Harwood
Engineering Company, Inc.
Harwood
Engineering Company, Inc.
Autofrettage is a method of developing favorably oriented stress distributions in pressure-containing bodies. The autofrettage increases the elastic strength of pressure-containing bodies, makes gross change in their resistance to fatigue and inhibits the rate of crack propagation. Autofrettage requires pumping equipment capable of generating pressure considerably in excess of that seen in actual service. Similarly, studies of fatigue and crack propagation rates require pumps capable of generating controlled cyclical high pressure to test specimens. This paper describes high pressure apparatus which has been supplied over a 30 year period to ordnance manufacturing agencies and to industry for these applications.
Early in the 1920's, Bridgman2 was responsible for the initiation and development of autofrettage equipment for the U.S. Army and Navy. He published an account of that experience which showed how readily he was able to apply his techniques in small scale high-pressure equipment to the large-scale equipment required.
The author, an officer in the U.S. Army (1939-1946) was in charge of the main autofrettage plant where approximately 60,000 cannon tubes were processed for World War II. This involved turning an experimental facility into a production plant with the attendant problems of some component redesign, maintenance, operation, training, safety, etc. To the author's knowledge this was the first production plant operating at pressures up to 10 Kb. The process used was the container method3,8 which made it possible to develop full autofrettage over the entire length of the gun tube using one application of pressure, resulting in a floor-to-floor 12 minute cycle to process a 90 mm tube. There were no records of self destruction of any gun tubes which had been autofrettaged. There were, however, records of a few failures of weapons heat treated to strength without autofrettage, particularly at proof firing.
Cannon tubes made prior to World War II and some made during that war were built of relatively low strength alloys. Some were heat treated to strength and the others autofrettaged. They were of heavy proportions and would fail eventually because of erosion. The course of failure was surface embrittlement, followed by fatigue cracking and subsequent erosion of the crack patterns by the hot, chemically active high velocity gases. These weapons would become unusable before the fatigue crack growth became critical. This picture began to change with the developing need for lighter weight tank, aircraft and other highly mobile weapons which were inherently more crack-sensitive. Ordnance designers turned to high quality, high strength steels in gun tubes of slim proportions. Fracture toughness and resistance to low cycle fatigue became an important preoccupation in ordnance research leading to more stringent forging specification. It is a a common practice to take test coupons from the ends of forging. While good test results from the coupons are hopefully indicative of the properties of the material between, they in no way guarantee the integrity of the material of the forging between the coupons. Autofrettage exposes the entire length to over-pressure. Thus it is a powerful quality assurance control, since the overstraining reveals defects manifested by bore and/or outside cracks and an occasional rupture of the forging.
Midway in World War II, the author, recognizing the emergence of low cycle fatigue problems, designed apparatus to produce cyclical hydraulic pressure in sections of gun tubes.4, 12 The tests showed substantial gains in the life of autofrettaged specimens. That work has been continued and expanded to include studies of crack initiation, growth, and fracture toughness5, 6 under the direction of Dr. Davidson at Watervliet Arsenal. The equipment generating cyclical pressure for these studies was designed and manufactured by Harwood Engineering Company.
Frequently, it is necessary to rough-machine forgings prior to autofrettage and in some cases a fair amount of machining is necessary to reach final dimensions, particularly in the case of gun tubes. Concern for the effect of the final machining on the distribution of the residual stress resulted in various investigations.7, 8, 9, 10
After World War II, the skills to design high pressure equipment were largely concentrated in a very few laboratories conducting high pressure research and in ordnance agencies. The introduction of polyethylene by ICI during World War II sparked the industrial interest in high pressure. After the war, the use of high pressure proliferated in other scientific disciplines and in industry.
Harwood Engineering, founded in 1948, from the beginning offered pressure generators, piping, valving and pressure measurement equipment with a working range to 14 Kb. Over the ensuing years, Harwood has supplied 14 Kb autofrettage apparatus to defense agencies and forging manufacturers world wide. Smaller 14 Kb pumping equipment has been found useful to processing firms and to research agencies in the U.S. and abroad. The sizes of equipment range from multiple units of 150 horsepower at one extreme to small air- and manually-operated units at the other. Ancillary equipment related to the autofrettage has also been supplied.
Ordnance agencies are not alone in their interest in low cycle fatigue. Similar apparatus has been supplied to the chemical processing industry. Cyclical pressure loads on apparatus containing as little as 2 Kb are sufficient to cause failure in the presence of stress raisers and metallurgical defects.
When pressure is applied to the bore, the cylinder expands as shown in Figure 1. The cylinder dilates elastically from a to b, plastically from b to c and upon the release of pressure, elastically from c to d. At c in Figure 1 the hoop stress distribution (e to f in Figure 2) is plastic, but in dropping the pressure from c to d (Figure 1) an elastic stress distribution g to h (Figure 2) is subtracted from the plastic leaving locked into the cylinder, the residual stress distribution i to j in Figure 2. At k the value of elastic stress equals the plastic, hence the residual stress is zero. A reapplication of any pressure less than the autofrettage pressure will produce elastic expansion following d toward c (Figure 1), essentially paralleling a to b. Pressure in excess of c will create additional permanent set.
