Problems with a liquid propane loading rack

[rocketscientist]

I have a difficult problem. Liquid propane is sent to two pipelines or a tanker truck rack. The suction pipeline to the pumps is uninsulated. There is no pressure control on the propane bullets. The pumps are fed 5 at a time to a pump through a single 4" sch 80 bottom connection for each bullet. The pipelines are buried except for the one to the truck rack. We have trouble pumping from the bullets as level drops below 45% when the pressure is about 110 psig or lower. The operators seem to think that ambient temperature is the source of their troubles. They have no trouble with the pumps when they operate below 70 F but by mid-afternoon in the spring or summer the pumps can't deliver to the truck rack. The buried pipelines still have trouble below 45% level in the bullets.

I looked at the situation and concluded the following problems after modeling the system with Hysys: 1) As the pressure drops, the saturation temperature rises until mostly vapor is flowing to the pump; 2)If flow can be maintained, the auto-refrigeration from pressure drop partially counteracts the heat transfer to air; 3) The heating of the liquid caused by pumping saturated liquid/vapor through the pump bumps the pipeline temperature up and produces more vapor.

One of the engineers seems to think that one approach to the problem is to change the flow controls at the truck rack. He wants a linear valve in place of an equal % valve and now wants to install 2 smaller valves in split range or sequence (I suggested the later) controlled by a new DP transmitter. I suggested moving the flow orifice upstream of the flow control valve. Hysys doesn't indicate choked flow in the line after the control valve even when I force the pump to the limit on suction vapor content.

I think the heart of this problem is the lack of pressure control on the bullets. There is no pump line back to the bullets. Instead, the operators flare the gas from the pump suction line to a flare.

What do you think? Any ideas?

 

[Compositepro]

The liquid propane in the bullets is typically at its boiling point so the suction head available to the pumps (NPSHA) is only the liquid head in the tank. This must overcome any pressure drop in the suction pipe and any increase in vapor pressure of the fluid in the pump and piping due to a temperature increase (sunshine, pump heating). For this reason pumps are usually located at the tank with a recirculation line back to the tank.
You could increase pressure in your bullets with a heated vaporizer.

 

[GHartmann]

You are on the right track, but not quite there yet.

In theory the pressure in the tank should be completely dependent on the temperature and not the amount of liquid in the tank. For 70 deg F the tank should be at your 110 psig and at 90 deg F the tank will be at 149 psig. Thus, the previous poster is correct that the available pump NPSH is set by the tank level.

If the pressure in pump discharge is greater than the tank pressure then there should be no flashing in the discharge line. (The tank level must be high enough to prevent flashing in the suction line; pressurizing the tank will just increase the tank temperature - unless you can determine that the vapor and liquid are not in equilibrium - which could be possible).

I want to think that at the higher temperatures there is increased flashing across the control valve and perhaps the control valve was not sized for two-phase flow. However, it was stated the pumps were not delivering which goes back to the NPSH problem.

As far as the pipelines are concerned its not clear if you are also pumping to them. If yes then this again seems to be a NPSH problem.

Good luck.

 

[TD2K]

It's not uncommon for propane bullets and the suction piping to the pumps to be uninsulated. It's also not uncommon for there to be no pressure control on the bullets. The tank design pressure is usually 250 psig or higher so within the range of ambient temperatures you don't have to relieve through the PSVs or whatever overpressure devices you have. As said, the propane in the tanks is at its bubble point in equilibrium with the vapor. If you can bring in higher pressure vapor you can initially have the vapor not in equilibrium with the liquid but it will equilibrate with time. When you do pump calculations on this type of system, you would assume saturated liquid and only take credit for any elevation head (less line losses) in estimating the NPSHA. Propane loading pumps are often can pumps so the height of the buried barrel can provide additional NPSHA. If they are conventional horizontal pumps, the NPSHA requirement will result in the suction vessels being significantly elevated. For can pumps, I will typically specify 0 NPSHA at the inlet flange and then let the pump vendor determine how much barrel depth they need to provide the NPSHA. I then make sure my pump suction piping is large enough so the line losses are less than any elevation gain I have. Basically, if the bottom of the bullets is 4' above the pump inlet nozzle centerline, I want my suction piping to take less than 4' pressure drop.

