Ok, I'm looking at this from the shaft end out.

1) is this a common plenum with all four fan feeding into a common area?

Options I can see:

1) set a greater dead band on the pressure switches for on/off operation of the second, third and fourth fans
2) Just install a fixed timer on the switchgear so that the fan runs for a min of say 5 minutes - gets you to max 12 starts/hour
30 Fit a simple ratchet to the fan motors to prevent reverse movement
4) Fit solenoid or low force spring powered louvres to the 2,3 & 4 fans. I'm astonished that you're wasting airflow this way - why??
Fit VFD drives and then run two fans at 50% each, then 60, then 70... then 3 fans at 50% flow etc etc. Many less starts.

Not sure why your engineering consultants are being so stubborn and not seeing what is happening in front of them - do they share the cost of these or something?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

@LittleInch, I appreciate the suggestions, it's great to have an independent review overlap with some of our ideas. Our engineers and independently the consultant has also batted around the idea of the timer and pressure switch adjustments. I initially suggested anti-rotation devices but am not sure if I have enough free shaft length to install them without changing the motors or mounting configuration. I'll pass these along to engineers on my team.

The units are a common plenum with all four fans mounted to one face. No louvers to eliminate a failure point, although the previous machines had louvers so I don't know why they had to be eliminated. It doesn't help that they oversized the units on top of everything else so that we would have margin for condensing coil degradation (we are in a salt air environment).

The reason the consulting firm and project team is being stubborn is likely because due to the issue we convinced our plant manager to stop work and pause their payments until the issue is resolved. We have several trains of these units and are replacing them one by one. They are in the beginning stages of replacing the final train and won't get their last draw until work is allowed to proceed, so anything we change will be to be retroactively installed on the previous machines.

I had a similar issue, but not identical.
The root cause was the setting of the overloads too close to the motor FLC.
We had three fans per coil which ran together.
One fan would nuisance trip.
The pressure of the other two fans would windmill the tripped motor backwards.
When the overload was rest, the motor would try to start.
As the motor was slowing and restarting in the correct direction, we could see the overload heaters glowing red hot inside the overload relays.
Two solutions;
1. The overload relays were set properly.
2. The operators were instructed to stop all fans in a set before resetting an overload.
Not your solution, I know.

Reversing power to a spinning motor is called plugging.
Plugging was used often with "U" frame motors.
Plugging was used with a special zero sped switch to stop a motor almost immediately.
The contactors were rated for plugging duty.
With the introduction and widespread use of "T" frame motors, plugging went away.
"T" frame motors tend to fail when used for plugging duty.
(Try to negotiate a quantity discount for replacement motors.)
There are two avenues for correction;
For an electrical solution, the motors should be replaced with "U" frame motors that are rated for plugging duty.
The contactors should be upgraded to contactors rated for plugging duty.

The mechanical solution;
Many similar schemes work well.
What is wrong here?
When staged motors are used on condensing units, each motor is mounted to a separate section of the condensing coils.
The separate an sections typically discharge to free air.
There is no way that the discharge from one or more fans will be able to easily back spin a de-energized fan.
If the fans are ducted, both supply and exhaust ducts should be separate so that there will not be reverse flow through an idle fan either due to positive pressure on the discharge side of negative pressure on the supply side.

You seem to be stuck with a very bad design.
The consultants seem to be trying to justify a bad design to avoid accepting responsibility for a bad design.

Possible remedies, with credit to LittleInch;
1) set a greater dead band on the pressure switches for on/off operation of the second, third and fourth fans
2) Just install a fixed timer on the switchgear so that the fan runs for a min of say 5 minutes - gets you to max 12 starts/hour
THIS WILL ADDRESS THE SHORT CYCLING, BUT NOT THE BASE ISSUES.
Thanks again to LittleInch;
30 Fit a simple ratchet to the fan motors to prevent reverse movement
4) Fit solenoid or low force spring powered louvres to the 2,3 & 4 fans. I'm astonished that you're wasting airflow this way - why??
I would modify this and use a gravity closed damper. This is the cheapest solution.

Where do you go from here?
It is pretty easy for the consultants to refuse to accept free anonymous advice from the internet.
You need to get this from an autoritative source.
#1. Contact the motor manufacturer and describe the operating conditions of the motor.
Ask for written confirmation that either the warranty will cover this use, or that this use will void the warranty.
#2. Contact the contactor manufacturer.
Ask for written confirmation that either the warranty will cover this use, or that this use will void the warranty.

Please share this with the consultants:
What's wrong with these statements?

