Read Etsion.JulyTT02 text version

Vol. 45 (2002), 3, 430-434


A Laser Surface Textured Hydrostatic Mechanical Seal

I. ETSION (Fellow STLE) and G. HALPERIN Technion Dept. of Mechanical Engineering Haifa 32000, Israel


A mechanical seal subjected to laser surface texturing over an annular portion of one of its mating rings is theoretically and experimentally investigated. The partial surface texturing provides a mechanism for hydrostatic pressure build up in the sealing dam similar to that of a radial step or face coning. Optimization of the surface texturing parameters to obtain maximum hydrostatic pressure effect is performed on a theoretical model. A test rig that allows friction torque and temperature measurements in a back-to-back double seal arrangement is used to validate the surface texturing effect. It is found that optimally textured seals generate substantially less friction and heat. Moreover, a simple unbalanced seal, limited in its pressure capacity, can be easily transformed by surface texturing to an equivalent balanced seal with much higher pressure capability. KEY WORDS Mechanical Seals; Hydrostatic Seals; Surface Texturing INTRODUCTION High pressure mechanical seals rely on hydrostatic pressure effects to generate fluid film stiffness that is essential for safe and reliable operation of the seal. The hydrostatic pressure build-up and the resulting axial stiffness is due to an increasing resistance to the fluid flow along its path from the high pressure to the low pressure side of the sealing dam. Such increasing resistance can be the result of a decreasing clearance in the form of a radial step for example, as described by Cheng et al. (1968) in a paper from 1968. Other means for generating axial stiffness were also sug-

Presented at the 57th Annual Meeting Houston, Texas May 19-23, 2002 Final manuscript approved January 12, 2002 Review led by William Marscher

gested and discussed in the literature e.g. various forms of orifice compensated seals (Cheng, et al., (1968), Adams, et al., (1969), Lipshitz, et al., (1978)), face misalignment (Etsion, (1978)), face coning (Cheng, et al., (1968), Metcalfe, (1978), Etsion, et al., (1980)) and wavy-tilt-dam seals (Young, et al., (1989)). Other works related to various aspects of high pressure or hydrostatic seals can be found for example in references (Doust, et al., (1986), Lebeck, (1988), Harp, et al., (1998), Divakar, (1994)). The ever increasing demand for higher pressures and speeds in modern sealing systems calls for more creativity and new ideas for better performing seals. Surface engineering offers interesting opportunities in the design of improved mechanical seals. In light of this a different approach to enhancing axial stiffness, in which the increasing resistance to flow is based on a change in the slip conditions along the flow path, was suggested by Etsion and Michael (1994) in 1994. In this paper the mechanism for axial stiffness consists of a partially porous face where the porous portion, with lower resistance to flow, is adjacent to the high pressure side of the sealing dam. It was found that an optimum partially porous seal is superior to an optimum coned face seal in terms of maximum axial stiffness for a given leakage. The practicality of material fabrication and incorporation of the porous component in the face components was, however, questioned in a discussion of this paper and is yet to be resolved. More recently Etsion et al. (1999) presented, and successfully used, laser surface texturing (LST) to enhance hydrodynamic induced axial stiffness of mechanical seals. The LST technology is based on a pulsating laser beam that, by material ablation process, generates thousands of micro pores on one of the mating surfaces. By controlling the laser beam parameters it is possible to very accurately control the diameter, depth and area density of the micro-pores and to obtain the optimum texturing that is predicted by a theoretical model. Indeed, LST has been successfully applied to mechanical seals used in operating pumps in the field (Etsion, (2000)) and has attracted attention of other researchers (Wang, et al., (2001)). Extending the potential use of LST in gas seals was also demonstrated recently (Kligerman, et al., (2001)). Compared to the practical difficulties associated with the idea of incorporating a partially porous face in a mechanical seal 430

A Laser Surface Textured Hydrostatic Mechanical Seal


mechanical face deformations the LST offers a fast and inexpensive solution that can be easily applied to existing ordinary seals to enhance their performance under high pressure conditions. The present paper aims at presenting an optimization study of a partially laser surface textured mechanical seal. It includes both the theoretical model and the experimental validation. The experimental results show how an ordinary unbalanced seal, that is limited by the manufacturer to pressures below 12 bars, can be transformed by partially LST to a high pressure seal that functions easily at double its original rated pressure. ANALYSIS Figures 1 and 2 present schematically the partial LST mechanical seal. The textured surface adjacent to the high pressure boundary extends from dp to do. The ratio is typically close to unity in mechanical seals and hence seal curvature may be neglected. This allows treating the LST surface as a collection of radial pore columns as shown in Fig. 2. Each column has a length B equal to the radial width of the sealing dam and a LST length Bp equal to (do-dp)/2. Assuming parallel seal surfaces it is sufficient, due to axi-symmetry, to deal with just one column. The hydrostatic pressure distribution over a single pore column is obtained from the dimensionless Reynolds equation:

3 P X (H X )

Fig. 1--Schematic of a partial laser surface textured mechanical seal.


