随着计算机与数值计算技术的迅猛发展，计算机模拟逐渐成为一种研究气固流态化的重要方法。在气固两相流的数值模拟中，MP-PIC（Multiphase Particle-In-Cell）作为一种粒子与网格混合的方法，兼具了效率与精确性：对颗粒的离散处理使其在计算、分析颗粒行为时比连续方法更具优势；以少量计算粒子代替大量属性相同的颗粒并以固相应力模型模拟颗粒间的碰撞显著降低了计算量并提高了计算效率。因此，本论文采用MP-PIC实现流态化反应器的快速模拟，并探讨进一步提高其模拟精度和效率的途径。作为研究基础，本论文首先整合了一套现有单相PIC代码与开源软件OpenFOAM构建了功能完整的MP-PIC计算程序FOAM-PIC，计算结果与已有软件一致验证了FOAM-PIC的正确性。固相正应力模型在MP-PIC中表征颗粒间的相互作用，而传统模型并未考虑介尺度结构的影响。为了提升MP-PIC方法的模拟精度，本论文引入EMMS（Energy Minimization Multi-Scale）理论，量化了气固流态化中的介尺度结构特征，构建了非均匀固相应力模型。以此模拟典型流态化反应器，证明了其精度优势和在粗网格与小粒子数下的计算稳定性。模拟对模型参数敏感性显著降低，能大幅提高MP-PIC模拟的计算效率。传统离散粒子方法中，颗粒的定位、颗粒属性在非规则流体网格上的插值复杂耗时。对此，本论文借鉴颗粒网格法与重叠网格法的思想，提出了一种精确高效的网格映射算法。借此气相与固相可以分别在不同的网格中计算并耦合，即在均匀笛卡尔网格中高效地跟踪颗粒并通过考虑两相网格中各单元被另一套网格切分出子网格的位置、重心信息，建立两相流场的精确映射。网格映射方法在保持计算精度的同时也能够大幅提高颗粒计算的效率，促进流态化反应器的快速模拟。应用本算法的MP-PIC模拟循环流化床提升管、鼓泡床、湍动床等均效果良好，验证了该算法处理复杂构体和非规则网格时的可行性和正确性。本论文还完成了FOAM-PIC程序的优化以及并行化，并分析了影响FOAM-PIC计算效率的因素。优化后的程序在对工业级的MIP提升管反应器以及煤气化炉的模拟中验证了其实用性。论文最后总结了本研究的主要成果，并展望了在模型、算法与程序实现方面的进一步研究。;With the rapid development of computer and numerical simulation technology, computer simulation has gradually become an important method to study gas-solid fluidization.Of numerical methods of gas-solid two-phase flows, the Multiphase Particle-In-Cell (MP-PIC) method, as a hybrid simulation method of particle and grid, is advantageous in accuracy and efficiency. The discrete processing of particles gives it more advantages than the continuous method in the calculation and analysis of particle behavior. A small amount of computational parcels is used to replace a large number of particles with the same properties and the solid stress model is used to simulate the collision between particles, which significantly reduces the computational burden and improves the computational efficiency. Therefore, MP-PIC method is adopted in this thesis to realize the rapid simulation of fluidized reactors and the ways to further improve its simulation accuracy and efficiency are explored.As the basis for research, a fully functional MP-PIC simulation program FOAM-PIC is firstly integrated by an existing single-phase PIC code and the open source software OpenFOAM. The simulation results are consistent with existing software to verify the correctness of FOAM-PIC. The solid stress model represents the interaction between particles in the MP-PIC method, while the influence of mesoscale structure is not considered in the traditional model. In order to improve the simulation accuracy of the MP-PIC method, the thesis introduces the Energy Minimization Multi-Scale (EMMS) theory to quantify the mesoscale structure characteristics in gas-solid fluidization and constructs the heterogeneous solid stress model. Typical fluidized bed reactors are then simulated with verification of this new solids stress model in terms of its precision and computational stability under the settings of coarser grid and smaller parcel number. The sensitivity of simulation to model parameters is significantly reduced, which can greatly improve the computational efficiency.In the traditional discrete particle method, the particle locating and the interpolation of particle properties to irregular fluid cells are complicated and time-consuming. In this thesis, a precise and efficient grid mapping method is proposed based on the ideas of particle grid method and the overlap mesh method. In this way, the gas phase and the solid phase can be calculated in different grids and then coupled, that is, the particles can be tracked efficiently in a uniform Cartesian grid, and the precise mapping of two-phase flow field can be established by considering the location and barycenter of each element in the two-phase grids segmented by another set of grid. The grid mapping method can greatly improve the efficiency of particle calculation and accelerate the simulation of fluidized reactor while maintaining the computational accuracy. Simulation results of the MP-PIC with the mapping method of a circulating fluidized bed riser, a bubbling fluidized bed and a turbulent fluidized bed are all excellent, which proves the feasibility and correctness of the method when dealing with complex structures and irregular grids.This thesis also completed the optimization and parallelization of FOAM-PIC program, and analyzed the factors affecting the computational efficiency. The optimized program verifies its practicability in the simulation of an industrial scale MIP riser reactor and a coal gasifier.In the end, the thesis summarizes the main achievements of this research and looks forward to the further research in the aspects of model, algorithm and program implementation.