US1097298A - Means for preventing surging in centrifugal compressors. - Google Patents

Description

J. G. GALLAN.

MEANS FOR PREVENTING SURGING IN 'GENTRIFUGAL OOMPRESSORS'.

APPLICATION FILED SEPT. 28, 1912.

1,097,298. Patented May 19, 191

2 SHEETS-SHEET 1.

' Fig. I.

O uuuuuu O m I IO O l4 0 Mug 0 I I i &

l ,iif' L--.

Witnesses: 4 5 Inventor.-

John Gcal larfi,

Q J.G.OALLAN. V MEANS FOR PREVENTING SURGING IN CENTRIFUGAL OOMPRESSORS. APPLICATION FILED SEPT. 28, 1912.

1,097,29 Ptented May 19, 1914.

\ 2 sums-51mm 2. F .5. I F1g 1.

Fig.7. Fig.8.

Irkkarnxm: W John G.Ca||ar1, 7 @W QM Boston, in the county of Suffolk, State of.

UNITED STATES PATENT OFFICE,

JOHN G. CALLAN OF BOSTUN, MASSACHUSETTS, ASSIGNOR TO GENERAL ELEGTRIG COMPANY, A CORYQRATION OF NEW YORK.

Specification of Letters Patent.

Patented May 19, 1914,

Applleationfiled se temt ras, 912. Serial No. 722,929.

To all whom it may concern i Be it known that I, JOHN G. GAnLAN, a citizen of the United States, residing at Massachusetts, have invented. certain new and useful Improvements 1n Means for Pre venting Surging in Centrifugal Compres-' sors, of which the following is a specifica- .tion.

This invention relates to centrifugal compressors for elastic fluids, and its object is to improve their construction and operation, whereby the surging, pulsating orp of said valve. The present invention deals with the matter further and points .out certain' fundamental conditions lieve are novel and highly valuable at the present state of the art.

In the accompanying drawings, Figure 1 is a sectional elevation of a portion of a' centrifugal air compressor showing a regulator embodying my invention; Fig. 2 IS a ufiingg efl'ect therein under certain load conditions will beprevented. In a prior.application, filed March 16th, 1910, Serial No. 549,748, I have discussed at some length the reasons* why this objectionable efi'ect occurs at light loads, and have disclosed one mode of cor recting it, namely, by inserting in the intake pipe of .the compressorv a throttle valve-. which follows the fluctuations in volume Q and pressure of the fluid, a damping device being provided to check violent'movements and from these derives certain remedies which I bevanes. These are designed to function best at a certain volumetric output and their efliciency falls off somewhat .With any departure from this, and particularly with re-. duced flow. At constant speed, which ought to generate constant pressure, this falling ofi of discharge vane efliciency exhibits tself by a drop of pressure when the volumetric output is reduced below a'jcertain point. This drop may be gradual or it may nearly all occur at once (break-down ,point) or it may assume characteristics between these extremes.

Fig. 3 shows a curve with a well marked break-down point ;Fig. 4 one with this sharpFbreak-down' point eliminated and the pressure reduction at partial load distributed; Fig. 5 shows what happens to the break-down point when the compressor speed is progressively reduced, thusreducmg the pressure. In this case any given condition, as for example the breakdown point or for another example the portion .of the curve which should be selected; as representing normal load conditions, will find its locus on a parabola of Which'the equation s .=-.-a constant.

on the unstable part of its characteristic.

Under such conditions the true remedy'for instability lies in throttling at a point nearthe machine, preferably onthe inlet but. alternatively on the outlet. The. re'asonforthis is that a ressure-volume curve fordischarge throng athrottled orifice necessarily slopes downward as shown by the curve ,below the axis ofV in Fig. 6, and acompressor characteristic plotted upward from thiscurve as a base line will partake of this downward'slope as shown in Fig. 6. That figure shows the result of throttling throu h a constant orifice; it will be seen that t e steadying action is extremely small at the early stages and becomes progressively greater as the flow increases, the action being approximately proportional to the square of the volume passing. The characteristic is seen to be so modified that the unstable portion becomes shorter than in Fig. 4 and the stable portion correspondingly moves back farther into the region of partial load, On the other hand the drop in pressure and consequent loss in the stable portion is highly objectionable.

As we shall see later,a simple throttled opening is not the best steadying device, but a study of it leads to appreciation of funda mental conditions. Therefore, we may continue it a little further. To determine the amount of throttling required 'for a given steadying effect we should appreciate that if there is just enough throttling to tip the actual working curve from an upward slope toa downward one at a point where the machine is to run, this will change its operation from unstable to stable conditions at that point. The amount of throttling required to do this obviously depends upon two things :first, inclination or tangent of the inherent or generated characteristic at the point in question, and second, the amount of flow at the point in question, since throttling is a function of volumetric flow. From these data we can determine an opening which will steady the flow and eliminate surging or pulsation at any one given point. It will beat once apparent, however, that the size of opening right for one point is.

probably too small for all points of larger flow and probably too large for points of smaller fiow. The only exception to this would be in the case of a characteristic with marked break-down point at which point exceptionally vigorous throttling would be called for and hence requires in many cases an opening actually smaller than that required at points of lesser flow.

It is evident that any throttling whatever produces inefficiency, and hence with the ideal automatic throttle we must apply at any point only just so much throttling as is necessary to insure stability, and where none is necessary must apply none. It has been thought that a check valve or relief valve floating on the incoming or emerging air stream and applying thereto an amount of throttling determined by the weight would efi'ectthe necessarysteadying, and in point of fact with commercial degrees of valve friction and inertia some steadying action is actually thus brought about. As indicated in Fig. 7, however, the action is exceedingly crude, imperfect and ineffectual, since it merely changes the position of the axis of ,V, or speaking in physical conception, merely imposes a constant pressure drop on the air while in point of fact the essence of the lief valve flouting on the inlet or outlet air stream does not fulfil the conditions of a suitable-and economical stcadying device.

