Study on deep learning of concrete failure criterion based on physics-information neural newwork
The concrete failure criterion is an important basis for the design and safety evaluation of engineering structures.It combines a new deep learning framework-a deep learning neural network based on physical information,using the concrete failure criterion function equation as a physical constraint to construct the corresponding representation term of the loss function,increasing the physical information drive between input and output,and more comprehensively reflecting the internal connections between various factors.Using a large amount of experimental data,train the deep learning model to establish a more accurate,applicable,and generalizable concrete failure criterion model.The results indicate that the physical information deep learning neural network model has good optimization recognition and generalization ability for the expression form and parameters of concrete failure criteria,providing guidance and reference for standard revision,engineering design,and finite element numerical simulation analysis and evaluation.
万方数据
维普
