28 replies to this topic

[gemhag]

Case: Water is heated by a hot oil heater. Capacity of the heater is 8500 kW.

Possible relief scenario for the downstream water safety valve is failure of automatic controls: meaning that the hot oil control valve to the heat exchanger fails open and based on it's cv it will give a lot of heat input (Q = m*cp*dT), even a lot more than the heat exchanger capacity.

Is the heat exchanger duty a limiting factor or should I size the PSV for the flow generated by the hot oil heat input?

Note: The duty of the exchanger is 7 000 000 kcal/hr; the hot oil flow calculated with a Cv of 400 can deliver a heat input of 14 000 000 kcal/hr 

Attached Files

•  hot oil exchanger.pdf   30.93KB   88 downloads

Edited by gemhag, 27 February 2013 - 01:07 PM.

[fallah]

gemhag,

 

Please upload a simple sketch of the system...

[gemhag]

Attachment added

[fallah]

gemhag,

 

Lack of information about the pressure and temperature of the outlet water in normal operation and during CV failure. Anyway, if there would be no any phase change (vaporization) in outlet water due to increasing hot oil flowrate, a small PSV is adequate to protect the water side against overpressure might be created by subsequent thermal expansion...

 

 

 

 

[gemhag]

I think my question is: do I have to use the heat exchanger duty (7 * 10^6)  or the hot oil CV failure (14 * 10^6)  for the scenario of thermal expansion (blocked in water in heat exchanger)

[Hugo P]

He gemhag

 

What need to do is using the heat exchanger capacity (Area) to calculate the heat input to the water side. The outlet temperature of hot oil will change in this situation giving u a heat flow different from what you are estimating. With the heat input you should be able to calculate the PSV depending on your configuration. Worst case escenario will be 14*10^6

 

Depending on the rigorosity required for your calculations u could do a full calculation of your exchanger. Using hot oil faillure flow and water conditions of during psv opening (probably higher pressure)

 

A simplification of this is using the formula Q= UA * dtml . You can asume UA its constant (calculate form exchanger design condition) and calculate the dtml for your relieving conditions.

 

Hope i was of any help

 

Hugo

[fallah]

gemhag,

 

For the scenario of the thermal expansion due to blocked in water, you should use heat exchanger duty...because CV failure is another unrelated scenario...

Edited by fallah, 22 February 2013 - 01:17 PM.

[paulhorth]

Gemhag,

What is the design pressure of the water side? What is the set pressure of the psv? What is the supply temperature of the hot oil?

 

If the oil is hot enough, and you block in the water, you can boil it. That means the PSV should be sized for STEAM not water, You can generate quite a lot of steam from 14 million kcal/hr. Calculate the steam flow at this heat input and at the set pressure of the psv. That is your sizing basis, if the oil is hot enough to boil the water at the set pressure.

 

Paul

[vikramltv]

gemhag

 

As per my understanding for your problem, following explanation will help U.

Case-1 :- Hot oil control valve failure- (duty=14*10^6)

    Water side valve can not be assumed blocked at the same time as it will be double jeopardy.

    Water outlet temp. will increase so water flow will be reduced by TCV.. No relief required.

Case-2 :- Water side flow blockage - (Duty = 7*10^6)

    Relief required.

[paulhorth]

Vikram,

I don't agree that it is "double jeopardy" to consider water blocked outlet AND full opening of the hot oil valve. The sketch doesn't show any temperature control of the water - but if the water flow is blocked, and the water temperature controller is downstream, it will see a falling temperature ( no flow coming through), so the controller will act to open the hot oil valve. So, one failure can result in the full heat duty applied to the blocked-in exchanger.

 

I have said repeatedly in posts to this forum - you need to be really careful about the concept of "double jeopardy" and where you apply it . The consequences of a single failure can cause other upsets by the normal and correct actions of controllers or shutdowns.

 

The exchanger in this case can be protected by a high temperature trip on the water outlet (inside any block valve) or on the exchanger, acting on a separate ESD valve on the oil supply. The oil control valve should be fail-closed on loss of air.

 

Paul

Edited by paulhorth, 25 February 2013 - 06:29 AM.

[fallah]

Paul,

 

[quote]...but if the water flow is blocked, and the water temperature controller is downstream, it will see a falling temperature ( no flow coming through), so the controller will act to open the hot oil valve...[unquote]

 

If water temperature controller is upstream, it will see rising temperature, so the controller will act to close the oil valve...and then it could be double jeopardy...

