Questions about electronic demand meters

[hmspe]

Is there a typical scan rate for reading demand? For electronic meters (typically Elster here) is the demand measurement a true moving average? Are motor starting spikes filtered?

Backstory: A local talk show host pushes reducing the size of air conditioning units to save on electric bills because "the meter will read the starting surge" and "the demand charge for people using demand meters will be based on the maximum surge". There may be good reasons to have smaller units that run longer, but having significant savings on electrical demand charges as the only stated reason is more than a bit of a stretch. I didn't find anything online that would be suitable to forward to the host (ie., in plain English) that would help him understand how demand meters really work. Just sending a mathematical analysis showing how little a 2 second 6X surge changes the average running load for a typical 10 minute run time would be TL;DR. FWIW, they completely ignore comparing running loads, and there are a number of brands where a 5 ton unit actually draws less power than a 4 ton unit.

 

[jim dungar]

Starting currents, based on seconds will not impact a demand reading that is based on minutes. There is no reason for filtering.
It would be a very strange utility billing structure that bills off of anything other than an average demand.

The unspoken issue is that few, if any, residential customers even incur demand charges at all.

This is similar to the old wives tale that it took so much current to start a flourescent lamp, that had to burn for several minutes to justify the cost of turning them on.
Take an actual fact, twist it a bit by using slanted analogies, and then state the result in a sensational manner; 'like current takes the path of least resistance'.

 

[Ingenieur]

It is an average over a fixed time period (varies based on utility: 10, 15, 30, ? Minutes)
say 15 minutes 900 sec
ac load when running 5 kw
ac load when starting 25 kw for 3 sec

the avg demand is 3/900 x 25 + 897/900 x 5 = 5.067 kw
so yes, it factors in, but so short little impact

surprised they use residential demand meters
what is the interval?
what is the penalty schedule?

 

[Iron_Ben]

The "host" is wrong on most everything. The industry standard is a 5 minute or 15 minute interval, usually the latter. Our 15 minute interval used to be "block" (12:00 to 12:15, 12:15 to 12:30, 12:30 to 12:45, etc., and use the highest one of the 2880 or so in the month/billing period). About ten years ago we changed to a "rolling" 15 minute interval (bill on the max 15 minutes interval of the month; 12:06 to 12:21, 5:10 to 5:25, whatever), which was of course financially advantageous to us.

It looks to me like the sampling rate of modern electronic meters is once per cycle at the slowest and often much faster. So yes, they should capture all the starting currents. But as you note, a few seconds here and there on a 900 second interval is barely significant. And the inrush will be largely reactive to boot.

Finally, residential customers are seldom - if ever - billed on demand. Ours never were. Our new remote read meters recorded and displayed the demand, but it was informational only. Residential billing was strictly on kWh, plus the monthly fixed customer charge of $10 or so.

 

[Electric-Light]

The host's recommendation is generally correct but his reasons are not. A/C units have a warm-up time. Oversized units are likely to short cycle and efficiency suffers from this.