A new neural network‐based model for hysteretic behavior of materials

Abstract

SUMMARYCyclic behavior of materials is complex and difficult to model. A combination of hardening rules in classical plasticity is one possibility for modeling this complex material behavior. Neural network (NN) constitutive models have been shown in the past to have the capability of modeling complex material behavior directly from the results of material tests. In this paper, we propose a novel approach for NN-based modeling of the cyclic behavior of materials. The proposed NN material model uses new internal variables that facilitate the learning of the hysteretic behavior of materials. The same approach can also be used in modeling of the hysteretic behavior of structural systems or structural components under cyclic loadings. The proposed model is shown to be superior to the earlier versions of NN material models. Although the earlier versions of the NN material models were effective in capturing the multi-axial material behavior, they were only tested under cyclic uni-axial state of stress. The proposed NN material model is capable of learning the hysteretic behavior of materials under even non-uniform stress state in multi-dimensional stress space. The performance of the proposed model is demonstrated through a series of examples using actual experimental data and simulated testing data.

Multiscale aggregating discontinuities: A method for circumventing loss of material stability

“…Another alternative approach is to pre-compute the response of the perforated unit cell or to train a neural network with the perforated unit cell. Recent work [30,31] has shown that neural networks can reproduce the constitutive response measured in experiments to a high degree of fidelity, particularly for monotonically increasing loads. In most applications of the proposed multiscale procedures, we envision that the loads are monotonic.…”

Review on the modelling of joint behaviour in steel frames

“…Dang and Tan [151] proposed an inner product-based hysteretic model for the application to piezoceramic actuators; Yun et al [152] as a model for hysteretic behaviour of materials;…”

A design‐variable‐based inelastic hysteretic model for beam–column connections

“…In this paper, a new NN-based cyclic material model developed by Yun et al [8,11] has been extended by adding the mechanical parameters from PPM for the application in structural design. The original NN-based cyclic material model can learn any complex hysteretic behavior [8,11].…”

“…where n indicates the nth load (or time) step, and M indicate the rotational displacement and moment, ,n = M n−1 n−1 and ,n = M n−1 n are the two internal variables for expediting learning capability of hysteretic behavior [8] and G i (DV 1 , . .…”

“…The original NN-based cyclic material model can learn any complex hysteretic behavior [8,11]. For training algorithm, a local adaptive learning scheme called RPROP (Resilient Back-propagation) algorithm is used [12].…”