US2949224A - Supersonic centripetal compressor - Google Patents

Description

Aug. 16, 1960 v. H. PAVLECKA 2,949,224

sUPERsoNIc CENTRIPETAL COMPRESSOR Filed Aug. 19. 1955 7 Sheets-Sheet l Aug. 16, 1960 v. H. PAvLEcKA sum-:asoma CENTRIPETAL COMPRESSOR 7 Sheets-Sheet 2' Filed Aug. 19. 1955 INVENTOR W40/Mae JZ P41/654 ef; L

rra/e/vfx Aug. 16, 1960 v.|vl.l=.1\vl.le:cK/=\ 2,949,224

SUPERSONIC CEMRIPE'ML COMPRESSOR Filed Aug. 19, 1955 7 Sheets-Sheet 3 EIM?.

IN VEN TOR. W40/me P41/foca BY 4MM/m4 #fraz/vnf.

Aug. 16, 1960 Filed Aug. 19, 1955 V. H. PAVLECKA SUPERSONIC CENTRIFETAL COMPRESSOR 7 Sheets-Sheet 4 INVENTOR. W40/Ame .PAU/z cr/9 rrafe/vfsa Aug. 16, 1960 l v. H. PAvLEcKA -suPERsoNIc CENTRIPETAL COMPRESSOR '7 Sheets-Sheet 5 Filed Aug. 19, 1955 Hc A #vor A Aug. 16, 1960 v. H. PAVLECKA 2,949,224

SUPERSONIC CENTRIPETAL COMPRESSOR I l '7 Shets-Sheet 6A Filed Aug. 19, 1955 5.. a.. 24M 5 ma@ MWWM n .m4/97. n.; PMM F Hf f2 WW m Aug. v16, 1960 v. H. PAvLEcKA suPERsoNIc CENTRIPETAL COMPRESSOR '7 Sheets-Sheet 7 Filed Aug. 19, 1955 .NN .SNN

, ...QSE

SUPERSONIC CENTRIPETAL COMPRESSOR Vladimir H. Pavlecka, Pacific Palisades, Calif., assgnor to American Machine & Foundry Company, a corporation of New Jersey Filed Aug. 19, 1955, Ser. No. '529,504

25 Claims. (Cl. 230-124) This invention relates to dynamic compressors of centripetal flow type, and more particularly to the supersonic centri-petal ow compressors.

This application for patent is a continuation-in-part of the parent application for patent, #217,347, led March 24, 1951, entitled Gas Turbine Power Plant, now U.S. Patent No. 2,804,747, issued September 3, 1957, which discloses and claims the entire gas turbine power plant. The supersonic compressor is disclosed and claimed in this application. The supersonic compressors having more than two stages and including a vector adjusting stage, originally disclosed in the parent case, is also disclosed in the continuation-impart case Serial Number 514,001, entitled Methods of Compressing Fluids With Centripetal Compressors.

1t is an object of this invention to provide a novel centripetal supersonic dynamic compressor in which compression is accomplished by means of oblique, reflected,V

and normal shocks singly or in any combination with each other, as well as subsequent subsonic diffusion, the oblique shock being produced by a leading at surface of a blade.

Still another object of this invention is to provide a novel centn'petal compressor having a prerotation stage and a plurality of contra-rotatable rotor stages, all rotatable compressor stages being supersonic on the input side and subsonic on the output side.

Still' another object of this invention is to provide means for synchronizing the rotation of the contra-rotatable compressor rotors of a supersonic cent'ripetal compresser for maintaining proper velocity vector relationships for compressed air owing between the compressor stages.

It is also an object of this invention to provide new methods of compressing fluids by accelerating the uid either to a supersonic or subsonic velocity and then compressing it by producing a single or a plurality of supersonic compression shocks by means of wedge-shaped blades, the lagging surface of the blades being parallel to the relative velocity of the fluid at its entry into the compression stages, the methods being applicable to the centripetal, centrifugal and axial methods of compressing gases.

Yet `another object of this invention is to provide a supersonic centripetal compressor in `which a stationary contra-prerotation stage is provided for accelerating a liuid to be compressed either to a subsonic or supersonic velocity at the exit from the contra-prerotation stage and then subjecting the accelerated iluid to a plurality of supersonic acoustic shocks in the succeeding contra-rotatable supersonic compression stages.

The novel features which are believed to be characteristic of the invention, both as to its organization and method of operation, together with further objects and advantages thereof, will be better understood from the following description taken in connection with the accompanying drawings in which several embodiments of the invention are illustrated as examples of the invention.

States Patent C) f drical member located on the other side of line A---A,`

It is to be expressly understood however, that the drawings are for the purpose of illustration and description" only, and are not intended as a definition of the elements of the invention. vReferring to the drawings:

Figure l is an axial vertical sectional view of a twostage centripetal supersonic compressor. Y

Figure 2 is a transverse cross-sectional view taken along linel 2 2, Fig. 1, of the contra-prerotation stage and two compression stages of the centripetal Vsupersonic compressor.

Figures 3, 4, and 5 are similar cross-sectional views of the modified versions of the centripetal supersonicV compressors disclosed in Figs. l and 2.

Figure 6 is an enlarged transverse cross-sectional view of compressor blades and blade reinforcing ribs.

Figure 7 is a vector diagram of tiuid velocities in a l two-stage centripetal supersonic compressor.

Figure 8 is the entropy-head diagram of the compres- Y sion stages and particularly, leading portions of the blades and of the flow channels formed by the opposing surfaces of the adjacent compressor blades. Y

Figure 15 is a side cross-sectional view of a leading edge of a compressor blade and supersonic air iiow over such portion of the blade.

Figure 16 is a perspective view of a iiow channel Vof l the iirst, rotatable compression stage.

Referring to Fig. 1, itdiscloses the axial vertical sectional view of the compressor portion of the gas turbine i power plant and of a portion of the rotatable toroidal combustion heat generator which interconnects the sec-v ond compressor stage with the first turbine stage not illustrated in Fig. 1. For a more detailed description and illustration of the entire power plant and especially of the turbine half of the power plant, reference is made" to the previous mentioned parent application #217,347,

now Patent 2,804,747, illustrating and describing the entire gas turbine power plant. The power plant is mounted on a composite frame consisting of an air intake duct portion 10, a mid-portion member 12, a front member 14, and a rear member comparable to the front member 14, whichv houses and supports the compressor and turbine portions of the power plant not illustrated in Fig. 1.

a unitary frame.

of bolts 26. The ring-shaped frame member 25 is closed off by means of a slanted flat ring 27. The compressor# turbine structure includes an inner shaft 31 mounted on two radial-axial bearings such as a bearing 34, only one bearing being visible in Fig. 1.

of the powervplant whichgdoes not appear in Fig.A 1.

A similar bearing 42 is used for mounting shaft 43` which l supports the outer side disc 54 and theV rst stage 64 of the compressor and the Vsecond stage of the turbine.

The first stage 64 of the compressor constitutes a part of the outer cylinder including ring 59 Iand a central cylindrical member 44 connected to the identicalrcylin- Patented Aug. 16,.

l The several frame members are peripherally bolted together at 17, 18, 26, and 20 so as to constitute The front end of the plant is cowled in a front cowl 21. The front member 14 is extended f forward by having a'ring-shaped member 25 bolted to it by means of a peripheral joint held together by'a plurality Y The second bearingis mounted on the extension of shaft 31 on the turbine sideY which bisects the power plant into two halves through the center of the toroidal combustion heat generator 35. The cylindrical member 44 is fastened to the identical cylindrical member on the nturbine side of the power plant by means of a plurality of bolts 62. Ring 55 is-proyided with a plurality of cylindrical recesses 56, which house a corresponding plurality of cylindrical pins 57, forming a sliding fit with the cylindrical recesses. The opposite ends of the pins 57 form radially sliding fit with the corresponding radial slots in the ring 58, the entire assembly acting as a torque-transmitting means and also as an expansion joint. Ring 58 is bolted to a substantially rectangular ring 59 which constitutes the outer hoop ring of the first compression stage 64 of the compressor. The inner side-surface ofthe hoop ring 59 is used for mounting the blades of the compressor all the way around the periphery of this ring. Rings 58 and 59 are made of titanium alloy which permits higher peripheral velocities than those obtained with steel rings. Side discs 54 and ring 55 on one side, and ring 58 and hoop ring 59 on the. other side, are interconnected by a pivoted ring 60 having a cross section identical to that of a dumbbell, the two cylindrical surfaces of ring 60 forming a sliding seat within the respective seats. When the rings 58 and 59 expand radially, their expansion exceeds radial expansion of side discs 54 and this differential expansion is transmitted and absorbed primarily by the elastic deformation of ring 60. The right sides, as viewed in Fig. 1, of the blades 64 of the first compression stage are welded to the left side-surface of an outer torque-transmitting hollow cylinder 44 composed of two rings forming a central bolted joint 62. The right side-surface of the hollow cylinder, including member 44, terminates in the plurality of blades in the second stage of the radial eentrifugal flow tubine, which is not illustrated in the figure.

