Research On Some Problems Of Real-Time And In-service Inspection And Health Monitoring On Bridge

After a bridge having beening constructed and opened to traffic, its materials will be deteciorated or aged gradually for the weather and environmental factors, and its strength and stiffness will degrade under the long-term of statistic and dynamic load, thus it will not only endanger the safety of the traffic, but also shorten the life span of the bridge. On the base of bridge structural inspection and health monitoring, safety assessment is important in bridge daily service.Most of long-term inspection is still based on periodical inspections by person in which limitation is obvious: (1) Damage can not be detected during the period of non-inspection; (2) Some location on the structure can not be touched by person; (3) More relative work and expensive; (4) Inspection result is more subjective.Though traditional inspection method is simple and easy, and there is a set of practical inspection and estimation codes as well, the inspection period is long and the result is undirect. Also closing the traffic which is always requested in traditional way will reflect the social and economic life. With the built-up of more and more long-span bridges, apparently, it is difficult to hold all the information that may globally reflect the bridge health condition and more difficult to estimate the safety factors and deterioration tracks systematically and directly by traditional visual inspection, maintenance and routine local inspection on such complicate and large-scale structures. It is necessary to set up a health monitoring system that can evaluate all the bridge structural capacity and safety for insurance of the bridge security and durability.Remotely and intellectually inspection, monitoring and safety estimation method is the tendancy of bridge management, it also gains the most attention, in civil engineering as well as relative disciplines. One of the critical and difficult problems is to set up scientific and practical elevation expert system from sea-scale monitoring data.Most of data process in bridge monitoring research focus on model analysis which usually calculates the structural dynamic characters or stiffness and softness based ondynamic test data, such as model, accelerometer, etc. In such techniques, mapping and comparing the difference between the limited test data to the global structural model by expanding or shrinking matrix, the location and extension can be assessed by the changes of stiffness, softness and dampness. Research and practice work on the process and analysis techniques based on dynamic test are progressing, especially by neural-net work and genetic algorithm technology.It should be noticed that dynamic model reflect the global character of the structure, while damage and destroy always happen on local, even destroy of the whole structure occurs by just one element. So it should monitor the elements and joints change in practical bridge monitoring system. The bridge safety and durability assessments or warning reports can provide to the manage department by monitoring the critical parts of the elements and joints, as well as the environment and traffic situation to avoid disaster by repairing in time.In spite of monitoring environment and traffic, there are two basic kinds of sensors in bridge monitoring system: one for dynamic monitoring, another for statistic monitoring, also it should be regarded as dispend each other in test. The goal of the bridge monitoring system is to evaluate the situation of the bridge from these sea-scale dynamically-changing data of large monitoring items.The calculation and design model can't totally and correctly consider the characters of the complicate structure, as well as various random factors in the construction and service, the test results are always far more different with the design results, also various environmental factors may change the characters of the materials and the structure. All these changes are uncertainty and can not be analyzed by traditional FEM model. Structural damage recognition is limit only based on calculation and design model analysis. So it is difficult to correctly assess the condition and characters of the structure which are changing timely based on model analysis by limited test data, especially as one part of the bridge management system, monitoring system should be able to sign out warning when there is damage.Single factor recognition is direct and simple in estimating the local situation of a single element, it can also provide warning report to the bridge manage system in time.But the changing and the damage in the structure are not equal in various monitor items, how to obtain the comprehensive information from multi-dimensional data system, how to assess the global health situation of the bridge from various monitor item and to improve the management and judge speed are the important perspects in promoting the capacity of the bridge manage system.With the development of structural real-time health monitoring technology, large volume and high dimensionality of structural and environmental data can be collected from continuous monitoring system by various more and more advanced and sensitive sensors. Reviewing the bridge health monitoring systems already built in the world, there is no really practicl data-based monitoring expert system to estimate the actual condition of the bridge. The difficulty is how to set up a precise model from the data to reflect the actual status of the bridge which is far from the design result for a large of random reasons. Retrace the traditional inspection and bridge management, it is concluded that: damages and changes exist in most of in-service bridges. Disaster can be happened in the whole scale, also can be happened in the elements. The practical bridge monitorning system should be based on monitoring of the elemental damage development, estimating the service condition of the bridge and giving the warning reports. Here the critical problem is to define the thresh-hold data for warning.This dissertation works on setting up in-service and real-time inspection and monitoring system. On the base of local structural and elemental real-time monitoring on damage of the bridge, the reflection on the global bridge securityby local damages and their development is analyzed. Also safety-estimation expert system by fuzzy logic control theory is established by evaluating the damage development and global main factor analysis based on time-series cubic data sheets. As well as research and practice work has been carried out on the in-service inspection of a bridge. The following are the main work in this dissertation:1. Structural damages exit and develop in real-time is presented. Local and elemental structure should be monitored in real-time monitoring system as well as the development of damages of the elements and joints. It is proposed that bridge condition assessment criteria should be established based on hereditarily using the traditional inspection andestimation codes and using fuzzy logic control theory, the membership criteria for bridge condition assessment and warning control can be established. In this dissertation, fuzzy logic control rules for cracks and their development speed in stress-cast concret structure are established, its feasibility in estimating and forecasting the bridge local damage and its development is illustrated via the practical monitoring data.2. A large volumn and high dimensionality of structural and environmental data collected by sensors are chuzzy logianged with time goes by. This group of data tables is called time-series cubic data sheets. From the aspect of the global bridge, by using the maximum data changing direction principle, the multi-dimentional data dimension reduction can be performed. On the base of elemental monitoring, a global main factor analysis technique based on time-series cubic data sheets is proposed in this dissertation. Via this technique, the critical and comprehensive information can be obtained rapidly from high dimensional timing-changed data, therefore it can promote the bridge managers' global judgement. In this dissertation, the safety of a bridge in different season is assessed via reai-time main factor analysis by using strain and cracks monitoring data, its effectiveness is illustrated via a practical monitoring data.3. Some practical research work is done on the in-service inspection of the bridge. On the base of analysis of the advantages of wireless sensors, an in-service inspection process is proposed. The results from the inspection on an in-service bridge verify that this technique is feasible and effective. Compared with the traditional inspection, the most advantages of the new technique based on wireless sensors are avoiding closing traffic, as well as collecting, storing and processing data automatically. Wireless sensors can be installed on the structure easily and for a long time, thus the in-service collection and process data can be transfered for a long distance, inspection can be long periodically as well. Data can reflect static and dynamic characters of the bridge for its statistical meanings.