停留时间分布(RTD)是化学反应工程学发展中的一个重要概念，在反应器的诊断和优化方面发挥着重要作用. 过去RTD常通过示踪实验获得，以示踪剂来追踪反应流体的流动；现今RTD的数值模拟也基于同一思路. 从流体的停留时间是反应器内流场属性的概念出发，提出了直接从流场数值模拟来导出RTD的算法. 一种是沿流线积分得到流体粒子的停留时间；第二种是基于流体粒子在流场中随机游动的概念，累计流体粒子历经各计算网格的停留时间. 以流场中一典型网格为例，证明了两种算法的等价性；用反应器的数值算例，验证了两种方法求出RTD的一致性. 还讨论了流场网格大小对准确导出RTD的影响.

Residence time distribution (RTD) is a smart concept in chemical reaction engineering, and has been playing an important role in diagnosis and optimization of chemical reactors. RTD is conventionally determined by experiment by using a tracer to track the motion of reacting fluids. Currently the numerical simulation of RTD also follows the same route. In this work, the numerical method of retrieving RTD directly from the simulated flow field is proposed in view that the RTD itself is the intrinsic property of the fluid flow in a reactor. RTD can be determined by following the streamlines of all entering fluid particles, or by accumulating the residence time increment within all control volumes that a fluid particle traverses in a random way. A numerical case of a radial flow reactor is presented to demonstrate the equivalence of these two methods.