Subway Tunnel Structure Deformation Monitoring Of Several Key Technologies

As China’s urbanization and rapid urban traffic congestion problem is getting worse, and the subway is an ideal way to solve the urban traffic congestion, almost all capital cities in the country are either already built a good subway, either being built or planned construction of the subway, and even some non-capital cities such as Suzhou, Changzhou and Wuxi have started the construction of the subway, so visible in the near future, the Metro will become a popular means of domestic cities to solve traffic congestion. In order to protect the safe operation of the subway, you need real-time grasp and understand the changes in the structure of the tunnel, and the need for safety hazards that may exist in the subway tunnel structure in a timely manner to govern, so the structure of the subway tunnel deformation monitoring is essential. This paper has studied the control network stability judgment, the basis of automated data adjustment processing, and monitoring data to predict of the subway tunnel structure deformation monitoring, specific contents are as follows:1)Manual monitoring in the subway tunnel structure, since the subway tunnel structural characteristics and requirements laid points, resulting in subsidence monitoring benchmarks in the subway network is only one degree of freedom often occur in the case. In this paper, the stability of the commonly used method to judge (the average gap method, tolerance method and the t-test) is equal to1and greater than1for the freedom of the net settlement basis for stability analysis and discussion.2) In the subway tunnel structure Automatic deformation, General adjustment software does not have a three-dimensional wire adjustment function, so it can only be separated from the plane and altitude adjustment calculation. Artificially separated one-dimensional, two-dimensional adjustment, s o in theory, it destroys the reference point between integrity and unity. If only the use of an automated total station monitoring, resection can be used for three-dimensional wire network adjustment processing. View of this situation, the horizontal direction, the zenith distance and slope away from the observed value as the initial value, the use of indirect adjustment to do rigorous adjustment and Accuracy Assessment, and finally deduced resection three-dimensional adjustment model. Software development due to the three-tier development mode into the interface layer, logic layer and data access layer, while significantly reducing the dependencies between the layers, the layers more independent, even if a layer has changed, non layer will affect the other layers, greatly improving the stability of the software, so this paper has used a three-tier structure mode program to develop three-dimensional adjustment of the resection.3) In the subway tunnel structure Automatic deformation, the structure of the subway tunnel short-term small amount of deformation monitoring data showed weak signal. Automated total station during the measurement, by the impact of vibration and speed of the operating metro, making the observation result contains a lot of noise, the performance for strong noise. This paper has used wavelet analysis filter to filter out the noise monitoring results. BP neural network is the most widely used, the most widely used neural network model, this paper uses BP neural network to predict the deformation monitoring data of the subway. BP neural network has advantages, but there are also the following disadvantages:easy to fall into local minimum, slow convergence and network architecture design blind. In this paper, wavelet analysis to transform BP neural network model is constructed wavelet neural network model, and then using wavelet neural network model for deformation monitoring data for prediction.The research in this thesis several key technologies used in the Nanjing subway tunnel structure monitoring, meaningful conclusions drawn as follows:1) When the monitoring network is equal to1degree of freedom, tolerance method for stability using good judgment, because the average gap method and t-test of hypothesis testing-bit value is relatively large and unstable.2) In the realization of the monitoring data adjustment function above the three-tier software development process; three-tier development model will be developed into a software program interface layer, logic layer and data access layer, the interface layer, logic layer, and more independent data access layer, so that the stability of superior software.3) Using wavelet analysis can be filtered in a lot of noise monitoring data, monitoring data extracted trends.4) Through the monitoring data to predict found BP neural network and wavelet neural network has high prediction accuracy; due WNN wavelet function through the BP neural network prediction accuracy slightly higher.