Harwood
Engineering Company, Inc.
Harwood
Engineering Company, Inc.
The controlled-clearance principle (3) makes it possible to reduce packing
friction and wear to a negligible amount; a slight leak, however, must
be tolerated for lubrication. This principle provides many new opportunities
for the design of process and laboratory pressure vessels. A few have been
exploited; others, though obvious, have only been suggested. This article
describes some examples of successful employment of the controlled-clearance
principle and discusses an obvious application which, though not yet in
use, is of general interest.
The stretch vessel can, however, be controlled by applying external
pressure. This is the purpose of the controlled clearance technique (Figure
1). When pressure P1 in vessel V is atmospheric,
piston S fits snugly at C. As pressure P1
is built up, cylinder V expands and the clearance increases. An
opposing pressure, P2, from a separate source is now
applied in the annular space around C, and its effect counteracts
that of P1. By making P2 sufficiently
great, the clearance can be reduced to zero, with seizure of the piston.
In practice, P2 is adjusted in accordance with P1
with a more or less constant ratio ranging from 0.4 to 0.8, depending on
the initial fit.
The free piston gage in its simplest and most familiar form is shown in Figure 2. The closely fitted piston of measured cross section is loaded with known weights. When sufficient pressure is developed to float the piston and its weights, the piston is rotated or oscillated about its axis to reduce friction. The pressure in the liquid is defined by the ratio of total load to effective cross-sectional area of the piston. As weights are added and the pressure is increased, leakage past the piston becomes more copious, until the floating time becomes too short for taking good measurements.
A major advance in design was made by Bridgman (1) in the construction of the re-entrant type of cylinder in Figure 3. The bottom tip of the cylinder containing the piston is exposed to the same pressure that is acting on the piston — i.e., the measured pressure. Regardless of proportions of the tip, the hole in it grows smaller with increasing pressure. Its usable range is limited by initial clearance. Bridgman used a mixture of glycerin, glucose, and water for the pressure fluid; this combination, no doubt, assisted in controlling initial leakage. This apparatus, in the hands of Bridgman, was used to establish the fixed points of the pressure scale above 50,000 to approximately 190,000 pounds per square inch.
The controlled-clearance gage (Figure 4) operates
well in the same range. Until now, the limiting pressure has been set by
the liquids, and those tested so far either freeze, become to viscous to
transmit pressure effectively, or are too dangerously volatile. White gasoline
(Amoco) has been used at 0° C. to 115,000 pounds per square inch and
at room temperature to 150,000. The piston, approximately 0.080 inch in
diameter, floats for a period of 30 minutes, falling in this time approximately
1/8 inch. This is a leakage rate of 1 cubic inch
per month. With a load of 1,000 pounds, the torque necessary to rotate
the piston is 8 inch-ounces, most of which is required for the thrust bearing
supporting the load rather tan for the controlled-clearance bearing.
In operation, the rock specimen, enclosed in a thin annealed copper tube to prevent absorption of the pressure-transmitting liquid (kerosene for moderate temperature and low-viscosity silicone for higher temperatures) is placed on the lower closure plug; the bomb is set on the platen, and the ram is moved upward until the piston makes up against the load cell. Confining pressure is built up to a desired value, putting the specimen under homogenous stress. The ram (with the bomb) is now either driven further upward for axial compression tests, or it is allowed to sink for the extension tests. In either case, there is a small relative motion of piston and bomb, and it is important that friction be held to a minimum. The friction is restricted to a region within the packing sleeve where the clearance is controlled by the packing pressure. The combined minimum leakage and friction occurs when the packing pressure is about one half the confining pressure. With this arrangement, friction of the moving piston is considerably less than 1% of the load. If the piston could be made to rotate between a pair of thrust bearings at its ends, the axial friction could be reduced much more.
Diameter of the piston in Handin's bomb is ½ inch. This is large enough to permit several kinds of experiments on a wide variety of materials. For example, in a space of the same dimensions, Bridgman has performed compression, extension, and shear tests on many metals and a few nonmetallic substances. The apparatus could be readily constructed to a large scale, increasing the number of possible subjects for investigation. For instance, it may be feasible to conduct routine triaxial tests on concrete at high pressures — in building structures such as large dams, the resulting economies could easily amount to millions of dollars.
The axial loading reported by Handin could be modified by the addition
of a torque. Still further, the specimen could be a closed tube with either
a positive or negative radial pressure gradient, axial tension or compression,
and a torque. With this variety of loading, nearly all possible combinations
of triaxial stress could be used to study elastic and inelastic behavior
of materials.
To meet this situation, the Harwood Engineering Company has manufactured
stirred autoclaves, essentially as shown schematically in Figure
6. Charge, H, is contained in vessel A. The stirrer shaft
passes through closure C and gland G, with clearance at B
controlled by auxiliary pressure P. The thrust exerted by the internal
pressure is taken up on the bearings T. The stirrer in a typical
unit is driven by a ¾ h.p. motor belted to the pulley M.
Leakage amounts to about 1 cc. per minute. The design of this autoclave
is not rigid; either much less or much more power input could be accommodated.
An important advantage is that practically no heat is generated because
the bearing is practically frictionless; hence, there is no special tendency
toward polymerization at this point.
All of the apparatus described here uses shafts or pistons which are unbalanced — in one way or another, axial thrust has to be absorbed. The axial thrust is countered by dead weights in the free piston gage, thrust bearings in the stirred autoclave, and the load cell press used with the triaxial test bomb. This condition exists wherever the piston or shaft terminates inside a pressure vessel. If a shaft were extended through, using two controlled-clearance bearings, both ends would be exposed to atmospheric pressure. Assuming that the shaft had the same diameter at both bearings, there would be no axial forces developed, regardless of pressure in the vessel, because there would be no unbalanced area upon which the pressure could act. Again, as with the apparatus described here, leakage would be low and the friction loss would be negligible.
Practically, the shaft needs to be long with respect to its diameter, and limberness is a virtue. Otherwise, tolerances for alignment of the bearings would tax the best toolmaker's talent. The controlled clearance during operation is approximately 2 microinches. Experimental data show that the length of the controlled-clearance path along the axis need not be more than 1/16 inch.
The balanced shaft offers opportunities for designing many useful devices,
particularly stirrers and mixers, and for positive displacement circulating
apparatus. Because friction is negligible, torque load within the vessel
can be measured accurately. This allows the sensitive detection of such
effects as viscosity changes which indicate progress of a reaction.
