Can a Thermal Relief Valve be Discharged Into the Pump Suction Line

[Berenger]

I have a facility set up as shown in the attachment. It is a crude oil pump station. Oil comes in, it is metered and it is pumped to another facility.

We have valves on the suction and discharge side of the pump. To account for thermal relief we have installed a TRV on both the suction and discharge piping.

Now, is it acceptable to discharge a TRV to the suction side of the pump as shown in the diagram attached? I know that a better option is a suction vessel or tank, but we don't have either at this facility. The suction piping is around 200 ft long.

Has anybody seen an arrangement like this? What is the norm.

Thanks.

 

[Latexman]

"We have valves on the suction and discharge side of the pump", yet the drawing does not show the discharge valve. Which is right?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[Berenger]

Latexman, thanks for the reply. I did not show the valves on the discharge line because I did not think it was important to the question. However, we DO have valves on the discharge line.

 

[Latexman]

So, how does the PSV relieve into a volume that is blocked in? It can't. If the suction and discharge valves are closed, the pump is blocked in. You'd have to relieve around all the discharge valves too, right?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[Berenger]

Yes, there is a TRV around the discharge valve too.

 

[Latexman]

Then, you may be golden. The backpressure in the new TRV's discharge piping will affect the opening pressure of a conventional safety valve. As such, the new TRV should be a bellows safety valve. Is the existing TRV a bellows safety valve or does it discharge to a partially full tank at essentially atmospheric pressure or something similar?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[LittleInch]

The key missing data here is design pressures and pipe class rating.

Unless they are the same then you would have no flow from the lower rated inelt/suction line into the higher rated discharge pipework.

Seeing only a small portion of the system makes it difficult to work this out, but cascading relief systems need to go down in pressure rating, not up. What is d/s of this inlet pipe?

It's also not normally great to pressure up a pump due to increased load on the seals.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Berenger]

The suction is around 100 psig and the discharge is 700 psig. The pipe ANSI ratings are higher on the discharge piping as can be expected (150 vs 600). However, the pump flanges are ANSI 900.

Also, don't overlook that there is a bypass around the pump. I assumed that even if we need to relieve when the pump is down, the crude can go through the bypass and we can use the discharge relief if necessary.

 

[LittleInch]

Errrr No.

You cannot relieve a low pressure system into a higher pressure rated systems as you have no control over the pressure in the downstream system.

Your discharge pressure PSV, assuming it is balanced bellows and not subject to downstream pressure, will be set a lot higher than 100 psig. Even if it is, your downstream pressure simply cannot be relied upon to be less than your upstream class rating ( #150).

Sure, a lot of the time it will be ok, but once, you could reach pressures of 1400 psig downstream the pump (otherwise why have a #600 system??) and that's what your upstream system could see in the sun.

Design this properly please.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

This sounds like it may be going off the rails. A PFD or P&ID may help us understand.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[LittleInch]

Latexman - that would always help, but I think this is just someone trying to be clever and not really understanding pressure system design.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[gocougs1]

Excellent question. In general it is not wise to have a TERV discharge to a pump's suction, as these systems may be isolated together. Ideally there should be no valves the process line it discharges into, to an upstream tank or downstream vessel.

If operating pressure is greater than 70% of set pressure, in liquid service, there is a risk that the valve may not reclose once it opens.

Consider setting the TERV to 20% above flange rating (allowable per ASME B31.3) and discharging to OWS.

 

[Berenger]

Sorry, was out of the country on vacation. Here is a PFD of what I've been trying to describe. The safety valve in question is the one in red.

As you can see, it is not discharging directly into the pump suction - there is a valve between the discharge and the pump suction. Also, note the bypass, so if there is thermal relief, it shouldn't have much effect on the pump suction side.

Please let me know what you feel.

 

[LittleInch]

Still wrong.

You cannot send a low pressure system relief into a higher rated system as the pressure in that higher rated system can exceed your lower rated one.

The line needs to discharge into the upstream system, which you haven't shown us or better still into a closed drain system.

I don't understand what the bypass sis doing or whay it is there, but actually makes things worse.

You really need to add where the pressure breaks are, but even then, see my previous replies - Don't do it. It is wrong.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[SNORGY]

The drawing looks like somebody's attempt to cascade all thermal reliefs towards what is, presumably, a buffer volume (the pipeline) big enough to take the incremental mass. But I agree with LittleInch, it's all going in the wrong direction, then - unless the entire suction system is rated to just as high a pressure as the pipeline. Some LACTs might be designed that way.

Presumably, the pump shown is a multistage centrifugal shipping pump. The bypass might be there so that operations can establish some pipeline pressure (maybe from an upstream booster, not shown if it exists) before starting up the main pump.

 

[SNORGY]

I've been in a few HAZOP reviews lately where part of the safeguarding against thermal expansion is afforded by an Owner's SOP (Safe Operating Procedure) that mandates the operators to crack certain valves open when the unit is down before completely closing the valves to block it in. It's a bit wonky for my tastes, but, in any event, that's what they sometimes do. Clearly, in this instance, either the pump suction block valve ought to be open or the bypass valve ought to be open for that to work, but again, the relief flow is probably going in the wrong direction.

The temperature at time of isolation might also be important. If you only block in when hot, maybe heat gain from ambient isn't enough to force the thermals anyway? That comment wouldn't apply if there are other mechanisms for heat gain besides ambient.

 

[vladrath]

Is there a tank upstream of the pump?

Typically in a situation like this I would want to send my relieved material back to the tank. But I am not sure if that is an option for you.

If all of these TRVs are in close proximity you might need to install a small catch tank.