Calibrating VAV boxes

[tbedford]

We have tried to shorten the construction time of a 6 storey hospital building. There are 1626 supply, return and exhaust boxes controlled by a Honeywell ComfortPoint system. In the majority of areas, there are 1 or 2 ceiling diffusers for each VAV box. During construction, we blanked off duct work to lower floors and started calibrating VAV boxes on upper levels as spaces had duct work and diffusers installed and became clean. So in essence the calibration was completed in 6 stages over 2 months. To complicate matters, when all duct and construction was finally complete, basement air flow was found to be marginal and not at design. In some areas, the boxes starve for air.

In addition, we are now discovering that the flow values for our boxes will drift over time and the sensed flow is different from the measured flow. The controllers require a new calibration factor. The new factor isn't radically different but it is enough for the supply and exhaust volumes to noticeably change over time. This change alters the building pressures. Doors don't close, airflow whistles through the building, etc. In some cases, volumetric offsets are changed from design to accommodate correct air flow movement.

The TAB technicians did visually verify VAV controller operation, pressure tube orientation, etc. But the calibration program did not allow any other commissioning of the VAV boxes. As far as I am aware, end-to-end checks from graphics to VAV box has not been completed.

There are no quadrant dampers at duct trunks to each floor. Should these be installed to allow for the TAB agent to proportionally balance the air to each floor?

Did the sequential method of applying flow reference values as each floor was complete add to our final error?

What is the correct sequence(s) for calibrating multiple levels of VAV boxes?

While I am in favour of maintaining first principles, is there a quick way to re-calibrate the building?

Thank you

 

[Lnewqban]

It seems that you have several concurrent problems.
1) Ventilation: Is OA supplied through the VAV boxes? If so, the minimum SA flow of each box should not limit the legal rate of ventilation.
2) Air balance: OA fans (constant air supply independent from thermal load demands) should be balanced against EX fans (continuos or variable or intermittent operation) in order to achieve proper exfiltration (including gravity relief dampers).
3) T&B for each box should be done at maximum and minimum openings: what physically happens in between can be electronically manipulated via T-stat.
4) Whether or not your AHU's are able to supply maximum thermal demand for the whole floor or building determines how T&B is done: by zones, by floors or by whole building. The heat gain calculation can guide you through the decision about what boxes will be partially open while boxes to be at 100% are balanced.
5) The more volume dampers, the easier the balance. That includes upstream of each VAV box (not too close as to create miss-readings due to turbulence) and each exhaust, main and branches.
6) Consider VFD controlled fans (SA, OA and EX) as another variable to be eliminated in the T&B.
7) Treat smoke evacuation per floor as a separate balance.

"Where the spirit does not work with the hand, there is no art." - Leonardo da Vinci

 

[tbedford]

Thank you Lnewqban for your input,

We completed the VAV box calibration work in stages, by floor about 3 months ago. The air supply was spotty and very often we did not have enough air to meet the maximum flow even though there may have been 125 Pascals(1/2 iwc) or more at each box. Now when we review flow values, we find the value sensed by the control program does not match our flow hood readings. It is out by 10 to 20%. The calibration factor we originally applied is unchanged. Today, I have checked controller grounding, shielded cable and several other electrical issues. All appear to be in line with a standard installation

1. Outdoor air is supplied through the air handlers with mechanical cooling, then each VAV box with reheat coil for local area temperature control. Each box has a min/max value for occupied and unoccupied times.
2. Air handler fans are controlled with a VFD. Pressure sensors located near end of duct run transmit a signal to the VFD. The VFD will modulate fan speed according to how many boxes are open and closed.
3. The air supply and exhaust values were provided by design engineers according to Canadian CSAz317. Pressure relationships for Type 1 rooms were included and from a design standpoint are sound. For a practical standpoint, our calibration values originally used for flow reference are not changed. However, in about 30 % of the current flow readings, either the supply, return or exhaust air volumes sensed by program have drifted from our original values and do not now match the calibrated flow hood.
4. The calibration factor was applied with only a single reference. We wanted to use minimum and maximum flow values for reference but due to time restraints we didn't. I am wondering if the single flow reference is our issue. We are going to re-calibrate all VAV boxes but I do not want to start the process until I am confident any lingering control or software issues are dealt with.

Thank you,

 

[Lnewqban]

I see, but I would treat the system as a fixed volume one, balanced by hood and Pitot tubes where non-accessible, before chasing the tail of the sensors and controls.
I know it is complicated and I don't claim that I know a magic solution; just trying to give you some suggestions to step back and remove as many variables as possible.

In my experience balancing big VAV systems, it is normal that the hood measurements differ from the airflow that each ring sensor registers.
The first can be affected by location of the hood, time used for each read and turbulence inside it, while the second is even more inaccurate, depending on degree of turbulence and asymmetric air flow, induced by dampers, elbows and duct transitions that may be too close to the box.

Verify the working pressure recommended by the box manufacturer; most boxes go crazy for pressures under 1" w.c.
Even when delivering +/- 5% cfm or more variations, I tend to trust hood measurements over box sensors: that was the only way prior to DDC.

Unless OA is forced by a dedicated fan into the variable speed AHU's, actual ventilation may become less than legal for low rpm's of the AHU's blower (which are commanded by the pressure sensor via VFD).
The pressure sensor should be located about 3/4 of the main duct.
If the AHU's are unable to supply max demand from all boxes at once, then your only option is to isolate zones or floors the best you can, measuring flows when all boxes of that area open at max or min at once.
If you install dampers for floor branches, you should use traverse Pitot measurements to verify that the floor has plenty of supply prior to start hood balancing.
If suspected, you should consider a possible design error regarding demand versus supply and discuss it with the Engineer of Record.

