The Application Of Optimization Of Grey Markov Model In The Building Settleme Prediction

With the urbanization aceelerated,various types of high-rise buildings have mushroomed from where they stand. Existing building security hidden danger, avoid to unnecessary economic losses and casualties, buildings should be conducted on a regular basis for settlement observation, sink down deformation data obtained, and the settlement deformation data processing analysis. With the attention and research on monitoring and prediction analysis of deformation, a variety of prediction methods have appeared at home and abroad, including linear regression analysis, time series analysis, kalman filtering analysis, the theory of wavelet analysis, artificial neural network model, support vector machine(SVM) method, the grey system model, markov chain, etc.A scientific and reasonable prediction model is beneficial to the safety of the buildings, and also for the appropriate decisions has great significance. In this paper, the monitoring content and significance of the building, the building subsidence monitoring technology and to meet the requirements of building subsidence monitoring and the research development of data analysis and forecasting model are discussed; Based on the basic theory of markov prediction to the discussion of markov process, the grey forecasting theory, discusses the precision of grey GM(1, 1) model and residual error GM(1, 1) model. Grey markov prediction model is established and its application in building subsidence prediction, the model to verify the gray markov prediction model in building subsidence prediction has certain feasibility. Focuses on the gray markov prediction model of two kinds of optimization method, optimization method of metabolism and residual error correction combined with the optimization method of metabolism, and the two optimized model to predict the settlement of the building, and compares trial analysis with other methods, the results show that the optimized model not only can accurately predict the range of data, and the prediction precision and accuracy are improved.