Node Rigid Domain Recognition Method Based On Monitoring Data

In the process of using the structure,the nodes are in the key parts where the stress is complicated and concentrated.Therefore,the node is extremely important to the overall structural security.The rigid domain of the node is uncertain.Obtaining the node rigid domain through monitoring means is helpful to establish an accurate structural finite element model and provide basic information for structural design verification and structural health monitoring.There is a certain correlation between the node rigid domain and the node response.As the load conditions and the length of the rigid domain change,the association relationship also changes.Therefore,a node rigid domain length estimation method based on multi-case node response is proposed.By determining the correlation and mapping relationship between the measured points and the corresponding rigid domain length,the length of the node rigid domain under multiple operating conditions can be estimated.The main research work of this article starts from the following three aspects:Research on node response and node rigid domain association rules based on Apriori algorithm.Fixed load conditions,change the length of the rigid domain,obtain the stress response of nodes at different locations and group them to form a database of rigid domain changes,and perform correlation analysis on them to obtain stress-sensitive positions to be measured that are sensitive to changes in the rigid domain;The stress to be measured position selected by the change database,the length of its rigid domain is fixed,the load conditions are changed,the node stress response is obtained and grouped to form a load change database.The Apriori algorithm is used to analyze the strong association rules of the database respectively,given the minimum support and minimum confidence,the strong association rule analysis of the rigid domain length and stress response change rate is performed to determine the sensitive position of the stress response change caused by the rigid domain length change.The sensitive position is used as the actual sensor placement position.Research on the identification method of node rigid domain based on random forest algorithm.Select the stress response change rate of multiple load conditions and multiple rigid domain lengths as training data,and use the random forest algorithm to determine the mapping relationship model between the stress response change rate and the K-node rigid domain length.Comparing the sensor positions obtained by using the random forest algorithm alone,the Apriori algorithm alone,and considering the random forest algorithm and the Apriori algorithm at the same time,considering the number and location of sensor placement points,carrying outnode rigid domain length estimation under multiple load conditions and clear training The influence of the amount of data and the number of measured points on the estimation accuracy of the node rigid domain length verifies the effectiveness of the proposed node rigid domain length estimation method based on the mapping relationship.Analysis and optimization of the parameter impact of the node rigid domain identification method.Taking the determined stress response change rate of the sensor to be arranged as the training set,the random forest algorithm is used to establish the node response and node rigid domain mapping relationship model,and the parameters of the random forest algorithm are considered,such as the number of models built,the maximum tree depth,The maximum number of nodes,the minimum child node size and the number of bins,and the impact of the amount of training data on the mapping model,so as to further optimize the mapping relationship model.Determine the influence of the number of construction models,maximum tree depth,maximum number of nodes,minimum sub-node size and number of bins,and the amount of training data on the estimation result of the mapping relationship model,and perform error analysis on the mapping relationship model to obtain the optimal mapping model.