Study Of Defect Data Visualization And Human-machine Interface Based On FSM

With the constant development of national industry, large steel structures areoccupying an increasingly important position in the national economy. Due to variety oflarge steel structures such as offshore oil recovery facilities, ship hulls, offshore platform,oil or gas pipeline and so on, are under complex environment of ocean and atmosphere for along time, they are abound to produced different degrees and types defects. These defectscan affect the mechanical properties and the safe operation of teel structure, and maybecome major safety accidents. So, steel structure defects detection, tracking and analysis,not only an important part of safety detection, but also an important means of getting safetyinformation, and can simultaneously provide original data and theroretical guide for furthersafety assessment and reliability analysis of steel structures, has important scientificsignificance and practical value.FSM (Field Signature Method) is a new kind of non-destructive testing techniquesstarted from1990s. Due to the features such as high sensitivity, convenience, non-invasivemeasurement and so on, it has unique advantage in the field of defects monitoring of steelstructures. This article design a set of detection system based on FSM, this system includehardware part and software part. The hardware includes power supply module, electrodemodule and data acquisition module. The software part includes data analysis, datavisualization, and data display on interface.First, in order to study field signature of circular and crack defects effect, this paperdesigns the type, size and layout of the two type defects, and designs the layout of electrodearray in the measurement area, and collect the voltage signal between two pins. Second, in order to quantitatively describe the correspondence between defect and field signature, thepaper introduces FC coefficient. The paper analyzes the relationship between defects’ depthfactor and FC value, and studies the causes of involve effects caused in practical.Last, the paper studies the recognition of cracks’ depth factor based on BP neuralnetwork, and dose actual verification using the trained e neural network. According to thegraphics and data showed by visualization software, it is showed that the neural network hashigh precision and can be used in actual detect on depth of cracks.