An Artificial Neural Network-Based Algorithm for Evaluation of Fatigue Crack Propagation Considering Nonlinear Damage Accumulation

Abstract

In the aerospace and aviation sectors, the damage tolerance concept has been applied widely so that the modeling analysis of fatigue crack growth has become more and more significant. Since the process of crack propagation is highly nonlinear and determined by many factors, such as applied stress, plastic zone in the crack tip, length of the crack, etc., it is difficult to build up a general and flexible explicit function to accurately quantify this complicated relationship. Fortunately, the artificial neural network (ANN) is considered a powerful tool for establishing the nonlinear multivariate projection which shows potential in handling the fatigue crack problem. In this paper, a novel fatigue crack calculation algorithm based on a radial basis function (RBF)-ANN is proposed to study this relationship from the experimental data. In addition, a parameter called the equivalent stress intensity factor is also employed as training data to account for loading interaction effects. The testing data is then placed under constant amplitude loading with different stress ratios or overloads used for model validation. Moreover, the Forman and Wheeler equations are also adopted to compare with our proposed algorithm. The current investigation shows that the ANN-based approach can deliver a better agreement with the experimental data than the other two models, which supports that the RBF-ANN has nontrivial advantages in handling the fatigue crack growth problem. Furthermore, it implies that the proposed algorithm is possibly a sophisticated and promising method to compute fatigue crack growth in terms of loading interaction effects.

Forecasting Helianthus annuus Seed Quality Based on Soil Chemical Properties Using Radial Basis Function Neural Networks

“…Radial basis function neural network (RBFNN) is a type of ANN which uses radial basis function as the activation function, it is used in prediction estimation, system control, etc. [15]. The actual topological structure of RBFNN is displayed in Figure 2.…”

“…spread is the expansion constant which represents the smoothness of the fitted curve. As the training of ANNs is mostly empirical [12,15], repeated modification of spread should be conducted during the training process for better simulation; mn is the upper limit number of neurons, which is set to 100; df is the displaying frequency of training process, which represents the number of neurons added between the two iterations, and it is set to 1. There are 60 groups of matched tested soil and Helianthus annuus seed data, and 20 groups in each experimental field.…”

“…Root mean square error (RMSE) and mean relative error (MRE) were used to validate the accuracy of all forecasting RBFNNs [15,17,18] (Table 1). …”

“…It avoids the disadvantages of back propagation neural networks (BPNNs), which is of low convergence speed and liable to fall into the locally optimal solution [12,15]. The setting-up of neural networks is based on the non-linear relationship between the input vector and the output vector.…”

Investigation on the Sensitivity of Ultrasonic Test Applied to Reinforced Concrete Beams Using Neural Network

“…The higher the weights, the higher the impact of the input node. It is used for modeling on prediction or estimation of strength of capacity of structures [23][24][25][26][27][28][29][30][31][32][33][34][35][36].…”