1、b= P. NASA I. mm -m GB C 3A ZE * 4 (/I *3C sz TECHNICAL NASA TM X-195% MEMQRAMDUM 1 COMPARISON OF EXPERIMENTAL 1 AND IDEAL LEAKAGE FLOWS THROUGH C. Yeb ad Reeves P. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-IDEAL LEAKAGE FLOWS THROUGH LABYRINTH
2、 National Aeronautics and Space Administration Technical Memorandum Washington, D. C. 20546 experimentally with room-temperature air for pressure differences of 0.2 to 1.4 inches of water (50 to 350 /rn) and pressure ratios of 0.9990 to 0.9999. The analytical leakage prediction method of Egli was sh
3、own to be valid for the range of test conditions. Leakage flow increased about 5 percent through both test seals when eccentricity ratio was changed from 0 to 0.72. The straight-through seal was insensitive to flow di- rection. Leakage flow for the stepped seal was 7 percent greater with flow over t
4、he step than with flow against the step. *For sale by the Clearinghouse for Federal Scientific and Technical Information Springfield, Virginia 22 15 1 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-COMPARISON OF EXPERIMENTAL AND IDEAL LEAKAGE FLOWS
5、THROUGH LABYRINTH SEALS FOR VERY SMALL PRESSURE DIFFERENCES by Frederick C. Yeh and Reeves P. Cochran Lewis Research Center SUMMARY A requirement for handling accurately metered quantities of cooling air in a NASA seal design application generated a need for the determination of seal leakage flows a
6、t small pressure differences (0.2 to 1.4 in. of water, 50 to 350 /m), and at pressure ratios of about 0.9990 to 0.9999. An experimental investigation was made to determine the leakage characteristics of three-stage straight-through and stepped labyrinth seals. The investigation was made at nonrotati
7、ng conditions, using room-temperature air. Leakage flow in both directions through the seals and the effect of eccentricity were studied. The analytical method developed by Egli proved to be valid for very small pressure differences. Based on the overall pressure difference, the leakage flows for bo
8、th the straight-through and the stepped seals increased by about 5 percent when the seal posi- tion was changed from an eccentricity ratio of 0 to 0.72. Leakage flow for the straight- through seal was insensitive to the direction of airflow. Leakage flow for the stepped seal was 7 percent greater wh
9、en the flow direction was over the step than when the flow direction was against the step. INTRODUCTION Labyrinth seals are widely used in rotating machinery to restrict or control the flow of fluids between adjacent chambers that are at different pressure levels. Studies of leakage through multiple
10、 -element labyrinth seals (refs. 1 to 4) have covered a wide range of pressure differences across the seals and a wide range of seal. geometries. However, the applicability of these studies to seals operating at very small pressure differences were not obvious. With a requirement for a labyrinth sea
11、l operating at Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-pressures as high as 100 psia (6.910 /m) and at pressure ratios as high as 0.9999, a study of the leakage characteristics was instituted. In addition to pressure djfference, the major fac
12、tors affecting leakage flow rate of labyrinth seals are the number, thickness, axial spacing, and radial clearance of the knife-edges. The various flow equations evolved to predict leakage flow rates (refs. 1 to 4) were based on specific geometric configurations and certain pressure ratio limita- ti
13、ons. The geometry of the present seal was most closely approximated by the study of reference 1. Therefore, a modified version of the analytical method was chosen as a means of correlating leakage flow rates. The purpose of this investigation was twofold. The first was to determine ex- perimentally
14、if the modified version of the analytical method is valid for predicting seal leakage flows with very low pressure differences. The second was to investigate the effect of eccentricity and flow direction on leakage through a labyrinth seal. To fulfill these objectives, two test seals were studied in
15、 a stationary (nonrotating) apparatus at room-temperature conditions. Both test seals contained three stages. One was a straight-through seal (constant-diameter land); the other had a single radial step in the land. Concentric and eccentric positions of the land with respect to the knife-edges and l
16、eakage flows in both directions through the seals were studied. The experimentally determined leakage rates were obtained at room temperature (75 F, 24 C), with air 5 5 2 pressures ranging from 30 to 100 psia (2.1X10 to 6.9X10 N/m ) and for pressure differences across the center knife-edge ranging f
17、rom 0.2 to 1.4 inches of water (50 to 350 /m). SYMBOLS area of throttling step to knife-edge distance radial eccentric displacement gravitational acceleration constant height of step in seal land number of knife-edges in labyrinth seal absolute pressure pressure difference R gas constant s axial spa
18、cing between adjacent knife-edges Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-T absolute temperature of air W weight flow rate of air a! flow coefficient y carryover correction factor A thickness of seal knife -edge 6 radial clearance between kni
19、fe-edge and land for concentric seal position E eccentricity ratio, e/6 Subscripts : d downstream exp experimental id ideal u upstream TEST SEALS The test seals are shown schematically in figure 1. Figure l(a) shows the straight- through (constant-diameter land) seal; figure l(b) shows the seal with
20、 a single radial step in the land. This step was located in an axial position between two of the seal knife-edges. This second seal will be referred to herein as a stepped seal. Each test seal had three stages. The knife-edges were on the inner part, and the land was on the outer part of the seals.
