US3731903A

US3731903A - Ball canister and system for controlling cavitation in liquids

Info

Publication number: US3731903A
Application number: US00095044A
Authority: US
Inventors: A Webb; J Tullis
Original assignee: Individual
Current assignee: FOWLER KNOBBE AND MARTENS
Priority date: 1970-12-04
Filing date: 1970-12-04
Publication date: 1973-05-08
Anticipated expiration: 1990-05-08
Also published as: DE2159963A1; GB1356318A; JPS5210522B1; IL38292A0; DE2159963C3; IL38292A; DE2159963B2; CA966394A

Abstract

Disclosed is a ball canister having a hollow case with an inlet and an outlet, the case having a portion which is frustoconical and is filled with tightly packed balls, with the diameter and length of the ball filled case portion being sized to significantly repress cavitation in a liquid flow system into which the canister is to be inserted, either alone or in conjunction with a control valve. A sizing method is disclosed and a particular construction example is explained for a high pressure water system.

Description

1 May8,1973

United States Patent 1191 Webb et al. v

[54] BALL CANISTER AND SYSTEM FOR FOREIGN PATENTS OR APPLICATIONS CONTROLLING CAVITATION IN LIQUIDS [75] Inventors: Albert A. Webb, Riverside, Calif.;

Great James Paul Tullis, Fort Collins, Colo.

[73] Assignees: Fowler, Knobbe & Martens, Orange,

k am m n a a T KMM C A T R n w w mw B H m A K r a mm 1 w mo mxF Em P yn wmm .mmm H S HAM U Calif.; Paul J. Tullis, Fort Collins, Co1o.; Fred H. Hanson, Riverside; Robert M. Massey, Crestline; Albert A. Webb, Riverside, Calif. part interest to each Dec. 4, 1970 Disclosed is a ball canister having a hollow case with [22] Filed: an inlet and an outlet, the case having a portion which is frustoconical and is filled with tightly packed balls, with the diameter and length of the ball filled case [21] Appl. No.2 95,044

portion being sized to significantly repress cavitation v in a liquid flow system into which the canister is to be inserted, either alone or in conjunction with a control valve. A sizing method is disclosed and a particular construction example is explained for a high pressure water system.

7 N2 1 5 I5 62 ll ,s 1 NH 1 45 I" c "r a .e S 1 m d Ld mm UN 55 .11.

5 Claims, 7 Drawing Figures [5 6] References Cited UNITED STATES PATENTS 2,323,839 Nixon l /1 18 X 1 0 J J J 1 M 0 L 7 w L M H a 1 H& m d a L 2.1.. .7 a W W -0 v 5 0 a 5 ET 4 0 L0 1b H 07 g r a PATENTEDHAY 8191s SHEET 3 BF 3 96 INVENTORS.

ALBEET A. W555 BALL CANHSTER AND SYSTEM FOR CONTROLLING CAVITATION IN LIQUIDS This invention relates to apparatus for controlling or repressing cavitation in a flowing liquid, and has particular reference to a ball canister and further reference to the ball canisters cooperation with a control valve in a liquid flow system.

Cavitation generally refers to the formation, growth and collapse of vapor cavities in liquid as the pressure is reduced to near or below the vapor pressure of the liquid by the flow conditions. The pressure reduction necessary to cause cavitation can originate from accelerating the liquid or from non-streamlined flow. In the latter case, turbulent eddies are created and the pressure reduction at the core of the turbulent eddies can be of sufficient magnitude to produce cavitation even when the average pressure in the liquid stream is considerably higher than the vapor pressure of the liquid.

The onset and severity of cavitation in liquid flow systems is affected by a large number of variables, but generally speaking, whether it be chemicals, oil, gasoline or water, at a selected point within a given system a critical liquid velocity can be determined empirically at which cavitation is insipient at a given pressure. Depending upon the energy available and the degree by which the critical velocity is exceeded, the cavitation may vary from light to heavy and from fine grained to coarse, or the system may go into supercavitation wherein the full stream is vaporized.

Once formed, the vapor cavities move downstream, become unstable and collapse, with the implosion of the cavitation bubbles creating extremely high local pressures which can be on the order of 10 to psi. The growth and collapse cycle of a vapor cavity can be of very short duration, such that cavitation can form and collapse inside a valve, turbine or other valuable structure thereby pitting and eroding the structure which caused it; or, depending on many variables and the degree of cavitation, the cavities may collapse within the stream or downstream against another structure. For example, in the case of super-cavitation, the vaporized zone may extend many feet downstream before collapsing.

The extremely high collapse pressures of cavitation bubbles, depending upon the energy in the system, can constitute an enormous and earth-shaking destructive force wrecking havoc with the hydraulic system and structures while creating sonic and vibratory levels of such magnitude that it is uncomfortable to be in the near vicinity, and at lesser magnitudes can constitute an insidious eroding force on expensive hydraulic equipment, or a useful phenomenon for cleaning, for mixing and homogenizing of liquid, or for the breakdown or synthesis of chemicals. in addition to destructive effects, the presence of cavitation can produce undesirable pressure and flow variations and can reduce the efficiency of turbines, pumps and control valves.

