Multiple Sump Pumps - Discharge Pipe 2

[civilman72]

Here's the situation: 6-unit condominium complex with sump pumps in the crawl space of each unit. Sump pumps all pump to one 1-1/2" discharge pipe (flowing south to north). This existing discharge pipe is about 180' long and installed with a negative (uphill) grade. The sump pumps at the south end of the line fail often and the discharge pipe clogs and needs cleaning annually. It's assumed that the most southerly sump pumps are burning out due to the distance they have to pump. It's also assumed that the pipes are clogging (mostly with sediment) due to the negative grade and lack of flushing velocity.

Solution being discussed: Replace existing discharge pipe to allow gravity flow through pipe, install cleanouts along pipe, and add new pump in north crawl space to pump discharge pipe fluid up to existing building outlet. There are openings in the wall between each crawl space, so re-installing the discharge pipe with a positive grade should be simple to accomplish.

Questions: Is the efficiency of the system improved by having a gravity-fed discharge pipe? Should we anticipate that the future pump replacement and maintenance will decrease enough with this new configuration to potentially justify the costs for a new discharge pipe and new pump? Any better solutions?

 

[katmar]

bimr, Your conclusions were all logical, what I was saying is that there was insufficient information to be certain. You obviously had these particular pumps in mind, and it seems that these are indeed what is installed, but no detail had been given to confirm that.

Based on the actual information given it would be fair to use the following logic:

Lines are clogging up implies low velocity = line too big.
Pumps are burning out implies insufficient back pressure = line too big

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[bimr]

Regarding "Lines are clogging up implies low velocity = line too big."

These pump applications do not pump continuously. The solids tend to settle out when the fluid stops as the pump shuts off. So you need a high velocity to resuspend the solids when the pump is restarted. Otherwise, the pipes will gradually foul over time.


Regarding "Pumps are burning out implies insufficient back pressure = line too big"

civilman72 has stated the pumps with the longest pipe length (and presumably the largest head loss) are the pumps failing. So this is really not plausible. If it was true, the pumps with the shortest pipe length would fail first.

 

[katmar]

With the extra info now available I am starting to see it bimr's way, but there is still information outstanding to be certain of a proposed solution.

If these pumps are similar to those in bimr's data sheet (and I must confess I did not envisage such small pumps) then it looks to me as though each pump will be operating in the top left corner of the curve and bimr is probably correct that these pumps are simply rattling themselves to death by being too far from the BEP. Especially if they are "cheap and nasty" pumps.

One thing that has not been mentioned is that the crawl spaces are flooding, so presumably this means that when the pumps are actually working they do have sufficient capacity.

We do not know what the static head is on the discharge pipe - 7 ft appeared somewhere but it is not clear if this is the static head. By making a few assumptions (dangerous), I come to around 36 gpm flowing through the 180 ft of 1.5" pipe and giving a friction head of 16 ft and a velocity of 5.6 ft/sec. This means that each of the 6 pumps is delivering only 6 gpm and the velocity in the pipe from the southmost pump will be only 1.0 ft/sec. The northmost pump would be a bit above 6 gpm and the southmost pump would be below 6 gpm. This would explain the silting up of the line at the southern end. The section of the header between the 1st and 2nd pumps could probably be only 3/4" and then from the 2nd to the 3rd a 1" pipe and only use 1.5" after the 3rd pump.

To do a proper design we would need the actual pump curve of the installed pumps, plus we would need the distances between each pump and the static head up to the discharge point. If the pumps are coping with the flowrates and the only real problems are the silting up at the southern end and the premature failure of the pumps then perhaps the answer is to use smaller pipes on the south end as described above and to invest in better quality pumps when replacing failed units.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[civilman72]

Katmar - I finally tracked down the exact type of pumps that are currently being used in the crawl spaces. It sounds like the condominium management company has been pulling the sump pumps off the shelf at a local Ace Hardware. The claim they use the 3/4 HP type - I have included a link to the 1/2 HP Ace Sump Pump. Based on the customer reviews, this is not the most reliable pump.

I could not find a pump curve, but specs for 1/2 HP Ace include:
65 gpm @ 5'
52 gpm @ 10'
35 gpm @ 15'

It appears the Wayne pumps may be similar. The performance specs for 3/4 HP Wayne include:
71 gpm @ 5'
58 gpm @ 10'
43 gpm @ 15'
18 gpm @ 20'

So, it looks like this pump (if placed in the south crawl space) cannot overcome the friction loss in the 150’ long, 1.5” discharge pipe, when trying to pump over 18 gpm.

Also, more specifics on pumps and pipe configuration:
There are five (not six) crawl spaces and pumps. Each unit is 30’ wide, so total building length is 150’. Each sump is placed in the middle of the crawl spaces, so the total distance between the south and north pumps is approximately 120’. The discharge pipe takes a 90 degree turn in the north crawl space and discharges on the east side of the building, so total length of pipe from south sump pump to discharge location is 150’. The discharge pipe elevation is just below the finished floor, the crawl space is approximately 4’-5’ in height, and the sumps are buried 2’-3’, so I’ve been assuming a total head of about 7’.

