Study Of Key Technique In Safety Monitoring System For Structure And The Application To The North Anchorage Foundation And Ground Of RunYang Bridge

The focus of this thesis is the key techniques of safety monitoring system for structure such as: arrangements of measured points, safety monitoring model and safety evaluation and the application to the anchorage foundation and ground safety monitoring system that is used in the working period of long span suspension bridge. The theoretic, experimental and applied studies of some concrete problems of these techniques are systematically performed in the thesis. To enhance the intelligent level of the monitoring system, computer numerical analysis is utilized largely. The main contents and research results are as follows:　1. By consulting the research results of similar geotechnical engineering, the combination of back analysis and monitoring in the monitoring of anchorage foundation and ground is suggested. The results of back analysis could be utilized to evaluate the safety of anchorage foundation and ground. The functions of the mechanical parameter errors about model errors and measure errors were built. With these functions, transferring process of model errors and measure errors to affect the outcome of back analysis was researched. The result shows that if measured points are located at the placements of high mechanical parameter sensitivity, the influence of these errors on the results of mechanical parameter back analysis can be diminished. Based on the conclusion and other disciplines of measured point arrangement, an optimal arrangement method of measured points is suggested for the first time. A program is coded and applied in an example of the RunYang Bridge. The result indicates that the optimal method is rational and easy to be put into practice.2. The utilization of statistical monitoring model as the forecast method of the behavior of anchorage foundation and ground is suggested. According to fiducial interval method, safety monitoring index can be built ultimately to detect the damage of structure based on the monitoring model. Least square adjustment method with window function is introduced to detect the damage trend as the improvement of fiducial interval method for the first time.　3. Thundering observed data could be caused not only by the failure of structure but also by the problem of monitoring system. It is difficult to estimate the reason of the observed data with traditional method. To solve this problem, a fuzzy cluster method based on grey relation degree is put forward. If the relation degree between thundering observed data series and other observed data series decreases dramatically, the probability of structure damage could be estimated as little.4. The usage of neural network to built monitoring model is researched in detail. The structure and principle of BP neural network are analyzed and some improvements are presented: An adaptive learning rate algorithm that can accelerate the training speed of traditional BP algorithm is suggested; The application of wavelet neural network in safety monitoring model is studied and one kind of wavelet neural network is built for that purpose; a n ew type of neural network, grey neural network, is suggested tocombine the capability of neural network and knowledge of human being. All these different kinds of neural networks could be functional in different working period of long span suspension bridge.5. To evaluate the safety of the anchorage foundation and ground, neural network is introduced to obtain the preliminary dada of back analysis. The process of back analysis will be accelerated through this way. Genetic algorithm is utilized to avoid the disadvantage of the use of preliminary dada. To diminish the influence of measure errors produced by inner measurement, a new target function for back analysis is put forward on the basis of grey relation degree.6. To evaluate the safety of the anchorage foundation and ground, finite element method techniques are applied to the analysis of RunYang Bridge's north anchorage foundation. The methods to simulate the process of digging are discussed. Safety monitoring model of north anchorage foundation is built. The calculation results indicates that the depth of parent rock influences the displacement of anchorage foundation deeply while the displacement is not sensitive to the change of the friction coefficient between the foundation and the groundsill due to the high weight of the anchorage. Some techniques studied in this thesis are also applied to the anchorage foundation.