Force Main Pumping Downhill 8

[tbs]

I have an unusual problem. I am designing a 10 MGD Lift Station and 30” DIP Force Main that is pumping downhill. I am a licensed engineer and experienced in pipeline design, but pumps and force main systems are not my specialty.

The Details--The proposed 30” DIP epoxy lined force main is 12,000 feet long and will connect to an existing 30” DIP mortar lined force main that is 6,500 feet long which discharges into a gravity line. There is a 17 foot drop from the lift station to the final discharge with an intermediate high point that is 7 feet higher than the lift station. The high point is only 1,800 feet downstream from the lift station. From the highpoint to the discharge there are several intermediate peaks that will prevent typical gravity flow. The pipe is epoxy lined to help prevent corrosion and air/vacuum release valves have been place at the appropriate peaks.

When the pumps are on, the line will be flowing full. When the pumps shut off, air pockets will develop at the peaks in the line. The potential for corrosion at the air pockets is high but has been minimized buy the use of epoxy lined pipe.

The concern is the pump performance. If the pipe remained full when the pumps shut off, the pumps would see a constant head when they came on. Unfortunately that is not the case. The pipe will have several air pockets which will have to be evacuated from the line which will cause the pipe to see highly varying heads and could produce shock waves in the line.

I have investigated two possible solutions to keep the line full when the pumps are off.

Solution 1 was to install a valve on the discharge that would close when the pumps are off thus preventing the line from continuing to discharge by gravity flow. This proved to be a possibility, but the criteria are the valve must be mechanical, no electricity, pneumatic, or hydraulic systems will be accepted by the client, and the failure mode of the valve must be open. Red Valve makes a spring loaded valve that fits the need, but they only make it up to 12”.

Solution 2 is to build a vertical “gooseneck” that will create a highpoint in the line to prevent it from draining. The force main would go vertical out of the ground just prior to the discharge to an elevation that was equal to the highpoint of the line and then turn 180 degrees to go back into the ground and discharge into the gravity line. This solution would create a manmade high that would keep the line full at all times which would prevent corrosion in the line and would allow the pumps to see a constant head when they came on. A decent solution, but no one likes the idea of a 30” DIP Force Main sticking up over 20’ feet in the air.

I have been told by some that pumping down hill in this situation is not a problem. If it is, then I have a hard time believing that this problem has not been encountered before and that with modern technology there is not a better solution than having a 20 foot tower of 30” pipe sticking out of the ground.

Any input would be appreciated.

Thanks
TBS

 

[stanier]

Hi Mou,

Variable speed pumps are not the answer. To illustrate this point if you plot your system curve against the pump curve with the HQ parameters for various speeds you will notice that there are portions of the system curve that are relatively flat near low flow rates. Now imagine trying to control the speed of the pump at a low flow rate where a small change in speed gives you extreme changes in flow.

The rule of thumb has been variable speed drives should be used where the static head represents 50% of the total head. Impossible to do when you have negative heads.

In porevious postings there are references to goose necks to provide the head on the pipeline. This is ok if the head required is small. turning the main into a pressure/gravity combination by having a vent at the high point is common however as Aussiemike mentioned there will be odour problems. But if you have a control valve at the end of the gravity section to maintain the line in a flooded condition then the pipe will remain full and the amount of air reduced considerably. On a very large main you could consider using reverse rotating pumps to generate power as a means of keeping the line full. Warman designed a such system for Bougainville Copper Mine to obtain energy from the tailings line.

Control valves at the discharge are another alternative in getting the hydraulic grade line above the pipeline profile.

Merry Christmas and a happy New Year

 

[mou]

Hi stanier,

Thanks for your response. Much appreciated. One other question. How would you design the pump head for such a system (assume the orginal scenario with no gooseneck, control valve etc). Would you assume the scenario where the pipe is fully pressurised (ie steady state equations). Or would you model it using a package like SewerCAD. I have been told that you would model this using pressure main to the high point (as its always full) and then gravity main (with bolted manholes) onwards from the high point. In this case the required pump head will be the difference from the high point to mean water level in the wet well plus friction losses in main to the high point. Is this right ?? I guess, what im asking is how 'best' to work out the design head of the pump under such a potential unstable hydraulic scenario.