The potential of having destructive cavitation occur in a liquid flow system places a large burden on the design and utility of such systems. In many instances the use of a multiplicity of large and expensive hydraulic control devices is required where otherwise a simple and single device would suffice and, even so, cavitation may be substantially repressed only over a very small useful range of pressures and flow rates, and in some cases may never be sufficiently repressed to prevent serious erosion and replacement problems. It would be ideal if cavitation could be controlled or limited by some simple device of relatively small size and universal application, so that the device is readily adaptable in design to many different types of installations and system requirements, is relatively inexpensive, and can be designed not only into new systems but can be added in existing systems where space is at a premium.

We have discovered that a ball canister can approach this ideal and, either acting alone or in cooperation with a control valve, can effectively control or suppress cavitation under different system requirements while providing a substantial pressure break. While ball canisters or devices resembling ball canisters have been used in gas systems as early as 1917, for example as illustrated in British Pat. No. 112,980 to Briggs et al., and have been used in water systems as early as 1398, for example as illustrated in British Pat. No. 24,148 to Hurst, so far as is known there has been no specific ap plication of ball canisters to the phenomenon of cavitation in flowing liquids, nor has the structure of ball canisters been particularly suitable for this purpose.

In accordance with our discovery, in a given liquid flow system, a ball canister if properly constructed will serve to receive the liquid stream at high pressure and discharge the liquid at a relatively low pressure while controlling or repressing cavitation in the liquid. Broadly speaking, such a canister comprises a case having an inlet and outlet, a plurality of balls filling the case along at least a portion of the length thereof, and means retaining the balls in the case. The balls are small relative to the diameter of the case so that it requires many balls to fill a cross-section of the case, and the case diameter and the case length portion filled by the balls are sized to significantly limit the occurance of cavitation in the liquid stream passing through the ball volume at the maximum designed volume flow rate of the system. The sizing depends upon the system requirements and the results desired.

More specifically, and in accordance with one embodiment of our invention, the canister case includes a tapered portion which gradually increases in diameter in the direction from the inlet toward the outlet, the balls are tightly packed in the tapered portion, and in one or more layers at the upstream end of the ball volume, the balls are welded together. We have found that the tapered construction maximizes the pressure reduction afforded by each layer of balls in relation to critical cavitation velocities of the liquid at the local pressures, thereby greatly reducing both the number of balls and the length of the canister while increasing its efficiency. For maximum efficiency while repressing cavitation, the diameter of the canister case at the upstream end of the packed balls is made sufficiently small to present a liquid velocity at the first layer of packed balls near the incipient cavitation velocity at the maximum designed volume flow rate of the system.

In accordance with our invention, a system for supplying liquid at a relatively low pressure from a high pressure supply while repressing cavitation in the liquid comprises a control valve connected to the high pressure supply and a ball canister connected on the outlet or downstream side of the control valve. The case diameter and the case length portion containing the ball volume are sized to significantly suppress cavitation at the control valve over a range of volume flow rates which includes the designed maximum volume flow rate of the system. In one embodiment of this system, the control valve includes an operator responsive to a hydraulic condition in the system, and means for sensing the hydraulic condition downstream of the packed balls in the canister.

The ball canister of our invention and its cooperation with a control valve is explained in greater detail in the following description made with reference to the ac companying drawings, in which:

FIG. 1 is a side elevation of a high pressure testing installation in a water system, illustrating the ball canister operating in conjunction with a control valve;

FIG. 2 is a graph illustrating cooperation of the ball canister and the control valve at different degrees of opening of the valve;

FIG. 3 is a graph illustrating the empirical determination of cavitation in a given system;

FIG. 4 is a modified longitudinal sectional view of the ball canister proper of FIG. 1;

FIG. 5 is a sectional view taken along line 5-5 of FIG. 4;

FIG. 6 is a fragmentary sectional view taken along line 66 of FIG. 5; and,

FIG. 7 is a sectional view taken along line 7--7 of FIG. 4.

Referring to FIG. 1, there is depicted a test installation in a high pressure water system wherein a four inch steel high pressure supply line 10 is anchored in a concrete slab 12 and runs to a manually operable plug valve 14. A four inch steel discharge or delivery line 16 is anchored in the concrete slab 12 at a distance from the supply line 10, and is connected to a similar plug valve 18, such that various apparatus can be connected between the plug valves 14 and 18 for testing and observation. The plug valves serve to shut down and open up the system and, if desired, help provide resistance in the line for control purposes. As shown, the outlet of the plug valve 14 at the high pressure supply is connected by means of a 4 inch to 2 inch reducer 20 to the inlet 22 of a control valve 24, which is a two inch Roto-Disc valve manufactured by Hitco, of Gardena, Calif, and generally described in U. S. Pat. No. 3,424,200. The outlet 26 of the control valve is connected to the inlet of a ball canister 28, the outlet of which in turn is connected to the inlet of the discharge plug valve 18 by means of a 4 inch steel pipe section 30. Intermediate support is provided by a pair of

adjustable jacks

32, 34 having saddles 36, 38 respectively, shaped to fit under the flanges shown.

