DEVELOPMENT OF NEUTRON TOMOGRAPHY AND ITS APPLICATIONS IN NONDESTRUCTIVE ANALYSIS

Transcription

1 Proceedings of the National Seminar & Exhibition on Non-Destructive Evaluation NDE 2011, December 8-10, 2011 DEVELOPMENT OF NEUTRON TOMOGRAPHY AND ITS APPLICATIONS IN NONDESTRUCTIVE ANALYSIS Ashish Agrawal, Yogesh Kashyap, P. S. Sarkar, Mayank Shukla and Amar Sinha Neutron & X-ray Physics Facilities, Bhabha Atomic Research Centre, Mumbai ABSTRACT Neutron tomography is a very powerful technique for nondestructive evaluation of heavy industrial components as well as for soft hydrogenous materials enclosed in heavy metals which are usually difficult to image using X-rays. It has found a variety of applications in aerospace engineering, nuclear industry, medicine, agriculture and other heavy industries. In our effort to use this technique for non-destructive testing, an experimental station for neutron tomography has been designed and developed at Beam-hole E12 of CIRUS reactor. It consists of a specially designed neutron beam collimator, shielding hutch, neutron sensitive imaging camera, sample manipulator system and control system. Several experiments have been carried out on this experimental station to study the feasibility of neutron tomography for various applications. Nondestructive detection and evaluation of location, shape and size of hydride blisters formed in pressure tube of PHWR reactor has been carried out along with quantitative determination of hydride concentration in it. Similarly the pellets as well as welded area of end-cap of PHWR fuel pin were imaged which shows imperfections in welded in the end cap and presence of defects in the pellets. Other samples which were analyzed using this facility are (a) turbine blade which is used in heat exchanger turbine systems (b) crack in SS Bar caused due to excessive tensile load applied on it, and (c) internal structures as well as distribution of chocked oil in the fuel injector of automobile engine. Keywords: Neutron Imaging, Neutron Tomography, Nondestructive Testing. PACS: Tv INTRODUCTION Neutron tomography is a form of computed tomography which is used to produce cross sectional as well as 3 dimensional images of objects non-destructively. The 3D image reconstructed by this method represents the distribution of neutron attenuation coefficient in the bulk of the object. These images can be used to characterize the objects for internal structure, detection of voids and cracks, shape and size and spatial distribution of various components in it. The more is the difference in the neutron attenuation coefficient of various components in the object, more clearly they will be visible in their 3D image. Neutrons possess high scattering and absorption cross sections with elements like Cd, Gd and most of the low Z materials specially Hydrogen hence neutron imaging is most suitable probe for investigating materials containing Hydrogen and other low-atomic-weight materials. High penetration of neutron in the bulk of heavy materials makes it a valuable tool for nondestructive examination of heavy industrial components which are otherwise difficult to image using other methods such as X-rays etc. THEORY Neutron tomography, an extension of neutron radiography, is a two-step method - tomographic data acquisition and slice or volume reconstruction via back projection. Projection images of the sample are acquired while rotating the sample along the axis perpendicular to the incident beam propagation. A collimated cone beam of neutrons with uniform cross sectional intensity distribution is made to incident on the object. Some of these neutrons interacts with the nucleus of the materials present in the object and get scattered or absorbed inside it and removed from the incident beam. The relationship between the incident neutron intensity (I 0 ) and the transmitted neutron intensity (I) is the simple exponential attenuation law - I = I 0 exp (-Σt) (1) Where (Σ) is total macroscopic cross section and t is thickness of the object. Transmitted neutrons containing important information about the object in its spatial intensity distribution are collected in a neutron sensitive imaging area detector CCD camera. After

