BFW NPSH available problem 6

[flinana]

Hi,

I am having a problem keeping my BFW pumps from having to stop due to low NPSHR.

50MW solar powered plant. 4 BFW pumps, 3 normally maintaining level in Steam generator drum, 1 stdby.
The pumps NPSHR at 100% (3300 rpm) is 3 m. We are having great troubles in having a NPSHA above 3 at high loads. We have implemented an interlock to protect the pump so when the NPSHR is below NPSHA we trip the pump leading of course a a Turbine trip when we are around 47 MW.

Normally the pressure drops at the pumps suction is derived by an increase in energy from field causing a cold flow of condensate into the deaerator causing a or large movement from the HP bypass valve, pressure drop while the temperature remains stable due to the inertia of the system. This pressure drops lead to a drop in NPSHR tripping the pump.

How can I control this???, We have tried being very careful with sudden changes in flow, slowing closing ST HP bypasses, Slow increase in field energy, all to no avail, somehow we cant get to 50 MW because of NPSH. It is driving me crazy. Please anyone help? ideas? I can send more info of course

 

[rmw]

What is your NPSHa when the system is not in distress.

Describe the piping between the water source and the pump suction. A diagram would be better.

How high above the pumps is the deaerator.

Is it a classical deaerator, heating section, storage tank?

If your inrush of condensate is quenching the deaerator and causing the pressure to drop, the saturated fluid in the piping between the storage tank and the pump suction will flash, pulling steam bubbles into the pump.

Knowing more about your system will help others offer some good Eng-tips.

rmw

 

[MJCronin]

Your DA should have been sized for the inrush of makeup water......was it ?

Do you have a DA water storage tank ? Was it sized for 7-15 minutes pump running time ?

Purchase a bigger DA water storage tank.

Locate the bottom of the tank at least 5m above the pump suction.

How large is the suction piping connecting to the pump inlet nozzle...? It should be one, maybe two line sizes bigger than the nozzle....

Trying to answer complex questions withoutsome kind of a sketch or picture is nearly impossible...

-MJC

 

[flinana]

Please find the P&ID attached. It is a storage tank situated about 22m above the BFW. When not distressed we have around 22m
We calculate the NPSH a and compare it with the calculated NPSHr at different speeds. When the NPSHa is less than the NPSHr we send stop command.

The problem is also that the comparison is done with 4 decimals!! eg. Npsha 3.2345 > 3.2346, maybe this could be a problem since there isn't even a delay or hysteris.

 

[Chance17]

Taking your word literally, 22m is at DA bottom and there would surley be a vortex or suction being starved of liquid. I guessing the 22m is a DA with 50% liquid level.

 

[rmw]

OK, we are gaining ground, but have some ways to go. What size piping is between the DA and the pump suction. Is it one pipe for all pumps, or individual pipes. I don't necessarily infer that from a P&ID.

And, more importantly, how do you calculate NPSHa? I understood you to say that you determine NPSHr from the pump curve. Do you add any margin to the pump curve?

22 meters is equivalent to the height of a 6 story building. Is your DA really that high?

When your pressure drops due to the inrush of condensate, how low does the pressure go? Does the DA go under vacuum? Does the DA have a vacuum breaker on it?

And, is this a solar power plant or a Solar power plant, Capital "S" Solar being a brand of gas turbine manufactured by Caterpillar little "s" being a sun powered plant. Just curious - it doesn't help solve the problem.

rmw

 

[flinana]

NPSH a = 100000/9.81/(-0.0022*Tsuction^2-0.2714*Tsuction+1007.2)*(Psuction-(0.000006*Tsuction^3-0.0012*Tsuction^2+0.0958*Tsuction-3.4232))

This is nothing more than -0.0022*Tsuction^2-0.2714*Tsuction+1007.2 being density
and Psuction-(0.000006*Tsuction^3-0.0012*Tsuction^2+0.0958*Tsuction-3.4232 being vapour pressure

NPSH r is calculated from the curve attached, at different speed different NPSH r.

The DA is 22 meters high. Individual Pipes are 6 inches.

When your pressure drops due to the inrush of condensate, the pressure drops around 0.4 barg. The DA is not under vaccum it does not have a vacuum breaker.
Temp around 177.6 ºC and pressure 8.2 barg, a 0.4 drop in pressure cause a drop of about 5 m in NPSH a.

 

[sshep]

flinana,

The drawing was helpful but lacks resolution on required detail.

1. How are the pump minimum flows being controlled?
2. What are the lines tieing into the pump suction from another page? Is this some injection of cooler condensate to give more NPSHa?
3. What are those block looking items in the suction line?Are they strainers with a dP measurement?

I have been in some root cause analysis investigations on BFW systems that look very similar where pumps were being regularly wrecked. Perhaps it is a common system for design problems- i.e. it is not simply sufficient that the overall NPSHa is calculated to be enough. This is boiling liquid so you also can't take hardly any pressure drop at the bottom nozzle of the drum going into the suction line. If you get even the smallest cavitation at that point (where very little liquid head is available), then your pump will be starved and cavitate regardless of your overall NPSHa calculation.

best wishes,
sshep

 

[flinana]

1. Minimum flow is controlled through a minimum recirculation.
2. These lines a dosing lines
3. Those blocks are strainers with a DP measurement.

I attach the PID in dwg.

I know that in other future plants the design will be different, the pipes have been increased to 10" and the DA placed 27 meters above. The main problem we have here is that it is very difficult to reach 50 MW so we are limited to 45/47 MW. I am now trying to tune the valve after the condensate pumps which controls the level in condenser which in turn affects the flow into the DA.

 

[sshep]

Sorry, I didn't have a program to open a dwg, but your answers make it pretty clear except the min flow.