As shown in FIG. I, the control valve 24 includes an electric operator 40. While in actual tests on this system a manually actuated electric operator was used, conventional electric or hydraulic operators are available for responding to a hydraulic condition in the system to automatically control the degree of valve opening, and in FIG. 1 there is illustrated a water pressure sensing line 42 running from the operator 40 to a location downstream of the balls in the canister, and in this case actually running to a pressure tap 44 on the bottom of the outlet section of the canister itself. Normally severe cavitation will interfere with a pressure or flow measurement downstream from a control valve,

and the present system offers advantages in this regard by suppressing cavitation at this nearby location.

When the pressure in the liquid is suddenly reduced, dissolved gases diffuse out of solution and collect in bubbles. Although this condition is not related directly to cavitation, it is important from the standpoint of satisfactory operation of a sensing device such as a flow meter located downstream of the canister. Accordingly, we provide a conventional gas relief valve 46 mounted on a gas pressure tap 48 located on top of the canister at a position downstream of the balls for venting such gases.

In the system illustrated, it is significant that the ball canister is located downstream from the control valve in order that the maximum flow control at the valve be realized while suppressing cavitation at the valve. With the canister upstream, the valve would have a very limited capability of controlling flow without being subjected to cavitation at high flow rates; but, with the canister located downstream of the valve, the canister provides sufficient back pressure on the valve such that higher flow velocities through the valve, hence higher pressure drops across the valve, can be realized without the valve being subjected to cavitation.

The distance of location of the canister downstream of the control valve can also be significant. Even aside from the obvious advantages of compactness of the system, convenience, and the desirability of using a short sensing line 42, under certain conditions a proximate location can be advantageous from the point of view of cavitation. From the latter point of view, if no cavitation is ever expected at the valve, the canister can be located at any convenient distance downstream of the valve. However, if some cavitation can be expected under certain flow conditions at certain degrees of valve opening, one would desire that the canister be located far enough downstream that the cavitation collapses before impinging upon the upstream layer of balls in the canister. This distance is dependent upon the valve opening and normally varies between about two to five times the inside diameter of the control valve outlet 26, with the distance S being measured from the downstream end of the throat S0 of the valve to the upstream end of the balls in the canister. For this purpose the throat of a valve is considered to end and its outlet begin at the point where the throat constriction merges with the pipe like portion normally provided as an outlet, characteristic of conventional valves. On the other hand, a relatively flat velocity profile normally requires at least about five diameters. Yet, if very heavy or super-cavitation may be expected under certain flow conditions and valve openings, even on a temporary basis as when a valve is in the process of being opened or closed, the canister should be close to the valve such that full development of the cavitation can not be realized due to the canister balls being interposed in the zone, thereby collapsing the super-cavitation before it can fully form and suppressing vibration. For this condition a desirable distance S is about two internal diameters of the valve outlet. Further, where significant cavitation is expected under certain conditions, and especially when the distance S is within about five diameters, it is desirable that the inlet and the first few upstream layers of balls be constructed of a highly erosion resistant material, such as stainless steel,

with the balls being welded together and to the canister case. We have found that about five normally is the optimum distance, all factors considered.

As can be seen in FIG. 4, the case of the ball canister is of welded steel construction and includes a detachable inlet section 52, a conical or tapered section 54, a cylindrical retainer section 56, and an outlet section 58, with the direction of liquid flow being indicated by the arrows in the figure.

The detachable inlet section 52 has an upstream connecting flange 60, a downstream connecting flange 62, a smooth or streamlined two inch internal diameter neck or inlet 64, and a conically tapered throat 66 which merges smoothly and is streamlined with the downstream end of the neck 64. The throat 66 of the inlet section extends downstream through the connecting flange 62 and terminates at the downstream end 69 of the conical section 54 so as to provide a smooth continuous cone, the diameter of which gradually increases in the downstream direction with the throat 66 defining the truncated apex of the cone.

The throat 66 in the inlet section contains the first or upstream four layers 71 of a volume of relatively small balls 70 which fill the conical section of the case and which are bare or exposed at the upstream end. These first four layers of balls are nested and extend .flat across the case with the center line 72 of the first layer being located in the throat 66 at a position where the inside diameter of the throat D is of a predetermined value, and with the fourth or downstream layer being located approximately flush with the planar flange face 68 defining the downstream end of the throat. The balls in the first four layers are welded or fused to one another and to the case, essentially in point to point contact so that there is no significant restriction of the passageways through the nested ball layers.

