Harwood
Engineering Company, Inc.
Effect of the Fluid Viscosity
RPM and Piston Configuration of Performance of Controlled-Clearance Deadweight
Tester
by D.H. Newhall, V.A. Zilberstein, and I. Ogawa*
*I. Ogawa was not of Harwood Engineering Company, Inc., but rather from
Tokyo Electron Limited, Tokyo, Japan
Controlled-clearance deadweight testers [1-4] are used widely as the primary
pressure standards above the manometer level. Our recent experimental studies
have shown [5] that the fluid viscosity is an important consideration in
evaluation of the effective area, since it affects float and stall curve
slopes. Experiments were made on DWT-300 and DWT-1000 with various fluids
such as white gasoline, light turbine oil DTE-24, Univis P12 and Spin-Esso
in a wide range of pressures up to 1.4 GPa.
It is shown (Figure 1) that the stall curves,
i.e. stall jacket pressure-measured pressure curves, for more viscous fluids,
such as a synthetic ester base oil P12, or DTE24 oil, and for "low viscosity"
fluid, namely white gasoline, do not coincide, though, conceptually, one
would expect it to be a single line. This finding is contradictory to what
was previously generally believed. Thus a choice of fluid for DWT becomes
a very vital problem. It will require both serious practical and theoretical
considerations in order to improve overall accuracy of DWT operating on
more viscous fluids.
Experimental results on the slopes of float curves plotted Figure
2 in the coordinates measured pressure — cube root of the fall rate
provide information which is shown to be sufficient to estimate viscosity
of the fluids under pressure or, at least their piezocoefficient of viscosity.
We estimated viscosity of white gasoline at room temperature from our results
obtained for a DWT-1000 at pressures up to 1.4 GPa (Figure
3).
For evaluation of the viscosity in this pressure range we used the following
relation:
m = xm³Pm
where x is a constant largely dependent on geometry and material properties
of the piston and cylinder, m the slope of a fall-rate curve, Pm
the measured pressure. The fall-rate curve slope was described empirically
on the basis of experimental data. The constant x was found by using a
boundary condition that the viscosity at 1.1 GPa should be the same as
found recently [6] for white gasoline in a high-pressure falling-body viscometer.
Since the falling body data were obtained in the range from 1.1 to 2.1
GPa there was some overlap between the DWT and viscometry data. Curvature
of the viscosity-pressure curve changes in the vicinity of 1 GPa where
the viscosity amounts to about 0.1 Pa·s. According to our data,
viscosity of white gasoline at room temperature changes by seven orders
of magnitude when pressure is increased from normal to 2.1 GPa (Figure
3). The viscosity at atmospheric pressure measured by Brookfield Laboratories
was 4.5 x 10-4 Pa·s.
DWT-1000 was run at various RPM, typically 0, 8, 16, 35, 80, and 160.
Surprisingly we found a significant effect of RPM on the performance of
the DWT in the range 0-35 RPM. The nature of the effect is yet to be studied.
From a practical point of view, determination of an optimum RPM versus
sensitivity of the instrument is highly desirable.
Until recently only plain cylindrical pistons (Figure
4a) were used in DWT. Alternate configurations were made (spherical
[7] and grooved Figure 4b and c) and tested to
.14 GPa in the DWT-1000. These pistons allowed variation in the pressure
gradient length and in the location along the cylinder of the point where
the pressure drops to atmospheric. It is shown that with oil these pistons
behave well indicating their potential for further studies of pressure
gradient effects. However, preliminary studies using white gasoline showed
the spherical and grooved pistons to be extremely sensitive to minute changes
in jacket pressure and to the relative positioning of the piston and cylinder,
resulting in poor reproducibility of the fall rate data with low viscosity
fluids. These alternate configurations appear promising as research tools
as well as practical new conceptions of piston-cylinder arrangements.
Footnotes
1 U.S. Patent 2,796,229 (1949).
2 D.P. Johnson and D.H. Newhall, Trans. ASME 75,
301-310 (1953).
3 C.O. Bennett and B. Vodar, High Pressure Measurement,
ed. by A.A. Giardini and E.C. Lloyd (Butterworths, Washington, 1963), p.
365.
4 P.L.M. Heydemann and B.E. Welch, Experimental Thermodynamics,
ed. by B. Leneindre and B. Vodar (Butterworths, London, 1975), Vol. 2,
Chap. 4, Part 3, p. 147.
5 D.H. Newhall, I. Ogawa and V.A. Zilberstein, Review Scientific
Instruments, 50 (August 1979).
6 L.H. Abbot, D.H. Newhall, V.A. Zilberstein, and J.F. Dill,
to be presented at the joint ASME-ASLE Conference, Dayton, Ohio, Oct. 1979.
7 U.S. Patent 3,630,071.
Harwood Engineering Company, Incorporated – © June, 1999
455 South Street, Walpole, Massachusetts 02081
phone: 508-668-3600
fax: 508-660-2276
http://www.ultranet.com/~harwood/
email: harwood@ma.ultranet.com
