Data Analysis And Information Management And Prediction Software System For Underground Engineering Monitoring

Aimed at the information construction for underground engineering, the monitoring methods, data analysis, modeling methods as well as 2-D monitoring information management and prediction system are discussed systematically in this thesis. Several aspects are studied as follows:(1) A 3-D non-touch convergence deformation monitoring technology for tunnel excavation based on the electronic total station is developed, and 3 methods for the calculation of 3-D coordinate are provided, such as directly coordinate computation method, 3-D survey adjustment method and measurement lines method. Corresponding software, operation procedures, applicability and cases are also supplied in the thesis.(2) Gross errors identification of stress and strain test for underground engineering is discussed. Because of absence of geometry verification condition similar to survey, these data cannot be treated according to the common adjustment principle. Wavelet analysis has been adopted to eliminate the accidental errors or systemic errors.(3) The general data modeling methods and examples for underground engineering are introduced. These models include logarithmic function, exponential function, hyperbolic function and system scientific models such as grey model, BP neural network, time series, kalman filter and so on. And ground settlement due to two box culvert pushing is simulated by Peck experiential formula.(4) BP neural network is adopted for the simulation of trend extracted from monitoring data by wavelet analysis. The prediction result is very well due to wipe off errors.(5) With high easy to use, reliable, well integration and high extensible, a 2-D monitoring information management and prediction system is developed for supervisor, construction worker and manage fellow. This software is already used in XiaoWan hydro-plant and includes four function modules, such as database management, data input and process, data modeling and prediction, graphics visualization and graph-properties dual link.(6) Seven plus one property objects are proposed to describe the complicated relations between hydraulic structures and monitoring instruments and between hydraulic structures themselves. System based on those objects has good scalability and can be used in slope, pit excavation and other geotechnical monitoring.(7) Object-oriented methods are adopted to develop the software system. Some common problems such as system architecture, interfaces, database design, integration with MS Office components, single document with multiviews and data structures are discussed during developing monitoring software system for hydraulic structures. And several data structure problems of monitoring data process, software robust, resetting initial values, cross queries on multiply tables and sensors replacement are discussed also.(8) Multidimensional data set are established by use of Xiao Wan monitoring data, and OLAP analyses as data mining are conducted on the data set.