Artificial Neural Network‐Based Model to Predict Burning Rate of Catalysed Composite Solid Propellant

ABSTRACT A prediction tool for the burning rate of composite solid propellants using an artificial neural network (ANN)‐based model is proposed. The methodology adopted can be divided into two parts (a) estimation of interaction between process variables using the Spearman rank‐order correlation method and (b) building an ANN‐based model to predict the burning rate from a trimmed dataset consisting of significant variables. A multilayer perceptron (MLP) neural network was fed with trimmed variables as input, and a backpropagation algorithm was used to solve the mathematical model in Python. ANN hyperparameters tuning was carried out using the Grid Search algorithm in Python. It was found that the ANN model can predict the average burning rate of a solid motor with high accuracy when compared with the average burning rate obtained from ballistic evaluation test motors. This methodology helps predict the burning rate from a propellant composition and mechanical and physical properties without firing ballistic evaluation motors (BEM).