 

We can see by changing the assumption the result will be different, then it is better at first the matter would be clarified by OP...

[paulhorth]

Fallah,

...but why would anyone put the temperature controller upstream of the heater? It wouldn't work there.

Paul

[neel_avi]

 Dear Mr. Paul,

 

Mr. Fallah might be referring to the location of temperature element on the u/s of the isolation valve at water side outlet of the exchanger. In that way he is correct. 

 

If its downstream your analysis is correct. 

 

Gemhag may post a better schematic or PID for our understanding. 

 

Regards

Avijit

[fallah]

neel_avi,

 

Of course, i meant the location of temperature element on the u/s of the isolation valve at water side outlet of the exchanger.

[paulhorth]

Avijit,

No - I don't agree with you.

If the temperature sensor is actually on the exchanger water side, then yes, it will sense the rising temperature when the exchanger is blocked in, and act correctly to shut off the oil, But if it is located somewhere on the water outlet piping, where there is no flow because a block valve has been closed, then it will will not detect a temperature rise, instead it will see a slow fall in temperature as the pipe cools down. It makes no difference whether the sensor is upstream or downstream of the block valve. This would still be true even if the closed block valve was on the inlet of the exchanger - but in that case, if the downstream piping was all open,  the exchanger pressure would not rise .

 

We need to see the PID for the exchanger with all instruments and valves, to give a complerte answer to the original question, All relief cases are like this - you need to study carefully the complete actual system.

 

Paul

[fallah]

 

Paul,

 

[quote]...But if it is located somewhere on the water outlet piping, where there is no flow because a block valve has been closed, then it will will not detect a temperature rise, instead it will see a slow fall in temperature as the pipe cools down. It makes no difference whether the sensor is upstream or downstream of the block valve...[unquote]

 

Would you please explain while hot oil flow is going to be continued how the blocked in water will face with a fall in its temperature...

[paulhorth]

Fallah,

 

Would you please explain while hot oil flow is going to be continued how
the blocked in water will face with a fall in its temperature..

 

The heat is applied only  to the water inside the exchanger. The water trapped inside the exchanger will get hot, very rapidly. This would be detected by a temperature sensor located on the exchanger itself. But if there is no water flow, the heated water cannot leave the exchanger and cannot reach the temperature sensor in the outlet piping. The water in the outlet piping is static and not heated. So this water,in the outlet piping, will gradually cool by heat loss. If the temperature sensor is located in the outlet piping it will see this temperature fall and will open the hot oil valve. This is true regardless of the location of the closed block valve.

 

Paul

 

Paul

[neel_avi]

Dear Mr. Paul,

 

you are right as well as Mr. Fallah - only the location for the temperature element has changed for you and him.The schematics attached may be referred to see whether my understanding is correct.

 

Please help me if its otherwise.

However, PID is requested from gemhag for further clarity.   schematic.xlsx   17.67KB   39 downloads

 

regards

 

Avijit 

[paulhorth]

Avijit,

No, you still haven't got it.

In your diagram named "Fallah" you show the temperature sensor on the outlet piping, but you label it "temperature will rise". . As I have already explained in two posts, this sensor on the outlet piping will see a falling temperature if the flow of water is stopped.It is only if he sensor is located on the exchanger shell itself, where the heat is applied, that the temperature will rise.

 

I'll try an everyday example.

The hot water to my bathroom is supplied from a heated tank in the next room, about 3 m away. The tank is heated by the central heating system. The outlet is blocked because the hot tap in the bathroom is shut. (The tank does not explode, because it is vented to atmosphere). This situation is exactly analogous to the case in the original post of this thread.

.When I go into my bathroom and turn on the hot water tap, do I get hot water instantly? No. I have to wait some seconds, because the water in the pipe from the tank is cold, and this cold water has to run through before the hot water comes. That is why the tank thermostat is on the tank itself and not on the outlet pipe.

If you have ever been in a hotel or large building and washed your hands with hot water, you will have had the same experience.

 

Paul

Edited by paulhorth, 26 February 2013 - 05:09 PM.

