Research on Non-destructive Automatic Identification Algorithm of Aircraft Structural Defects Based on Machine Learning

Nondestructive testing is a new comprehensive applied technology discipline. It can control the product quality in processing and manufacturing and ensure the safe operation of in-service equipment. Nondestructive testing is widely used in machinery, metallurgy, electronics, chemical industry, railway, shipbuilding, nuclear energy, aviation, aerospace, electric power industry and other industries. In the civil aviation system, the aircraft has to undergo a pressurization and decompression in each lifting process, which is easy to fatigue the materials around the rivets fixing the aircraft surface and form radioactive micro cracks. When the crack grows to the critical crack length, it will expand rapidly, and finally lead to structural fracture and failure. The fracture and failure of this structure is also one of the most important reasons for air crash accidents. The research on structural defect detection technology of aging aircraft has been paid more and more attention by aviation organizations all over the world. How to detect the crack defects in the multilayer structure of aging aircraft has always been a difficulty in the field of nondestructive testing. Pulsed eddy current technology is an electromagnetic nondestructive testing technology which can effectively detect defects in multilayer structures. The formula of pulse eddy current penetration depth is deduced theoretically. It is concluded that appropriately reducing the pulse excitation frequency and increasing the duty cycle are conducive to the detection of deep defects. The experimental system and rectangular sensor are designed, and the frequency and duty cycle of the excitation signal are optimized. The inner defects in multilayer structure are tested, and the necessary data processing is carried out for the weak detection signal. The experimental results show that the pulse eddy current testing technology can effectively detect the inner crack defects. Pulsed eddy current technology will play an important role in the field of aviation nondestructive testing. In this paper, aircraft skin defects are taken as the research object, and the recognition algorithm of structural defect inspection robot is studied. A new method based on fuzzy support vector machine to automatically identify rivet crack defects in magneto-optical image of aircraft skin is proposed.