Controls in a Clean Room System

[friartuck]

Hi Folks

I need some help on how to control the AHU in a clean room system.

I have constant volume regulators on the terminal HEPA filter boxes which are designed to hold a constant flow rate regardless of the supply system pressure. But heres the BUT...

I would like to control the AHU with a velocity probe and Inverter (VSD) to the fan motor.(to save energy)

My question is, that the controls will in effect be 'open loop.' and there will be a conflict between the inverter control and the CV boxes.

How do I set up the controls? If I set the AHU to perform a constant velocity, what stops the CV regulators from going wild and closing if the airflow from the AHU is a tad too high.

Am i better setting the inverter VSD drive to control from a set supply duct static pressure rather than a set velocity. That way, the CV's will adjust to the main duct pressure.

What I would like to do is save energy by having the fan ramp up gently as the HEPA filters get dirty.By doing a simple set supply duct pressure control, i won't achieve that.

I have considered monitoring the terminal HEPAS and setting the supply duct pressure such that it will adjust in steps according to the highest terminal filter pressure drop. i.e if I monitor 5 terminal hepas, if the fan supply pressure is set initially at say 400Pa, i can adjust its set point in step with the hepa filter incresae. So if the hepa PD goes up by 100pa, I simply increase the control point by another 100Pa to give me 500Pa at the supply fan outlet.

Any suggestions



Friar Tuck of Sherwood

 

[imok2]

Tuck this article may be of interest for you

http://www.hpac.com/microsites/networkedcontrols/clark_9901/clark_9901.htm

 

[SAK9]

The supply terminal HEPAs will load up evenly.So there is no need to monitor the pressure drop across each filter.The simple way of controlling this sytem is by duct velocity control.Static pressure control for such a system is complicated as it needs to be reset in line with filter loading.Some prefer to use room pressure to control the AHU speed if it is serving a single large space.The static pressure in the duct for such a system is constantly changing(increasing).Therefore it requires a dynamic balancing system such a CAV box.Manual balancing with VCDs may not work properly with a varying duct pressure. You can employ a common CAV box for a no of filters in a room.
The pressure drop across the CAV box in such a system will be minimal.

 

[KenRad]

Are the "CV regulators" dampers in the filter boxes, or fans?

I think you're on the right track with using a fan VFD to maintain a constant velocity pressure. The CV regulators will no longer be needed, and should be disabled or you will not have stable control. The system will adjust for filter loading, and your fan energy will be reduced. An added benefit of the VFD is the soft-start that is easier on all drive components.

---KenRad

 

[quark]

The most critical thing is monitoring the pressure drop across HEPA filters. Door openings drastically change the differential pressure readings and this may create some nuisance.

SAK has a good point that HEPAs load evenly. This is due to the fact that efficiency of HEPA increases as it gets loaded. So there will be more air flow through other HEPA filters and they eventually get loaded. This is called as Plenum Effect or something(I couldn't recollect now)

We have been doing flow control with velocity sensors and VFDs for ages and no problems so far. You need not change the settings periodically. You can optimize your savings continuously yet you can have the required flowrates.

Main drawback from your proposed control, as far as energy optimization is concerned, is that you can have optimum energy usage at 400 and 500Pa as you reset the pressures, but not when the pressures are 401, 402.... 498 and 499 etc.

I agree with Kenrad that CV Boxes are redundant.

Regards,

 

[friartuck]

I put the CV boxes on each terminal filter on the advise of the client who had used them before. He liked the benefit of the reduced commissioning time since all is needed is to set each CV box and ...voila, (in theory anyway) the system is set up. Hence they are fitted and I don't think the client would be too pleased to throw away 30 CV boxes.

I think in hindsight the CV boxes looked good, but may be redundant. Is there a strategy to keep them in use and still use VSD drives on the AHU?? i.e. as i suggest, set the controls to provide a nominal static pressure (300Pa would cover the distribution system) and then adjust on the worst HEPA filter pressure drop.

I know its not a straight forward option, but can anyone think of an alternative without trashing the VC boxes.



Friar Tuck of Sherwood

 

[quark]

I don't know what is pulling you back from using a velocity control. Is it a cost factor?

If I presume that your CV boxes are constant volume flow boxes(just fan filter units), then I reckon you are thinking of keeping a constant differential pressure across the secondary fan(or CV fan) so that you can get constant volume flowrate across the HEPA irrespective of its loading.

But how will you ensure that the increase in static pressure due to HEPA loading gets replicated in the common duct work? In other way, the CV fan suction pressure will be constant but its discharge pressure may vary to overcome HEPA loading(i.e CV fan running left on the curve to its initial point) You may have to place your pressure sensors after the CV fan discharge. If you have multiple HEPAs this becomes a problem.

If you go with constant velocity, as soon as the CV box fan runs to the left of its original point, there is a reduction in the flowrate and thus you will get clear indication.

We generally use FFUs to supplement the pressure where ever necessary. This way the main AHU fan size can be optimized. As you have already sized your main AHU fan considering the filter loadin, what I suggest you is to put off the CV fans initially and start them when your main AHU fan runs at full speed. This way you can also eliminate inefficiency of smaller fans.

