Using Machine Learning to Predict the Stress-strain State of a Ring Plate

This article is devoted to the problem of prediction of the stress-strain state of a ring plate using neural networks. The article shows how to train the neural network to estimate the von Mises stress of a ring plate under external uniform pressure. Machine learning, one of the six disciplines of Artificial Intelligence (AI) without which problems of having machines acting humanly could not be accomplished. Applications of ANNs to engineering structures arise in a variety of industries such as engineering, automotive, space structures, etc. ANNs allow to develop models e.g. for the stress-strain state estimation of some type of solids. Thus, the development of machine learning methods for predicting the behavior of engineering structures is urgent. The paper describes the scheme for using machine learning in the stress-strain state analysis of a ring plate. Additional input parameters of the data set are the following: outer radius, inner radius, thickness of the plate, Young’s modulus, Poison’s coefficient, and pressure load. The training set is generated by the finite element method. Initial parameters of the training set have been randomly generated. The artificial neural network merges numerical and one-hot input layers. One hot The developed regression model allows to predict von Mises stresses for a ring plate. The developed model allows to predict the von Mises stress with 10% of the mean absolute percentage error. The key advantage of an artificial neural network is the speed of prediction. The ANN predicts the von Mises stress almost instantaneously (milliseconds) comparing the finite element method