Use of Machine Learning Methods in Syntactic Foam Design

Due to the viscoelastic nature of polymers, material characterization is a major challenge in designing syntactic foam microstructure. As the properties of syntactic foam shows strong nonlinearity, it is overwhelming to test each composite material for every application case under the combined effects of temperature and strain rate. Machine learning methods can help by using the existing datasets to predict properties over a different combination of parameters. This article focuses on building an artificial neural network (ANN) based architecture to help in predicting properties and compositions of viscoelastic materials. The high density polyethylene (HDPE) syntactic foam is used as a case study material. Four types of HDPE syntactic foams were tested using dynamic mechanical analysis (DMA). Then, ANN was used to build the master relation of viscoelastic properties with respect to frequency, temperature, particle volume percentage and strain. The master relation for storage modulus was transformed to time domain relaxation function and used to predict the stress-strain relations to calculate modulus. The predicted and measured modulus values show good agreements for both tested and extrapolated compositions. These results show that machine learning methods can help in designing composite materials and reduce the requirement for generating experimental data over a large number of loading conditions.