大型跨海桥梁的船撞风险评估、智能监控与安全防护研究

For the prominent risk of ship collision with large-scale sea-crossing bridge in complex sea environment and the drawbacks of existing security protection methods, this project firstly aims to identify the deep factors, which lead to the ship bridge collision hazard, through big data analysis of ship navigation path on the sea in typical bridge area. Based on that, a 2D risk model of ship yawing will be established and the method of building a monitoring system on anti-ship collision by integrating different sensing techniques will be studied. Secondly, based on multi-source heterogeneous data fusion, ship dynamic identification theory and estimation method of ship bridge security situation will be considered, including the development of intelligent algorithm for target recognition and tracking and proposing the judgment criterion of abnormally navigated ship as well as predicting the movement variation of ship . Thirdly, the method of interception control on runaway ship and disaster evolution analysis will be investigated, involving the construction of optimization algorithm for dynamic interception, the establishment of a coupling mathematical model for intercepting process to reveal the motion control mechanism and evolution rule of ship intercept. Fourthly, combined with the experimental analysis, the evaluation method of the damaged state of the bridge under the collision with low-speed ship will be discussed, characteristics of the dynamic response and damage mechanism of beam and pier will be revealed and the judgment criterion as well as appraisal procedure of damaged components will be proposed. Finally, based on the system integration and application fusion of research results, the problem of anti-ship collision with sea-crossing bridge will be solved as well as the scientific basis and technical support for the protection of the collision safety of large-scale offshore projects will be provided.

本项目针对大型跨海桥梁在复杂海域环境下船撞风险突出的特点和现有安全防护方法的不足，首先，基于典型桥区海域的航迹大数据分析，离析和揭示导致船撞桥危险的深层次因素，建立船舶偏航的二维风险模型，研究集成不同感知技术构建防船撞监测系统的方法；其次，基于多源异构数据融合，研究船舶动态识别的理论及船桥安全态势估计的方法，包括发展目标识别跟踪的智能算法，建立异常航行船舶的判断准则，预测船舶的运动状态；第三，研究失控船舶拦截控制及灾情演变分析的方法，包括创建动态拦截的优化算法，建立拦截过程的耦合数学模型，揭示船舶拦截的运动控制机理及演变规律；第四，结合试验分析，研究在船舶低速碰撞下的桥梁受损状态的评估方法，揭示梁和墩构件的动力反应特征及破损机理，建立构件的受损判定准则及评估方法。最后，通过研究成果的系统集成及应用融合，解决跨海大桥的防船撞难题。本项目将为保障大型海上工程的防船撞安全提供科学依据和技术支撑。

本项目以复杂海域环境下的大型跨海桥梁为工程背景，针对大型跨海桥梁在台风频发、浪流多变、航船密集桥区海域的船撞桥风险突出的问题和现有防灾能力的不足，从风险识别、智能监控和安全防护三个方面，系统开展了海域桥梁主动防船撞的理论方法与关键技术研究。.在风险识别方面的研究中，通过现场实测和事故数据统计识别找到了海域船撞桥致灾因素的分布规律。获得了过桥船舶的航迹带宽度及其分布模型。建立了多层次模糊综合风险评价模型，计算出典型跨海桥梁的年船撞风险平均指标为4.22，提高了风险识别水平。基于概率论和船舶操纵理论建立了考虑多种风险源的船撞桥概率模型，并以海上长桥为例经过精确分析后，对比传统分析，采用该模型得到撞击风险是现有规范计算结果的近20多倍。.在智能监测方面的研究中，针对海域复杂多变的环境，基于以Darknet网络为基础的Resized-YOLO算法，研发了一种海域船舶实时识别方法。构建了船舶深度特征矩阵，同时加入了最大池化层，并引入局部神经元的活动创建竞争机制，从而大大提高了检测的精度和效率，对海域船舶目标的识别准确率可达95%。基于计算机视觉定位算法、KCF算法、状态识别优化算法以及风险评估理论，提出了一种异常航态船舶识别方法，并研发了一套适用于跨海大桥的船撞桥预警系统软件。.在安全防护方面的研究中，考虑了船-桥-海的动力演化过程和水位变化及冲刷效应，建立了船撞桥全尺度精细化数值模型，并设计完成了相对应的船撞桥模型试验，研究了构件整体和局部的线性及非线性力学行为演变机理，找到了桥受撞能量主要成分分布在0～62.5Hz范围的规律，撞击转移能量大部分耗散在桩基。提出建立了基于磁基弹性体的分级防护方法，研发了一种针对不同吨位船舶撞击的磁基弹性体防撞设施，并通过实验表明该防护装置可使桥墩受撞的最大应力减小80.9%。构建了船-桥-海一体化安全防护的方法，并已应用在宁波舟山港主通道跨海大桥中。