Control Valve Sizing for Steam Systems

Before discussing the sizing of control valves for steam systems, it is useful to review the characteristics of steam in a heat transfer application.

Steam is supplied at a specific pressure to the upstream side of the control valve through which it passes to a heat exchanger, also operating at a specific pressure. Steam passes through the control valve and into the steam space of the equipment where it comes into contact with the heat transfer surfaces. Steam condenses on the heat transfer surfaces, creating condensate. The volume of condensate is very much less than steam. This means that when steam condenses, the pressure in the steam space is reduced. The reduced pressure in the steam space means that a pressure difference exists across the control valve, and steam will flow from the high-pressure zone (upstream of the control valve to the lower pressure zone (the steam space in the equipment in some proportion to the pressure difference and, ideally, !alancing the rate at which steam is condensing. The rate of steam flow into the equipment is governed !y this pressure difference and the valve orifice size. Should, at any time, the flowrate of steam through the valve !e less than the condensing rate (perhaps the valve is too small , the steam pressure and the heat transfer rate in the heat exchanger will fall !elow that which is required" the heat exchanger will not !e a!le to satisfy the heat load. #f a modulating control system is used, as the temperature of the process approaches the controller set point, the controller will close the valve !y a related amount, there!y reducing the steam flowrate to maintain the lower pressure required to sustain a lower heat load. (The action of opening and closing the valve is often referred to as increasing or decreasing the $valve lift$" this is explained in more detail in Tutorial %.&, $'ontrol (alve 'haracteristics$ . 'losing the valve reduces the mass flow. The steam pressure falls in the steam space and so too the steam temperature. This means that a smaller difference in temperature exists !etween the steam and the process, so the rate of heat transfer is reduced, in accordance with )quation *.&.+. Equation 2.5.3

,here. /eat transferred per unit time (, (0 1 s 2 . 3verall heat transfer coefficient (, 1 m4' 5 . /eat transfer area (m T6 . 6ean temperature difference !etween the steam and secondary fluid (4' The overall heat transfer coefficient (2 does not change very much during the process, and the area (5 is fixed, so if the mean temperature difference (T 6 is reduced, then the heat transfer from the steam to the secondary fluid is also reduced. Top

Saturated steam flow through a control valve

5 heat exchanger manufacturer will design equipment to give a certain heat output. To achieve this heat output, a certain saturated steam temperature will !e required at the heat transfer surface (such as the inside of a heating coil in a shell and tu!e heat exchanger . ,ith saturated steam, temperature and pressure are strictly related" therefore controlling the steam pressure easily regulates the temperature. 'onsider an application where steam at 78 !ar g is supplied to a control valve, and a given mass flow of steam passes through the valve to a heat exchanger. The valve is held fully open (see 9igure %.:.7 .

#f a ;<&8 valve is fitted and the valve is fully open, the pressure drop is relatively small across the valve, and the steam supplied to the heat exchanger is at a fairly high pressure (and

temperature . Because of this, the heating coil required to achieve the design load is relatively small. 'onsider now, a fully open ;<:8 valve in the steam supply line passing the same flowrate as the ;<&8 valve. 5s the valve orifice is smaller the pressure drop across the valve must !e greater, leading to a lower pressure (and temperature in the heat exchanger. Because of this, the heat transfer area required to achieve the same heat load must !e increased. #n other words, a larger heating coil or heat exchanger will !e required. 9urther reduction of the valve size will require more pressure drop across the control valve for the same mass flow, and the need for an increased heat transfer surface area to maintain the same heat output.

Fig. 6.4.1 Flow t roug a fully o!en "ontrol valve ,hatever the size of the control valve, if the process demand is reduced, the valve must modulate from the fully open position towards closed. /owever, the first part of the travel has only a small regulating effect, with any percentage change in valve lift producing a lesser percentage change in flowrate. Typically, a 78= change in lift might produce only a &= change in flowrate. ,ith further travel, as the valve plug approaches the seat, this effect reverses such that perhaps a &= change in lift might produce a 78= change in flowrate, and !etter regulation is achieved. The initial part of the control valve travel, during which this lowered control effect is seen, is greater with the selection of the larger control valves and the accompanying small pressure drop at full load. ,hen the control valve chosen is small enough to require a $critical pressure drop$ at full load the effect disappears. 'ritical pressure is explained in the Section !elow. 9urther, if a larger control valve is selected, the greater size of the valve orifice means that a given change in flowrate is achieved with a smaller percentage change in lift than is needed with a smaller control valve. This can often ma>e the control unsta!le, increasing the possi!ility of $hunting$, especially on reduced loads.

Criti"al !ressure
The mass flow of steam passing through the valve will increase in line with differential pressure until a condition >nown as $critical pressure$ is reached. The principle can !e explained !y loo>ing at how nozzles wor> and how they compare to control valves. 'onsider an almost perfect orifice, such as a convergent-divergent nozzle shown in 9igure %.:.*. #ts

shape, if designed correctly to match the upstream and downstream pressure conditions and the condition of the supplied steam, will allow it to operate at high efficiency.