21、In this study, both the land and the knife-edges were stationary (nonrotating). In most practical labyrinth seal applications, one of these two components would rotate while the other remained stationary. At the inception of this study, the exact geometry of the seal for the design applica- tion whi
22、ch generated the need for this study had not been established. However, a decision to use a balanced-pressure labyrinth seal for this application had been made. (A balanced-pressure seal is defined as one in which a pressure difference of zero is maintained across the knife-edges of the seal. ) Also
23、 the monitoring of pressure dif- ference would be done across an internal knife-edge of the labyrinth seal, as shown by the indicated pressure tap locations of figure l(a). This latter decision was based on the fact that, during the operation of the machinery on which the design application seal. wo
24、uld be mounted, the chambers outboard of the seal on either side could be subjected to unequal and unpredictable airflow influences. Such influences could cause relatively Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-I Land diameter, 0 I Measure-
25、Dimension ment in. m I P - +- nife-edTe Seal nife-edge ring ife-edge thickness, A I Radial clearance, b 7 LLand ring (a) Straight-through seal. - - - -4- Seal pressure taps Step height, h (b) Stepped seal, Figure 1. - Schematic of test labyrinth seals. large changes in the indicated pressure differe
26、nces across the seal compared to the magnitude of the actual pressure difference across the seal. Previous studies on labyrinth seals, such as reference 1, have shown that a stepped seal was generally superior in performance to a straight-through seal. But the presence of steps in the land (fig. l(b
27、) will introduce flow disturbances that could mask the very small pressure drops between adjacent seal chambers that would exist very near balanced -pressure conditions. Furthermore, the magnitude of these disturbances would probably differ with the direction of leakage flow through the seal. Anothe
28、r concern associated with measuring very small pressure differences on the stepped seal, but not investigated in this study, was the change in pumping effect due to a change of radius on the seal land surface. Although this change in radius is small, at high rotational speeds the effect on pumping a
29、ction can be significant in comparison to the measured pressure differences. In contrast to a stepped seal, a straight-through labyrinth seal design (fig. 1 (a) could provide identical adjacent chambers that have the same flow characteristics re - gardless of the direction of leakage flow through th
30、e seal. The constant diameter of the Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Stationary (nonrotating) part of seal Rotating part of seal F) Pressure-sensing taps Outer seal balance air Test blade cooling air Inner seal balance air Chamber 4-1
31、 5; 63 7: 4 Engine - Figure 2. - Cross-sectional view of balanced-pressure labyrinth seal system for turbine cooling research engine. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-land would also avoid differences in pumping effect over the axial l
32、ength of the seal. Against these favorable factors for the straight-through seal must be weighed the inher- ently larger leakzge flow of this type of seal. In the experimental phase of this study the effects of many of the factors discussed in the preceding paragraphs with regard to the straight-thr
33、ough and stepped labyrinth seals were evaluated. The two test seals were identical in all features with the ex- ception of the step in the land of the stepped seal. Prior to the conclusion of this study, the decision was made to use the straight-through seal in the design application that initiated
34、this study. A detailed description of this design application is given in the appendix. The specific seal is shown in figure 2. TEST APPARATUS The test apparatus (fig. 3) consisted of a pressure tank and an air supply system. The test seals were installed in the pressure tank for stationary (nonrota
35、ting) testing at room-temperature conditions. Pressure Tank The pressure tank is shown schematically in figure 4. The tank was a right circular cylinder approximately 24 inches (0.61 m) in diameter and 5 inches (0.13 m) in height. The parallel base and cover and the cylindrical side of the tank were
36、 separate pieces. At assembly, the base and cover were held in position by tie-bolts and spacers. The side was sealed to the base and cover by O-rings installed in circumferential grooves in the edges of these flat plates. Several screws installed in a radial direction anchored the side to the base.