2 NDE 2011, December 8-10, all such projections are acquired; filtered back-projection method is used to reconstruct the slices or whole 3D volume of the object. The detailed description of the tomography reconstruction algorithms can be found in Kak and Slaney [1].Using various image processing techniques desired volume and features in the reconstructed volume can be highlighted. EXPERIMENTAL SETUP We have been engaged in development and experiments on neutron and X-ray tomography since more than 10 years. Earlier we have developed neutron tomography system at APSARA reactor India and conducted several experiments [2, 3, 4]. To improve resolution, contrast of the images apart from faster image acquisition, a new experimental station has been setup for neutron tomography at CIRUS reactor India. This system consists of a thermal neutron source, 4 axis sample manipulator system, imaging area detector and data acquisition as well as control system. Thermal neutrons were extracted from a reactor (CIRUS reactor) beam hole E12. A specially designed collimator with L/D 125 has been used to collimate the beam into a cone of size 120 mm diameter at the sample stage. Neutron flux of 3x10 6 was measured at the sample position with a beam divergence of 1 0 approximately. Sample manipulator system consist of rotation motion ( ) with a precision of and three axis translational motions (X, Y, Z) for remotely placing the sample at the suitable location in front of the beam. A neutron imaging camera consisting of Neutron sensitive scintillator screen LiF(ZnS), to convert the transmitted neutron beam in to electromagnetic radiation, mirror and CCD camera was designed and developed in house and has been used for image acquisition. The spatial resolution of camera was 200 µm with 16 bit dynamic range. The rotations of sample are synchronized with image acquisition and controlled through user developed software. RESULTS AND DISCUSSION Several neutron tomography experiments have been carried out on this experimental station for example visualization and analysis of hydride blisters in Zircaloy pressure tube of PHWR reactor, quantitative analysis of hydride concentration in Zircaloy up-to few ppms, visualization of end cap of PHWR and BWR fuel rod, shape and defects in fuel pellets, Visualization of cooling channels in turbine blade, cracks in SS bar due to tensile load and nondestructive evaluation of fuel injector of automobile engine and choked oil in it. Samples were placed over a rotational stage and rotated with an angular step depending upon the size of the sample and resolution required. Radiographic projections were collected for a given time usually for few seconds at each rotational position and several projections (typically of the order of 10 3 ) were collected while rotating the sample in A full tomographic scan took 4-8 hours. 1. Neutron tomography of blisters in PHWR pressure tube Hydride blisters formation in Zircaloy pressure tubes of pressurized heavy water reactor (PHWR) is a major problem which hinders the safe and uninterrupted operation of reactor. Though an individual blister is unlikely to defy leak before break criteria of pressure tube design, an array of blisters joined together by delayed hydride cracking (DHC) may form a crack larger than the critical crack length, which can cause catastrophic failure of the pressure tube [5, 6].Since pressure tube is the primary containment for the hot pressurized coolant, Fig. 1 : A schematic of experimental set up as well as configuration of various components.

3 338 Ashish Agrawal et.al : Proceedings of the National Seminar & Exhibition on Non-Destructive Evaluation Fig. 2 : (a) 3 dimensional volume image of pressure tube showing three blisters (b) enlarged view of blister (c) volume image showing migration of hydride towards blister. catastrophic failure of pressure tube will result in complete loss of coolant, which if unabated for long period may eventually lead to release of activity. In view of this, it is important to study and examine the blister formation process in pressure tubes and quantify the hydride concentration in it so that the severity of the problem can be identified. A sample of Zr-2.5Nb pressure tube section of length ~ internal diameter 100 mm, external diameter 105 mm and length 150 mm was gaseously charged with 100 wppm of hydrogen [6]. Three hydride blisters 120 apart from each other were grown on the outer circumference of tube sections [5]. Size of blisters formed was measured to be 2 mm approximately. This sample when images using neutron tomography three hydride blisters in the sample containing hydrogen concentration of wppm can be clearly seen in the reconstructed 3D volume images(figure 2a-2c). This figures also shows the shape and size of the blister formed which shows that blister has a shape of an elongated ellipsoid while its size is approximately 2 mm. Information about size of blister, its shape, its depth, presence of cracks inside blister and orientation of crack with respect to hoop stress, migration of hydrogen towards blister are required for safety assessment of the pressure tubes. Since the shape and size of the blisters formed depends upon shape and size of the contact tip, thermal boundary condition and tube section dimensions, varying these, the dependence over various parameters can be calibrated for their effect on the shape and size of the blister. 2. Quantitative Analysis of Hydrogen Concentration in Zircaloy Since hydrogen has a relatively higher attenuation coefficient for neutrons, its small presence in zirconium alloy affects the resulting grey values in the image significantly. Hence concentration of hydrogen can be detected, quantified and calibrated against the grey values in the reconstructed image. Various samples of Zr-2.5 Niobium charged with different concentration of hydrogen (25, 50, 75, 100 wppm) were prepared and investigated using neutron tomography. It is clearly seen that the concentration difference of 25 wppm can be clearly distinguished in the reconstructed 3D images in Fig 3(a-c). To calibrate the concentration of hydrogen in the samples against the image intensities (grey values) in the reconstructed image, histogram of the 3D volume image of all the samples were calculated separately. Equal area was selected in the reconstructed image of each sample. Fig. 4(a) shows the plot of histograms of each sample. It is seen from the histograms (figure 4a) that the most probable grey value (Mode of the distribution) in the reconstructed images keeps on increasing with the increment of hydrogen Fig. 3 : (a) 3D dimensional volume image of coupons charged with different hydride concentration of 25, 50, 75 and 100 ppm (b) Another view showing clear difference in grey values (c) a slice image

