Research On Source Characteristics And Pattern Recognition Of Acoustic Emission For Crane

As special and large mechanic-electrical equipment,crane is widely used.Its safe operation is closely related to social and economic development.There is a large number of lifting equipment in use in our country with a fast growing speed.By the end of 2007,there had been 957,900 cranes in use,which has increased by 72.3% compared with the same period in 2003.Conducting non-destructive testing(NDT) plays an important role in guaranteeing the safe operation of cranes.In the past few years,acoustic emission(AE) technology as a new NDT method has been widely used in such fields as pressure vessels,aerospace engineering,etc.Compared with such conventional NDT as ultrasonic testing(MT),magnetic particle testing(MT), radiographic testing(RT) and penetrant testing(PT),AE testing(AET) has many advantages such as being sensitive to active defects,whole monitoring the structures or equipment in a testing,having short testing period as well as high efficiency.But in the NDT of cranes,it is still at the initial stage.Moreover,it lacks cognition of AE sources in the crane operation field.Therefore,it is urgent to study AE sources characteristics in the crane operation field and to seek for an efficient AE sources recognition method.It is also the premise in drafting AET standard for crane and conducting field inspection.By referring to the studies of the 11th Five-yea Plan of the China Key Technologies of R&D Program(No.2006BAK02B04),the dissertation studies all kinds of classical AE sources characteristics and their recognition methods in the working process of crane.The tasks finished are as follows:(1) Through the tensile testing of the Q235 and Q345 steel both material and welding;the AE characteristics of the four specimens were acquired.The results indicate that:the AE behavior corresponds with the internal damage of the material; the yield point of the specimen can be clearly observed from the AE RMS voltage curve and the energy rate curve,and the weld specimen yields many times especially for Q345 steel with two-yield phenomenon,which can not be observed in the stress-strain curves. (2) On the basis of AE sources characteristics acquired from the AE inspection testing in the process of tensile testing of the Q235 and Q345 steel and bending testing of box beam and trough specimen,the author conducted a destructive testing on large structural—box girder of the crane with surface cracks in the weld.AE phenomena in the testing process are also inspected.Compared with the results of the stress and metal magnetic memory testing on the area of prefabricated surface cracks in each levels,the author acquired the AE sources characteristics of surface cracks propagation in the weld and plastic deformation,including AE location features,parameter distribution features and wave spectrum characteristics.(3) By conducting many crane field AET,the author systematically obtained six kinds of classical AE characteristics of AE sources from the working process of overhead travelling cranes and portal bridge cranes.They are noises caused by moving vehicles and trolleys,brake in lifting and descending,structure friction, peeling of the oxide and paint,raindrops and electrical equipment noises respectively. Meanwhile,the author finished measuring the attenuation curve of the crane girder and linear location testing.The results show linear location testing can accurately locate the AE sources of box girder and truss girder.(4) The author studied the application of wavelet analysis in processing crane AE signals.The rules of how to select the suitable wavelets for AE signal processing were analyzed,and Daubechies wavelet was selected for the crane AE.According to the Mallat arithmetic,the maximum decomposition level of wavelet analysis was also formulated.Moreover,the author fixed the frequency bands on each decomposition level of the signal and proposed the characteristic extraction method for crane acoustic emission resources based on the energy spectrum coefficients of wavelet analysis.It is proved that the method can accurately recognize the four AE sources including surface cracks propagation,plastic deformation,structure friction as well as moving vehicles and trolleys.(5) The author studied the application of neural network in the pattern recognition of crane AE wave signal.Based on studying BP network structure and algorithmic selection,the improved BP algorithm was proposed.Combining the extraction of AE signal characteristics based on wavelet analysis,the author designed and trained a neural network which can accurately recognize classical AE sources of crane working process.By recognizing the signal acquired from actual crane AE testing,its reliability was validated.The research results in this dissertation will contribute to the comprehensive understanding of AE sources and lay a foundation for the standard drafted of AE examination and evaluation of cranes.It is of vital importance and utility value in promoting the application of AE technology in the field of cranes.