Traceability of Acoustic Emission measurements for a proposed calibration method – Classification of characteristics and identification using signal analysis. (January 2015)

• Record Type: Journal Article
• Authors: Griffin, James
• Is Part Of: Mechanical systems and signal processing. Volume 50/51(2015)
• Page Start: 757
• Page End: 783
• Publication Date: 2015-01
• Subjects: Single grit scratch tests -- Acoustic Emission -- Force -- Time–frequency domain -- Tensile tests -- Signal analysis -- Neural networks -- Fuzzy-c clustering -- CART rule based system and Simulink models; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2014.04.018
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: When using Acoustic Emission (AE) technologies, tensile, compressive and shear stress/strain tests can provide a detector for material deformation and dislocations. In this paper improvements are made to standardise calibration techniques for AE against known metrics such as force. AE signatures were evaluated from various calibration energy sources based on the energy from the first harmonic (dominant energy band)[1, 2] . The effects of AE against its calibration identity are investigated: where signals are correlated to the average energy and distance of the detected phenomena. In addition, extra tests are investigated in terms of the tensile tests and single grit tests characterising different materials. Necessary translations to the time–frequency domain were necessary when segregating salient features between different material properties. Continuing this work the obtained AE is summarised and evaluated by a Neural Network (NN) regression classification technique which identifies how far the malformation has progressed (in terms of energy/force) during material transformation. Both genetic-fuzzy clustering and tree rule based classifier techniques were used as the second and third classification techniques respectively to verify the NN output giving a weighted three classifier system. The work discussed in this paper looks at both distance and force relationships for various prolonged Acoustic Emission stresses. Later such analysis was realised with different classifier models and finally implemented into the Simulink simulations. Further investigations were made into classifier models for different material interactions in terms of force and distance which add further dimension to this work with different materials based simulation realisations. Within the statistical analysis section there are two varying prolonged stress tests which together offer the mechanical calibration system (automated solenoid and pencil break calibration system). Taking such a mechanical system with the real-time simulations gives a fully automated accurate AE calibration system to force and distance measurement phenomena.