Typically,uneven force application can easily lead to localized stress concentration and uneven de-formation in composite steel structures,even structural instability and damage.To ensure the safe use of composite steel structures,this article put forward a simulation for predicting elastic deformation under uneven force application in composite steel structures.Firstly,according to the entropy increase principle as well as the first and second laws of thermodynamics,we analyzed the internal energy distribution of composite steel structures under uneven force applica-tion,thereby clarifying the distribution of elastic deformation in composite steel structures.Based on the mechanical a-nalysis in the process of elastic deformation,we analyzed the phenomenon of rigid vibration during the elastic deforma-tion,and then used the generalized force to describe the vibration mode of the composite steel caused by elastic de-formation,thus obtaining its dynamic parameters of elastic deformation.After that,we input these parameters into a Back Propagation(BP)neural network and trained them to enable the BP neural network to predict the elastic de-formation of composite steel structures.Finally,through reverse weight correction,we optimized the prediction accuracy of the BP neural network,thus achieving an accurate prediction.Experiment results prove that the proposed method has accurate prediction results,and the predicted distribution of elastic deformation is consistent with the actu-al situation.This method provides an important guarantee for the widespread application of composite steel structures.
关键词
熵增原理/热力第一定律/热力第二定律/刚性振动/神经网络
Key words
Principle of entropy increase/The first law of thermodynamics/The second law of thermodynamics/Rigid vibration/BP neural network
维普
万方数据