Numerically Trained Ultrasound AI for Monitoring Tool Degradation

Abstract

Monitoring tool degradation during manufacturing can ensure product accuracy and reliability. However, due to variations in degradation conditions and complexity in signal analysis, effective and broadly applicable monitoring is still challenging to achieve. Herein, a novel monitoring method using ultrasound signals augmented with a numerically trained machine learning technique is reported to monitor the wear condition of friction stir welding and processing tools. Ultrasonic signals travel axially inside the tools, and even minor tool wear will change the time and amplitude of the reflected signal. An artificial intelligence (AI) algorithm is selected as a suitable referee to identify the small variations in the tool conditions based on the reflected ultrasound signals. To properly train the AI referee, a human‐error‐free data bank using numerical simulation is generated. The simulation models the experimental conditions with high fidelity and can provide comparable ultrasound signals. As a result, the trained AI model can recognize the tool wear from real experiments with subwavelength accuracy prediction of the worn amount on the tool pins.

Defect‐Free Sound Insulator Using Single Metal‐Based Friction Stir Process Array

“…A sound wave around 0.9 MHz was selected as the operating wave for the performance demonstration as it's considered a long‐enough wavelength to eliminate the microstructure‐induced dispersion effect. [ 29 ] It also fits the fundamental frequency of the commercially available air‐coupled acoustic piezoelectrical transducers. A total of 20 FSP parallel paths were introduced on the long aluminum plate along the width direction.…”

“…The collected temporal data and the time trigger were sent to Tektronix MDO 34 oscilloscope with a consistent time range window at a 1 GHz sampling rate. In the collected signals, the reflected pulses from the upper and lower surfaces of the scanned sample were located for a further calculation to obtain the dynamic (effective) bulk modulus following the equation: [29]…”

“…The collected temporal data and the time trigger were sent to Tektronix MDO 34 oscilloscope with a consistent time range window at a 1 GHz sampling rate. In the collected signals, the reflected pulses from the upper and lower surfaces of the scanned sample were located for a further calculation to obtain the dynamic (effective) bulk modulus following the equation: [ 29 ] , where , is the magnitude of the emitted pulse from the transducer, and are magnitude of reflection from the upper and lower surfaces of the workpiece, c is the speed of sound in the sample, Z is the acoustic impedance of the scanned sample, and is the reference acoustic impedance of the ambient material (Nu‐clear cutting oil in this study). The experimental profile from the FSP sample was repetitively assigned on the plate in the numerical simulation which contained the base material regions, SZ, and HAZs.…”

Prediction of Tool Remaining Useful Life Based on NHPP-WPHM

“…At present, the tool remaining useful life prediction is mainly based on failure information, performance degradation information, or wear amount. Wiener process, Gamma process, or competitive failure model are applied and combined with the failure threshold [2][3][4][5][6][7][8][9][10]. However, collecting a large amount of life data for mechanical products, such as machine tools, in a short period of time is difficult compared with electronic products.…”