Detecting Cutout Shape and Predicting Its Location in Sandwich Structures Using Free Vibration Analysis and Tuned Machine-Learning Algorithms

This paper deals with the detection of cutout shapes and the prediction of their location using machine learning algorithms. To this end, a series of simulation studies were performed for constructing a dataset containing different cutout shapes in different locations. Firstly, free vibration analysis of sandwich structures was performed using COMSOL. Natural frequencies of sandwich structures with different-shaped cutouts at different locations are obtained as a result of conducted simulation studies. Since the location and shape of cutouts significantly cause a change in the natural frequencies of sandwich structures, the first six natural frequencies of the sandwich structure chosen to be used as input for the machine learning algorithms. Afterward, different machine learning algorithms were employed to classify and predict the cutout shapes and locations using MATLAB. Furthermore, the Bayesian optimization algorithm has been adopted in order to tune the hyperparameters of the machine learning classification algorithms and improve their performance. The results of this work showed that the best cutout shape detection can be conducted by the tuned SVM machine learning algorithm with classification accuracy reaching 98.4%. Moreover, multiple machine learning algorithms have been adopted for predicting the location of the cutouts. The location prediction study gave an outperformed performance where the R-Squared metric was 1 for all tested machine learning algorithms. Also, the results of the location prediction study showed that the best RMSE value was 0.00049 that has been obtained using the Gaussian process. In summary, this study can be used as a reference by the designers to develop an opinion about the created SS before design and production, which can help in saving time and money.