이종센서를 이용한 자기 벡터 카메라 개발 및 응용

Abstract

Some of the direct causes of damage to a wide range of structures from large-component structures such as airplanes, thermoelectric and nuclear power plants, oil refineries, and manned space stations, to ultra-small members such as integrated circuit (IC) packages and very large scale integration (VLSI), are material degradation, corrosion, and the initiation and growth of fatigue. In particular, defects within large-component structures may cause huge losses of human lives, economic losses, pollution, and damage to the reputation of a nation. If such defects are detected at an early stage and steps such as maintenance and replacement of components are taken to correct or avoid such damage, not only is safety guaranteed during the design lifetime but the lifespan can also be extended from the perspective of damage tolerance. The initiation and growth of defects can be divided into cases where introduced defects grow and cases where defects are created and grow by a random cause after the operation of the structures or devices. Further, in each case, harmful defects are to be detected and fixed through pre-service inspection (PSI) and in-service inspection (ISI). With respect to PSI and ISI, among the non-destructive tests and evaluation techniques that detect whether a defect exists and if it does, then detect its location, size, form, and dispersion as well, techniques that are better at detecting surface flaws are methods that use electromagnetic principles, such as magnetic particle testing method (MT), magnetic flux leakage testing method (MFLT), and the eddy current testing method (ECT) and surface inspection methods like the penetrant testing method (PT). The magnetic particle testing method requires pre-conditioning and post-conditioning processes before and after the test and continuously uses magnetic particles that are consumable, and although it can assess the length of a flaw, its disadvantage is that it cannot assess the depth or the width of a defect. The magnetic flux leakage inspection method is more advanced than the magnetic particle inspection method in terms of quantitative evaluation and can detect flaws and their location, form, and size by measuring the dispersion of the magnetic field via magnetic sensors; this method is being developed further and applied in developed countries such as Japan, Germany, and the U.S. However, because the magnetic field dispersion is measured using a high-accuracy scanner, problems such as detection speed, spatial resolution, and the liftoff limit along with the temperature dependence arise and need to be dealt with quickly. Moreover, the test cannot be applied in cases of paramagnetic material or metals when ferromagnetic materials are mixed with paramagnetic materials. The eddy current testing method energizes the subject with an induced current, and by measuring the impedance difference according to whether or not a flaw exists, it can evaluate whether a flaw exists as well as its location, form, and size. However, because this method uses the same scan method as the MFLT, there is a limit to the detection speed and spatial resolution, and arrays numerous sensors taking into consideration the inter-sensor interference. Moreover, in cases where ferromagnetic materials are mixed with a paramagnetic material, the presence of sectional magnetic material may be misunderstood as a mitotic defect or a large defect may be mistaken for a sectional magnetic material. The penetrant testing method uses osmotic pressure; it is used in the diaphragm and the rotor journals in nuclear power plants. However, similar to the magnetic particle inspection method, this method also requires pre-conditioning and post-cleaning; further, it is hard to implement automation or quantitative evaluation by using this method. Therefore, a national brand of non-destructive testing equipment, which does not use consumables such as magnetic particles; does not need pre-conditioning or post-processing for tests; has low power consumption; can detect surface flaws, dihedral defects, and internal defects within metallic structures composed of ferromagnetic, paramagnetic, or a mixture of ferromagnetic and paramagnetic materials at high speed and high spatial resolution in a wide area irrespective of the structural form; can be interpreted quantitatively; and most of all, has excellent field applicability, needs to be developed using only domestic technology. The technology developed for this purpose is the magnetic camera. The magnetic camera, which observes the dispersion of the magnetic field, is one of the cutting-edge IT technologies that can be applied in the field of non-destructive tests, and thus far, many studies on the development of such a technology have been conducted. The flaw detection capability of the magnetic camera depends on the sensor size (0.2 mm), array spacing (0.52~1.04 mm), and sensitivity (approx. 100 μT), and as a result, the use of this camera is limited to the detection and evaluation of flaws that are more than 1~2 mm in length. Therefore, according to the current camera technology, flaws that are more than 2 mm in length can be detected, and to detect flaws that are 50~500 µm in length, a nano-magnetic sensor implementation is mandatory for increasing the spatial resolution and sensitivity of the sensor. Further, in the quantitative evaluation of defects, because only the vertical dispersion of the magnetic field of the specimen is utilized, if the diversification of information, or real-time visualization of the three-dimensional magnetic field vector is realized and a non-destructive quantitative evaluation algorithm is developed using the abovementioned technology, relatively smaller flaws are expected to be detected and evaluated with greater accuracy. Therefore, in this thesis, we aim to develop the technology that measures three-dimensional magnetic field vectors to acquire more information, array more sensors, and have greater spatial resolution, higher sensitivity, and lower noise ratio to detect micrometer-length flaws including temperature aging. Meanwhile, a three-dimensional magnetic field vector dispersion can be realized by scanning the available three-dimensional magnetic sensors onto two-dimensional flat surfaces and reconstructing it by acquiring the triaxial magnetic field vector’s spatial dispersion. The core of such technology is in semiconductor processes that accurately align the sensor-type face of the sensors of the same type and specification along the XYZ-orthogonal coordinate axes. Hence, to align tens and thousands of magnetic sensors in a matrix form, limitations with respect to spatial resolution and minimum wiring must be solved. Therefore, in this paper, we propose a technique for acquiring a real-time video of a three-dimensional magnetic vector by expanding the existing two-dimensional sensor array technology. In other words, the existing real-time magnetic field visualization device (magnetic camera) densely arranges hall sensors two-dimensionally along a wafer, and by cross-wiring power cables and signal cables along the wafer according to each one’s original patented technology, the components vertical of the sensor can be visualized in real time. Further, according to the original technology described above, along with hall sensors, magneto resistance sensors (MR, GMR, and AMR) can be arranged using similar technology. Meanwhile, unlike the hall sensors, the MR, GMR, and AMR sensors measure the components horizontal of the sensor surface; therefore, by using a dual structure that arranges hall sensors for the Z-axis measurement and magneto resistance sensors for the XY-axis measurement, an X-Y-Z-axis triaxial magnetic field vector component can be measured using the dense array.