Simple gravity question 8

[jack880]

Hi. I have a water tower,and i need to install a valved pipe as shown in the sketch to allow me to take samples of the water.

How can I size the pipe to get the flow I need from gravity?

Is it as simple as plotting the system curve for the outlet pipe and the flow will be whatever flow corresponds to the driving head (water depth)?

Thanks a lot

 

[pennpiper]

Use a 3/4" pipe.
You will have all the static head from the water level in the tank plus all the static head from the water level in the pipe. Just stand back a little when you open the valve. Oh, be sure you use a "globe" valve so you can throttle the flow.

 

[zdas04]

You've drawn a typical new-guy sample line. If you are in a climate that ever freezes, the dead leg between the tower and the valve will freeze. It happens every year when winter sneaks up on us. I address that issue with a cable operated ball valve (just fabricate a piece of lightweight plate in an "L" shape with a hold at the intersection that fits the ball valve stem and put cables in the ends of the "L") located as close as possible to the tower. Then put a globe valve on the end to throttle the flow to the sample bottle. After you pull the sample, you can shut the ball valve and re-open the globe valve to drain the line.

You need to think about what else you may want to use this line for in the future. You may need to use it as a drain, or to inject chemicals, or some other odd thing that you can't pre-engineer. Consequently, while 3/4" (or 3/8" for that matter) would give your adequate flow for a sample, I usually design this sort of thing as 2-inch (maybe with a 2X3/4" bushing in the end to get a manageable sample size).

David

 

[jack880]

Thanks for the replies. I take it neither of you know the answer to my question then?

David, I didn't fully explain the exact scenario this "new-guy" line will be operating in as it wasn't relevant to my question. I am aware that water can freeze, thank you. Although I am fascinated by how you usually design your sample lines, I'd be more interested to hear how to calculate the flow rate. If you wanted to inject chemicals into a tank I'd advise you to inject them into a pipe that flows into the tank rather than out of the tank, or you might find that they don't ever reach the tank. That sounds even worse than a typical "new-guy" chemical dosing system.

I think that I can plot the system curve of the pipe, and the point on this curve that corresponds to the head provided by the level in the tank will give me the flow. I was asking for confirmation of this, that's all.

 

[zdas04]

You never said exactly what flow you need, so no I can't tell you how to size the line for an secret flow rate. If you want to fill a one liter sample bottle in less than 10 minutes the line could be 1/4". If you need 10,000 gpm for reasons that escape me, it would have to be a bit bigger.

Good to know that you realize that water freezes. The way you drew the picture would guarantee that your sample line would freeze, I was just trying to help avoid a problem that I see every day. Sorry if I stepped outside the contract parameters--oh wait, we don't have a contract and I have the latitude to comment on whatever the hell I want to comment on. You just can't get good free help anymore.

David

 

[jack880]

I wasn't asking you to size it for me, I was aking about how to work it out for myself so I can size it based on the secret flow rate that only I know.

The picture was merely a simplified diagram to illustrate my question. It wasn't a "new guy"'s engineering drawing. I am far from a new guy, I've just never had to perform this particular calculation before so was checking I was doing it right. You obviously don't know the answer to my question, but despite that you obviously relish the opportunity to patronise someone who you see as less experienced than you.

 

[vpl]

**Equipping myself with full fire-proof armor and an anti-dragon breath shield, I lower my visor, and step gingerly onto the field ... **

I think that I can plot the system curve of the pipe, and the point on this curve that corresponds to the head provided by the level in the tank will give me the flow. I was asking for confirmation of this, that's all.

You could do that, but it is overly complex and is likely not going to give you the result you desire.

**Do I hear teeth gnashing? Smell sulpher? I back quickly away, shouting:**

Most people size sample pipes the quick and dirty way -- they look at the size of the sample they need and put on a big enough pipe to handle that flow yet not cause them operational problems. They don't spend time calculating it. They definitely don't do a complex calculation.

If the dragon is polite, I might feel inclined to tell him how I would calculate it.

Your choice: feel frustrated at how no one is rushing to solve your problem or act nice -- and realize that anyone who posts is volunteering their time and services to help you. "Please's" and "thank you's" go a long way.



Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[zdas04]

Pat,
You are once again (sitll?) my hero (heroine?).

