6 replies to this topic

[ppister]

Dear all,

 

I am a process engineer with 4 years of experience in project designs and I was asked to evaluate an existing system of a propane/propene sphere that is the feed of a distillation unit.

 

The problem is that, according to the unit operators, in very hot days or when the sphere pressure is very low (under 10 bara), the sphere pump (feed pump of the distillation unit) shows evidence of cavitation or in most cases complete vaporization of the suction line.

The operators told me that when the problem was detected, they opened a suction vent line to the atmosphere and the gas exiting was hot, indicating that the suction line was filled with gas. If the fluid was in the liquid state, it would expand, decreasing temperature and even freezing the line.

So the problem is more vaporization of the fluid than cavitation.

 

System data:

The fluid is around 65% propene, 33% propane and 2% C4 (volume basis). There are 2 spheres, connected to each other by an equalization line on top, that operates normally at 12 - 14 bara and have a minimum liquid level of 4,5m and a maximum around 11m.

The pump is located in a distance of approximately 300 m, and according to my calculations the accidents sum an equivalent length of 600 m. Medium flowrate is 35 m3/h.

The sucction line (12"/10") is exposed to the sun and the ambient temperature vary a lot, reaching 35 - 40 C in hot days.

The whole line is insulated, but there are some parts where the insulation is damaged and some accidents (valves and filter near the pump, but located under a roof) are not insulated. I honestly do not know if the insulation is the proper one or if the thickness is correct.

 

Question 1:

I am considering that the mixture inside the sphere is saturated at its bubble point, as in LPG spheres. Is that correct? Or is it possible that this mixture is stored in a subcooled point without sphere refrigeration?

 

Question 2:

If the insulation is perfect, the temperature at the pump suction nozzle is the same as the sphere bulk liquid. In that case, the NPSHa is calculated based only on elevation difference and head loss, as the term "P - Pv" in the NPSHa equation is equal to zero. Am I right?

 

According to my calculations, the head loss is very small (around 0.006 bar) and even at the lowest liquid level the NPSHa is sufficient (considering that the insulation is perfect).

However, when I consider the minimum liquid level and that the fluid heats in the piping, my calculations show that a difference of only 1 C between the sphere bulk liquid and the fluid in the piping is sufficient to reach the mixture bubble point. And a difference of 2.5 C to reach the dew point and vaporize completely.

I do not know how much temperature is gained considering inexistent insulation.

 

Question 3:

Does anyone have a sugestion to improve this system? I have some thoughts but I do not know which is the most viable:

 

- Improving the insulation system (I do not know how).

- Locating the pump to a place closer to the sphere so that the suction line is not exposed so much to the sun.

- Installing a booster pump right next to the sphere so that it raises the fluid pressure and so the bubble point temperature

- Changing the mixture to a closer sphere and with a higher elevation to the ground (there are some in the area). But according to my calculations, increasing the elevation difference in 1 m have little impact on the temperature difference needed to start vaporization.

- Pressurize the sphere. According to my calculations, if the sphere was always mantained at 14 - 15 bara, the vaporization problem would pratically disappear as it would need to gain too much temperature to vaporize. But I do not know if this solution is possible. If it is, how would be done? I only heard of pressurized LPG sphere with nitrogen.

 

Sorry for the long text.

 

Thank you.

 

[S.AHMAD]

Ppister

Question 1 - I agree with you

Question 2 - I have the same thought.

Question 3 - Could please send some schematic diagram with all relevant data shown? You need to identify the ROOT cause before making any modification to the existing unit.

Edited by S.AHMAD, 10 December 2013 - 04:01 AM.

[curious_cat]

Some thoughts:

 

(1) What's the NPSHr of the pump?

 

(2) Is it only a startup issue? What may happen is stagnant fluid heats and vaporizes in suction line? 

 

(3) There's also an acceleration head at startup which may induce vaporization. After venting is problem solved or does it recur? 

 

(4) What's the thickness / material used for insulation? 

 

(5) Do you have Temp. probes at any points in tank ./ suction lines?

Edited by curious_cat, 10 December 2013 - 08:40 AM.

[ppister]

S.AHMAD,

I don't have any diagram. All available data is presented in my topic.

I believe that the root cause is vaporization of the liquid in the suction line due to inappropriate insulation. But I am not 100% sure yet, because there are some hot days in which the problem does not occur.

 

curious_cat

1) NPSHr = 1,8 m (already considering hydrocarbon)

2) This is a valid point. But the problem occurs when the pump has been operating for a while.

3) After venting the suction line the problem is solved and the pump operates normally. I don't know what you mean by "acceleration head at startup". Can you please explain me or show me some reference?

4) The insulation material is a type of foam. Thickness is around 1".

5) There is only one at the sphere but it does not look so reliable. In a day that the pump was operating normally I took a pyrometer and measured the temperature of the sphere and the suction line near the pump (in a spot that there was no insulation). Considering the instrument error, the values were the same.

 

I am trying to think of other possible variables because there are some hot days that this problem does not occur, even with low liquid level in the sphere.

 

What do you think?

 

Thanks,

[curious_cat]

Another idea is that it's a composition related issue?  Can there be some intermittent lights that flash?  That'd explain a lack of a pattern. 

[S.AHMAD]

ppister

1. One possibility is the pump eye is too small that causes too high pressure drop.

2. This can be solved by changing the pump with double suction.

3. Installing inducer  can also help.

[Art Montemayor]

Ppister:

 

Before going on any further, it would wise and thoughtful of you to submit some additional and required information:

 

• a detailed, isometric sketch of the pump's suction piping (here, I'm assuming you are using a centrifugal pump);
• the rated storage temperature (and pressure) of the Propane+Propene mixture.

 

The reason for the above needs is, I believe, obvious.  If you are storing at ambient temperatures, then your storage pressure should correspond to the saturated vapor pressure of the mixture.  You cannot have gas in the sphere except at the top vapor space.  That is the nature and scope of the storage system.  Any conventional centrifugal or gear pump should easily pump this mixture out - as long as it is in the liquid state.  Temperature is not an issue, as long as the saturated liquid exists.  Therefore, you should not require insulation.

 

Await your response.