Edward Valves. Asbestos-Free Stem Packing for High Temperature Valves. V-Rep 82-1

Transcription

1 Edward Valves Asbestos-Free Stem Packing for High Temperature Valves V-Rep 82-1

2 By R. L. Schweitzer, Supervisory Development Engineer, Flow Control Division and D. A. Goodman, Former Development Engineer, Flow Control Division First Published 1982 International and Federal government restrictions on the use of asbestos have begun to impact valve manufacturers. Since most high pressure, high temperature valves use a stem packing which contains asbestos, it has become important that a suitable replacement material without asbestos be identified. Graphite-based packings have recently received widespread attention and many new materials have recently been proposed by packing manufacturers as a substitute for asbestosbased packings. Before a substitute packing material could be approved for Rockwell valves, it was necessary to test the various materials at the most extreme conditions to which they would be subjected. This article describes the test program which was conducted by the Flow control Division of Rockwell International for glove and gate valve applications. Those materials which equalled or surpassed the performance of asbestos-based packings were approved for use in Rockwell valves. Introduction There has been widespread publicity on the impact of government restrictions on the use of asbestos and asbestos-containing products, This impact was immediate and serious in industries involving high, volume use of asbestos-construction, insulation, etc. The impact on valve manufacturers has been slower in coming. Most high pressure, high temperature valves have traditionally used a packing material in the stuffing box which contains asbestos. (See Figure 1.) The publicized health hazards of asbestos are related to inhalation of free asbestos fibers. It is difficult to visualize the packing in a valve as a significant source of free asbestos fibers in the air in a power plant or any industrial installation. The packing is confined, and, if it is to perform its sealing function, there can be little wearing away of the material by valve stem cycling. For these reasons, until fairly recently, there has been little buyer resistance to asbestos-based packings in valves; however, some users already have no asbestos clauses written into their specifications. A strict interpretation of no asbestos rules required that a substitute be found. The Flow Control Division of Rockwell International undertook an extensive one and a half-year program to develop acceptable alternate packing materials which did not require the use of asbestos. It would seem that graphite-based packings, which have received widespread acceptance, would be the obvious selection as a substitute for asbestos. However, the capability of the packing to maintain a tight stem seal is such an important valve performance feature, and since asbestosbased packing has a history of decades of laboratory testing and field experience, a major change could not be made without substantial testing. To be sure, we had test data and field experience with asbestos-free packings which have graphite as the base material, Our data indicated that these packings had pressure limitations which prevented their use as across-the-board substitutes for asbestos-based packings. For example, the new Rockwell Edward Univalve uses graphitic packing material along with asbestos-based top and bottom rings. (See Figure 2, pg. 3.) Extensive testing showed this construction was essential to prevent extrusion of the graphitic material at the very high pressures at which many Univalves operate. Because of the asbestos problem, many other materials have been proposed by packing suppliers as substitutes for conventional packings. It is extremely important that these packings be tested under the most extreme temperature and pressure conditions at which they will be used. Cycle life must be determined and tests conducted to determine the stuffing box STEM, 2 GLAND, 3 GLAND FLANGE, 4 GLAND BOLTS, 5 ASBESTOS-BASED PACKING RINGS, 6 JUNK RING, 7 BONNET FIGURE 1: Rockwell Edward Valve Stuffing Box with Asbestos. 2

3 clearances required to prevent extrusion of the packing. Resistance to corrosion must also be determined. This article describes the test program which provided the basis for selection of asbestos-free packing materials and the arrangement of the packing within the stuffing box Background For many years Rockwell Edward globe valves have used a wire-reinforced asbestos-based stem packing material in fossil and nuclear power generating stations for high temperature-pressure applications. A typical stuffing box is shown in Figure 1. Customer acceptance was widespread; however, stem leakage was reported on some nuclear services where borated water was used. In 1971, a laboratory test program was conducted by Rockwell to compare the performance of several packing materials [1] 1. It was concluded that a graphite-based packing had better sealing performance and required less maintenance than the traditional asbestosbased packing. It must be noted, however, that these tests were conducted at a maximum pressure of 2500 psi (172 bar) and maximum temperature of 650 F (343 C) with limited stem cycling. To take maximum advantage of the graphite materials, it was necessary to redesign the stuffing box, and this was done during development of the new Univalve line [2]. For rotating/rising stem valves with all-graphite packing, it was discovered that at 4200 psi (290 bar) and above, the graphite would extrude between the gland and stem, causing serious leakage. This could occur at a very low number of stem cycles and in most cases the gland could not be retightened to effect a seal. Closer clearances in the stuffing box area provide some success, but this success involves increased cost and greater risk of stem scoring. There may be an acceptable compromise of these factors for low and intermediate pressures, but a better solution was necessary for high pressure service GLAND, 2 STEM, 3 GLAND BOLT, 4 DIE-FORMED GRAPHITIC PACKING RINGS, 5 ASBESTOS-BASED PACKING RINGS, 6 YOKE, 7 BONNET 1 DIE-FORMED GRAPHITIC RINGS, 2 STEM, 3 RETAINING WASHERS FIGURE 2: New Univalve Packing Arrrangement for Class 1690 and Numbers in brackets designate references listed at end of paper. FIGURE 3: Cyclic Test Fixture. FIGURE 4: Close-Clearanced Washers with Three Graphitic Rings 3

