Study On Application Of The Grey Model-Time Series Analysis (GM-TSA) Combined Model In Deformation Prediction Of Buildings

With the development of technology and the enhancement of Chinese economic strength, there are more and more over-ten-floor, tens of floors or even over hundred floors buildings in the cities. To ensure these buildings'security and determine the deformation state, we need to carry out long term precise deformation monitor. For all kinds of reason, due to the deformation being not predicted or effectively controlled it tend to causes accidents. Being able to avoid economic losses and casualties, building deformation monitor and prediction has been an important aspect of construction engineering disaster prevention and reduction.Combining the engineering practice, we research and analyse methods of building deformation analysis and prediction in this dissertation. Then some suggestion and methods are presented. The main content are as follows:1. According to characteristics that the aging building deformation is monotonous and random, we introduce grey model-time series analysis dynamic combined model to prediction of building deformation. Grey model-time series analysis combined model embodies the capability of gray model on processing small-stylebook data as well as the capability of time series model on processing random series. The introduction of equal dimension new information model and equal dimension filling model to combined model furtherly improves the capability of dynamicly processing data. The introduction of AICC determination order criteria improves accuracy of small-stylebook series order determination.2. We discuss on modeling non-equal interval series, and analyse relation between period and prediction step number, determine optimal interpolation period in the general case. According to the case that we predict building deformation after struction completion basing on the data of loading period, a method that interpolated series data is multiplied by correction coefficients is presented to improve prediction precision.3. Combining the specific engineering example, based on building deformation of earlier construction period, we establish grey model-time series analysis (GM-TSA) dynamic combined model to predict the later deformation of building, concurrently we also develope MATLAB program to actualize the model calculation. At the same modification method. It can be seen from the calculation result that grey model-time series analysis dynamic combined model is able to predict building deformation with good precision, and the improvement presented in this dissertation availably improve precision.4. Finally, to explore route for the study later in this field, we also present suggestion and prospect on the method and research approach on analysis and prediction of building deformation.