Hierarchical Redundancy-Based Monitoring and Robust Seismic Control Systems

[9909247]研究了位于自然灾害易发地区的具有不确定动力特性的桥梁、塔楼、烟囱、水坝和建筑物等民用结构的基于yanga冗余的监测系统。这种分层智能监控系统被设计成一个集成的实体，由三个主要软件组件组成，这些组件之间实时无缝交互，并与物理结构接口。这些组件将数据处理和决策能力组织到屏蔽层的层次结构中，每一层都采用分散的信息处理单元拓扑结构。层的屏蔽提供了数据抽象，每层中的分散拓扑创建了冗余机制。反过来，这些因素提高了监控系统的整体性能和可靠性。监控系统的要素被称为多层专家模块（MEM）、多层识别模块（MIM）和多层控制模块（MCM）。多层专家模块协助进行结构健康监测（在“正常”时间和地震期间），并使用神经模糊逻辑技术采用先进的损伤检测和评估。多层识别模块主要负责维护系统模型的活动数据库，并对其进行适当的更新。多层控制模块中开发的控制信号由控制系统硬件在实际结构上实现，使用诸如主动肌腱，主动变刚度系统，主动质量阻尼器，主动基础隔离器等设备，或这些设备的组合。通过基于模糊逻辑的决策支持，实现了MIM中的模型选择和MCM中的控制算法选择。该项目的主要目标是开发一个复杂的专家系统，该系统可以监测结构健康状况，并且可以在不准确的感官数据下检测和适应需要纠正措施的故障，并建立一个可靠的识别系统，该系统与物理结构接口，并且可以在存在结构不确定性的情况下实时合成准确的结构模型。调查涉及结构监测实践的几个关键方面，将有助于发展更安全、更可靠的基础设施。

9909247YangA redundancy-based monitoring system for civil structures such as bridges, towers, chimneys, dams and buildings with uncertain dynamic properties located in regions that are susceptible to nature disasters is investigated. This Hierarchial Intelligent Monitoring System is projected as an integrated entity consisting of three principal software components which interact seamlessly with each other in real time and are interfaced with the physical structure. These components have data processing and decision-making capabilities organized into a hierarchy of shielded layers, with each layer employing a decentralized topology of information processing units. The shielding of the layers provides data abstraction and the decentralized topology in each layer creates a mechanism of redundancy. In turn, these factors improve the overall performance and reliability of the monitoring system. The elements of the monitoring system are referred to as the Multilayer Expert Module (MEM), the Multilayer Identification Module (MIM), and the Multilayer Control Module (MCM).The Multilayer Expert Module assists in structural health monitoring (both in 'normal' time and also during earthquakes) and employs advanced damage detection and assessment using neuro-fuzzy logic techniques. The Multilayer Identification Module is primarily responsible for maintaining an active database of system models, which are updated appropriately. The control signals developed in the Multilayer Control Module are implemented by the control system hardware on the actual structure, using devices such as active tendons, active variable stiffness systems, active mass dampers, active base isolators, etc., or combinations of these. Model selection in the MIM and control algorithm selection in the MCM are achieved via fuzzy logic based decision support.The main objectives of this project are to develop a sophisticated expert system which can monitor structural health, and which can detect and accommodate failures needing corrective action despite inaccurate sensory data and to build a reliable identification system which interfaces with the physical structure and which can synthesize accurate structural models in real time, despite the presence of structural uncertainties.The investigation addresses several key aspects of structural monitoring practice and will contribute to the development of safer and more reliable infrastructures.