3 P Z (H Z )



where the dimensionless parameters are:

P =

p-pa po -pa ;


x rp ;


z rp ;

H(X, Z) =

h(x,z) c

and pa is the low pressure boundary at the inner diameter di. The film thickness distribution, h(x,z), over a single pore with a diameter 2rp (see Fig. 2) and method of solving the Reynolds equation with its proper boundary conditions (also shown in Fig. 2) are detailed elsewhere e.g. (Etsion, et al., (1999)) and, hence, are not repeated here. Note that the dimensionless pressure boundary conditions are P=1 and P=0 at X=0 and X=B/rp, respectively. The hydrostatic opening force, W, acting to separate the seal mating surfaces is obtained by integrating the hydrostatic pressure distribution over the sealing dam area and can be related to the dimensionless average hydrostatic pressure Pav by:

Pav =

4W (d2 -d2 )(po -pa ) o i


Fig. 2--A radial pores column and boundary conditions.

(Etsion, et al., (1994)), it is relatively very easy to partially texture a portion of the face width using the LST technique. This provides a real opportunity to achieve hydrostatic induced axial stiffness in high pressure seals using a simple, very accurate and reliable technique. Contrary to expensive machining of very accurate and tiny steps, coning, waviness etc. or relying on complex thermal and

The main parameters affecting the hydrostatic average pressure are the pore area density, Sp, representing the percentage of the seal face area between do and dp (see Fig. 1) that is occupied by the pores, the pore depth over diameter ratio, = hp/2rp, the seal clearance ratio, = c/2rp, and the textured length ratio = (do - dp)/(do - di). Figure 3 presents the effect of pore density Sp on the average pressure Pav for a typical case of = 0.02 and / = 5. As can be seen the average pressure varies almost linearly with the pore density. At Sp = 0 the average pressure equals 0.5 as would be expected from a linear pressure drop across an untextured sealing dam. At Sp = 55% the average pressure increases to 0.64 amounting to a 28% increase in the opening force compared to an untextured



Fig. 3--The effect of pores density, Sp, on the dimensionless average pressure, Pav.

Fig. 5--The effect of pore depth over seal clearance ratio, /, on the dimensionless average pressure, Pav.

pendent of the seal clearance c. Hence, in the above range of values the effect of / on Pav is due to the pore depth hp alone. In the limiting case of = 0 the authors can assume a uniform pressure po over the dam portion from do to dp and then a linear pressure drop to pa at di. Accordingly the upper limit of the dimensionless average pressure Pav is: (Pav)max = + (1 - )/2 [3]

which for = 0.6 for example yields Pav = 0.8. Investigation of the effect of the sealing dam width on the average pressure Pav revealed that this effect is very small and can be neglected. Hence, the results in Figs. 3-5 are typical for any practical seal. EXPERIMENTAL SETUP AND PROCEDURE

Fig. 4--The effect of textured length ratio, , on the dimensionless average pressure, Pav.

seal. This feature of the LST hydrostatic seal is actually equivalent to reducing the seal balance ratio. It may therefore enable to transform a simple unbalanced seal to an equivalent balanced one and by that extend the range of its operating pressure capacity. Figure 4 shows the effect of the textured length ratio on the average pressure for the case of = 0.1 and Sp = 55%. The results are presented for a wide range of the ratio /, which is actually the ratio of the pore depth over the seal clearance hp/c. A clear optimum of exists ranging from about 0.55 to 0.65 as / varies from 1 to 10. As can be seen a value of = 0.6 would be a good design choice to optimize the effect of LST in any hydrostatic mechanical seal. It can also be seen from the figure that the rate of increase of Pav with / diminishes above / = 5. Figure 5 presents the effect of the ratio / = hp/c on the dimensionless average pressure for a range of clearance ratio values, = c/2rp, from 0.01 to 0.03. As can be seen the average pressure is not sensitive to variations in in this range of values. This means that for a given pore diameter 2rp the average pressure is inde-

