Re: PRV control (activating when pump turns off)

Bummer, man.

Bentley has something that can help folks in your situation. It's called "BE Emplyable" and we provide free training to  folks who are out of work to help them be more employable. Go to the link

 http://www.bentley.com/en-US/Community/Academic/Be+Employable/?BI=homepage&v=promo+beemployable

With this, you can come back better than ever.

Good luck.

Ben/Mal,

Thanks for the insight, but all modeling is on hold becuase I just got furloughed (until the economy improves).

Running the scenarios individually works smoothly, but adding controls (on/off, closed/active, etc.) and alternate demands causes the system errors.

See ya,

Chris

Hi Chris,

I think in this case you may need to post your model up to the forums (WTG and WTG.MDB zipped) to have a look at, as the solution is going to depend on a few things.

To clarify though, it sounds for the refill pumps there is another potential condition to make them shut off before the tank gets above 2929.8'?  My approach would be the same, have a PRV active control and Pump Off control both using the same condition (or 1 control with a Pump Off/PRV active compound action).

I'm not sure if there are cases where you have controls with potentially conflicting actions, but that is where Priority comes in.  A control with a higher priority number (ie. 5 takes priority over 4) overrides the action of any controls with lower priority numbers that also have an action on the same element.  In the case of equal priority, is it the last control in the list of controlsthat affect a particular element, with a condition that evaluates to TRUE, that takes priority over all other controls with an equal (or no) priority

Normally a pump not being able to deliver any flow (in this case the domestic pumps) doesn't cause a network unbalance condition, but it can upset other things to cause this.  What and how won't know unless look at model.

As far single vs multiple scenarios.........I've seen and used both approaches, and the all-in-one scenario is still common, it originated back when there was no scenario management in modelling software, and the modellor would otherwise have been forced to create separate model files and run them individually.

My personal approach is to tackle different design events (ie. AD, MD, FF, PH) in different scenarios, as you are evaluating different elements of the network capacity.  You also have far more control over the starting boundary conditions in the model (eg. Initial Tank levels).   You may want to start your MD and FF scenarios with the tank near TWL......whereas for refill or AD you may wish to have the model start with the tank near BWL.

In the multi-scenario approach you can also mix-and-match EPS and Steady State scenarios.  For evaluating tank refilling capacity, you need EPS, but if you are trying to have a look at minimum customer/network pressures under Peak Hour.......I'll use a separate Steady State scenario in which I tell the model explicity which pumps are on/off, what the PRVs are doing, what tank levels to assume.......typically with "worst-case" combinations to give me the worst possible pressures (eg. upstream transfer pumps off + downstream transfer pumps on + tanks at minimum operating level).   If you use EPS to try to look at network pressures at peak hour, it's a bit hit and miss as to whether you'll see something akin to minimum network pressures or not at the Peak Hour timestep.

Ben,

Making the condition the same makes sense, but how should I control the PRV when the tank elevation is less than 2929.8?  I'm running into this issue on other system pumps.  The pumps are active until a condition is met and then shuts off, but the 0.1 hour time to shut off causes a system error.  Reducing the time step to 0.25 or 0.5 hour doesn't work.

For example, domestic pumps are running until a fire flow is activated (via demand curve) @ hour 7.  The compute window returns an error @ hour 7 that the domestic pumps can't handle the fire flow (causing a network unbalances code).  I'm trying to document an average day, max day, fire flow, peak hour and refill requirement in one 48-hour scenario (a real-world situation).  How is it possible to have domestic pumps operating, but immediately turn off once a flow increase is detected?

Or should I just keep the individual scenarios (fire, refill, max, etc.) and compute?

Thanks,

Chris

I still think Mal is on the right track.

Instead of making the Pump On/Off as the Condition for the PRV becoming Active, just use the same Condition that turns the Pump Off as the identical Condtion that makes the PRV Active.

Then the Pump will turn Off, and the PRV become Active, both in the identical timestep.

Use this for Eg. 

