Filter backwash from elevated tank 3

[OneManWolfpack]

I'm working on a small surface WTP design. Instead of providing a pumped filter backwash system, we're looking at running a line from an on-site elevated tank and installing a flow control valve.

One concern I have is the flow control valve wearing out. It will have to provide a significant energy loss - 80 psi inlet pressure and 5 psi outlet pressure - to give the desired flow. According to the valve manufacturer's literature, this would subject the valve to a "critical" cavitation index; however, the manufacturer's rep tells me that since it will be operated only a few times per day, it will be fine.

Another concern is that if the valve somehow fails, or is tampered with, the filter media could be completely washed out due to excessive backwash flow rates.

An incremental pressure drop before the fcv would solve both of the above problems. What is the best and least expensive way to do this?

I thought about running an extremely undersized line from the elevated tank to the fcv to dissipate some of the energy (and lower capital costs), but I can only go so small before velocity and surge pressure become an issue. I've always assumed that 7 fps is the maximum velocity you should design for, but since energy loss is a benefit here instead of a detriment, can I go higher?

If anyone has any experience with this I would appreciate your input.

Regards

 

[JedClampett]

How about a sleeve valve just to burn energy?

 

[bimr]

Not really sure that you have a problem. Where did you come up with the 5 psi? 5 psi is probably not adequate pressure to backwash a filter. You probably want 20-30 psi.

Where is the filter backwash piped after the filter. You will have some friction loss in that pipe.

The filter backwash effluent valve is a rate set valve for the backwash flow. There will be pressure loss across that valve as well.

Install a pressure reducing valve to obtain 20-30 psi and it will be acceptable.

http://www.cla-val.com/pdfs/e-90-01.pdf

 

[OneManWolfpack]

The filters are gravity sand filters. The backwash water flows up through the filters (opposite of normal flow), and exits through backwash troughs.

I estimated the head loss through the filter to be around 5 ft (2.2 psi) based off the attached chart. I'm assuming that backwash flow will have similar head losses as normal operating flows. In my mind, this is conservative. Since the bed will be fluidized during backwash the head loss will likely be reduced. If I'm wrong about this, or you know of a better way to estimate the loss through the filters, please let me know.

Adding that 2.2 psi to other friction losses between the flow control valve and the filters, plus the elevation difference between the backwash trough and the flow control valve, gives me about 5 psi at the outlet of the flow control valve at the design backwash flow.

The inlet pressure of 80 psi is simply the static elevation plus friction losses at the design backwash flow between the valve and elevated tank.

A PRV and FCV are essentially performing the same function in this case - providing an energy (pressure) loss by throttling to achieve a target flow (which has a corresponding pressure). With no throttling, the filters would see twice as much flow as they need, and the media would get washed out the backwash troughs. One FCV is capable of throttling the flow, but it would be barely cracked open, which makes me uncomfortable.

What I want to do, is configure the system so that the FCV only has to slightly throttle the flow. This way, the valve would be under less stress, and in the event of a FCV failure or tampering, the filters would not be damaged.

I'm not sure what a sleeve valve is, but using it just to "burn energy" is the right idea. I'm hoping to come up with a more permanent and inexpensive way than a valve, such as an orifice plate, or using very small pipe diameters. Any guidelines for the maximum velocity I can design for in ductile iron pipe?

Thanks for your help!

 

[bimr]

Sorry, I thought you were asking about a pressure filter.

You should be asking for a control valve that will allow the maximum backwash flow (15-20 gpm/square feet of filter bed area) to pass through the valve. If you inject the backwash flow of 15-20 gpm/square feet of filter bed area, it will simply overflow the backwash wiers. Any extra pressure will be dissipated.

You don't have to be concerned about the discharge pressure on the outlet side of the valve.

The backwash flow should be such a large flow that there should be significant pressure losses in the piping from the source to the filter. There should be a significant pressure drop at the valve to reduce the flow to 15-20 gpm/square feet of filter bed area. Any excess head will be dissipated in the filter.