The left portion of shaft 31 is provided with an integral side disc 65, which forms a similar expansion joint with a hoop ring 66 provided with a fluid-accelerating surface 67 located in the exit channel of the compressor.

Shaft 31 is mounted in two bearings, such as a bearing 34, made preferably of bearing aluminum alloy. The front end bearing 34 consists of a steel ring 36, the flat inner surface 37 of which forms a sliding axial lit with shaft 31 but is keyed to the shaft to prevent its rotation around the shaft. The outer surface 38 of ring 36 has a spherical surface constituting the sliding surface for the similarly shaped sliding surface of bearing 34. The frame member 25 is provided with an oil gallery 39 connected to a source of lubricating oil, this gallery communicating with oil grooves 40 which run parallel to shaft 31. Means are also provided for sealing the entire periphery of bearing 34 on its sides to prevent excessive sideway oil leakage from the grooves 40. The position of the bearing ring 36 is fixed on shaft 31 by `an end nut 41 so that the engagement of the two spherical surfaces fixes the longitudinal position of shaft 31 with respect to frame member 25. Similar bearings 42 are used for mounting shaft 43 by means of a spherical ring 44, ring 44 being provided with va shoulder for longitudinal location of shaft 43. This longitudinal location is obtained through a ring gear 46 which is held in fixed position by a ring nut 47 and circumferential splines between shaft 43 and the ring gear 46. Ring gear 46 is geared to a plurality of circumferentially positioned pinions 48 which revolve around their respective shafts such as shaft 53 mounted in a ring resembling somewhat a squirrel cage solidly bolted to a frame member 14. The squirrel cage ring 49 is provided with four large openings which house four synchronizing gear assemblies including gear ring 46, pinion 48, gear 50, and a ring gear 51 provided with gear teeth on its inner surface. The ring gear 51 is splined to shaft 31 and is held in a fixed axial position by a ring nut 52 threaded to a shaft 31. The synchronization of the two shafts is obtained by` transmitting any differential torque between the shafts 31 yand 43 through the synchronizing gears 46, 48, 50, and 51. Pinion 48 and gear 50 constitute one solid piece mounted on the same shaft 53. Since shafts 53 prevent the rotation of the synchronizing pinions 48 and 50 around the two gear rings 46 and 51, the gear assembly will act as a synchronizing means between the shafts 31 and 43.

The second stage of the compressor is welded on one side to the hoop ring 66, `and on the other side to a torque transmitting hollow cylinder 68 similar to the hollow cylinder 44.

Ambient air enters the air duct 10, passes through the compressor contra-prerotation stage 69, the rotatable compression stages 64 and 63, and then enters the toroidal combustion heat generator 35, which is described more fully in the parent case 217,347, now Patent 2,804,747, and divisional application S.N. 606,451, filed August 27, 1956. 'Ihe combustion chamber revolves with the second stage 63 of the compressor and the first stage of the turbine.

The compressed air which by-passes the combustion heat generator and the heated gases from the combustion heat generator enters a radial input duct into a radial centrifugal flow turbine having first and second radial stages mounted in the same manner as the compresso'r stages. 'I'hese two contra-rotating turbine stages rotate in opposite directions the two stages of the compressor, the torque developed by the turbine stages being transmitted to the compressor stages through the hollow cylinders 68 and 44. 'I'he two contra-rotating turbine stages are rotating at two different angular velocities, the ratio between these two angular velocities being held constant by the synchronizing gear train between the two rotors of the turbine and the two rotors of the compressor. As will be explained more fully later in this specification, proper vectorial relationships are maintained in the compressor and the turbine by maintaining the angular velocities ratio constant between all rotating elements of the power plant. The angular velocity of the compressor is controlled by a governor geared to a ring gear 30 which controls the rate of fuel supply furnished to the combustion chamber. This, in turn, controls the temperature of gases leaving the heat generator, thereby controlling the angular speed o'f the turbine and compressor. The governor, therefore, does not maintain this speed constant but adjusts it to maintain a constant Mach, number throughout the plant, which is especially desirable because of the supersonic type of compressor.

The joint application of the present applicant and Frederick Dallenback, filed August l2, 1950, Serial No. 179,028, now Patent No. 2,712,895, entitled Centripetal Subsonic Compressor, discloses a centripetal Subsonic compressor. The following appears in the above application under subtitle Fluid Dynamics and Thermodynamics of the Centripetal Compressor: In any rotary compressors having continuous compression cycle, as differentiated from intermittent cycle (reciprocating compressors), the only way that the process of compression can'be accomplished is by first accelerating an elastic fluid to as high velocity as possible to obtain as high a kinetic energy as possible from the ambient potential energy, state and then converting or transforming this kinetic energy into potential energy (pressure). The above statement leads one to an immediate conclusion that the higher is the velocity of the accelerated air prior to diffusing or shock co'nversion of the kinetic energy into pressure (if air is the uid medium), the higher will be the final pressure obtained with the compressor. This being the case, the next logical question is whether there is any practical limit to this velocity. Such absolute limit does exist, and it is within the supersonic velocity range. The absolute velocity'limit is determined by the overall gain in compression obtained in the first rotor stage which is also a function of the Mach number of the absolute velocity at the entry to the first compression ratio obtainable in the first rotor.

order of 1 to 1.3, the gain in the first compression stage is maximum. Beyond 1.3, the absolute pressure in the gap between the prerotation stage and the first compression stage becomes so low that the first compression stage cannot overcome this drop in pressure with any additional vincrease in compression. Stated differently, the higher the Mach number in the gap, the lower is the gap pressure. Therefore, the rst stage before producing any gain must overcome lower and lower pressure as the Mach number increases. There comes a point at which any further increase in the Mach number begins to reduce the overall compression ratio. This will be discussed more fully in connection with the compression ratio curve illustrated in Figs. 8 and 9. lt is one of the requirements of the supersonic compressors that, in order to operate supersonic compressors at optimum performance characteristics, it is necessary to operate such compressors at substantially constant Mach number. In some applications, it is desirable to operate the compressors at variable speed, and when this is the case, then it becomes more advantageous to use subsonic compressors whose characteristics permit their operation at Variable speeds without materially affecting their performance. The disclosed compressor is a supersonic compressor.

irrespective of whether the compressor is subsonic or supersonic, the maximum obtainable compression ratio is a function of the maximum peripheral velocity which can be attained safely With the existing high strength alloys. As illustrated in Fig. 9, the higher is the peripheral velocity, the higher is the maximum compression Therefore, one of the factors limiting this ratio is the maximum allowable proportional limit of metals used for making the compressor ro'tor. For conventional steel, this limit is reached when the peripheral velocity is of the order of 700 feet per second in centripetal compressors with ringY rotors, while with the newly developed titanium alloys this limit can be extended up to 850 feet per second, It is even higher for the new vacuum-processed steels. Supersonic compression of gases is accomplishedrby means of aerodynamic shocks induced in the channels formed by the blade surfaces. The sharp entry edge, and particularly the leading or pressure side of the rotor bladeis used for producing an oblique shock and the converging compression channel defined by the opposing surfaces o'f two adjacent blades produces a normal compression shock. There also may be a reflected shock following the oblique shock. Adiabatic compression of gas raises its temperature and its sonic velocity;V therefore, the supersonic compressor, after producing a series of shocks and the concomitant supersonic shock'compression, becomes a subsonic compressor where ,addii tional compression is obtained through additional sub-` sonic diffusion of the compressed fluid.

Referring to Fig. 2, it illustrates a cross-sectional view, normal to the axis of rotation, of the prerotation stage and two supersonic compression stages.

Prerotation stage of any dynamic compressor must expand the ambient air pressure to a sub-atmospheric pressure thereby accelerating the sttationary ambient air mass to a velocity whose vector is acceptable to the first stage of the co'mpressor. The relative entry velocity, into the first stage, which is known in the art as velocity W1, should be as high as possible. In the disclosed supersonic centripetal compressor, the Mach numbers Mach number where a1 is the local velocity of sound) of the relative velocity W1 can reach values of the order of 1.5 to 3.0, depending on the fluid used. Therefore, the configuration of the entry channel, which is the channel deiined by the convex surface 202 of the prerotation blade 204 and must have apre-acceleration region for gradual acceleratio'n of 'the air mass crossing the plane'deiined b y the] outer periphery of the prerotation stage, and then the f acceleration region which enables one to each the rela" tive velocity in the input of the rst'compression stage A having a Mach number ofthe order of 1.5 to 3.0. The

pre-acceleration region is necessary -to avoid rst, in the first half of the entry region, excessive friction losses because of large area exposed to the ilow, and second,to avoid ow separation in the intermediate region of the channel where maximum turning of theflo'w occurs and Y where it blends into the accelerating channel proper.