F ig. S exemplifies what is really wanted. Here the full line gives the generated characteristic of the compressor as measured at" the outlet. An ideal throttling action may 'be assumed which at any given volume superposes upon this characteristic a downwardly sloping (throttled) curve exactly neutralizing at that pointthe upward slope and leaving the net characteristic perfectly horizontal. Obviously we may over-neutralize to produce extreme stability or under-neutralize to allow for line drop as in an over compounded electric generator. For simplicity, we will consider exact neutralization alone. To realize exactly or approximately the required conditions necessitates two things. First, an arrangement or structure such that at any given flow the automatic throttle will be closed down to an orifice giving the required horizontal characteristic; this for any one point may be accomplished by the weighted check valve or equivalent. Second, a device which will hold this orifice at approximately this value through the period of one cycle of variation under the existing conditions of line resistance and capacity; this-can beaccomplished for any one periodicity by a. dash-pot set to permit valve vmotion at not more than a certain rate.

\Vith the combination thus formed and the exact proportioning and adjustment thus defined, we have, for any given mean volumetric delivery, a condition of throttling such that any momentary volumetric increase causes at the throttle exactly as much net pressure drop as it-causes net pressure rise at the discharge vanes and converscly momentary volumetric decrease causes through decreased throttling an increase of net discharge pressure exactly coinpensating for the decrease of that pressure due to impaired discharge vane efliciency. If this condition can be. realized at each point of the curve, We shall have established an ideal steadying device since at no point will the throttling and consequent impairment of efliciency be greater than is absolutely required, and on the other hand the operation will be perfectly steady through-.

blower characteristic to have been-determined and plot-ted so that the slope many points as are required along the curve.

From these a cam (or equivalent mechanism such as a wrist plate linkage) will be laid out such that when the valve is opened to a given orifice the weight applied to it will be that shown by the computation in every case. Thus when a given volume is passing, the valve will assume a position such that thethrottled orifice which it fixes is exactly that required for/ the steadying action necessary at that point. It then is only necessary to maintain this valve in position by a dashpot still enough so that it will not depart by more than a permissible amount from the theoretical orifice during the period which would be occupied by one cycle of rise and fall of pressure as described in my previous patent application. Under these conditions disturbances in the blast line or elsewhere cannot start the cyclic surging which would otherwise take place. Fig. 1 illustrates the application of these principles in practical form. The intake 1 of the air compressor 2 is provided with a throttle valve 3 mounted on a stem 4 which has at its lower end a dash-pot The valve is a disk located in an inverted trusto-conical sect-ion 6 of the intake, so that as it rises and falls it enlarges and diminishes the annular opening between its periphery and the wall of the section 6. The stem 4 passes up through the top of the elbow 7 in the intake pipe, and carries a cam 8, whose back 9 is parallel with the path of movement of the stem, and bears against anti-friction guide rollers 10v mounted on a standard 11. A lever having two arms 12, 13 at right angles is pivoted to the upper end of the standard. The-vertical arm 12 carries a roller 14 bearing on the cam. the horizontal arm 13. lVith an increase in the volume of air passing through the intake, the valve rises and the cam tilts the lever, throwing the arm 13 upward and thereby shortening the eife'ctive distance of the weight 15 from the pivot, and lessening the pressure of the roller 14 against the cam,

A weight 15 is adjustable alongso that the valve is more sensitive to slight changes in volume when working in the upper part of the section 6. The cam shown happens to have an ogee curve, but the exact curve is determined by the'calculation hereinbefore mentioned.

In accordance with the provisions of the patent statutes, I have described the principleof operation of my invention, together with the apparatus which I now consider to represent the bestembodiment thereof; but I desire to have it understood that the apparatus shown is only illustrative, and that the invention can be carried out by other means.

What I claim as new and desire to secure by Letters Patent of the United States, is 1 1. In a centrifugal compressor, means for producing stability of operation which comprises a conduit through which the fluid handled by the compressor, flows, a valve movably mounted in the conduit, a weight that tends to oppose the movement of the valve in one direction, a device connected to the valve to move therewith, said device acting to vary the effect of the weight on the valve, and means for damping the movement of the valve.

2. In a centrifugal compressor, means for producing stability of operation which comprises a conduit through which the fluid handled by the compressor flows, a valve movably mounted in the conduit that has a tendency to close and is opened by the action of the fluid, a weight that opposes the opening of the valve, a cam connected to the valve to move therewith, said cam acting to vary the effect of the weight on the valve, and a dash-pot for damping the movement of the valve.

3. In a centrifugal compressor, means for insuring stability of operation which comprises a conduit through which the fluid handled by the compressor flows, a valve arranged in the conduit that has a tendency to close and is opened by the action of the flowing fluid, a stem for the valve, a movable weight .that opposes the opening of the valve, a cam carried by the stem which acts to vary the eltect ofthe weight on the valve, and a daslrpot for damping the movement of the valve that has one of its members connected to the valve stem.

4.. In a centrifugal compressor, means for preventing surging which comprises an intake pipe having an inverted frusto-conical section, a throttle valve movable in said sec- 7 tion, a weighted lever resisting the opening movement of said valve, and a cam on the stem of said valve adapted to actuate said lever and lessen the retarding effect of said weightasthethrottle opens.

In a centrifugal compressor, means for preventing surging, which comprises an mdetermined from the characteristic curve of 0 said compressor.

In witness whereof, I have hereunto set my hand September, 1912.

JOHN G. GALLAN.

Witnesses:

FLORENCE S. HARRIS, PERRY BARKER.

Images