Pressure-Sensing Line Problems and Solutions

H.M. Hashemian, Analysis and Measurement Services Corp. and Dr. Jin Jiang, The University of estern !ntario "ages# 1234 Improper pressure-sensing line design or installation is often found to be the cause of poor sensing system accuracy and response time. Heres how to identify and solve those pesky pressure sensor problems in short order. Sensing lines $also referred to as impulse lines% are used to ena&le the location of pressure transmitters a'ay from the process &eing measured so as to reduce the temperature effects on the transmitter(s performance and operating life. High am&ient temperatures can affect a transmitter(s mechanical components and also shorten the life of its solid)state electronics. *ocating a transmitter a'ay from the process can also reduce the adverse effects of vi&ration and facilitate access to the transmitter for replacement or maintenance. +igures , and - illustrate t'o different vie's of sensing lines. As these figures sho', sensing lines connect a pressure transmitter to the process. Depending on the application, there may &e one or t'o sensing lines for each transmitter. .oth li/uid)filled and gas) filled sensing lines are used in industrial processes. *i/uid)sensing lines typically contain the process li/uid or oil, depending on the sensing line(s design and application. 0as)sensing lines typically contain steam, air, nitrogen, or other gases, and there is sometimes a transition in sensing lines to another medium, such as oil or 'ater. A diaphragm, &ello's, or condensate pot is used in the sensing line for the transition from one medium to another.

1. Get it right the first time. 12ample of a proper pressure transmitter installation. Source: Analysis and

easurement Services

!orp.

2. Isolation valves required. The typical pressure)sensing system design uses a com&ination of isolation valves. Source: Analysis and easurement Services !orp.

Sensing lines are typically made of small)diameter $on the order of ,.3 cm to - cm% stainless steel, car&on steel, or copper tu&ing in thic4nesses of a&out - mm. Tu&ing is preferred over piping &ecause it can &e installed in one piece, reducing the possi&ility of lea4s. Sensing lines vary in length, depending on the application. They can &e as short as a fe' meters or as long as -55 or 655 meters, and average ,5 to 35 meters. .ecause the length of sensing lines affects the overall response time of a pressure)sensing system, attempts are often made to ma4e sensing lines as short as possi&le. Sensing line installations are usually designed to allo' for the lines( thermal e2pansion and vi&ration 'ithout deformation, to ensure drainage &y gravity, and to allo' the lines to vent themselves. +or li/uid)sensing lines, self)venting is accomplished &y sloping the sensing line do'n'ard so that any gas or air in the line can vent to the process. The slope of a sensing line might &e a&out ,5 cm per meter. 7f the sensing line cannot &e sloped, a high)point vent must &e provided for li/uid)sensing lines and a lo')point vent must &e provided for gas)sensing lines.

Sensing Line Problems

Sensing lines may encounter a num&er of pro&lems that can affect the accuracy and response time of the pressure)sensing system. Discussions of the typical pro&lems found in po'er plants follo'. Referen e leg boil-off. The e2ample in +igure 6 sho's ho' a reference leg &oil)off can cause sensing line pro&lems. 7n the figure, a differential)pressure transmitter is &eing used to measure the fluid level in a vessel containing 'ater at the &ottom and steam at the top. 8ormally, the transmitter is cali&rated 'ith the assumption that the reference leg is filled 'ith a 'ater column of 4no'n height. The condensate pot helps ensure that this assumption is satisfied &y condensing the steam into 'ater at a 4no'n point in

the system. During certain plant transients or accident conditions, the am&ient temperature may increase, and pressure may decrease until it causes the 'ater in the reference leg to flash to steam. This causes the level information to lose accuracy. !ne remedy is to use isolation diaphragms or isolation &ello's in the sensing lines $+igure 9%.

3. Steam! s ene. *evel measurement instrumentation must include a condensate pot if steam is present. Source: Analysis and easurement Services !orp.

4. Isolated s!stems. *i/uid)level measurement re/uires isolation diaphragms in the sensing lines. Source: Analysis and easurement Services !orp. Level measurement "roblems. *evel measurement pro&lems can occur 'hen noncondensa&le gases &ecome dissolved in the reference leg of sensing lines. 12perience has sho'n that the dissolved gases may reappear during a rapid depressuri:ation of the process &elo' a certain pressure. More specifically, dissolved gases that accumulate over time during normal operation can rapidly

come out of solution and displace 'ater from the reference leg. This reduces the reference leg level and results in an erroneously high level indication. #oids$ blo %ages$ lea%s$ and free&ing. ;oids, &loc4ages, lea4s, or free:ing in sensing lines can cause errors in pressure measurements and can also affect the dynamic response of the pressure)sensing system. Although sensing lines are usually designed to avoid these pro&lems, they still occur in industrial processes#

"oids. Air or gas entrapped in li/uid)sensing lines can cause false pressure readings, sluggish response, and e2traneous noise as a result of acoustic resonances. +or e2ample, in differential)pressure measurements, an air poc4et on the lo') pressure side can cause the pressure indication to &e higher than normal. 7t can also add a delay in the transmission of the pressure information. "urging air from voids is difficult. Though one 'ould e2pect air poc4ets to dissolve in the li/uid under the high pressures common in industrial pressure measurements, the pro&lem of voids persists.