In constructing our test canister we used one-half inch diameter cast iron grinding balls, and for the first four layers in the throat of the entrance section, the balls were coated with 0.0005 inches of copper and furnace braised so as to fuse or weld them together and to the wall of the throat. Cast iron grinding balls, while not ideal, are relatively inexpensive and readily available, and are adequate for many applications.

As seen in FIGS. 4 and 7, tightly packed balls nest in polygon patterns and a layer of nested balls will not at all points around its perimeter reside in contact with a circle, thus leaving edge voids, for example as shown at 74 in FIG. 4. Improved performance in the critical first four layers 71 may be expected by first welding the balls together in a larger nested configuration and then turning or milling them down to exactly fit the throat 66, thereby achieving ideal layers with certain partial balls disposed around the perimeter so that no voids are left, such as partial ball [76. If desired, all ball volumes in the canister could be constructed in this manner.

As indicated, the conical case section 56 of the canister matches the conical throat 66 of the inlet section 52 to form a continuous truncated cone. The conical section has a downstream connecting flange 78 to which the upstream flange 62 of the inlet section is detachably bolted, with a gasket 80 disposed between the abutting flange faces to seal them. The conical section 54 expands to a predetermined diameter D, at its downstream end, such that the volume of balls 70 has a length L as measured between the center lines 72, 82 of its first and last ideally nested layers. Given predetermined upstream and downstream diameters D D,,, the angle 0 of the cone is determined by the length L which is in turn determined by the number of nested layers of balls and the diameters of the balls.

The cylindrical retaining section 56 is welded to the downstream end of the conical case section 54, and is filled by a volume of larger retaining balls 84 tightly packed therein and in abutting relationship with the last or downstream layer of the relatively small balls 70. As also seen in FIGS. 5 and 6, the retaining balls 84 include a downstream or terminal layer of balls 86 which forms the base of a pyramidal nest 88 of retaining balls, welded together, centered within the volume of retaining balls 86, and pointing upstream. A retainer plate 90 abuts the terminal layer 86 of retaining balls and has a hexagonal opening 92 centered therein. A plurality of ribs 96 are welded in said opening and extend across said opening in spaced parallel relationship to divide the opening into a plurality of openings. The retaining balls in the terminal layer 86 are tightly nested and fixed in a pattern which includes rows of adjacent balls having their center lines aligned with said ribs, as typically indicated at 96, with the balls in the row abutting the rib in point contact and being welded thereto. This provides solid support for the base layer 86 of the pyramidal stack of retaining balls 88, while providing a minimum of interference with liquid flow by preventing the balls from seating in the openings between the ribs.

As an aid in fabrication and for further solid support, a shoulder plate 98 abuts and is welded to the retainer plate 90 on its upstream side. The shoulder plate 98 has a hexagonal opening 100 therein which is larger than and is aligned with the hexagonal opening 92 in the retainer plate 90, so as to form a peripheral shoulder 102 extending around the periphery of the hexagonal opening 92 on the upstream side of the retainer plate. The terminal layer 86 of retaining balls is nested in a pattern which exactly fits the hexagonal opening 100 in the shoulder plate 98. Thus, in combination the shoulder plate and the retainer plate 90 form a dished grill matching the terminal layer 86 of retaining balls. In fabricating our test canister, we used one and onefourth inch diameter cast iron grinding balls for the retaining balls. The pyramidal nest 88 was formed and welded to the shoulder and retaining plates 90 by coating the balls with 0.0005 inches of copper and furnace braising them to one another and to the shoulder plate and retaining plate.

As can be readily appreciated, the retainer construction must be strong in order to prevent any balls from getting into the liquid flow system downstream, and at the same time should provide a minimum and known interference with liquid flow. The retainer structure shown supports the downstream end of the volume of small balls 70, and provides a method whereby a retaining plate having a plurality of openings can be used without the problem of the small balls seating in or obstructing such openings and altering the flow characteristics, as is the case in conventional canister construction. Being held in position by the retaining balls in point to point contact therewith, the flow through the last layer of small balls is not substantially impeded nor are the balls able to reach an opening in the retainer plate.

The outlet section 58 includes a dished head I04 welded to the downstream end of the cylindrical retainer section 56. A stub outlet 106 of 4 inch inside diameter is welded centrally to the dished head and carries a connecting flange 108 flush with its downstream end.