Cheers and have a Happy and Safe X-Mas.

Mark

 

[stanier]

Hi Mark,

If you had to live with no gooseneck or control valve and you were going to use bolted manhole covers then the calculations need to be two fold. One to be able to fill the line for there will be a point where it will drain from the high point. Air ingress through rubber ing joints leaks, incorrectly fitted manhole covers etc. thus the near stalled head of the pump must be great enough to get liquid into the line. The when the line is in operation your system curve will be depressed by the gain in head as the liquid runs downhill.

Really it is not a desirable way of designing the pipeline. All sorts of consequences arise. column separation causing extreme surges greater than predicted by the Joukowsky equation causing pipe bursts, damage to pump and valve seals, damage to valve seats, failed linings of pipes and pipe collapse.

If your client is insisting on this have the good grace to walk away from the job for the embarassment now will be far less than when the pieline fails.

 

[mou]

Hi Stanier,

Thanks again for your feedback. Wasnt really thinking about actually constructing the downhill portion with bolted manholes but only simulating the partial full situation by using fake bolted manholes. See the following and in particularly Tom Walskis notes:

http://www.haestad.com/hmicom/listserv/archive/default.asp~action/mess/messid/19692/searchid/48721

The problem is that if i use this method and adopt a pump to pump under this situation, when i put the same pump under a full pressurized situation the pump head is vastly different (ie higher head, lower flow). Off course this latter situation is the worst case (ie during wet weather) but i wont achieve the required pump flows. And i guess the same is inversely true: if i design a pump assuming a full oressure situation and then use this pump under the part full case (ie simulated with bolted maholes in SewerCAD) i get the pump pumping against a very low head (ie the pump may be pumping past its allowable range). Im not sure if im missing something in my logic or that im going around in a circle !!

The control valve option is not totally out, but i still am determined to work out what to do with a situation like the above. Im sure im missing something simple im my thinking !!!

Cheers

Mark

 

[stanier]

Mark,

I have a license for Sewercad & Watercad but prefer to use AFT's Fathom and Epanet for that matter. Haestad have too much glitz and glamour for my liking.

What you describe is what is happening. If there is no vent in the line the benefit of the hydraulic head (syphon)in the gravity section will be seen by the pump. As I said you have to fill the line. After stopping the line will empty , albiet slowly, depending upon the amount of air that can get into the system.

I have just designed a negative system. DN250 4.2 km long. The only way I could get stable operation was to lift the HGL above the pipeline elevation. To do this I added a pressure sustaining valve to the outlet of the line. Once you do this the added head allows the use of a variable speed drive.

 

[mou]

Hi Stanier,

Ahh...so you had a PSV on the end of your main which basically prevented the main from emptying ie if the pressure starting dropping, the valve started to close and then it actually closes when the pump stops (or thereabouts). I guess you modelled this using a water hydraulic package like EPANET to work out what upstream setting you required to maintain the HGL above the pipe. With the pump starting (or pump at low flow), the valve may not open until the upstream pressure reached a certain pressure.

This is starting to make making more sense.

Cheers

Mark

 

[HydroTripp]

Having designed numerous WWPS of various sizes, I cannot see that water hammer will be any problem with the low pressure head you will be dealing with. You should check the pump design for various conditions to make sure it does not runnout and overload or alternately run too far left on the curve. Take a look at the following conditions:
1. Design for 1800LF of Forcemain and the lift to reach the peak you described. Under this condition ignore the remainder of the force main. This ignores the "recovery head" and simulates the pump starting with an empty force main.
2. Check the pump using the elevation of the discharge and the full force main length for friction loss. This condition assumes the pipe is full and ignores the peak. This is called using the "recovery head" and will be true only after the pump runs long enough to completely fill the force main.
3. Check the pump using the elevation of the peak but still use the full length of the force main for friction. This is the worst case for pressure head.

If your selected pump and motor is ok for these three conditions, and if the rate of flow is acceptable for all three conditions (For condition 3 flow will be low, for conditions 1, & 2 flow will be higher), you are done.
We designed a pump for 1/2 recovery head for GUC in 1986 and it worked like a charm. Don't create the site conditions by implementing that gooseneck idea. Deal with the conditions you have.