This paper presents the latest results obtained from Acoustic Emission (AE) monitoring and detection of cracks and/or damage in diaphragm couplings, which are used in some aircraft and engine drive systems. Early detection of mechanical failure in aircraft drive train components is a key safety and economical issue with both military and civil sectors of aviation. One of these components is the diaphragm-type coupling, which has been evaluated as the ideal drive coupling for many application requirements such as high speed, high torque, and non-lubrication. Its flexible axial and angular displacement capabilities have made it indispensable for aircraft drive systems. However, diaphragm-type couplings may develop cracks during their operation. The ability to monitor, detect, identify, and isolate coupling cracks on an operational aircraft system is required in order to provide sufficient advance warning to preclude catastrophic failure. It is known that metallic structures generate characteristic Acoustic Emission (AE) during crack growth/propagation cycles. This phenomenon makes AE very attractive among various monitoring techniques for fault detection in diaphragm-type couplings. However, commercially available systems capable of automatic discrimination between signals from crack growth and normal mechanical noise are not readily available. Positive classification of signals requires experienced personnel and post-test data analysis, which tend to be a time-consuming, laborious, and expensive process. With further development of automated classifiers, AE can become a fully autonomous fault detection technique requiring no human intervention after implementation. AE has the potential to be fully integrated with automated query and response mechanisms for system/process monitoring and control.
This paper presents the results obtained during the development of a semi-real-time monitoring methodology based on Neural Network Pattern Recognition of Acoustic Emission (AE) signals for early detection of cracks in couplings used in aircraft and engine drive systems. AE signals were collected in order to establish a baseline of a gear-testing fixture background noise and its variations due to rotational speed and torque. Also, simulated cracking signals immersed in background noise were collected. EDM notches were machined in the driving gear and the load on the gearbox was increased until damaged was induced. Using these data, a Neural Network Signal Classifier (NNSC) was implemented and tested. The testing showed that the NNSC was capable of correctly identifying six different classes of AE signals corresponding to different gearbox operation conditions. Also, a semi-real-time classification software was implemented. This software includes functions that allow the user to view and classify AE data from a dynamic process as they are recorded at programmable time intervals. The software is capable of monitoring periodic statistics of AE data, which can be used as an indicator of damage presence and severity in a dynamic system. The semi-real-time classification software was successfully tested in situations where a delay of 10 seconds between data acquisition and classification was achieved with a hit rate of 50 hits/second per channel on eight active AE channels.
Proc. SPIE. 5392, Testing, Reliability, and Application of Micro- and Nano-Material Systems II
KEYWORDS: Principal component analysis, Statistical analysis, Pattern recognition, Manufacturing, Inspection, Control systems, Data acquisition, Acoustic emission, Signal generators, Picture Archiving and Communication System
In order to develop an effective and accurate way to monitor and control the quality of fiberglass products, Acoustic Emission (AE) signals, generated during compression of fiberglass samples, were studied and analyzed using neural network based pattern recognition software. Distinguishable patterns were found in samples manufactured under different conditions and compositions, which resulted in different product quality. AE waveform features, such as absolute energy, average frequency, duration, and rise time were analyzed and the features showed strong dependence on the sample tested. This made sample classification possible and definitive and therefore a classifier was developed and applied to data collected from additional test samples. Finally, an AE system for the evaluation of fiberglass insulation was designed and built. It is expected that the developed system will be used as a quality control tool in industrial production of fiberglass insulating material. In this paper we will discuss the AE data collection and analysis, classifier development, and give an overview of the inspection system developed.