As best seen in FIG. 4, the canister includes a pressure tap 110 located upstream from the first layer of balls and communicating with the interior of the neck 64 of the inlet section. A similar pressure tap 112 is located immediately downstream of the last layer of relatively small balls. The other pressure tap 44 is located at the bottom of the outlet section 58 and communicates with the interior of the dished head 104, and the air tap 48 is located in the same position but on the opposite side of the canister of what would be the top side of the canister in operation since the air bubbles tend to collect at the top. The pressure taps serve in conjunction with pressure gauges (not shown) to monitor pressures at strategic locations in the canister during operation to evaluate the performance of the canister under different flow conditions; or, as in the case of the downstream pressure outlet 44, they may serve control functions when connected to a hydraulic control device, such as to the control valve operator 40 by the pressure line 42. The pressure tap 44 being located at the bottom also serves as a means of draining the canister. When not in use, the pressure taps and air tap would be plugged, for example as illustrated in FIG. 4.

In constructing the canister loose retaining balls 113 are poured through the open upstream end 69 of the conical section 54 and surround the pyramidal stack 88. The canister case is vibrated or struck with a hammer until the retaining balls become tightly packed and immobile. After the cylindrical section 54 is filled approximately flush with its upstream end, the 1% inch diameter balls 70 are poured into the open end of the conical base portion 54 and the case is again vibrated or struck with a hammer until these balls settle against the larger retaining balls and become tightly packed and immobile. One-half inch diameter balls are added until they fill the conical case portion 54 and reside approximately flush with its open upstream end, after which the detachable inlet section 52 containing the first four layers of welded balls is attached by bolting its downstream flange into the upstream flange of the conical section 54.

As would be expected the relatively large retaining balls 84 and the relatively small balls 70 will not reside entirely in nested layers as shown in FIG. 4 merely for purpose of illustration, except where they are welded together. The small and large balls will settle into some of the void areas shown around the perimeter of the layers in FIG. 4, or the layers will spread to fill some of these voids, and some of the small balls may enter certain adjacent void areas in the volume of larger balls, as illustrated for example at 116, 118. Nevertheless, the balls both in the conical and retaining section will approximate layers, and are tightly packed and immobile.

The relatively small balls filling the conical portion of the canister should have a diameter which is relatively small compared to the diameter of the case so that many balls are required to form a given layer filling a cross-section of the case. Seven nested balls in a layer probably would be sufficient, but in our test canister we used nineteen for the upstream layer at a position where the case diameter D was 2.7 inches. We believe that the balls should be as small as practicable consistent with their not acting as a strainer, such that the passageways between the balls tend to be clogged by deposits from the liquid medium flowing through them, and consistent with keeping the cone angle 0 small.

It is desirable to keep the cone angle 0 relatively small in order to insure streamlined flow through the entrance section of the case without boundary separation at the location where the cone merges with the neck 64. In our test canister, an angle 0 of about l5 r worked fine, but the degree this angle could be increased while maintaining performance is uncertain, and decreasing the size of the balls say from A to A inch diameter would greatly increase the cone angle 0 for the same number of ball layers. Consistent with the above considerations, we have found that the balls should have a minimum diameter of about one-half inch while, of course consistent with case diameter, larger balls could be used thereby increasing the length of the volume of balls in the case accordingly.

While it is possible to mix the size of balls so that the void ratio is gradually decreased toward the upstream portion of the canister, this is not considered a satisfactory substitute for a tapered canister because it would be diffieult to construct and control, its operation would be difficult to predict and the canister would have a tendency to act like a strainer and plug up.

The small balls 70, as well as the retaining balls 84, should be approximately spherical and ideally should be constructed of a very hard material highly resistant to erosion and to corrosion by the liquid medium. While we have used cast iron grinding balls in our tests due to their low cost and ready availability, and while these balls would render acceptable performance in many liquid system applications consistent with replacement costs, better balls are available such as rejected stainless steel ball bearings or ceramic grinding balls such as those available from Coors Porcelain Company of Golden, Colo. for which there are methods of welding and fusing the balls together and to the case without significantly reducing the passageways through the ball areas.

There are of course many types and varieties of liquid flow systems where flow rates or pressure differentials are sufficiently high that our ball canister is useful, and within any given system there are a large number of variables which can affect cavitation. For example, some of these variables are the pipe diameter, the upstream pressure, the downstream pressure, the properties of the liquid medium such as density and viscosity, the ambient temperature and the vapor pressure of the liquid at that temperature, the air pressure, the gas content of the liquid, and generally the size and shape of any obstruction, constriction or bend in the path of liquid flow in the system, for example, the size and shape of a control valve orifice at different degrees of valve opening. Fortunately, for most systems, many of these variables are of known value and remain constant or substantially constant so that a mean or average value of many variables can be used for design purposes. Thus, in a given system into which the canister is to be inserted, and for a given liquid at a location where the environmental conditions are known,

the major variables will be variations in pressure and flow rates and, where a given control valve is used in conjunction with the canister, changes in the degree of valve opening. Normally, the upstream pressure is known and is approximately constant, and system is designed to deliver a maximum volume flow rate at a much lower downstream pressure while controlling or repressing cavitation.