[neel_avi]

Dear Mr. Paul,

 

apologise for my ignorance, because I have still not understood the fact. If there is a constant flow of the oil, theoretically the blocked water should attain equilibrium with the hot oil which will be greater than or at least equal (not a probable scenario though - to me it should always increase) to the present temperature.

 

"Water in the tap is cold first" - It might be the reason for poor insulation?

 

I understand that I might be dragging this thread a way too much deviating from the actual post, but an enlightenment would help immensely.

 

regards

Avijit

[thorium90]

Sorry to interrupt with an off topic reply, but anyone noticed the OP hasn't bothered to reply since 5 days ago...?

[ankur2061]

Fallah /Paul / Avijit,

 

I have joined this discussion late and this has been a very enlightening discussion so far. The purpose of joining this discussion was to share an operating experience related to PSVs on heat exchangers or to be precise on coolers which utilize cooling medium (water in most cases) to cool the hot process fluid.

 

We had a hot glycol cooler (PHE) where the process hot glycol was being cooled from around 120-130°C to about 40-45°C using cooling water. Because the service was critical and the process fluid tended to be dirty an installed standby cooler was provided along with the working one. A preventive maintenance schedule was put in parctice where the operating cooler was taken out of line after every 2-3 weeks of operation, drained and isolated completely from both process and cooling water side and cleaned by the maintenance people before being certified clean for renewed service.

 

It so happened that during one such switchover from operating to standby the operator forgot to isolate & drain the cooler taken out of line from the process side and only isolated it from the cooling water side. Maintenance was scheduled to takeover the cooler the next day for cleaning. In the night we had an emergency breakdown in the plant and all the maintenance crew were diverted to ensure that the plant could be restarted. The next complete day was devoted by maintenance in bringing the unit back onstream. In all this chaos, the cooler was forgotten conveniently. After the unit stabilized, the maintence crew was informed by operations to find time to clean the out-of-service cooler as per the schedule. Maintenance asked operations whether the on-line cooler was functioning allright and the same was confirmed by operations. In view of this response, maintenance asked operations to defer the cleaning by a further few days.

 

The 6th day since the off-line cooler was supposedly isolated and maintenance having still not taken up the cleaning, an operator reported to me that the PSV (3/4" X 1") on the cooling water side had probably popped and there was a steady trickle of cooling water from the PSV outlet (open to atmosphere). On checking we found that the process side valves for the off-line cooler were still open and that the line upto the outlet block valve on the cooling water side had become considerably warm. After taking all the necessary actions such as isolating the process side and draining the cooling water and process side and having ensured that the PSV was reset the investigation was carried out wherein it became clear that the operator had made a mistake of not isolating the process side.

 

Why this long story and how is it related to this thread? It certainly is.

 

Avijit's contention is that the cold side will be in equilibrium with the hot side during a blocked-in condition. That is absolutely right. However, this process of getting in equilibrium with the hot fluid is very slow due to the inertia of the blocked in water mass to get heated up due to temperature gradients caused by convective heat transfer. It took six days for a relatively small mass of water trapped in the heat exchanger and the connected 6" cooling water outlet line of about 3 meter length to get heated up to the point where the thermal expansion of the trapped water caused the thermal relief valve to pop.

 

Coming to Gemhag's specific problem of how to size a relief valve on the water side which is being heated by hot oil, is that it cannot be a normal thermal expansion valve due to a blocked-in condition such as a 3/4" x 1" valve which is seen in most small heat exchangers (coolers). It will be a thermal expansion valve alright but a bigger size considering the heat duty on hot oil control valve failure mentioned as 14 MKcal/h based on the formula:

 

W = B*Q / 500*SG*Cp

 

where:

 

W = Relief flow rate, gpm

B = Coefficient of cubical exapnsion @relieving temperature, 1/°F

Q = Heat duty for the exchanger on control valve failure, Btu/h

SG = specific gravity of the water at relieving temperature

Cp = Specific heat of water at the relieving temperature, Btu/lb-°F

 

For this kind of large heat duty the relief rate would certainly be higher than what can be handled by a conventional 3/4" X 1" thermal relief valve. I guess it could be be a 2" x 3" thermal relief valve.  

 

Hope this helps.

 

Regards,

Ankur.    

