· Table of Contents
· Fundamental & Applied Research
Thickness evaluation of pipes using density profile on radiographsSung Sik Lee and Byung Gyu Jang
Korea Inspection & Engineering Co., Ltd., Seoul, Korea
Young H. Kim
School of Mechanical Engineering, Sungkyunkwan University, Suwon, Korea
Dept. of Civil Engineering, Daejon University, Daejon, Korea
Korea Infrastructure Safety & Technology Corporation, Goyang, Korea.
Keywords : inspection of pipeline, radiograph, density profile, thickness profile.
A new general equation for determination of wall thickness from measured film density profile in a radiograph was derived. Using this equation, thickness and film density profiles for a non-insulated and vacant pipe were calculated. As a result of calculation, they appeared continuous and symmetrical around the center of pipe. An experimental testing on a non-insulated carbon steel pipe with artificial flaws of different depth is carried out with Ir-192 and industrial film. The measured density profile on the radiograph shows good agreement with calculated one. Comparing both density pfofiles, it was found that the scattered radiation is increased as penetration thickness is increased. The result of the present work shows that thickness variation of the curved specimen can be evaluated by measuring density profile on a radiograph.
Nondestructive evaluation of remaining wall thickness of pipes is one of fundamental issues in the assessment of structural integrity of chemical plants. Ultrasonic testing has been widely used for measuring pipe wall thickness, however, most of pipelines in chemical plants is insulated and insulation should be removed before ultrasonic testing is applied. The removal of insulation requires expense and normally is restricted to sampled small areas, therefore it may be difficult to detect local small metal loss with the ultrasonic testing. Therefore, radiography as a non-contact method has been widely used to overcome these problems.
The tangential radiography, one of well known film radiography technique for this purpose (Lee and Kim 1998), has some distinctive merits compared with ultrasonic testing. However, its inspection range is also restricted to relatively small area where the wall thickness image is shown on a radiograph and relatively small size of pipes which are less than about 8 inches (OD), which is due to the limited penetrating power and the growing scattering effect with increment of pipe size when the radioisotope, Ir-192 is used as a radiation source. Additionally, because thickness evaluation is performed directly on a radiograph with a measure, it is required to minimize geometrical unsharpness for the enhancement of resolution and sharpness on the radiograph. Therefore it needs long distance between radiation source and film as possible as and it will requires much more exposure time to obtain proper density on the radiograph, which may also result in much more absorbed dose to the radiographer and also lower productivity in the work.
Conventionally, wall thickness has been evaluated by comparing film density of radiograph to that of step wedge placed beside specimen. However, this technique can be used for the limited film density and wall thickness, and it is hard to use this technique to the specimen with thickness variation in wide range. For the specimen with curvature such as pipes, the penetration thickness is largely varied. For example, the maximum penetration thickness is several times of the wall thickness. Therefore the conventional method could not be utilized for the evaluation of wall thickness of pipes.
In this study, a new general equation is suggested to determine the thickness profile with measurement of density profile in a radiograph by considering the type of radioisotope and its radioactivity, the characteristic properties of industrial film, source to film distance and build up factor. By using this equation thickness profile and density profile are calculated about non-insulated pipes. Finally, with radiographic testing about a pipe having several artificial notches, the applicability of density profile method is examined.
The radiation intensity decreases exponentially, depending on linear absorption coefficient, as its penetration distance increases (Cullity 1977). The linear absorption coefficient also changes according to effective energy of radiation and chemical composition of test material (Cullity 1977). In considering the intensity of radiation incident on unit area of detector in broad beam condition, the intensity decreases inversely as the square of the distance from the source. On the other hand because incident radiations on unit area of film are composed of directly penetrated radiation and scattered radiation, the build up factor must be considered in broad beam condition (Halmshaw 1991). Therefore considering mentioned factors, the intensity of penetrated radiation incident on unit area of film may be expressed:
where I0 is intensity of incident radiation, B is build up factor, m is linear absorption coefficient, x is thickness of test material and d is source to film distance.
Considering the characteristic curve of industrial film, the film density(D) increases exponentially with logarithmic exposure(log E) in the density range of 0 - 3.5 depending on film speed and film gradient (Lee et al 1999). Therefore in film radiography the measured density(D) on a radiograph may be expressed:
where E0 is exposure, which is radioactivity(Ci) multiplied by time in the case of gamma ray, d is logarithmic film gradient and fs is film speed. As a result, the density profile on a radiograph may be calculated by considering the source to film distance, thickness of test material and its linear absorption coefficient, film speed, film gradient and build up factor due to scattered radiation. Because it is not easy to calculate the variation of build up factor with thickness, it may be difficult to calculate exactly the density profile with simply theoretical calculation. Nevertheless, the equation (2) may give systematic understanding about radiography in which how each factor will affect to the test result and the trend of density profile with thickness.