Fig. 6.4.2 # "onvergent$%ivergent nozzle Such a nozzle can !e thought of as a type of heat engine, changing heat energy into mechanical (>inetic energy. #t is designed to discharge the required weight of steam with a given pressure drop, and with minimum tur!ulence and friction losses. #n the convergent section, the steam velocity increases as the pressure falls, though the specific volume of the steam also increases with the lowered pressures. 5t first, the velocity increases more quic>ly than the specific volume, and the required flow area through this part of the nozzle !ecomes less. 5t a certain point, the specific volume !egins to increase more rapidly than does the velocity and the flow area must !ecome greater. 5t this point, the steam velocity will !e sonic and the flow area is at a minimum. The steam pressure at this minimum flow area or $throat$ is descri!ed as the $critical pressure$, and the ratio of this pressure to the initial (a!solute pressure is found to !e close to 8.&? when saturated steam is passing. 'ritical pressure varies slightly according to the fluid properties, specifically in relation to the ratio of the specific heats cp 1cv of the steam (or other gaseous fluid , which is termed the adia!atic index or isentropic exponent of the fluid, often depicted !y the sym!ols $n$, $>$ or $ $. ,ith superheated steam the ratio is a!out 8.&&, and for air a!out 8.&+.

9or dry saturated steam, using )quation %.:.7-

'learly, the mass flow through the throat of a given size is at a maximum at this $critical pressure drop$. To achieve a greater flow, eithera. !. The velocity would have to !e greater, which could only !e reached with a greater pressure drop !ut this would also increase the specific volume !y an even greater amount, orThe specific volume would have to !e less, which could only !e the case with a lesser pressure drop - !ut this would reduce the velocity !y an even greater amount.

Thus, once the critical pressure drop is reached at the throat of the nozzle, or at the $vena contracta$ when an orifice is used, further lowering of the downstream pressure cannot increase the mass flow through the device. #f the pressure drop across the whole nozzle is greater than the critical pressure drop, critical pressure will always occur at the throat. The steam will expand after passing the throat such that, if the outlet area has !een correctly sized, the required downstream pressure is achieved at the nozzle outlet, and little tur!ulence is produced as the steam exits the nozzle at high velocity. Should the nozzle outlet !e too !ig or too small, tur!ulence will occur at the nozzle outlet, reducing capacity and increasing noise-

#f the nozzle outlet is too small, the steam has not expanded enough, and has to continue expanding outside the nozzle until it reaches the required downstream pressure in the low pressure region. #f the nozzle outlet is too large, the steam will expand too far in the nozzle and the steam pressure in the nozzle outlet will !e lower than the required pressure, causing the steam to recompress outside the outlet in the low pressure region.

The shape of the nozzle (9igure %.:.+ is gently contoured such that the vena contracta occurs at the nozzle throat. (This is in contrast to a sharp-edged orifice, where a vena contracta occurs downstream of the orifice. The vena contracta effect is discussed in more detail in Tutorial :.* $@rinciples of 9lowmetering$ .

Fig. 6.4.3 & e "onvergent$%ivergent nozzle 'ontrol valves can !e compared to convergent-divergent nozzles, in that each has a high-pressure region (the valve inlet , a convergent area (the inlet !etween the valve plug and its seat , a throat (the narrowest

gap !etween the valve plug and its seat , a divergent area (the outlet from the valve plug and its seat, and a low-pressure region (the downstream valve !ody . See 9igure %.:.:.

Fig. 6.4.4 & e "onvergent$%ivergent !rin"i!le in a "ontrol valve <ozzles and control valves have different purposes. The nozzle is primarily designed to increase steam velocity in order to produce wor> (perhaps to turn a tur!ine !lade , so the velocity of steam leaving the nozzle is required to remain high. #n contrast, the control valve is a flow restricting or $throttling$ device designed to produce a significant pressure drop in the steam. The velocity of steam passing out of a control valve throat will !ehave in a similar fashion to that of the steam passing out of the throat of a convergent-divergent nozzle" in that it will increase as the steam expands in the diverging area !etween the plug and seat immediately after the throat. #f the pressure drop across the valve is greater than critical pressure drop, the steam velocity will increase to supersonic in this area, as the pressure here is less than that at the throat. @ast this point, the steam passes into the relatively large cham!er encased !y the valve !ody (the low pressure region , which is at a higher pressure due to the !ac>pressure imposed !y the connecting pipewor>, causing the velocity and >inetic energy to fall rapidly. #n accordance with the steady flow energy equation (S9)) , this increases the steam enthalpy to almost that at the valve entrance port. 5 slight difference is due to energy lost to friction in passing through the valve. 9rom this point, the valve !ody converges to port the steam flow to the valve outlet, and the pressure (and density approach the pressure (and density in the downstream pipe. 5s this pressure sta!ilises, so does the velocity, relative to the cross sectional area of the valve outlet port. The relative change in volume through the valve is represented !y the dotted lines in the schematic diagram shown in 9igure %.:.&.

