String Testing of a Lube Oil System

[RogerSAM]

Hello everyone,

I have a specific case that I needed some insight on since I am a relatively new engineer who is still learning.

We conducted a string test to check the response of a lube oil system that consisted of two pumps - 1 main and 1 standby. The main turbine driven pump was manually tripped and the system transient response during the stand-by's cut-in period was verified. The test was successful and header pressure did not drop to alarm levels during the test.

Now the question is around 5 minutes or so before the test, both the pumps were operational. The stand-by pump was then shutdown to conduct the test. My senior is telling me that since the stand-by pump was just recently shut off, it doesn't simulate the real behaviour because during the normal operation, the stand-by would be off for a long time. So although in the test, it generated pressure fast, it wont necessarily do it in the actual real world scenario. He's talking about breaking of the inertia of the fluid etc. that just doesnt make sense to me.

Is he right?

 

[seuenergy]

hi,
Could you provide the P&ID of the lube oil system?
Currently, I could not get your senior's point. Is he talking about the filling conditions of the inlet pipeline of stand-by pump is different for long term shut down and short term shut down?

 

[RogerSAM]

hi,
Could you provide the P&ID of the lube oil system?
Currently, I could not get your senior's point. Is he talking about the filling conditions of the inlet pipeline of stand-by pump is different for long term shut down and short term shut down?

Please find the P&ID attached.

Both are the same.

Here is the full breakdown of the problem:
We recently experienced a tripping on the main lube oil pump motor. This led to the whole machine tripping within 1.5s while the standby pump recovered the pressure within 1.8s.

So obviously we suspected problems with the accumulator and turns out they only have an oil capacity of 21 liters while being responsible for both the hydraulic oil and the lubricating oil. Not to mention, their location is not very ideal as their oil will have to go through a bunch of filters and exchangers before serving their purpose.

We conducted the string test back in 2022 which somehow passed so we were looking for reasons as to why the transient response of the LO system was adequate then but it isnt now. Thats when my senior mentioned the reasoning that standby pump had been recently taken offline before the test which could contribute to the overall pressure recovery in 2022.

Are there any other deficiencies that I am missing in the P&ID that may be contributing to such an erratic transient response of the LO system? Something that is not in line with industrial best practices?

p.s for your convenience, i have highlighted the LO lines in the P&ID

 

[LittleInch]

The standby pump is steam driven according to the p & ID so there may be a difference between starting from warm and starting from cold.

That sure is a complex lube oil system you have there.

When it tripped were the accumulators at the right pressure and volume? If you start with too high pressure you don't fill the accumulator with the right amount. Ditto too little pressure and you fill more, but it won't maintain the pressure.

What sort of accumulators are they?
Whatever they are they are clearly about a quarter the size you need if together they are only 21 litres/ 5 us gal when your pumps are capable of 285 gpm, that's 1.1 seconds supply....

Not surprising you tripped.

 

[RogerSAM]

The standby pump is steam driven according to the p & ID so there may be a difference between starting from warm and starting from cold.

That sure is a complex lube oil system you have there.

Standby is motor driven. By our data, the auto-pump start (APS63QA) right at the discharge of the pump normalized 1.5s after actuation. So that is a time of 1.5s for the stand-by to start building pressure. But The LO system should still be designed for 4-6s transient oil supply if standby is motor driven. >15s if its ST driven.

Its a vintage machine originally installed somewhere in the US back in 1960-70s before being transferred to where Im at around 1990s. The original P&ID from back in 1970s doesnt even have the accumulators. God knows who had that changed. We have no surviving data on the accumulators. The original P&ID shows standby as ST driven and idling at 1800 rpm. Afaik ST driven pump has never been operationally kept as standby in this machine until ST was replaced by motor in 2022. Standby has always been motor driven.

When it tripped were the accumulators at the right pressure and volume? If you start with too high pressure you don't fill the accumulator with the right amount. Ditto too little pressure and you fill more, but it won't maintain the pressure.

What sort of accumulators are they?
Whatever they are they are clearly about a quarter the size you need if together they are only 21 litres/ 5 us gal when your pumps are capable of 285 gpm, that's 1.1 seconds supply....

Accumulators are bladder type. Yes, thats the conclusion we derived that 1.1s is insanely low. Also, the accumulators are being pre-charged to 50 PSIG which is exactly what the Stand-by pump cut-in point is at. By our estimate of all the pressure drops in the circuit, a tripping pressure of 8 PSI at the farthest end equates to 30 PSIG at the accumulators inlet. What should be the usual pre-charge pressure? In one paper I read it should be 80 - 85% of the standby cut in. In another it was 90% of the minimum system pressure at the accumulator inlet.

Also, what other deficiencies exist in the system? The few I was able to figure out are:
1. Location of the accumulators. Oil has to pass through filters and exchangers to go to the hydraulic circuit or to the brgs.
2. Size of the accumulators is inadequate.
3. Pre-charge is not correct. I asked the operators responsible and the only data they had was that this is how it's always been done.
4. Stand-by pump cut in set point should be increased too to allow more time for the stand-by to build pressure.

Not surprising you tripped.

Yes, not at all. I am just trying to accumulate (no pun intended) all the reasons that contributed.

