Acoustic emission characterization of failure modes in composites with ANN Chandrashekhar Bhat a , M.R. Bhat b , C.R.L. Murthy c, * a Department of Mechanical Engineering, Manipal Institute of Technology, Manipal 576119, India b Department of Aerospace Engineering, Indian Institute of Science, Bangalore 560012, India c Department of Aerospace Engineering, Indian Institute of Science, Bangalore 560012, India Abstract Noise suppression in acoustic emission data was attempted by developing and using artificial neural networks (ANN) and with the long-term objective of in-flight monitoring. In-flight experiments conducted earlier and the noise characteristics outlined therein were taken as basis for their simulation in the laboratory. Simulated noise sources were classified through both supervised and a combination of un-supervised and supervised training of ANN. AE signals were generated by fatigue spectrum load tests on CFRP specimens and their failure modes were characterized. Finally, simulated noise and the actual signals were mixed and re-classified into their respective classes. The results obtained are encouraging and the methods and procedures adopted confirm the feasibility of the approach for field applications. Ó 2003 Published by Elsevier Science Ltd. Keywords: Acoustic emission; Signal characterization; In-flight monitoring; Noise suppression; Artificial neural network 1. Introduction Information retrieval from AE data requires that it is free from noise. Background noises normally exists in a laboratory environment and it is more so in a field condition. While the type of noise sources is context dependent, in general, background noises of a test en- vironment can be classified as mechanical, hydraulic, cyclic or electromagnetic (EMI) interference noise. Further, these noise sources may be external or internal to the component under test. In the context of in-flight, sources of noise in an aircraft in-flight are various and are difficult to identify [1–3]. Aircraft structures com- prise of a large number of bolts, fasteners and plates, which move relative to one another due to differential structural loading during flight. This leads to bolt hole rubbing (friction noise) as well as crack face rubbing (friction noise––when a crack is present) and fretting. While bolt hole rubbing leads to undesirable frictional noise, crack face rubbing may be useful noise as it in- dicates the presence of a crack. This noise along with the structural flexural noise is termed as ‘‘airframe or structural born noise’’ due to load changes. Jet engines produce appreciable noise and so also the airflow. Landing gear operation vibrations and hydraulic system vibrations give rise to extraneous noises. All these kinds of noises can be grouped under ‘‘mechanical noises.’’ Another category of noise comes from ‘‘electromag- netic interference (EMI)’’. This is due to the avionics (radar and communications) systems, relays and possi- ble bus switching between AC and battery power sour- ces. Further, EMI may arise from switching ON/OFF of various other electronic and electrical systems like mo- tor startups, sparking etc. inside the aircraft. In general, noise elimination can be through spatial and parametric filtering. Spatial filtering technique can be used with multi channel system through guard sen- sors. This method has limited scope, as all AE moni- toring systems may not have enough numbers of channels available. And, parametric filtering is very popular though it requires characterization of the noise sources, which may not be that straightforward a pro- cess. In this method, various waveform parameters of noise signals are studied and proper threshold for fil- tering the signal based on one or more combination of parameters (rise time, ring count down, energy, event duration, frequency etc.) range is set at the time of signal * Corresponding author. Fax: +91-80-3345134/3341683/3342085. E-mail addresses: bhat_chandra@yahoo.com, chandra.bhat@ mit.manipal.edu (C. Bhat), mrb@aero.iisc.ernet.in (M.R. Bhat), crlmurty@aero.iisc.ernet.in (C.R.L. Murthy). 0263-8223/03/$ - see front matter Ó 2003 Published by Elsevier Science Ltd. doi:10.1016/S0263-8223(03)00068-0 Composite Structures 61 (2003) 213–220 www.elsevier.com/locate/compstruct