基于人流模拟的教育建筑空间布局疫情防控效率比较研究

基于人流模拟的教育建筑空间布局疫情防控效率比较研究

Comparative Study on Epidemic Prevention Efficiency of Educational Space Layout Based on Agent-based Simulation

作者：张茜 刘畅

指导老师：刘宇波 邓巧明 周玄星 梁凌宇

00 研究背景 / Background

在当前疫情的大背景下，本研究借鉴已发表的高校防疫措施，将关注点集中在道路的人流量上，以华南理工大学五山校区为例，利用GAMA平台建立校园基础模型，模拟多个防疫措施下的校园人流轨迹，计算道路人流密度，并以此为基础衡量防疫措施的有效性，为高校制定防控管理措施提供依据。

In the context of the current epidemic, referring to the published prevention measures related to the resumption of school, taking the Wushan campus of South China University of Technology as an example, the GAMA platform was used to build the campus foundation model in this study, to simulate the life trajectory of campus population under the control of multiple epidemic prevention measures, and we calculated the density of road population to measure the effectiveness of epidemic prevention measures, so as to provide a basis for universities to make prevention and control measures.

结合对模拟数据的分析与讨论，本研究提出了组团式和密路网校园布局模式，并将该模式应用于浙江大学紫金港校区再设计的研究探索中，借助GAMA平台对新校园内学生出行行为进行模拟，比较不同校园布局模式的防控效率差异。

Based on the analysis and discussion of the simulated data, this study proposes the campus layout mode of group and dense road network, and applies this mode to the research and exploration of the redesign of the Zijingang campus of Zhejiang University. Then, the GAMA platform was used to simulate the students’ behavior in campus and compare the difference in the prevention and control efficiency between the campus layout mode and the traditional campus layout mode.

01 建立模型 / Build the Model

在华工2019届硕士研究生胡凯所建立的“GTSMC （Gama_based Traffic Simulation Model of Campus）校园交通仿真模型”的基础上，为适应此次研究的特性，本研究在“道路人流密度计算方法”、“代理人行走规则”、“代理人选择目标及出发时间规则”等几个方面做出了调整。

Based on the "Gama_based Traffic Simulation Model of Campus" established by Hu Kai, a graduate student of the Future Laboratory, in 2019, and in order to adapt to the characteristics of this study, adjustments about "road crowd density calculation method", "agent walking rule", "agent selection target and departure time rule" and other aspects were made.

在进行道路人流密度量化时，本研究依据《校园建筑与环境疫情防控手册》中多次提到的“人与人之间的距离应至少保持1m”，把“1人/米”设为判断道路人流密度是否超标的临界值。计算式为“该条道路上的人数/道路长度”，此值大于1人/米时，即视作道路人流密度超标。考虑到现实生活中行人都是靠马路右侧行走，同一段道路上两个方向的人流距离为一个道路的宽度（大于1米）。因此，本研究将不同方向的两股人流分开统计。

When quantifying the density of road pedestrian flow, the study referred to the "Distance between people should be kept at least 1m" mentioned many times in the Campus Building and Environment Epidemic Prevention and Control Manual, and set "1 person/meter" as the threshold value of whether the density of road pedestrian flow exceeds the limit. The calculation formula is "the number of people on the road/the length of the road". If the value is greater than 1 person/meter, the density of people on the road is deemed to exceed the standard. Considering that pedestrians always walk on the right side of the road in real life, people’s distance on two walking directions is the width of the road, which is more than 1 meter. Therefore, only the crowd density on one-way roads was counted in the study.

为了反映上下课与用餐时人流的集中情况，本研究对教学楼的学生人数进行了重新分配，同时按照行走路程计算代理人的出发时间。

In order to reflect the concentration of people during class and meal, this study allocated the number of students in the teaching buildings, subdivided and modified the teaching buildings and catering buildings, at the same time, the departure time of the agent is calculated by distance according to the schedule.

代理人出发时间示意图

02 基于华工的前期研究 / Preliminary Study Based on the Wushan Campus of South China University of Technology

本研究首先搭建了华工的GAMA模型，分别进行了正常情况，以及应急时期的错峰上课、控制上课人数、食堂送餐和道路疏导等措施的模拟。

This study firstly built the GAMA model of South China University of Technology, simulated the normal situation and the measures of off-peak classes, class attendance control, canteen distribution and road flow diversion in emergency periods.

在模拟的过程中，措施强度并不完全与道路人流密度降低程度成正比，达到一定的程度后，道路人流密度的值很难再继续降低，例如，错峰上课措施中把学生分为3批和4批的结果差别不大。除了客观必然性，这可能与校园布局方式有关。华工的综合教学楼较为集中，上课时全校的学生需要从四面八方聚集到同一个地点，这将很难达到分流的目的。从减少通勤时间，增加可选范围上考虑，组团式的校园布局模式将更有利于校园防控。

In the simulation process, the intensity of measures is not completely proportional to the reduction degree of the road density. After reaching a certain degree, the value of the road density is difficult to continue to decrease. For example, there is little difference in the result of dividing students into three batches and four batches in the staggered class measures. In addition to objective necessity, this may have something to do with the layout of the campus. The comprehensive teaching buildings of South China University of Technology is relatively concentrated, so all the students need to gather at the same place from all directions to attend classes, which will be difficult to achieve the purpose of diversion. From the perspective of reducing commuting time and increasing the range of options, the group layout mode will be more conducive to campus prevention and control.

