Welding is a key manufacturing process for many industries and may introduce defects into the welded parts causing significant negative impacts, potentially ruining high-cost pieces. Therefore, a real-time process monitoring method is important to implement for avoiding producing a low-quality weld. Due to high surface temperature and possible contamination of surface by contact transducers, the welding process should be monitored via non-contact transducers. In this paper, airborne acoustic emission (AE) transducers tuned at 60 kHz and non-contact ultrasonic testing (UT) transducers tuned at 500 kHz are implemented for real time weld monitoring. AE is a passive nondestructive evaluation method that listens for the process noise, and provides information about the uniformity of manufacturing process. UT provides more quantitative information about weld defects. One of the most common weld defects as burn-through is investigated. The influences of weld defects on AE signatures (time-driven data) and UT signals (received signal energy, change in peak frequency) are presented. The level of burn-through damage is defined by using single method or combine AE/UT methods.
The spline component of gearbox structure is a non-redundant element that requires early detection of flaws for preventing catastrophic failures. The acoustic emission (AE) method is a direct way of detecting active flaws; however, the method suffers from the influence of background noise and location/sensor based pattern recognition method. It is important to identify the source mechanism and adapt it to different test conditions and sensors. In this paper, the fatigue crack growth of a notched and flattened gearbox spline component is monitored using the AE method in a laboratory environment. The test sample has the major details of the spline component on a flattened geometry. The AE data is continuously collected together with strain gauges strategically positions on the structure. The fatigue test characteristics are 4 Hz frequency and 0.1 as the ratio of minimum to maximum loading in tensile regime. It is observed that there are significant amount of continuous emissions released from the notch tip due to the formation of plastic deformation and slow crack growth. The frequency spectra of continuous emissions and burst emissions are compared to understand the difference of sudden crack growth and gradual crack growth. The predicted crack growth rate is compared with the AE data using the cumulative AE events at the notch tip. The source mechanism of sudden crack growth is obtained solving the inverse mathematical problem from output signal to input signal. The spline component of gearbox structure is a non-redundant element that requires early detection of flaws for preventing catastrophic failures. In this paper, the fatigue crack growth of a notched and flattened gearbox spline component is monitored using the AE method The AE data is continuously collected together with strain gauges. There are significant amount of continuous emissions released from the notch tip due to the formation of plastic deformation and slow crack growth. The source mechanism of sudden crack growth is obtained solving the inverse mathematical problem from output signal to input signal.
Based on the micromechanical approaches, high frequency elastic waves are generated in carbon fiber/polymer composites due to three main damage sources as fiber breakage, matrix crack or delamination. The occurrence of damage mode depends on the ratios of energy release rates. Simultaneous generation of damage modes such as multiple fiber breakage called as fiber fragmentation influences the characteristics of propagating elastic waves which are used for detecting, locating and understanding damage modes for acoustic emission method. Understanding the wave characteristics via experimental methods is difficult as the control of damage mode sequence is a challenge. In this paper, wave propagations due to single or multiple fiber breakages positioned at various locations along the laminate and through thickness in different composite lay-ups are studied using dynamic finite element models. Fiber breakage is defined as a point load with the rise time of 1 μsec. The load amplitude is identified using stress-strain curve of carbon fiber. The amplitude of matrix cracking is defined as crack opening displacement as a boundary load to the finite element model. The positions and amplitudes of damage modes are varied to understand the characteristics of waveform signatures. Each lamina is defined as a different plate and orthotropic material properties are entered as input to the model. The paper shows that multi-mode simultaneous damage in composites causes complex waveform generation, which makes pattern recognition based on amplitude and frequency real time a challenging task.