Development Of Remote And Automatic Monitoring System For Safety Of Urban Underground Engineering

Control cycle of safety of underground engineering is generally long and theamount of monitoring data is very huge. Monitoring control can’t be timely andefficiently carried out because of the wicked field environment and heavy workload ofdata diagnosis process. Remote transmission technology and computer technologyhave great advantages in solving these problems.Based on comprehensive and systematic summary about the research anddevelopment status, remote and automatic monitoring system and its diagnosismodule for safety of urban underground engineering is developed and applied topractical engineering. The main work and conclusions are as follows.1. Based on the Java language, monitoring system is developed. Some functionsincluding data acquisition, remote real-time display and query and alarm informationautomatically publishing are achieved by a unique development mean.2. Based on the Matlab, diagnosis module is developed. Some functions such asmeasuring points arrangement plan view, warning value setting and view, landsubsidence simulation and prediction, groundwater level change curve drawing andhorizontal displacement simulation prediction are achieved.3. Monitoring system and diagnosis module are applied to expansion supportingairport transportation center project of Tianjin Binhai International Airport and westdistrict foundation deformation monitoring project of Tianjin Hongze City, and theirreliability and practicality are proved.4. GM(1,1) grey model is combined with adaptive filtering model and rollingmodel respectively by variance combination theory to obtain two combined models:grey adaptive filtering combined model and grey rolling combined model. The twocombined models are applied to simulation and prediction time series of subsidenceand horizontal displacement.5. Four basic models and two combined models are used to simulate and predictmonitoring data. Basic models include GM(1,1) grey model, adaptive filtering model,BP neural network model and rolling model. Combined models include grey adaptivefiltering combined model and grey rolling combined model. Accuracy of simulationand prediction data is analyzed. The result shows that the four basic models cansimulate and predict the time series of subsidence and horizontal displacement well. The hypothesis that variance combination theory can improve the stability of themodel effectively is proved by contrast of the error fluctuation range of the combinedmodels and the basic models.