Fig. 6.4.5 & e "onvergent$%ivergent$"onvergent valve 'o%y ,hen the pressure drop across a valve is greater than critical, noise can !e generated !y the large instantaneous exchange from >inetic energy to heat energy in the low pressure region, sometimes exacer!ated !y the presence of supersonic steam.

Valve outlet velo"ity( noise( erosion( %rying an% su!er eating effe"t
<oise can !e an important consideration when sizing control valves, not only !ecause it creates increased sound levels !ut !ecause its associated vi!ration can damage valve internals. Special noise-reducing valve trims are availa!le !ut, sometimes, a less expensive solution is to fit a larger valve !ody than required. 'omplicated equations are required to calculate noise emitted from control valves and these are difficult to use manually. #t is usually considered that the control valve will produce unaccepta!le noise if the velocity of dry saturated steam in the control valve outlet is greater than 8.+ 6ach. The speed of sound in steam will depend upon the steam temperature and the quality of the steam, !ut can !e calculated from )quation %.:.* if the conditions are >nown (6ach 7 . speed of sound . Equation 6.4.2 ,here' . Speed of sound in steam (m 1 s +7.% . 'onstant of proportionality . Steam isentropic exponent (7.7+& - saturated, 7.+ - superheated A . 8.:%7 & the gas constant for steam (>0 1 >g T . 5!solute steam temperature (B 5 less accurate !ut useful method to estimate whether noise will !e a pro!lem is !y calculating the velocity in the valve outlet port. #n simplistic terms and for dry saturated steam, if this is greater than 7&8 m 1 s, there is a chance that the valve !ody is too small (even though the valve trim size suits the required capacity . /igher velocities also cause erosion in the downstream valve !ody, especially if the steam is wet at this point. #t is recommended that the maximum exit velocity for wet steam is :8 m 1 s in the outlet port. 5nother result of dropping steam pressure across a control valve is to dry or superheat the steam, depending upon its condition as it enters the valve. Carge degrees of superheat are usually unwanted in heating processes, and so it is useful to !e a!le to determine if this will occur. Superheated steam (and dry gas velocities, however, may !e allowed to reach 8.& 6ach in the outlet port" whereas, at the other end of the scale, liquids might !e restricted to a maximum outlet velocity of 78 m 1 s.

E)am!le 6.4.1 & e valve outlet velo"ity an% %rying * su!er eating effe"t
5 control valve is supplied with dry saturated steam from a separator at 7* !ar g and used to drop steam pressure to : !ar g at full load. The full load flowrate is 7 +88 >g 1 h requiring a B vr of ?.+. 5 ;<*& (7D valve is initially considered for selection, which has a Bvs of 78 and a valve outlet area of 8.888 :E m. ,hat is the steam velocity in the valve outletF

;etermine The degree

the of

state drying

of

the

steam

in

the

valve

outlet from

at

!ar

g.

and superheating

can !e

calculated

the following

procedure-

9rom steam ta!les, total heat (hg in the upsteam dry saturated steam at 7* !ar g . * G?G >0 1 >g 5s the supply steam is in a dry saturated state, the steam will certainly !e superheated after it passes through the valve" therefore the superheated steam ta!le should !e used to quantify its properties. 2sing the Spirax Sarco we!site steam ta!les, it is possi!le to calculate the condition of the downstream steam at : !ar g !y selecting $Superheated steam$ and entering a pressure of $: !ar g$ and a total heat (h of * G?G >0 1 >g. By entering these values, the steam ta!le returns the result of superheated steam at : !ar g with 7%.E degrees of superheat (::* B . (9urther details on how to determine the downstream state are given in Tutorial *.+ $Superheated steam$. Specific volume of superheated steam, : !ar g, ::* B is 8.+E7 ? m 1 >g (from the steam ta!le .

#t is necessary to see if this velocity is less than 8.& 6ach, the limit placed on valve outlet velocities for superheated steam. The speed of sound (6ach 7 can !e calculated from )quation %.:.*. Equation 6.4.2 5 value of 7.+ is chosen for the isentropic exponent $g$ due to the steam in the valve outlet !eing superheated. A T is is the gas the constant a!solute for steam temperature 8.:%7 of & >0 ::* 1 >g B

Therefore the speed of sound in the valve outlet-

5s the steam is superheated in the valve outlet, the criterion of 8.& 6ach is used to determine whether the valve will !e noisy.

5s the expected velocity is *?E m 1 s and a!ove the limit of *&G.& m 1 s, the ;<*& valve would not !e suita!le for this application if noise is an issue. 'onsider the next largest valve, a ;<+* (!ut with a *& mm trim . The outlet area of this valve is 8.888? m* (see Ta!le %.:.7 .

