Research On Grey Model Based On Cumulative Method In Bridge Construction Monitoring

With the construction of the country's top strategy "The Belt and Road",for the integration and efficient allocation of resources to achieve the depth of the market,to achieve economic cooperation in a more extensive and deeper level to each region,to build birdge infrastructure is essential.In the construction of modern bridge,the prestressed concrete continuous beam bridge is widely used because of its good overall performance,low cost and mature construction technology.The prestressed concrete continuous girder bridge is usually constructed by cantilever casting method.In order to meet the design requirements and ensure the safety of construction,the bridge construction monitoring is essential.Because the cantilever placement method is more complex and has many factors,the vertical displacement of the beam can not be accurately expressed by the structure calculation.Therefore,in the bridge construction monitoring,it is necessary to establish a mathematical model to analyze the error,find the law of error,so as to guide the construction more accurately.The grey theory model has many advantages,such as less sample data and higher operation accuracy,and has good compatibility.Therefore,it is widely applied in the elevation error analysis of bridge construction monitoring.But in the follow-up study of grey theory model,it is found that when using least square method to estimate parameters of the model,the prediction effect presents different degrees of morbidity.Therefore,how to improve the grey theory model and increase its stability and prediction accuracy,whether in the field of bridge construction monitoring or other engineering fields,has good engineering significance and practical value.This paper studies the improvement of line shape prediction method in bridge construction monitoring,which includes the following aspects:(1)Analyze the research results in the field of bridge construction monitoring and grey system theory modeling at home and abroad,and clarify the main problems and improvement space of the application of Grey Theory in construction monitoring.(2)The purpose and significance of construction monitoring are systematically summarized and analyzed.In addition,the monitoring content,structure calculation method and error control in construction monitoring of prestressed concrete continuous beam bridge are introduced in detail.(3)Aiming at the application of error analysis model in elevation prediction in bridge construction monitoring,we improved the parameter estimation method and prediction formula of GM(1,1)model.The concept of cumulative method is combined into the parameter estimation of GM(1,1)model,and the connotative prediction formula of the model is deduced from the defined type equation of the model,replacing the whitening response formula of the original GM(1,1)model,and the GM(1,1)model based on accumulative method is established.Through the study of matrix expression in parameter estimation,the related properties of the improved model are obtained.(4)Combined with the port bridge project,the elevation of the forecasting process,using GM(1,1)model,Kalman filtering method,based on the accumulative method of improved GM(1,1)model of three kinds of prediction methods for error analysis,and compare with the actual measurement of elevation,the prediction effect of three methods of analysis.In order to make the evaluation of the prediction more reliable,the author uses the relative error method and the posterior difference method in two ways to evaluate.The result of relative error detection shows that the average relative error of GM(1,1)model is 0.0313%,and the average relative error of Kalman filtering method is 0.03%,and the average relative error of GM(1,1)model based on cumulative method is 0.0247%.The test results show that the improved cumulative based GM(1,1)prediction model is more accurate,and the prediction accuracy of the traditional GM(1,1)model is increased by 21%,which is 17.6%higher than that of the Kalman filtering method.The results of the post test difference test also show that the GM(1,1)model based on the cumulative method has the best prediction effect.It shows that the GM(1,1)model based on the improved accumulation method has a better application effect in the process of linear error analysis in bridge construction monitoring.