[fallah]

Paul,

 

As long as the hot oil is flowing and the temperature sensor is located on the outlet piping before isolation valve, the heat loss from static hot water in outlet piping to the ambient (water to air heat transfer by natural convection via some resistances such as tube wall, insulation,... ) during blocked outlet could be compensated by heat gain of the same static hot water via trapped hot water in the exchanger (water to water heat transfer by natural convection), otherwise the temperature sensor would be located far from the exchanger that normally not to be done so and to be located as close as possible to the exchanger. Then the temperature sensor will not see the temperature fall such that opens the hot oil valve more than that of normal condition.

[neel_avi]

Dear Mr. Ankur,

 

Greatly appreciate your experience and explanation on the subject.

 

I believe that if hot oil flow continues, the temperature will not drop in the piping in the blocked outlet condition. Thus, I guess opening of oil flow control valve due to fall in water outlet temperature may not be a credible scenario if the TI is upstream of the water outlet block valve . 

 

You have further advised sizing of the TRV for 14 MMKCAL/hr of heat duty which reflects to control valve failure at the time of blocked outlet valve. I was unable to understand the reason  for doing the same. What I understand from such sizing is that we are seeing the chances of the control valve failure or the controller failure (zero reading) simultaneously or may be after the water side block valves were isolated. I was of the opinion that these two events were not related, so both happening at the same time fall under the double jeopardy description. To me sizing would be sufficient with the max heat duty of the exchanger (7 MMKCAL/hr).

 

I guess, you are of the opinion that after some time of closed valve operation, the oil control valve also fails to safe mode (fail open - assumed), then the TRV sizing might be required to be sized on the basis of 14 MMKCAL/hr. 

 

I would be glad if you could explain.

 

Regards

 

Avijit

[paulhorth]

This has gone on long enough. I will try one more time to make my basic argument here, then this will be my last post on the thread unless the original poster comes back,

 

Where are you, Gemhag?

 

I contend that it is not correct to assume that the hot oil valve will not open on control action if the hot water flow is stopped, because that assumption relies on a lot of detail about exactly where the sensor is located in this particular case. It obviously makes a difference if the temperature is located 1 cm from the exchanger or 10 m from the exchanger. We don't know where it is so we should make the most conservative assumption,

I contend that the safe basis for the PSV sizing is that the full hot oil flow is applied to the blocked-in exchanger, and this will probably lead to not just thermal expansion but boiling of the water (this depends on the psv set pressure and the oil supply temperature, which we don't know). If you want to adopt a less conservative basis then you have to be in a position to prove it.

 

Amijit says

"Water in the tap is cold first" - It might be the reason for poor insulation?

Yes, of course that is the reason, it is exactly what I was saying, the water in the outlet pipe will cool due to heat loss when there is no flow. If there is good insulation, the cooling will be slower but it will still get cooler unless it is so close to the exchanger that heat conduction along the pipe wall and the water is sufficient, This effect will not extend further than about 6 inches along the pipe.

I have just measured the temperature along my hot water pipe, about 2 hours after I last ran the hot tap.

Readings: Hot tank 50 C, 1.5 m distance, 31 C, 3 m distance ( at the tap) 21.1 C. Room temp was about 19 C. The water temperature of 50 C in the tank didn't get very far down the pipe.

 

Amijit says

What I understand from such sizing is that we are seeing the chances of the control valve failure or the controller failure (zero reading) simultaneously or may be after the water side block valves were isolated. I was of the opinion that these two events were not related, so both happening at the same time fall under the double jeopardy description.

You are wrong.We are not talking about control valve failure but normal controller action. I've explained why three times now how such a "double jeopardy" can easily happen from normal control action. In any case with psv sizing, if in doubt, be conservative.

 

Fallah says

the heat loss from static hot water in outlet piping to the ambient (water
to air heat transfer by natural convection via some resistances such as
tube wall, insulation,... ) during blocked outlet could be
compensated by heat gain of the same static hot water via trapped hot water in the exchanger (water to water heat transfer by natural convection)

 well it could..........but we have no basis to assume that, so we should not make that assumption the basis for a smaller relief case.

 

Amijit says

the oil control valve also fails to safe mode (fail open - assumed),

No - this hot oil valve should fail closed, like all valves which add energy to a system. It is not air failure which gives rise to this relief case. I think Ankur is referring to the valve sticking in the open position which is another type of failure.

 

We will see if the original poster comes back. Meanwhile, have a nice day, guys.

 

Paul