The ;<+* !odied valve will !e suita!le !ecause the outlet velocity is less than 8.& 6ach allowed for superheated steam. The same procedure can !e used to determine the conditions of the downstream steam for other upstream conditions. 9or instance, if the upstream steam is >nown to !e wet, the downstream condition might !e wet, dry saturated or superheated, depending on the pressure drop. The allowa!le outlet velocity will depend on the downstream steam condition as previously outlined in this section, and o!served in )xample %.:.*.

Erosion
5nother pro!lem is the possi!ility of erosion in the valve !ody caused !y excessive velocity in the valve outlet. #n )xample %.:.7, due to the drying and superheating effect of the pressure drop from 7* !ar g to : !ar g, the steam is in a dry gaseous state containing a!solutely no moisture and erosion should not !e an issue. Simplistically, it can !e guaranteed that the steam leaving a control valve is superheated, then *&8 m1s is an appropriate limit to place on the outlet velocity. Sometimes, when saturated steam is supplied to a control valve, it will !e carrying a certain amount of water and the steam may !e, for example EG= or E?= dry. #f it has Hust passed through a properly designed separator it will !e close to 788= dry, as in )xample %.:.7. ,ith anything more than a small pressure drop and wet steam, the steam will pro!a!ly !e dried to saturation point or even slightly superheated. #f the supply steam is dry and1or the valve encounters quite a large presure drop, (as in )xample %.:.7. , the steam steam will !e more superheated.

Equations for sizing "ontrol valves.

'ontrol valves are not as efficient as nozzles in changing heat into >inetic energy. The path ta>en !y steam throgh a valve inlet, the throatand into the valve outlet is relatively tortuous. #n a control valve a great deal more energy is lost to friction than in a nozzle and !ecause ...

. .

The outlet area of the valve !ody is unli>ely to match the downstream pressure condition. The relationship !etween the plug position and the seat is continually changing. . tur!ulence is always li>ely to !e present in the valve outlet.

#t seems that control valves of differing types may appear to reach critical flow conditions at pressure drops other than those quoted a!ove for nozzles. Aestricted flow passages through the seat of a valve and on the downstream side of the throat may mean that maximum flowrates may only !e reached with somewhat greater pressure drops. 5 !all valve or !utterfly valve may !e so shaped that some pressure recovery is achieved downstream of the throat, so that maximum flow conditions are reached with an overall pressure drop rather less than expected. 'omplicated valve sizing equations can !e used to ta>e these and other criteria into consideration, and more than one standard exists incorporating such equations. 3ne such standard is #)' %8&+:. 2nfortunately, the calculations are so complicated, they can only !e used !y computer software" manual calculation would !e tedious and slow. <evertheless, when sizing a control valve for a critical process application, such software is indispensa!le. 9or example, #)' %8&+: is designed to calculate other symptoms such as the noise levels generated !y control valves, which are su!Hected to high pressure drops. 'ontrol valve manufacturers will usually have computer sizing and selection software complementing their own range of valves. /owever, a simple steam valve sizing equation, such as that shown in )quation +.*7.* for saturated steam, is perfectly adequate for the vast maHority of steam applications with glo!e valves. 5lso, if consideration is given to critical pressure occurring at &?= of the upstream a!solute pressure, a glo!e valve is unli>ely to !e undersized.

9or simplicity, the rest of this Tutorial assumes critical pressure for saturated steam occurs at &?= of the upstream a!solute pressure. 9or example, if the pressure upstream of a control valve is 78 !ar a, the maximum flowrate through the valve occurs when the downstream pressure is78 !ar a x &?= . &.? !ar a

)qually, critical pressure drop is :*= of the upstream pressure, that is, a pressure drop ratio of 8.:*. 5s shown in the previous text, once this downstream pressure is reached, any further increase in pressure drop does not cause an increase in mass flowrate. This effect can !e o!served in 9igure %.:.% showing how, in the case of a glo!e valve, the flowrate increases with falling downstream pressure until critical pressure drop is achieved.

Fig. 6.4.6 & e mass flowrate t roug a steam valve in"reases until "riti"al !ressure is rea" e%

Sizing a "ontrol valve for a steam eat e)" anger is a "om!romise 'etween+
7. *. 5 smaller pressure drop that will minimise the size (and perhaps the cost of the heat exchanger. 5 larger pressure drop that allows the valve to apply effective and accurate control over the pressure and flowrate for most of its travel.

#f the pressure drop is less than 78= at full load, three pro!lems can occur-

;epending upon the controller settings and secondary temperature, and system time lags, $hunting$ of the temperature around the set value may occur !ecause the valve is effectively oversized" small changes in lift will cause large changes in flowrate, especially in the case of a valve with a linear characteristic. Aunning loads are often much less than the full load, and the valve may operate for very long periods with the valve plug close to its seat. This creates a ris> of wiredrawing, (erosion caused !y high velocity water droplets squeezing through the narrow orifice . ,iredrawing will result in a reduced valve service life.