Quote:

it should take 30A for ~30 seconds to trip

Quote:

the motor is rated for locked rotor current for 23 seconds.

Note: The current profile and the motor and overload heating profiles are different during plugging as compared o a normal start.
Our observations were that the time to stop the motor was greater than the time to start a motor from standstill.
Both the I²t and the RMS heating were thus greater during a plugging event.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

It does seem to me to be a poor design. Exhaust louvers and timers seem the best solution.

Currently, the first stage fan will be mainly drawing air in reverse though the other fans rather than though the coil. That is why it runs 100 percent (wasting energy).

Thanks for the discussion on plugging duty- I couldn't find too many resources on this and have no experience with them. Our contactors are ABB AF09-30-10-13 which don't appear to be rated for IEC AC-4 utilization category. Could this explain why we've gotten trips despite not appearing to exceed the thermal trip current/time?

Quote (waross)

What's wrong with these statements?
Quote:
it should take 30A for ~30 seconds to trip
Quote:
the motor is rated for locked rotor current for 23 seconds.

Doh! I had it right in front of me and didn't notice those don't line up.

Quote (waross)

Note: The current profile and the motor and overload heating profiles are different during plugging as compared o a normal start.
Our observations were that the time to stop the motor was greater than the time to start a motor from standstill.
Both the I2t and the RMS heating were thus greater during a plugging event.

I did some back of the napkin calcs from a rotor loss formula I had in a textbook and approximated that ~2.5x more rotor heat was generated during the reverse start, so your experience supports that math. For my personal edification, does the deep bar effect play a factor in this due to being a function of slip? i.e. since slip > 1?

Our HVAC engineer met with the design firm and they recommended installing a refrigerant bypass line and spring valve to effectively derate the unit. This is an extensive change to plumbing, but our engineer is pushing for either that or for adding time delays to the fan starts and installing gravity closed damper on top of the unit to prevent backdraft. I'll post updates as we get changes in the field.

You've gotten advice from very knowledgeable people already. I haven't read through it careully. I'm just going to throw in a quick anecdote and some thoughts.

The anecdote. We have a variety of fans that when in standby (while the sister fan is running) rotate backwards due to leakage of the discharge "backdraft damper" (which to me means check valve). We have occasionally had both overload trips of the starter and instantaneous trips of the molded case breaker on these fans. I can understand the overload trips as related to prolongation of the starting current. I have a hard time understanding the instantaneous trips, but it has happened multiple times and does seem to be correlated to the positions that are observed to be rotating backwards in standby due to leaking backdraft daumpers....

So I am forced to abandon my simple mental model that the motor is drawing roughly LRC and developing LRT while rotating backwards. Instead I imagine that if I projected the current vs speed backwards past zero speed, it may well be the case that the current ends up slightly higher than locked rotor current. For that matter I imagine that if you project the torque vs speed curve backwards past zero speed you might end up with less than locked rotor torques (for some curve shapes).

Quote:

30A for ~30 seconds to trip versus the 30A for the ~5 seconds it takes for the fan to stop reverse rotation and accelerate to 1800rpm

On the basis of my previous above comments, I would question that on 2 more points (aside from the already-noted contradiction that you're going beyond 23 seconds).

1. Is it really LRC = 30A, or is it possible the motor is drawing slightly more than LRC while started rotating backwards?
2. On what basis was 5 seconds calculated. Did you use Locked rotor torque (it may be less). Did they consider fan loading torque? I recall that for an axial flow machine the bhp vs flow curve is highest at shutoff and decreases as flow increases (of course that curve has to be scaled down for lower speeds). So it stands to reason that if the fan is fighting against a reverse DP than the bhp may be even higher than we would calculate for starting against shutoff (which is a lot worse than starting with a flowpath available).

Quote:

I contend that the frequent starts (30/hr vs NEMA guidance of 19/hr), rotor heating presented by reverse-rotation starts, and stress on the thermal overload mechanism are causing trips and will impact long term performance.

It is true that this kind of duty can result in long term degradation of the rotor circuit. However since you said it started right away upon installation I think there's something more to it than rotor degradation (rotor degradation would tend to get worse over time)

I think you really need to stop the fans going backwards AND stop all that lost airflow.

I still can't believe they didn't fit louvres of any kind to a common air plenum.

The refrigerant bypass idea won't solve any of those issues which will just get worse.

