Where to Put The DP Sensor in Constant Volume Systems

[71CHOPS]

But how do you use valve position for speed drive hz if you have a hospital with 10 different AHU’s with 10 different load profiles?

Or....better yet, a campus wide chilled water plant....with 10 different buildings, with 10 different AHU’s each?

[heatingman]

But how do you use valve position for speed drive hz if you have a hospital with 10 different AHU’s with 10 different load profiles?

Or....better yet, a campus wide chilled water plant....with 10 different buildings, with 10 different AHU’s each?

I wouldn’t.


Sent from my iPhone using Tapatalk

[jayguy]

If a valve is 95-98% open, speed up the pump. You usually don’t wait until it gets to 100%. they may not get to 100% so you pick an arbitrary number close to 100%. There are drawbacks like any other system. Such as inaccurate valve positions. Of course, a DP sensor can be off too.

as the “most open” valve drops to 90% or something, you slow down the pump.

[Cagey57]

Wayne,
Here's my take on your question;
I believe that the function of DP control of large(er) hydronic loops/systems is to maintain the correct flow at all of the connected coils.
With that as a qualifier I was taught, and later confirmed with the TAB pros, is that for the secondary pump to be be able to deliver correct flow (or in this case DP) to the most remote coil the DP has to be either just before or at the last or most remote coil and piped to measure the DP delivery pressure to that coil.

Keep in mind, I have been a Controls Guy since 95' so I could be working off of a basic misconception.

[Cagey57]

If a valve is 95-98% open, speed up the pump. You usually don’t wait until it gets to 100%. they may not get to 100% so you pick an arbitrary number close to 100%. There are drawbacks like any other system. Such as inaccurate valve positions. Of course, a DP sensor can be off too.

as the “most open” valve drops to 90% or something, you slow down the pump.

Jayguy,
I remember seeing Sequence's that called for that exact type of control. My big question, which still pretty much remains unanswered to this point is, can that strategy co-exist with DP control (I say no for now) ? I deal with CE's that , as far as I can tell, do NOT know what the term (or reference) "Cv" means (simply put it is GPM flow at 1 PSI of pressure drop).

[71CHOPS]

CV rating, as I understand it, is a ratio of pressure drop vs. flow rate across a valve.

On smaller stuff, there is a “close” relationship to actual “GPM”, but you can’t use it as a flow rate only.

[jayguy]

Jayguy,
I remember seeing Sequence's that called for that exact type of control. My big question, which still pretty much remains unanswered to this point is, can that strategy co-exist with DP control (I say no for now) ?...

I would also say no.

[Gdh33]

Jayguy,
I remember seeing Sequence's that called for that exact type of control. My big question, which still pretty much remains unanswered to this point is, can that strategy co-exist with DP control (I say no for now) ?

You can do anything you want in controls , but why would anyone do this? I would do it one way or the other. Dp is the simpler standard way to do it. With more information and computing power, seems Ashrae and Enginèers are going towards demand control ( valve opening).

[heatingman]

CV rating, as I understand it, is a ratio of pressure drop vs. flow rate across a valve.

On smaller stuff, there is a “close” relationship to actual “GPM”, but you can’t use it as a flow rate only.

Cagey is right

Cv value is the number of gpm required to have a 1 psi drop across the valve, or coil, etc..

Its useful to know, because it can be used to calculate current gpm from a current pressure drop.

I used to know the formula, and could probably dig it up, but now days I just use Bell and Gosset app to do the math. The tab called unknown pressure drop will do the math. Enter the Cv rating on the top, then enter 1 as the PSI, then enter the measured pressure drop, and hit enter again and it will fill in the blank flow number, and also spit out the Cv.



So lets say the listed Cv of a valve is 15.

And I measure 2 psi drop across the valve, now I know the current gpm through the valve is 21.2 gpm or very close to it.

This also works and perhaps more accurately for chiller barrels for example.

The MFG will have a design pressure drop at a design flow rate.