The system will not control well at low heat loads, effectively reducing the $turndown$ capa!ility of the valve.

Sim!le sizing routine for glo'e valves in steam servi"e

The flow and expansion of steam through a control valve is a complex process. There are a variety of very complex sizing formulae availa!le, !ut a pragmatic approach, !ased on the $!est fit$ of a mathematical curve to empirical results, is shown in )quation +.*7.* for glo!e valves throttling saturated steam. The advantage of this relatively simple formula is that it can !e used with the aid of a simple calculator. #t assumes that critical pressure drop occurs at &?= of the upstream pressure. Equation 3.21.2 ,heres . 6ass flowrate (>g 1 h Bv . (alve flow coefficient (mI 1 h !ar @7 . 2pstream pressure (!ar a

. @ressure drop ratio .

@* . ;ownstream pressure (!ar a ,ote+ #f )quation +.*7.* is used when @* is less than the critical pressure, then the term within the !rac>et (8.:* - !ecomes negative. This is then ta>en as zero and the function within the square root sign !ecomes unity, and the equation is simplified as shown in )quation %.:.+. Equation 6.4.3 5lternatively, valve-sizing or Bv charts can !e used. Top

Terminology
<ormally Bvr Bvs . the . 9ull full lift value of the value capacity valve will !e stated for for a using the an particular term B vs, thus5ctual lift required stated application valve

6anufacturers give the maximum lift Bvs values for their range of valves. /ence the B v value is not only used for sizing valves !ut also as a means of comparing the capacity of alternative valve types and ma>es. 'omparing two ;<7& valves from different sources shows that valve $5$ has a B vs of 78 and valve $B$ a B vs of ?. (alve $5$ will give a higher flowrate for the same pressure drop.

-ringing toget er t e information for steam valve sizing

'ertain minimum information is required to determine the correct valve size-

The pressure of the steam supply must !e >nown. The steam pressure in the heat exchanger to meet the maximum heat load must !e >nown.

The difference !etween the a!ove criteria defines the differential pressure across the valve at its full load condition.

The heat output of the equipment must !e >nown, along with the enthalpy of evaporation (hfg at the wor>ing pressure in the heat exchanger. These factors are required to determine the steam mass flowrate.

E)am!le 6.4.2.
5 control valve is required for the application shown in 9igure %.:.G. The shell and tu!e heat exchanger manufacturer specifies that a steam pressure of & !ar a!solute is required in the tu!e !undle to satisfy a process demand of &88 >,. ,et steam, at dryness 8.E% and 78 !ar a, is availa!le upstream of the control valve. )nthalpy of evaporation (hfg at & !ar a is * 78?.*+ >0 1 >g.

Fig. 6.4.. Control valve on steam su!!ly to a s ell an% tu'e eat e)" anger

/etermine t e steam flowrate.

9irst, it is necessary to determine the steam state for the downstream condition of & !ar a. By entering wet steam at 78 !ar a, and 8.E% dryness into the Spirax Sarco we!site wet steam ta!le, it can !e seen that the total heat (hg held in the 78 !ar wet steam is * %EG.7& >0 1 >g. The heat exchanger design pressure is & !ar a, and the total heat in dry saturated steam at this pressure is * G:?.%& >0 1 >g (from the steam ta!le . The total heat in the 78 !ar steam (due to its $wetness$ , is less than the total heat in saturated steam at & !ar, and so the lower pressure steam will not contain enough heat to !e totally dry. The dryness fraction of the lower pressure steam is the quotient of the two total heat figures.

The steam flowrate can now !e determined from )quation *.?.7, where h fg is the enthalpy of evaporation availa!le after accounting for wet steam.

Equation 2.0.1

/etermine t e !ressure %ro! ratio 1 2 at full loa%

/etermine t e require% 3vr The pressure drop ratio at full load is larger than 8.:*, so critical conditions apply and )quation %.:.+ may !e used to find the required Bvr. Equation 6.4.3

5 ;<*& control valve with a Bvs of 78 is initially selected. 5 calculation can now !e carried out to determine if noise is an issue with this sized valve passing wet steam in the valve outlet. The speed of sound in the valve outlet-

A . 8.:%7 & >H1>g (the gas constant for steam The temprature of wet steam at & !ar a is the same as dry saturated steam at the same pressure" T . :*& B. The speed of sound in the wet steam in the the value outlet-

The

noise

criterion

for

wet

steam

in

the

valve

outlet

:8

5s this outlet velocity is higher than :8 m 1 s, the ;<*& control valve might7. *. 'reate an unaccepta!le noise. 'ause unreasona!le erosion in the valve outlet.

The ;<*& control valve will therefore !e unsuita!le for this application where wet steam passes through the valve outlet. 3ne solution to this pro!lem is to fit a larger !odied valve with the same B vs of 78 to reduce the wet steam outlet velocity.

'onsider Ta!le %.:.7 to determine the minimum sized control valve with an outlet area greater than 8.88* ** m.

