Steam trap and temperature controlling valve 2

[kszyho100]

Hi

I have a question. At a production plant I noticed hot water system heated by 6 bar g steam. Downstream of a heatexchanger was a temperature controlling valve and parallel to that valve was installed a steam trap. Could anyone explain me if it makes any sense bypassing controliing valve with a steam trap?

Best regards

Chris

 

[quark]

What is the control signal? Is it temperature of condensate? What type of trap it is? IMHO, I don't find any reason why a control valve is to be placed downstream of a HX unless you are using sensible heat of condensate and thus subcooling it. Thermostatic steam traps can handle condensate subcooling and this may help incase of control valve failure.

Regards,

 

[kszyho100]

There is a sensor in a hot water pipe, and the steam trap I think is termostatic.

Regards

Chris

 

[hacksaw]

sounds like pretty normal piping configuration. you don't want condensate accumulation upstream of the valve...

good luck

 

[katmar]

It sounds totally wrong to me. If the valve and trap were upstream of the exchanger I would agree with hacksaw that this is normal, but not if they are downstream.

It does make sense to have a steam trap on the outlet of the steam coil, because you want to get rid of the condensed steam.

But it makes no sense to put a control valve in parallel with the trap. Assuming the valve is responding to the temperature signal from the water pipe, what action does it take? If you want to heat the water more do you open the valve wider? This would have the opposite effect to that desired. Opening the valve will lower the pressure of the steam in the coil, thus lowering its condensing temperature and therefore decreasing the heat transferred to the water.

The valve should be on the INLET to the steam coil. Then as the water temperature goes down the valve will open and INCREASE the steam pressure in the coil, thus increasing the condensing temperature and increasing the heat transferred to the water in the tank.

Either you have not described the situation correctly, or the person who put this system together did not understand what they were doing. In defence of that person I must admit many people have a misunderstanding of steam systems. I have often seen operators opening the bypass around a steam trap to "improve" the operation of a reboiler. As described above, this can easily have the opposite effect to what is wanted.

 

[abeltio]

in very low pressure, saturated steam applications for heating the practice is to put the control valve DOWNSTREAM of the HE - i've seen this arrangement mostly in hospitals.

the valve is put downstream to maintain a certain liquid level... i.e. ensure that all the latent heat is used in the HE, and admit some sub-cooling.

the first time i saw this disposition i was also surprised... the engineer that was my mentor at that time said:
the difference in size between the 2 valves:
1. to control LP steam
2. to control liquid
is reason enough to control downstream, and we do not have budget for the pressure drop required across the valve upstream.
whereas, if we put it downstream gravity does most of the job.

regarding the steam trap... in this particular application it may be required to maintain a dry HE.
So the trap will drain liquid until there is only steam flowing, then the control valve will take over...
i.e. the steam trap will most likely operate during start-up and shutdown of the system.

If the duty of the HE does not require condensation of all the steam, and if dry steam is required downstream of this service this sounds like a reasonable arrangement.

These are just appreciations that need confirmation thru an analysis of the complete P&ID.

Just wanted to express that having the control valve downstream of the HE is not crazy.

HTH


saludos.
a.

 

[25362]

Abeltio, not only it is not crazy, in fact there are some operating companies that very definitely prefer condensate-outlet control, albeit its apparent disavantages, some of which may be precluded already at the design stage.

Among the frequently emphasized disadvantages of condensate control: (a) sluggish response; (b) unpredictable performance; (c) thermal shock; (d) corrosion by the condensate level kept in the exchanger if CO[sub]2[/sub] and oxygen are present.

Although oversurfaced exchangers are more readily handled by condensate-oulet control systems, cases of hammering have been reported as a result of level of condensate surging within the exchanger.

Having said that, condensate flooding could also happen with steam-control under reduced loads, steam on the tube-side, and a steam trap incapable of removing the condensate because of lack of sufficient upstream pressure, through no fault of the trap.

As you see there are pros and cons, and ways to control a steam-heated exchanger vary. It is issues like these that make the subject much more interesting.

Hasta la pr[ó]xima.

 

[katmar]

Hi Abeltio,

You are 100% correct - it does make sense to put the valve on the outlet when dealing with low pressure steam because the valve for controlling condensate is much smaller than for controlling LP steam. In this situation the condensate backs up in the heat exchanger, which in effect becomes a "variable area" device as the condensate blanks off tube area from the steam.

But in the situation described by Chris the steam trap would remove the condensate and the control valve would see steam. This would result in live steam being lost via the valve to the condensate system.

It makes sense to have either the valve or the trap, but not both.

I think the designer of this system was trying for a "belt & braces" system, but he hooked the braces under his feet instead of over his shoulders!

regards
Katmar

 

[25362]

It seems to me the trap would have been installed not to function as an in-service steam trap.

