Intelligent Material Extrusion Defect Prediction: CNN-SVM Integration

In this study, they propose a sophisticated model that uses a hybrid technique that combines a Convolutional Neural Network (CNN) or Support Vector Machine (SVM) to anticipate material extrusion faults in additive manufacturing processes. Excellent recall, precision, and F1-score values are displayed by the model for a range of material extrusion defect classes. The precision of the model is 98.47% to 98.69%, indicating its accuracy in detecting base bulging, weak fiber bonding, deformed edges, and under-extrusion. In line with this, recall values ranging from 98.53% to 98.70
^%
demonstrate how well the model captures the bulk of pertinent cases in each fault class. Recall and precision are nicely balanced, as seen by the model's consistently high F1-scores of 98.60%, which validates the model's efficacy in defect identification. The material extrusion defect prediction achieved an impressive overall accuracy of 98.60% using a dataset of 3864 pictures. The support values, which also maintain a balanced distribution among defect classes, demonstrate the model's exposure to a range of datasets. What's interesting is that the model achieves a support proportion that shows how well it can generalize to different fault situations. The model's 99% overall accuracy demonstrates both its potential for control of quality in additive manufacturing processes and its dependability in real-world applications. The Macro Average, Weighted Average, or Micro Average metrics-all of which are 98.60%-all attest to the model's continuous excellence in defect prediction. This all-inclusive material extrusion quality assurance automation system is provided. The very precise and adaptable defect prediction model presented in this work improves additive manufacturing quality control. The model is presented by the hybrid CNN-SVM technique as a helpful tool for ensuring structural integrity & quality in extruded products, enhancing the general dependability and efficacy of additive manufacturing processes with its excellent performance measures.