4 Several combinations of other packing materials were also tested. The final design for Class 1690 and 2680 Univalves, as shown in Figure 2, consisted of two graphite rings preformed from ribbon, with an asbestos-based ring above and below the graphite rings. The graphite packing ensures sealing at all temperatures and pressures, and the asbestos-based packing acts as a wiper and anti-extrusion ring. For Class 4500 valves, however, it was necessary to use all asbestos-based packing rings to get satisfactory packing cycle life, because at the maximum working pressure of 10,800 psi (745 bar), extrusion occurred after a few stem operations. The new stuffing box design requires significantly less packing volume than with older designs. Figures 1 & 2 compare old and new stuffing boxes for the same size and pressure class Univalve. Primary evaluations of the new packing chamber design were conducted at the maximum cold working pressure of each pressure class, because room temperature cycling at maximum pressure was found to be the worst case for packing extrusion, Additional tests at temperatures up to 1250 F (675 C) were conducted on final designs. Sixty-three different room temperature and hot tests were conducted with a minimum of 2500 room temperature operating cycles being completed on each stem size. This is more cycles than most stop valves will see during the life of the valve. Needs for packing adjustment were minimal, with gland tightening required from FIGURE 5: Types of Packing Construction lnterbraided Fiber Packing Also known as lattice-braided and cross-braided construction, this type of packing is formed by passing each strand diagonally through the packing cross-section at a 45 angle. There is thus no cover to wear through, and the packing can provide satisfactory service in high-wear applications, even at very high pressure levels. Square Braided Fiber Packing Each strand crosses over and under oppositedirection strands at a 45 angle. Especially suited for low-pressure service, particularly if the application also involves high speed rotary motion. FIGURE 6: Carbon Microstructures Graphitic Crystal Structure known as a Monoclinic C microstructure. Ten carbon atoms in crystal. Layers of atoms are free to shear along cross-hatched plane. Crystal can be, but is not always, cubic in shape. FIGURE 6 (cont.): Face-centered Cubic Construction of Carbon filaments and diamonds. Fourteen carbon atoms in crystal. No natural shear planes. Crystal is always cubic. 4

5 not at all to three times during the 2500 operating cycles. The stuffing box design developed for the new Univalve gave excellent performance, but it did not completely eliminate the use of asbestos, Increased demands for the elimination of asbestos caused the packing test program to be extended into a search for stuffing box designs and/or nonasbestos materials which would retain sealing and cyclic life equal to or better than the existing stuffing box FIGURE 7: Reference Asbestos-Free Packing Configuration of Interbraided Carbon Fiber Rings and Die-Formed Graphitic Rings 1 Interbraided Carbon-Fiber Wiper Rings, 2 Stem, 3 Die-formed graphitic rings The Test Program In the development of this asbestos-free packing, it was established that the following requirements must be met: No packing containing asbestos could be used. The packing had to be capable of sealing fluids up to 1250 F (675 C) and up to 10,800 psi (745 bar) without requiring more than absolutely minimal maintenance. These conditions (though they need not be met simultaneously) represent the maximum allowable temperature for Rockwell Type 316 stainless steel valves and the cold working pressure (CWP) for ANSI Class 4500 valves. The packing had to be able to resist attack from a wide range of fluids (such as water, steam, and so-called sour petrochemicals) without deterioration or loss in sealing and retention abilities. Valve operating torques with asbestosfree packing should be similar in magnitude to those of valves with asbestos-based packings. Armed with these requirements, a search for a suitable packing was started. Obviously, the temperature requirements precluded the use of many packings based on organic materials such as PTFE and silicone. Further, conversations with numerous packing manufacturers showed us that there would be no off-the-shelf solution to the problem; that is, no manufacturer known to us had subjected his products to the tests that we needed. However, two vendors did suggest a few plausible packing combinations, which were tested. I. Cycle Tests To test stem packing combinations cyclically, a special test fixture was designed and fabricated as shown in Figure 3. This test fixture duplicated the packing chamber design and dimensions of the new Rockwell Univalve, which is offered in ANSI pressure classes 1690, 2680, and 4500, and uses a rotating/rising stem. (Such stem motion was assumed to present much more stringent conditions for packing operation than a rising-only stem, which is standard in many larger valves. This assumption would later be verified experimentally.) A rotary electric actuator was then mounted on the fixture to provide stem motion against the test pressure and was set up to turn the stem through one complete cycle every 90 seconds. Previous experience with certain graphitebased materials indicated that packing failures (blowouts) were usually due to extrusion of the packing through the annular clearance between the valve stem and gland. This extrusion is a function of the test pressure, and is not dependent on the fluid temperature unless this temperature is high enough to cause chemical breakdown of the packing. For this reason, all of the cycle tests described in this report were performed at room temperature so that the highest rated pressure in each ANSI class could be applied to the packing combination being tested. The first area that we examined was whether sharply reduced stem-to-gland 5

