The Research On Application Of Data Fusion In Steel Pipe Damage Identification System

Because of the good characteristics of the steel pipe, it has played an important role in the current industrial production and people's normal life, which owns steel pipe a good reputation "industrial skeleton". While enjoying the convenience that steel pipe provides us with, we should also be attention to varieties of accidents caused by the poor quality during the production of steel pipe or the damage during it's overuse every year. In order to avoid the frequent accidents, it is very significant to study the damage identification of steel pipe.The magnetic flux leakage detection which is widely used as one of the numerous detection methods of nondestructive testing is chosen as the research object, according to the research on the damage identification technology of steel pipe in this paper. Then putting forward applying data fusion and wireless sensor network technology to steel pipe damage identification system, and with mobile communication technology as the supplement to realize the remote function of it. The main work of this paper is as follows:(1) The main contents and technology methods of nondestructive testing technology are introduced at the beginning of this paper, then the magnetic flux leakage detection is chosen as the main research object of this paper after combining the characteristics and application scope of these various technologies. After that the corresponding relationship between the MFL signal and pipe damage parameters in magnetic flux leakage detection technology is analyzed. And make analysis according to the advantages and disadvantages of different magnetization and data acquisition schemes of magnetic flux leakage detection, then the method in which permanent magnet magnetization as the magnetization scheme, and Hall sensor as the data acquisition scheme is selected for the damage identification system.(2) The data fusion technology is introduced in detail, including it's concept, characteristics, levels,structures, common methods and so on The two-level data fusion algorithm which included data-level and decision-level is designed based on the research of the specific characteristics during the design of steel pipe damage identification system. Data-level fusion is aimed at using median filtering fusion algorithm to fuse single Hall sensor's multiple measurements in order to improve the veracity of the collected signal; GA-BP neural network is used as the decision-level fusion algorithm to fuse the peak value and pulse width of the axial and radial components of MFL signal to obtain the specific parameters of steel pipe damage. Then the selected data fusion algorithms are introduced in detail. At the same time, their specific implementation methods are introduced and the performances of BP neural network and GA-BP network are compared.(3) The hardware and software of steel pipe damage identification system are designed. Steel pipe damage identification system is based on ZigBee network, the hardware design mainly includes the design of magnetic circuit, model selection of sensor, design of terminal nodes and gateway nodes; and as for the part of software design, the starting point is chosen on the basic of data flow within the system,then the introduction of the software design of every important section that the data has flowed by is made. Finally Lab VIEW is used to design the computer interface of steel pipe damage identification system, EXE file and Installer file are generated in Lab VIEW to facilitate the transplant of the damage identification system at the same time.(4) The performance test and experiment of the whole steel pipe damage identification system are carried out in the laboratory environment. The results show that the whole steel pipe damage identification system is easy to use and stable, and it is capable of high accuracy and betterpracticability.