#lockages. .loc4ages occur in sensing lines 'hen the chemicals that are used to treat the 'ater and sludge solidify or 'hen other contaminants accumulate. They also occur due to o&structions caused &y isolation and e/uali:ing valves that are improperly aligned or seated or due to sensing lines &ecoming crimped. A partial &loc4age is detrimental only to the dynamic response time of the pressure)sensing system and does not normally affect the static output of the transmitter. .ut 'hen the &loc4age completely &loc4s the line, the pressure information is totally lost.

$eakage. "ressure)sensing lines provide many opportunities for lea4age to occur. A sensing line may have a root valve, one or more isolation valves, an e/uali:ing valve, or other connections that can give rise to lea4s, especially under high 'or4ing pressures. Any significant lea4age or loss of fluid in a sensing line can cause a false pressure indication.

%ree&ing. 7n cold 'eather, free:ing can occur in fluid sensing lines if the sensing line(s heat tracing, 'hich is used to prevent free:ing of the fluid, is aged or damaged. This pro&lem can go undetected if the free:ing causes a normal operating pressure to &e loc4ed into the system.

'ommon Sensing Lines

<edundant pressure transmitters in some processes sometimes share a sensing line. The pro&lem 'ith common sensing lines is that they can cause a common mode failure if there is a lea4, &loc4age, or void in the common leg. 7n addition to common mode pro&lems, the dynamic response times of a group of pressure transmitters that share a sensing line may &e dominated &y the response time of the most compliant transmitter on the common leg. The most compliant transmitter in most cases could &e the slo'est)responding transmitter. This could cause all transmitters on the common sensing line to &e as slo' as the most compliant transmitter.

(oise from Sensing Lines

8oise arises in sensing lines &ecause of process fluctuations, vi&ration in the sensing line, acoustic resonances, steam line resonances, control system malfunctions, and resonances caused &y undissolved air poc4ets in li/uid)filled sensing lines. To reduce the effect of noise, mechanical snu&&ers are sometimes used in pressure)sensing lines. Snu&&ers reduce the effect of noise &y increasing the dynamic response time of the pressure)sensing system. Therefore, they must &e used cautiously in those cases 'here response time is important. An alternative to snu&&ers is electronic lo')pass filters 'ith ad=usta&le response times. These filters can provide any level of noise reduction> ho'ever, li4e snu&&ers, they increase the system(s response time. !ne advantage of electronic filters is that they remove not only any mechanical or acoustic noise in the system &ut also any electrical noise. Another advantage is that they can &e designed to have a precise roll)off fre/uency. The disadvantage of electronic filters is that, unli4e snu&&ers, they do not protect the sensing element of the pressure transmitter from mechanical fatigue caused &y the e2cessive high)fre/uency vi&ration that process pressure fluctuations impose.

Sensing Lines) *ffe t on Pressure +ransmitters) Res"onse +ime

The response time of a li/uid)filled sensing line has t'o ma=or components# a sonic delay and a hydraulic delay. The sonic delay corresponds to the time that it ta4es for the pressure signal to travel at the speed of sound through a completely filled $solid% sensing line from the process to the transmitter. A 'ater)filled sensing line of a&out 65 meters has a&out -, milliseconds of sonic delay. The sonic delay is also referred to as acoustic delay. The hydraulic delay in a sensing line depends mainly on the volume of fluid that must move in the sensing line in order to &ring a pressure change from the process to the transmitter. +igure 3 sho's a sensing line leading to a pressure transmitter that has a sensing element that must move a portion or all of the distance 2 in order to indicate the applied pressure. The distance 2 traveled &y the sensing element depends on the pressure transmitter(s design.

,. *as! al ulation. Transmitter compliance is determined &y the physical installation of the component and its volume. Source: Analysis and easurement Services !orp.

7n some pressure transmitters, such as some manufactured &y <osemount, the sensing element is a diaphragm that moves very little to indicate the applied pressure. 7n other transmitters, such as some manufactured &y .arton, the sensing element is a &ello's that must move an apprecia&le amount to indicate the applied pressure. The movement of the sensing element re/uires a corresponding movement of the fluid in the sensing line. Therefore, the longer the sensing line, the more time is re/uired for the fluid to move the re/uired distance and also overcome the additional resistance to flo'. The distance 2 'as used to illustrate the relationship &et'een sensing line delays and a pressure transmitter(s design characteristics. 7n reality, ho'ever, the controlling factor in a sensing line(s hydraulic delay is the volume change inside the transmitter, not the distance 2. Another controlling factor is the pressure that is re/uired to induce the volume change. The parameter that com&ines these t'o factors is the transmitter compliance, 'hich is defined as the ratio of the transmitter volume change to the pressure change that is re/uired to attain the volume change. Transmitter compliance is a characteristic parameter of a transmitter that should &e specified &y the manufacturer. Ta&le , sho's the compliances of three different pressure transmitters. +or the transmitter 'ith a larger compliance, a larger volume of fluid must move through the sensing line in order to indicate a given pressure change. +or such transmitters, the response time of the overall pressure)sensing system from the process to the transmitter output is a strong function of the sensing line(s length and diameter.

