Proceedings of PVP2006-ICPVT-11:, 2006 ASME Pressure Vessel and Piping Conference ‘July 23-27, 2008, Vancouver, Canada PVP2006-ICPVT11-93072 Bolt Anti-Seize Performance in a Process Plant Environment Warren Brown ‘The Equity Engireering Group ‘Shaker Heights, OH, USA iwbrown@equityeng.com ABSTRACT This paper details recent testing that was performed as an extension of eater work on nu factor and high temperature Break ‘out performance of selected an-seize poduets. Comparison is made between previous results obsined on C-Si bolts (ASTM. A193 B17) with more recent txt on 516 Statless Stel bolts (ASTM A193 BEM), The Bolt nut factor vesus temperature and the required Break ‘ul torque aller one week at elevated tempersture are detailed, In addition, common theories reganding the unsuita anti-seize containing graphite on stainless steel andthe use of milk of magnesia an antiseize are quantitatively tested by comparison with standard anseize prods. The test methods used ae designed to closely mimie actual bolt sembly in a proces plant envionment The paper. therefore presets useful information that will enable sore accurate assembly of toed flanged joins on pressure vessels and piping in any process plant environment INTRODUCTION Bolt antiseize procucts play an important role in the operation of pressure vessd and piping joints in process pants. During joint assembly they reduce the ition at the nut fave and thereby allow the user tose the desired bolt by using measurement of Torque. In adkltion, &gand.ant-seize enables pipe-iters or bolermakers to disassemble bolted joins without needing to eut bolts ‘or split mts. This greatly reduces the effort and Gime required (0 complete maintenance activities on bolted joints and results in ‘uemendons net savings, due to reduced timaround duration and associated increase in plant aailbility and productivity, Inthe pas the end user has had to rely on manufacturer supplied data or reference material, such as Bickford (1 for fommentary on the effectiveness of the different bolt aniseize Products available and tables of fetion (nut factors. However, a5 noted in Bickford, there is ¢ wide variation inthe performance of ‘Similar” antiseize products. Table | is e-produced trom Bickford Luc Marchand Thi TIRL /Ecole Polytechnique Montreal, QC, Canada luc.marchand@polymtl.ca ry LaFrance TTRL Ecole Polytechnique ‘Montreel, QC, Canada thlafr@polymtlca and demonstrates the range that ean be expected forthe nut factors of several typical ani-seize product Table 1 ~ Typical Nut Factors (from Bickford [1]) Trenton ons Fertrocsi 0 ve FettrNswne | 098 [ois Maine | at ans Neersste | 04 ai ‘Te listed values for different lubricants inthe above table are bviously obtained fiom testing, towever the limits of the applicability of such nut factors or an explanation of the westing conducted to obtain them vas not given. Such a table is, thereire, only useful if the end-user assumes the values apply t thet pplication and if they use one ofthe listed lubricans Ideally, an end user should beable to request the ant-seize performance characteristics from the manvfactuer, However, this bis proven to be an unrealistic expectation and, as will be shown later in this paper, the manufacturers themselves are offen not aware ofthe performance charactestics of their own aniseize products when applied ina process plant environment, It shouldbe noted tat this i not, genealy, the fault of the ant-scize manufacturer, They will, most appropriately, quote test results obiined fiom standardised testing. The available standardised tests (uch as SAE-JI74 [2] and 180-1604? [3) are focussed on automotive or aerospace applications and as such the thread forms and holt sizes around which the tests were designed are different to those commonly used in pressure vessel and piping bolted joint applicatiens. The results are, therefore, ‘often not applicable to process plant applications and this ea ead Inappropriate nut factors and inaccurate sexing of holt loadsThe refining and proces industries have not developed their ‘vm standardised tess and, therefore it is impossble wo sles a bolt antiseize based on comparative nut factor or disassembly performance for applications in those industries. Not only are the bolts and tightening