Material constitutive model prediction method based on PSO-LSSVM

Abstract

The invention discloses a material constitutive model prediction method based on PSO-LSSVM. The method can be used for establishing a material constitutive model of various alloy materials in a mechanical property experiment, and comprises the following steps: firstly, obtaining material mechanical property indexes through the mechanical property experiment, recording experiment parameters, and dividing obtained data into a test sample and a training sample; optimizing key model parameters in an LSSVM (Least Square Support Vector Machine) by utilizing a PSO (Particle Swarm Optimization) algorithm and using the training sample; substituting the model parameters to predict the material constitutive model by using an LSSVM algorithm, and analyzing a prediction result by using a test sample. According to the method, the constitutive relation prediction model is established for the alloy material based on the PSO-LSSVM algorithm, the prediction precision is good, error evaluation indexes are maintained within a small range, and a new method is provided for establishing the material constitutive model.

Description

technical field [0001] The invention relates to the field of establishment of metal material constitutive models, in particular to a method for predicting material constitutive models based on PSO-LSSVM. Background technique [0002] In the forming process of various alloy materials, the constitutive model is an important basis for mastering the rheological behavior of the material and designing the hot forming process. An accurate material constitutive model can intuitively characterize the dynamic response of material flow stress to material mechanical performance indicators and experimental parameters such as deformation temperature, strain rate, and strain. However, due to the characteristics of multi-element alloys and the multi-field coupling in the actual processing process, the deformation mechanism is complicated and the process control is difficult. In the constitutive model, it is multi-factor and highly nonlinear. [0003] At present, the main methods used to es...

AI Technical Summary

Problems solved by technology

However, in the process of optimizing the parameters of the fruit fly algorithm, it depends on the initial position and direction selection, and needs to combine experience to achieve better prediction accuracy, and the overall stability of the prediction is insufficient.

Images