So those become the benchmark. Its not exactly Cv, buts its used the same way.


Sent from my iPhone using Tapatalk

[WAYNE3298]

Cagey57 I'm with you on DP sensing at the most remote coil. I would add on a large enough system that demand can vary which coil will be starved to have more than one sensor.
I'm not criticizing control using control valve position because that will also work. I think both are better than 2/3 down the duct. The problem I have with control valve Cv is you rarely if ever see taps to measure pressure drop. If you have the taps it will work.

[flange]

I stop by every few years to see what's up.

The manner in which I was taught is as follows:
There is a specific design pressure drop for your control valves. That is how they are sized. Many new jobs are sized for ten psid at design flow, however this is a guess based upon the piping layout, and will change slightly due to field installed conditions, which may vary. Maintaining pressure differential was typically measured at 2/3's down THE LONGEST anticipated run. This would account for the proper flow and head at the pump, and ensure all other runs had enough. BUT it needs to be verified at start up.
How we always did it was to open all control valves via the front end. Have the balancer set the pump at the desired flow, as well as branch circuits. We would then back down ten to twenty percent of the control valves/flow to give us diversity ( on speed controlled pumps), since in most cases the building wouldn't be at 100% on all units. Once we achieved our diversity, we would then measure ACTUAL differential pressure and speed of the freq drive. This would become our setpoint and reference. We might give it a hair more, but not craziness. At that point, you should be able to look at the pump curve and see where you stand with respect to actual design, knowing that on a design day, you might just need that added twenty percent you allowed for diversity, but whatever you had left would be what you got on a design day with all zones calling. Once you set your design differential pressure, the balancer can go about his business. A cursory check at the end to verify things still looked ok is a nice touch.

I have used this procedure on buildings up to sixty stories with very good results. On a few jobs we ran into pumps without enough left for design days and had to do impeller changes, but not often.

[WAYNE3298]

jayguy your post about controlling pump speed with control valve position has stuck in my mind. I was skeptical at first because my thought was if the control valve wasn't properly selected you would have problems. It occurred to me that as long as the coil is properly conditioning the air that really doesn't matter. You have convinced me using control valve position is the best method. Since all control valve positions are monitored the control program simply has to select the most open valve to satisfy.
Never too old to learn. Thanks for your input.

[jayguy]

Your welcome! Of course, I didn’t invent it. There is a downside. Obviously, every actuator has to have a feedback signal..which is more money. And this means that every actuator could overdrive the system if it sticks or becomes flakey instead of just one sensor to fail or get flakey.

the major upside is that the system may run a lower pressure differential than design saving money. Also, you flow a bit more water in places you might not have before and this keeps the headers “at temperature”.

[WAYNE3298]

What I like about it is the constant optimum operating pressure. Water balance of every coil is also unnecessary which means circuit setters are not needed. That all means payback. Excess water flow shouldn't be a problem because the control valves on the coils that don't need the water will throttle back and push the water to the coils that do need it.

[71CHOPS]

One problem I see would be system design.

If you have a system that’s poorly designed, and doesn’t flow correctly, the valves closing still may not push water to the ones that need it.

Or if you have multiple AHU’s on a take off from the main loop, you would still have flow issues. I.e. if the first AHU is calling for 100%, and the 3rd is also calling for 100%, the first one on the run is likely going to take the majority of the flow from the others. In that case, you would still need balance valves.

Also, if you have multiple loops interconnected, balance would still be an issue.

Interesting concept, but I think, it has positives and negatives.

[WAYNE3298]

Any of the flaws you mentioned CHOPS would be a problem no matter how the loop(s) are controlled. Controlling with valve position I think is better used on smaller coils such as VAV's and FPB's. My preference on AHU's and other big coils is discharge temperature combined with circuit setters. A little mismatch on a small coil isn't that big a deal but the same problem on a big coil can hurt you. Excess flow on heating or cooling coils doesn't add much capacity and should be avoided.