14 replies to this topic

[hianbotech]

I would like to receive comments regarding with a relief case to be considered in our hot Oil system.

 

We have a typical Hot Oil system (closed loop) with a Waste heat recovery unit supplying the heat for 10 different users connected to a supply header.  Each heat exchanger has a relief valve in the hot oil side set at 250 psig and sized for tube rupture case.

 

Additionally the Hot Oil expansion drum has a relief valve also sized for tube rupture and set at 80 psig. For the design case of this PSV the single largest flow from one of the heat exchangers was used (the same flowrate in lb/hr).

 

The hot oil supply and return headers are designed for 250 psig and the hot oil expansion drum is designed for 80 psig.

 

I would like to know your comments regarding with the relief case for the PSV to be installed in the hot oil expansion drum. I think the relief valves installed in the discharge of the heat exchangers will react very fast and will be the first protection of the equipment and the hot oil system. Why is the same case (tube rupture) used for the heat exchangers and the expansion drum? I know that the expansion drum has less design pressure than the users. Is it a credible scenario?, if the answer is yes how the flow rate should be calculated.

 

I appreciate your comments

Edited by hianbotech, 02 June 2015 - 05:08 PM.

[ahmadikh]

 What is the design pressure of the cold side of exchangers?

 

The relief valves on the exchangers are set at 250 psig so cannot protect the upstream drum. So, where are the design break points?

[AymanAyadi]

PSV on hot oil system could be designed for tow credible scenarios:

- Blanketing gas PCV failure,

- Tube rupture of process heat exchangers: Tubes rupture case need to be considered when hot oil circuit pressure is less than the process stream pressure.

 

It is strange that pipings and equipments in the same system and the same pressure range (no HP-LP interface) have different design pressure !! 

As mentioned by ahmadikh the PSV set @ 250 psig cannot protect the expansion tank. 

 

You should re-rate your expansion tank to the limit design conditions and redesign your PSV accordingly.

[hianbotech]

The design pressure in the cold side of the heat exchanger with highest pressure is 1480 psig.

 

The design breaks on Hot oil system are located in the inlet nozzle of the Hot Oil Expansion Drum, and the block valve located in the suction of the Hot Oil Recirculation pump.

All the supply and return header with their equipments in the hot oil system are designed at 250 psig. The hot Oil expansion drum and the suction of the hot oil recirculation line are designed for 80 psig.

[hianbotech]

The Hot Oil expansion Drum was designed at 80 psig. The criteria was based to limit the maximum suction pressure in the hot oil recirculation pumps in order to avoid shutoff conditions higher than 250 psig in the hot oil headers (discharge of the pumps).

[ahmadikh]

The Hot Oil expansion Drum was designed at 80 psig. The criteria was based to limit the maximum suction pressure in the hot oil recirculation pumps in order to avoid shutoff conditions higher than 250 psig in the hot oil headers (discharge of the pumps).

 

This is not true! 

 

In this close loop the suction vessel cannot reach design pressure unless the blanketing pressure control valve fails or there is a tube rupture in the exchangers. So, for calculating the max suction pressure what you should have considered is only the max liquid level and not the design pressure. So, there was no need to design the headers for such a huge design pressure. Remember for this pump the source and destination points are the same and the pump head is only equal to the friction pressure drop plus the pressure drop considered for sizing the control valves for the required flow in the headers, and not even the difference in static heads.

 

The reason that I asked for the cold side design pressure was to see that the PSVs on the hot oil side are sized for the catastrophic tube rupture or only for tube leaks. But in your case since 10/13 rule is not met, so it is sized for the catastrophic tube failure. However, you CANNOT consider the same flowrate for sizing the PSV on your drum as it is set at a lower pressure. A surge/dynamic pressure analysis is required to accurately size the PSV on the drum, otherwise it will become a huge PSV, if not two. If you do not do the tube rupture study, the driving force you have to consider is:   cold side operating pressure - 80 psig.

 

Hope this helps.

Milad

[Bobby Strain]

Looks like you have a major problem.

Bobby

[hianbotech]

Thanks to all for your time to answer my questions.

 

A dynamic analysys is in progress now, to determine flows and type of relief valves to be installed in the heat exchangers.

 

Regarding with the maximum suction pressure in the pump, the criteria to calculate the shutoff conditions of the pump based on the design pressure of the vessel was indicated by client in the design basis and specifications of the project. Credible or not, these are our basis to be followed.

[shan]

The PSVs set at 250 psig will start releasing when the expansion drum pressure reaches 80 psig (before its PSV open) and the hot oil circulation pump differential pressure is 170 psi.

[fallah]

hianbotech,

 

Please upload a simple sketch of the system you described...then you will certainly get better responces on your query...

[paulhorth]

Hianbotech,

 

In my opinion, what you have described is a safe design, perhaps a little conservative, but not greatly so.

 

First I would say that the design pressure of 250 psig is reasonable, because the piping engineers usually call for 300# flanges for hot oil service, regardless of the design pressure, to allow adequate flange tightening against leakage. Your 250 psig will be 300# at the design temperature of the system. You will not save cost by going to a lower pressure.

