Study On Crowd Evacuation Simulation Models For Semi-Submersible Accommodation Platform

At present,in urban life and special industrial activities,there are increasingly scenes where a large number of people gather.For example,large public buildings,transportation hubs,and offshore support platforms with relatively narrow space.Once the pedestrians are organized and evacuated improperly,serious consequences and major social impacts will happen.Therefore,research on pedestrian evacuation have become increasingly significant.The crowd evacuation simulation model can study the state of movement and evacuation characteristics of pedestrians in various evacuation scenarios,and the simulation results can be a reference for making an evacuation plan.In recent years,many scholars have proposed many pedestrian flow models and made some significant progress,but there still are many problems to be studied.Therefore,it is necessary to improve original models of pedestrian flow for different purpose.This paper reviewed the research on pedestrian flow and introduced the basic framework and solving methods of two typical models of pedestrian flow.Then,two types of pedestrian flow models were induced and improved respectively.The improved models were both applied to simulate the pedestrian evacuation in typical evacuation scenarios and offshore support platforms.Finally,the simulation results of the two models were compared.(1)Floor field cellular automata model considering pedestrian obstructionThis paper proposed a floor field cellular automaton model which considered pedestrian obstruction.The static field of the model was constructed by the shortest evacuation time.At the same time,in order to consider the obstacle between different types of pedestrians in the bidirectional pedestrian flow,the fast sweeping method for calculating the static field was improved.In this model,different types of pedestrians regarded each other as movable obstacles,leading to the decrease of the value of local static field.Then,this model was applied to typical evacuation scenarios,sucn as corridors,T-shaped channels,orthogonal channels and barrier-free rooms.Specifically,the influence of δ,the increment of local background field value,on evacuation state of pedestrians with different average density was studied.Furthermore,the relationship between evacuation variables was analyzed and the effect of exit width and exit position on the evacuation state and evacuation results were researched,which reproduced self-organized phenomenon of pedestrians.The results showed that δ affects the evacuation state of pedestrian flow to some extent.When considering this influence,the model can better reproduce the selforganized phenomenon of pedestrian flow and improve the evacuation efficiency of pedestrian flow with the middle and low density in the channel.In different evacuation scenarios,the average density was not the same when phase transition occurred,reflecting the difference of traffic efficiency of pedestrian under different evacuation scenarios.The results also showed the width of the exit has a significant impact on the evacuation time.When the exit width was less than 2 meters,the evacuation time and the exit width have a inverse relationship.Besides,evacuation efficiency is relatively higher by setting the exit near the central axis of the boundary.At last,this model was applied to the actual evacuation scenario.The process of pedestrian evacuation was simulated,and the simulation results were compared with the experimental results.The results showed that the simulated evacuation time,pedestrian evacuation route and the location of the congestion phenomenon are consistent with the experimental results,which verifies the rationality of the model.(2)Fluid dynamics model of pedestrian flowIn this paper,the RDUO model proposed by Huang and the non-local pedestrian flow model proposed by Zhang were comprehensively summarized.Based on the two models,an improved fluid dynamics model of pedestrian flow was proposed.The governing equations of the model included a conservation equation with non-local flux,a constraint equation of path selection of pedestrians derived from the user conservation condition of the traffic flow of dynamic balance distribution,and a eikonal equation solved by the fast sweeping method.To solve the model numerically,a program framework of joint modeling was established by combining computing core of PDE module of COMSOL with MATLAB programming.Afterwards,the model was used to simulate the typical evacuation scenarios,such as corridor channels,T-shaped channels and orthogonal channels.And,the typical evacuation characteristics of pedestrians in each evacuation scenario were obtained.The results showed that this model successfully reproduced the self-organized phenomenon of pedestrian flow,and the simulation results were basically consistent with the experimental results and related literatures,which verified the rationality of the model.Lastly,this model was applied in the evacuation scenario of sea life platform.The results were compared with the results of the evacuation experiment and the simulation evacuation time was only 2% different from the evacuation time obtained by the experiment.It meant that this model had high simulation accuracy under the premise of reasonable parameter selection.The congestion location obtained by the simulation was well consistent with the location of the congested area in the evacuation experiment,which successfully simulated the evacuation process of pedestrian flow in evacuation scenario of sea life platform.The crowd evacuation simulation model proposed in this paper had relatively high simulation accuracy and computational reliability,which can be a reference and theoretical basis for crowd evacuation simulation and facility design in actual engineering.