A drawing of the test rig used for the experiments is shown in Fig. 6. It consists of a pressure vessel (Cheng, et al., (1968)) that holds two identical unbalanced seals in a back-to-back arrangement. The cavity between the seals is filled with distilled water and pressurized by means of an accumulator and nitrogen gas. A shaft (Adams, et al., (1969)) having a diameter of 32 mm is held in the chuck of a milling machine spindle and turned by the machine electrical motor at a slow rate of 750 rpm. The pressure vessel is resting on a thrust bearing arrangement (Lipshitz, (1978)) that allows free rotation of the pressure vessel when friction torque is transmitted to it from the rotating shaft via the rotor and stator rings of the two seals. Such free rotation is, however, prevented by an arm attached to the pressure vessel and pressing against a load cell that provides on-line friction torque measurement. The low speed of 750 rpm was selected to eliminate any substantial hydrodynamic effect, which can result from the surface texturing (Etsion, et al., (1999)), and thus assuring detection of hydrostatic effects only on the measured friction torque. The sealed water temperature was measured by a thermocouple attached to the outer wall of the pressure vessel. Burgmann MG1 type seals, rated by the manufacturer at a maximum operating pressure of 12 bar, were used for the test. The mating faces of each seal are made of tungsten carbide and form a

A Laser Surface Textured Hydrostatic Mechanical Seal


Fig. 7--Friction torque vs. sealed pressure for untextured and textured seals.

Fig. 6--Test rig arrangement.

sealing dam with an inner diameter of 33.7 mm and an outer diameter of 42.1 mm. The balance diameter is of 32 mm, resulting in a balance ratio of 1.23. The spring load of each seal is adjusted to 1.1 bar. Thus, the average contact pressure, pc, between the mating faces of each seal is equal to [(1.23 - 0.5) (po - pa) + 1.1] bar or in its dimensionless representation: Pc = 0.73 + (1.1 - pa)/(po - pa) Four seals were used in the experiment. Two of the seals were left untextured for base line results. The other two had their stator rings textured with the following parameters: pore depth hp = 6 µm, pore diameter 2rp = 60 µm (resulting in = 0.1) and density Sp = 55%. The textured length ratio was = 66%. These parameters were selected based on the theoretical results in Figs. 3-5 in order to maximize the hydrostatic pressure effect of the LST seals. The maximum average pressure in this case would be by Eq. [3] (Pav)max = 0.83 and in this case the dimensionless contact pressure is: Pc = 0.4 + (1.1 - pa)/(po - pa) indicating an equivalent balance ratio of 0.9 compared to the 1.23 in the standard untextured seal.

Assuming that the friction torque of the seals is proportional to the contact pressure we can expect a reduction of about 45% in the friction torque of the LST seal compared to the standard one as the pressure differential (po - pa) increases. The test procedure includes the following steps: The two identical seals were assembled in the pressure vessel, which was then closed and filled with the water. The starting sealed pressure was adjusted to zero barg and the electrical motor was turned on. The friction torque was monitored and as soon as it stabilized after about 10 minutes it was recorded and the sealed pressure was increased to 5 barg. This procedure was then repeated with steps of 1 bar increments. The test was terminated when the friction torque showed erratic behavior at about 5 Nm indicating inception of severe face contact, or when the sealed pressure reached 23 barg which is close to the 25 barg limit of the pressure vessel. Each full test was repeated at least 5 times. TEST RESULTS The measured friction torque vs. the sealed pressure is presented in Fig. 7 for the untextured base line and for the textured case. The torque values shown (average of five tests) are for one seal only and were obtained by dividing the actual torque reading from the double seal arrangement by two. As can be seen the textured seals produce substantial lower friction torque compared to the untextured seals. At a sealed pressure of 5 barg the reduction of the friction is more than 50 percent from 0.18 Nm to 0.09 Nm. The two lines diverge as the sealed pressure increases indicating larger hydrostatic effect at higher pressures, as would be expected. The test with the untextured seals had to be terminated at 12 barg due to a too high friction torque reading of 5 Nm (for the two seals or 2.5 Nm for one seal) that indicated the inception of severe face contact. Indeed the manufacturer limits the use of these highly unbalanced seals to no more than 12 bar. The corresponding friction torque of the textured seal at the 12 barg is only 0.16 Nm, hence, a reduction of more than 90 percent. The textured seals could be easily operated up to the 23 barg limit of the test and at this relatively high pressure the friction torque was only 0.5 Nm,



It was shown that a simple unbalanced seal, having a balance ratio of 1.23 and limited in its pressure capacity to 12 bar, can be easily transformed by surface texturing to an equivalent balanced seal with much higher pressure capability. The LST seals showed substantially less friction torque compared to the standard seals. As was predicted the benefit of the LST increases with increasing sealed pressure due to the greater reduction in the contact pressure between the mating faces. Good correlation was found with the predicted reduction in friction torque that can reach 50% and more. ACKNOWLEDGMENT The research reported here was supported in part by Surface Technologies Ltd. and by the German-Israeli Project Cooperation (DIP). This support as well as permission by Surface Technologies to publish the work and the help of Mr. V. Brizmer and Mr. A. Pascal in running the computer code and test rig is gratefully acknowledged by the authors. REFERENCES