In designing our canister for test in the water system illustrated in FIG. 1, we assumed a maximum volume flow requirement of 800 gallons per minute at an approximately constant upstream pressure P of 660 psig. set by the pressure in the supply line 10 in conjunction with a full open setting of the plug valve 14. It was considered desirable to deliver this water over a range of volume flow rates under the control of the control valve 50 to an approximately constant downstream pressure P of psig. We noted from the basic relationship that liquid velocity is equal to volume flow rate divided by the area of the water stream, that a pipe having a 2 inch internal diameter would result in a stream velocity of about 81.5 feet per second which, while a high velocity, should not result in cavitation caused by the mere acceleration of the water under streamlined flow at the local pressure.

Generally, where the canister is to be used in a system including a control valve upstream of the canister as illustrated in FIG. 1, the next design consideration is to select a control valve all of the components of which are of adequate pressure rating to accommodate the maximum working pressure with a safety factor, and with the valve having a high velocity rating so as to minimize the size of valve required to obtain the maximum flow rate, thereby normally to minimize the cost of the valve and its appurtenances. Many valves are not constructed to operate at a water velocity of 81.5 feet per second. The flow-through" type valve normally is best in this regard. As this expression is used here, flow-through" refers to the type of valve where the stream passes through the valve in a substantially straight path as opposed to a tortious path, and where, when the valve is wide open, it essentially provides no constriction and matches the pipe. Also, the flow-through type valve generally will handle higher liquid velocities without cavitation at a given operating pressure. A good example of such a valve is the Hitco Roto-Disc valve illustrated in FIG. I having a 2 inch internal diameter throat 50 which matches the pipe when wide open.

Another reason for selecting a valve type and size which will provide the desired flow rate at high velocity, is that the liquid velocity issuing from the valve outlet should be near the critical cavitation velocity of the liquid at the operating pressure at the first layer of balls in the canister in order to take full advantage of the maximum energy which can be dissipated by the first layer of balls without cavitation, so that a minimum of reductions or increases in the internal diameter of the inlet section of the canister or other pipe upstream from the first ball layer is required to achieve such velocity.

The next step normally to be employed in design is the determination of the size of balls to be used and the size of the canister portion containing the balls. Ball size has already been discussed; hence, using /2 inch spheres in our test canister, the exercise was to calculate the diameter D,, at the upstream layer of balls, the diameter D,, at the downstream layer of balls, and the length L required to achieve the desired pressure reduction at maximum flow rate.

To calculate the upstream and downstream diameters D D laboratory measurements should be made in order to ascertain the critical cavitation velocity V at different operating pressures. FIG. 3 generally illustrates such a laboratory determination. When a sensor such as an accelerometer is attached to the exterior of an orifice being tested, the sensor output will increase as the liquid flow velocity is increased at constant pressure, the measurement normally being made by measuring the liquid velocity upstream of the orifice at constant upstream pressure with changes in flow velocity being obtained by adjusting the downstream pressure, although the measurement can be made in a reverse fashion at least for purposes of determining incipient cavitation.

As illustrated by the curve in FIG. 3, as velocity is increased at constant pressure, the sensor output will first gradually increase on account of increasing flow noise until a point is reached where the sensor output begins to rise rapidly, this being the point of incipient cavitation with the flow velocity at this point being referred to as the critical cavitation velocity, V Further in crease in the flow velocity will result in a rapidly increasing level of cavitation and finally, if there is sufficient energy in the system, in super-cavitation where, if the cavitation collapses some distance downstream, the sensor output may and frequently does decline somewhat because of its more remote location from the area of collapse. Critical cavitation velocity will vary with pressure, and it is found that the point of incipient cavitation will occur at lesser velocities as the local pressure is reduced.

In the design of our test canister, the critical cavitation velocity of the upstream layer of balls at a local pressure of about 660 psig. was estimated to be about 41 feet per second, based on previous laboratory experience. For the downstream layer of balls at the retainer, the critical cavitation velocity was estimated to be about 7.2 feet per second as measured in the stream issuing from the last layer at an assumed local pressure of 20 psig. Using the mentioned relationship between velocity flow rate and area, it was calculated that the upstream diameter D should be about 2.7 inches and the downstream diameter D, should be about 6.7 inches.

However, as to the downstream diameter D of the balls, it was decided to allow for a safety factor and use a value of 10 inches because of the retainer construction, on the theory that at worst the retainer could be considered to be a device blocking off part of the area of the last layer of small balls. This accounts for the reason that the total open area provided by the grilled opening 92 in the downstream retaining plate was made approximately equal in area to a 6.7 inch diameter circle as originally calculated for the downstream layer of small balls residing at D With the upstream and downstream diameters D D of the cone set at 2.7 inches and 10 inches respectively, we then proceeded to find the length L of the ball volume required to provide the necessary pressure drop by using the following energy equation, with the left side of the equation indicating in psi the energy in the water system upstream less the energy in the system downstream, and the right side indicating in psi the summation of the energies dissipated by the respective layers of balls in the ball volume:

The following values appear in the energy equation:

Q Volume flow rate in gallons per minute.