+able 1. *-am"les of om"lian e values for re"resentative "ressure transmitters. Source: Analysis and Services !orp.

easurement

+urthermore, the response time of transmitters 'ith larger compliances is more significantly affected &y any void or o&struction in the sensing line. +igure ? sho's ho' the response times of representative pressure transmitters are increased as a function of sensing line &loc4ages. This data 'as o&tained in la&oratory e2periments in 'hich a snu&&er $+igure @% 'as used to simulate sensing line &loc4ages for the tests. 7t is understood that the snu&&er may not correctly simulate the effect of a real &loc4age in a pressure)sensing line. As such, the data in +igure ? only serves as an illustration of the effect of sensing line &loc4age on transmitter response time.

.. /ifferent res"onses. The response time differences &et'een different transmitter designs 'ith a sensing line &loc4age can &e profound. This figure is &ased on la&oratory test data. Source: Analysis and easurement Services !orp.

0. Give a blo %age the old shoulder. Snu&&ers are used to simulate sensing line &loc4ages. Source: Analysis and Services !orp.

easurement

7t is o&vious from the data in +igure ? that different transmitters are affected differently &y &loc4ages, depending on their compliance value. +or e2ample, the response time of the .arton transmitter sho'n in +igure ? increases &y almost -55A 'hen the &loc4age advances to near ?3A of the diameter, 'hile the response time of the <osemount transmitter increases &y only a&out ,5A for the same amount of &loc4age. +igure B sho's po'er spectral densities $"SDs% for a pressure)sensing system that 'as tested in a la&oratory e2perimental setup 'ith and 'ithout air in the sensing line. The effect of the void in the sensing line is manifested &y a resonance on the "SD and a lo'er &rea4 fre/uency. At this lo'er &rea4 fre/uency the "SD roll)off &egins measuring that the transmitter(s response time is larger 'ith the void in the system.

1. Sensing an air lea%. The po'er spectral density $"SD% for a pressure)sensing system can identify air lea4age in a sensing line. Source: Analysis and easurement Services !orp.

+igure C compares t'o "SDs for a pressure transmitter tested in a po'er plant &efore and after the sensing line 'as cleared of a &loc4age. Clearly, the &loc4age in this case increased the transmitter(s response time &y at least an order of magnitude.

2. Line blo %ages also sensed. The po'er spectral density $"SD% from online testing of a .arton transmitter can identify a sensing line &loc4age. Source: Analysis and easurement Services !orp.

#alidation of (oise 3nal!sis +e hnique for 4nline /ete tion of Sensing Line Problems
The validity of the noise analysis techni/ue for online detection of sensing line &loc4ages has &een esta&lished &y numerous la&oratory and in)plant demonstration tests involving a variety of pressure transmitters. $See H.M. Hashemian, D*ong)Term "erformance and Aging Characteristics of 8uclear "lant "ressure Transmitters,D U.S. 8uclear <egulatory Commission, 8U<10EC<)3B3, FMarch ,CC6G%. Ta&le - sho's representative results of such tests that involved a .arton pressure transmitter. The transmitter 'as tested alone, then 'ith 65 meters of sensing line tu&ing, and finally 'ith a snu&&er in the sensing line. 7n each case, the response time of the transmitter and the attached sensing line 'as measured &y the conventional ramp method and &y the noise analysis techni/ue. The noise 'as generated for this e2periment in a la&oratory test loop that 'as designed to simulate process fluctuations for research purposes. As sho'n &y these results, the noise analysis techni/ue identifies the response time of the transmitter and its sensing line 'ith good accuracy and accounts for the effect of sensing line length and the &loc4age $simulated &y the snu&&ers% on the response time.

+able 2. Re"resentative results of validation of noise anal!sis te hnique for res"onse time testing of "ressure transmitters and asso iated sensing lines. Source: Analysis and easurement Services !orp.

Solving Sensing-Line Problems

<emedies that remove voids andEor &loc4ages in sensing lines are to periodically &lo' do'n, &ac4 fill, or drain the sensing lines. Another solution is to test or monitor for the presence of voids or &loc4ages in the sensing lines online. This can &e done &y using the noise analysis techni/ue for in)situ testing of pressure transmitters( response times descri&ed earlier. 7n fact, one of the main advantages of response time testing 'ith the noise analysis techni/ue is that its results 'ill include the effects of sensing lines. That is, any response time result for pressure transmitters that is o&tained &y the noise analysis techni/ue 'ill inherently account for the length and diameter of sensing lines as 'ell as for any &loc4ages, voids, lea4s, or free:ing that may &e present in the sensing lines.

'H.M. Hashemian (hash)ams-corp .com* is president and !+, of Analysis and

easurement Services !orp. Dr. Jin Jiang

(--iang)uwo.ca* is a professor of electrical and computer engineering at the .niversity of /estern ,ntario.