Research on anti-impact performance of steel fiber reinforced concrete based on finite element and machine learning

Abstract To predict the impact resistance of steel fiber reinforced concrete (SFRC), 50 specimens with different fiber lengths and different fiber contents were loaded using ABAQUS finite element software to obtain data in this paper. Two machine learning (ML) models, backward propagation-artificial neural network (BP-ANN) and support vector machine (SVM), were used to train the data. The results show that in the prediction of the impact resistance of steel fiber reinforced concrete by this model, the deviation of the predicted values from the real values is small, and the two models are well fitted. To further optimize the model, the parameters of the prediction model were adjusted using the whale optimization algorithm (WOA) in this paper, and the accuracy of the optimized model was significantly improved. After optimization, the WOA-BP-ANN and WOA-SVM models have better generalization ability and higher prediction accuracy than the WOA-SVM model.