I would leave the electronic adjustments and calibrations for after the manual balance is as close to specs as possible.
Hope you can achieve more solid and consistent results soon.
You are welcome

"Where the spirit does not work with the hand, there is no art." - Leonardo da Vinci

 

[tbedford]

Thanks again Lnewqban,

Do the VAV boxes require 2 reference airflow measurements at calibration or is one airflow low reference sufficient?

Thx,

 

[Lnewqban]

This article explains it better than I could:

http://www.facilitydynamics.com/assets/1/15/VAV_Airflow_Control_Simens2013.pdf

"Where the spirit does not work with the hand, there is no art." - Leonardo da Vinci

 

[EnergyProfessional]

I'm surprised in a hospital where you control pressures you use pitot-tube VAV boxes. Usually air-vales are used. We also use them for offices. They use vortex-flow stations are are factory-calibrated and almost immune to duct-effects. Especially at low flowrates they are way superior.

VAV operation depends on flow measurement. As long as you use devices that are not accurate, you don't have good VAV operation.

 

[ChasBean1]

1. No. You should NOT need to install balancing dampers in the MP duct work, unless the duct work were grossly mis-sized. Velocity between 500 fpm and about 1800 fpm is generally acceptable.
2. No. The sequential method was correct. As long as there is sufficient (but not excessive) pressure in the ducts feeding the upper level VAV boxes, and they were in their specified pressure ranges when calibrated, this follows NEBB standard and should not have been an issue.
3. You did this correctly. In theory, calibration of VAV boxes while in their control range (e.g., 0.3” to 3”) will stand even if you calibrated at 2” and later find an acceptable static set point of 1”.

The pitot tube airflow stations are tested and (in my opinion) true, so long as the transducers are of the proper range and the velocity pressure produces a repeatable reading, which tends to happen between about 1/4-7/8 of the transducer range.

Yes, boxes on short runs and elbows won’t perform well.

The design's overall building volume of OA and EA should be reviewed if not already done.

 

[tbedford]

Thank you ChasBean1,

Except for basement and some areas on first floor, the velocities were generally above 500 fpm. Having said that, there were times when we just packed up because control and Cx were playing with the systems and our air flow dropped off to zero or near zero. While the theory indicates that the boxes should be capable of calibration within a specified pressure range, from the practical side of the work I don't agree. We are having the least problems with the smaller systems that had a constant pressure throughout our calibration.

I did ask to have fan speeds fixed for our calibrations, however due to other ongoing testing and work, this was impractical. The schedule could not be met. The article reference supplied by Lnewqban was very illuminating.

I can state as a fact that controls is incomplete. We have returned on numerous occasions (and still do) to "re-calibrate" boxes that were registering zero...no flow. We discovered the VAV damper closed! I don't know why they didn't look at their graphic representation of VAV damper position and I have not yet observed end-to-end checks from Honeywell. All end-to-end checks are completed at the computer with "problem" boxes being investigated by ourselves.

As indicated our current headache is the "drift" away from our original calibrated values. In some instances, this ends up with a significantly different flow value than originally measured...20% difference! I have long suspected a control issue or VAV boxes with an improperly sized controller for the pressure ring.

There are 4 systems. Three of these are 100% O/A. The one giving us the most problems utilizes a free-cooling set of modulating dampers. This one also happens to serve the lower three floors of the building.

When I get to the bottom of this vexing little problem, I will surely post a solution.


Thanks,

 

[lilliput1]

Confirm the duct pressure loss calculation for the worst run is correct. Confirm design CFM to meet cooling load is correct. VAV boxes can only hold up to 3"wg. Try to limit them to see max 2"wg upstream pressure to minimize noise. Now know that the total pressure loss from the fan discharge to each supply air terminal must match the worst case supply pressure loss of the system. This is because the pressure at the discharge of each supply air terminal is athmospheric. This means that there has to be dampers to kill off excess pressure in runs with less rersistance. Providing dampers at each branch connection of the supply and return air riser would allow the pressure in each branch duct to be limited, ideally to 2" wg. This would also allow the designer to designate the correct pressure classification on the branch ducts, risers, mechanical room ductwork etc.
If the pressure in the branch is more than 3"wg the VAV boxes would leak out more air.
VAV box manufacturer should provide for each size VAV box the formula for controls to implement to properly calculate CFM based on measured diferential pressure on the unit integral flow (differential pressure)measuring element.

 

[lilliput1]

Also if it is critical to maintain pressurization in each floor then the total supply to the floor would need to be measured and the total return from the floor be controlled to a fixed differential from the total supply. Air flow measuring devices and return air control damper would be required. Provide control overrides if smoke control is required to pressurize adjacent floors and exhaust the fire floor.

 

[ChasBean1]

If it's proportional control, auto-calibrate boxes once per day.

What gets me: some parts of today's balancing can sometimes be BS. It's fine to throttle LP dampers to proportion flow among diffusers downstream of a VAV box. Example: you have a conference room that has four 250 cfm diffusers. You read them, and they are 200 cfm, 100 cfm, 400 cfm and 300 cfm. You throttle the BDs to get 250 each, which is fine.

BUT. A box that is properly installed with some lengths of inlet and outlet duct will have a Pitot type flow station: the concept of P = 4005 (dP)^5 [English units, pressure in inches w.c.] doesn't change whimsically. This converts to an accurate flow, as long as the pressure transducer is good and in its reading range.

If I take a balometer (flow hood) and read a vastly different value that causes me to have an adjustment (sometimes called "K") factor outside 0.8 to 1.2, I'd likely have an issue with the verification/balancing method rather than the installed transducer.

I posted this thought without reconfirming the original post. My apologies if I derailed the conversation.