Figure 2 represents full autofrettage which occurs when plastic flow has gone completely through the wall. In partial autofrettage, plastic flow has gone only part way through the wall and the outer part of the cylinder wall remains in an elastic state of stress. The process is essentially the same for partial autofrettage. This description of the influences on material properties of the autofrettage process is oversimplified for the sake of brevity. For a more complete discussion, the reader is referred to references 1, 2, 13, 14, 15, 16, 17.
Timoshenko16 pointed out that favorably oriented residual stresses, even with pieces of equipment with complex geometry and unknown stress raisers, could be developed simply by overloading the part once into the plastic or partially plastic stress state. It is important that the overloading be in the same direction as the normal loading of the piece. otherwise the residual stress would be harmful.
The units described here are single-acting with pressure generated in one direction only and double-acting wherein two high pressure pistons are driven by a common double-acting low pressure cylinder. Single-acting units have an intermittent delivery. Except for the compressibility of the high pressure fluid, double-acting units pump continuously.
When an intensifier makes a new stroke, it has to first compress the high pressure fluid to the system pressure before it can start delivery fluid to the system. Thus, a pulse in delivery is generated. Harwood devised and patented drive systems in which the pulse is essentially eliminated18, 19 but this refinement is not necessary in autofrettage because the pulse does not present a problem.
There is a fair choice of stroke limiting devices for effecting the reversal of the intensifier action, some quieter than others. In Figure 4 the flow control valve (FCH) consists of a variable throttle shunted by a pressure compensating valve. The flow control valve offers essentially infinite variation in the pumping rate. The throttle valve can be remotely adjusted either manually from the control panel, or it can be made responsive to controllers to slow the pumping rate at higher pressure if desired.
Electric-motor-driven (25 to 250 horsepower) Vane type pumps are generally used (PF) for they are quiet, reliable and forgiving of the haphazard maintenance so often existing in production shop conditions. It is protected against overload by the pressure regulator (RR) installed close to the pump which is preset at the factory.
The return lines are collected and directed through the fluid conditioning system consisting of a water-cooled heat exchanger with a thermal regulator to conserve cooling water and a differential filtering system not unlike that used in automobiles. The oil reservoir is generally of 100 or more gallon capacity.
The drive system described above is not of radical design, but has proved to be dependable. The components are designed for easy maintenance in heavy industrial areas. The drive system operates in the 100-140 bar range with components rated at 200 bars, assuring long trouble-free service. Experience has shown that sophisticated servo-valves with characteristic small orifices etc. require too much maintenance in the factory environment to be practical.
The high pressure manifold contains external check valves, providing simple substitutions and eliminates the stress concentration brought about by the cross drilling required for internal check valves usually found in low pressure pumps. The high pressure packings are carried on a floating piston, unattached to the high pressure ram, reducing maintenance. All moving packings have longer service when held tightly in position with a reasonably high intake pressure (350 to 700 bars). The high pressure fluid is generally a mixture of water and ethylene glycol, fed to the intake manifold by a pump, not shown in Figure 3.
Originally manganin cells were supplied to measure the high pressure, but the autofrettage service proved too demanding for what is essentially a laboratory instrument. It was replaced in this service by the rugged bulk modulus cell.20, 21
For obvious safety reasons, the pumping system is remotely operated at the isolated control panel. The typical operation is semi-automatic, the operator sets the desired autofrettage pressure on the high pressure recorder or digital indicator. If it is desired to hold the high pressure for a predetermined time, the operator sets that on the timer. When the desired pressure is reached it is held for the preset time and then is automatically released.
Essentially the same operation and circuitry with minor variations are applied to large single-acting autofrettage units that are driven by oil hydraulic pumps.
The smallest 14 Kb units are panel-mounted with a small sized intensifier driven by air operated oil pumps. The controls are manual but the circuitry provides for making multiple strokes. These units are used for many purposes, from pressure testing and small scale autofrettage of components, to general research purposes, and for deadweight tester panels, etc. Variants of these units are used as the last stage in gas compressors to 14 Kb. These units have been widely used and have been made in quantity since 1948.
The cyclical pressure is generated with a single acting intensifier but without a discharge check valve. The driving pressure is controlled by the action of the servo-valve on a pilot actuated relief valve, varying the driving pressure according to commands of the electronic control box. It compares the output of the bulk modulus cell with the program in the controller and adjusts itself accordingly. A simple modification of the above allows the system to follow computer commands.
Systems have been built to accommodate specimens of about 50 cc void space to those of about 12 liters, and in pressure ranges from 70 bars minimum varying to 14 Kb maximum. Speeds of cycling up to 60 cycles per minute in small specimens have been reached. On larger equipment, the problems of the inertia of the moving part of the intensifier becomes an obstacle to fast cycle rates, but if the wave shape is not critical, and using relatively small specimens of minimal void space, 19 strokes per minute have been achieved.