Thevblades, therefore, assuming the shape of curved or cambered blades which begin with a circular leading edgeY 208 having a radius 209 whose center is positioned on the radial line 210 of the entire compressor circle 2,11; The

magnitude of radius 209 is not critical andis primarily dictated rst by the ability of the cylindrical surface to divide the yair ow equally into each pre-acceleration flow channel, and secondly, to avoid oW separation in the process, the latter consideration setting the maximum length that can be assigned to radius 209.

The cylindrical leading edge zus of the made thenu blends into two sloping planes 212 and 213 on the opposite sides of the cylindrical leading edge, these planes forming two equal angles, 214 and 215 with the radial line 210. The magnitude of these angles is of the order', yof 8 to 12. The right side of the blade 204, as viewed in Fig. 2, then blends into a circle having a radius 216 whose magnitude is not especially critical and is Yprimarily determined by the rate of acceleration desired in the subsonic acceleration region of the flow channel.l The shape of the flow channels is defined by the iiat inner side-surfaces 1137and 112 (see Fig. l) and by opposed surfaces of the blades. curved median axis line 218 bisecting arc 219 (an arc of outer periphery 211 between radial lines 210) at point 220 at the entry into the channel. Axis line 218.also bisects a line 223, the length of which is equal to the minimum width of the flow channel between the surfaces of the blades. between the subsonic and supersonic acceleration regions of the entire channel. sumes the shape of a supersonic expansion nozzle dened by two straight surfaces 224 and 225 which form equal obtuse angles 226 and 227 with line 223. The median@k line 218, between points 221 and 222, is a straight line and is normal to line 223. The supersonic expansion nozzle terminates at the inner periphery 201 of the p rerotation stage where all blades terminate in trailing edges From a purely theoretical point of view these 223. edges should be razor-sharpY edges to avoid any excessive wakes if one is to consider the entire problem of iiuidr,

dynamics only in terms of ideal ow. However, mechanical strength and stiffness requirements make `it necessary to end the trailing edges as cylindrical surfacesVK having a small radius. This, obviously, produces a -srnallfn degree of turbulence, the overall effect of which is benelicial rather than detrimental if not carried to an extreme@` These small wake turbulences diminish therate of growth p of the thickness of the boundary layers in -the succeeding Y compression stage, which would be far more detrimental v than the small venergy loss due, to the wakes. ToV utilizef thisbeneficial wake eiect in the succeeding stage, the air gap 229 should be made as small as mechanically possible.

The inner end of the yarc produced by radius 216 forms.,v a tangent connection with a substantially straight line 217 which extends to line 223 where it makes a junction with line 224 of the supersonic nozzle. Line 225 terminates at line 223 at which point it is tangent to arc 202. Y Arcs curves; the onlyfrequirement is that no separation is introduced by these surfaces, and the general proportioni'ngQ i of the channel is such thatrthe maximum rate of accelera-1QV by the concave surface 206 of theidentical blade 207 tion in the. subsonic region of the channel isdelayed The flow channel will have al Line 223 constitutes the boundary. line Beyond line 223 the channel asasa-igea! suiciently to avoid needless frictional losses in the early part of the channel. The blades can easily be made hollow `todecrease the weight of the compressor.

The mean flow line 218 stops at point 222, at the inner periphery of the contra-prerotation stage, and then it is deviated by an angle 230, which is known as the flow deviation angle. 'Ihis angle isy formed by twolines 218 and 231, line 231V designating the direction of actual ow after the exit from the prerotation stage, and therefore constitutes the direction of an actual velocity vector C1 of the fluid flow in the interstage gap 229. Therefore, in order to obtain proper vectorial relationship with respect to the first rotor, the mean ow line 218 of the prerotation stage must be turned by the same angle in a counterclockwise direction direction as viewed in Fig. 2. Stated differently, angle 230 thus becomes the correction angle for the geometry of the supersonic region of the prerotation stage.

Fig. 2, on the right side, also illustrates another version of the prerotation stage in which the supersonic nozzle of the straight type is replaced with a supersonic nozzle in which the acceleration is produced in a shorter distance by means of two curved surfaces 231 and 232 originating at the minimum width of the flow channel. The advantages of this nozzle are: (l) shorter length and therefore, smaller loss due to friction; (2) more uniform velocity distribution across the adjacent portion of the subsonic region; and (3) greater ease of introducing the correction angle 230. The first advantage results from the shorter length of the nozzle because the expansion is more rapid. The second s due to the fact that lines 233 and 234 are substantially straight lines which form equal angleswith the median line 218, which is also a straight line at this part of `the channel. Therefore, equal degrees of convergence of the two channel surfaces will produce equal degrees of acceleration. Hence, except for the boundary layer, there is a uniform velocity distribution. These surfaces, together with the viscosity effect, Will tend to equalize the flow velocities even though these velocities are not uniform at the plane of entry of fluid into this region because of the velocity differences produced by the surfaces 235 and 236. A further refinement may be introduced into this region by making the angle of line 233 with respect to line 223 smaller than the angle of line 234 to make the velocity distribution, except for the boundary layer, exactly uniform at line 223.

The first supersonic compression stage consists of hollow blades 250 having sharp leading edges 251. To avoid excessive wear, vibration and bending, the leading edges are actually cylindrical surfaces having small radii. Surface 252 of the blade is a flat surface parallel to the relative velocity vector W1. It will be referred, at times, as the lagging surface of the wedge-shaped portion of the blade. It terminates as a tangent .to a preferably cylindrical surface 253 having a radius 254. Surface 253 merges into an arcuate surface 258, which terminates at an edge line 259. The remaining surface 270 of the blade, from edge 259 to the cylindrical tip 251, is a iiat surface, and it will be referred to, at times, as the leading surface of the wedge portion 270-251-252 of the blade. 'I'he channel consists of a compression channel from a point 260 to a point 261. .Point 260 is on the outer periphery 262 of the compression stage, and the position of point 261 is determined Iby the position of the boundary between the supersonic compression portion ofA the flow channel and the subsonic, substantially constant velocity channel, this boundary being defined by the position of the reflected shock plane 265. Point 263 is a median point on the line defining the position of an oblique shock plane 264. From -point 261 to point 266, the latter lying on the inner periphery 267 of the compression stage, the channel is a subsonic duct whose normal-.to-ow cross-sectional areas are made constant to maintain constant iiow velocity. The median ow in the `channel is designated bythemedian-flow line 268` which joins the above mentioned points. The sharp edges 251 and the leading surface 270, or the pressure side .of the blades encounter the fluid moving against them with the relative velocity W1. This produces an oblique shock, the latter producing a stationary wave `front 264 or an oblique compression wave existing in the plane defined by an infinite number of lines identical to line 264. Because of the presence of the inclined, leading plane 270, which forms an angle 269 with the velocity vector W1, the gas flow is deflected clockwise by an angle equal to angle 269 (this angle is equal to the angle formed by the wedge-shaped leading portion of the blade), with the result that .it will follow the line joining the points 263 andv 261 at a reduced velocity but higher pressure than the pressure existing in front of the oblique shock wave front. If the relative velocity W1 is sufficiently high, the oblique shock will have sufficient velocity and energy to be reflected from surface 252 to produce a reflected shock wave front 265. The reflected shock again reduces the velocity of the flow stream and converts its kinetic energy into potential pressure energy so that the pressure on the inner side of the reflected shock wave front is higher than on its outer side. The maximum relative velocity W1 and the angle 269 are proportioned to produce only two shocks 264 and 265, with the result that all flow beyond shock 265 is subsonic, the reduced subsonic velocity vector W2 existing throughout the remaining part of the channel. This remaining part, i.e., vfrom point 261 to point 266, is merely a turning channel, the degree of turning being determined by the magnitudes of W2 and U3, which is the peripheral speed of the second stage, and the minimum feasible angle of entry of W2 into the compression channel of the second stage. The wave fronts 264 and 265 may or may not be straight planes, depending on the conditions of ow. Irrespective of the configuration of the shock fronts, the functioning of the compressor will be the same.

In order to make. the subsonic portion of the flow channel merely a turning channel, its cross-sectional area, normal to Ithe direction of ow should remain constant at all stages. Stated differently, area A1 should be equal to area A2, Fig. 16. To obtain this, it becomes necessary to reduce the axial width of the channels, as illustrated at 114 `and 116 in Fig. l. It is to be noted that the transverse dimension of the subsonic portion of the channel, as it appears in Fig. 2, is increasing very rapidly as one progresses inwardly toward the inner periphery 267 which is necessary topmake the blades terminate in sharp trailing edges 256. Thus, the channel, as it appears in Fig. 2, creates an impression as if it were a diffusing channel, and the correction illustrated in Fig. 1 is the correction which lis necessary to transform it into a constant velocity channel.` lt is also to be noted Ithat the above configuration increases the hydraulic radius of the channel toward the exit which is favorable for diminishing friction and the thickness of the boundary layer because of Kthe continuous increase in the Reynolds number. It is advantageous to maintain the velocity W2 constant since greater compression ratio can be obtained in .this manner in the second stage. This is so because the compression ratio of any stage is a function of the local Mach number to the second power.