By-passing the TCV on the condensate with, say, a bi-metallic thermostatic trap, could be used just for short periods of time for temporary trapping, as in start-up when air is being expelled out from the system and during warming up.

There are thermostatic traps of a liquid expansion type with an adjusting nut, designed to shut down at an adjustable set point temperature, keeping the upstream pipe flooded.

Although kszyho100 didn't inform us whether there are valves in line with the trap to stop its functioning after some time, I assume there are for "maintenance" reasons. I'd appreciate if kszyho100 would tell us whether the trap is "cold".

 

[quark]

Katmar and Abeltio,

I agree with you guys about steam control rather than condensate control for better process heat transfer but I won't straight away deny the condensate control for the reasons given by 25362. To my logic, it seems condensate control is a better way to minimize heat exchanger stall.

Further, it is totally not true that opening a condensate valve would reduce pressure, thus condensing temperature and subsequently quantity of heat transfered.

Let us see an example.

Pressure,bar(abs) Sat. Temp.deg.C Latent HeatkJ/kg
7 164.94 2065.74
6 158.826 2085.763
5 151.831 2108.025

These data from steam tables indicate that steam has higher latent heat at lower temperatures. That is why low pressure steam is generally recommended when heated medium temperature is not a concern. Presence of condensate downstream side of HX indicates latent heat removal by heated fluid and this indicates proper quantity of heat transfered.

Regards,

 

[katmar]

Quark,

The rate of heat transferred by a heat exchanger is modelled using the equation

Q = U x A x dT

Q is heat transferred in unit time
U is heat transfer coefficient
A is heat exchanger surface area
dT is temperature difference between hot and cold fluids

This shows that the rate of heat transfer is directly proportional to the temperature difference, and your example shows clearly that this will decrease as the pressure decreases.

While your example shows that more heat would be given up by each kilogram of steam if the condensing pressure were lower, it does not show that the RATE of heat transfer would increase. As I have shown above, it would in fact decrease.

If the control valve is used on the outlet side of a heat exchanger and it is opened too wide there is the risk of passing live steam to the condensate, and in this case NO latent heat is transferred.

 

[25362]

I sensed this would become an interesting thread.

It is clear that a change in load due to temperature variations in the process fluid, would involve manipulating the (smaller and cheaper) CV in the condensate outlet line, the rate of heat transfer will be modified by partial flooding of the exchanger tubes.

However, consider, please, the common cases in which heat loads vary due only to changing process fluid flow rates.

In these situations the lower process-side film coefficient will govern the rate of heat transfer. The OHTC (U) will change almost linearly with fluid velocity, and so will the rate of heat transfer.

By keeping the steam pressure constant, and so also its temperature, following Q = A x U x LMTD, the exit process fluid temperature, even without being directly controlled, will nonetheless be well regulated since A and LMTD will remain practically unchanged.

I think, thus, that the inherent sluggishness of response when having a CV on the condensate outlet, on a horizontal heat exchanger, wouldn't be so important under these conditions.

What do you think ?

 

[katmar]

Hi 25362,

This is an interesting scenario that you have raised here, and I believe you are correct. Overall, my feeling is that controlling on the condensate can work and it will definitely be cheaper than controlling on the steam, but as others have pointed out it is less responsive.

My point in my earlier posts was not that it was wrong to control on the condensate, but that it was wrong to have a control valve AND a steam trap on the condensate. The only possible reason to have both is the situation you described where the trap is set to drain cold condensate at start up.

regards
Katmar

 

[quark]

Katmar,

I hope I am not playing the devils advocate here but pushing the discussion a bit so as to get familiarized with unseen corners.

I have never seen such a type of control told by the OP but just thinking of various possibilities. Coming to your question, the log mean temperature difference do vary with varying temperatures, the effect will be relatively less as this is a condensation heat transfer. Further, with low pressures, steam velocity increases and hence steam side convective coefficient.

My prerequisite is "if we get condensate out of the heat exchanger".

I do agree with the live steam wastage and other good points.

Regards,

 

[25362]

What I said regarding heating liquids or gases with steam, can also be applied to boiling organics.

In this case the thermal load would usually depend on the process fluid's flow rate. Since, as indicated by quark, the OHTC and LMTD wouldn't change much, as long as the pressures on both sides are kept constant the heat load could be controlled by varying the exposed surface.

A vertical unit, with its linear relationship between exposed surface and degree of flooding may still be well-controlled by acting on the condensate, although this system may respond more slowly than direct manipulation of steam flow.

Let's not forget that in many process-vapor condensers the most effective way of control -while the coolant flow rate is constant- is by varying the heat-transfer area manipulating the condensate by a pressure-controlled valve to get partial flooding of the condenser. The level of condensate in the condenser shell-side is a good indication of the process heat load.

More comments are invited.