David

 

[katmar]

Wow jack880, you are awfully sensitive for an old-hand engineer. Most of the old engineers I know are far too thick skinned to take offense at David's response. I opened your sketch, and my immediate reaction was "Oh, another student asking homework questions". Anyone who thinks that the driving head in this instance is the height of water above the pipe opening hasn't been around very long.

You must see David's response in the correct context. If you take the trouble to read some of the past posts here you will see that David is extremely experienced and knowledgeable, and gets involved with some really deep engineering discussions. And he is highly respected in this community for his willingness to share his knowledge and experience in a very open way.

He has adjusted his approach in his first post above to accommodate someone who is clearly a piping newbie. If you find his approach "patronizing" that is your problem. David was obviously pitching the level of his response in line with the level of your question. One of the very first lessons engineers need to learn is that you have to ask the right question to get the right answer. It seems you skipped that lesson.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[JLSeagull]

If you have an iPhone, download an app called PressureDropCalc. I think that I paid $0.99 USD. Inputs include the flow rate, pipe inside diameter, fluid density, viscosity, roughness factor and equivalent length. The results include the reynolds number, Fanning friction factor, pressure drop and fluid velocity.

(Now wearing my Nomex)
Often people calculate continuous liquid sample lines to obtain an acceptable lag time. The calculations address the pressure drop and volume within the tubing over the distance etc. This can be more complicated if the sample is returned to the process.

This example appears to be a grab sample application. The density of the water is mostly constant except for some adjustment associated with the temperature and elevation if not near sea level. I gather that this is an atmospheric tower. Balancing the units desired you multiply the height of the tower above the sample connection by the density to obtain pressure.

The length and piping or tubing area establish the volume.

 

[tr1ntx]

http://www.engineeringtoolbox.com/nozzles-water-discharge-capacity-d_1526.html

 

[TenPenny]

Rather an odd exchange.

How do I do something that I'm not explaining?
Here's how I'd do it.
I don't care how you'd do it, I want you to tell me how I should do it.

Typically, if you know the inlet pressure and the desired flow rate, you can calculate the appropriate pipe size.

Since the OP knows his required flowrate, and must knw the inlet pressure, he must be able to calculate required pipe size. So why is he asking how to do it, without wanting to be told how to do it?

I'm confused, but it's Friday.

 

[BigInch]

Seems he doesn't know what pressure to use.

I think he'll need a pipe size giving a flowrate that flushes the line in a reasonable amount of time to get a fresh sample from the tank, whatever that time might be??

Initial pressure will be height of fluid over the sample point x density as Seagull says. With that pressure and the flowrates from the chart in the link above, you can size the smallest possible pipe diameter for the indicated nozzle flowrate making sure it doesn't lose more than 75% of that pressure and bracket the ultimate diameter choice between that and the largest diameter based on the time to get the fresh sample and you've got it.

Move along folks. Nothing to see here.

"I am sure it can be done. I've seen it on the internet."

"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[micalbrch]

A star for David at Katmar. Agree with Katmar. Looks like student posting. An engineer would know how to calculate the flow in dependence upon height and pipe diameter.

 

[Latexman]

Can you just dip the container into the water to get the sample? That's how we do it on our water towers.

Good luck,
Latexman

 

[BigInch]

Knew a guy in high school that fall off a water tower.

"I am sure it can be done. I've seen it on the internet."

"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[jmw]

OK, I just couldn't resist interupting even though this thread has great promise.

It has been my experience of client/solution innovator discussions that the real problem has to be ferreted out.

I assume that at intervals, perhaps a week? an operative is sent to sample the water in the tank and has to climb the ladder. Scenario one is the BigInch syndrome. The operative will sooner or later contrive to fall and if he survives, thus avoiding a moderate payout to survivor, he and his personal injury lawyer will live well for the rest of their days.
Option two is that unsupervised, he will decidee to goof off for the available time and simply fill the sample bottle at the nearest tap.

So the real problem is "How do I collecta 500ml sample from the water and deliver it to an operative at the bottom of the tower without flooding the area flushing the pipework, having to install drains, the operative getting soaked, etc and how do I ensure that the operative really will flush the lines instead of grabbing the first 500ml he can get which, if this is a week in the sun between samples, is probably a zoo, a primordal soup.

