Research On Gravity Quay Structure And Health Monitoring Forecasting Model Based On Support Vector Machine

Through health monitoring and model prediction for the gravity quay structure, its deformation law can be understood well and the early warming can be given timely, thus serious safety accidents in quay operation can be avoided. Considering one kind of gravity quay as the research background in this paper, the deformation law of quay structure is studied for further deformation prediction through field monitoring, data analysis and support vector machine. The main contents are as follows:1、Theoretical analysis for multiple linear regression, time series analysis and gray prediction model are carried out respectively; According to their application in practical engineering, their prediction and measured values are compared with each other, then the limitation of these three prediction methods are analyzed.2、The deformation mechanism of quay and its main influence factors are analyzed, parameters and kernel function of support vector machine for predicting quay structural deformation are defined, quay structural deformation prediction model based on SVM is built in MATLAB. Considering horizontal displacement of soil in quay’s typical position as the object of research, parameter values are obtained by training some collection data; The other observation data is used for prediction, then relative error of observation and forecast values can be obtained. The analysis shows that the prediction result is very close to actual monitoring data.3、The effects of kernel function σ and penalty coefficient C for quay deformation prediction is studied through numerical analysis method. The result shows that high precision prediction result can be obtained with the optimal combination of kernel function and the penalty coefficient, and the model can fully play the role of prediction.