Advanced Sensors and Modelling for Next-generation Bridge Management

英国的交通网络，包括桥梁，对其经济至关重要。桥梁对经济的重要性在桥梁短暂停止使用时最为明显，例如：(I)2011年的Hammersmith立交桥和(Ii)2012年M1上3号交叉口附近的一座桥梁：(I)由于突然发现锈蚀的钢筋而关闭，以及(Ii)在桥下起火后关闭。这些关闭造成了巨大的干扰，由此给个人/企业带来的中断旅行的成本非常高。对负责的机构来说，维护英国长期运营的大型桥梁库存是一项挑战。目前桥梁管理的最新水平是由受过适当培训的检查员定期对桥梁进行目测检查。现行系统的局限性在于，它没有提供关于在役桥梁行为的定量信息，因此，决策仅基于视觉信息，这可能是主观的，取决于检查员的经验。该项目提出的改进是在现有视觉检查的基础上增加新的传感和数据建模。该项目利用了这样一个事实，即在当前的桥梁管理实践中，特定地理区域的桥梁往往会在几天或几周的时间内一起进行检查。这意味着有机会在检查区块的第一天在所有要检查的桥梁上布置定制的、易于安装的传感器，像往常一样进行检查，然后在最后一天抬起传感器。这样做具有成本效益，对现有检查制度增加的成本相对较少，但将提供获得结构行为数据的机会，以补充可视信息。然而，要实施这一新系统，需要解决两个具体领域的挑战：(A)数据收集；(B)如何以较低的财务和业务成本记录足够高质量的数据。传统的传感系统价格昂贵，需要现场供电，因此与当前的桥梁检查制度不兼容。在这个项目中，我们提出了一种新的系统来解决这个问题。(B)数据的交互和存储；如何将记录的数据转换为对决策有用的信息。桥梁在荷载作用下的挠度或振动情况可以作为其状况的指标。因此，桥梁移动方式的变化可能表明桥梁状况的变化/恶化。不幸的是，一座给定桥梁的移动幅度也受到温度等环境因素的显著影响。因此，需要一种方法将潜在的损伤或劣化对桥梁运动的影响与环境变化引起的运动变化区分开来。这种对数据的“清理”通常被称为数据标准化。在这个项目中，这种分离将使用一种新的计算机算法/模型进行，该算法/模型将作为项目的一部分进行开发。拟议系统的好处是更有效地管理桥梁，减少交通干扰。例如，拥有上述两座桥梁(一)和(二)的性能数据基线很可能会减轻困难。因为(I)问题可能会更早暴露出来，而(Ii)大桥本可以更早重新开放。在定期桥梁检查期间使用传感器数据是对当前做法的一步改变，因为获得的定量信息将使有限的桥梁维护预算得到更好的指导，并有助于提高桥梁对冲击事件的应变能力。

The UK's transport networks, including bridges, are critical to its economy. The vital importance of bridges to the economy is most evident when a bridge is briefly out of service, e.g. (i) the Hammersmith flyover in 2011 and (ii) a bridge near junction 3 on the M1 in 2012: (i) was closed due to the sudden discovery of corroded steel tendons, and (ii) was closed following a fire under the bridge. These closures caused massive disruption and the resulting cost to individuals/businesses for disrupted journeys was very significant. Maintaining the large UK bridge stock permanently in service is challenging for the responsible organisations.The current state of the art in bridge management is periodic visual inspection of the bridge by suitably trained inspectors. The limitation of the current system is that it provides no quantitative information on the in service bridge behaviour, therefore decisions are based on visual information only which can be subjective depending on the experience of the inspector. The improvement proposed in this project is to add novel sensing and data modelling to existing visual inspections. The project exploits the fact that in current bridge management practice, bridges in a given geographic area tend to be inspected together over a period of days or weeks. This means that there is an opportunity to lay out customised, easy to mount sensors on all the bridges to be inspected on the first day of the inspection block, carry out the inspections as normal then lift the sensors on the last day. Doing it this way is cost-effective, adding relatively little cost to the existing inspection regime but will provide the opportunity to obtain structural behaviour data to supplement visual information. However to implement this new system challenges in two specific areas need to be addressed:(a) DATA COLLECTION; how to record data of adequate quality with low financial and operational cost. Conventional sensing systems are expensive, require on site power and therefore are not compatible with the current bridge inspection regime. In this project we put forward a novel system to address this.(b) DATA INTREPERATION & NORMILISATION; how to convert the recorded data into information useful for decision making. How a bridge deflects under load, or how it vibrates can be indicators of its condition. So changes in how the bridge moves can indicate change/deterioration in a bridge's condition. Unfortunately the magnitude of a given bridge's movements are also significantly influenced by environmental factors such as temperature. Therefore a method to separate the effect of potential damage or deterioration on the bridges movements, from those movement changes caused by environmental variation is required. This 'cleaning' of the data is often referred to as data normalisation. In this project, this separation will be carried out using a novel computer algorithm/model which will be developed as part of the project. The benefits of the proposed system is more efficient bridge management with reduced traffic disruption. For example it is likely that having a baseline of performance data for the both bridges (i) and (ii) above would have mitigated the difficulties. Since for (i) the problem may have come to light sooner and for (ii) the bridge could have been reopened earlier. This use of sensor data during regular bridge inspections represents a step change to current practice, as the quantitative information obtained will allow better direction of limited bridge maintenance budgets and facilitate greater resilience of bridges to shock events.