It must be remembered that fatigue data at the "mortal" end of the life cycle relation is being studied where failure occurs in 103 to 105 cycles, and the specimens are full scale in size. A different approach is required for isolating endurance limit, where the specimens are smaller, the pressure i much lower, and faster cyclic rates are required because so many more cycles are required to establish endurance limits.
The container method3 lends itself to high rates of production; the tube may be fully autofrettaged or not, as needed. Lower strength alloys may be used but when they are, a more complete autofrettage is usually needed. The containers themselves offer a first line of defense in a catastrophic tube failure. When the tube and containers are properly proportioned, the tube is free to shrink longitudinally inhibited only by the friction of the packings. The tube and container profiles are proportioned to provide the desired degree of autofrettage. Once the containers are made, the amount of autofrettage may be changed with alteration of the tube profile only. Since the tube is finished machined after autofrettage, the designer has considerable latitude in designing the before autofrettage tube proportions to produce the desired amount of autofrettage.
A bore mandrel is designed to carry the packing thrust in tension and the transverse load of two-dimensional hydrostatic pressure. The mandrel will increase in length elastically, and the tube will shrink longitudinally permanently during autofrettage. Allowances for these dimensional changes must be incorporated in the design. In the case of gun tubes, the tapered, usually discontinuous, outer surface can be larger in bore diameter than the rest of the gun tube to diminish the amount of machining in the powder chamber end after autofrettage. This makes it easier to remove both muzzle and breech packings through the breech end. It also reduces the breech end wall ratio thus reducing the difference in the pressure required to autofrettage the muzzle and breech ends. If the muzzle end, because of the taper of the tube, is not strong enough to carry the pressure needed to process the heaviest part of the tube, additional metal can be left on the muzzle before autofrettage and machined off later.7, 8, 9, 10, 11, 12
When the mandrel is designed with larger proportions at the breech end larger longitudinal stresses are developed in the mandrel, so that it is sometimes necessary to autofrettage in two steps, autofrettaging the heavy end of the tube separately. It must be noted that the larger the longitudinal stress in the mandrel, the lower the maximum autofrettage pressure which can be used with mandrels of the same physical properties due to the nature of yielding in combined stress. Failure of mandrel would be the "pinch-off" type. If the pinch-off is appreciable, the failure accelerates and the mandrel will become two projectiles.
The longitudinal load from the packings can be carried by an external press, as with the container method, or in the case of gun tubes, a container can be used on the muzzle end to limit the bore expansion and thus increase the amount of autofrettage pressure that the muzzle can stand and still limit the bore expansion.
The challenge in the design of the tooling for autofrettage has many parameters including economy. The autofrettage of experimental equipment requires a different approach than tooling for high production rates of a standardized part. Large equipment requires a different approach than does small equipment. Otherwise, tooling is a minor challenge in the engineering use of the theories of elasticity and plasticity, fundamental considerations of the metallurgy of available materials, and their properties in relation to the high pressure environment and practical engineering design.
All of the above types of tooling are used in the autofrettage of pressure vessels although it is common to simply autofrettage the vessel by a predetermined over-pressurizing for partial autofrettage prior to the final machining of the bore. When cross drilling is needed for windows, etc., it should be done prior to autofrettage. When it is desired to autofrettage and pressure test liners prior to shrinking into place, in absence of a press to carry the end thrust, the mandrel method may be used. The container method can also be used, if the quantity of parts to be manufactured justifies the expense. Again, the container offers a first line of defense against a catastrophic failure. Components for high pressure equipment such as tubing, valves, tees, crosses, or elbows, pressure cells for pressure-measuring devices, etc. are autofrettaged simply by pressuring at predetermined pressure sufficiently above service levels to ensure plastic yielding. As in all autofrettage processes, due allowance for the permanent bore expansion is made in the pre-autofrettage sizing to minimize the final machining to size requirement.
Care should be taken to ensure safety to personnel and the environment. The protective steps should be taken after a full consideration of the energy in the system including the energy stored in the high pressure fluids and in the deformation of the pressure-containing equipment.
In the early 40's, the author made some experiments in rifling small arms by driving a swaging button with the male counterpart of the desired rifling through the bore with hydraulic pressure. The final contour of the button, using the best lubricant then available, resulted in exceptionally fine rifling, well within specification. Subsequent firing from a vise showed the smallest scatter that had been observed in tests of that weapon model. Unfortunately, it was considered at that time uneconomical. Incidental to swaging the rifling, the barrels were permanently expanded and thereby autofrettaged. This technique of producing autofrettage was subsequently successfully developed as an alternative method of autofrettage of larger tubes at Watervliet Arsenal under the direction of Dr. Davidson.23 It does not, however, have the virtue of testing the full forging. The expansion is localized by the swaging contour of the button and tends to hide local defects in the surface.