4 NDE 2011, December 8-10, Fig. 4 : (a) Histograms of grey values in the reconstructed slices of Zr coupons charged with 25, 50, 75 and 100 ppm (b) calibration of hydride concentration with respect to average grey value in the reconstructed image. concentration in the volume. This is because with more and more hydrogen in the sample, larger number of voxel in the image possesses higher grey values hence the mode of the histogram distribution shifts towards higher values. A variation on peak heights in the histogram is attributed to the nonuniform distribution of hydrogen in the sample as described in the previous section. A graph showing the variation of hydrogen concentration value with respect to mode of the distribution has been plotted in Fig.-4(b) which was linearly fitted to show the direct proportionality between the two quantities. 3. Nondestructive Experiments on PHWR Fuel Rod Pressurized heavy water reactor (PHWR) consists of a fuel bundle with large number of fuel elements each of which is hermetically sealed by two end caps welded to cladding tube by pressure resistance welding. [7] Several cases have been witnessed for the failures in the weld of end caps. A defect/failure of a single end cap weld needs removal of 8-12 fuel bundles from the channel. It is essential to have highly reliable end cap welding which guarantees integrity of the fuel pin. The nondestructive examination of the quality of this weld using neutron tomography has been carried out which provides high resolution 3D images of the inspected volume of the sample. Defects and imperfections can be clearly visualized in the welded zone of the end cap (figure-5a, 5b). In the same experiment the quality of fuel pellet in the fuel pin was also examined. In the 3D image of several pellets one pellet was found having some crack (figure-5c, 5d). which may propagate during operation of the reactor and cause serious safety problem. 4. Neutron Tomography of Fuel Injector Fuel injection system for admitting fuel into an internal combustion engine has become the primary fuel delivery system used in automotive petrol engines which has almost completely replaced carburetors. A fuel injection system Fig. 5 : (a) 3D volume of Zircaloy cladding of fuel pin (b) image showing weld of the end cap (c) 3D image showing crack in the pellets (d) a slice image of the cracked pellet.