techniques in a process plant different to automotive or aerospace aplication but the conditions over which fnteseize products are used are also differnt, For example, presire vessel and piping belied joints are not assembled in a controlled environment and so the ambient temperature during assembly can ‘ary from 40°C to +S0°C (0° to +120°F) or worse. ln ain, 2 ‘sommon practice in refineries iso tgiten a bolted joint as it heals up to between 175°C to 230°C 350°F to 4S0°F), in order to regain lost, bot load due to gasket relaxation (hottorque). So mut factor daa can he required, for one join alone, over a 270°C (S00°F) temperature range. “The various factors that affect antistize performance in the ‘process plant envionment heve not been comprehensively studied for ‘wie range of lubricants to date Is common to neglect the eect ‘of temperature, bolt size arount of lubricant applied, bolt material, joint surface finish and vation in “generically simile” lubricants For example, the guideline on pressure vessel and piping joint sembly, ASME PCC-I [4], uses a single tition Faior of 0.16, ‘wich equates to nut fat cF around 0.2 depending on bolt size, for allnon-coated bolt applications. Iwill be sen later inthis paper that, the use of a single value fy all applications may result in gross _under-loading oF overloading of joint. The lack of standardised testing for anti-seize performance led to the dovelopment of two simple test procedures at TIRLIEeole Polytechnique in Montreal, Canada during 2002, These tess were specifically aimed at developing performance test messurs for ati= seize products in process plant (pecially, oil refining) applications, The testing corsisted ofa nut-factor test, performed at five different temperatures and a Breakout tes, performed after ageing ofthe est assembly ia an oven to simulate acto hoted joint, ‘operation. The focus of the inital est program was to provide comparative test data on diffrent antiseize products using results zaered with est methods that elosely matched aetual proces pant joint assembly and operation. Results fiom the iit testing on ASTM A193-B7 studbots were presented in Brown [5], While this bolt ype is the most common wsed in the refining industy, the behaviour oF atiseize products with stainless steel materials was Known to be diferent. Since stainless sto suse in high temporatut, sd therefore critical applications in refining, the orginal text methods were recently extended t include testing oF 3165S (ASTM A193:B8M) bolts, ‘TEST METHOD SUMMARY The first series of tess were performed on 78 neh diameter UNC, ASTM AI93-B7 studbols with ASTM AI94210 nuts and ASTM F436 through-hardened washers. The second series of tess were conducted using 7/ ich diameter UNC, ASTM A193-B8M studbols with A194-8M nuts and 308 Stainless Steel washers. The nut factor tess were conducted using a load eel t© measure the applied load and a train-gauged ring spanner to measure the applied torque (Fig. 1). The bolts and nuts were lubricated at only one ed ‘and tightened though a solid metal elinder. A new bolt and nut ‘were used For each test. The einer Ws heated to maintain the es surfaces atthe requied test temperature. Forth tests conducted a emperatres above ambient, the hols were ist soaked in an oven at the tes temperature for one to two hours prior to testing, Six bal ‘were tested fr each lubricant and each temperature value. The dats collection syst allowed the recording of over $0 tst points for teach ofthe hols tested, Each of the nt factors fora given lubricant, and temperature are therefore obtained from the satisteal treatment ‘of over 300 test points [Nut Factor tests were conducted at 4 total of five different temperatures: -5°C_Q3°F), 25°C (TSF), 40°C (105°F), 100°C (210°F) and 200°C (90°F), Ten different types of lubricants, from a variety of manufacturers, and two special lubricant mixtures were ‘ested. The lubricant are identified as following: CEI » ceramie based lubricant (max. temp: 1650°C/3000°F, Nut Factor: NA.) CUI - copper based lubricant (max. temp: 982°C/1800°F, Nut Factor.0.16) CU? - copper based lubricant (ax. temp: 982°C/800°F, Nut Factor: 0.12!) GRI - Graphite based lubricant imax. temp: 1095°C/2000°F, Nut Factor: N.A.) ‘MOI - molybdenum disulfide based lubricant (max. temp: 400°C/750°F, Nut Factor: 0.10) 'MO2a-- molybdenum disulfide based lubricant (max. temp: 13168C/2400°F*, Nut Facior.: 0.08 ) 'MO2b - molybdenum disulfide based lubricant (max. temp: 1316°C/2400°F*, Nut Facor.: 0.08 *) MOS - molybdenum disulfide bised lubricant (max. ‘temp: 1316°C/2400°F*, Nut Factor. 