·Table of Contents
Digital X-ray TV System in serial production inspection of welds
V. Mezhuev, A. Kuznetsov, A. Akhmetov, A. Kruteev
PJSCo " Mashinostroitelny zavod ", Electrostal, Russia.
During the weld quality estimation it is important to determine the weld geometry parameters, namely, the dimensions of weld defects, such as pores, undercuts or gas channels as well as the weld reinforcement value.
Methods of traditional X-ray radiography in principle allow to determine basic parameters of weld defects: size of defects shadow images, location depth in the weld and defect size in the X-raying direction.
As for the determination of the defect shadow size - it is rather simple and causes no difficulties during X-ray examination.
The determination of the depth of defect location is not a technically complex task either. Its limited application in routine weld examination is explained only by multiple operations and time consumption.
Determination of the defect size in the direction of X-raying is rather difficult because of the need to measure the values of transparency difference of small parts of X-ray film and to take into consideration many factors affecting the stability of this procedure. This is the reason why this parameter can be determined only under laboratory conditions but not in industrial shop environment.
Development of X-ray TV system with modern computer processing of video-signals gives the opportunity to perform such measurements in real time. This idea was realized at PJSCo "Mashinostroitelny zavod" for the weld inspection of nuclear fuel assemblies. JSC "MSZ" is the Russian leader in manufacturing nuclear fuel for Nuclear Power Stations in Russia and Europe. In 1996 at JSC "MSZ" traditional X-ray radiography was substituted by X-ray TV System MU80F designed under special order by Philips Industrial X-Ray GmbH (Hamburg).
This system was designed and initially used for the search and size determination of shadow defects only. However during the operation at JSC "MSZ" there were developed more complex methods of welds examination. It became possible due to variety of functions of image processing software of this system.
Basic characteristics of X-ray TV System MU80F
- System MG164 including X-ray tube of 160 kW with focal spot of 0.2 mm;
- Image intensifier PRS220 with resolution 12 - 21 Lp/cm as function of magnification and TV camera of CCIR standard (625 lines);
- PC with PXV2000 program to process and archive images;
- Monitors displaying live images and images produced after digital processing;
- Relative sensitivity 2 %;
- Dimensions of inspected objects:
|max thickness for stainless steel
|max thickness for zirconium alloys
|| 10 mm.
Measurement of a defect size in the direction of X-raying
The advantages of image digital processing in X-ray TV system are mostly evident in case of defect measurement in the direction of X-raying.
The main idea is rather simple - to measure the image brightness of the defect and compare it with that of defectless area. But in film X-ray technique it is a very approximate and times consuming procedure. On the contrary, in digital X-ray TV System there are various possibilities to perform these operations very precisely within short time.
In System MU80F this is provided by two interconnected forms of X-ray images presentation:
Raster image is a result of digital processing of live X-ray image. Program PXV2000 allows to perform integration of consequential images to reduce "noise" effects and carry out different digital image filtration to increase contrast and border sharpness of individual elements of X-ray image. These operations are necessary to ensure stability and accuracy of brightness measurements.
- live shadow image on display;
- raster digital image in format TIFF, 770 x 573 pixel; the value of one pixel is 0.04-0.1 mm in the object plane depending on the magnification. Brightness of video image can vary from 0-black to 256-white.
Fig 1: Image of defect and groove-type penetrameter with graph amplitude of brightness versus distance along the marked axis.|
Brightness measurements of video image can be carried out in two ways.
- Brightness evaluation along the line marked by the operator on the image and presentation of the results in form of the graph (see Fig.1) - amplitude of brightness versus distance along the marked axis.
Graph evaluation along the line makes measurements more obvious especially when working with extended defects.
- Determination and statistical evaluation of brightness histogram (Fig.2,3) of the definite image area selected by the operator. The histogram of a number of pixels with certain brightness value as a function of a brightness amplitude. The program generates average brightness value and other statistical parameters.
Fig 2: Weld image with histograms of background and defect regions.
Fig 3: Weld image with total histogram of background and defect region.
Brightness evaluation by means of a histogram has some advantages - it gives more accurate results of measurements because large number of image points is taken into account thus reducing "noise" effects.
It is possible to apply two techniques for region selection:
- Selection of two regions: that of the defect and that of neighbouring background (Fig.2). It is recommended to select the background region from two opposite sides of the defect and compute the mean value.
- Selection of one total region if the dimensions of the defect are small (Fig.3). In this case it is necessary to obtain a histogram with two brightness peaks, corresponding to background and defect. The brightness values of background and defect are obtained as average of each peak of the histogram.
The other important point of this method is that brightness dependence on thickness of material is nonlinear. Hence, thickness measurements must be properly calibrated. For calibration the following tools were used:
As a result experimental relationships were determined and approximation functions were calculated (Fig.4) for thickness change vs. the difference of brightness of the initial thickness (background) and local thickness change (defect).
- groove-type penetrameter (Russian standard GOST7512-82) with groove depth 0,1 - 1,75 mm ;
- thickness plate standards 0,1 - 12 mm.
Fig 4: An example of the curve of local thickness decrease as a function of "relative brightness difference" of the defect and background regions (material thickness 9,0 mm).|
The functions of thickness decrease and increase were determined for the initial weld thickness in the range from 2 to 14 mm the pitch being 0,5 mm.
For each function optimal System parameters were determined: X-ray tube voltage, brightness, contrast, image intensifier format and others.
Optimal parameters and the effects limiting the method
As a result of examinations the concept of procedure was developed, which provides stable measurement results and an acceptable measurement error.
During investigation the following effects were studied:
In particular, the following was determined:
- optimum parameter of the System:
- X-ray tube voltage,
- image intensifier scale,
- opening of the diaphragms of X-ray tube and image intensifier,
- parameters of live image brightness and contrast,
- influence of the shadow defect size,
- influence of a set of defects neighbouring to the examined one.
if there are other defects near the evaluated one of a comparable or larger size, the thickness X-rayed conventionally decreases. As a result the image becomes lighter, less contrast, that leads to the decrease of absolute brightness difference, in a smaller degree - of a relative difference. Nevertheless, the effect of this factor can not be evaluated as a systematic error, because of the random nature of real defects.
- influence of the shadow defect size (in the plane, perpendicular to X-raying):
during inspection by plates (10 mm wide) the effect of the defect length was found when brightness was measured in the background region. It was determined that in case of defect width being more than 10 mm, measurement error due to total non-uniformity of brightness achieve 50% of the measured value. That's why 10 mm is considered to be the limit plane size of defect which can be measured by this method. At the same time this value is sufficient to inspect weld reinforcement in most cases;
- influence of a set of defects neighbouring to the examined one:
The error of the method was evaluated with the help of test objects of different thickness. The total relative error of the method does not exceed 17%. This value meets the requirements of the standards for measuring equipment.
As a result of investigations the method of defect size measurement in the direction of X-raying was elaborated.
Total X-rayed thickness is up to 14 mm;
The material of parts welded together is stainless steel of any type.
The basic types of examined welded joints are butt and lap joints.
Defects of welds:
- When examining the value of local thinning of the weld material the defects can be the following: pores, non-penetrations, gas channels, burns, undercuts, craters, blowholes;
- When measuring the value of local thickening of the weld material the weld reinforcement was observed.
This procedure has already been used to help in many nonstandard situations in weld examination. The procedure is not complicated and can be performed by X-ray inspector of common skills and does not take more than 5 minutes from the moment of a defect founding out.