I attach my automated sample collection solution as per the attached diagram.
All you need are three suitable solenoid valves, a colection fo relays, some time delay energise and some time delay off, a key operated start button and a few sundries.

At the start all the valves are closed and the pipework between the valves is full of air.

On start SV1 opens allowing air to escape via the top valve (the U is to prevent rain, dust and debris collecting and falling into the pipe when the valve opens and the purpose of the closed valve is to stop bugs etc taking up residence.)
A limit switch (closed when the valve is open) initiates a time delay off relay which opens SV2. This valve now admits an amount of water into the vertical pipe which, closed at the bottom, causes the water to enter the vertical rising section above. It will not go above the water level in the tank but the timer will close after a suitable interval anyway such that only sufficient water enters to deliver a good sample at the bottom.
SV2 closes and initiates a time delay on relay. While this is timing out the water in the pipe flows down to the bottom with the trapped air rising through it.
After a suitable interval the water will have collected at the bottom of the pipe.
The timer expires and SV3 opens.
If the operative is awake he will collect 500ml of it.
A second timer will hold the valve latched open to allow the pipe to fully drain and then close.
SV1 now also closes.

The operative may, if so instructed use the initial sequence to flush the pipe and collect his sample from the second sequence.
Power can be frrom wind turbines, solar panels etc. I'd suggest 1/2" 3/4" is adequate since we now don't care what the flow rate is only that we colelct a good sample and deliver it safely to the operative.
Because the water is collected at the bottom of the pipe before the valve opens we only have the head of water in the pipe to worry aboiut so no massive flow velocities or soaked eemployees.
Employee now does not have time to goof off because there is no time allowed for climbing ladders and admiring the view. Plus, suitable systems can be added in to log the opertaives visits, ideentity and that he did what he was supposed to do.

JMW

http://www.ViscoAnalyser.com

 

[JStephen]

"To take samples of the water" to me, implies drawing off a liter of water at periods, and practically any size of pipe would do it, with the main problem being freezing as noted. This assumes a typical water tower of 100' or 125' high, which should give high flow rates.

Presuming there is something more to the problem, you can use Bernouli's equation including local losses for pipe friction, fittings, entrances, and exits, applied between the tank surface (velocity and pressure = zero at that point) and the pipe exit (pressure = zero at that point) and solve for velocity. You may need to adjust friction factors and repeat.

This is an energy-based solution and it is possible that you will be unknowingly assuming that absolute pressure goes negative at some point up the line when doing this (not generally feasible) or that gauge pressure goes negative at some point (in which case, your assumptions for the solution are invalidated). So as a check, remove the vertical line from your diagram, and calculate flow rate assuming the flow just runs out of the tank at the point of connection, and use whichever is smaller.

 

[BigInch]

"remove the vertical line from your diagram, and calculate flow rate assuming the flow just runs out of the tank at the point of connection, and use whichever is smaller."

Unless head loss in the downcomer is equal to the height of the bottom of the tank, your method will yield initial outlet pressures and resulting flowrates possibly much smaller than when including the downcomer head loss as a contributing factor to the potential flow.

"I am sure it can be done. I've seen it on the internet."

"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[jack880]

To all the people that have replied in answer to my question, thank you very much.

My original question was asking about how to carry out the calculation relating the water depth in the tank, and the speed of flow through the pipe (assuming I can calculate a system curve for the pipe).

Our client has a specific procedure sorted for taking the sample. Although I appreciate most people's responses anyway, I wasn't asking for advice on sampling procedures.

To TenPenny:

"How do I do something that I'm not explaining? Here's how I'd do it. I don't care how you'd do it, I want you to tell me how I should do it."
Do you not understand? I was asking how to carry out a calculation, and got a reply telling me how to ensure a sample line doesn't freeze. I wasn't asking how to take a sample. It's not a difficult concept.

"Typically, if you know the inlet pressure and the desired flow rate, you can calculate the appropriate pipe size."
Yes, I am aware that you can carry out this calculation, which is why I was asking how to do it.

"Since the OP knows his required flowrate, and must knw the inlet pressure, he must be able to calculate required pipe size. So why is he asking how to do it, without wanting to be told how to do it?"
Nobody told me how to do it. Most people just answered questions I hadn't asked.

Again, thanks to all the people who have genuinely tried to help me.