Differential Pressure Transmitter in Chilled water loop 3

[Manu68]

Hi,

Are there any conditions or lessons learned for deciding the no of differential pressure sensors (DPS) required in a closed loop chilled water piping that have a primary and secondary loop that have VFD. If the DPS is connected at the far end terminal unit say FCU, how would it affect the 100% load requirement of the terminal units in the nearer side. To manage this multiple DPS cases are there any lessons learned from experts.

Thanks,

 

[EnergyProfessional]

You have 3 choices:
- if you have one larger coil (meaning the largest dP inc. pipes), tha tis hwer eyou locate the dP sensor
- if none of the coils really isdifferent (say you have just 20 FCU), put it 3/4 of the main run
- if you have multiple large branches (i.e. serving 2 buildings), you may need mutlipe sensors

If you implement static pressure reset based on valve positions, the dP setpoint will be increased when a coil doesn't get enough flow (is 100% for long). So no worry.

 

[DRWeig]

Hi Manu68, welcome to eng-tips.

I will respectfully disagree with HerrKaLeun on this one. The last coil in the piping run is the location I have always used, and results have been always good. This most-remote sensor guarantees that all coils closer to the pump will have more than the differential pressure they need, whether they are large or small.

The piping loop cannot have coils with differing design pressure requirements unless coil booster pumps are employed. It's just one loop and the differential at each coil is the pump differential minus friction losses.

If you locate the sensor 3/4 of the way to the last coil, the coils beyond the sensor may have too little pressure at times of high load and the system will not be able to compensate for this.

Only one sensor is needed. Reliability is quite high with these devices. However, they do fail sometimes. Failure mode is almost always going to be zero output, so you could conceivably put two sensors on that most-remote coil and control to the one that is known to be working when one fails. That would be the only reason for more than one sensor, but I've never seen it done.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[EnergyProfessional]

DRWeig: it probably doesn't matter so much, especially if he uses static pressure reset it is kind of self-commissioning. In your case the furthest coil, likley often was the one with larges pressuredrop.

It isn't jsut about fuctioning, but if you can runt he pump a bit slower on average, you save a lot.

What really is important is that when balancing the static presseure setpoints and ranges are set correctly. What I see the balancers never talk to the controls contractor, and the balancers have no clue.

 

[DRWeig]

Thanks HerrKaLeun, good points.

You're right about the disconnect between balancers and controls folks. That could lead to the old controversial discussion as to whether balancing valves are needed at all -- but that's a topic for another thread.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[Manu68]

Thank you HerrKaLeun & DRWeig for helping with your experience.
In the past multiple DPT gave problems for the pump control, eventually had to resort to the one prior to the last RTU/highest pressure drop RTU.

 

[Drazen]

Drweig,

The problem with having sensor at the end of long loop would be too much pressure at the coils closest to pumps, which cannot be overcome with coil balancing (even automatic balancing) valves. such scenario can take place in case of partial load, of course, which is always the worst case for balancing analysis.

Of course, measures must be taken to supply adequate dp to the most distant coil. if automatic balancing valves are used, that can be solved by applying minimum dp on the most distant coil (at design stage), so that balancing valve can slightly open when not having enough flow.

3/4 is arbitrary figure taken by analogy with vav systems. in most cases that should be verified during calculation stage if possible.

i strongly believe, and it is confirmed in practice almost every day, that vsd pumps and automatic balancing valves, while relieving lot of commissioning burden, put significantly more burden on design work, design have to be more precise and more elaborated, or new technologies will not only be waste of money, but they can even show worse results than traditional techniques.

 

[DRWeig]

Agreed, Drazen. The overpressure at nearby coils can disrupt control, particularly if the loop is really big or long. I haven't had such a problem arise in my experience, but I've never done controls on really huge loops. If you forget the automatic balancing valves or PICCVs or other what-not, the pressures along the system loop are easy to calculate.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[EnergyProfessional]

Drazen; good points, especially in large systems. The Pressure indep. valves have 50 psi or more in pressure resistance. So in small to medium sized systems I never overcome that.i laso tend to keep pipes relatively large to have less dP, that way the furthest coil still gets much P, as opposed to reverse return.

DRWeig: yeah we had the discussion on balancing valve. I tend to agree with you to entirely skip them, or to use PICV. but not the traditional balancing valves.

Location of sensor is kind of arbitrary. ASHRAE says at largest coil, or 2/4. I also heard 3/4. why not 3/5, or 4/6 (I use that for 3/4 of main pipe since no contractor can calcualte 2/3 inc. branches anyway. but having it at last coil sounds reasonable too. It depends on the main pipe dP, and how much the coils differ.

 

[Jazzerman]

I've recently retrofit variable speed pumping onto a couple decent size hydronic systems. In each project I've put 3 or 4 dp sensors in, because I was not sure what branch or area would be the worst case run (but I had some educated guesses), and wanted to get a picture of the overall pressures in the system. In both projects (one is chilled, one is heating) is has ended up being one certain sensor that is consistently the minimum, and the pumps just have to control to that one. So the other sensor simply became informational. As long as the coil loads on the system have demands that rise up and down following similiar profiles, it doesn't matter so much where you put the sensor, you can always adjust the setpoint to make the system happy at the worst case run, even if the sensor is in the middle, or even near the pumps. Where you end up having the need for multiple sensors and maybe wanting to watch the min of all the sensors is where you have coil loads that peak at very different times of day.

 

[Drazen]

diligent commissioning or maintenance engineer can set up system with virtually any sensor position with detailed monitoring of system branches over some time, than applying artificial sensor setting which will satisfy all needs.

the question is, however, what is achieved that way in terms of efficiency. it can easily come to situation where most of large network has excessive pressure all the time, number of local controllers operate at the edge of usable working range, there can be noise and apparently increased maintenance costs as well.

it really depends on size and complexity of system. i firmly believe that most of those issues should be solved at design stage, cut currently we have very aggressive picv market representatives who approach to clients directly splashing them with promises how picv' solve all issues by themselves.

i have enough work so that i don't rely on hopes that clients will pay for more detailed design, but i really see two things in practice: increased maintenance costs and failed energy efficiency expectations, which i believe come out of reasons stated above.

it happened to me more than once that picv representatives were not able to provide specific answer to specific question about hydronic loops, but stay with their stories how everything will be resolved by itself...