This study proposes a helical gear defect detection system based on Symmetric Dot Pattern (SDP) and Convolutional Neural Network (CNN). By converting the collected time-domain signals into SDP images and processing them with different delay coefficients and weighting coefficients, followed by deep learning training using CNN, the method's accuracy was analyzed to achieve monitoring and detection of helical gears. Experimental results show that this method has high accuracy in detecting helical gear defects. When the training parameters are at their optimal values, the detection accuracy can reach 91.25%. Different epochs demonstrate that setting the epochs to 200 achieves the highest accuracy, while too many or too few epochs lead to a decrease in model accuracy. The delay coefficient significantly impacts the model’s accuracy, with the highest accuracy of 90.6% achieved at a delay coefficient of 3. The weighting coefficient also shows considerable importance, with the highest accuracy of 90.3% at a weighting coefficient of 80. This study indicates that the helical gear defect detection method based on SDP and CNN can achieve high-accuracy detection results with appropriate parameter settings, demonstrating significant application potential.
27 December 2024 in The International Journal of Advanced Manufacturing Technology
Tool breakage in CNC machining often leads to reduced productivity and increased maintenance costs. This study proposes a non-contact tool breakage detection method using spindle current signals captured by an SCT013 current sensor. The sensor easily attaches to the motor line without any hardware modification and provides real-time current signals for frequency domain analysis. Fast Fourier Transform (FFT) is employed to extract spectral features, particularly focusing on high-frequency energy spikes at the moment of breakage. A total of 20 experiments were conducted, and consistent spectral anomalies were observed. Additionally, deep learning models including ANN, DNN, and CNN were compared for automated detection performance. The results indicate that the proposed system can reliably detect tool breakage by identifying frequency domain anomalies that emerge within 1–3 s after the actual event, based on processed current signals. While the inference time of deep learning models ranges from 15 to 58 s, the detection mechanism captures the breakage characteristics early in the signal, enabling timely tool condition evaluation.
Gears play a pivotal role in power transmission systems. Over prolonged periods of high-load transmission, the issue of gear lifespan has become a challenging one to address. Monitoring systems are vital in today's smart manufacturing landscape, contributing to safety and sustainable development. Presently, there is a dearth of research focused on the vibration of gears with surface modification. Consequently, this study has devised a novel defect detection model for surface modification gears. It employs time-domain, frequency-domain, and discrete wavelet transform (DWT) analysis in combination with deep neural network (DNN) deep learning to validate the effectiveness of this approach. The study deals with gears in two states: undamaged and wear. Vibration data was sampled for 800 seconds at a rate of 1024 Hz. Leveraging Artificial Intelligence, this research opens the door to innovative applications and services in the field of gear manufacturing. Notably, DWT exhibits a significant reduction in the amplitude of vibrations between surface modification helical gears and surface modification spur gears. Furthermore, from the DNN training and accuracy, it is evident that there is no substantial difference in accuracy between epochs 30 and epochs 100.