HVAC seems to be full of cowboy firms who never last more than a few years, fold and then restart as a different name. IME.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

The back flow dampers, (gravity dampers) should fix most of the problem.
Apart from starting issues, you should have backflow dampers from an efficiency standpoint.
The air flowing back through the unused fans has already absorbed heat from the condensor coils.
This is mixed with the incoming cold air and the air through the condensing coil is warmer than it should be.

Quote:

Our HVAC engineer met with the design firm and they recommended installing a refrigerant bypass line and spring valve to effectively derate the unit.

It gets worse.
You should be suggesting a more experienced design firm.
Try the dampers first.
No operators, just simple gravity operated dampers.
The increase in efficiency may well solve the short cycling issue.
Your stages are coming in too soon because of the inefficiency of the back flowing air.
First things first.
STOP THAT DAMNED BACKFLOW.
Imagine a submersible pump in a well.
The discharge line has a leak and 50% of the water through the pump passes directly back into the well.
Not only do you have a similar inefficiency moving air though the condenser coils, but you have the added inefficiency of passing warmed and less effective air through the condenser coils.
FIRST THINGS FIRST
STOP THAT DAMNED BACKFLOW.
When more stages of fans come online the dynamics of the backflow change and the increase in condensing efficiency may be be significantly greater than the simple arithmetic of 2 fans plus 1 more fan may suggest.
Going from 3 fans to all 4 fans is probably increasing the condensing efficiency enough to cause the greatest part of the short cycling.

Solve the problem by reducing the rating of the unit?
Give you less capacity than you have paid for because the consultants do not understand the dynamics of air flow?

I am thinking about a truck where the throttle linkage becomes bent and the engine cabnnot develop full torque.
These guys would weld a stop across the transmission quadrant so that the driver could not use top gearand claim the the truck would now have enough torque.
First rule of trouble shooting.
Identify the problem.

Sorry for the rant but I have been called on too many trouble calls where the first two or three guys tried (unsuccessfully) to treat the symptoms.
This has all the marks of fixing the wrong thing, one more time.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

Something like this:
Back Draft Damper

The price is in Canadian Dollars, expect to pay less.
These should be on all fans.
The #1 fan will need one if it trips off for any reason.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

Reversing fan while it is still spinning in opposite direction takes much more current than LRC. Self closing dampers with contactors designed for multiple motor starts is the way to go.

Muthu
www.edison.co.in

You already covered why it's a screwed up installation in your description and analysis.

Personally, I would install a 15-20hp VFD and run the 4 motors in parallel all the time using PID control from a pressure sensor.

Lionel's tip seems like a simple solution without dampers, back flow and heavy duty contactors and more importantly not thermally stressing the motors.

Muthu
www.edison.co.in

More than one way to skin that cat.

Question now is who pays???

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

Quote (LionelHutz)

Personally, I would install a 15-20hp VFD and run the 4 motors in parallel all the time using PID control from a pressure sensor.

I would too, except for the cybersecurity requirements that come with installing a digital VFD in a nuclear power application. :(

Quote (LittleInch)

Question now is who pays???

Ideally the outside firm, but somehow we always manage to get stuck with the bill for the mistakes of others.

I've passed all of the advice in this thread to our HVAC engineer and his top preference is for dampers and a time delay to prevent excessive cycling during certain conditions. We'll see if he gets his way.

Dampers and a time delay is probably the simplest and cheapest retrofit alright. Not sure why your HVAC engineer is only now getting involved?

What sort of review of the design happened BEFORE it was purchased and installed? did it go through any sort of design review or HAZOP? If not why not?

Not stopping all that air flow from going in reverse is a pretty big miss IMHO.

You might find that the HVAC firm goes very quiet and then simply goes into liquidation once they work out that the extra costs are more than the money you're holding back on them. My experience is that HVAC seems to be full of such firms. But maybe I was unlucky. But it means any warranties etc are worthless.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

rc10
Thank you very much for the feed-back.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

> in a nuclear power application

I know some guys from that industry. Ever been to an EPRI Lemug meeting?

Quote (electricpete)

I know some guys from that industry. Ever been to an EPRI Lemug meeting?

Yep! Our company refused to pay for travel this year but I've been in the past and found them highly valuable.

Update- our site HVAC engineer lost the battle and the only thing that the project team agreed to do was to remove the magnetic trip from the breaker because it was "redundant" to the external magnetic trip mechanism (there were two for some reason after they made this change). I doubt this will reduce our nuisance trips but given that summer is over it'll be 9 months before we can test their "fix". Don't know what to say except this is one of the downsides to having design engineers as a corporate function and the knowledgeable site engineers too busy to be able to provide constant input to projects.