Feb 6 2019
It can be costly and time-consuming to optimize advanced composites for specific end applications, where manufacturers have to test several samples to achieve the best formulation.
Research by Nikhil Gupta, associate professor of mechanical and aerospace engineering promises to reduce the cost and boost the efficiency of materials testing by combining traditional dynamic mechanical analysis (DMA) with artificial neural networks. (Image credit: New York University)
Researchers at the NYU Tandon School of Engineering have developed a machine learning system using artificial neural networks (ANN) that can extrapolate from data obtained from a single sample, thus rapidly formulating and delivering analytics on theoretical graphene-enhanced advanced composites.
The research, headed by Nikhil Gupta, associate professor of mechanical and aerospace engineering at NYU Tandon, with PhD student Xianbo Xu and collaborators at 2D graphene materials manufacturer GrapheneCa, is described in “Artificial Neural Network Approach to Predict the Elastic Modulus from Dynamic Mechanical Analysis Results,” which will feature on the inner cover of the journal Advanced Theory and Simulations.
Dynamic mechanical analysis (DMA) and tensile tests are extensively used to characterize the viscoelastic properties of materials at various loading rates and temperatures. However, this needs a detailed experimental campaign including a large number of samples.
The Tandon team found an approach to bypass this process by developing an ANN-based technique that constructs a model and subsequently offers data from DMA—a test of the response of a material to a given temperature and loading frequency (a measure of load applied in cycles)—to estimate its response to any other pressure and temperature combination. Gupta described that ANN judged from measures of the ability of the samples to store and dissipate energy under various conditions.
Testing materials under different conditions during the product development cycle is a major cost for manufacturers who are trying to create composites for numerous applications. This system allows us to conduct one test and then predict the properties under other conditions. It therefore considerably reduces the amount of experimentation needed. Applying an artificial neural network approach to predict the properties of nanocomposites can help in developing an approach where modeling can guide the material and application development and reduce the cost over time.
Nikhil Gupta, Associate Professor, Mechanical and Aerospace Engineering, NYU.
Working with the researchers at NYU Tandon’s Department of Mechanical and Aerospace Engineering, we have developed a new method for predicting the behavior of thermosetting nanocomposites over a wide range of temperature and loading rates. Furthermore, the same approach can potentially be applied to predict a behavior of thermoplastic materials. This is a critical step towards advanced composite production.
Dr Sergey Voskresensky, Head of Research and Development, New York Production Facility, GrapheneCa.
The U.S. Office of Naval Research provided grants to support this work.