Two-dimensional structural deformation monitoring method and device based on dynamic decision-making and neural network

Abstract

The invention discloses a two-dimensional structural deformation monitoring method and device based on dynamic decision-making and neural network, including obtaining strain data and structural deformation values ​​at different positions and different times of the two-dimensional plane structure; constructing a data set; constructing a deep neural network model and performing the model Training; conduct dynamic decision-making analysis on the strain data collected in real time to obtain dynamic decision-making results; input the strain data at different times into the deformation reconstruction model based on convolutional neural network and the deformation prediction model based on time convolutional network, and according to the The deformation reconstruction model of the convolutional neural network is used to reconstruct the deformation, and the deformation reconstruction amount at each monitoring position is obtained; the deformation prediction is performed according to the deformation prediction model based on the time convolutional network, and the deformation prediction results at each monitoring position are obtained; according to the deformation The reconstruction quantity and deformation prediction results are combined with the sensor coordinate position information, and the deformation field of the entire planar structure is fitted by the cubic spline interpolation method. Invention deformation reconstruction accuracy is high.

Description

technical field [0001] The invention relates to the field of structural deformation prediction, in particular to a two-dimensional structural deformation monitoring method and device based on dynamic decision-making and neural network. Background technique [0002] Under the disturbance of various complex and changeable field environments such as external wind load, vibration and shock, large structures will produce dynamic deformation, which will seriously affect the performance and life of the structure. There is an urgent need to reconstruct and predict structural deformation for health monitoring and deformation compensation. [0003] Accurate acquisition of structural deformation is the basis for structural damage assessment, residual life prediction and vibration control. The current commonly used structural deformation reconstruction methods mainly include: modal superposition method and Ko displacement theory. The modal superposition method needs to use the modal a...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve any of the deficiencies of the above-mentioned prior art, aiming at problems such as poor monitoring flexibility of the two-dimensional structure under external wind load, difficulty in system modeling, and low accuracy of deformation reconstruction, it provides a two-dimensional structure based on dynamic decision-making and neural network. A three-dimensional structural deformation monitoring method and device, the present invention firstly collects the discrete sensor measuring point information on the plane, builds a data set and a deep neural network model, and performs model training; through principal component analysis, K-means clustering and absolute average change Rate and fuzzy entropy determine whether to perform reconstruction or prediction tasks

Images