a gd Thermal insulation and pipe stress An often-overlooked function of insulation in piping designs is to mitigate weather effects L, ©. Pong, Peng Engineering, Houston, Texas; and TEL. Peng, The M. W. Kellogg Co., Houston, Texas Phermal insulation is mainly nsed to reduce heat loss and noise level. Tt is also wed to prevent burn injuries. However, a lesser-known yet izaportant function of thermal ingulation is to reduce pipe streze Tn 2 petrachamical/refinery complex, piping systems are vonstantly subjected to the abuses of weather and environmental changes. Occasional rain showers, for instance, can generate very high pipe stresses, Repeated occurrences can eventually lead to pipe failure. Some typieal situations when a seemingly innocent rain showar may damage a pipe are diseusted. Some field problems can actually be solved with @ simple appliea- tion of insulation. Unfortunately, the engineer who rolies caly on.a computer to design and analyze siping will mies out on this kind of common s2ne. A case history. Fig. 1 shows a piping system used to, transfer a hot ges mixture from the primary reformer to the secondary reformer in an ammonia fertilizer plant. The gas mixture was operating at approximately 41,500°F and 506 pai, The main portion of the piping was constructed with thick, internal refractory-insu- Jation to reciuce the pips metal tomperatitre to approx imately 200°F. This section of the piping is called the cold-wall portion, ia contrast to the externally insu lated hot-wail portion, whose metal temperature is lose to the fluid temnerature of 1,600°F: Carbon stee! js used for cold-wall piping, and alloy steel is used for hhot-wall piping, This plant was bnilt in the early 1970s. Most of the piping was designed using a cold-wall approach due to ‘economic benefits of using common carbon svect mate- ial and the desire to reduce thermal expansion. Hot- wall construction was used only at the piping segments connecting to the reformers, Four transition joints can- nected cold-wall pipe ta the: hot-wail pipe. After operating without problems for the first. 10 Steers etomer Fig. 4. Ammonis plat rofornerplping. years, maintenance was performed on the piping to ‘repair the refractory and to replace the hat-wall spi ciakalloy pipe. Strangely, after this revamp, the miter elbows nt poiots A and B near the qvtd-wall/hot-wall junctions developed leaking cracks about every four ‘months. The revamp eontractor was called in to inves- tigate the problem. “As exxpecied, their first etep was to inpat the system into a computer for a sfrese analysis. However, all the computer indicated was that everything was in good shapo. So the contractor modified some springs based on the computer analysis, fronically, the original design, was probably done without the help of a sophisticated computer program. After spending thousands of dol- Jars replacing the spring hangers, the system still faith- Fully failed about every four months. You ean. bet it was very frustrating for the plant engineers. After this exercise, the plant engineers decided to get help from a largo contractor, For unknown reasons, the original contractor was not called. A large contrac: tor naturally has a greater depth of engineers, The con- tractor first performed a series of heat-transfer caleu- lations to cheek the cold-wall section’s metal wall temperature. They understood that a good analysis needs good data, With the newly calculated metat wall temperature, they made a refined siress analysis of the system. Again, the computer said everything was in good order. Nothing conld be done, or needed io he done. HXDROCARRON PROCESSING 438 184114OF | agoT sel = Fetactoy avilon Het-wa Fig, 2. Coki-wall 1s hot-well junctions in thea} corel arangament sand bj wrong arangsment. However, out of professional conscience, or maybe to Justify the fee for their service, they recommended that all the spring hangers be replaced with eonstant-effart springs. Result: the same frequent failure except the system changed inta a pile of twisted spaghetti piping. ‘The constant-offort spring hangers bad a diffieuit time holding the system together. ‘The problem was later solved by a mall modifica- tion in the insulation arrangement. It all started with a casual look at photos and some casual discussions with the plant engineers. Subsequently, all the origi- nal spring hangers were reinstalled. The fancy con- stanteffort epring hangers were removed and destined to fill the warchouse, Cold-wall/not-wall junctions, Joints connecting the cold-wall pipe and hot wall pipe require special arrange- ‘ments in both pize material and insulation, Justlike a roof contractor who knows how éo treat a shingle-to- wall junction, an experienced piping contracior knows exactly how the eold-wall to hot-wall junction must be constructed, A small mistake in the detail by the roofer normally results in constant roof leakage. Fig, 2 shows two arrangements of eold-to-hot june- tions, The two do not really appear different to the inex- porienced aye; however, the conaequence is the differ ence beliveen failure and safety. Fig. 2a shows the correct arrangement, and Fig. 2b shows the olten-used wrong arrangement, ‘The original system in Fig. 1 was, constructed with the arrangement shown in Fig, 22, yet the drawing somehow had indicated Fig. 2b, Due to the revamp contractor's inexperience, the junctions were consiructed aa Fig. 2b, ennsing constant cracks of the pipe, The pictures are not obvious, but onee the tempora- ture profiles are constructed using some common sense, they become very clear. In case 2a, the pipe wall tem perature decreases gradually from 1,500°F to the desien, cold-wall temperature of 200°F. Ensure that the wall temperature at the dissimilar wold location is below SOU’F to avoid high