Katmar – I was thinking about upsizing the pipe in increments as it heads north, just as you suggested. I’m trying to determine if this would be a better long-term option than installing a separate header (2’-3”) for the two south pumps and separating them from the three north pumps.

 

[DubMac]

Consider the effects of parallel pumping in this combined header; they cause many unforseen problems. You could be pumping some of the check valves shut.

Heat buildup due to running at shutoff can transmit back through motor shaft since this is most likely a small close-coupled pump and the impeller is attached to the motor shaft.

 

[katmar]

Thanks for the requested info. With all the facts on the table we can now see where the problem is, but the solution is not easy.

To make the description easier, and hopefully clearer, let me give the pumps the item numbers P1 thru' P5 with P1 being at the South and P5 at the North.

With 5 off of these 3/4 HP pumps connected to a single 1.5" manifold there is a serious mismatch between their duties. It is aggravated by the fact that the line from P5 to the eventual discharge is only 30' and not the 120' I originally took it to be. If the discharge line was long relative to the distance between the pumps it would make the individual pump duties more equal.

My estimate is that with all 5 pumps working the total flow will be about 70 gpm, but the individual flows vary between 30 gpm from P5 and less than 5 gpm from P1. This puts P1 way up on the left top corner of the curve and it will be running for long periods under very stressful conditions. It partially explains why the line is silting up but by my calculations when all the other pumps have emptied their sumps and shut off and P1 is still running its flow rate should pick up to around 30 gpm. I'm surprised this isn't enough to clear out the line.

The proposal to install a 4" header will certainly equalize the pressures, but it means that the velocity on the south end will be very low. I suppose it would take a long time to silt up with such a big diameter. One factor to bear in mind if you do use this method is that the pressure drop will be very low when only one pump is running and without the 7' of static head these pumps could run off the curve on the right hand side and possibly burn out more quickly than they already are. Check on the static head and make sure the pumps are OK there.

If you are unhappy with cvg's proposal of running individual pipes from each pump to a new sump with a new larger pump there is a halfway method that I believe would also work. You could run separate 1.5" lines from each pump through to the position of P5, and then joint them all there into a single 3" line which runs to the discharge point. This could be done with no new pump. All the pumps running together would give about 200 gpm with the flows varying from P1 at 30 gpm to P5 at 60 gpm and giving velocities in the 1.5" sections of between 4.5 and 9.0 ft/s.

Because P5 will discharge directly into the 3" line the concern remains of it running off its curve if the static head is too low. A disadvantage of this scheme is that when only one pump is running the velocity in the 3" section will be low - between 1.4 and 2.4 ft/s depending on which pump is running. The pumps would be pumping at much higher rates than they are now, so changing to this arrangement will mean that the pumps pump for a much shorter fraction of the time and therefore there will be a lower chance of them running at the same time and achieving an increased velocity. With this arrangement the pumps, particularly P1/P2, will be running for a fraction of the time they currently run and should therefore last proportionally much longer.

All these calcs are based on the pump curves being the same as the 3/4 HP Wayne pump data given.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[cvg]

the take away from this is that each section of the header should be sized to match the assumed flow and the pump curve(s) that are producing the flow. Sizing should maintain minimum velocity while keeping the head loss to a minimum. There may be numerous options for your piping system, you need to evaluate them to determine the recommended one. Don't forget minor losses which in these types of systems can be significant.

An additional issue which has not been discussed is the potential for air in the force main. Since these types of systems are rarely profiled, it is not uncommon to have high and low points in the pipes. The high points collect air which may impede flow and increase head loss. The low points collect sediment also resulting in increased head loss.

 

[civilman72]

Lots of good discussion and explanations here... Thanks again.

It appears there is no cookie-cutter, easy answer here, without having actual groundwater inflow information in each crawl space. This type of monitoring is overkill as far as I'm concerned.

It's very difficult to discharge from the crawl space to the exterior of the building under the middle Units (#2-4). But I can combine the two pumps in the most southerly crawl spaces and create a new exterior discharge point out from Unit 5. This will decrease the total distance to pump to 50'-60', which should eliminate a lot of the concerns discussed above, and offer a cost-effective solution.

 

[ccfowler]

civilman72,

Since the sump pumps were selected primarily on the basis of availability, convenience, and price, it is reasonable to expect that the check valves were selected for the same set of reasons leading me to suspect that they are simple brass swing check type. It seems likley that these check valves probably do not seal well due to crud in the flow. The troublesome pumps may be suffering from having their sumps refilled by significant check valve leakage in addition to the ground water inflow. Since these pumps are already working against greater effective heads than the others and they may be re-pumping the same water, their actual number of on-off cycles and number of working hours may significantly exceed those of the other pumps.

Another possibility involving the check valves is that consideration may have been given to the presence of crud in the flow, and reasonably inexpensive soft-seated spring-loaded check valves were chosen. In this case, the cracking pressure is likely to be relatively large since they would likely be a readily available type normally used in a pressurized domestic water system where their 2-3 psi cracking pressure would be relatively insignificant. Such a cracking pressure represents roughly 4 to 6 feet of head just to overcome the cracking pressure with even more head required to overcome the restriction of the flow due to the configuration of the check valve.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.