Further, speed control w.r.t pressure is a square root control where as it is linear with flow control(ofcourse, flow is a function of (dp)^1/2 but measuring error gets cancelled as it is dp. Only instrument error comes into picture here)

Regards,

 

[SAK9]

Quark,
Please could you elaborate on the last sentence of your post?Are you meaning to say that velocity control is superior as the relationship between air flow and fan speed is linear?That is an interesting point I have never thought of.

cheers

 

[quark]

SAK,

Actually I pointed out the accuracy in readings in my earlier post. Generally any error in a reading is the sum of measuring error and instrument error. In a dp sensor, the measuring error can be minimum.

Further a flow control is a better one and the following example shows this.

Say you have to control a fan which was designed for a duty point of 4000cfm and 100mm static.

The theoretical fan laws say(the exponential coefficients differ in actual case)
Q1/Q2 = N1/N2 and SP1/SP2 = (N1/N2)²

A decrement of 1Hz from full frequency results in a new flowrate of 4000(49/50) = 3920cfm. Reduction in flowrate is (4000-3920)x100/4000 = 2%

Pressure will be 100(49/50)² = 96.04 or the pressure reduces by 3.96%.

If we can control the flow by 2% decrements, it needs 3.96% minimum decrement for a pressure control. That means even if you require to control the pressure by 2%, you can't do it.

Having said this, I don't totally rule out pressure control. But for a constant volume flow systems it is better to go with flow control. If you have variable volume flow systems then pressure controls may work better than volume controls.

What I want to emphasize, in both cases, is that it is painful operation if we have to reset the control signal frequently. It is better if we choose a control signal that need not be reset.

Hope I tried my best to answer your question.

Regards,

 

[friartuck]

I think I might not have explained the system very well and this is confusing the issue. I do not have terminal fan boxes. The system is as follws:

input air handling unit (rated at about 6m3/sec)
fan withe inverter control (however this is controlled is yet to be decided)

distribution ductwork serving about 30 terminal filter grilles (without fans)

each filter grille has a Trox type RN constant volume regulator. (go to Trox.de, click on the bottom LHS, on the word English and then follow the trail through products to control units to constant volume boxes type RN). The site is German with an english option to view.

The CV boxes comprise of a sensor and a bellows which inflates when the airflow increases. This closes the damper gradually to achieve the desired pre-set airflow.

Now if I have for argumants sake 6 big boxes at 1m3/sec each and set them to 1m3/sec each, then theoretically I could set my AHU velocity sensor to provide a constant 6m3/sec. BUT..... if the CV boxes are not accurate in there controllability (i.e. within 10% as these generally are), then I might get a flow rate of 6 plus or minus 10% i.e. 5.4 or 6.6 m3/sec (with me so far)

Now, if the CV units attain 5.4m3/sec as they might within the given tolerances, the AHU will see 5.40m3/sec as a deficiency and it will increase the fan duty/velocity in order to try and achieve 6m3/sec. The CV boxes will then try to address the additional airflow by closing down and throttling back the airflow....and so the AHU will speed up to compensate...in a spiral of incontrollability. I think you get the picture.

So because the AHU control (by constant velocity probe control) is not directly linked to the CV box operation, they will have some conflict. I therefore need to make one system be able to interact with the otherin some way to prevent this conflict.

Conversely, if the CV boxes deliver more air due to there tolerance of accuracy, they will try to allow more air to flow into the rooms than the AHU is set to deliver, and it will consequently drop in speed creating a deficiency of airflow and the CV boxes will adjust accordingly delivering less air than they were set to achive. They would pulse up and down.

Does anyone appreciate this effect and more to the point, is there a solution??

Friar Tuck of Sherwood

 

[lilliput1]

For large systems, we typically control the fan VFD to maintain a set point minimum static pressure at a selected remote point in the ductwork & let supply air boxes maintain constant supply air CFM to individual rooms and ther HEPA terminal diffussers. This is the most flexible arrangement because with the boxes you can do the following:
. Do airflow tracking of air in & air out of the room to maintain design CFM differential for pressurization
. Have an occupied & unoccupied mode for individual rooms that can be overriden for say 1 hour intervals when an ovveride button on theroom temperature sensor is pressed.
. If additional boxes are added you don't have to reset the AHU CFM if you have excess available from diversity.

You are correct in being concerned about possible control hunting or fighting with both the box & fan controlling for CFM airflow.

For your case however since you do not have terminal boxes consider removing the Trox unit, just use the manual volume damper (typically furnished w/ HEPA filter diffusers) and control the VFD for the CFM. The HEPA filters should load up evenly so once balanced to compensate for the variation in ductwork configuaration to each, the Trox controls would be redundant.

 

[SAK9]

Friartuck,

If you set the AHU flow as 6m3/s,there will be a flow of 6m3/s in the main duct.Assuming the 1st two Cv boxes allow and 1.1 m3/s each thru them the remaining 4 CV boxes need to be content with the balance 3.8 m3/s.Let us assume this gets equally divided into 0.95 m3/s and flows out thru the 4 boxes.