The compression technique used in the second stage is identical to the Itechnique used in the rst stage, and therefore does not require a detailed description. Oblique and reflected shocks `are used which create the respective compression wave fronts 275 and 276. The main difference between the first and the second stages resides in the fact that the ow channel on the other side of the wave front 276 is a diffusion channel, the side walls 118 and 120, Fig. l, of the channel being parallel to each other, Iwhile the distance between the opposed surfaces of'the blades 280 is increasing in centripetal direction. The rate of diffusion is determined by 91 the use to which the compressed air is subjected. In the disclosed compressor, the compressed air is ducted into the combustion chamber 35, Fig. 1, which revolves at the same angular velocity as the second compression stage. It leaves the compressor at a relative velocity W4 which is considered to be of proper magnitude to produce the desired degree of swirl in the combustion chamber, which in turn is determined by the rate of combustion desired in the combustion chamber.

Therefore, the degree of diffusion is a function of the rate of combustion. The relative velocity W1 is, in this case, radial, and therefore, the abolute swirl velocity C114 is of the same magnitude `as the peripheral velocity U4 at this radius, i.e., C111=U4, and the absolute velocity C1 forms a leading tangle fy with the radial line. The shape of the blades in the second stage is determined in the supersonic region in the same manner as the shape Y of the blades of the first stage in the same region. The two differ in shape, however, because W1 W3. As to U1 and U3, U1 may be made greater than U3 if the angular velocities of the two rotors are made equal. However, it is more advantageous to make U1=U3. This will be discussed more in detail in connection with the vector diagram of the compressor illustrated in Fig. 7.

In Fig. 2, a compressor is disclosed which uses oblique and reflected shocks in both stages to obtain the compression of gases. This compressor obtains the highest compression ratio at the highest thermodynamic elticiency for the obtained compression ratio, and therefore, olers advantages over the compressors illustrated in Figs. 3, 4, and 5. However, the additional compressors illustrated in the above figures have some of the advantages of their own, as will be described later.

Referring to Fig. 3, it discloses a two-stage compressor havin-g a stationary lsupersonic contra-prerotation stage and two supersonic compression stages which utilize only an oblique shock for obtaining compression. Since only oblique shock compression is used in the compression stages, the relative velocity W1 need not be especially high, and therefore, the prerotation stage may be a cornpletely subsonic stage. The shape of the blades 300 would then differ from the blades illustrated in Fig. 3 only in the lower part of the blades. While in Fig. 3 the flow channel terminates in the supersonic expansion nozzles of the types previously described with Fig. 2, the subsonic acceleration channel would then continue to the very end by gradually narrowing the dimension of the flow channel in the plane of the drawing, i.e., in the plane normal to the axis of rotation. A subsonic prerotation stage of this type is illustrated in Fig. 10.* The fluid to be compressed leaves the prerotation stage at a subsonic or a supersonic absolute Velocity C1, and then enters the ow channel of the first stage at a relative supersonic velocit-y W1, which must be a supersonic velocity irrespective of whether C1 is subsonic or supersonic. The leading edges (on the pressure side) of blades 3011, 302, etc., create oblique shocks, the wave fronts of which are illustrated by inclined lines 303, 304, etc. The configuration of the blades in this stage differs from the configuration of the blades illustrated in Fig. 2 by the omission of an edge 259 between at surface 270 and curved surface 258 since no reflected shock is used in this compressor. Angle 311 must have the magnitude which will produce only an oblique shock of compression with the entry velocity W1. The equation for this angle is given later in this specification, in connection with the discussion of Figs. 11 through 14. The remaining portion of the flow channel preferably is a constant velocity channel, as in the case of Fig. 2, since introduction of any diffusion in this channel would only produce the thickening of the boundary layer `and reduce the exit velocity W3, thus lowering the compression ratio obtainable in the second compression stage. Stated differently, greater compression ratio is obtained by maintaining W3 as high as possible and therefore, any attempt of obtaining some compression'b'y diffusion in the iirst stage would produce a disproportionate loss of compression in first stage, i.e., the edges 282, Fig.`2 now are eliminated since there is no reflected shock. The oblique shock Wave fronts are illustrated by lines 306 and 307 inrthis figure. From then on, i.e., from point 308 to point 309, the channel is a subsonic diffusion channel where the compression obtained by the oblique shock 307 is increased subsonically to a still higher pressure by converting relatively high kinetic energy to potential energy through the process of subsonic diffusion. The advantages of this compressor, disclosed in Fig. 3, are several. Although it does not produce as high a compression ratio as that in Fig. 2, it nevertheless produces a somewhat lower compression ratio at the highest thermodynamic eciency because of Very nearly isentropic process of compression by the oblique shocks.4 It also will operate quite effectively and ellciently in the subsonic region, and -therefore will pull in into and through the-subsonic region without any difliculties. This is so because of the absence of sharp transverse edges 259 and 282, Fig. `2, which are apt to cause a certain amount of separation in the subsonic region, in the compressor disclosed in Fig. 2.

Fig. 4 discloses a compressor which utilizes oblique, reflected, and normal shocks. In this case, the relative velocity W1 of the uid at the exit from the prerotation stage must be high in order to produce three shocks in the compression stage. The prerotation stage, therefore, must have a supersonic nozzle, identical to the nozzle disclosed in Fig. 2, which will produce supersonic velocity W1. Sides 400 and 401 of blades 402 and 403 are identical flat surfaces parallelto the relative velocity W1. Angles 404 and 405 are computed according to the formula y given in the latter part of the specification to produce oblique, reflected, and normal shocks with the given 4supersonic lrelative velocity W1. The wave fronts of these shocks are 406, 407, and 408 respectively. One` of the requirements of the compression channel utilizing normal shock is that its minimum width must be at thel position where the appearance of .the normal shock is desired, which is the case in Fig. 4. From then on, the channel may be either a diffusing channel or a constant velocity channel,V the latter offering the advantages mentioned in connection with Fig. l2, i.e., higher compression ratio is `obtained in the second stage when W2 isk Vmaximum; since W2' is maximum, W3 is also manmum.

pression Vratio because lof the use of .three shocks, butl this high compression ratio is produced at a" -decreased thermodynamic eiciency. This compressor also presents `some structural limitations in that the angles 404, 405, etc., are small and therefore the leadingk edges ofthe blades are less rigid. The pulling-in characteristic of :the compressor disclosed in Fig. 4 is inferior to the pullin characteristic of the compressor disclosed in Fig. 2, ,Y

land even more inferior to the pull-in characteristic of the compressor disclosed in Fig. 3.

Fig. 5 discloses a compressor in which the prerotation stage has a supersonic nozzle, and the compression stages y produce oblique and normal shocks. The compression ratio of this compressor can -be as high as that of the compressor illustrated inFig. 2 and also in Fig. 4 if the relative entry velocities, W1, are equal in all cases.

All compression stages utilizing normal shock require sharp edges 409, 414, Fig. 4, 500 and 501, Fig. 5, ex? tending through the entire length of the channel, to locate and keep the position of the normal shocks in xed positions with respect tothe blades and at the minimum width of the channel. This will be discussed more fully 1n connection with Figs. 11 through 14.

Fig. 6 illustrates the method of reinforcing compression blades by means of thin vanes 600 and 601V placed midway across the blades 604 and 605 to prevent centrifugal deflection of the leading edges 602 and 603 of the blades. The leading edges 606 and 607 of the blades must be sharp edges, and these edges should be positioned behind the oblique shocks, where the Mach number is lower. These vanes are placed in slots, such as slot 608, with which they form a tight t. After their insertion into the slots, the vanes and the blades are brazed together. The vanes constitute integral parts of the blades. When the com# pressor blades are especially wide, several vanes may be used with each blade.

Fig. 7 illustrates the velocity vectors of all velocities present in the compressor of the type illustrated in Fig. 2. All of these velocities are also illustrated in Fig. 2 in proper relationship with respect to their actual location in the compressor stages. Therefore, the description of these vectors will be given in connection with Figs. 2 and 7. The lluid to be compressed enters the prerotation stage in radial direction with an absolute subsonic velocity C0, this velocity being taken at the outer periphery 211. In the prerotation stage, C is transformed into an absolute supersonic velocity C1 leaving the contra-prerotation stage at its inner periphery 201 at an angle of approximately 25 which is an optimum angle, and is determined by the angular position of the mean flow line 218 at the exit from the prerotation stage. Generally, from a theoretical point of view, this angle should be as small as possible, in order to produce as large a relative velocity W1 as possible and large C11. As it Will be demonstrated in con-Y nection with Figs. ll through 14 (Equation 5), the compression of any stage is a direct function of the product of the peripheral velocity, U, and the projection of the absolute velocities, C1 and C2, at the entry and exit of compression stage, upon the peripheral velocity U. Therefore, the smaller is the angle l, the greater is the stage compression. However, this angle cannot be made equal to zero because no fluid would enter the compression stage under such conditions, and there would be nothing to compress. Geometric solution of this problem -involves composing the shapes of the blades of the prerotation stage, and more accurately stated, the shape of the flow channel of the prerotation stage and its proper angular relationship lwith respect to the entry channel of the first stage, the angle and the shape of the latter being dictated primarily by the peripheral velocity U1 and 131. The minimum value of ,31 is of the order of 15. This angle is an inverse function of the radius of the compressor, and the above value is for the compressor in which the outer diameter of the iirst compressor stage is 32.5 inches. With the increase of the radius this angle decreases since there is a corresponding decrease in the curvature of all stages. The outer periphery of the rst stage travels at the peripheral velocity U1, the value of which should be as high as possible, and its limit is determined solely by the strength and weight of the available metals. For an ordinary steel it is of the order of 700 feet/second, and

for titanium alloys it is of the order of 850 feet/second,

maximum. For vacuum-processed steels it .may be as high as 1000 feet/second. The relative velocity, W1, then merely closes the velocity triangle. As mentioned before, ,81 is then the angle between W1 and the tangent line 900 drawn through theV point or origin 901 of the vectors. This point lies on the outer periphery of the first stage. The absolute radial velocity of the uid is C1. The significance and the limits of the' Mach numbers will be discussed in connection with Figs. 8 and 9.

