摘要
为分析船舶在自由航行海域航行的事故风险,提出自由航行海域船舶碰撞概率计算方法。首先,从船舶自动识别系统(AIS)轨迹信息提取船舶航行特征点,利用航迹聚类算法对自由航行海域的船舶行为特征建模,并采用合适的剪枝方法对聚类树进行剪枝,获得某目标海域船舶的航迹分类和航路尺度特征。然后,根据航路中船舶交通流的分布特征,利用IWRAP理论对船舶碰撞概率进行蒙特卡洛仿真,计算不同航路中的船舶碰撞概率值。由试验结果可知,根据航迹聚类算法将该水域划分为10段航路;该水域的最危险航路为T1和T9,最危险的汇聚点是P1,最危险船舶种类为货船;该水域船舶碰撞事故频率为1.18起/a。仿真试验结果表明,航路聚类结果和船舶碰撞概率与目标海域实际情况一致。
To analyze the risk of ship collision accident when ship sailing in free navigational sea area,a method of calculating the collision probability was work out. First,feature points were extracted from ship AIS data,and using the track clustering algorithm to model the ship behavior,and an appropriate pruning method was adopted for clustering tree pruning to realize track classification and get characteristics of ship route. Then,on the basis of feature of vessel traffic flow distribution in the route,and using the theory of IWRAP,Monte Carlo simulation of ship collision was finished,the collision probability was calculated for each route. Finally,A case study was carried out for a certain actual sea area. The experiment results are as follows： the sea area is successfully divided into 10 routes,the most dangerous route is T1 and T9,the most dangerous intersection is P1,and the most dangerous ship type is cargo ship,the frequency of ship collision is 1. 18 / a. The simulation results show that the track clustering result and collision probability calculation conform with the actual navigational features and situation.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2015年第1期53-59,共7页
China Safety Science Journal
关键词
自由航行海域
航迹聚类
蒙特卡洛模拟
碰撞概率
船舶自动识别系统(AIS)
free navigational sea area
track clustering
Monte Carlo simulation
collision probability
automatic identification system（AIS）
