Density measurements of high-velocity metallic sprays using x-radiography and high-speed photography
C E Lloyd1, W G Proud1 and A Rimmer2
1 SMF Group, Cavendish Laboratory, Madingley Road, Cambridge, CB3 0HE
2 AWE Aldermaston, Reading, Berkshire, RG7 4PR.
1
Abstract. Understanding and measuring particular flow is important in a wide range of industrial applications. In this paper, a combination of flash x-rays, 70 ns duration, and high-speed photography were used to observe density variations within high-velocity, >500 m s-1, metallic sprays. The absorbance levels in radiographs were related to a known line of sight mass. Using the spray dimensions, as measured in high-speed photography images, an average spray density was found. Additionally, the spray velocity was obtained from the high-speed photographs, allowed density vs. time plots to be produced. The densities typically fluctuated within the sprays, suggesting that higher density zones were present within them. Maximum densities varied from ~320 to 820 kg m-3. The density vs. time traces obtained from the x-ray images were compared directly with independent density vs. time measurements obtained by commercially available piezoelectric sensors. The density values measured by the piezoelectric sensors were similar to of the x-radiography technique. It is suggested that slight differences in the density values stemmed from oversimplifications in the assumptions used in the analysis of the piezoelectric sensor response.
1. Introduction
If a shock wave reaches a free surface, the resulting interaction between the shock wave and the solid/vacuum interface results in particle ejection. Efforts have been made to model the complex mechanisms responsible for ejecta production.[1,2] However, quantitative experimental data are required to check the validity of these models. Density distribution measurements within ejecta clouds would assist in understanding the evolution of ejecta. The ejecta cloud has hydrodynamic properties and may travel at velocities of a few kilometres per second.[3-7] Over the past two decades, the dynamics of ejection has attracted the attention of both experimental and theoretical physicists.[1, 3-8] Intricate techniques such as interferometry, [7,3,9] holography [4] and x-radiography [1] have been used to study the behaviour of shock loaded surfaces. Flash x-radiography is a particularly useful technique for the study of fast-moving ejecta clouds. The short duration of an x-ray pulse from modern flash x-ray units (typically <50 ns) makes it possible to obtain high-resolution images of objects travelling at velocities of up to a few km s-1. The relatively large field of view that is available with most x-ray systems also allow for the whole shape of the dynamic event, at a given stage in its development, to be recorded. In addition, since the amount of x-radiation absorbed by any object relates to the line of sight mass of the object, x-radiographs can give a two-dimensional projection of the mass. Mass distributions within thermal spray jets [10], high density plasmas [11] and hypervelocity objects [12] have all been studied in this manner.
Recent research efforts at AWE and Los Alamos have concentrated on developing a piezoelectric sensor system to obtain density values of sprays.[5,6,8] The sensors of interest are produced by Dynasen Inc, and consist of a small (1.2 mm diameter, 0.4 mm thick) PZT-5A sensing element. When impacted along the PZT’s poled z-axis, the PZT element undergoes a polarisation pz along the z-axis that is linearly related to the stress σzz. Since the sensor arrangement is largely resistive in nature when terminated at Z = 50 Ω, the stress σzz can be related to the time-integrated output voltage V(t) of the sensor,
(1)
where A is the area of the PZT element, and d33 is the piezoelectric charge coefficient along the z-axis. For PZT-5A, d33 = 374 pC N-1. If the impacting material transfers all of its momentum to the piezoelectric sensor, the density ρ(t) can be related to the stress σzz(t) and impacting velocity u,
. (2)
However, little quantitative data has been published on the behaviour of the sensors in spray environments. Comparing the sensor measurement with those obtained by the well-established
x-radiography technique will allow the validity of the new diagnostic technique to be investigated.
This paper presents a method of measuring the density distribution within a high-velocity,
>500 m s-1, spray cloud, using x-radiography and high-speed photography. The density measurements could then be compared with density measurements obtained from the piezoelectric sensors. Unlike the ejecta sprays produced in more traditional ejecta experiments, the gas gun technique allows greater control over the particle size distribution, velocity and diameter of the spray. The suitability of the
x-radiography and high-speed photography density measurement method could then be assessed before being applied to less predictable ejecta sprays.
2. Experimental procedure
An aluminium powder and Microballoon powder mixture was prepared and packed into hollow acrylic sabots at a density of 403 ± 17 kg m-3. A single-stage light gas gun was used to accelerate the sabots down a 25 mm diameter gas gun barrel. A sabot stripper at the end of the barrel stopped the sabot and allowed the powder to travel onwards, thus producing a spray.
A DRS Hadland Ultra-8 high-speed camera and an Scandiflash 150 keV x-ray unit, which produces polychromatic x-ray pulses of 30 ns FWHM were used to obtain density measurements of the spray. Both systems were triggered by a stress gauge system at the sabot stripper. X-ray images and high speed photographs of the spray were obtained at orthogonal angles. Agfa Ortho CP-G Plus Medical
x-ray film encased in Agfa Ultra-Vision Detail x-ray cassettes with image intensifier screens were used to obtain x-ray images of the spray.
Figure 1 Schematic of the experimental arrangement.
The calibration for the relation between the intensity measured by x-ray film and the density of the material through which the beams traversed was achieved by imaging wedges of the aluminium/Microballoon mixture. The mass of powder per unit area could be deduced for the whole wedge, since the dimensions of the wedge and the packing density of the powder were known. In this fashion, each intensity level on the x-ray image could then be converted to its corresponding mass per unit area. Each film was calibrated separately, to avoid systematic errors. Information regarding the dimensions of the spray was obtained from the high-speed camera images. Spray density measurements for given areas of spray imaged on the x-ray film could then be obtained by dividing the equivalent mass per unit area by the thickness of the spray at that point.
