10 replies to this topic

[ramlalithravi]

Hello All,
I already read all the discussions regarding this Mixing of Two Pressure Streams topic in this Forum, but still i need to clarrify some of the doubts.

Upon my unclear ideas, I have classified 3 scenarios of Mixing stream conditions as given in the sketches.

1. Consider two pressure streams from different pump discharge at 10 bar & 30 bar respectively as mentioned in the schematic at datum level (0 Elevation). At this instance what will be the Pressure of the exit stream 3, If it is to be pressure of the lowest Pressure of the inlet stream. How the stream 1 has lost its pressure from 30 bar to/less than 10 bar.

2. At same above Condition, except Pump 1 at 10 m elevation and Pump 2 at 0 m elevation. My question in this condition, Is only the Dynamic (Velocity energy) will get reduced because of Mixing of Two Streams or Static head also?.

3. What will happen if incase Pump 1 discharge stream has Ejector at the mixing point? I have come across somewhere that when mixing two streams with an Ejector is an exceptional case, in that case What will be effect on exit stream??

Finally Mixing of two Streams at different pressure will leads lowest pressure of the Inlet Streams at exit stream, Is it applicable only for Gas or Liquid/Two Phase fluid also???

Various scenarios are all given in the attachment.

Hope i have posted all the queries regarding this Topic, Pl give ur insights on this to avoid any further doubts..


regards,
ramesh

[ramlalithravi]

Sorry somehow the attachment did not get attached.. here i have enclosed...

Attached Files

•  Mixing of Pressure Streams.doc   60.5KB   332 downloads

[kkala]

Attached "scheme.xls" is hopefully a way to solve the problem by trial & errors through an algorithm ( which can be improved). Probably a numerical application will dispel vague points. Following is pointed out.
1. Pressure of both discharge streams is reduced (due only to friction in Case 1) on their way to the connection point ( No 3 on scheme), where they approach a common value.
2. The two pumps would set the pressure at the connection point to a value lower than 10 Bara, setting their flows accordingly and following their performance curves (curves should be available, or algebraically expressed).
3. You have to know destination pressure (P4 on scheme) in order to specify flows and pressure at the connection point (as indicated on scheme.xls).
Note: Since pipe rougness can increase up to 10 fold during pipe life, calculated values may not be precise, but of course better than nothing.

Attached Files

•  scheme.xls   24.5KB   367 downloads

Edited by kkala, 29 September 2011 - 12:16 PM.

[ramlalithravi]

Mr.Kkala
Thanks a lot for ur effort and in depth explanation, which gave me more thoughts and intend to ask more questions to clarrify myself.

As u said in ur explanation, Streams are reduced to common value @ 3 due to frictional loss. How possible it would be reduced to that much pressure loss for stream 2, because the discharge of the pump is too high compare to other stream, which is at 40 bar for stream 2 and only 10 bar for stream 1.

If incase the discharge pressure of each pumps are adjusted based on pressure at connection point, the flow rate of centrifugal pump will also vary with respect to discharge pressure of the pump. In that case we can't achieve required flow rate at consumption point 4. The stated procedure in excel sheet applicable if it is a reciprocating pump but for centrifugal pump the principle is totally different, since i have considered different capacity pumps.

I guess pressure loss would occur on mixing and it results in exit stream pressure equal/less than lowest pressure of inlet streams. If i am worng, pls correct me.

Also i have read somewhere in the forum related to this topic, mixing of two pressure streams with an ejector is exceptional case. In what way it is so??

Awaiting for ur more insights.

Once again thanks for ur usefull calculation sheet.

[kkala]