&a'le 6.4.1 &y!i"al valve outlet areas /,15 $ /,244 "ontrol valves #t can !e seen from Ta!le %.:.7 that the smallest valve required to satisfy the maximum outlet velocity of :8 m 1 s for wet steam is a ;<%& valve, having an outlet area of 8.88++* m. Therefore, due to wet steam passing through the valve outlet, the size of the control valve would increase from, in this instance a ;<*& (7D to ;<%& (*JD . 5 !etter solution might !e to fit a separator !efore the control valve. This will allow the smaller ;<*& control valve to !e used, and is preferred !ecause-

#t will give !etter regulation as it is more appropriately sized to handle changes in the steam load. #t will ensure dry steam passes through the control valve, there!y reducing the propensity for erosion at the valve seat and valve outlet. #t will ensure optimal performance of the heat exchanger, as the heating surface is not thermally insulated !y moisture from wet steam. The cost of the smaller valve and its actuator plus separator will pro!a!ly !e the same as the larger valve with a larger actuator.

Sizing on an ar'itrary !ressure %ro!

#f the apparatus wor>ing pressure is not >nown, it is sometimes possi!le to compromise. #t should !e stressed that this method should only !e used as a last resort, and that every effort should !e made to determine the wor>ing pressures and flowrate. 2nder these circumstances, it is suggested that the control valve !e selected using a pressure drop of 78= to *8= of the upstream pressure. #n this way, the selected control valve will more than li>ely !e oversized. To help this situation, an equal percentage valve will give !etter operational performance than a linear valve (this is discussed in more detail in Tutorial %.& $'ontrol valve characteristics$.

Sizing on an ar!itrary pressure drop is not recommended for critical applications.

& e ig er t e !ressure %ro! t e 'etter5

#t is usually !etter to size a steam valve with critical pressure drop occurring across the control valve at maximum load. This helps to reduce the size and cost of the control valve. /owever, the application conditions may not allow this.

9or example, if the heat exchanger wor>ing pressure is :.& !ar a, and the maximum availa!le steam pressure is only & !ar a, the valve can only !e sized on a 78= pressure drop (K& - :.&L 1 & . 8.7. #n this situation, sizing on critical pressure drop would have unduly reduced the size of the control valve, and the heat exchanger would !e starved of steam. #f it is impossi!le to increase the steam supply pressure, one solution is to install a larger heat exchanger operating at a lower pressure. #n this way, the pressure drop will increase across the control valve. This could result in a smaller valve !ut, unfortunately, a larger heat exchanger, !ecause the heat exchanger operating pressure (and temperature is now lower. /owever, a larger heat exchanger wor>ing at a lower pressure !rings some advantages-

There is less tendency for the heating surfaces to scale and foul as the required steam temperature is lower. Cess flash steam is produced in the condensate system leading to less !ac>pressure in the condensate return pipewor>.

#t is important to !alance the cost of the valve and heat exchanger, the a!ility of the valve to control properly, and the effects on the rest of the system, as explained previously. 3n steam systems, equal percentage valves will usually !e a !etter choice than linear valves, as low pressure drops will have less effect on their operating performance. Top

Types of steam heated heat exchangers

This su!Hect is outside the scope of this Tutorial, !ut it is useful to have a !rief loo> at the two main types of heat exchanger used for steam heating and process applications. & e s ell an% tu'e eat e)" anger Traditionally, the shell-and-tu!e heat exchanger has !een used for many steam heating and process applications across a !road spectrum of industries. #t is ro!ust and often $over-engineered$ for the Ho!. #t tends to have an inherently high mass and large thermal hysteresis, which can ma>e it unwieldy for certain critical applications. Shell-and-tu!e heat exchangers are often greatly oversized on initial installation, mainly !ecause of large fouling factors applied to the calculation. They tend to have low steam velocity in the steam tu!e, which reduces-

Tur!ulence. The sheer stress !etween the flowing steam and the tu!e wall. /eat transfer.

Cow sheer stress also tends not to clean the tu!e surfaces" hence high fouling factors are usually applied at the design stage leading to oversizing. ;ue to oversizing, the actual steam pressure after installation is often much less than predicted. #f this is not anticipated, the steam trap might not !e correctly sized and the steam tu!es might flood with condensate, causing erratic control and poor performance.

