Two HVAC chillers with VFD failures and compressor failures 1

[jmbelectrical]

In a few days, I will be inspecting two HVAC chillers. They are arranged in an n+1 configuration in which only one chiller operates at a time. They are estimated to be approximately 15 years old. I do not know if the chillers are located outdoors or indoors. With respect to the VFDs, I do not know if they are controlling the compressors or the fans (If the chillers are air-cooled).

Within a three to four month window, each chiller experienced failures reported to be as follows:

Chiller #1:
-VFD pre-charge capacitors

Chiller #2:
-VFD SCR/IGBTs
-VFD gate driver
-One compressor (Failure was reportedly determined via insulation resistance testing of each compressor)

Repairs have already been performed. Aside from lightning, a surge, and/or end-of-life, what else could these failures be attributed to? Are there any other potential power quality issues, such as high THD, that I should be taking into consideration?

Any assistance is greatly appreciated.

 

[itsmoked]

As usual those error codes are fairly useless. For instance there's no such thing as "pre-charge capacitors". There are capacitors that get "pre" "charged".

That only happens on an initial power-on. So for that complaint to appear it likely means the failure happened on a power-up. Why would you ever power up a chiller? They tend to be turned on and stay that way until a power-failure or service work.

Gate driver and IGBTs toasting can certainly be caused by a shorted motor winding. You often get a mass of error messages because the standard failure mode is for the large monolithic power semiconductor module to crack like an egg which severs and mis-connects all sorts of sensor and driver control signals.

The death of ~15 year old VFDs I wouldn't spend much time on.

What I would do is check for any new loads added in the last few months to the facility. Any welders? Soft starters? BIG Lasers? Did the plant switch over to LED lighting? These are all things that can put VFDs into an electrical tizzy.

I'd also confirm that both those VFDs have correctly sized 3% source reactors in series with them. Those reactors will protect VFDs from a lot of network mayhem and conversely will protect the plant network from a lot of the ugly distortion power factor caused by the drives.


Keith Cress
kcress -

http://www.flaminsystems.com

 

[jmbelectrical]

Keith,

Thank you for the response. It was extremely helpful.

Do you happen to know of any whitepapers or other authoritative literature that addresses the average lifespan of a modern VFD? Everything I've found so far has been anecdotal.

Thanks again.

 

[itsmoked]

Hi jmb..; Sorry I don't have any VFD lifetime data but jraef will likely come by and whatever he sez in that regard is The Facts since his focus for decades has been VFD support.

I've kinda heard ~seven years is the expected lifetime. I've recently replaced several 20 year old drives that died in a plant fire so I know they can last a lot longer than 7 years. The only drives I've ever seen that died before-their-time were killed by the installer (don't ask me how I know) and by failed motors.

What takes them out is:
1) Conductive dirt+moisture (pollution)
2) User interface buttons/display failure
3) End-of-life of the DC bus capacitors
4) Failed motors
5) Malevolent network voltage excursions EMI

Keith Cress
kcress -

http://www.flaminsystems.com

 

[jraef]

Keith hit most of the highlights.

The thing I will add on VFD #1 is that what can fail in the pre-charge circuit are the pre-charge RESISTORS. The pre-charge circuit is there to prevent the capacitors from damaging themselves and the rectifier components by trying to pull too much current when initially energized. So a pre-charge current limiting resistor is placed in series with the capacitors for a second or two when power is first applied, then the resistor is bypassed with a relay (or contactor depending on the size) so that it doesn't burn out by being in the circuit all of the time the drive is running.

What can happen is that, if the drive has power cycled to it too often (duty cycle wise), or too many times (life cycle wise), the resistor will fail and that then either causes the drive to go into a fault to stop you from energizing it again and causing the capacitors to blow. In most drives, the pre-charge resistor and circuit is designed to last about 1,000 operations. If the VFD is powered down and back up once per month, that's 83 years. But if it is powered down every day, that's less than 3 years. So I would be investigating procedures and/or the control system to see if there is something cycling power to these drives. In some retrofit cases, people leave an old motor starter in place ahead of the VFD because that is already tied to the Building Management System and they want to leave it that way. But what they don't know is that will shorten the life of the VFDs.

On VFD #2, I would guess that the motor started to fail first, then that took out the VFD transistors, which then takes out the gate firing board and maybe the rectifier. I would be checking to make sure the motor was rated to be used with a VFD and if not, change that or add at least a DV/DT filter to the output. If the motor was truly "Inverter Duty" (more on that below), then also check the wiring. One set of wires per conduit from the VFD to the motor, ground WIRE in the conduit, not using the conduit as the ground path, if it is anything other than STEEL conduit, then shielded VFD cable, etc. etc.

PS regarding "inverter Duty" motors; There is no real "official" definition of what that means, so some motor suppliers play a little fast and loose with what they say in that regard. You want to look for them stating that the motor meets NEMA MG-1 Part 31 specifications for motors run from inverters (or similar IEC specifications that I can't recall at the moment). If they make other unsubstantiated claims, it likely not true. For example a VERY well known motor mfr is now marketing a cheap junk line of motors to equipment OEMs that they call "inverter ready". Those motors have a very high failure rate when connected to VFDs, but they usually outlast the 1 year warranty that the OEM provides on their equipment.


" We are all here on earth to help others; what on earth the others are here for I don't know." -- W. H. Auden

 

[FacEngrPE]

OP indicated the failure was in a chiller compressor motor. Usually these motors are integral to the compressor, and are cooled by the returning cool refrigerant. While there are a few made for use with VFD's it is unusual to find a label on the units stating that they are VFD compatible.

https://vrfwizard.com/vrf-compressor-types/[/URL]]The Danfoss inverter compressor solution for commercial HVAC applications allows a VRF system to modulate the cooling capacity precisely between 25% and 100% when a single compressor is used, and between 12.5% and 100% with tandem compressors (Modulation range for VZH compressors). The unit adjusts to the varying loads that the building needs during the day.