Dear Forum Members,

We are in the process of stydying various options to impement arcflash technologies for our switchgear. Please clarify on the following points,

1. Can the light sensor based arcflash detectin scheme be implemented in MCC

2. As of now no one has used this technology in india, hence we are skeptical about the reliability of this method.

Please share you valuable information on this.

You should be able to use a light sensor in any place where it can detect an arcing condition fast enough to be able to trip the upstream protective device.

I feel this is the major problem of these sensors, as who can empirically predict where an arc will occur within a piece of equipment?

Installation of light sensors on MCC buckets is possible and has been done. We installed a system on LV MCCs and it operated properly.

The recommendation though is to have light plus current.

Since the MCC buckets are designed for operator use (opening the doors, etc), you want to monitor the current as well.

This leads to the configuration of the MCC incoming section - whether it is MLO or if you have an incoming breaker, etc.

How do you know it worked properly? Did you create different arcs for testing?

This maybe the primary reason I am leary of field installed 'light sensors'.

Even with high speed light detectors devices you still have the deal with mechanical opening time of the PD device... in LVPB this can be up to 6 cycles or more.

Testing optical sensors

I use a camera strobe to test. A medium sized independent strobe. Built in camera strobes may be too weak. A vendor told me a laser pointer works too, mine does not work, must be too low power.

Same here, I use the flash from my old Maxium 9000

Normally, the relay will not operate unless there is fault current at the time of the flash. This is to prevent tripping when people take photos.

I think this is a fairly limited test. Yes, it does show if the sensor can react to the light from a camera flash unit, but it does nothing to test the sensor's ability to see an arc flash inside of the equipment. This is similar to using a jumper from 'hot to ground' to test a ground fault system.

Is the light output from your flash calibrated to be equal to some amount of incident energy? Can it be used to locate "blind spots" inside of the equipment (i.e. sensor above breaker, arc underneath breaker)?

Testing with flash

You are correct the test is mostly qualitative. I have used the flash with 2 brands of optical arc flash relay. What the camera flash does show is that the relays react at a variety of light incident angles and distances, in a given indoor ambient light situation.

So far, only bench testing in my experience. Actual installations will come soon. Will test the installed version with a camera flash. That will be qualitative with some quantitative distance and angle measurements.

Suggestions for testing, verifying optical sensors?

Note, the optical sensing relays are not suitable for areas where the sensors will be exposed to direct sunlight.

Reflections

Testing by camera flash is OK.

I find questions of blinds spots or obstacles somewhat irrelevant, because in case of an arc fault there is so much light and reflections inside of the equipment that the sensors will activate.

I have examined some switchgear arc test videos frame by frame, and even when the doors stay closed, there is so muct light OUTSIDE of the equipment that the camera is blinded. Not to say anything about the light intensity inside.

If low power arc faults, possibly phase-to-earth faults with restricted current, are an issue, then the location of the light sensors/fiber may be an issue.