•  If Tank > 2929.8 Then Pump 1 Status = Off
• If Tank > 2929.8 Then PRV Setting = 105 psi

*EDIT*  Just had chance to look at your diagram closely.......looks almost exactly like one of my first few WaterCAD models (with same small things causing errors/warnings that had me scratching my head and spending hours trying to diagnose).............if only I knew then where I had been going wrong....so many hours would have been saved! ;-)

First, you should remove the check valves from the pump pipework.  I know there may be 1 there physically, but a model pump element automatically incorporates an in-built check valve.  Having extra check valves like this in the model will often cause disconnection errors.

It also appears (hard to tell) you have controls on the pipe upstream on the PRV, rather than the PRV itself......again this will cause problems (probably disconnection errors).  Make the controls instead directly on PRV by either activating (by using a PRV "Setting"or "Grade" Action) or Closing it.

Not knowing the system exactly, I don't think you need 2 PRVs?  I think you are just using 1 to control the outlet pressure discharge back to the main system.  The PRV you have nearest the tank I think will just need to be an Open/Shut valve (ie. use a TCV, or just a plain pipe), as it doesn't need to duplicate the pressure control the first PRV nearest the main system is already doing?

Anyhow, quick and dirty example model here of best guess of what you are looking to do.  The bypass lines on my model are optional but allow for any potential staggered pump operation to allow flows to be drawn through any non-operating pump station, and head off any potential disconnection errors from this.  I could have also used a single control with a single condition, but a compound Action, but instead just showed it the "simple" way with 4 separate controls for 4 simple actions.

Mal,

Attempt #2 - PRV1 and refillpump 1 are at the top, PRV2 and refillpump2 are in the middle and the tank is at the bottom.  During a fire event (6,000 gpm for 4 hours) or domestic flow, the tank empties through the two PRV's for system supply.  The requirement is that the tank must be refilled within 24 hours of a fire event, necessitating the refill pumps.  The model works without the controls, but we're trying to make it as realistic as possible (with PRV's, pumps switching on/off as demand changes).

Changing to simple controls creates a "Reference to a deleted or inactive condition element" error.  May havta try Composite conditions/actions?

So you have a logical control on the PRVs that says something like "set PRV pressure setting to 105psi when the pump is off"?

I think the problem is that the PRV control isn't checked until after the pump goes off - giving you that 0.1hr delay (logical controls are checked every 1/10 of a hydraulic timestep, so if your hydraulic time step is 1 hr then this explains the 0.1 hr delay).

I can't really tell without seeing the model or a schematic, but do you actually need a control on the PRV? Could you just set it to 105 psi initially and leave it at that, without using a control?  

Alternatively, you can try evaluating the PRV controls as "simple controls". You can't set a priority for "Simple Controls" and they can't have an "Else" clause, but they are evaluated as part of each series of iterations for each time step. That means that if the controls say the pumps turn off at a certain time step (e.g. 23.4 hrs), the PRV should open at that same time step (23.4 hrs).

Note that simple controls can sometimes cause adverse effects (for example the calculation engine can find it harder to converge on a solution if a pump or PRV is flipping from open to closed over successive iterations within the same time step), but it might give you the results you are looking for.

Let us know if that helps.

Regards,

Mal Sharkey

Bentley 

A little model background:  Tank at 2715 elevation, 2745 max (initial elev 2729.3), 16" discharge line. PRV2 (and refill-pump2) @ 2420, PRV1 (and refill-pump1) @ 2320.  Start the model with pumps on and PRV's closed and everything runs smoothly.  Once the tank reaches 2929.8, the pumps are supposed to turn off (first control).  Once the pumps turn off, the PRV's are supposed to activate to 105psi (second control).  I get pump turnoff (23.4 hours), but it takes 0.1 hour for the PRV's to activate, which causes a system disconnect because of prv/pump error.

Comparing the results table shows that PRV2 is activating @ 23.5 hours, whereas PRV1 is not (both pumps shutoff @ 23.4 hours).  I'm trying to make the model more realistic, but will discontinue if there's no easy solution. The EPS is 24-hours long.

Any idea on the potential fix (validate shows no problems)?

Thanks,

Chris