6 clearances (which at present may reach a relatively large value at extremes of manufacturing tolerances) would prevent this extrusion and subsequent blowout of graphitic packing rings. This was accomplished by machining close-clearanced metal washers and installing them above and below three graphitic rings, as shown in Figure 4. Test results, however, were mixed. At the lowest test pressure of 4200 psi (290 bar), the washers were able to retain the graphitic packing. At higher pressures, however, the washers became less and less able to maintain the preload on the graphitic rings, resulting in extrusion past the washers, and ultimately, failure. Another method of retaining graphitic packing (unsuccessfully as was later found out) involved placing two rings die-formed from knitted stainless steel Wire mesh above and below two graphitic rings. The combination was marginally successful only at the lowest test pressure of 4200 psi (290 bar), CWP for Class 1690 Univalves. Higher test pressures produced blowouts of the graphitic rings within a few cycles, which was not acceptable. After these two failures, braided rope packings (see Figure 5) were investigated, These are available from many manufacturers in dozens of formulations, ranging from uncoated hemp to carbon fibers lubricated with molybdenum disulfide. The following packings were chosen for evaluation: 1) Square-braided graphite fiber packing 2) Square-braided carbon fiber packing 3) Interbraided carbon fiber packing It should be noted here that graphite and carbon fibers, despite the fact that both are forms of the element carbon, have very different physical properties. Graphite displays a crystal microstructure that is prone to shearing through the crystal s center plane (technically known as monoclinic C); carbon fibers show a face-centered cubic crystal structure which has no natural shear planes. (This is the same microstructure found in diamonds.) For this reason, carbon fibers are much stronger than graphite fibers. Figure 6 shows a comparison between these two crystal structures. There is also a considerable difference in construction between a square-braided fiber packing and an interbraided fiber packing. In both constructions, individual fibers are twisted together to form strands, In square-braided packings, the strands are then passed over and under oppositedirection strands at a 45 angle, as shown in Figure 5. This construction, is preferred where there is high-speed rotary motion and low pressure. Interbraided packings, on the other hand, are formed by passing each strand diagonally through the body of the packing at an angle of about 450. This prevents one strand from wearing down to a core as can happen with square-braided packings. First to be tested was the graphite-fiber packing above and below two pressed graphitic rings. The combination displayed acceptable performance up to about 4000 psi (276 bar). Pressures higher than this, while not forcing any failures by blowout, caused almost continual seepage of water past the packing. Retightening of the gland bolts did not stop this seepage. Similarly, the square-braided carbon fiber packing showed the ability to retain graphitic center rings at test pressures up to about 5000 psi (345 bar); thereafter, it too leaked almost constantly. A packing combination of interbraided carbon fiber end rings and graphitic center rings was then tested. This combination required one gland bolt preload readjustment which was sufficient to prevent leakage up to the 2500-cycle criterion at 4200 psi (290 bar). Subsequent testing at 6700 psi (460 bar) and 11,000 psi (760 bar) confirmed this pattern, In one test, the packing survived more than 7800 cycles at 6700 psi (460 bar) and needed only one gland preload readjustment! Such performance, even for asbestos-based packing, would have been outstanding; for asbestos-free packing, it was unprecedented. The condition of the stems after all these tests was also outstanding, appearing to be almost micro-polished in the stem packing contact area. As mentioned earlier, it had been assumed that a valve with a rotating/rising stem would be harder on the packing than would a valve with a rising-only stem. To check this, a Rockwell size 3, Class 2500 Equiwedge gate valve (rising-only stem) was modified to allow mounting of an electric actuator; two tests of the interbraided carbon fiber packing/graphitic packing combination as shown in Figure 7 (from here on, referred to as the reference 6