I'm curious you don't have a problem with cooler temperatures and the pumping issues only come up when the ambient temperatures go up. Higher ambient temperatures means higher suction pressures. When you are loading a truck, where is it vented to? What is the pressure inside the truck? Would a higher suction pressure to the pump reduce the dP the pump operates against and causes the pump to run out on its curve? Does the loading control valves you mention maintain a flow or do they run wide open? A higher flow would increase the NPSHA requirement and could get you into a cavitation issue where you wouldn't see it at a lower suction pressure.

With respect to your Hysys modeling:

"As the pressure drops, the saturation temperature rises until mostly vapor is flowing to the pump"

Lowering the pressure does not increase the saturation temperature. Also, any significant amount of vapor in the suction line (indicative of insufficient NPSHA) is going to result in the pump not pumping. I think you are pushing your simulation, just because Hysys will send a significant amount of vapor to a pump doesn't mean the system in the field operates that way.

"The heating of the liquid caused by pumping saturated liquid/vapor through the pump bumps the pipeline temperature up and produces more vapor"
I'm not sure what you are modeling here but a pump simply won't pump that much vapor.

 

[GHartmann]

TD2K said much better what I was trying to say.

It could be that on a hot day, once any liquid gets into the truck it creates too much back pressure for the pump to overcome.

Typically a propane loading station would vent vapor back to a compressor that would then send the vapor back to the tank. This type of system could load trucks completely without any pumps.

TD2K was further on the mark about the higher suction pressure causing the pump to run out its curve if you have a fixed discharge pressure.

This seems to be the case feeding your underground pipes.

 

[Compositepro]

Most propane pumps will eventually run okay if the piping is designed in a way that won't vapor lock the pump. There will often be some vapor generated in the suction piping. If it can pass through the pump, then no problem. But if the vapor displaces liquid back to the storage tank then the pump will not prime.

 

[TD2K]

Do you pump simultaneously to the pipelines and the truck rack?

When you say you can't deliver to the truck rack during hot weather, is there no flow to the truck rack or you can't get the desired flow or there is some other operational issue that prevents you from delivering to the truck rack?

 

[SNORGY]

What kind of pumps are being used and what colour have the bullets and suction piping been painted?

 

[rocketscientist]

Although the idea of the tank static head is a good one, I don't see how you overcome the vapor pressure of the propane (97.5% by mole) once the pressure drops.

I agree with you that the vent system for the truck racks is not very good. Ideally, you would vent to flare until the liquid flow is established. This would get you through the difficult time with the tanker at say 50 psig during startup.

The trouble is that the operating engineer thinks the pump problem is with the truck rack flow control valve. He thinks there is a vent line at pump discharge that opens at about 270 psig preventing the pump from achieving a high pressure thereby resulting in a low dP across the FCV. Otherwise the CV would be small with 400+ psig across the valve; with a flow of 600 gpm propane.

I ran another simulation with Hysys. This time I purposely ran the pump to the limit with the bullet feeding it at about 110 psig and 78 F. Above that temperature you're trying to pump gas. I set the ambient temperature at 78 F and the molar vapor liquid ratio (V/L) was about 5% at the inlet to the pump. The control valve saw about 3% at the inlet and about 6% at the tanker truck. What happened was the gas caused the pump to heat the liquid raising its temperature to about 105 F even though there was obvious auto-refrigeration.

 

[rocketscientist]

Here's a problem I did win acceptance for: moving the flow orifice. Right now it's downstream of the flow control valve; dumb right? Now, if I can talk them into a more forgiving type of flow measurement. A vortex shedder might work or possibly a vortex. I'm still thinking about that.

 

[GHartmann]

Lets try this again. No matter the pressure, temperature, or level, the propane in the tank will be a saturated liquid (with the temperature and pressure dependent variables). The head in the tank is necessary to protect the pumps from vapor in the suction.

However, as one poster suggested, propane pumps are able to pump this saturated liquid.