 

[LittleInch]

Pre charge pressure, I.e. bladder pressure when no liquid in the accumulator I think should be a bit lower than the min pressure you need, so the 90% sounds about right in order that you can actually increase the amount of oil stored within the accumulator still at a decent pressure.

So your point 3 I think is correct and this should be less than 30 psi.

If your normal op pressure is 65 psi, the you don't want the second pump starting too early either though so your point 4 is maybe not correctmm if the trigger point from that new pump start is actually 50 psi.

Any idea why you have two,sets of lube oil coolersmone after the other?

Just d/s of those filterss? K2541, what exactly do those two ROs and the control vavel actually do? I can't see why they are there.

 

[georgeverghese]

Strange, are those "N2 bottles" these bladder accumulators ?
And your rundown tank appears to be on gravity based feed into the lube oil circuit - the ones I've seen are pressurised rundown tanks.
If the rundown tank level is low in this gravity setup, then the feed pressure at the lube oil supply header could also be low ? There is no low oil level switch on this rundown tank either.

 

[RogerSAM]

Pre charge pressure, I.e. bladder pressure when no liquid in the accumulator I think should be a bit lower than the min pressure you need, so the 90% sounds about right in order that you can actually increase the amount of oil stored within the accumulator still at a decent pressure.

So your point 3 I think is correct and this should be less than 30 psi.

If your normal op pressure is 65 psi, the you don't want the second pump starting too early either though so your point 4 is maybe not correctmm if the trigger point from that new pump start is actually 50 psi.

In this case, shouldnt we want the standby to cut in early? The earlier it cuts in, the fast it will build the required pressure. By our data the difference between trip actuation and the pressure recovery was less than 0.4 seconds. Even a 5 PSI increase in cut-in set point could make a substantial difference. Is my thinking correct in that regards? (we will be testing this in the upcoming shutdown though)

Any idea why you have two,sets of lube oil coolersmone after the other?

Just d/s of those filterss? K2541, what exactly do those two ROs and the control vavel actually do? I can't see why they are there.

Originally only the A/B units of coolers were in service. However due to operational requirement, the D unit was installed and some genius installed it downstream of the accumulators instead of right next to the A/B units. The reports I read indicated they had a physical space issue and this was the only option left.

Header pressure in normal operation is 17-20 PSIG. ROs & the VPR-2 control that header pressure. These are flow regulators. Pretty common in lube oil designs. The issue could be in the VPR as well but we havent gotten around to inspect it as we're waiting for a shutdown opportunity. But I doubt its that.

Strange, are those "N2 bottles" these bladder accumulators ?
And your rundown tank appears to be on gravity based feed into the lube oil circuit - the ones I've seen are pressurised rundown tanks.
If the rundown tank level is low in this gravity setup, then the feed pressure at the lube oil supply header could also be low ? There is no low oil level switch on this rundown tank either.

rundown tank is designed to only supply oil when the trip has actuated in case of complete lube oil failure. A tripping back in 2000s had damaged the bearings so they decided to install the rundown tank to atleast save the bearings. It is gravity based, and the calculations are solid, I've checked.

 

[georgeverghese]

Okay, thanks for reminding me-the rundown tank is only meant for emergency cooling of bearings during a total stop.
Drawing shows no check valve or trip valve on the rundown line leading into the main oil supply, so what stops oil from reversing into this rundown tank in normal operation ? Without a check valve or similar, oil pressure in the header would equalise with rundown tank static head pressure.

 

[LittleInch]

Set pressure for the start of the spare could be higher but is acompromise between rapid starting and starting due to a small blip in pressure. If you are confident that in normal operation the lube oil pressure is constant plus minus a couple if psi then adjust the start pressure accordingly.

 

[RogerSAM]

Okay, thanks for reminding me-the rundown tank is only meant for emergency cooling of bearings during a total stop.
Drawing shows no check valve or trip valve on the rundown line leading into the main oil supply, so what stops oil from reversing into this rundown tank in normal operation ? Without a check valve or similar, oil pressure in the header would equalise with rundown tank static head pressure.

Apologies if the comment sounded demeaning, wasnt my intention. I really appreciate all the help. Rundown tank is pressurized and its height is designed to have a max head of 6.5 psig. The overflow line at the top ensures 3% recirculation to keep the oil in the rundown tank at optimum temp. I believe the same line is being used as both inlet and outlet. I dont see an issue with that given that its designed to start supplying oil when the header pressure reaches 6.5 psig decreasing. Perhaps you can correct me.

Although now that I look at rundown tank more closely, I do see two problems:
1. No check valve upstream of the rundown tank. What will ensure all the oil travels to the machine and not towards the pumps?
2. Oil from the accumulators may go towards the rundown tanks. As the circulation is kept at 3% of pump discharge, oil demand for accumulators is now 285 gallons + 3% of 285 = 293.5 GPM.

Set pressure for the start of the spare could be higher but is acompromise between rapid starting and starting due to a small blip in pressure. If you are confident that in normal operation the lube oil pressure is constant plus minus a couple if psi then adjust the start pressure accordingly.

Yes, and according to our evaluation, 55 PSI should be a good compromise.