Study On The Optimize Dynamic Model Of Grey And In The Application Of Deformation Monitoring

Dynamic grey forecasting model is the development and perfection for grey forecasting model. It follows the principle of new information priority, and produces the fixed dimensions (equal dimension) and replacement dynamic grey forecasting model. For new information occupies absolutely proportion, this model has solved the effect of origin error effectively, and make the result of forecast more practical, thus improve the prediction accuracy of grey model greatly. But for keep the dimension invariance, the traditional model has to abandon the old information compulsively so that it leads to the following three issues:1. The different dimensions lead to different accuracy of predicted, and the influence is bigger.2. The information of out dimensions is give up indiscriminately so that the useful information drains abundantly.3. The limit of dimensions lead to less amount of data to set up model, as a result forecast distortion.Based on the above three problems, this paper has started in-depth and detailed research. Firstly, aiming at the problem of dynamic model dimensions, this paper has used the optimizing test for different dimensions of prediction model, and obtained the range of optimal dimension for the dynamic model of grey. Secondly, aiming at the problem of useful information drain abundantly, this paper has come up with a new dynamic model of grey, which is'optimize dynamic prediction model of grey'. This model follows the principle of that new information data is priority and the effective utilization maximizes, sifts the old information through specific mathematical method reasonably, keep the effective data for modeling enough, and then resolve the problem of that lack of data lead bad prediction, in addition, expand the dynamic prediction from original equal dimension new information to unequal and uncertain dimensions, and the range of application is enlarged. Finally, verified the correctness of the model by the subsidence monitoring of a certain section, in the second line of WuHan subway.Paper work and the main research results are as follows:1. Optimal dimension experiment of the dynamic forecasting model of grey To reduce human factors and make test more objective, the data of test are all randomly generated by the Matlab software. Then paper do the traditional dynamic grey prediction by selecting 5-13 dimension data, and compare the prediction results in different dimensions, finally to jump to the following conclusions:(1) Too high or too low dimension will lead to predict undistorted, the different dimension has great Influence on forecasting results.(2) Suitable scope are 7-11 dimension.(3) By selecting experiment of the stage dimensions, finding that suitable dimension have phase change.Therefore when using dynamic grey model to predict, should do the periodic dimension experiment.2. Research on modeling method of optimize dynamic prediction model of greyThe main research is establishing the smooth functions, selecting the associated threshold, and the setting method of screening function. Its main modeling process is that: first, establishes a second or a polynomial fitting function through the fixed dimensions of new discrete sequence; second, determines the associated threshold by the squared residuals minimum standards; the third, Screens function by the residual value calculation. For the dynamic grey prediction model can selects and eliminates the old information reasonably, this model can take maximum advantage of the useful information, and offset the weakness of Insufficient data for the dimension modeling, so it can improve the accuracy of deformation monitoring and prediction correspondingly. Finally, this paper verifies the validity of this method by the simulation experiment.3. Research on application of the optimize dynamic model of greyPaper has done a text of forecast for subsidence monitoring, which bases on the 24 section of second line for WuHan subway, from HuQuan station to MingDu Garden station. The data for text has been picked three part of the excavated interval, which is the front part, middle part and back part, and each part has three different periods, which is earlier period, middle period and later period. The result for text is that(1) Depend on the different period of excavated interval, the experiment of the stage dimensions can improve the accuracy of prediction.(2) The optimize dynamic prediction model of grey has the advantage of high precision, thus it can apply in some high accuracy projects, such as subway, tunnels under the river and crossing sea bridge etc.4. Base on the optimize dynamic model of grey, develop the deformation Monitoring Data Management and Period SystemFor manage the deformation monitoring data better conveniently, base on the study of the optimize dynamic model of grey, develop the deformation monitoring data management and period system。In allusion to present deformation monitoring technology, as a design idea to cyclical pattern repeated measurement and dynamic data processing method and as a form to organization that relations to database tables and data structure, the article selection of Microsoft.NET as a development platform, the new language C# as the develop language. The system selects SQL 2005 database management system for data storage and uses ADO.NET database access interface technology to achieve the establishment of the database, delete, update and query, expatiated the system module structure and the specific functions and system Design and Implementation. Now it has been applied to the part of the monitoring of WuHan subway successfully.Finally, the paper expounded the shortcomings in optimize dynamic prediction model of grey, and further improvement and model perfect is prospected.