An application of Artificial Neural Network (ANN) to predict the friction coefficient of nuclear grade graphite

In scientific research, computer modeling techniques are widely used. Artificial neural networks are now well established and prominent in the literature when computationally based methodologies are used. New advancements in these domains have benefited and continue to benefit the materials science and engineering research community, with new applications and levels of sophistication appearing regularly. However, with this greater utilization comes a growing tendency for neural network approaches to be misapplied, reducing their potential effectiveness. Due to its good prediction quality, an artificial Neural Network (ANN) is a valuable mathematical tool for solving complex scientific and technical problems. The study presents an approach to predicting the tribological properties of nuclear grade graphite using ANN. The data required to predict the frictional behavior of nuclear grade graphite was developed experimentally and compared with ANN software (Alyuda Neuro intelligence) for analysis and verification, taking input variables such as temperature, time, and sliding distance, and friction force. The study proceeds experimentally to train the neural network, test, and validate. The study concluded that the developed ANN model and backpropagation Alyuda Neuro intelligence could predict the frictional characteristics of nuclear grade graphite with a correlation coefficient of 0.9995 and a mean absolute error of 0.0030. © 2022