0.16) [NII - nickel based lubricant (max. temp: 1315°C/2400°F “Nut Factor NA.) [N12 nickel based lubricant (max. temp: 1370°C/2500°F, Nut Factor: 0.153) (MHI ~ Magnesium Hydroxide & SAE 30 Oil Mix CCI ~Calcium Carbonate & SAE 30 Oil Mix Notes: {= Nut Factor for stainless stel application 23 prostated performance to 1500°F at 75°F. farther information available from manufacturer his value taken from can, itrature fists 760°C/1400°F as limitThe list values for maximum temperature lms and nut factors wore taken from the wailble manufacturer literature for the product being tested. A notaion of N.A. indicates tht the data was ‘ot readily available from the manufacturer. Ic should be notd that MO2a and MO2 were idenical products with different methods of application (MO2a = paste and MO ~ aerosol). In addtion, MOS ‘vas the manufacturer's high emperature version ofthe MO2a pase Additional measurements were conducted on the ASTM: A193 BEM bolts using two “home-made” blends, one of engine oil tnd magnesium hydroxide and « second of engine oil and calcium feurbonate. These wo products are commonly found in antacid Produits (Such as Mik of Magnesia) which ae used in refineries for high-iemperature stainless seel applications that are found to Be Aittcult to disassemble. These products are believed to enable ease of disassembly. based on industry experience. The two blends were idemitid as MHI and CCI wspectvely. ‘The Breakout tongue tests examined the effet of prolonged ‘exposure to temperature on the lubricating capabilites ofthe ani- Seize products. This test was conducted by assembling the ASTM, ‘A193:B7 bolts lbrication epplied to all contacting surfaces of the ‘ut, washer and bolt) onto a sold carbon stel (grade 1020) eylinder (Fig. 2). The ASTM A193-BEM bolts were asembled on 4 3048S ‘pliner, so tat thermal expansion did not affet the bot foad during the test” The eylinder was then placed nan oven (Fig. 3) a a high temperature for the period of one week. The temperature used forthe ‘A193.B7 bolts was held at 315°C (680°F), which represents a typical, te high temperature applicaion for these hols. Since the ASTM. A193-B8M bolts are used in higher temperature applications, the hold temperature was inereved to 485°C (880°F). Afioe a week at ‘emperature, the cslinder wis removed from the oven, allowed to «vol and then the torque recuired (© disassemble the bot fom the cylinder was recorded (using the same suain-gauged ring spanner fiom previous tests) [Breakout torgue test results were also obtained by testing ASTM 193-87 bolts that vere in the "as-received" condition (ie hot lubricated, other than a fight coating of oll Fe by the ‘manufacturing proces). These test resuls were used asa meas of Whether or not the ani-seize was effective in reducing the required Breakout torque when compared to not using ant-seze Unfortunately, due to bolt end nut galling, @ similar baseline test ould nat be conducted on the ASTM A193-B8M bols, we CLL Ot Nt Loe MON | Mor Figure 2~ Breakout Torque Test Fixtures Figure 3 Break-out Toque Test Oven NUT FACTOR TEST RESULTS ‘The bot load caresponding toa given applied torque is often caeulated fom the following equation: F=THK.D) a where: T= applied tongue (Nim, fb), D= bolt nominal diameter (00, ),F= ole load (N, Ibs). K = dimensionless "Nut Factor" This simple equation does not filly account for the effets of bolt diameter or thread pitch on the achieved load. However, itis eneraly considered sulicient for use in a process plant ‘environment, where the accuracy achieved by torque is considered to be 230%. The obtained nut factors for each lubricant and each temperature tested are shown in Fig, 4.1 is evident from the graph that the actual temperature during jpint assembly can have a significant effect on the nut factor. Forsome anti-seize products the Aiference in achieved load fora given torque applied during winter 23s compared to summer may be as high as 30% fo $0% This would ‘can that tightening a joint wih alow mit factor would resin 30% ‘o 50% eror in obtained load jn aon to the normal =30% So the cual bolt lad may be 20% to 80% lower than the desired level due to torque enor and the effect of temperature on the nut factor. Obviously if this is not accounted for during joint assembly then it will be very dificult to improve joint integrity. Since this sort of variation i undesirable it may also be a reason for choosing one Aniseie over another. For most of the antiseize products Fisted, there is a significant variation between the ambient temperaure nut factor and the hot- ‘orgue nut factor (at 200°C, 392°F), indicating a need to revise the specified torque value for a joint ding the hottorque stage of ‘ightening, However, since most joints sould not require hottorquethe resus foe B7 bolts seem to suggest that an average nut factor of round 0.16 to 0.18 would be appropriate, This compares poorly 10 the values for nut factors quoted by the manufacturers, which all tended tobe lower and would therefor lead 10a specifed torque that, ‘would underload the bolt In fic, oF the seven tested only one manufacturer quoted nit factor that was within 10% of the measured value teeny Tempore ot) Figure 4 - Nut Factor Results ~ B7 Bolts * htventy Temperaire“C CF Figure 5 - Nut Factor Results ~ BM Bolts “The results for the BEM bolts are graphed in Fig, 5. The fis noticeable difeence is thatthe nut factors for BBM bolts ae all considerably higher, with a commen ambient value of 021 for the ‘vee similar antiseize produts. The variation in nut Tatoe forthe hot-orgue ease is even more pronounced with the B8M bolts and in all eases the nut factor increased, This indicates tat the bots would be under-loaded if the same lorque value was used for hottorgue &s vas used for ambient iehtening. For example, iF IZ was used, the hattorgue bolt load would be only 67% ofthe ambient Toad withthe same torgue vale. Onee again, the manufacturer supplicd nu factors were considerably lower than the ts results BREAK-OUT TORQUE TEST RESULTS Bolt aniseie produ serve thee basic functions 1, Lubricate the bolt during assembly in oder achieve ‘more uniform preioad 2. Prevent galling of the mt, washer and bot 3. Facilitate bolt disassembly Since most aatirseize products ase an oil base, it should be possible for them to achieve even lubrication and prevent galing ‘during assembly. Therefore, the major difference between antisize products ean occur at the disassembly stage. The Breakout test Deformed in this testing is designed t) determine which antiseize product has beter performance during disassembly by measuring the peak torgue required to disassemble the bol. In addition, a visual inspection ofthe nut, washer and bolt surfaces was conducted to Look for galling nal cases it was found that there was a direct correlation between the amount of galling seen and the required disassembly torque “The Breakout torque ress forthe BT bolts are presented in Fig. 6 It can be seen that five ofthe nine ant-seize products made it easier to disassemble the bolt, by eonparison 10 the hols without ‘niseize, OF hose Five, only one lowered the requited torque by half and the other four lowered the requted tongue by approximately 25%, Another interesting point to note from this graph, is thatthe performance of the peeduets appears to be inverse the listed maximum temperature of the product, with te lowest temperature rated product performing the bes. lp ERE wal 1 i | «| i a, = Figure 6 - Broak-out Torque Results ~ B7 Bolts The break-out tongue results for the BBM bolts are shown in Fig.7. The results are not directly comparable tothe B7 results asthe assembly torgue forthe bolts was much lower, due tothe lower yield (of BM bolts. Infact, the break-out torgue results forthe majority of the BEM bolts are around 25 to 3 times the assembly torque, which isa significant increase. It canbe seen tat the trend for dhe results i similar to the 7 bolls, wit the NIZ end MO3 outperforming the GRI fubricant. However, it does seer that the NID antiscize is slighily beter than the MO2, which isthe inverse of the BY bolt results, This is possibly due tothe small pereentage of graphite in the "MO? antseze, as graphite in aniseize products is