(1) Adams, M. L. and Colsher, R. J., (1969), "An Analysis of Self-Energized Hydrostatic Shaft Seals," Trans. ASME Jour. Lubr. Tech., 91, pp 658-667. (2) Cheng, H. S., Chow, C. Y. and Wilcock, D. F., (1968), "Behavior of Hydrostatic and Hydrodynamic Noncontacting Face Seals," Trans. ASME Jour. Lubr. Tech., 90, pp 510-519. (3) Divakar, R., (1994), "Sintered Silicon Carbides with Controlled Porosity for Mechanical Face Seal Applications," Lubr. Eng., 50, pp 75-80. (4) Doust, T. G. and Parmar, A., (1986), "Hydrostatic Effects in a Mechanical Face Seal," ASLE Trans., 29, pp 467-462. (5) Etsion, I., (1978), "Nonaxisymmetric Incompressible Hydrostatic Pressure Effects in Radial Face Seals," Trans. ASME Jour. Lubr. Tech., 100, pp 379-385. (6) Etsion, I. And Sharoni, A., (1980), "Performance of End-Face Seals with Diametral Tilt and Coning ­ Hydrostatic Effects," ASLE Trans., 23, pp 279-288. (7) Etsion, I. and Michael, O., (1994), "Enhancing Sealing and Dynamic Performance with Partially Porous Mechanical Face Seals," Trib. Trans., 37, pp 701-710. (8) Etsion, I., Kligerman, Y. and Halperin, G., (1999), "Analytical and Experimental Investigation of Laser-Textured Mechanical Seal Faces," Trib. Trans., 42, pp 511-516. (9) Etsion, I., (2000), "Improving Tribological Performance of Mechanical Seals by Laser Surface Texturing," in Proc. 17th Int. Pump Users Symp., pp 17-22. (10) Harp, S. R. and Salant, R. F., (1998), "Analysis of Mechanical Seal Behavior During Transient Operation," Trans. ASME Jour. Trib., 120, pp 191-197. (11) Kligerman, Y. and Etsion, I., (2001), "Analysis of the Hydrodynamic Effects in a Surface Textured Circumferential Gas Seal," Trib. Trans., 44, pp 472-478. (12) Lebeck, A. O., (1988), "Contacting Mechanical Seal Design Using a Simplified Hydrostatic Model," Trib. Int., 21, pp 2-14. (13) Lipshitz, A. and Etsion, I., (1978), "A Modified Face Seal for Positive Film Stiffness," ASLE Trans., 21, pp 356-360. (14) Metcalfe, R., (1978), "Predicted Effects of Sealing Gap Convergence on Performance of Plain End Face Seals," ASLE Trans., 21, pp 134-142. (15) Wang, X., Kato, K., Adachi, K. and Aizawa, K., (2001), "The Effect of Laser Texturing of SiC Surface on the Critical Load for the Transition of Water Lubrication Mode from Hydrodynamic to Mixed," Trib. Int., 34, pp 703-711. (16) Young, L. A. and Lebeck, A. O., (1989), "Design and Testing of a Wavy-TiltDam Mechanical Face Seal," Lubr. Eng., 45, pp 322-329.

Fig. 8--Water temperature vs. sealed pressure for untextured and textured seals.

a value that was obtained with the untextured seals below 11 barg. Hence, from the point of view of maintaining the same friction torque the pressure capability of the textured seals is substantially greater than that of the standard untextured ones. Figure 8 presents the temperature of the sealed water as was measured on the outer wall of the pressure vessel. As can be seen the textured seals run cooler throughout most of the range of test pressures. Here again the two lines diverge showing a larger effect at higher pressures. At 11 barg for example the textured seals run at about 31°C compared to 38°C in the untextured case. The reason for the increasing temperature at higher pressures in the textured seals case is due to the longer accumulated test time required for more pressure steps and the lack of sufficient cooling to remove the heat generated by the friction. This resulted in a steady increase of the wall temperature with test time and, hence, meaningful temperature comparison can be made only up to 11 barg. CONCLUSION A mechanical seal subjected to laser surface texturing over an annular portion of one of its mating rings was theoretically and experimentally investigated. The partial surface texturing provides a mechanism for hydrostatic pressure build up in the sealing dam similar to that of a radial step or face coning. Optimization of the surface texturing parameters to obtain maximum hydrostatic pressure effect was performed on a theoretical model. A test rig that allows friction torque and temperature measurements in a back-to-back double seal arrangement was used to validate the surface texturing effect.



5 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Microsoft Word - HYDROTEX Filter Point _FP-PET_ General Specification - 6 3 09 doc.doc
Microsoft Word - Section 7_Bill of Quantities.doc