Theory, History, and Results of Sealants for Subsea Service

Abstract

Summary This paper discusses the theories of repair sealants on ball valves andtesting done on various sealants to determine their basic characteristics. Preliminary penetration, chemical, and flow testing results are presented, along with a discussion of field problems. Introduction The use and application of sealants to ball valves (Fig. 1), both on landand offshore, is critical to their long-term operation and reliability. Inspite of this critical need, little formal testing has been done, documented, and distributed to the industry on this subject. This paper reflects a firststep on behalf of several companies to develop technology to supportstandardized guidelines for the specification and use of these products. Example of a Failed Field Application Fig. 2 shows the seat ring pulled in a 24-in. ANSI 600 ball valve operatedat 500 to 600 psi on dry gas in the Far East. The pipeline was similiar to 500miles long, with 220 miles offshore and 280 miles onshore. Roughly 300 valves(12 through 30 in.) were used in this project. As Fig. 2 shows, the sealantmaterial is so caked and hard that it completely bridged the seat ring (see the Glossary for definitions of terms). It literally supported the seat ring offthe ball in the valve. The leakage past this seat would be so prolific that anobserver would have questioned whether the valve was closed. In this particularcase, the line could not be closed off to allow the insertion of a pig in thepig trap. Fig. 3 shows what the seat ring looked like after it was removed andcleaned. Immediately inside the dark band is a small face groove that allowsdistribution of the sealant around the face of the seat. Small holes aredrilled in the interface groove to intersect a second small outer groove, whichcommunicates with the external sealant injection fittings. The sealant flowsthrough the sealant injection fittings, into the outer groove, through thesmall holes, into the face groove, and around the seat/ball interface. (Thedifficulties of flowing along this path are discussed later.) Fig. 4 shows atest laboratory report on a sample of the sealant material returned to Houston. The material was an asphaltic material that resembled tar. It dissolved inmineral spirits and completely melted at 150 to 175 degrees F. Thus, if theline had high-temperature production, the sealant would just melt anddisappear. We show this example to emphasize that this is not a minor problem. Imagine having to disassemble and repair three hundred 12- to 30-in. ballvalves along a 500-mile pipeline. This would be a disaster, especially in theunderwater portion. Sealant Chock Valves A secondary problem with sealant injection is the occurrence of check-valvefailures within injection fittings. Typical check valves within fittings allowthe sealant to push a ball off the seat. The ball is normally held against theseat by a coil spring. After the sealant passes the ball, it then passesthrough the coils of the spring, through a drilled hole, and into the outergroove of the seat ring, as discussed above. The sealant material occasionallyis so stiff that, instead of going around the ball, it pushes the ball downuntil the spring goes to stack height. Thus, the sealant flow path iscompletely closed off and the flow is stopped. Fig. 5 shows a comparative testbetween Injection Fitting X, which ports the sealant through the spring, and Injection Fitting Y, which ports ports the sealant through the spring, and Injection Fitting Y, which ports the sealant around the spring. The pile ofsealant adjacent to Fitting X was the amount passed before the fitting bridgedover. The pile adjacent to Fitting Y is a complete gun load. There was noindication of any bridging during the test. Fig. 6 shows a second test whereapplication of pressure was continued on Fitting X. The injection fitting wasdestroyed at approximately 10,000 psi. psi. It is important to note here thatthe sealant material was loaded with teflon repair chips. The sealant did nothave the light consistency of cup grease, which is usually injected into ballvalves without problems. Although penetrometer readings were not taken at thatparticular test, it was considerably softer than some of the readings seen oncommercially available sealants after 220 degrees F testing. Preliminary Testing of Commercial Products Preliminary Testing of Commercial Products To determine the basic characteristics of available sealants, sampleswere secured from all but one of the known suppliers. They are referred to as Brands A through E. Penetration Testing. The samples were tested by the procedures specified Penetration Testing. The samples were tested by the procedures specified in ASTM Spec. D 217–60T. The testing basically consisted of preparing astandardized sample and dropping a standardized penetrometer cone (Fig. 7) intothe sample. The depth to which the cone sank into the sample was measured in1/10 mm and recorded. Fig. 8 shows the basic design of the penetrometer cone. The change in the design from a 15 to a 45 degree angle penetrometer cone. Thechange in the design from a 15 to a 45 degree angle from the centerline is at areading of about 150 (0.590 × 254). The readings on the samples ranged from 220to 43. To put these in perspective, standard automotive multipurpose grease(cup grease) is perspective, standard automotive multipurpose grease (cupgrease) is specified to have a reading of 265 to 295. Children's modeling claywas tested to have a reading of 38. Table 1 summarizes the data recorded. Thesequence of testing provided 45 days of heating at 140 degrees F, 1 day at 32degrees F, and 7 provided 45 days of heating at 140 degrees F, 1 day at 32degrees F, and 7 days at 220 degrees F for the same samples. New samples werenot used for the succeeding tests. Samples D and E probably best illustrate thedifferent characteristics of the available sealant material. Sample D startsout relatively thick, but not quite as thick as modeling clay. After 140degrees F testing, Sample D softened. When removed from the oven after 220degrees testing, it was a liquid. In actual service, Sample D likely would haveprovided no sealing capabilities. When it cooled after the 220 degrees Ftesting, Sample D had a penetrometer reading of 24 and was extremely hard for asealant material (considerably harder than modeling clay). When stirredthoroughly, it became almost as soft as it was originally. Sample E wasconsiderably softer, with a penetrometer reading of 185. A reading of 171 afterthe heat testing indicates that it is formulated for stability. The finalreading of 220 after stirring indicates that Sample E becomes somewhat softerwhen stirred, but not significantly. Testing in Mineral Spirits. As the data in Table 2 show, Sealant Acompletely dissolves in relatively mild mineral spirits. It had two of theworst mechanisms: with too much heat, it becomes hard to pump; and with thewrong fluid, it may dissolve and disappear. Flowing Testing. Sealants A and E were determined to represent the extremesof available sealants for flow testing. *Now with Merriworth Services August 1991 P. 277