5 340 Ashish Agrawal et.al : Proceedings of the National Seminar & Exhibition on Non-Destructive Evaluation Fig. 6 : (a) 3D volume image of fuel injector of automobile engine (b, c) other views of 3D volume (d) 3D image showing distribution of chocked oil only. Fig. 7 : (a) 3D volume image of lower part of turbine blade (b) Enlarged image of encircled section (c) 3D volume image of Upper part of turbine blade. atomizes the fuel by forcibly pumping it through a small nozzle under high pressure. Various designs of fuel injectors has been developed for various applications to optimize power output, fuel efficiency, emissions performance, ability to accommodate alternative fuels, reliability, drivability and smooth operation, initial cost, maintenance cost, diagnostic capability, range of environmental operation etc. Nondestructive evaluation of a fuel injector can provide the design engineer a valuable insight of the tool to examine its quality as well opportunity to optimize various parameters. We have carried out neutron tomography experiment on a fuel injector sample and able to visualize its each and every internal component (Figure-6). The fuel choked in various parts can also be seen clearly in the reconstructed 3D images. 5. Neutron Tomography of Turbine Blade Turbine blade, a component of turbine section of a gas turbine, is responsible for extracting energy from the high temperature, high pressure gas produced by the combustor. They are very crucial component of gas turbines which has to survive in difficult environments like high temperatures, high stresses, and a potentially high vibration environment. All three of these factors can lead to blade failures, which can destroy the engine hence turbine blades are carefully designed to resist those conditions. Exotic materials like super alloys and different methods of cooling, such as internal air channels, boundary layer cooling, and thermal barrier coatings are used for make effective cooling in turbine blades. Feasibility of neutron tomography has been studied for nondestructive visualization of internal structures and defects in cooling channels in a turbine blade which can provide the valuable information regarding its design, cooling efficiency and overall performance. Figure 7 shows the volume images of a sample turbine which clearly shows its internal structure and design of cooling channels. Such images can also be used to find out defects in cooling channels or with improved resolution, defects caused by difficult environment during operation. 6. SS Bar Crack due to Tensile Load The strength of a material is its ability to withstand an applied stress without failure. Excessive load applied on a given material beyond its strength, causes its failure and micro-cracks get started to appear in the volume of material. To study the feasibility of applicability of neutron tomography in the study of visualizing cracks in the material caused by excessive

6 NDE 2011, December 8-10, Fig. 8 : (a) 3D volume image SS bar with crack (b) another view of bar showing crack (c) slice of volume near crack showing its propagation. loading, we carried out experiments on a stainless steel bar (100mmX100mmX25mm) which has a crack caused by excessive tensile load applied on it. The results shows that the crack of size upto minimum 200 micron can be visualized clearly (figure-8a-8c). CONCLUSION Neutron tomography is valuable tool for nondestructive analysis of objects from a wide range of applications. It is the most suited method to visualize the low density materials enclosed in the bulk of tool made of high density materials. Due to high sensitivity to hydrogenous materials or similar low Z material neutron tomography is best tool to detect them in the matrix of high Z material such as Zircaloy. Tools from various industries can be inspected for 3 d visualization of internal structures, quality assurance and reverse engineering. ACKNOWLEDGEMENTS We acknowledge the constant support from the staff of Reactor Operation and maintenance divisions of CIRUS reactor particularly from Dr. A. K. Sahu, Sri N. Ramesh, Sri Anil Bhatnagar, Sri R.N. Singh, Sri J.L Singh, and Sri N. K. Mondal. We acknowledge the constant support and guidance of Dr S. Banerjee Chairman AEC, and Dr. S. Kailas Director Physics group. REFERENCES 1. Kak, A. C. and Slaney, M., Principles of Computerized Tomographic Imaging. New York: IEEE Press(1988). 2. Amar Sinha, A.M. Shaikh, A. Shyam,Nuclear Instrum. Method. B. 142 (1998) Amar Sinha Rev. Sci. Instrum., Vol. 71 No. 3, March P.S.Sarkar, A.Sinha, Y.kashyap, M.R.More,B.K.Godwal, Nucl. Instr. Meth. A 524, pp (2004) R. N. Singh, R. Kishore, T. K. Sinha and B. P. Kashyap; Jl. of Nuclear Materials vol. 301, (2002) R. N. Singh, R. Kishore, T. K. Sinha, S. Banerjee and B. P. Kashyap; Materials Science and Engineering A 339, (2003) J. L. singh, Arun Kumar, V. R. Nair, E. Ramadasan, S. Majumdar and K. C. Sahoo ; Proceedings of 14 th World conference on NDT New Delhi, (1996) Vol. 2, pages