摘要
提出大型实训船火灾情景下的人员逃生路线规划方案。首先,通过PyroSim软件建立“育德”轮船舶三维模型,对学员生活层(甲板D)中的船舱宿舍、图书阅览室、被服间三种典型场所进行火灾情景模拟仿真,并将舱室内平面环境离散化为二维栅格图。接着,优化蚁群算法,设计曲线圆滑的路径,使逃生路径更符合人员行走规律;将单一起点单一终点改为单一起点多个终点,实现不同情景下的多条逃生路径规划。然后,基于Pathfinder及规划的路径,模拟船舱宿舍情景下的人员疏散,结果显示,在有路径指引下人员疏散时间由107 s缩短至96 s,疏散效率提高。最后,将甲板D分为8个区域,分别针对三种典型场所的火灾情景,确定了各区域在0、120与240 s时的最优疏散路径及可选择疏散路径,并据此制作逃生指导手册。
This paper aims to plan evacuation paths in fire scenarios to address the issue.Firstly,the PyroSim software is utilized to establish a 3D model of the training ship"YU DE",focusing on deck D,which serves as the student living floor.Three typical fire scenarios,such as cabin dormitory,library,and linen room,are then selected for simulation.Afterwards,the pots-fire flat ship cabin environment is discretized into a two-dimensional raster map.Secondly,the enhanced ant colony algorithm is employed to design smooth curved paths that align better with pedestrian movement patterns,ensuring efficient escape routes.By replacing the single starting point and endpoint with a single starting point and multiple endpoints,multiple escape routes can be planned based on changing scenarios.Thirdly,based on Pathfinder and planning path,the evacuation of people from cabins scenario is simulated.The results show that the evacuation time is reduced from 107 s to 96 s under the path guidance,and the evacuation efficiency is improved.Finally,deck D is segmented into 8 regions,and optimal evacuation routes for each region at 0,120,and 240 s are determined for three typical fire scenarios,ie,cabin dormitory,library,and linen room,respectively.An evacuation manual is then developed to offer guidance for evacuating the"YU DE"ship in fire scenarios,with the aim of enhancing safety and evacuation efficiency.
作者
王敬钰
赵强
徐生磊
WANG Jingyu;ZHAO Qiang;XU Shenglei(Navigation College of Jimei University,Xiamen 361021,China)
出处
《集美大学学报(自然科学版)》
2025年第4期364-373,共10页
Journal of Jimei University:Natural Science
基金
福建省自然科学基金项目(2023J01326)。
关键词
实训船
三维模型
火灾情景
疏散路径
蚁群算法
training ship
3D model
fire scenario
evacuation routes
ant colony algorithm