Glad to know that "scheme.xls" has been useful, ramesh, let us see the questions according to my understanding. Static difference is temporarily neglected (see note on scheme.xls).
1. As u said in ur explanation, Streams are reduced to common value @ 3 due to frictional loss. How possible it would be reduced to that much pressure loss for stream 2, because the discharge of the pump is too high compare to other stream, which is at 40 bar for stream 2 and only 10 bar for stream 1.
- Expect entrance losses (not so high, included in frictional ΔP) at point 3 due to velocities change. You may mean this by "pressure loss on mixing". At a very small distance downstream of connection No 3 there will be one stream of uniform flow and pressure. I see now that this pressure can be a bit higher than 10 bar a, but at any case lower than shutoff pressure of centrifugal pump 1, otherwise this pump will not deliver flow.
- The pressures have to be same at 3; centrifugal pump 2 is supposed to increase flow, unless a valve is partially closed at its discharge. In this way pump discharge pressure will decrease and frictional ΔP will increase in its discharge line (2 to 3). Increased flow of pump 2 will shift its operating point beyond best efficiency point, towards end of curve. This is generally not recommended, pump motor has to be sized accordingly.
2. If in case the discharge pressure of each pump is adjusted based on pressure at connection point, the flow rate of centrifugal pump will also vary with respect to discharge pressure of the pump. In that case we can't achieve required flow rate at consumption point 4. The stated procedure in excel sheet applicable if it is a reciprocating pump but for centrifugal pump the principle is totally different, since i have considered different capacity pump
-Thinking that this is the behaviour of centrifugal pumps, I assume stated flow inadequacy is from centrifugal pump 1. You may consider selection of reciprocating pump 1, developing (practically) constant (or even regulated) flow under any discharge pressure. Then check performance of pump 2. Although you have to install PSVs on pump 1 discharge, things can be much simpler in total (especially if required flow of pump 2 is much higher than of pump 1).
3. I guess pressure loss would occur on mixing and it results in exit stream pressure equal/less than lowest pressure of inlet streams. If i am worng, pls correct me.
-See point 1 above. Mixing itself does not cause loss of pressure, entrance losses (or exit losses, from discharge pipes point of view) do (limited extent for liquids). For example, suppose 1" water pipe connected to 10" water pipe of 3 m/s velocity. If velocity of 1" pipe v is 3 m/s, no entrance losses (neglecting the change in flow direction from 1" pipe). If v=10 m/s (say, this is too high), entrance losses could be (v2/(2g), g=gravity acceleration) 10^2/20-3^2/20=4.6 m H2O or 0.46 kgf/cm2. Pressure is same for these two streams (1" & 10") just downstream of connection point.
Above is based on some "extension" of Perry's formula, 7th edition, 6-17 (expansion or exit losses), advice welcomed from others.
4. Also I have read somewhere in the forum related to this topic, mixing of two pressure streams with an ejector is exceptional case. In what way it is so??
-The ejector in this case would "inject" stream 2 (of pump 2) into stream 1 at the mixing point 3. This would cause an additional frictional ΔP in the line 2 - 3, contributing to increased entrance losses mentioned. But this ΔP may not be enough (see example of 10 m/s pipe). Ejector exit velocity must have a limit.
-Why pump 2 has been specified for 40 bar at discharge? If really needed, flow 2 will be reduced by increasing frictional ΔP in line 2 - 3. Do you want to connect discharge of a new pump to an existing circuit of an existing pump? If so, which pump is new?
5. It is noted that connecting two dissimilar centrifugal pumps (both operating simultaneously, not operating / standby) to a common discharge line is not recommended (*). They generally develop different discharge pressures, changing as flow changes, probably causing even instability of each pump flow. I have also heard of reduction of pump flow under this arrangement (due to “channeling”). Some of us avoid placing even identical pumps of same service in parallel (when one is run down, pumps are somehow dissimilar).
6. So connection of centrifugal pumps 1 & 2 to a common discharge, even distant, has drawbacks. At the beginning I thought scheme under discussion represents an existing situation or just thinking for future. However it is indicated to avoid this scheme (as is) in a new “project”.

Note(*): Same discharge pressure P at their "normal" flow could be tolerated (but P changes with flow, etc). However this is not our present case.

Edited by kkala, 02 October 2011 - 09:19 AM.

[S.AHMAD]

Hi Ramesh

For all three cases, Pump-1 (10 bar) is running BUT not pumping. This will eventually leads to pump overheating due to no flow. If we want to install this type of pump (1 pump of higher pressure) in parallel, we need to install flow control valve on both pumps or at least on the higher pressure pump.

The pressure at point-3 (P3) is the system pressure which is related to downstream conditions and independent of pump characteristic curve. Let say the destination is point 4 with pressure P-4 and elevation Z4 and DP3-4 is he line pressure drop: P3 is given by:

P3 = P4 + (Z4 - Z3) + DP3-4
(All in meter of liquid)

Hope the above helps and enhance your understanding of the subject matter.

Edited by S.AHMAD, 03 October 2011 - 09:51 PM.

[ramlalithravi]

@Ahmad
I understand ur concept that based on required pressure at the destination (P4), the pumps will function to its revised head and flowrate according to Head vs Flow curve . When high pressure (40 bar discharge) pump connected with low pressure pump (10 bar discharge) will deliver more volumetric flow rate with lower head, so low pressure pump will not pump any liquid and still it is in working mode. Which is actually not desirable for the pump safety point of view and this kind of situation shoud be avoided.

But the purpose of this discussion is to assess the behaviour of two pressure on mixing and how the pressure will reduce to same level in the downstream. Incidentally in the above case two pressure streams are from centrifugal pump outlet, and it can be solved by any of the methods suggested above. Incase if pressure streams are from anyother sources like closed drain header (Operating pressure 1 bar) collecting drain from vessel (Operating pressure 10bar) and at elevation 20m. How fluid will behave in this scenario??