Is the minimum flow just an orifice plate? If so, then how much flow (and dP)? Changing this to another method (either control valves, or an automatic recirc valve) could be a cheap piping and control fix to unload the pumps and reduce the NPSHa. Just a thought.

best wishes,
sshep

 

[flinana]

there is only orifice plates, not sure what the dp is, wasn't my design.
95% of the trips we have are caused by NPSH protection. Today for example the problem arose when one of the HP ST extraction to one HP preheater opened after closing due to high level. This caused the NPSHa to lower below the NPSHr.
At around 50 MW the NPSH are very close together and a slight change in load, or operation affects the NPSH protection. It is causing us a fortune, we are talking that an hour at 50MW is worth 15000€

 

[rmw]

For what reason do you have strainers in your BFP suction lines? Very bad idea, even with the strainers clean. What part of the process dictates that you need to strain BFW? Can you strain on the inlets to the DA rather than the outlet?

I'd start trying to remove every resistance possible in the BFP suction lines, and I'd begin with the strainer baskets or screens. I would look for other non essential items in the suction lines that could be removed or altered to reduce the DP in the suction line. The luxury of a block valve may be more than outweighed by the money you are losing in lost generation due to the trips.

YOu posted another thread about having trouble controlling heater levels. Are the heaters creating these condensate surges that are knocking your pressure down? Seems like the answer might be in tuning your heater level controls or finding out what is making those levels hard to control.

I can't open the .dwg drawing here. I'll have to do it from my work computer.

Second, and I am just asking for information-not necessarily suggesting- when compared to a turbine trip, lost MWs, etc, could you live with a little cavitation for short periods during these periods. Seems like 15,000? (sorry, I am not familiar with that symbol) per hour might buy quite a few pump impellers during your periodic turnarounds.

Speaking of pump impellers, have you looked at alternate impellers. Sulzer might have a pump that would run at a different speed (or a rotating ass'y for this pump) that would need less NPSHr. Seeing that the pump speed is 2500 rpm indicates to me that it might be ST driven; yes-no?

You are giving great information and I think we might be gaining on getting somewhere.

rmw

 

[rmw]

Then I remembered I worked from home today so my work computer was still fired up in my office. Looking at the .dwg P&ID, I can see that the only purpose that the valve in the line seems to serve is to allow you to open the strainer. So, if you determine that you don't need the strainer as per the last post, you don't need the valve either. Put a spool piece in its place.

Can't see heater #3 on this diagram, but does it have a subcooler too? If so, you might want to consider bypassing it a little to reduce the FW flow through it and reduce the subcooling. That will help your condensate be warmer and less prone to quench your DA when it hits.

rmw

 

[flinana]

I meant 15000 euros! and the problem is that these pumps since they are multistage are very delicate and dont feel comfortable in breaking them up just yet.

These pumps are from Sulzer with a VFD. The pumps run up to 3300 rpm. I have done a polynomial where i plot a NPSHr for each curve give, 1500, 1900, 2300, etc.. and use that equation for comparison versus NPSHa.

for example NPSHr for 1500 is around 1m and for 3300 3m. These values I take from the pumps curves choosing a normal flow.

Regarding the heaters, yes that is a problem specially since some of the evacuation piping i a bit undersized. Although LP heater 1, 2 and 3, trip the turbine for high levels, heaters 5 and 6 cause changes in pressure and temperature in deaerator since the heater 6 evacuates to 5 and this to the DA, both have a of course an emergency line to the condenser which in turn suffers high levels when receiving flow from heaters emergency lines.

The valve controlling this level is the one that feeds flow to the DA through lp heaters 1, 2 and 3 so basically I need this valve to not only control the condenser level but also make sure that this inrush condenser flow to DA is a stable as possible, AHHHHHH, complicated huh? all our heaters have a subcooler btw.

As for the strainer I have thought of that, I am just waiting to make sure the water is 100% clean and also the plant is new so i don't want to modify too many things still under warranty. But in future my plan is to remove as much as possible from the line.

I like the idea about bypassing LP heater 3, I didn't think of that. Heater 2 goes to Heater 3 so I guess you mean bypassing some of this flow so as to get a a bit higher temp favouring npsha in DA. Am i on the right track?

thanks a lot by the way!! You are great.

 

[BenThayer]

if the orifice is sized for minimum flow at max speed, that is likely too large for normal operation. since min flow is likely proportional to rpm's, you may need a very small orifice. that is, if you are at high speed, it is because the flow is required to the boiler.

confirm the flow through the orifice and then confirm the NPSHr for the total flow through the pump (delivered to boiler plus flow through orifice).

 

[flinana]

Please find the rest of the involved PIDs for those interested.

 

[flinana]

now hp

 

[flinana]

and condensate

 

[MJCronin]

Where is this plant located ?

 

[flinana]

Plant is located in Spain. Thank you all for the replies.
After spending hours working out a solution with your help being very very useful I have two people to blame for this. One is Bill Gates and the other is me for not checking. It so happens that according to microsoft and I quote

"In Microsoft Excel, when you add a third-order polynomial trendline to an xy (scatter) chart, and you display the equation, both the trendline and the equation displayed on the chart are incorrect. "

When I calculated the NPSHd I only used around 2 decimal places, it so happens that one solution is to use many more, I have uses up to 8. Now it makes more sense that instead of going down to 4m at around 177ºC, 9,8 barg at approx 2900 rpm,I know get 14 m!!!

I still have the fluctuations problems of course but I have more margin to work on. Now I need to work on the heater level problems which need to be properly tuned I guess and studied. But thats another thread. If anyone needs it, I have the NPSHd formula which is pretty handy.

thanks!!!