Next, I understand the reasons for dropping to 80 psig  in the expansion drum, arising from the general rule of adding the pump shutoff head to the design suction pressure to define the design discharge pressure. If you had 250 psig as the suction design pressure, then it would be possible in principle to reach (250 + 170) = 420 psig on the discharge, if an ESDV in the discharge or somewhere in the circuit had closed (which could happen in the event of an upset such as a tube rupture) while the pump continued to run. It would probably be transient, but it is possible. You could try to eliminate this scenario by various lines of argument and dynamic modelling, but it is simpler and thus clearer not to go that route.

 

You could go lower than 80 psig in the expansion tank, so long as the design pressure is above the supply pressure of the blanket gas, but there would be little to gain from that. As it is, the flanges on the vessel would have to be 300# to comply with the pipe spec. Pinhole leaks in the  exchangers would pass to the tank and be discharged safely (though difficult to detect).

 

Now, to come to the tube rupture scenario. A tube rupture should open the relief valve at the exchanger and pass all the gas to flare. The pressure at 250 psig (plus accumulation) would drive the oil out in both directions, pushing it to the expansion tank, and eventually if the gas was still flowing after the entire oil header had been emptied, the gas would break through into the expansion tank where it would have to be relieved.  The flow to the expansion tank would have a pressure drop of 170 psi to drive it through the large piping of the hot oil system. It seems likely (but you can check) that the flow that could be sustained with this DP would be higher than the tube rupture could supply, so the pressure in the exchanger would fall. This would close the RV at the exchanger. Thus all the gas through the tube rupture would have to be relieved from the expansion tank.

Note - if you had a bursting disc at the exchanger instead of a RV, the flow woud continue through it at lower pressure and so the flow to the expansion tank would be less.

 

Note also, that for the above scenario to happenm the tank has to be large enough to receive most of the oil in the system. This is not usually the case, and so the relief valve would be passing oil when it first opens.

 

Mitigating factors:

This scenario assumes that no instrumented actions have happened to stop the gas flow or to isolate the exchanger. It is good practice to fit a PSHH on the exchanger shell and to fit ESDVs on the hot oil side inlet and outlet, and to provide an ESDV somewhere in the gas system to shut off the feed - these valves actuated by the PSHH. If you have these features, and you check the system integrity level then you could possibly eliminate the expansion tank relief scenario. You could estimate how long it would take to empty the hot oil system as described, and evaluate whether it is credible that no operator action occurs in this time. High high level in the tank would also provide an additional protection.

Making these arguments to convince the client and your colleagues can be difficult, and in the end only offers a small cost saving in the RV at the expansion tank.

 

Obviously the PSHH should also cause a shutdown of the hot oil pumps, avoiding the discharge high pressure referred to above, with due regard to the required integrity level.

 

I would add that if you had used a rupture disc instead of a RV, the ESDVs would be also useful to avoid losing all the hot oil into the flare system when the gas pressure falls.

 

Good luck

Paul

[Bobby Strain]

Paul obviously missed the point that gas can accumulate in the piping system at high pressure. Liquid will be displaced first. Maybe the drum can hold all the liquid. But when the gas breaks through, it will flow at a much greater rate than it is being introduced through the ruptured tube. Then what? Rupture!

 

Bobby

[paulhorth]

Bobby is absolutely right.

I did miss the situation that he describes. It is important to protect the expansion drum from overpressure from the gas breakthrough.

I would suggest that there are two approaches to take to provide this protection.

 

First

The protections that I listed earlier should be installed. The source of gas to the ruptured exchanger should be shut off, and the exchanger should be isolated by ESDVs on the hot oil inlet and outlet lines. This provides two independent means to shut off the gas. These actions would be caused by high-high pressure in the exchanger, and high-high level in the expansion drum, thus, two independent instrument signals. Together the integrity of this system should be sufficient to avoid relying on the relief valve on the drum.

 

Second (alternative)

The protections described for the first option should be fitted, but instead of relying on integrity, the expansion drum design pressure should be raised to the system design pressure, eliminating the overpressure case. To deal with the pump shut-in head problem, described in the original post, the discharge system could be protected by an ESDV on the pump outlet (a large valve), OR dispensation to vary from the client standard should be requested on the basis that the high discharge pressure was only possible if a tube rupture occurred, and then the pump would be stopped by the instrumentation described above.

 

Sizing basis for the relief valve

If the drum has a lower design pressure than the system, the relief valve on the drum should be sized for the volumetric flow of hot oil which would be displaced from the exchanger by the tube rupture.  The valve sized for this liquid flow will be more than sufficient for the equivalent gas flow from the ruptured tube, but not necessarily for the gas breakthrough case which Bobby has correctly pointed out. That case has to be prevented by other means.

 

All comments on my proposals would be welcome. Reader should judge them carefully as I have already been wrong on this problem .

 

Paul

[hianbotech]

Thanks to all for the comments.

 

We already discussed with client, and tube rupture will be the sizing case for the relief valve to be installed in the expansion drum. The type of safety valve to be installed in the heat exchanger will be a pin rupture.

 

 

 

[paulhorth]

hianbotech,

 

That's great, so long as you also include all the protective features that I have described. Unless they are all present, your system is unsafe.

 

I agree that a buckling pin device is a better choice than a relief valve for the exchanger. Not a rupture pin - that's different.

 

Paul