Theory And Experiment Research On Non-distructive Test Of Con-casting Slabs Based On Pulsed Eddy Current

The Pulsed Eddy Current Test (PECT) technique, which receivedrecognition and application in aviation industry, is an entirely newNondestructive Test (NDT) technique rising in recent decades. It not only can beused to detect the material surface flaws but also to check the sub-surface flawscompared with the visual detecting method. It enables the technicians to reducethe touches of the material and absolutely improve the detecting accuracy.Furthermore it provides electronic engineers a wider spectrum and a higherdetecting speed. Thus it is extremely suitable for the automatic detection of themetallic material (MM). Additionally, it has been broadly applied to detect theflaws during the process of metal forming, and the fatigue crack of aircraftstructure, pressure vessel, and petroleum tube in the midst of their service stages.The whole treatise, which based on the latest achievements of PECT both fromdomestic institutes and overseas ones, dedicates to study the surface cracks andsub-surface deficiencies of the continuous casting slabs. The main content of theresearch is as follows:1. Analytic modeling and theory of the NDT of the conductive platebased on PECThe magnetic vector potential (MVP), magnetic filed of the exciting coil, theinduced electromotive force (IEMF) due to the conductive plate and the electricflied in the conductive plate are derived in view of the Dodd-Deeds eddy currentresponse analytic module. Under the first class cylindrical boundary condition,the IEMF of the pick-up coil is inferred, the same as under the step currentexcitation. For the seeking the transient PEC response due to a flaw in aconductor plate, the MVP of the flaw field is got by the dipole current module.The electric field and magnetic field distribution due to a flaw can be calculatedwith the help of the dyadic Green’s function. The results of the calculationsprovide the theoretic foundation for the design of the PEC detecting system.2. The PEC defects automatic-detect system The PEC defects automatic-detect system (PECDADS) consists of the PECexcitation source based on Labwindows/CVI (CVI), the eddy current probes,and the embedded OMAP3530PECDADS of the continuous casting slabs. Inthe PECT, the stable and continuous adjustable excitation source of the voltage,frequency and duty will be dominant to acquire the PEC responses due todifferent conditions. In the most present designs, an arbitrary signal generatorplus the power driver, which suffers badly repeatability, hardly debugging andcarriage, is selected for the excitation source. The hardware components of thepulsed signal generator based on C8051F340, power driver and so on, and thesoftware testing platform of CVI are developed to satisfy the rigorous excitingparameters. To be suitable for the industrial product pipeline, the embeddedPECT detect system based on OPMAP3530is developed in this paper. Itconstitutes the hardware equipments of the DAQ data acquisition board (DAB),motor driver board, OMAP3530mainboard, the defect casting slabs separatingdevice, and so on. Accordingly, the programs of the data acquisition andcommunication of the DAB have been compiled. Moreover, the embeddedpulsed eddy application program, the fast dual-core data communication onesbased on the DSPLINK module and the data process algorithm of the waveletdenoising in the C64+DSP core are developed in the Linux platform.3. The study of fast pre-processing algorithm of PEC response basedwavelet theory (WT)In the casting product, the PEC response is influenced by the noise of themeasurement, the surface oxide of the slabs, the lift-off effect and theenvironment of the production line, when the PECDADS works under theenvironment at high temperature (above1000oC) and high humidity. Because ofthe poor SNR, it is challenging to extract the flaw features. So many strictoptimizations are applied for the C64+DSP core of the OMAP3530on the basisof WT, including data acquisition, communication, and algorithm, and so on.For achieving the pre-processing at full speed (51kHz), a practical denoisingalgorithm named “Fixed Sliding Window (FSW)” is proposed.4. MM defects classification and quantification based on spectrum analysis and wavelet transform(1) The responses due to different MM. Three specimens (Q235, T2, andAl(6061) are made to analyze the PEC response characteristic of MM with thediverse conductivity and permeability.(2) PEC defects classification and quantification based on spectrum analysis.A square exciting signal has the performances of the wide spectrum, abundantflaw response. Its basic and harmonic amplitude components include defectinformation. Hence the FFT spectrum of the MM defect response is launchedaccording to the embedded DSP advantage. After the basic and odd harmonicamplitude component features have been extracted and applied to train theclassification and quantification of the defect, the unknown specimens will beclassified and quantified.(3) PEC defects classification and quantification based on spectrum analysisand wavelet transform. A new PEC feature extracting and classifying algorithmentitled “the algorithm based on wavelet and spectrum analysis” is proposed.Using the multiresolution of the wavelet, the PEC response is decomposed intothe wavelet domain. Then the spectrum of the approximate (low-frequency) anddetail (high-frequency) components are analyzed and used to extract the low-and high-component features. So it is more precise and reliable for the defectclassification compared with the direct spectrum analysis. The research revealsthat the features of the low-frequency odd harmonic amplitudes from the basicfrequency (100Hz) to the lower frequency (1.3kHz) and the high-frequency oddharmonic amplitudes from the1.3kHz to6.1kHz can both be used to realize thedefects classification. But in the quantification experiment, it is more accurateand high linear using the low-frequency amplitude spectrum.