thermal stress due to different: expansion rates hetween carbon steel and high-alloy tel Conversely, the temperatare in case 2b drops much faster than in 2a. The facal point is atthe location called the exposing point, This is the highest temperature point, 112 MYDROCARHON PROCESSING MAY 298 008 » Fig. 2. Thermal bowing sect. exposed to open air that can ke quenched by @ rain showor, In 2a, the temporature at the exposing point is about 00°F, wheress in 2b, the temperature at the expos- inugpointis close to 1,500°F. It is clewe theta much higher thermal stress will be generated in case 2b, ‘Thermal stress, The thermal stress caused by a tem- perature gradiont or discontinuity normally does uot protluce any gross distortion. Therefore, itis. often over- looked. However, if we appreciale how high a stress can be generated, we would pay more attention to it ‘Tho magnitude of the thermal stress can be rougily estimated by: S=EaT a where § = thermal stress, psi ~ expansiva rate, in/inPF B =modalus of elasticity, psi ‘f =tomporature difference, °F For carbon steel pipe, a 500°F discontinuity will mean a 105,000-psi thermal stress, ‘The stress will be ‘even greater for a stainless steel pipe due to its higher expansion rate. This kind of stress greatly exceeds the safe reference value of twice the yield strength and should not be ignored. It is difficult to estimate exactly how much ofa temperature gradient ean be generated by arain shower, but any local area with a metal tem- perature of 50°F or higher should be protected. Showers. In 2 potrochemical/refinery complex, some locations are especially susceptible to rain shower dam- age: High-temperature flange connections. The ASME L313 piping code contains # elause that stipu- lates that the design temperature of uninsulated flanges, including these on fittings and valves, can ws0 90% of the fluid temperature (par 301.3 2). Thus, some hightemperature flanges are purposely designed with- out insulation covering. Although the situation har changed since the 197s energy crisis, many uninsu- lated high-temporatore flanges are still in the field These can easily develop leaks and cracks alter repeated rain showers, It should be cautioned, how- over, that insulating these flanges now might make ‘them unacceptable to the code due to a design tempar alure increase Expansion joints, Tt could be for ease of inspection or just due to logistical problems in construction, but many expanaior joints are not insulated. Once itis built thai vay, the plat engineers will not change it, even ifit dovelops cracks, They are very often ineensed at knowing that the simple application of insulation will solve the problem. They thought that it was something they bad to suffer through dve to technological impossibilities, that prevent these joiria foom being properly constructed, High-temperature valves on the cold-wall sec- tion. Due to practicality, vaives used in the cold-wall portion of hot-fTuid applications often use hot-wall, design (meaning without internal refractory insula- tion). These valves are often lefé uninanlated for the same reasons as high-temperature flange connections. If insulation is not desirable, then at least some type of rain shield should be provided, Thermal bowing, Por thernal stresses, we are con~ cored mainly with the high-temperature areas. These ato the areas that can create high enough thermal stresses to cause cracks, However, in some systems, even though the temperature is not high, another ther- mal effect may create a different kind of problem. This, is the lesser-known hosving effect. For example, assume we have a 16-in. gas line that is not insulated and operates at 200°: During a sum- mer shower. the pipe’s top may suddenly quench to 100°F while the bottom maintains 200°F. This 100°F quench on the top produces chrinkzaga of 0.00065 inn. of pipe surface. This ebrinkxage will bend the pipe into an are with @ radius of curvature equal to R 16/0.00065 = 24,615 in. This bowing effect (Fig. 8) can potentially lift the ends of a 200-ft tong pipe up 7 in Although the actual lift will be greatly reduced by the pipe’s weight, its significance cannot be ignored. ‘Damage esused by thermal bowing is often very ghostly. Ie normally happens without anybody actually seeing it In the above example, when the shower starts, the ends move up and possibly tear off sone supports oF small conneetians. However, when the rain stops or when the temperatures even out, the pipe reiurns inno- cenily to its initial position, It leaves the damage with- clue of the cause. . out giving any The authors ‘Uang-Chuan Peng, PE, isthe presisemt of Feng Engineering, Houston, Texas, He previcusty Worked fr Tha MAW. Keliogg Co. Poster Wheeler, Brown & Root Bechtel, Taiwan Power and others, He has 30 yasrs of experience in piping stress ‘analysis anc engirasring. tes tho orginal author Of the program NUPIPE. and hae developed tho BENGS and SIMPLEX sores of pipe stress pro prams. Mr. Peng has partorrsd traubtashootiny er piping and Mas taught piping eagineoring coursos fo many couniries. Nr Per received hie 148 dogrea in machanical angiagering from Kanes ‘State University, and #s a registred professional snineur i Texas tnd Castornia. Taen-Leong (Alvin) Peng, RE, isa senior sngt ngor with Tie MN. Kellogg Co, Houston, Texas Fe has mare than 10 years of experiane In ple stress analysis anal coniputor aojteation dove oper, After working several yoars as a sess lenginesr at Kolong, Me Rong fs curranty rvaiven lathe dovalopment of @ wise scone of anginees- lng computor applications. He obtained his BS degree in mectanteal engiveering fron The Unk versity of Taxas at Bustin, and recetved his MEA degree from The Univarsity of Saint Thornes, Houston, Texas, Mets {Erepistored professianal enginogr i Texas.