7 packing combination ) in this valve were then performed. At 6200 psi (428 bar) both tests were quite successful, sealing for 2500 cycles without requiring any gland bolt preload adjustment. Microscopic examinations of the stem indicated that virtually no abrasion had occurred. Once we determined that two interbraided carbon fiber packing rings could retain two graphitic rings successfully, the question was raised as to whether four carbon fiber rings (without graphitic rings) would show even better performance. Three cycle tests of this combination in the rotating/ translating stem test fixture at 4200 psi FIGURE 8: Size 1/2 Figure Univalve in High-Temperature Steam Loop FIGURE 9: Stem Corrosion Caused by Interbraided Carbon Packing (Without Zinc) (290 bar) indicated that this was not advisable. None of the three tests was able to proceed past 1000 cycles without continuous leakage and frequent gland bolt readjustments. This confirmed our hypothesis that in a four-ring packing chamber such as in the new Rockwell Univalve, the center graphitic rings perform a majority of the sealing function and the end rings primarily retain the graphitic rings. Another area of concern was that of hightemperature cycling. The vendor of the interbraided carbon packing recommended temperature limits of 650 F (345 C) in the presence of oxygen in the flow stream and 1250 F (675 C) in steam (or any other reducing atmosphere). This capability had to be verified and was. Several new production alloy and stainless steel Univalves were obtained; the valves standard fibrous asbestos and graphitic packing rings were removed, and the reference packing combination was installed as per Figure 7. Each test valve was then mounted in an externally heated steam loop capable of producing fluid temperatures in excess of 1500 F (815 C) at 2000 psi (138 bar). (A photograph of a size 1 2 Figure Univalve In this loop is shown in Figure 8.) Conditions were brought to the maximum allowed for each valve and then turned through as many as 1000 stem cycles. In no test was any steam leakage noted, nor were any gland bolt readjustments required. II. Corrosion Tests It is well known that graphitic-based packing materials can initiate galvanic corro- 7

8 sion on various steels in the presence of water and steam. This is very apparent on 400 series stainless steels, which are commonly used to make stems for high-temperature valves. This corrosion problem (which also occurs with asbestos-fiber packing rings) was solved with the addition of a sacrificial anode such as zinc dust or zinc washers between the packing rings. Because the manufacturer of the interbraided carbon packing did not have any data on corrosion of stainless steels induced by his packing, we addressed the problem with appropriate tests. These tests consisted of complete packing stacks (representing both the present and reference configurations) tied around various types of stainless steel bars such as Types 316, 416, and 17-4PH. The assemblies were then immersed in room temperature tap water and left to stand. After a period of time the packing assemblies were removed and the bars were checked for corrosion. It then became apparent that those assemblies without either zinc washers or zinc dust in the graphitic rings had initiated severe corrosion of 400 series stainless steel under the rings. In addition, severe corrosion was also observed beneath the interbraided carbon fiber rings as shown in Figure 9. The current production combination of wire-reinforced asbestos packing with zinc-dusted graphitic rings showed virtually no corrosion. The problem was solved when the vendor of the interbraided carbon fiber rings agreed to impregnate this packing with zinc dust during its manufacture. Repetition of the short-term tests indicated that the reference packing combination now had corrosion resistance at least as good as the current production packing combination. Further confirmation occurred after longterm tests simulating a wet layup situation were completed. This condition can occur in valves that are tested after assembly for backseat leakage (this must be done with little or no preload on the packing) and then capped with end protectors. The packing thus becomes saturated with no means for evaporation. If the valve is then left unused for long periods of time (as in a customer s warehouse), stem corrosion is a likely outcome unless there is protection against it. Figure 10 shows a test fixture design to simulate this action for the reference and production packing configurations. Packing chamber dimensions duplicated a size 2, Gland (2), 2 Two Pressed Graphitic Rings (All with Zinc Dust), 3 Body, 4 Pressurization Port, 5 Stem, 6 Retaining Ring, 7 Two Interbraided Carbon Rings, 8 Two Fibrous Asbestos Rings, 9 Two Pressed Graphitic Rings (All with Zinc Dust), 10 Gland Bolt (4), 11 Retaining Ring FIGURE 10 Corrosion Test Fixture 8