华南理工大学五山校区模拟

同时，由于华工校园道路选择较为单一，使得部分道路承担了较大的交通压力。出于这点考虑，在后面的校园设计中，本研究将采用密路网的模式，更好的起到人流分散的作用。

At the same time, As a result of the college campus road choice is relatively single, some roads bear great traffic pressure. For this reason, in the later campus design, this study will adopt the model of dense road network to better play the role of crowd dispersion.

华南理工大学无干涉条件下全天内道路超标次数统计图

03 紫金港校区再设计 / Zijingang Campus Redesign

综合浙江大学紫金港校区现状中存在的问题，本研究在重新规划设计的过程中，着眼于以下几个要点：维护原生的环境风貌，注重不同层级景观的合理尺度；规划方格形道路网，以小街区密路网的布局倡导出行多样、步行友好的校园道路体系；圈层模式的功能分区，建筑按照开放性顺序布置，各组团配备完整的公服设施，相互交叉缩短步行距离；校园周边的建筑功能多样化，减少基地与城市的割裂。

Based on the existing problems in the Zijingang Campus of Zhejiang University, in the process of re-planning and design, the study focused on the following points: maintaining the original environmental style and restrict the reasonable scale of different levels of landscape; planning a grid-shaped road network, using dense road network in small blocks to advocate a campus road system with diverse travel and walk-friendliness; the functional division follows a circle mode, buildings are arranged in an sequence of openness, each group is equipped with complete public facilities to shorten the walking distance. The architectural functions around the campus are diversified, reducing the separation between the base and the city.

多层级交通流线分析图

圈层模式功能布局分析图

浙江大学紫金港校区重新布局总平面图

组团/防疫单位分布图

04 紫金港校区的人流模拟实验 / The Simulation Experiment of Zijingang Campus

本研究搭建了紫金港校区GAMA模型，并进行了和华工相同的无措施和施加防疫措施的模拟过程。随后，选取了紫金港校区的一个防疫单位，再次进行了上述模拟实验。

The study built the GAMA model of Zijingang Campus, and carried out the simulation process same as the Wushan Campus, both non-measure and anti-epidemic measures. Subsequently, an epidemic prevention unit in Zijingang Campus was selected and all the simulations were carried out again.

浙大紫金港校区GAMA模型

浙江大学紫金港校区某组团平面图

从整体上看，紫金港组团的各项数据都为最优，说明了将学校拆分为几个部分，各自独立运行时的效果最为突出。

On the whole, the data of Zijingang Group performed better, indicating that when dividing the campus into several groups, and each group runs independently, anti-epidemic measures work most effectively.

华南理工大学五山校区数据统计图

浙江大学紫金港校区数据统计图

浙江大学紫金港校区某组团数据统计图

紫金港校区相较于华工，在各个措施的发挥效用上更为突出，但在人流密度值的统计上略高。这里可能与我们的代码设置有关，也可能是客观结果，需要进一步的模拟验证。

Compared with the Wushan Campus of South China University of Technology, Zijingang Campus is more prominent in the effectiveness of various measures, but the density of people is slightly higher. This may be related to our code settings, or it may be an objective result, which requires further simulation to verify.

05 结论 / Conclusion

这次模拟实验以及校园规划都是在防疫的角度上进行，密切结合社会现实需求。通过模拟数据的证实，初步肯定了控制上课人数、错峰上课以及食堂送餐三个措施的防疫有效性。同时通过对两个校区的横向对比，让我们看到组团式和密路网的校园布局模式在适应各类应急预案中的优势。

The simulation experiments and campus planning were carried out from the perspective of epidemic prevention, closely combined with the social needs. Through the confirmation of simulation data, the effectiveness of the three measures, such as controlling the number of students, taking classes at wrong peak and delivering meals, is preliminarily confirmed. At the same time, through the horizontal comparison of the two campuses, the study shows us the advantages of the campus layout mode of group and dense road network in adapting to various emergency plans.

本次的实验结果对于疫情期间高校的复学部署具有较强的实用意义，也提供了有支撑力的实验数据和方法。但由于各种限制，没有做到更为真实和全面的仿真，与实际情况或有偏差。在未来希望可以通过多代理建模这种较为直观的方法，辅助设计师、规划师和利益相关者，为综合判断设计合理性提供科学依据。

The results of this experiment have an alternative practical significance for the re-deployment of universities during the epidemic, and also provide supporting experimental data and methods. However, due to various restrictions, a more realistic and comprehensive simulation has not been achieved, and there may be deviations from the actual situation. In the future, we hope that multiple intuitive methods such as multi-agent modeling could assist our designers, planners, and stakeholders to provide scientific basis for the comprehensive judgment of design rationality.