant component of high-grade carbon fiber product series due to their excellent mechanical properties. The industrial preparation of PAN fibers typically involves an aqueous suspension polymerization process. In the production of acrylic fibers, the polymerization process prior to spinning is a critical step that determines product quality. Among the factors influencing this process, the mixing effect is vital for enhancing product quality. Uniform mixing ensures the consistent distribution of reactants, thereby maintaining the stability of polymerization rate, and significantly influencing molecular weight distribution and the overall quality of the fibers.
To enhance the mixing effect within the reactor, this paper employed computational fluid dynamics (CFD) technology to model and conduct numerical simulations of the aqueous suspension polymerization process of acrylonitrile. Based on experimentally measured kinetic and rheological data, complex transport models and reaction kinetic models, such as the reaction rate equations and reaction heat equations for various substances, were transferred into Fluent software using C language to create user-defined function (UDF). The resulting CFD model integrates coupled flow mixing, mass and heat transfer, and polymerization reaction processes. The numerical simulation method was utilized to investigate the distribution patterns of velocity fields, components change, temperature fields, and polymerization products within the reactor during the polymerization process. By utilizing moment equations, the infinite-dimensional terms in the dead polymer reaction rate equations were converted into finite dimensions, enabling the solution of polymerization product quality indicators such as molecular weight distribution index and molecular weight.
The simulation results briefly demonstrate the patterns of temperature distribution and monomer distribution within the polymerization reactor, highlighting the complexity and inhomogeneity of the rheological behavior of the reaction medium inside the reactor. This underscores the importance of reactor structural design. In comparing the simulation results with actual production data, the errors in inlet and outlet temperature rise, monomer conversion rate, molecular weight, and molecular weight distribution index are all within 15%. This, to a certain extent, validates the feasibility and accuracy of the numerical calculation model. This method also provides important theoretical reference for the enhancement of polymerization processes and the design and scaling-up of reactors. 

Key words: acrylonitrile, reactor, rheological behavior, computational fluid dynamics, reaction kinetics, quality index

摘要： 聚丙烯腈纤维（PAN）的工业生产通常采用水相悬浮聚合方法，为了提高丙烯腈聚合反应器内混合效果，基于实测的流变和动力学数据，以PAN溶液聚合过程为研究对象，建立了一个耦合流动混合过程、传质传热过程和聚合反应过程的CFD模型，使用数值计算方法探究了聚合过程中反应器内速度场、浓度场、温度场以及聚合产物的分布规律。对比CFD模拟计算结果与实际工业生产数据，其进出口温升、单体转化率和分子量分布指数误差在15%之内。该模型实现了对分子量分布指数和分子量等聚合物质量指标的求解，为工业聚合过程强化和反应器设计提供参考。

关键词: 丙烯腈, 反应器, 流变行为, 计算流体力学, 反应动力学, 质量指标