The vector diagram for the second stage is derived in the same manner. The fluid leaves the first stage with relative velocityfWz and absolute velocity C2. The first stage has a peripheral velocity UZ at its inner periphery 267, Fig. 2.Y At this stage, it encounters the oppositely rotating second stage, the outer periphery of which rotates at the peripheral velocity U3, which is opposite to U2 because of the contra-rotation of the second stage with respect to the rst stage. This produces a relative velocity W3 at which the fluid enters the second stage. The radial velocity Cr2 is increased and,ltherefore, C12 C11. Shock compression and dilusion in the second stage reduce all velocities, the degree of the nal reduction being .determined by the degree of diffusion in the ditusion channel, which may be Vmade greater or smaller, according to the final use of the compressed gas. As mentioned previously, the exit absolute velocity C4 may be radial or follow the direction of rotation, as illustrated by the two triangles in Fig. 9, its direction being determined by the angle the mean ilow line 279 makes with the inner periphery 28S of the second stage.

4 Optimum operating conditions for the supersonic centripezal compressor Fig. 8 illustrates a temperature-entropy diagram of a compression cycle of a dynamic compressor with a prerotation stage. This diagram will be discussed primarily in connection with the performance of the centripetal compressor. However, it is equally applicable to all other types of compressors, axial and centrifugal, if they are provided with stationary prerotation stages.

The atmospheric pressure, P0, is the pressure at which the air enters the prerotation stage; the entry velocity is C0, and is equal to zero, which means that the ambient air is at a standstill. This point is indicated as point 0 on the diagram. The first stage of the compressor is the stage which accelerates this ambient static air from velocity CU to velocity C1, which may be either supersonic or subsonic velocity, depending upon the pressure differential created by the tirst and second compressor stages.

In Fig. 8, this velocity, according to Bernoullis theorem, produces a corresponding pressure drop to P1, which is lower than the atmospheric pressure, this loss in pressure corresponding to the gain in the kinetic energy of the air entering the compressor. This loss in pressure head fis expressed by y ft. ibs. Cfzg 1b. (2)

The higher is the exit velocity C1 from the prerotation stage, the larger is the value of AHCl, and it becomes necessary to determine what is the optimum value of C1 for obtaining maximum compression ratio. Conditions may arise when this loss in pressure may be so high that the recovery otthis pressure in the tirst compression stage either would not be possible altogether, or only a minor recovery would be obtainable, in which case the advisability of using the prerotation stage, at first glance, may be questioned altogether. As will appear later, the presence of the prerotation stage is more than justified under all circumstances, and that the only parameter which requires a closer scrutiny isthe value of C1 which controls, at least insofar as the practical operating range is concerned, the pressure ratioA of the entire compressor. Thus, this velocity determineswhether the prerotation stage should be subsonic or supersonic, and whether the first compression stage utilizes only ran oblique shock or oblique-and-reected shocks for compressing the air.

The local Mach number of velocity C1 is expressed by C C' zus-5l.: L 1 3) c2l ao2. }012 where Mc1=local Mach number of C1 ac1=local sonic velocity, at point 1, Fig.v 8 a11=sonic velocity at point 0, Fig. 8 k=ratio of specific heats,

Cp Cv The air, after being accelerated to velocity C1, enters the irst compression stage, which accelerates this air to higher absolute exit velocity Cz'and at the same time compresses it to pressure p2, this state of the air being illustrated by point 2 in the diagram. The rise in pressure is denoted by the positions of the respective points on the constant pressure lines P through p8. The pressure gain obtained in the rst compression stage cor-Y t.

responds to the pressure difference between p2 and p1. This pressure gain :is used for deriving a pressure head, AH131 and mechanical head AL1 from the following equa- AHp1=pressure head obtainable in the first compression stage g: acceleration due to gravity Cul=projection of C1 on U1 U1=peripheral velocity of the outer periphery of the iirst stage f Cu2`=projection of C2 onUz U2=peripheral velocity of the inner periphery o'r' the rst stage C2=absolute exit velocity of air at the exit from the rst stage C1=entry velocity into the first stage and exit velocity from the prerotation stage.

Also, AHPO, which is the net pressure head gainlin the first stage, ie., actual pressure gain above static atmosp heric pressure, is derived from the equation l 2 Anpfoul U1- cu2 U2-922-} (a) where the terms are the same as those inthe Equations 2 through 5 Besides producing the above pressure head gain, the rst stage also increases the kinetic energy head by the amount AHK, which isV Y and, therefore, the total head rise in the rst stage above static atmospheric stage, which is represented by the vertical distance between points 0 and 3, and designated by H0 in Fig. 24, is A The most important quantities` are AHU, AHp0 and AL since they determine'the net performance of the first stage. The quantities are plotted against Mach number ofthe velocity C1, in Fig. 9, which will be discussed more in detail later. Y,

l The compressed air leaves the rst'compression stage with an absolute velocity C2 which is larger than C1, the

AHk: (7)

velocity appearing in Fig. 8 is the velocity head AHC2 which is equal to Y AH02 2g (9) have any prerotation stage.

Therefore, point 4 corresponds to the stagnation prmsure of velocity C2, the corresponding pressurerline being ps.

MFM

It is to be noted here that the absolute exit Velocity C2 should be larger than C1 for optimum overall performance, which will produce a large compression ratio in the second stage. That C2 should be larger than C1 follows from the Velocity diagram in Fig. 7. This velocity, C2EC3 is a function of ,83 and of the peripheral velocity U3 of the succeeding stage and optimum compression ratio of the entire compressor is obtained when C2 is as high as possible and larger than C1. As stated previously, C2C3 is maximum when a is of the order of 15. It may be also stated here that C3 may be made smaller than C1, but this can be done only at the exf pense of the overall compression ratio'. Y v

The compression cycle in the second stage begins -at point 2 rather than point 4, since point 4 represents the stagnation condition of stage 1. The mechanical head of the second stage, ALZ, is supplied by the mechanical energy delivered to the second compressor stage by the first stage of the turbine. The pressure head of the second stage, AHM, is from point 2 to point 6, and subsequent subsonic dilusion produces an additional gain in pressure head, AHd, which is C72 C62 l AHd- 2g Y (10) thus raising the pressure head from point 6 to point 7.'

The absolute exit velocity, C7, whose Mach number is of the order of `0.2, is represented in the diagram as the velocity head AHq-- (11) thereby reaching stagnation pressure p8 at point Y8. The pressure ratio p11/p0 is the total compression ratio of the compressor.

The above head-entropy diagram isrfor the two stageA compressor. For the three stage compressor, the diagram should be continued in similar manner for one additional stage.

Before concluding the description of Fig. 8, lit is noteworthy to point out the important role being played by the prerotation stage in the compressor. VjThe prerotation'stage increases the velocity C1 and its peripheral component C111 as indicated in Fig. 7. Since the mechanical head AL1 is equal to ALFtculUi-cnus 5) it is obvious that AL1 is a function of C111. The elect of C1 on AHO is illustrated in Fig. 9, and extrapolation cf this curve to Mc1=0, when AHO is also equal to Zero, indicates that there is a very rapid decrease in AHU with the decrease in C1. This dotted region corresponds to the region of operation when the compressor does not Low compression ratio in the first stage obviously will iniluence the remaining stages as well.

Proceeding now with the description of Fig. 9, examination ofthe AHpO and AL1 curves discloses that these heads increase continuously with the increase of Mc1 or with the increase of C1. reaches its maximum at Mc1f-v0.9, and then decreases to a constant value when M1.=2.4 or 2.5. Fig. 9

also includes two additional curves, one for AHCB and the,

The curve for AHO, however,V

15 other for ALT. AHc2 is the velocity head of the absolute exit velocity from the first stage, and is equal to while ALT is the total mechanical head of the compressor, and is equal to limited reduction in the air velocity through the first stage when there is only an oblique shock, and what is more important, the two contra-rotating stages produce the summation of U2 and U3, the resulting relative velocity W3 becomes a very large supersonic velocity. Therefore, the second compression stagemay be of the type in which the oblique as well as the reflected shock become possible. This combination of the stages is disclosed in Fig. 10.

