Automated System for Magnetic Particle Inspection of Railway Wheels

This paper presents a new developed automated system used for flaw detection and classification on railway wheels. The described system uses the principle of non-destructive method based magnetic particle inspection. Our proposed and finally constructed automated system is used in industrial applications where non-destructive testing of railway wheels is desired. As a part of testing line, it is possible to detect both surface and subsurface flaws. Using this system, it is possible to test about 15 wheels per hour. The system consists of mechanical parts (frame), inspection parts (cameras) and application software. Application software includes advanced signal and image processing methods used for efficient flaw detection. Using our proposed automated system it is possible to safely detect flaws presented at any part of railway wheel. Introduction High safety standards required in the management of railroad lines demand the inspection of railway wheels directly after production in order to detect the presence of surface cracks that could seriously affect the integrity of the railway, and therefore passengers’ safety. During the last two years, we have been developing an automatic system for surface defect detection on railway wheels. The main goal was to develop the highly reliable system based on image processing algorithms that gives a warning of surface flaws to prevent possible future accidents. The system is based on magnetic inspection where the railway wheel is put into the magnetic field and the cracks cause a magnetic stray field. By this technique the cracks are visible and they can be easily recorded using high speed digital camera. The mentioned testing of railroad wheels is based on magnetic particle inspection. Magnetic Particle Testing Magnetic particle inspection (MT) processes are non-destructive methods [1] for the detection of surface and sub-surface defects in ferrous materials [2]. They make use of an externally applied magnetic field or electric current through the material, and the principle that the magnetic flux will leave the part at the area of the flaw. The presence of a surface or near surface flaw in the material causes distortion in the magnetic flux through it, which in turn causes leakage of the magnetic fields at the flaw. This deformation of the magnetic field is not limited to the immediate locality of the defect but extends for a considerable distance; even through the surface and into the air if the magnetism is intense enough. Thus the size of the distortion is much larger than that of the defect and is made visible at the surface of the part by means of the tiny particles that are attracted to the leakage fields. Magnetic particles are usually applied as a suspension in water or paraffin. This enables the particles to flow over the surface and to migrate to any flaws. The most sensitive technique, however, is to use fluorescent particles viewed under UV (ultraviolet) light. System proposal As described in previous section, our system is based on magnetic particle testing. As our system is developed for inspection purposes of railway wheels we set initial requirements that have to be finally fulfilled. The first requirement was to scan the whole surface of railway wheel in different wheel diameters. The desired wheel diameter is within 500 – 1300 mm. In general, the system had to be flexible to scan the surface in different volumes. This also corresponds to magnetization. Different wheel diameters require different magnetization current. Another requirement was to scan the surface automatically, without human intervene. For this reason we proposed new highresolution digital cameras that were automatically driven and scanned the surface of the wheel. With these cameras, it was necessary to scan all corners and drapes of the wheels. The last main requirement was to propose and implement efficient signal processing methods used to safely detect all flaws in exact sizes. The minimal size of detected flaw was determined to 1 mm in length and 0.3 mm in thickness. These values are based on standards defined in rail industry. As described system was as a part of complex inspection line used for subsurface defect detection using ultrasonic non-destructive testing it has to be also prepared to inspect wheels in adequate speed. System overview The system requirements mentioned above have been kept in mind during the system proposal. As a basis for railway wheel inspection was the magnetic particle method. Based on this, we propose settings of inspection line. The main part of proposed inspection line the coil used for magnetization of inspected wheels was used. Fig. 1. System configuration for railway wheel testing The configuration of used coil has to correspond to correct magnetization [3] of railway wheel. The whole volume of railway wheel has to be magnetized. To obtain such magnetization we decided to rotate the railway wheel during the magnetization process (i.e. during the current was applied to the coil). The Fig. 1 shows the proposal of coil configuration and location of railway wheel in the coil. In the developing stage, we simulated the magnetic field to find out the optimal position of railway wheel in magnetic field for the best defect visibility. In Fig. 3., the streamlines of magnetic flux density represents the magnetic field. The strongest magnetic field (flux density) can be visible at the bending of coil. It means, in case the railway wheel has sizable diameter than is better magnetized. As we need to reach the appropriate magnetization of all predefined wheels (all required diameters), it was necessary to use the second spiraled coil located near Fig. 2. Configuration of coils 1 coil 2 coil to the centre of the railway wheel (see Fig. 2.). The second spiraled coil was used to improve the total magnetization of railway wheel and make the magnetic field stronger in relation to definitions mentioned in standards. Finally, the main goal was to detect flaws in all directions Fig. 3. Magnetic flux density To make the decision about the proper configuration of coils and location of railway wheel, the first measurements were performed. For measurement, six measured points was determined. These can be seen in Fig. 4a. The results of magnetic intensity H measurement are visible in Fig. 4b.. Pricne mereni sonda horizontalne 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4