Because in real time application, i observed flashing across these lines from high pressure vessel to closed drain header, having potential problem causes "Ice forming" over the closed drain header for a distance. What is the remedy to avoid ice forming on the line and how the pressure is reduced to closed drain header pressure??

@kkala
In simulation softwares, if u analyse two pressure streams, the exit stream pressure is considered as lowest pressure of the inlet stream why??. I understood ur explanation in centrifugal pump case, can u tell ur views on closed drain system, What is the behaviour of the fluid in such cases mentioned above??

[kkala]

In simulation softwares, if u analyse two pressure streams, the exit stream pressure is considered as lowest pressure of the inlet stream why??.
I understood ur explanation in centrifugal pump case, can u tell ur views on closed drain system, What is the behaviour of the fluid in such cases mentioned above??
Opinion / interpretation on these two matters are as follows.
1. If you directly combine two liquid streams of different pressure into a common line (without any pressure reducing element at their end) simulator is "confused" and selects the lowest pressure, probably issuing a warning. This is what I assume, since simulator cannot "judge" the case. In reality the pressures of the two streams converge (abruptly or smoothly) into a common value as they approach to the connection. This in the case when the two streams flow towards the junction point. Let us see other cases, depending on pressure distribution.
α. No flow in stream A (noted by S.AHMAD): All discharge line from A to junction point at shutoff of pump A.
β. Junction point pressure higher than pump A shutoff pressure (due to pump Β: Reverse flow towards pump A. This can cause damage on pump A, but (as far as I know) limited reverse (i.e. channeling) flow can be tolerated.
In all cases, pressure of two streams approach a common value in the vicinity of junction point.
2. A drain of constant pressure upstream is simpler in calculation than a centrifugal pump (assuming full drain lines). Junction (i.e. connection) develops common "converging" pressure. Flow from A to junction point C of two drains can be to that direction (if pA>pC), zero (if pA=pC), or even reversed (if pA<pC) (p=pressure).
Check valves are placed to prevent back flows, but check valves are not always reliable. As previously mentioned, centrifugal pumps operating simultaneously are not recommended to have common discharge, especially if these are not identical.

Edited by kkala, 05 October 2011 - 03:35 PM.

[S.AHMAD]

As I have mentioned earlier, pressure at point 3 is independent of upstream pressure (40 bar).

For sake of discussion, let say P3 = 5 bar and P2 = 40 bar. The pressure drop = 35 bar. This pressure drop is equivalent to the line pressure drop from point 2 to point 3.

1. In the blowdown case there is always a blowdown (BD) valve and checkvalve. The BD valve will take the pressure drop significantly as compared to checkvalve and the pipeline.

2. If we open the BD valve more, P3 will increase due to higher pressure drop downstream. P2 remains at 40 bar.

3. If the upstream stream at 40 bar is saturated liquid, adiabatic expansion (Flash) will occur at the BD valve. Temperature drops and may result in icing depending what is the final temperature.

4. This is what we sometime called AUTO REFRIGERATION. This is always true when the fluid is pressurized liquid of gas such as LPG, NGL etc.

5. Similar concept in refrigeration unit.

6. Simulation software defaults to lower pressure.

7. Mixing must be at the same pressure. The mixing pressure is the back pressure of downstream unit. The valve from point 1 (10 bar) and at point 2 (40 bar) will drop the pressure to 5 bar. delP1-3 = 5 bar and delP2-3 = 35 bar.

Edited by S.AHMAD, 05 October 2011 - 07:30 PM.

[ramlalithravi]

@Ahmad

From ur statement observed, pressure will drop from point1 5 bar and at point2 35 bar, Is it is merely due to mixing of liquids, incase if there is no pressure reducing elements on the line before joinig the common stream? If i do prob the actual pressure in the plant when two pressure streams mixing, will it give the same results as u mentioned?

[S.AHMAD]

For liquid, from pressure drop equation, the pressure drop is proportional to square of flow rate.

DelP = f(L/D). V²/2g where f is the Moody friction factor

At steady state, the system will arrive at a point where the system hydraulic will reach an equilibrium.

The following equations shall be met:

DelP1-3 = 10 - P₃ = f₁(L₁/D₁)V₁²/2g = K₁V₁²/2g

DelP2-3 = 40 - P₃ = f₂₍L₂/D₂)V₂²/2g = K₂V₂²/2g

DelP3-4 = P₃ - P₄ = f₃(L₃/D₃)V₃²/2g = K₃V₃²/2g

W₃ = W₁ + W₂

You have 4 equations with 4 unknowns (W1, W2, W3 and P3). The above equations can be solved.

The above equations are without fittings and valves - pipeline only.

P₄ is known as well ( e.g atmospheric)

Edited by S.AHMAD, 06 October 2011 - 02:11 AM.