Application Research On The Deformation Monitoring Data Analysis Based On The Improving Dynamic Gray Model

Deformation monitoring makes on a comprehensive analysis of the deformation of the deformable body to accurately determine the future tendency of deformable bodies by using observation data obtained, thus providing a scientific basis and security for the building project design, construction and operation. The grey system theory mainly concentrates on the system of the small sample, poor information and uncertainty, which is known by generating information on the part of the known to excavate the important information in the systematic observation data, and ultimately the whole system to get the exact understanding and description. Deformation monitoring is used as a method to monitor and forecast the construction project and ensure the safety in the construction and operation, while as a discrete sequence of deformable body, the cumulative observed data of deformation monitoring, itself has a gray character, so the grey theory can be effectively applied to engineering analysis and forecasting of deformation monitoring.Based on the application requirements, the article focused on deficiencies of gray model in the analysis and prediction of deformation monitoring importantly to find out the corresponding improvement measures. The improved model is applied to the analysis and forecasting of deformation monitoring in tunnel, finally verified the effect of the improved prediction model. The main contents of this article are as follows:(1) The article briefly introduced the purpose and significance of deformation monitoring, the content and classification of deformation monitoring, the technological development status of deformation monitoring and the common modeling analysis, and emphatically focused on the application status of the gray system theory and optimization method, the basic concepts of the theory of gray and the modeling mechanism of the gray model.(2)Dimension selection problem has been the research hotspot of dynamic gray forecast model. Carrying out the preferred test of the dimension before using the dynamic grey modeling to draw the appropriate dimension dynamic range of gray modeling.(3)Before using dynamic gray prediction model, the author used Grubbs guidelines to inspect the gross error of observation data and used the least square polynomial fit method to generate the gap data. Eventually, the author eliminated the gross error interference of model prediction accuracy effectively.(4)Observation data is interferenced by external factors or factor of system in the deformation monitoring easily, thus causing data fluctuation, in turn, leading to poor smoothness data modeling. Using three detesting smoothing to process the discrete data to solve this problem, Finally this will not only effectively weaken the interference of human subjectivity or chance to discrete data, but also enhances the regularity of the discrete data.(5) In order to keep the data dimension invariance, the traditional dynamic gray model has to abandon the old information, which will inevitably lead to limited amount of data modeling, forecasting results easily distorted, and information on the outside dimensions without screening discarded, resulting in losing a lot of useful information, the article makes improvements to the dynamic grey model by the specific mathematical methods. The improved model can screen the old information initiatively, which not only can solve the problem of distortion of model prediction results, but also chang the gray model from the dimension dynamic gray forecast to the dimension uncertain, and effectively solve the problem of the applicability of the model. Finally the improved model is applied to deformation monitoring analysis and forecasting. The comparison was made between the prediction accuracy of the traditional GM 1,1 model and the dynamic grey model, and then verify the validity of the improved model.