The above demonstrates that different types of stages may be combined with each other to produce the sought results, from which the following table results.

SUPERSONIC OB SUBSONIC PREROTATION STAGE where U3=peripheral velocity of the outer periphery of the second stage Cu4=projection of the absolute exit velocity, C4, from the second stage on U4 U4=peripheral velocity of the inner periphery of the second stage In Equation (l2), the directional signs of the velocity vectors have been taken into consideration, and therefore, it is applicable only to the vector diagrams as they are used here.

Examination of the curve for ALT discloses that ALT increases slowly with the increase of Mc1 or C1. Therefore, it would appear that Mc1 should be made at high as possible for obtaining maximum overall compression. This, however', is not so for two reasons. One reason is the increase in friction losses at high velocities, which do not appear in Fig. 9. Another reason is of purely mechanical nature: the shape of the leading edges of the cascades assume the form of wedges having progressively larger includes angle (angle 269 in Fig. 2) with the result that the concave surface forms a progressively smaller angle with the inner side of the wedge. This will produce separation at subsonic velocities during the starting period, which will create starting diiiculties Therefore, the maximum limit for Mc1 is somewhere between 1.5 to 2.0. Line 900 indicates the point for the twostage compressor illustrated in Fig. 10. Y

If the maximum limit for Mc1 is in the region indicated above, there is also a minimum limit which is in the region of Melee] to .9, although the compressor will produce compression even in the region lower than Mc1=-.7, but the second stage will need a large subsonic diiuser to convert very high kinetic energy into piressure. The above minimum values for Mc1 definitely indicate that the prerotation stage may well be a subsonic stage. This will require blades in both compressor stages which would produce oblique. shocks only, which is especially the case for Mc1=.7, since the kinetic energy of air will be insufficient to produce an oblique as well as the reflected shock. This type of compressor has distinct advantages in that, although the overall compression ratio may be somewhat lower, its operating characteristics become much more ilexible in terms of rotational speeds. IFor example, such compressor will pull in during starting periods with greater ease and lower rotational speed because of the shape of the blades which produce flow channels substantially devoid of separation dangers at lower speeds.

An additional combination of stages is alsopossible. If the prerotation stage is subsonic, the `first stage in all likelihood `will be only an oblique shock stage because of low velocity C1. However, since there is only a O=oblique shock stage OR=oblique and reflected shock stage ORN=oblique, reflected, normal shock stage No stage is suggested which utilizes only the normal shock since the efficiency of such compressor would be quite low in comparison to the suggested combinations. It also has mechanical limitations due to the difficulty of anchoring this shock at the at surface of adjacent blade, although this type of compression is possible in connection with the centripetal type of compression.

Supersonic compression by means of supersonic compression shocks Although it has been known that the supersonic compression shocks constitute eiectime means for obtaining eicient compression of gases within extremely limited space, the subject invention constitutes the first application of the supersonic compression shocks to centripetal compressors for obtaining compression of gases at high eiciency.

The disclosed methods of supersonic compression are based upon the principle of inducing and localizing the geometrical position of the shocks to the input side of the ow channels. Stated differently, the blade structure is such that there is a highly selective production of the desired shocks and elimination of all parasitic shocks; the elimination of the parasitic shocks is imperative for obtaining supersonic' compression of gases, and the disclosed method of compression is impossible without such elimination. Therefore, the disclosed channel structures are adapted not only to produce supersonic compression shocks in proper geometric relationship with respect to the geometry of the flow channels in the centripetal compressors, but they are also constructed to eliminate the parasitic shocks in adjacent channels and in the gap between the stages. This will appear more clearly in the course of the discussion which is to follow.

Fig. 15 discloses the cross-sectional view of a twodimensional, leading edge of a blade in a free supersonic stream of flow at supersonic velocity W1. This leading edge 1500 is formed by the intersection of two flat surfaces 1501 and 1502, which thus form a wedge-shaped structure 1503. It is to be noted at once that one of the flat surfaces, i.e., surface 1501, which is the lagging, or the suction, part of the blade in the disclosed compressors, is parallel to the velocity W1 and therefore, apart from the frictional eiects, is incapable of producing any disturbance in the stream, i.e., it does not produce any supersonic compression shocks which would otherwise interfere with the desired shocks.

The other surface 1502, which is the leading side, is inclined at an angle with respect to the direction of iiow and, therefore, forces the ow to deflect from the direction of vector W1 to the direction of vector W2, the

where p1=gas density before the shock p2=gas density after the shock;

Wn1=normal velocity component of W1; which forms 90 angle with the plane of the shock (see Fig. 15); it is a function of W1;

Wn2=normal velocity component of W2; which forms 90 angle with the plane of the shock; (see Fig. 15). The momentum equation must be satisfied also; it is and, from 15 p1=absolute pressure ahead of the shock,A

p2=absolute pressure after the shock;

Wt1=compcnent of the stream velocity W1 in the plane of the shock (see Fig; 15)

Wt2=component of the stream velocity W2 in the plane of the shock.

The third equation which must be satisfied is the equation of conservation of energy which is K p l 1 K p2 l 2 K1 p.+2W1 K-1 E+W2 17) where K is the ratio of specific heats.

From (14) and 16) it follows that:

W'tf-:Wt2` (18) and from Fig. 15

W2 oos Wlcos a (19) Using where a is the sonic velocity before the shock, it follows p12Wh12=p12W12 Sill ,9:p1M1ZIP1 sin2 S From the above equation, it is possible to derive the fundamental equations of shocks, which are:

l 22- )K*1 2( 2 Krl pl-(Tl K+1 M1 sin- 2K) (21) p1 2 1 p-z K-ll M12 sin2 0+ 2 22) and from (16) and (18) & I W112W1 tan a (23) p2 P17111 Wtz Wal tain i? therefore,

tan (19-) 2 1 'K-l) (24 tan K-}1\M12 sinl i? 2 Equations (23) and (24) are general equations and are applicable to all compression shocks (oblique, reflected,

normal). For normal shocks 7l' 1?-2- and 5 0 v channel.

18 the flow, therefore does not change direction and the shock is normal to the ow. Therefore, it is called the normal shock. The normal shock gives the maximum obtainable pressure ratio for a given Mach number in front of the shock.

For the normal shock, from 2l where subscript n means that the pressure ratio is for the normal shock and MaK-ow and for the Mach number behind the normal shock 2 Mlll 2K 2 Figs. l1 through 14 illustrate the application of the oblique, reflected and normal shocks in four possible combinations, which are (l) Oblique shock only (Fig. l1).

(2) Oblique in combination with normal shock (Fig. l2). (3) Oblique and reected shocks (Fig. 13).

(4) Oblique, reflected and normal shocks in combination (Fig. 14).

In Fig. 11, one of the important features is that surface is parallel to W1. This prevents generation of any parasitic shocks, which would interfere with the entire phenomena of compression. has the same, significance as in Fig.` l5. Its value can be derived from Equation (24); this value is not critical. Meeting of the at surface 1100 with a cylindrical convex surface 1101 establishes an edge 1102. The position of this edge is chosen so that the oblique shock terminates at this edge. 'From then on, the channel may be a constant velocity subsonic channel or a subsonic diffusion In either case, it may be either a straight or a curved channel, depending on the desired direction 0f W2.

In Fig. 12, is so chosen that W2 is still supersonic, and therefore, the lower at surface 1200 of the wedge is parallel to W2. When the fluid reaches the narrowest portion of the channel, which is indicated by points 12111 and 1202, a normal shock appears at this portion, changing the flow to velocityWB, which is subsonic.

Y lIt may be shown that the relationship between M2 and S is expressed by `the equation The above equation is solved for S after making M2 larger than 1 for the condition illustrated in Fig. 12.

The shocks illustrated in Fig. 13 are oblique and reflected. The reilected shock may be` considered as no more than the second oblique shock, and therefore-what has been said of the oblique` shock illustrated in Fig- 1l also applies to the reflected shock illustrated in Fig. 13. The reected shock is anchored. at edge 1300. From then on, the channel is either a subsonio diffusion ora constant velocity subsonic channel. Fig. 14 merely combines everything that has been said before of all previous channels, since in this case all three shocks are present. W1 should be higher here than in any previous case to justify the use of such complex shock pattern. The invention thus discloses the novel methods of compressing uid in a centripetal supersonic compressor in which the uid is first accelerated to a supersonic or subsonic absolute velocity C1 in the stationary contra prerotation stage. The direction of velocity C1 is such that it tends to be in opposition to the direction of the peripheral velocity U1 of the first compression stage. Obviously, there cannot be an angle of 180 between C1 and U1, and this angle in practice is of the order of 160. However, what is meant by the expression that C1 tends to be in opposition to U1, is that U1 points almost in the opposite direction to C1; i.e., that the contra-prerotation stage contra-prerotates the uid with respect to U1. This increases W1 still further, which enables one to obtain very high relative velocities (Mach number of W1 of the order of 2.3) although the peripheral velocity of the compression stages is not especially high, of the order of 830 feet per second. The same principle is used between the succeeding stages, whichare contra-rotating with the result that W1 is obtained by the vectorial addition of U2 and U3 (see Fig. 7, and especially triangle W3-WZ-U3-U2), which obviously increases the value of W3 approximately 2.5 times as compared to W2, i.e., 250%. This obviously produces much more intense shocks than those which lwould have been obtainable with C1. lt is to be noted also that the supersonic compression shocks are produced only on one side of the leading edges of the blades (see Figs. ll through 14, for example), which is accomplished bymaking one plane surface, such as surface 1100 in Fig. 11, parallel to the direction of fiow of the fluid with the result that this plane surface does not produce any compression shock, and only the opposite surface of the wedge, such as plane surface 1200 in Fig. l2, is made to produce the compression shocks by being inclined at an angle to the direction of flow.

