Vacuum in gravity flow

[mcmg]

Hi. I've read a lot of valuable information regarding this topic on this forum, but I am still quite bothered with the problem I am currently handling. Hope you can all help me. Posted this topic too on cheresources.

The scenario is: A closed loop system with 46wt% water and 54wt% ethylene glycol used for cooling the process side: a pump takes in -20C liquid from a tank (both at ground) and supplies the liquid to 2 heat exchangers (HE1, HE2) located approximately +12m above ground. The outlet fluid (-14C) is used as cooling jacket fluid for accumulator tanks (T1, T2) located below the exchanger (located at ~8m above ground). The exiting fluid from these T1/T2 (at -10C) flows to a refrigeration system located at ground to cool back the fluid to -20C. This goes back to the tank (N2 blanketed and vented to atm), via dip pipe (all the way below near the tank bottom).

Observations and Data:
1. By design, the flow rate should be approximately ~18,000 kg/hr and discharge pressure at pump should be ~320 kpag. However, the flow we are getting is ~25,000 kg/hr and discharge pressure is 230 kpag. We tried to throttle the valve outlet of the pump, and we are still experiencing high flows as compared with the theoretical (even if the valve was nearly 80% closed).
2. We conducted pressure profiling of the system and was surprised by appearance of vacuum at certain points (at elevation of 7.4m). However, at top point (HE1, HE2), pressure was ~100 kpag. Temperature during this time was <-20C.
3. The vacuum pressure was -35 kpag. Fluid temperature at that time we got that value was 10C (ref unit was just shut down). When we opened the points, we experienced a drop in flow rate (from 27,000 to ~25,000 kg/hr).
4. At fluid temperature 35C, we opened up the drain points again, we experienced alternate vacuuming and liquid draining (every 5s) for system 1 (T1) and continuous discharge for system 2 (T2). The flow dropped again to ~25,00 kg/hr (from 27,000 kg/hr), discharge pressure increased to 220 kpag (from 200 kpag). We then tried to throttle the pump discharge valve while both drain valves were still opened to atmosphere. We did not see any changes in the discharge pressure and flowrates.
5. In all tests, we do not observe any pump vibration.

Questions:

1. What cause/s the vacuum pressures in the line?
- My initial guess is that siphoning effect is occurring. And such, the higher flow rate. On theoretical calculations (assuming allowed design DP across the heat exchangers), I should only get ~1 bar dP across the system, but my pump is designed for ~3bar dP, hence higher flow. Is it correct to assume that my line is "fully filled" and the equations I used are valid?

- I back calculated from discharge point (where I used back Pressure = static head of liquid) and got some vacuum pressure value, and should not pose for any vaporization at low temperature, VP is almost 0 barg.

2. Should I be concerned with the appearance of vacuum pressures in the system if my system is attaining more flow? Velocity concerns and or vacuum concerns for my piping and equipment?

3. Most puzzling for me is that with increase in temperature I did get some vacuuming and discharging actions on the drain valves which I have not experienced with the cold temperature. I checked on the vapor pressure and at 35C is -0.95 barg. Flow rate was higher (which on reading references, higher flowrates should "flush out" the vapor in the system, if any). I am assuming that this may be due to "relatively lower back pressure on the discharge end" as the density is relatively lower at 35C. Is there any relation to viscosity changes (Nre at low temp <10,000 vs Nre 35C >50,000).

It's quite a lengthy description, but thanks in advance to everyone.

 

[itdepends]

1. Your pump needs to be designed with enough head to overcome the highest static head, but under steady state as you've seen- you do get head recovery on the down pipes. The minimum head recovery you can achieve is the difference between the vapour pressure of the liquid and the atmospheric pressure inside your tank. Higher levels of head recovery are possible if there is a component providing backpressure (e.g. control valve or heat exchanger) downstream of the drop in piping elevation.

2. You need to check the pressure rating of your equipment. Unlikely to be an issue for your piping (unless it's fibreglass or plastic) - or very large diameter. May be an issue for larger diameter vessels designed for low pressures.

3.Unlikely to be temperature related unless you were operating close to the boiling point. More likely the higher flow flushed out the air you allowed into the system, re-establishing the vacuum and setting up a cycle of air ingress giving a partial loss in the siphon effect, followed by discharge, air is then flushed from the down leg, vacuum recovers and draws more air in etc.

As a chem eng/metallurgist the first part of any answer I give starts with "It Depends"

 

[LittleInch]

Your problem to me stems from the fact that this is not a closed system. Your tank sits at atmospheric pressure and is effectively an end point. A closed system would allow you to set a higher static pressure than atmospheric to achieve positive pressure at all points in the system.

Your pump is oversized in terms of head and will be working beyond its power rating. You might be lucky and fin that the motor is capable of operating at your higher flow, but if you look you will find the running amps being higher than they should be.

IMHO, what you should be doing is inserting a proper control valve, (not an isolating valve) on the return line back to the tank and then control on the flow that the system was designed for (18,000 kgs/hr). This will force the pressure in the system back to what was designed and I suspect all your problems will go away.

What tends to happen in design is that pressure drops are over estimated and the pump and system designed accordingly. If every one adds 15 to 20% as an extra then it rapidly grows and the system then doesn't work as required.


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

 

[mcmg]

Thanks to your replies.

@itdepends: im still confused with #3. the difference in velocities between the flows at high and low temperature is just about 10%.. is there like an equation that may be used to check this?

@littleinch: i think the tank is not suited to handle high pressures (will still find the vendor drawings). is it possible to use an orifice instead?

 

[katmar]

The frictional pressure drop from the pump to the highest point is only about 20 kPa, but the friction loss from the high point down to where you are finding a vacuum is about 200 kPa. The fact that these are so different, together with the fact that your design delivery pressure for the pump is so different from what you find in practice, lead me to believe that your first step should be to compare each element of the design case with the pressure profile you have measured. A cynic might suggest that it is surprising that the system is working as well as it is.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[LittleInch]

The tank would not be affected as this would be d/s of the control valve.

Your system is just running too fast / too much flow. Slow it down and raise the pressure by adding a pressure drop at the end and it will all work.

A control valve is much better than an orifice, but you need something to back up the flow which is just running away with itself.

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