Transmission mains should be sized for 3-5 ft/sec velocity. In building water mains can be sized for 7-9 ft/sec.

If you have a rotary wash, it will need approximately 50 psi.

Be sure you install a filter-to-waste-step as it is important to water quality.

In regard to reducing water pressure, the simple method is to install an orifice plate.

http://www.nesc.wvu.edu/pdf/dw/publications/ontap/2009_tb/filter_backwash_DWFSOM83.pdf

http://www.mrwa.com/OP-Filtration.pdf

 

[OneManWolfpack]

I'm not concerned about the pressure on the discharge side of the valve; rather, I'm concerned about the pressure gradient across the valve. Based on my analysis, the valve will have to impart a 75 psi pressure drop. To do this, it will be barely opened and thus susceptible to cavitation problems.

Also, I'm concerned that the valve could fail or be tampered with. If it is the only obstruction in the line and it is removed, the flow will be around 12,000 gpm or 68 gpm/sf. This will damage the filters.

What is the reason for sizing transmission mains at 3-5 fps and building water mains at 7-9 fps?

 

[bimr]

What is the size of your piping and the backwash flow?

You seem to be forgetting about the velocity head across the valve (V2/2g). According to the Bernouli equation, the total energy in the fluid is equal to the pressure plus the velocity head plus the static head. As you traverse the pipe run, the total energy may be converted to velocity and then back to pressure. The total energy on the discharge side of the valve is equal to the velocity head of V2/2g plus the pressure. Velocity head drops to zero inside the filter.

http://en.wikipedia.org/wiki/Hydraulic_head

If your line pressure reads 5 psi after the valve, the total energy in the fluid is actually higher, so you will not taking 75 psi across the valve.

Regarding the flow through piping:

The maximum flow that you should expect through piping is about 15-20 ft/sec. If the flow exceeds that, then the friction head gets to be too high and thus a maximum flow is self limiting.

Regarding "What is the reason for sizing transmission mains at 3-5 fps and building water mains at 7-9 fps?"

These are generally accepted economical values based on experience. Review crane's technical paper 410 flow of fluids.

http://www.flowoffluids.com/

At 5 ft/sec, you have a headloss of 50 feet per mile. At 9 ft/sec, you have a headloss of 10 feet per 100 feet.


Regarding failure of the control valve:

You should also have an on/off valve to activate the backwash water supply in addition to the control valve.


Regarding reducing the pressure:

If you remain concerned with reducing the water pressure, a vary simple method is to install an orifice plate in the line. You can size an orifice plate to reduce the water pressure to any value that you desire.

http://www.mcnallyinstitute.com/13-html/13-12.htm

http://www.ajdesigner.com/phporifice/orifice_equation_flow_rate.php

 

[OneManWolfpack]

"If your line pressure reads 5 psi after the valve, the total energy in the fluid is actually higher, so you will not taking 75 psi across the valve."

I'm not following you here. If the pressure is 80 psi before the valve and 5 psi after the valve, then there is a 75 psi pressure drop across the valve. The other two components of Bernouli's equation - V2/2g and elevation - are equal upstream and downstream.

A 80 psi inlet and 5 psi outlet results in a "severe" cavitation index according to the attached literature. This is why I'm concerned about the pressure drop across the valve.

The line size is one of the things I'm trying to figure out. The "generally accepted economical values" you gave make sense if in a pumped system, since higher velocities = higher headloss = more $, but in this system that is not the case.

Maximum backwash flow is 3645 gpm (20 gpm/sf). There will be 418 ft. of line between the elevated tank and the filters. There is a 160 ft elevation difference between the elevated tank low water level and the backwash troughs. Theoretically, I could install as small as an 8-inch line (20 ft. of headloss per 100 ft. = 84 ft). This would be great if I can get away with it since 8" pipe is a heck of a lot cheaper than 16-inch pipe (6 fps and 6.9 ft head loss per 1000 ft).

