Application Research On LSSVM Deformation Prediction Model Of Deep Foundation Pit Based On Phase Space Reconstruction Theory

With the accelerating pace of urbanization of our country's economy and society,high-rise,super high-rise building basement and urban rail transit project is also developing rapidly.In the process of construction,the project will face the problem of the deformation.If the deformation exceeds the set value,it will cause disaster.In order to ensure the normal construction engineering and the safety of surrounding buildings and people,analyzing and accurately forecasting the trend of dynamic deformation of deformable body has become an indispensable part of engineering construction.To address the problem of error of the data time series,the difficulty to reflect the Internal variation law and the randomness of time series reconstruction for single variable,wavelet thresholding is used to denoise the time series,then the theory of phase space reconstruction is adapted to reconstruct the denoising data,finally the grey least squares support vector machine model based on phase space reconstruction is established;The particle swarm optimization algorithm is also used to find the optimal parameters of the model.A comparative study of engineering example is made by using GLSSVM,SVM and LSSVM models,respectively.The results show that the LSSVM prediction model based on phase space reconstruction not only have high accuracy and fast convergence,but also provides a new idea for the error analysis of the data time series and the multi-dimensional reconstruction of the deformation analysis.