D The internal diameter at the cone at the centerline of the upstream layer of small balls expressed in inches.

D The internal diameter of the cone at the centerline of the last layer of small balls expressed in inches.

P,, The upstream pressure at the canister inlet in psi.

P The pressure downstream from the canister in psi.

d The diameter of the small balls in inches.

N The number of layers of small balls, equal to L divided by 0.866 d L The distance between the centerlines of the first and last layers of small balls expressed in inches.

K,, What is assumed as a ball constant for the onehalf inch diameter balls, having a value of 0.01 57.

Substituting the values Q 800 gallons per minute, D,, 2.7 inches, D inches, P 660 psig, and P psig. into the left hand side, the energy equation indicates a pressure drop of 653.5 psi across the volume of small balls in the canister. The summation was solved as an iteration by assuming various values for L which resulted in an integer value for N, the number of ball layers, until a correct value was found which caused the summation of pressure drops across the successive layers to add up to the 653.5 psi pressure drop calculated across the ball volume indicated at the left side of the equation. As it turns out, this is approximately true when L equals 13 inches such that N equals 30 layers, with the calculation indicating a pressure drop across the first layer of balls of about 189 psi and rapidly decreasing pressure drops across subsequent layers until the drop across the last or thirtieth layer is on the order of 1 psi.

The test canister was constructed in accordance with the above calculations and it worked remarkably well considering that we ignored the pressure drop across the retainer and outlet section of the canister, which should be added to the right side of the energy equation to valuate whole canister operation. As measured this pressure drop turned out to be about psi. The performance seemed remarkable for other reasons as well, such as the matter of assuming critical cavitation velocities from previous laboratory experience rather than actually measuring them at the laboratory. P and P were both assumed as constant at about 660 psig. with the control valve 100 percent open. P actually varied from about 720 psig. to about 674 psig. as the valve was opened from 5 to 100 percent open, with P varying accordingly from about 20 psig. to 24 psig. Further, our value for the ball constant K,, was determined from the measurements made on 1% inch cast iron grinding balls tightly packed in a 4 inch diameter cylinder, and we are uncertain as to whether the ball constant K might vary from layer to layer depending upon the diameter of the balls and the diameter of the canister at the layer. Of course, these factors can be measured in the laboratory although we have not yet done so ourselves.

In calculating K, we used the following ball constant equation in relation to measurements of the flow through a cylindrical canister tightly packed with r inch diameter balls;

b= P 0.866 a, D/L

In the ball constant equation P the upstream pressure was 539 psig, P the downstream pressure was 51 psig, d the ball diameter was one-half inch, D the case diameter was 4 inches, L the length of the ball volume was 17 inches and Q the volume flow rate was 450 gallons per minute.

As can be seen from the energy equation calculations where the pressure drop across the last or ten inch diameter layer of small balls is on the order of 1 psi, if a cylindrical as opposed to conical canister case were used, the cylindrical canister case might require between 600 and 700 such layers of balls to achieve the same pressure drop, as compared to the mere 30 layers of balls required by the conical canister, when both operate at the same cavitation intensity.

FIG. 2 is a graph illustrating P as actually measured at different degrees of valve opening. P was measured at the longitudinal position indicated in FIG. 1 and was a measure of the downstream pressure on the valve or the upstream pressure on the canister, both being roughly the same. Also, the measured volume flow rates in gallons per minute corresponding to various points on the curve are shown in the figure. The pressure differential across the valve (P P as well as the pressure differential across the canister (P P are only approximate from the figure because the mentioned minor variations in the upstream and downstream pressures are not taken into account.

The system illustrated in FIG. 1 was instrumented with accelerometers (not shown), and with pressure gauges (not shown) at various locations including the longitudinal positions indicated at P P and P Flow measurements were made downstream of the canister.

In the tests, the accelerometer readings remained in the vicinity of l0,000 to 20,000 inches/sec. peak to peak, except in the 20 to 25 percent valve opening range where some readings approached 50,000 to l00,000 inches/see, which appears to be about the percentage valve opening where about half the pressure drop in the system was across the canister, and indicates the possible presence of some fine-grained cavitation at the valve. Also, large quantities of air were discharged through the orifice 48.

By contrast, the system was otherwise a supercavitating liquid flow system. Thus, in the same system with the canister removed, and with the upstream pressure at about 700 psig, the pressure measured at the valve outlet varied from a vacuum of 9 inches of mercury at a 15 percent valve opening to a vacuum of 17 inches of mercury at a 40 percent valve opening, with the flow rates varying from 445 gallons per minute to 937 gallons per minute at these openings. Without the canister, the vibration and noise level was so great in the vicinity that it was not considered safe or desirable to maintain the valve open at 40 percent or to open it further.