9 Class 1690 Univalve. After pressurization to the Class 1690 Cold Working Pressure, the test fixture was drained, capped, and left to stand for several months. At the end of the test period, an examination of the stem revealed no visible corrosion on the stem half in contact with the reference configuration and only, microscopic pitting underneath the production packing combination. This confirmed the short-term test results. III. Friction Tests As stated earlier, one of the project objectives was to develop an asbestos-free Packing system for the Rockwell Univalve that did not increase the valve s operating torques from the current packing configuration. Carbon is noted for its outstanding lubrication properties, so torque increases were not expected, but tests were run anyway for confirmation. The quantity that we use specify a packing s friction properties is C f, the coefficient of packing friction. It is defined as the tangential friction force divided by the packing pressure normal to the stem surface. The current production packing combination of wire-reinforced asbestos rings and pressed graphitic rings has been measured in the past to have a C f range of approximately to (This ranging is due to the amount of packing wear already experienced and minor variations in each ring s construction.) However, in friction tests of the reference packing combination, the coefficient of packing friction ranged from to 0.035, a 40% drop from the production packing configuration. This range was found to further decrease as cycles were accumulated: at, 1000 cycles, the range was to 0.017; at 2500 cycles, the range was to This is believed due to the micropolishing effect described earlier. Summary Due mainly to strict requirements of certain Federal agencies, some Rockwell valve customers are beginning to request that all valves be free of asbestos. For high-temperature, high-pressure valves, this presents an immediate problem: wire-reinforced asbestos fiber packing is used extensively. The Flow Control Division identified this problem and initiated a program to find a solution. Interbraided carbon fiber packing with zinc appears able to improve on the performance of fibrous asbestos packing in all areas; it has a longer service life, less friction, and identical resistance to stem corrosion and elevated temperatures. It must, however, be used as an end ring material only, pressed graphitic rings must be used between these end rings for satisfactory performance in applications up to (and including) ANSI Class Acknowledgements The authors wish to extend their thanks to E. A. Bake for his contributions to this article and for his guidance during the test program. References 1. R. D. Norden, ROCKWELL EDWARD UNIVALVE BORATED WATER STEM PACKING TESTS, Rockwell Technical Article V-Rep. 72-1, 2. E. A. Bake and R. L. Schweitzer, UNI- VALVE EVOLUTION ANOTHER ADVANCE, Rockwell Technical Article V-Rep

10 Flow Control Division Edward Valves Flowserve Corporation has established industry leadership in the design and manufacture of its products. When properly selected, this Flowserve product is designed to perform its intended function safely during its useful life. However, the purchaser or user of Flowserve products should be aware that Flowserve products might be used in numerous applications under a wide variety of industrial service conditions. Although Flowserve can (and often does) provide general guidelines, it cannot provide specific data and warnings for all possible applications. The purchaser/user must therefore assume the ultimate responsibility for the proper sizing and selection, installation, operation, and maintenance of Flowserve products. The purchaser/user should read and understand the Installation Operation Maintenance (IOM) instructions included with the product, and train its employees and contractors in the safe use of Flowserve products in connection with the specific application. While the information and specifications contained in this literature are believed to be accurate, they are supplied for informative purposes only and should not be considered certified or as a guarantee of satisfactory results by reliance thereon. Nothing contained herein is to be construed as a warranty or guarantee, express or implied, regarding any matter with respect to this product. Because Flowserve is continually improving and upgrading its product design, the specifications, dimensions and information contained herein are subject to change without notice. Should any question arise concerning these provisions, the purchaser/user should contact Flowserve Corporation at any one of its worldwide operations or offices. For more information about Flowserve Corporation, contact or call USA FLOWSERVE CORPORATION FLOW CONTROL DIVISION Edward Valves 1900 South Saunders Street Raleigh, NC USA Toll- Free Telephone Service (U. S. and Canada) Day: After Hours Customer Service US Sales Offices Phone: Facsimile: Facsimile: Visit Our Website Flowserve Corporation, Irving, Texas, USA. Flowserve and Edward Valves are registered trademarks of Flowserve Corporation. V-Rep /03 Printed in USA