基于WSN的船舶危险品运输风险预警模型与仿真研究

发布时间：2018-08-31 18:49

【摘要】：船舶运输是世界上最主要的运输方式之一,其中有超过半数以上的船载货物属于危险品。危险品在运输过程中安全控制的复杂度高,管理难度大,事故时有发生。一旦事故发生,就会对生命、财产和环境造成重大威胁。船舶危险品运输的安全问题一直是世人所关注的热点问题。尽管目前有关部门对危险品运输的安全管理做了大量工作,但是仍旧难以彻底解决危险品运输的安全问题。对船舶运输过程中货物的安全状态进行有效监控,实现风险预警,保障运输安全,是当前亟待解决的一个现实问题。本文针对船载包装危险品,围绕其运输途中的状态监测和风险预警问题开展了以下几方面的研究工作： (1)将无线传感器网络用于对舱室内表征货物安全状态的环境信息的监控,从而达到掌握危险货物安全状态的目的,设计了适用于船上实施的无线传感器网络拓扑结构,并通过实船实验验证了其可行性。 (2)将时间序列分析方法引入危险品安全状态的预测问题中,在基于单传感器的货物安全状态信息的分析、预测中,针对传统的差分方法在时间序列平稳化过程中容易造成有价值信息丢失,影响计算精度的问题,提出了一种指数平滑法和自回归移动法的组合方法；在基于多传感器的信息融合研究中,提出了基于协整理论的船载危险品在途安全状态信息融合处理方法,该方法发挥多传感器的优势,可对危险品在途状态信息进行较为准确的预测,同时具有异常状态的诊断功能。 (3)为掌握船载危险品处于异常状态下的危险态势演变过程,提出将计算流体力学数值模拟方法引入到船舶包装危险品运输风险监测预警问题的研究中,采用该方法在船舱内有危险气体产生的情况下,对危险气体的泄漏、扩散进行了数值计算和动态仿真。同时,结合相应的物理实验对方法的可行性予以验证,实验结果表明,计算流体力学数值模拟方法应用于船舶包装危险品运输风险预警问题可行、有效。 (4)为指导船上人员对货物异常状态进行应急处置,提出一种将无线传感器网络监测和计算流体力学数值模拟有机结合的新方法,实现对船载包装危险品异常状态的态势判定。该方法一方面将无线传感器网络作为危险货物状态的监测手段,另一方面采用计算流体力学方法事先对运输过程中可能出现的典型异常状态做数值模拟,最后运用模式识别方法将二者有机结合,实现对船载危险品异常状态的态势判定。在该模式识别的相似性度量问题研究中,针对大数据量情况下现有的动态时间弯曲算法复杂度高、计算效率低的问题,提出了一种基于动态时间弯曲的模式距离滑动窗口算法。
[Abstract]:Shipping is one of the most important modes of transportation in the world, and more than half of the cargo on board belongs to dangerous goods. The safety control of dangerous goods in the course of transportation is complex, difficult to manage, and accidents occur from time to time. The whole problem has always been a hot topic in the world. Although the relevant departments have done a lot of work on the safety management of dangerous goods transportation, it is still difficult to completely solve the safety problem of dangerous goods transportation. This paper focuses on the state monitoring and risk warning of Shipborne dangerous goods in transit. The following aspects are studied in this paper:
(1) Wireless sensor network is used to monitor and control the environmental information in the cabin which indicates the safety status of the cargo, so as to grasp the safety status of dangerous cargo. The topology of wireless sensor network suitable for ship implementation is designed, and its feasibility is verified by the real ship experiment.
(2) The time series analysis method is introduced into the prediction of dangerous goods'safety state. In the analysis of the information of goods' safety state based on a single sensor, an exponential smoothing method and an exponential smoothing method are proposed to solve the problem that the traditional difference method can easily cause the loss of valuable information and affect the calculation accuracy in the process of time series smoothing. In the research of information fusion based on multi-sensor, a method of information fusion processing of ship-borne dangerous goods in-transit safety state is put forward, which takes advantage of multi-sensor and can predict the dangerous goods in-transit information accurately and diagnose the abnormal state at the same time. Break function.
(3) In order to grasp the evolution process of dangerous situation of ship-borne dangerous goods under abnormal condition, the computational fluid dynamics numerical simulation method is introduced into the study of dangerous goods transportation risk monitoring and early warning. In the case of dangerous gases in the cabin, the leakage and diffusion of dangerous gases are counted. At the same time, the feasibility of the method is verified by the corresponding physical experiments. The experimental results show that the CFD numerical simulation method is feasible and effective in the early warning of shipping packaging dangerous goods.
(4) In order to guide the shipboard personnel to deal with the abnormal state of cargo, a new method combining wireless sensor network monitoring with computational fluid dynamics numerical simulation is proposed to realize the situation determination of abnormal state of Shipborne packaging dangerous goods. On the other hand, computational fluid dynamics (CFD) method is used to simulate the typical abnormal state which may appear in the process of transportation. Finally, the pattern recognition method is used to combine the two methods to judge the abnormal state of dangerous goods on board. In the case of high complexity and low computational efficiency of existing dynamic time warping algorithms, a mode distance sliding window algorithm based on dynamic time warping is proposed.
【学位授予单位】：大连海事大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：U698;U695.292

【参考文献】