Study On Visual Nondestructive Detection And Evaluation Method For The Steel Wire Rope

As one of the most important flexible components,the steel wire rope(SWR)is widely used in the lifting and transportation systems,such as mine hoists,cranes,elevators,cableways.Due to its working characteristics,the SWR body,especially the surface of the rope body,is inevitably damaged due to abnormal friction,scratch,collision,rope biting,bending fatigue and other problems in the service process.If it cannot be detected or replaced in time,the SWR may even fail,can lead to major safety accidents in the hoisting and transportation system,threatening the life and property safety of users.Therefore,it is important to use reliable and efficient nondestructive detection method to improve the safety of the rope.However,it is difficult to detect the damage of SWR because of its complex structure,diverse working environment and the limited detection approaches.To date,the non-destructive detection method of SWR still has some shortcomings,such as low intelligence,poor detection effect,incomplete damage information,and unable to make reliable estimation and evaluation of SWR damage in time.Based on this,this thesis focuses on the image acquisition and processing,damage visual recognition,damage quantitative estimation,state evaluation and visual dynamic acquisition and detection experiment of SWR by adopting visual method as the main means,and introducing swarm intelligence optimization,deep learning,finite element,mechanical stretching and other methods and technical means.The research results are expected to solve the problems of small damage target,large interference,complex morphology and difficult evaluation,so as to replace manual inspection to achieve efficient detection and evaluation of SWR surface damage.The main research contents are as follows:1)Firstly,the wire rope basis and the theoretical basis based on data-driven methods are introduced.Secondly,combined with the actual situation,the working conditions and common damage types of SWR are analyzed to master their common characteristics.Finally,the overall scheme of visual inspection and evaluation method is designed.And the scheme is analyzed,designed and explained on the basis of establishing the visual state model framework of cumulative damage on the whole rope surface,which paves the way for the later work.2)In order to obtain clear and standardized image data of wire rope,the efficient acquisition and standardized processing method of SWR image is investigated.Firstly,according to the visual perception theory,the visual perception model of SWR is established,and the relationship between the perception parameters is deduced.Then,the basic architecture of wire rope circumferential image acquisition system is designed to provide basis and reference for different practical applications.On the basis of the above,the image data normalization processing method is proposed,including pose adjustment,region of interest(ROI)extraction and filtering denoising.The wire rope image preprocessing under non normalized pose and illumination is realized.Finally,the dynamic vision acquisition system of SWR is established and the dynamic acquisition experiment is carried out.The results show that: the dynamic acquisition of wire rope image is greatly affected by the exposure time.So it is necessary to set a smaller exposure time for shooting under a larger light intensity,and zoom lens imaging is needed for small target damage.3)To solve the problem that it is difficult to accurately detect the surface damage of SWR,a data-driven visual recognition method for surface damage of SWR is studied.Firstly,based on the introduction of the data-driven method and the key algorithm theory,a data-driven visual recognition method architecture for rope surface damage is proposed.Then,the rope surface damage diagnostic classification method based on feature engineering is studied,including multi factor influence analysis,multi texture feature extraction and fusion,hybrid algorithm design and diagnosis results.Meanwhile,the influences of feature operator,feature dimension and algorithm performance are explored.Secondly,to further improve the intelligent degree of damage classification,an end-to-end diagnostic classification method of rope surface damage based on the improved Le Net-5 is proposed.In this method,the Le Net-5 structure is further improved to realize the rope damage diagnosis classification,i.e.,the accurate diagnosis of rope damage.Then,to realize the location and identification of wire rope surface damage,the recognition method of rope surface damage based on the improved yolov3 is studied.Finally,the damage simulation experiment platform is built,and the visual dynamic acquisition experiment system is used to carry out the dynamic detection and recognition experiment of SWR surface damage.The results show that the end-to-end algorithm has higher detection accuracy and more intelligent process compared with the feature engineering algorithm.In particular,there is no need for artificial feature extraction process,which is more suitable for engineering applications.4)To grasp the health status of SWR with surface damage,the status evaluation method of wire rope based on quantitative cumulative damage of surface is studied.Firstly,the state evaluation strategy of wire rope with surface cumulative damage is proposed,and the relationship between visual inspection and state evaluation of wire rope damage is established.Secondly,combining with image processing technology,a quantitative damage evaluation method is proposed.Then,based on this strategy,in order to build the state accumulative library,a finite element model is established which can simulate and analyze the stress-strain distribution of different damaged wire ropes under tension load,and the effect of damage on mechanical properties is qualitatively explored.Subsequently,the tensile test of the wire rope under the cumulative breakage and wear damage is carried out,the mechanical properties of the wire rope under the surface quantification cumulative damage are proved,and the actual remaining mechanical strength under different damage is quantitively acquired.Finally,based on the test,simulation results and relevant regulations,the SWR health status is evaluated.The results show that the proposed evaluation method can grasp the state of SWR surface damage,and the indexes given in the regulation can be used as a rough reference value because different types,structures and damage conditions of SWR in practice.There are 127 pictures,20 tables,and 177 references.