The velocity in the 8-inch would be 23 fps. Is this too high in this situation? If so, why is it too high, and what is the maximum velocity that I can design for?

 

[JedClampett]

I talked to my guy who knew about sleeve valves. They're made by Bailey and pretty much no one else. Here's their website:

http://baileyvalve.com/?pg=b10

 

[bimr]

I would recommend that you maintain the pipe velocity at less than around 10 ft/ sec. That is a typical velocity for water mains. If you go higher than that, you will have to do a stress analysis on your piping and pipe supports because of the high dynamic forces imparted by the higher fluid velocities.

A velocity of 10 feet per second will then require the use a 12-Inch size pipe. Steel pipe has a headloss of 2.8 feet per hundred feet at that velocity.

Install an orifice plate of 5.6-Inch diameter in the 12-Inch backwash line. The orifice plate will generate a headloss of 40 psi when the pipe is flowing at 3645 gallons per minute. If you want additional headloss, reduce the size of the orifice plate.

http://www.mcnallyinstitute.com/13-html/13-12.htm

http://www.ajdesigner.com/phporifice/orifice_equat...

I have never used one of the bailey valves. It looks like something that is used at dams not water filters. Would expect that the Bailey valves are extremely expensive.

Upstream of the proposed control valve, you have a total fluid energy equal to 80psi. The pressure can't be higher than the water level in the storage tank. Downstream of the valve, you have 5 psi plus the velocity head.

I still fail to understand where you came up with the 5 psi on the discharge of the valve. Ask the filter manufacturer what the recommended backwash pressure should be. You also have a headloss as the water flows upwards through the filter media. The water pressure for backwash should generally be at least 20 – 30 psi. If you have a rotary wash, that water supply will need to be approximately 50 psi.

The backwash enters into a tank that is open to the atmosphere. As long as you control the rate of flow to 3645 gallons per minute, the excess pressure will be released to the atmosphere.

Regarding failure of the control valve:

You should also have an on/off valve to activate the backwash water supply in addition to the control valve. Review the Leopold schematics. Note the flow controller on the backwash supply as well as the on/pff valve:

http://www.fbleopold.com/products/controls/

 

[QualityTime]

Given all of the transmission losses, the headloss is the pressure drop just before and just after the valve. Put two PRVs in series in line if you want longevity[highlight #E9B96E][/highlight]

 

[stanier]

Hi

Mitech will build a custom tortuous path orifice plate with butterfly control valve to reduce the pressure if you think it a concern. Mitech are owned by Spirax Sarco now I believe. I have used these to let down pressures of 100bar. Noise could be an issue but suitably applied lagging will solve that challenge.

Else you could use a tortuous path control valve such as by CCI. These are used for HP steam turbine bypass but I think that is over the top for your application.

I would go to Fisher, Valtek, Maisoneillan, Dezurik etc with the process conditions and seek their advice on a control valve for the application. If the valve was fail safe closed then you overcome that challenge.

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/

 

[bimr]

What are you fellows talking about? Nothing is more reliable than a fixed orifice plate with no moving parts or service required.

 

[semo]

As BIMR mentioned, this is starting to go way overboard. KISS method.

First, I wouldn't really be concerned with pressure. The concern is flow. You don't want the backwash rate higher than recommended for the type of underdrain and media installed. The pressure will reduce accordingly.

On extremely small systems with lack of funds I have seen and even installed a ball valve that was partially closed to set the flow rate. The operator (handwheel/lever) was then removed so that someone else did not adjust it. A second valve was installed upstream just for open/close operation during backwash. It can be manually or automatically operated.

An orifice plate is cheaper than a ball valve and untamperable (?); but, the ball valve can be adjusted in the field to obtain the flow rate desired. It can also be adjusted if your system variables change in the future.

No need to purchase expensive flow control valves.

 

[QualityTime]

Go and talk to your local singer valve representative if you are really concerned about cavitation of the valve. They should be able to steer you in the right direction about cavitation.