Search for: All records

Additive Manufacturing (AM) is a crucial component of the smart manufacturing industry. In this paper, we propose an automated quality grading system for the fused deposition modeling (FDM) process as one of the major AM processes using a developed real-time deep convolutional neural network (CNN) model. The CNN model is trained offline using the images of the internal and surface defects in the layer-by-layer deposition of materials and tested online by studying the performance of detecting and grading the failure in AM process at different extruder speeds and temperatures. The model demonstrates an accuracy of 94% and specificity of 96%, as well as above 75% in measures of the F-score, the sensitivity, and the precision for classifying the quality of the AM process in five grades in real-time. The high-performance of the model could not be achieved with the values usually used for printing temperature and printing speed, only in addition with much higher values. The proposed online model adds an automated, consistent, and non-contact quality control signal to the AM process. The quality monitoring signal can also be used by the AM machine to stop the AM process and eliminate the sophisticated inspection of the printed parts for internal defects. The proposed quality control model ensures reliable parts with fewer quality hiccups while improving performance in time and material consumption. 

Metrology experiments can be limited by the noise produced by the laser involved via small fluctuations in the laser’s power or frequency. Typically, active power stabilization schemes consisting of an in-loop sensor and a feedback control loop are employed. Those schemes are fundamentally limited by shot noise coupling at the in-loop sensor. In this Letter, we propose to use the optical spring effect to passively stabilize the classical power fluctuations of a laser beam. In a proof of principle experiment, we show that the relative power noise of the laser is stabilized from approximately 2 × 10⁻⁵Hz^−1/2to a minimum value of 1.6 × 10⁻⁷Hz^−1/2, corresponding to the power noise reduction by a factor of 125. The bandwidth at which stabilization occurs ranges from 400 Hz to 100 kHz. The work reported in this Letter further paves the way for high power laser stability techniques which could be implemented in optomechanical experiments and in gravitational wave detectors.

 

Network intrusion detection systems (NIDS) for Internet-of-Things (IoT) infrastructure are among the most critical tools to ensure the protection and security of networks against malicious cyberattacks. This paper employs four machine learning algorithms and evaluates their performance in NIDS considering the accuracy, precision, recall, and F-score. The comparative analysis conducted using the CICIDS2017 dataset reveals that the Boosted machine learning techniques perform better than the other algorithms reaching the predicted accuracy of above 99% in detecting cyberattacks. Such ML-based attack detectors also have the largest weighted metrics of F1-score, precision, and recall. The results assist the network engineers in choosing the most effective machine learning-based NIDS to ensure network security for today’s growing IoT network traffic.