& e !late 1an% frame2 eat e)" anger

@late heat exchangers are a useful alternative" !eing relatively small and light, they have a small mass and are extremely quic> to respond to changes in heat load. ,hen properly designed, they tend not to foul, !ut if they do, they are easily disassem!led, cleaned and recommissioned. 'ompared to shell-and-tu!e exchangers, they can operate at lower pressures for the

same duty, !ut !ecause of their high heat transfer characteristics, and a lower requirement for oversizing, they are still smaller and less expensive than a compara!le shell-and-tu!e exchanger. @late heat exchangers (when properly engineered to use steam are therefore more economically suited to high pressure drops across control valves than their shell-and-tu!e counterparts. This can give the advantage of smaller and less expensive control valves, whilst minimising the cost of the heat exchanger itself. Menerally, it is !etter to design the system so that the plate exchanger operates with critical pressure drop (or the highest possi!le pressure drop across the control valve at full load. #t must !e stressed that not all plate heat exchangers are suita!le for steam use. #t is very easy to !uy a heat exchanger designed for liquid use and wrongly assume that it will perform perfectly when heated with steam. 'orrect selection for steam is not Hust a matter of pressure1temperature compati!ility. @roper expertise is availa!le from !ona fide manufacturers, and this should always !e sought when steam is the prime energy source. Steam sizing e)am!les using " arts The required $flow coefficient$ (Bvr may !e determined in a num!er of ways, including calculation using )quation +.*7.* or )quation %.:.+ or via computer software. 5n alternative method of simple valve sizing is to use a Bv chart, 9igure %.:.?. 5 few examples of how these may !e used are shown !elow-

Saturate% steam
E)am!le 6.4.3 Criti"al !ressure %ro! a!!li"ation
Steam demand of heat exchanger . ?88 >g 1 h Steam pressure upstream of valve . E !ar a Steam pressure required in heat exchanger . : !ar a Aeference steam Bv chart (9igure %.:.? 7. ;raw a line from ?88 >g 1 h on the steam flow ordinate. *. ;raw a horizontal line from E !ar on the inlet pressure ordinate. +. 5t the point where this crosses the critical pressure drop line (top right diagonal draw a vertical line downwards until it intersects the horizontal ?88 >g 1 h line.

:. Aead the Bv at this crossing point, i.e. Bvr N G.&

E)am!le 6.4.4 # non "riti"al$!ressure$%ro! a!!li"ation
Steam demand of heat exchanger Steam pressure upstream of valve Steam pressure required in heat exchanger Aeference steam . *88 >g 1 h . % !ar a . & !ar a Bv

chart

(5ppendix

5s in example %.:.+, draw a line across from the *88 >g 1 h steam flow ordinate, and then draw another line from the % !ar inlet pressure ordinate to the 7 !ar pressure drop line. ;rop a vertical line from the resulting intersection point, to meet the *88 >g 1 h horizontal and read the B v at this crossing point i.e. Bvr N +.?

E)am!le 6.4.5 Fin% t e !ressure %ro! 1672 a"ross t e valve aving a 8nown 3 vs value
Steam demand of heat exchanger . + 888 >g 1 h Steam pressure upstream of valve . 78 !ar a Bvs of valve to !e used . +% Aeference steam Bv

chart

(5ppendix

;raw a horizontal line from + 888 >g 1 h to meet at the B v +% line. ;raw a vertical line upward from this intersection to meet the 78 !ar horizontal line. Aead the pressure drop at this crossing point, @

7.%

!ar.

,ote+ #n the examples, to convert gauge pressure (!ar g to a!solute pressure (!ar a simply add $7$ to the gauge pressure, for example, 78 !ar g . 77 !ar a.

Fig.

6.4.0 Steam 3v " art Su!er eate% steam To size a valve for use with superheated steam refer to )xample %.:.% and the superheated steam chart, 9igure %.:.E. E)am!le 6.4.6 The following example shows how to use the chart for 7884' of superheat- follow the respective steam flow line on the left to the vertical line which represents 7884' of superheat, then draw a horizontal line across as normal from the resulting intersection. By doing this, the graph introduces a correction factor for the superheat and corrects the Bv value.

Fig. 6.4.9 # su!er eate% steam sizing " art Sele"ting a "ontrol valve for steam servi"e The previous Section covered the procedure for sizing a control valve !ased on the flowrate it needs to pass, and the pressure drop across the valve. 9rom this data, the B vs value of the control valve can !e

o!tained. Aeference to the appropriate product literature will provide the information needed to select the required valve size. 'ontrol valve selection requires several other factors to !e ta>en into account. The !ody material must !e selected to suit the application. (alves are availa!le in cast iron, SM iron, !ronze, steel, stainless steel, and exotic materials for very special applications, for example titanium steel. The design and material of the control valve must !e suita!le for the pressure of the system in which it will !e fitted. #n )urope, most valves have a nominal pressure !ody rating, stipulated !y the letters $@<$ which actually means $@ression <ominale$. This relates to the maximum pressure (!ar gauge the valve can withstand at a temperature of 7*84'. The higher the temperature, the lower the allowa!le pressure, resulting in a typical pressure 1 temperature graph as shown in 9igure %.:.78. #t should !e noted that the type of material used in manufacturing the control valve plays an important part in the pressure 1 temperature chart. Typical limiting conditions are-

