Research On Fuzzy Pattern Recognition For Welding Damage On Earplate Connection Of The Guyed Mast

Mast structure is widely used in the important lifeline engineering and the national economic construction,but the collapse incidents of guyed masts at home and abroad have caused great social impact and economic losses.The survey found that a lot of accidents of the guyed masts were caused by welding fatigue of the ear plate connection of the guyed mast structure.So it is of great importance to diagnose the ear plate damage timely.Considering the complexity of the influence of random variations of cable force of the guyed mast structure under wind load on the ear plate surface strain,a technology to diagnose the welding crack length is proposed which is by measuring the cable force and the strain on the ear plate surface to identify the damage degree of the mast structure using the two-step fuzzy pattern recognition method.And the technology is systematically studied by finite element method as well as experiment.Specific contents are as follows:1.Based on the sensitivity analysis of the strain of the ear plate surface to the welding crack length,the technology is proposed which is by using the two-stage fuzzy pattern recognition method through measuring the strain on the earplate surface;2.The finite element model of the whole guyed mast structure and local finite element model of ear plate are established.Through the wind vinduced response analysis of mast structure finite element model and selection of the key points of ear plate surface,the pattern library of cable force-key point strain-weld crack propagation length are established;3.The effective principle of identification is determined,and the membership function for identigying cable force and the crack propagation length is put forward.In order to demostrate the effectivity of the recognition method,simulation identification analysis is performed;4.A group of full-scaled lug joint model with welding crack are designed and fabricated.Taking the measuring strain of key points of these different models under different cable force as the object to be identified,the welding crack length of the different models are identified by two-stage fuzzy pattern recognition method based on the pattern library.Results show that it is effective and feasible to identify the crack propagation length of the welding crack on earplate of the mast structure by proposed technology which has a great practical walue in engineering.