The disclosed methods also indicate that a single or a plurality of different shocks may be used for obtaining the desired degree of compression.

The specification refers to stator 69 (Fig. l) as a stationary contra-prerotation stage. This term, as used in the specification and claims, means a stationary device inducing an absolute exit velocity C1 which has a radial, centripetal flow component C1-1 and a peripheral component Cu1 (see Fig. 7). Cu1 has the direction which is directly opposite to the direction of the peripheral velocity U1, and it is for this reason that stator 69 is referred to as the contra-prerotation stator, i.e., it is the Istator which produces velocity component C111, contra to 1. fore, Cu1 is contra to the direction of rotation.

It should be understood also that the flow of the fluid being compressed, obviously, takes place only when an external power is furnished to the two compressor shafts 31 and 43. This is discussed at great length throughout the preceding portion of the specification The claims also refer to the centripetal flow stages. Such flow need not be a purely radially directed centripetal ow, such as C11, but also includes such flows as C0, C1, W1, C2, W2, W3, etc., may (C1, W1, etc.) or may not (C0, W4, C11, Cm) include a peripheral component, but all of which include the centripetally-directed purely radial component.

What is claimed as new is:

1. A centripetal compressor for dynamically compressing a working fluid, said compressor comprising a stationary, centripetal flow contra-prerotation stage having two side-walls and a plurality of cambered blades mounted between said side-walls; said blades being shaped to define acceleration channels for accelerating said uid, at the exit from said stationary stage, to a maximum absolute velocity of said fluid within said compressor when said compressor is in operation; and a plurality of centripetal ow compression stages, each stage having a. plurality of cambered blades defining respective centripetal flow channels; said blades having wedge-shaped leading portions, formed by first and second flat surfaces, for compressing said fluid by means of supersonic shocks when said compressor is in operation; said first surface being a leading surface and said second surface being a lagging surface of said blades, said lagging second surface being parallel to the direction of flow of said fluid at the entry of said fluid into the respective compression stage, said direction of flow being represented by the supersonic relative velocity of said fluid at the entry into the respective compression stages when said compression stages are rotated at their normal, operating angular velocities, said leading surface creating at least an oblique supersonic compression shock in said fluid, said shock extending diagonally into the flow channel from the tip of said wedge-shaped leading portion to the inner portion of said second flat, lagging surface.

2. A centripetal compressor for dynamically compressing a fluid, said compressor comprising a stationary centripetal flow contra-prerotation stage having cambered blades positioned between two sidewalls, the adjacent blades and said side-walls forming, or defining a plurality of flow channels, each channel including a first portion for accelerating said fluid to a centripetally directed subsonie velocity and alsecond portion, following said first portion, for accelerating said fluid to a centripetally directed supersonic velocity, and at least two concentrically mounted contra-rotatable supersonic compression stages, each compression stage having a plurality of cambered blades having wedge-shaped leading portions, each wedgeshaped portion including a first, leading flat surface and a second, lagging fiat surface. said first surface forming a sharp angle with the relative velocity of said fluid at the entry into the respective compression stages, for producing supersonic compression shocks in said fluid when said compression stages are rotated at their normal, operating, angular velocities.

3. The centripetal compressor as defined in claim 2 in which said second surface makes a sharp angle with said first surface, and a third cylindrical surface meeting said second surface and forming an obtuse angle with said second surface, said first surface producing an oblique compression shock and said second surface producing a reflected compression shock.

4. A centripetal compressor for dynamically compressing a fluid with they aid of supersonic shocks created in said fluid, said compressor including a stationary centripetal flow contra-prerotation stage having two opposing side-walls and a plurality of blades uniformly dis- U1 represents the rotation component, and, there- 4' tributed around the periphery 0f Said Side-Walls, the OP posing surfaces of said blades and of said side-walls defining a plurality of centripetal flow channels, said flow channels having a subsonic acceleration portion for said fluid beginning at the outer periphery of said stage, said subsonic acceleration portion blending into a centripetal flow supersonic acceleration region for said fluid terminating at the inner periphery of said stage, said supersonic region being defined by a plurality of supersonic nozzles uniformly distributed around the periphery of said stationary stage for discharging said fluid at an absolute velocity C1, said nozzles being formed by the opposing surfaces of said blades adjacent to the inner periphery of said stage, and a centripetal flow compression stage rotatable with a peripheral velocity U1, said velocity C1 having a purely radial component Cr1 and a peripheral cornponent Cu1 having the opposite direction to that of U1.

5. A centripetal compressor for compressing a uid, said compressor including a stationary centripetal flow contra-prerotation stage followed by a centripetal ow compression stage receiving a fluid at a supersonic relative velocity W1 when said compression stage is rotated at its normal, operating angular velocity, said compression stage including a plurality of blades having wedgeshaped leading edges producing oblique shocks in said fluid, each of said edges being formed by an intersection of a first fiat leading surface and .a second flat lagging surface, said second lagging surface being parallel to the direction of the relative velocity W1, and said second surface being parallel to W2, where W2 is the velocity of said fluid upon emergence vfrom said shock, the two surfaces forming a sharp angle, the magnitude of said angle being a function of the relative local Mach number Mwn at the entry into said compression stage.

6. A centripetal compressor for compressing a fiuid with the aid of supersonic acoustic shocks, said compressor including a stationary centripetal flow contra-prerotation stage followed by a centripetal fiow compression stageV capable of receiving said fluid at a supersonic relative velocity W1, said stage including aV plurality of blades, the opposing surfaces of said blades defining flo-W channels, each blade having a leading edge formed by the intersection of a first leading plane surface and a second lagging plane surface, the second, or lagging, surface being parallel to the direction of said relative velocity W1, and the first surface being at an `angle with respect to the first surface, said first leading surface, upon encountering said fluid, producing an oblique supersonic compression shock in said iiuid, said shock extending, -from the edge of one blade to the inner portion ofl the second surface of adjacent, upstream blade, said angle having any value capable of producing said oblique shock, and being a function of the relative local Mach number.

7. A supersonic centripetal compressor having at least one centripetal flow compression stage, said stage comprising two side-walls, a plurality of blades mounted between said side-walls and spaced from each other and proportioned to receive a compressible `iiuid at a relative supersonic velocity W1, the opposed surfaces of each pair of adjacent blades and said side-walls forming four boundaries of a flow channel, each channel beginning with a first flat surface, on a corresponding blade, parallel to said relative velocity W1, said first surface merging with a second convex surface, the junction line between said first and second surfaces producing an edge normal to the direction of flow of said iiuid through said channel and extending across the depth, or width, of said channel, and the opposed surface of said channel beginning with a leading edge produced by the junction of a third, leading flat surface on an adjacent blade with a surface on said adjacent blade which is parallel to W1, said leading third fiat surface being inclined to said Velocity W1 for producing an oblique shock in said fluid when said' fluid enters said compression stage at said relative velocity W1, said oblique shock terminating at substantially said transverse edge.

8. The supersonic centripetal compressor as defined in claim 7, in which the remaining portion of said channel is defined by said convex surface on one side, and a concave surface on the opposing side, said convex and concave surfaces, in combination with the two side-wall surfaces, completing the boundaries o-f said channel, said remaining portion of the channel having a cross-sectional area for producing constant velocity channel.

9. A centripetal compressor for dynamically compressing said fluid, said compressor comprising a stationary centripetal flow contra-prerotation stage at the entry of said fluid into said compressor, said stage having means for centripetally accelerating the flow of said iiuid through said stage, and at least two contra-rotatable centripetally compressing compression stages, each compression stage having -a plurality of cambered blades, each blade having a leading flat surface and a lagging fiat surface to produce a wedge-shaped leading portion of the blade, the leading fiat surface producing at least an oblique supersonic compression shock in said fiuid upon its entry into the respective compression stage when said compressor is in operation.

10. A centripetal compressor as defined in claim 9, in which said compression stages also include the additional means for producing a normal supersonic compression shock in said fluid.