thought to reat ‘with stainless steel at high temperatures and cause galling, This theory is also supported by the fact that the NII aniscize also‘contains a small percentage of graphite and it ean be seen thatthe performance was worse than the NI produc z 3 3 Average Release Toque Val (Nm o 888888 Soe 15 2 tocar) 5 a é Figure 7 = Break-out Torque Results - BBM Bolts The testing performed on the MIE and CC1 mixtures didnot suppor the belief that they provide beter break-out performance than standard ant-seize products. The break-out torque for both of these products was comparable w the GR ant-seize, but worse than the ‘other products tested. It coald be that this diffrence in industry ‘experience and test result is ewe to other factors, suchas the fact that the base for dhe mixture was oil versus water for normal Milk of Magnesia products. This coud afect the break-out results oF could result in uch lower bok loads being achieved inthe fel, which would lead 10 an apparent (but false) improvement in break-out performance, Alternatively te laboratory testing may be missing a Secondary or longererm elect For example, there could be a 4iference in the coking behaviour or oxidation of the bolt ‘interface duc tothe Magnesitm Hydroxide tht was not found bythe short-term laborstory testing ‘As mentioned previously, there was a diect correlation between the amount of galling seen and the require break-out torque This ean be clearly seen by comparing the posiest photographs of the different products, The MOL anti-seize (Fig. 8) had the best Treak-out tongue onthe 7 bois andi is clearly evident that there is no galling and the final surace finish has a polished appearance. Conversely there is galling evident on the NII eylindes, nut and washer surfaces (Fig. 9). The diflerence was less dramatic with the BEM bolts, however it was stl apparent that mare galling exist on with the GRE ant-seize (Fig. 10) as eompared to the NIT antiseize ig. 11) atthe 9% oF yield torque value re 10 - BBM Break-out F ee Figure 11 ~ B8M Break-out Test Cylinder NI1‘CONCLUSIONS: ‘These tests demonsiate dhe need for product specific and application specific testing for all antiseize products in order to determine an appropriate nut factor valve, The tests also demonstrated the worth of having a practical, comparative test method for the Selection of an appropriate antseize product, For selecion of an antiscize, the test method must consider all ‘Significant variables that wil influence the performance of an anti= Seize in the Held, as it was shown that the performance ofthe anti= seize products tested was not consistant between assembly perfornance and disassembly perfomance. In addition, the testing didnot find a "prfeet” anti-seize product among thse tested and so ‘nowedge of the test results allows “application spectie™ decisions to be made, “The testing also demonstrated the lack of availability of stutable data from the antiseize manufucures for process plant applications. It should be nated, however, thatthe manufacturer of several of the products tested has since updated their product datasheets to include these test results. This further coafiems the willingness of the manufacturers supply test data if suitable test, method is available. It isthe process industry that should push for such standardised testing of aniseize products. Use of aniseize produets without appropriate test results will signtieanty contribute to pressure vesse and piping bolted joint leakage and extended unit tumaround periods, due to difficulty disssembling joins. The cost to the process industry per day, far ouleighs the cost of ‘comparative an-seize testing REFERENCES [1] Bickford, JH. S. Nassar 1998, "Handbook of Bolts and Bold Joints", Marvel Dekker, New York, USA, []SAE-1174, SAE International, 1998, “Targue-Tension Test Procedue for Steel Threaded Feseners~lnch Series SAE, USA [3]180-16047, 180, 2005, “Fasteners - Torqueelamp force testing” International Organization for Standardization [4] ASME PCC-1. 2000, “Guideline for Pressure Boundary Bolted Flange loin Assembly", Amzrican Society of Mechanical Engineers, NY.USA, {5} Brown, W. 2004, “ficient Assembly of Pressure Vessel Bolted Joints, Proceedings ofthe ASME PVP 2004, ASME, San Diego USA. 478, 163-168