When a test specimen is an insulated pipe as shown in Figure 1, the penetration thickness will change depending on the incident angle(q) of radiation into the pipe. So, the penetration thickness will be determined as a function of r0,ri,r which are the radius of outside diameter, the radius of inside diameter and the distance from the center of pipe respectively.
where ti is thickness of insulation, k is its equivalent factor compared to carbon steel. Therefore thickness and density profiles may be calculated with easy through the equation (2) and (3).
|Fig 1: Geometrically different penetration thickness due to different incident angles into a pipe.|
In broad beam condition, the radiation passing through the center of pipe will penetrate simply twice of the nominal thickness of a pipe. On the other hand, the radiation passing through the pipe at an incident angle with the pipe center line shall penetrate relatively long thickness. The penetration thickness increases until the incident angle reaches to the tangential angle in which radiation pass through the inside diameter of pipe tangentially. After passing the tangential angle, the penetration thickness decreases rapidly with enlargement of the incident angle.
When the pipe is surrounded by uniform insulation material, the penetration thickness can be calculated as the same way. Because the insulation material has different radiation attenuation property compared with metal pipe, its equivalent factor must be considered to calculate the thickness and the density profiles exactly.
The radioisotope, Ir-192 (9.5 Ci) is used as a radiation source. The pipe whose OD and nominal thickness are 165mm and 10 mm is used as a test specimen, where the flat bottom holes with different depths such as 20, 30, 40, 50, 60, 70% compared to the nominal thickness are fabricated. In radiographic testing the notched surface is toward to the film and the radioisotope is located to the center of pipe with 600 mm of SFD. A lead plate of 5 mm thickness is disposed under the film to minimize the influence of scattered radiations from the bottom wall of cement. The test pipe is irradiated until the optical density becomes 3.0 in the center of pipe after development with automatic film development processor which can maintain the developing time and temperature as constant. The film densities on the radiograph are measured by densitometer having 0.01 density precision.
Fig 2: Test specimen and its radiograph.
The test pipe having artificial notches and its radiograph is shown in Figure 2. It is found that the density difference is large between the notched area and the sound area. The density profiles measured and calculated circumferentially along the pipe in the radiograph are shown in Figure 3. In the calculation, it is assumed that the equivalent factor of insulation material is 0.15 which is ratio of the linear absorption coefficient of an arbitrary insulation material to the linear absorption coefficient of pure iron for 300 keV radiation. As a result, it is found that the calculated density profiles are continuous and symmetrical for the non-insulated and the uniformly insulated pipe. In On the other hand, the measured density profile is very continuous and symmetrical in the sound area of the pipe and the density difference is large and its profile is discontinuous over the notched area. On comparing the measured density profile to the theoretically calculated density profile, it is shown that the discrepancy is increasing as it go to the far distance from the center. This is because of scattered radiations which is not considered quantitatively in the theoretical calculation. So, the discrepancy is closely related with the growing effect of scattered radiation with increment of penetration thickness as shown in Figure 4.
|Fig 3: Density profile measured on a radiograph and theoretically calculated density profile.|
|Fig 4: Experimental value of build up factor for different penetration thicknesses.|
Figure 5 is plotted with data extracted from the dotted box in Figure 3 and from a step wedge which is tested with the pipe specimen. Theoretically the slope depends on linear absorption coefficient and logarithmic film gradient. The slopes obtained by least mean square are 0.017 and 0.019 for the pipe and the step wedge respectively. Supposing the average energy of Ir-192 is about 300 keV, the linear absorption coefficient of pure iron is 0.087/mm and the logarithmic film gradient is 1.80±0.15 (Agfa D5), the slope will be in the range of 0.027~ 0.032. However, the test material is not the pure iron but carbon steel containing light elements of relatively small linear absorption coefficient, and although the energy spectrum of radioisotope Ir-192 is complex the main line spectrums are 310, 470 and 600 keV, which constitute almost 90 % of its radiations. Therefore, considering the linear absorption coefficient decreases with increment of energy, the measured slope must be lower than the calculated slope that is based on the pure iron and mono-energetic radiation of 300keV.
|Fig 5: Logarithmic density variation with thicknesses of a reference step wedge and with circumferential thickness change of pipe.|
Fig. 6 is evaluated results about the depth of notches in longitudinally direction of the test pipe through the density - density relation based on the step wedge. The evaluated values are thicker than the real thicknesses excepting the center of pipe. This is because the radiation source is centered on the pipe and the radiation can not penetrate perpendicularly into the pipe excepting the center, so the penetration thickness will be longer than the real nominal thickness. Therefore, to evaluate the thickness of pipe exactly, it is necessary to correct the increment of penetration thickness due to the radiation incident angle.
|Fig 6: Measured thickness of flat bottom holes fabricated on a pipe using thickness-density relation obtained from a reference step wedge.|
In Figure 7, the evaluated and calculated thickness profiles are shown, which are evaluated through the density-thickness relation obtained from the dotted box in the Figure 3 and theoretically calculated from the sound pipe respectively. Although the evaluated and the calculated profiles correspond very well In the center of pipe, the difference between the calculated and evaluated is increasing as it go to the far distance from the center. This is due to growing effect of scattered radiation with increment of penetration thickness, which shows the possible range to be applicable to the pipe by density profile technology. Therefore it may be concluded that it needs at least two shooting of radiography spaced 90 degree apart in a single plane around the pipe circumference to know the exact thickness profiles of pipe in the broad beam condition.
|Fig 7: Comparison of geometrically calculated thickness profile and evaluated thickness profile through measured density profile on a radiograph.|
|© AINDT , created by NDT.net|||Home| |Top||