Even with the canister installed, some minor cavitation also was occurring at the valve in the opening range of percent and below; however, the test structure was of such heavy steel construction and the flow noise was so great that whatever cavitation there might have been was not at a sufficient energy level to be apparent in the measurements.

Also, it is apparent from FIG. 2 that if the maximum desired flow rate were achieved at substantially less than a 100 percent valve opening, such as say 30 percent open, and the canister were sized accordingly, the valve could then be opened 100 percent to cause a sufficiently excessive flow rate and pressure drop across the canister to cause cavitation in the canister in order to clean it of any scale or deposits. This appears to be practical because, as can be seen from the flow scale in FIG. 2, most flow control by a valve occurs at valve openings below 40 percent.

We claim:

1. In a liquid flow system, a system for supplying the liquid at a relatively low pressure from a high pressure supply while repressing cavitation in the liquid, comprising a control valve having an operator for opening and closing the valve to varying degrees to vary the liquid volume flow rate through the valve, means coupling the inlet of the control valve to the high pressure supply, a canister, and means coupling the inlet of the canister to the outlet of the control valve, said canister including a case, a volume of balls filling the case along a portion of the length thereof, and means retaining the balls in the case, with the diameter and length of said ball filled case portion being sized to significantly repress cavitation at the control valve over a range of volume flow rates which includes the designed maximum flow rate of the system.

2. In a liquid flow system, a system for supplying the liquid at a relatively low pressure from a high pressure supply while repressing cavitation in the liquid, comprising a control valve having an operator for opening and closing the valve to varying degrees to vary the liquid volume flow rate through the valve, means coupling the inlet of the control valve to the high pressure supply, a canister, and means coupling the inlet of the canister to the outlet of the control valve, said canister including a case, a volume of balls filling the case along a portion of the length thereof, and means retaining the balls in the case, with the diameter and length of said ball filled case portion being sized to significantly repress cavitation at the control valve over a range of volume flow rates which includes the designed maximum flow rate of the system, with the diameter of the canister case at the upstream end of the ball volume being sufficiently small to present a liquid velocity near the incipient cavitation velocity at that location at the maximum designed volume flow rate of the system, and wherein the canister case increases in diameter going downstream.

3. A system for supplying liquid at a relatively low pressure from a high pressure supply while repressing cavitation in the liquid, comprising a control valve, means for coupling the inlet of the control valve to the high pressure supply, a canister, and means coupling the inlet of the canister to the outlet of the control valve, said canister including a case, a volume of balls filling the case along a portion of the length thereof, and means retaining the balls in the case, with the diameter and length of said ball filled case portion being sized to significantly repress cavitation at the control valve over a range of volume flow rates which includes the designed maximum flow rate of the system, with the diameter of the canister case at the upstream end of the ball volume being sufficiently small to present a liquid velocity near the incipient cavitation velocity at that location at the maximum designed volume flow rate of the system, and wherein the canister case increases in diameter going downstream, and wherein the system is sized to accommodate the maximum designed volume flow rate when the valve is only partially open, whereby the valve may be opened further to deliberately produce cavitation in the canister to clean the balls.

4. The system of claim 2 wherein the means for retaining the balls in the case has a plurality of openings therein the total area of which openings is at least about equal to the cross sectional area required for the downstream end of said ball volume to operate at or near incipient cavitation at said relatively low pressure and at the maximum designed flow rate of the system.

5. A system for supplying liquid at a relatively low pressure from a high pressure supply while repressing cavitation in the liquid, comprising a control valve, means for coupling the inlet of the control valve to the high pressure supply, a canister, and means coupling the inlet of the canister to the outlet of the control valve, said canister including a case, a volume of balls filling the case along a portion of the length thereof, and means retaining the balls in the case, with the diameter and length of said ball filled case portion being sized to significantly repress cavitation at the control valve over a range of volume flow rates which includes the designed maximum flow rate of the system, and wherein the upstream end of the ball volume is disposed downstream of the throat of the control valve a distance which is equal to about two to five times the inside diameter of the control valve outlet.

3%??? UNITED STATES PATENT oF zcn' CERTIFICATE OF CORRECTION ?c1 ;;2nt K0; 3,731,903 I Dated Mav 8. 1973 Inventor) Albert A. Webb and James Paul T11111 ppears in the above-identified patent It is certified that error a hereby corrected as shown below:

and that said Letters Patent are Col. 5 line 2 it'ive normally" should be -.,-five diameters normally-- Col. 6 line 67 "an" should be -any-- Col. 11, line 12' "(D D should be .--(D D Signed and sealed this 8th day of January 197L|..

SEAL) Attest: I v

EDWARD M.FLETGHER,JR. RENE D TEGTMEYER Atte sting Officer Acting Commissioner of Patents