600 GPM is a pretty high flow through a 4" suction line to the pumps.

Perhaps you need to look at the pressure dynamics in the truck. First the initial liquid to the truck must start vaporizing to pressurize the truck. Then once the truck reaches about 110 psig the liquid will stop vaporizing. The remaining liquid will begin to compress the gas until the desired pressure is reached. At this point some gas may have to be vented to fully load the truck.

With such a high loading rate the truck may not initially be in equilibrium.

Seems to me a good engineering project would be to set up a vapor recovery system to avoid your losses to the flare header (sorry for the sidetrack)

 

[Compositepro]

Propane is USUALLY a saturated liquid. In the mornings the liquid will have cooled at night and the vapor space will warm much faster than the bulk liquid, so the tank pressure will be higher than the vapor pressure. This same effect can be achieved with a small heated vaporizer, as I mentioned before. Warm gas stays at the top of the tank and does not quickly equilibrate in temperature with the liquid. Temperature stratification can easily occur in both the vapor and liquid phases.

 

[rocketscientist]

To GHartmann; I agree. That's part of the problem (I think). A hot truck has a high backpressure. They probably don't have it vented properly. (I've been trying to watch a truck loading for 3 months. I guess I'll wait until fall. All I have now are some of the PI trends.)
To GHartmann; always a liquid --- I don't think so. The physical state of a compound, especially a mixture, will depend on temperature and pressure. Bottled up in a bullet, the liquid propane will equilibrate with its vapor. But, there is no such thing as a closed system. Does an ice cream bar stay solid inside a thermos --- nope. As for a bullet feeding a pump the propane will gradually convert to gas as the bullet is emptied. The pressure drops and there is less liquid propane to evaporate and provide cooling.

To GHartman: I think you have a point regarding the tanker truck. As I said the vent system is inadequate.

FYI: I'm talking to Rosemount about the vortex shedding flow meter. It shows real promise in increasing the measurement span while reducing the effect of vapor in the liquid propane.

 

[GHartmann]

Sorry, I should clarify. When there is liquid in the tank it will normally be saturated liquid.

I suppose when you pull 600 GPM from the tank, you can depressure the tank quicker than the fluid can vaporize itself, but I think the liquid vaporizes quickly to match the lower pressure. I suppose then the cooler liqud temperature equilibrates with the vapor at some lower pressure at a temperature lower than ambient (auto-refrigeration).

But, none of the above helps explain your pumping situation.

I don't believe the control valves and metering are your issue. But your revision to the metering station is good. See also attached reference regarding mass (corilois) meters.

Somehow the higher ambient may be causing a pressure in the pump suction or impeller eye that is higher than the pressure in tank and blocks the suction flow. See this article:

http://findarticles.com/p/articles/mi_m3251/is_9_227/ai_n25030741/

It appears to be pretty common that the trucks are loaded without any vapor return to correctly "inventory" the feed to the trucks for custody control. If you start with a atmospheric truck the liquid can fill the truck to compress the vapor fill to 15/265 psia or 6% of original volume.

Good luck.

 

[rocketscientist]

GHartmann: I agree. The client is making me change the loading station. I don't think there will be much of an improvement in the pump problem, if any.

My plan, at the moment, is an old one. I will try to line up the pump manufacturer's tech support for a visit. They were not too happy with the pump installation so said the area foreman. We'll promote the manufacturer's ideas while I look into one of my own.

I will go back to the three (?) units feeding the bullets. I think improving the pressure control feeding the bullets could reduce the NPSH problems at the pump. If I can get them to feed the bullets while they're being drawn down, I can reduce the fall-off in bullet pressure from a jump off a cliff to a gradual descent.

The truth is that this is a minor problem at the refinery. I have 25+ other projects and a staff of engineers with 25+ projects each that demand my attention at the moment.

 

[MortenA]

The strange thing to me is the 45% level in the bullit tank and the 70ºF ambient limits.

The problems at the high temperature would best be explaind as many have noted by high pressure in the trucks caused by an empty hot vessel on the truck. The pressure in teh bullit will be much lower since this has a lower temperature due to large amount af liquds in the bullit.

But that you have to have a combination of low level and high ambient seems strange unless its not sharp 45% but correlated to ambient temperature - but "around" 45%

Best regards

Morten