基于复杂网络理论的超大型浮动平台系统灾变动力学研究

The very large floating platform is cutting-edge equipment in ocean engineering strongly desired for natural resource exploitation in deep seas. It is a typical complex multi-modular system with massive size of structures, fluid-structure interaction, flexibly connection among modules and operating in violent environment. This project will study the catastrophic phenomena of the floating system aiming at the development of pre-warning methodology by using the complex network theory in order to improve the design safety and reliability. The study will firstly develop the modeling methods of nodes and links which will be eventually integrated into a network dynamic model based on the network theory. Then the catastrophic dynamic characteristics of the floating platform will be investigated, including the phenomena of jump, synchronization, and network synergic effect, in order to understand the catastrophe mechanisms and features. Further investigation will focus on virtual network reconstruction techniques by mapping response time series onto virtual networks. The pre-warning and prediction methods of catastrophic events of the floating platform will be developed by capturing and extracting the key topological features of the virtual networks in time evolution. The project aims to provide a new theoretical analysis tool for catastrophe identification, applicable to catastrophe dynamics forecasting of complex systems in many engineering fields.

超大型浮动平台是立足深海资源开发利用的具有前瞻性和探索性的重大高端海洋工程装备。其结构具有大尺度、流固耦合、多模块连接、作业环境恶劣等特点，是一个典型的复杂多模块系统。本项目运用复杂网络理论来设计超大型浮动平台系统的灾变识别和预警机制，提高系统的安全性和可靠性。首先建立超大型浮动平台系统的节点和连接等效力学模型及集约合成系统网络动力学模型；运用网络灾变理论，研究超大型浮动平台系统动力学响应的阶跃现象及形成机理、系统网络协同效应同步化及响应变迁等灾变动力学特性，分析超大型浮动平台系统灾变特征；然后开发基于响应时间序列的虚拟网络重构技术，通过观察抽取虚拟网络拓扑特征提出超大型浮动平台系统的灾变预警方法。通过对超大型浮动平台这一典型的复杂网络结构开展系统性灾变动力学研究，以期为灾变领域提供新的理论分析及预警方法，该方法可广泛应用于其他复杂工程结构的灾变动力学研究，具有广泛的应用价值。

项目以超大型多模块浮动平台为研究背景，系统性地开展工程网络结构的动力学建模以及基于复杂网络重构技术的工程结构灾变预计识别方法研究。围绕项目的关键科学问题，开发了网络化工程结构的模块化快速建模方法，并通过南海科学试验平台实测结果对比，验证了新方法的可靠性；基于网络动力学模型，提出了超大型多模块浮动平台连接器的关键参数多目标优化方法，并指导设计我国第一型多模块科学试验平台的关键部件连接器；开展了不同海况及系统参数下多模块浮动平台的非线性动力学响应特性，揭示了耦合网络系统发生阶跃灾变时的频率转换、同步破缺、协同破缺等复杂动力学现象；提出了符号网络、重现网络等一维和多维时间序列的复杂网络重构技术，突破噪音干扰、多源时间序列混杂等技术难题；针对耦合系统网络灾变特征，基于提出的复杂网络重构技术，开发了超大型多模块浮动平台灾变识别及预警机制，并将该方法应用于癫痫等复杂疾病的诊断上。本项目建立的非线性网络结构建模分析方法和灾变识别预警方法不仅限于海洋工程领域，还可以拓展到其他复杂工程结构领域。