Figure 2 Example of a calibration graph relating the relative absorbed x-ray intensity to the equivalent mass per unit area for aluminium/Microballoon mixtures of known density.
As a complimentary method of measuring the density of the spray, a piezoelectric sensor was placed a few centimetres away from the end of the barrel, and sampled the spray. The voltage response was recorded on a 500 MHz oscilloscope. Equation (2) was used to obtain density measurements from the voltage signal. The sensor also detected the time of arrival of the spray, and could therefore be used to measure the spray velocity.
3. Results and Discussion
Spray velocities were obtained from high-speed photographs (figure 3). For firing pressures of
50 MPa, the typical spray velocities were found to be ~750 m s-1. The images showed that the sprays were directional and reproducible.
An x-ray image of a spray taken at the same instant as the central frame in figure 5 is shown in figure 6. Since the x-ray intensity across the exposed film was not uniform, the absorbed intensities had to be normalised against a background “null” intensity in order to retrieve density values. Densities of the spray taken across various cross-sections are shown in figure 6. The maximum spray density detected by the x-ray image was 480 kg m-3. X-ray images of other spray experiments with the same powder mixture found slightly lower maximum densities of ~320 kg m-3.
Figure 3 High-speed photography sequence of a gas-gun produced spray.
Cross-sections of the density profile of the sprays are shown in figure 3. As the velocity of the spray was known, density vs. time plots could be produced. From the radiographs it was clear that the densities fluctuated within the sprays, suggesting that clumps were present within them. This method allowed the number, size and density of those zones to be measured.
Figure 4 X-ray image of a gas gun-produced spray, with various density cross-sections.
Density vs. time plots were also obtained by the piezoelectric sensors (figure 5). All the plots taken were of similar shapes, showing a fast initial rise in density at the front of the spray jet, as seen in the x-ray images. The sensors consistently gave density readings of ~800 kg m-3, which were slightly higher than the densities obtained by the x-ray imaging technique. Several factors could be responsible for the sensor’s higher density measurements. Assumptions made in the analysis, such as the linear constitutive equation related stress and polarisation, and the complete momentum transfer between the spray and the sensor might not be valid.
The piezoelectric sensor appears to be unresponsive at later times during the spray impact. Damage accumulation in the sensor element during impact will degrade its performance.
Figure 5 Density vs. time measurements obtained from piezoelectric sensors.
4. Conclusions
Reproducible, low density, high-velocity sprays have been produced. These have been studied in several ways. Two-dimensional projections of the high-velocity spray densities have been obtained using high-speed photography and x-radiography. The images showed that some degree of clumping occurred within the spray. Maximum density measurements varied from ~320 to ~820 kg m-3.
A direct measurement of the density using piezoelectric sensors was used. The sensors consistently gave density measurements of ~800 kg m-3. By comparing their performance with the x-radiography and high speed photography technique, the suitability of the piezoelectric sensors for spray environments can be assessed. This research is on-going and forms a part of a larger study.
Acknowledgements
The authors are grateful to AWE Aldermaston, UK for its financial support. EPSRC is thanked for grants in support of the high-speed camera system.
References
[1] J. Chen, F. Q. Jing, J. L. Zhang, D. Q. Chen, and J. H. Wang, Journal of Physics-Condensed Matter 14, 10833-10837 (2002).
[2] Q. F. Chen, X. L. Cao, Y. Zhang, L. C. Cai, and D. Q. Chen, Chinese Physics Letters 22, 3151-3154 (2005).
[3] J. R. Asay, L. P. Mix, and F. C. Perry, Applied Physics Letters 29, 284-287 (1976).
[4] D. S. Sorenson, R. W. Minich, J. L. Romero, T. W. Tunnell, and R. M. Malone, in Shock Compression of Condensed Matter-2001, Pts 1 and 2, Proceedings; Vol. 620 (2002), p. 531-534.
[5] W. S. Vogan, W. W. Anderson, M. Grover, J. E. Hammerberg, N. S. P. King, S. K. Lamoreaux, G. Macrum, K. B. Morley, P. A. Rigg, G. D. Stevens, W. D. Turley, L. R. Veeser, and W. T. Buttler, Journal of Applied Physics 98 (2005).
[6] W. S. Vogan, W. W. Anderson, M. Grover, N. S. P. King, S. K. Lamoreaux, K. B. Morley, P. A. Rigg, G. D. Stevens, W. D. Turley, and W. T. Buttler, in A New Spin on an Old Technology: Piezoelectric Ejecta Diagnostics for Shock Environments, Baltimore, MD, 2005.
[7] J. R. Asay and L. M. Barker, Journal of Applied Physics 45, 2540-2546 (1974).
[8] C. S. Speight, L. Harper, and V. S. Smeeton, Review of Scientific Instruments 60, 3802-3808 (1989).
[9] J. R. Asay and T. G. Trucano, International Journal of Impact Engineering 10, 35-50 (1990).
[10] H. Tamura and M. Itaya, Journal of Thermal Spray Technology 9, 389-393 (2000).
[11] M. Desselberger, T. Afsharrad, F. Khattak, S. Viana, and O. Willi, Applied Optics 30, 2285-2286 (1991).
[12] F. Mizusako, K. Ogasawara, K. Kondo, F. Saito, and H. Tamura, Review of Scientific Instruments 76 (2005).
© British Crown Copyright 2006/MOD