1l A centripetal compressor for dynamically compressingV a fiuid, said compressor comprising a stationary centripetal ow ncontra-prerotation -stage having a plurality of cambered blades uniformly distributed around the periphery of said stage, said blades being mounted between two side-walls, said blades and said side-walls defining a plurality of centripetal flow channels for accelerating said'iiuid, at least first and second concentrically mounted supersonic compression stages, means for rotating said first stage in a clockwise direction and said second stage in a counter-clockwise direction at angular velocities having a fixed ratio with respect to each other, and a plurality ofA cambered blades and side-Walls for supporting said blades in each compression stage, said last-recited blades having wedge-shaped leading portions formed by leading and lagging flat surfaces, said leading' surface producing an oblique supersonic shock in said fluid, and a ridge portion at the downstream end of said flat lagging surface for anchoring a reflected supersonic shock produced by said lagging surface.

12. A dynamic compressor for compressing a fiuid in a centripetal direction with the aid of supersonic acoustic shocks, said compressor comprising a stationary centripetal flow contra-prerotation stage for accelerating said fluid in a centripetal direction, and at least two centripetal liow compression stages rotatable in two opposite directions, each stage having a plurality of cambered blades having wedge-shaped leading portions formed by first and second intersecting surfaces, said first surface being a leading surface and said second surface being a lagging surface, the second surface being p-arallel to the direction of flow of said fluid along said second surface, and said rst surface forming a sharp angle with said second surface for producing an'oblique shock in said fluid, and a third concave surface meeting said first surface and forming an obtuse angle with said first surface and also forming a ridge at the line of intersection with said first surface for anchoring -at said ridge a reflected shock produced by the second surface in said iiuid.

13. A centripetal compressor comprising a first rotor including a first side-'disc and a first hoop-ring connected to said first side-disc, a first set of blades, one end of said first set of blades being connected to said first hoop-ring, a first hollow cylinder having first and second ends, said first end being connected to the other end of said first set of blades, said first hollow cylinder being connected to a source of power for rotating said first rotor, first means rotatively supporting said first side-disc; and a second rotor including a second -side-disc adjacent to said first side-disc and a second hoop-ring connected to said second side-disc, a second set of blades, one end of said second set of blades being connected to saidsecond hoopring, a second hollow cylinder connected to the other end of said second set of blades, said second hollow cylinder being connected to a source of power for rotating said second rotor, and second means rotatively supporting said second side-disc, said first and second rotors being in concentric relationship with respect to each other.

14. A centripetal compressor comprising at least first and second hollow cylinders including, respectively, first and second centripetal fiow compression stages integrated, respectively, into said cylinders, and first through fourth side-discs and first through fourth bearing means for supporting said cylinders and said side-discs in rotative and concentric relationship with respect to each other.

l5. A centripetal compressor for dynamically compressing a fiuid, said compressor including a first, or outer, centripetal flow compression stage having a plurality of centripetal compression channels, each channel having a converging upstream portion terminating in a throat, said upstream portion being defined by two sidewalls and first and second fiat transverse walls both inclined in the direction of rotation of said first compression stage and also converging toward each other; the downstream ends of said first and second walls terminating at and defining said throat; said first wall being the lagging wall and said second wall being the leading wall when viewed in the direction of rotation of said compression stage; a stationary centripetal flow contrapre-rotation stage surrounding the outer periphery of said first compression stage; said stationary stage having a plurality of How-accelerating channels having downstream portions turned, or cambered, in the direction opposite to the direction of rotation of said first compression stage; the degree of the inclination of the respective first and second transverse walls and the degree of turning of the flow-accelerating channels in said stationary stage being proportioned to discharge said finid from the stationary stage and into the first compression stage at a supersonic relative velocity parallel to the leading transverse wall of said compression channel and at an angle to the lagging transverse wall of the compression channel when said compression stage is rotated at its normal, operating, peripheral velocity, said lagging transverse wall producing at least an oblique shock in said fluid upon its entry into said rst compression stage.

16. The compressor as defined in claim in which the downstream end of said second wall merges into a convex transverse wall, along a first transverse line, and the downstream end of said first wall merges into a concave transverse wall, along a second transverse line, the two transverse lines defining the position of said throat, said throat producing a normal shock in said fluid.

17. The compressor as defined in claim 15 which also includes a transverse, leading convex wall making a first junction with the downstream end of the first, or leading, wall; and a transverse, lagging concave wall making a second junction with the downstream end of the second, or lagging, wall; said oblique shock extending from the upstream edge of said lagging fiat wall to said first junction, said first junction producing a reflected shock in said fiuid, said refiected shock following said oblique shock and extending from said first junction to said second junction.

18. The compressor as defined in claim 15 which also includes a transverse, leading convex wall making a first junction with the downstream end of said fiat leading wall, and a transverse lagging concave wall making a second junction with the downstream end of said fiat lagging wall, said two junctions and adjoining surfaces of the two side-walls determining and defining the throat portion of said channel, said oblique shock extending from the upstream edge of said lagging fiat wall to the downstream portion of said leading fiat wall, said lastmentioned wall producing a reflected shock following said oblique shock and terminating at the second junction, said reflected shock being followed by a normal shock located at the throat of said compression channel.

19. The compressor as defined in claim 15 which also includes a second compression stage following said first stage, said second stage having a plurality of compres sion channels possessing the same compression means as the compression means of the channels of the first stage, and means for rotating said first and second compression stages at two opposite angular velocities having a fixed velocity ratio with respect to each other.

20. The compressor as defined in claim 15 in which said compression channel, on the downstream side of said throat, has a constant cross-sectional area for conveying said fluid at constant subsonic speed through the downstream portion of said compression channel.

21. The compressor as defined in claim 15 in which the compression channels of the first compression stage also include downstream portions each defined by said side-walls and a first, leading, convex, transverse wall on one side and a second, lagging, concave, transverse wall on the opposed, or other, side; said convex and concave walls being cambered in the direction of rotation of said first compression stage; a second compression stage; means for rotating said first stage at a peripheral velocity -j-Ul and said second stage at a peripheral velocity nU2 having a fixed ratio to -i-Ul; a plurality of compression channels in said second stage, each having an upstream portion defined by two side-walls and inwardly converging toward each other leading and lagging flat transverse walls inclined in the direction of rotation of said second stage, the degree of camber, or curvature, of said first convex wall and of said second concave wall of the first stage and the degree of inclination of the flat walls of the second stage being proportioned to discharge said fiuid from said first stage and into said second stage at a relative entry velocity into said second stage parallel to the leading fiat wall of the second stage and at an angle to the fiat lagging wall of the second stage when said means is rotating said first and second stages at their respective normal operating peripheral velocities -I-U1 and -U2, said fiat lagging wall of the second stage producing at least an oblique shock in said fiuid upon the entry of said iiuid into said second stage.

22. The compressor as defined in claim 2l in which the downstream portion of the compression channels of the second compression stage are shaped as diffusion channels.

23. The compressor as defined in claim 15 in which the downstream portions of the flow-accelerating channels of the stationary stage terminate in supersonic nozzles.

24. A centripetal fiow compressor including at least first and second concentric rotors, each rotor including first and second side-discs, said side-discs each having an axial opening for accommodating bearing means for ro-` tatively supporting said first and second side-discs on a frame of said compressor, and a hollow mechanical member axially separating and inter-connecting the outer peripheries of said first and second side-discs, said member including at least one centripetal flow compression stage rotatable with its rotor.

25. A centripetal compressor comprising at least first and second contra-rotatable rotors; each of said rotors having: a first side-disc rotatively supporting one end of said rotor, a second side-disc rotatively supporting the other end of said rotor, at least one centripetal How compression stage having a plurality of blades oriented to produce centripetal compression of fiuid when said compressor is in operation, said blades being supported by and being coupled to said first side-disc at one end of said blades, and a hollow rotatable member support ing with its adjacent end, and being coupled to, the other end of said blades, said hollow member extending along the rotational axis of said compressor and being coupled to the second side-disc of said rotor, and bearing means for rotatively supporting said first and second Side-discs.

References Cited in the file of this patent UNITED STATES PATENTS 1,075,300 Moss Oct. 7, 1913 2,258,793 New Oct. 14, 1941 2,258,794 Way Oct. 14, 1941 2,318,990 Doran May 1l, 1943 2,344,366 Price Mar. 14, 1944 2,357,778 Beaver Sept. 5, 1944 2,417,693 Lange Mar. 18, 1947 2,435,236 Redding Feb. 3, 1948 2,471,892 Price May 3l, 1949 2,579,049 Price Dec. 18, 1951 2,628,768 Kantrowitz Feb. 17, 1953 2,659,528 Price Nov. 17, 1953 2,689,681 Sabatiuk Sept. 21, 1954 2,727,679 Pica Dec. 20, 1955 2,756,561 Morley July 31, 1956 2,804,747 Pavlecka Sept. 3, 1957 FOREIGN PATENTS 594,203 Great Britain Nov. 5, 1947 813,337 France Feb. 22, 1937

Images