Fig. 6.4.14 #n e)am!le of 7,25 tem!erature * !ressure limiting "on%itions The design thic>ness and !ody Hointing methods also have an effect. 9or example, an SM iron valve could have a @<7% rating and may also !e availa!le with a slightly different design, with a @<*& rating. Cocal or national regulations may affect the limits, as may the type of connection which is used. # " e"8list of t e ma:or fa"tors to 'e ta8en into a""ount w en sele"ting a "ontrol valve for steam servi"e in"lu%e+ 7. *. +. :. &. %. G. ?. E. 6ass flow or volumetric flow to !e considered (typically maximum, normal or minimum . 9low medium (this may affect the type of material used for the valve !ody and internals . 2pstream pressure availa!le at maximum, normal and minimum loads. ;ownstream pressure for maximum, normal and minimum loads. Bv value required. @ressure drop across the valve at maximum, normal and minimum loads. Body size of valve. Body material and nominal pressure rating. 6aximum differential pressure for shut-off.

78. 'onnection required. ,hich pipe connections are required on the inlet and outlet of the valveF Screwed or flanged connections, and which type of flange, for example, 5<S#, )< 78E* or ;#<F 77. 6aximum temperature of the medium flowing through the valve. 7*. 5ny special requirements, for example, special gland pac>ing variations" hardened valve seat and plug, soft seats for a!solutely tight shut-off" and others. <ote- 6anufacturers restrict the lea>age rates of control valves to agreed limits and 1 or they are sometimes the su!Hect of national standards. 5lso see point 7G. 7+. ;etails of the application control requirements. This is explained in more detail in Tutorial %.&. Briefly, an application needing on 1 off control (either fully-open or fully-closed may require a valve characteristic suited to that purpose, whereas an application calling for continuous control (any degree of opening or closing , might perform !etter with a different type of valve characteristic. 7:. 6ethod of actuation and type of control to !e used" for example, self-acting, electric, pneumatic, electropneumatic. 7&. <oise levels. #t is often a requirement to >eep noise !elow ?& dB5 at 7 m from the pipe if people are to wor> unprotected in the area. Beeping the same size internals !ut increasing the size of the connections may achieve this. (6any control valves have the option of reduced trim variants, alternatively special noise-reducing trims are availa!le, and 1 or acoustic lagging can !e applied to the valve and pipewor>. (alves for critical process applications should !e sized using computer software utilising the #)' %8&+: standard or national equivalent. 7%. @ressure drops, sizes of valve !ody and noise level are related and should !e considered. #t is good practice to >eep the downstream steam velocity in the valve !ody typically !elow 7&8 m 1 s for saturated steam and *&8 m 1 s for superheated steam. This can !e achieved !y increasing the valve !ody size, which will also reduce the velocity in the valve outlet and the li>elihood of excess noise. #t is possi!le to consider a saturated steam exit velocity of 7&8 m 1 s to *88 m 1 s if the steam is always guaranteed to !e dry saturated at the valve inlet. This is !ecause, under these circumstances, the steam leaving the control valve will !e superheated due to the superheating effect of reducing the pressure of dry saturated steam. @lease note that these are general figures, different standards will quote different guidelines.O1O liP 7G. Cea>age and isolation. 'ontrol valves are meant to control flowrate rather than isolate the supply, and are li>ely to lea> slightly when fully shut. 'ontrol valves will !e manufactured to a standard relating to shut-off tightness. Menerally, the !etter the shut-off, the higher the cost of the valve. 9or steam control valves, a lea>age rate of 8.87= is perfectly adequate for most applications. 7?. Turndown. 2sually expressed as a ratio of the application maximum expected flow to the minimum controlla!le flow through a control valve. 7E. Aangea!ility. 2sually expressed as a ratio of the valve maximum controlla!le flow to the minimum controlla!le flow, !etween which the characteristics of the control valve are maintained. Typically, a rangea!ility of &8-7 is accepta!le for steam applications. *8. #t would !e wrong to end this Tutorial on control valves without mentioning cost. The type of valve, its materials of construction, variations in design and special requirements will inevita!ly result in cost variations. 9or optimum economy the selected valve should !e correct for that application and not over-specified. Top

5ppendix 7 Saturated steam valve sizing chart

Top

5ppendix * Superheated steam valve sizing chart

Top

,hat do # do nowF

The printa!le version of this page has now !een replaced !y The Steam and 'ondensate Coop Boo> Try answering the Questions for this tutorial (iew the complete collection of Steam )ngineering Tutorials

'ontact 2s

'ontrol /ardware- )lectric1@neumatic 5ctuation

'ontrol (alves 'ontrol (alve 'apacity 'ontrol (alve Sizing for ,ater Systems Control Valve Sizing for Steam Systems 'ontrol (alve 'haracteristics 'ontrol (alve 5ctuators and @ositioners 'ontrollers and Sensors

;elate% Content
'ontrol (alve Sizing
2se the calculator to size steam valves.

Feature

The Steam and 'ondensate Coop

5 comprehensive !est practice guide to saving energy and optimising plant performance, this !oo> covers all aspects of steam and condensate systems. 3rder your copy today