Whether a CV box passes 1.1 or 0.95 m3/s,there is no feedback from the boxes asking the AHU to speed up as theses are mechanical devices.The Trox CV boxes are nothing but self balancing dampers which regulate according to the duct pressure.Ther is no measurement of airflow in the box.When there is insufficient flow into the last(index) CV box,all it can do is to open 100%.It can not influence the AHU flow.

I have commissioned a few systems(exactly the same as you described here) with Trox RN CV boxes as well as Phoenix Venturi Valves.I have not come across any hunting
problems.The +/-10 % tolerance on airflow is applicable only for a certain pressure range.If you exceed the pressure range(quite likely in the brnches near the AHU),the tolerance range can be much higher than 10%.Since it appears your system has been installed,put it to work and see how it responds.
Quark thanks for the illustration.you hve driven home the point!

 

[Icanfix]

Why can't you just use the volumn regulators to modulate the VFD?

 

[quark]

I think Friartuck seems to have a valid point here. Unlike PD devices, a centrifugal fan can't pump in required volume rate continuously. If the inaccuracy of CV boxes create extra pressure drop then the fan flowrate gets reduced and VFD tries to speed up the fan and subsequently a bad control occurs. But one has to confirm constant flowrate at all inervals between two pressure resets. This is because all the cleanrooms require constant flowrates despite of filter loading(this is compulsory atleast in Pharmaceutical Clean Rooms). This is the main drawback for pressure control.

If I still go with a flowrate control, then I will first balance the system manually taking readings from all the terminal units and set all the boxes at values a bit higher than what I require. Say, instead of 1m³/sec it will be 1.1m³/sec.

Regards,

 

[ChasBean1]

Friartuck,

I was looking over the Trox literature and apparently the device can be equipped with an electric setpoint adjustment actuator. Do your CV boxes have this option?

If so, you can first maintain duct static pressure at some point in the system (typically two-thirds of the way down the main duct) sufficient to satisfy all CV devices with clean terminal HEPAs. Then, if any one actuator approaches a 100% open command, the duct static pressure set point can be incrementally adjusted upward to assure constant flow at all CV devices. You could likewise use the static pressure setting or the actuator positions to trigger alarms or warnings.

Personally, I would not be in favor of a velocity, volume, or velocity pressure control for the central system in series with these CV devices. The Trox units control flow - the central system should only control static pressure.

Best of luck, CB

 

[walkes]

Friartuck,

In the original post you said you wanted to ramp the fan up slowly as the HEPA gets loaded.

That doesn't make sense to me as the CV boxes should adjust as the HEPA gets loaded to ensure constant velocity. As the CV boxes adjust the shouldn't system static pressure remain relatively constant? (Assuming that the system is set for full flow with HEPA's fully loaded). So it would seem to me that as CB said maintaining the static pressure should suffice with the CV boxes.

 

[friartuck]

Hi walkes

Lets take a typical HEPA with an initial PD of say 100Pa. The airflow req'd might be 200l/sec. Now at start up, the CV box will adjust to give 200l/sec. Since the sytem is clean, the overall supply ahu fan duty might typically be 100Pa (for the HEPA) plus say 200Pa for the main distribution ductwork. The fan ststic pressure will initially be 200+100=300Pa. hence the fan duty will be 200l/sec at 300Pa.

As the HEPA dirties, (gradually up to say 500Pa), then the CV box will see a drop in flow rate as the fan is not capable of overcoming the newer-higher HEPA pressure drop.

The CV box will open up to try and meet the demand but if the fan has not been subsequently adjusted (speeded up)then the system will under perform.

I suggest that a pressure probe is used at the HEPA end to measure the HEPA pressure drop (in this case it would see a pressure of 500Pa since the filter has dirtied) and the main fan control can subsequently be re-adjusted to control at the 500Pa plus the original system drop of 200Pa. (hence the fan outlet static pressure control point would be 700Pa when the system was dirty.)

Does that make sense??


Friar Tuck of Sherwood

 

[lilliput1]

For your case however since you do not have terminal boxes consider removing the Trox unit, just use the manual volume damper (typically furnished w/ HEPA filter diffusers) and control the VFD for the CFM. The HEPA filters should load up evenly so once balanced to compensate for the variation in ductwork configuaration to each, the Trox controls would be redundant.

Measuring the CFM supply & controling the fan VSD without the TROX units would be more accurate CFM & relative pressurization control. It also will use the least energy.

 

[imok2]

Have you looked at the possibility of averaging all the CV box pressure drops to reset the cfm supply? Anyway I think you should read this article, it may give you more insight

http://ateam.lbl.gov/Design-Guide/dghtm/helpcontents1.htm

 

[quark]

friartuck,

If you want to go with pressure control, then I think CB's approach will give you better control though energy optimization may not be perfect(because you are fixing 2/3rd of pressure as constant irrespective of flowrate). Monitoring pressures at HEPA end may be a futile exercise.

Please let us know when you finally do it as I am very much interested to learn about it.

Regards,