37、 Inlet ports for airflow were located at the center and near the outer radius of the base. Similarly, outlet ports for airflow were located at the center and near the outer radius of the cover. The two test seals could be mounted interchangeably in the pressure tank. The knife-edge ring was mounted
38、on the cover of the tank; the land ring was mounted on the base of the tank. The mounting of the knife-edge ring on the tank cover was such that the ring could be shifted in a radial direction to produce eccentricity. The partially assembled pressure tank with the parts of the stepped test seal in p
39、lace is shown in figure 5. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Pressure tap locations U Thermocouple Air exit Air Figure 3. - Schematic of test apparatus for seal leakage flow tests. Air Supply System The airflow diagram for the labyrinth
40、 seal leakage tests is shown in figure 3. Lab- 5 2 oratory air at 75 F (24 C) and at a supply pressure of 125 psig (8.610 N/m ) was introduced through a filter and a pressure regulator to the inlet of the rotameter which was used to meter seal leakage flow. Downstream of the rotameter, the air suppl
41、y line branched into two inlet lines which lead to the base of the pressure tank. Two air lines attached to the cover of the tank served as exits. A total of four airflow control valves, one in each of the inlet and exit lines, were used. By choosing inlet and exit lines from separate tank chambers,
42、 the air entering the tank through one of the inlet lines would always have to pass through the seal to reach the exit line. The airflow direction from chamber A to chamber B is referred to in this report as the normal flow direction. Similarly, the airflow direction from chamber B to chamber A is r
43、eferred to as the reverse flow direction. Air pressure in the tank and the quantity of airflow through the tank (and thus through the seal) were controlled by the combined manipulation of the pressure regulator and the inlet and exit control valves. Provided by IHSNot for ResaleNo reproduction or ne
44、tworking permitted without license from IHS-,-,-12.0 in. (0.305-rn) kpressure 5.0 in. (0.13 rn) Anchor screw 1 (a) Pressure tank with stepped labyrinth seal installed. (b) Detail of straight-through labyrinth seal installed in pressure tank. Figure 4. - Cross-sectional schematic of static test press
45、ure tank with test seals installed. lnstru mentation The test apparatus was instrumented to measure seal leakage flow rate, air pressure levels, and pressure differences between various locations on the test seals. Seal leak- age flow rate was measured by means of a rotameter located in the air supp
46、ly line up- stream of the pressure tank (see fig. 3). Temperature and pressure of the air were measured at the inlet to the rotameter. Pressure level of the air was measured also in chambers A and B in the pressure tank. All measurements of pressure level were made on calibrated precision pressure g
47、ages. The pressure differences across the center knife-edge of the test seal were measured by pairs of pressure taps at four locations equally spaced around the circumference of the seal in chambers C and D (see figs 3 and 6). These pressure taps were located midway between the knife-edges and were
48、flush with the surface of the knife-edge ring. The pressure difference across the seal assem- bly was measured by one pair of pressure taps located across chambers A and B (fig. 3). Each pair of pressure taps was connected to a water manometer. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Figure 5. - Static test pressure tank. Figure 6. - Labyrinth knife-edge ring mounted on pressure tank cover. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-PROCE
