Brain fade time - measuring over time.

zagmund

I'm in the middle of a project involving monitoring my domestic electricity consumption. I am wired in (safely) to pulse counters on a couple of the meters I have set up an Arduino to do the counting, and am getting what look to be pretty accurate pulse counts. I know that there are 2,000 pulses per kWh from this specific meter.

I can see from the attached graph snippet that the base load (if that's the correct term) on one circuit is about 6 pulses per period measured, and the other is either 8 or 9 per period measured. I think the period measured in this case was 10 minutes. I can see occasions when someone had the toaster on and usage increased temporarily.

I've run the numbers for longer periods and I know that one circuit accounts for (say) 1.2kWh in a regular day and the other for (say) 1.6kWh in a regular day. Since I'm paying for kWh this is of interest for me.

The area where I'm a little confused (yes, I'm not an engineer) is that the script which is logging the data is also doing calculations based on the time interval between pulses. I *know* this is useful, but I just can't remember why. Can someone explain to me what the time between pulses is going to indicate to me that the count of pulses in a 10 minute interval isn't?

In other words, when the washing machine is on and we move from 10 pulses per 10 minutes to 1,750 pulses per 10 minutes . . . I can tell the demand is high. What do I learn from knowing that the interval was 0.2s instead of 15s on average?

Thanks,

z

ted1

zagmund wrote: »

I'm in the middle of a project involving monitoring my domestic electricity consumption. I am wired in (safely) to pulse counters on a couple of the meters I have set up an Arduino to do the counting, and am getting what look to be pretty accurate pulse counts. I know that there are 2,000 pulses per kWh from this specific meter.

I can see from the attached graph snippet that the base load (if that's the correct term) on one circuit is about 6 pulses per period measured, and the other is either 8 or 9 per period measured. I think the period measured in this case was 10 minutes. I can see occasions when someone had the toaster on and usage increased temporarily.

I've run the numbers for longer periods and I know that one circuit accounts for (say) 1.2kWh in a regular day and the other for (say) 1.6kWh in a regular day. Since I'm paying for kWh this is of interest for me.

The area where I'm a little confused (yes, I'm not an engineer) is that the script which is logging the data is also doing calculations based on the time interval between pulses. I *know* this is useful, but I just can't remember why. Can someone explain to me what the time between pulses is going to indicate to me that the count of pulses in a 10 minute interval isn't?

In other words, when the washing machine is on and we move from 10 pulses per 10 minutes to 1,750 pulses per 10 minutes . . . I can tell the demand is high. What do I learn from knowing that the interval was 0.2s instead of 15s on average?

Thanks,

z

Ignore the time. The pulse is giving you quantified units. Slower pulses just mean that you are using less electricity.

Bruthal

Well a pulse every 1.8 seconds means there is a 1kw load.

The shorter a time period you measure the pulses in, the more resolution you have in a graph of the loads throughout the day.

But if you simply count the number of pulses in an hour for example, you will have the average kW load for the hour, which = the kWh usage for that hour. But you wouldn't have recorded any upward or downward trends for that hourly period.

So the only extra info you have by measuring shorter time periods, is higher resolution of real time loads.

ted1

Bruthal wrote: »

Well a pulse every 1.8 seconds means there is a 1kw load.

The shorter a time period you measure the pulses in, the more resolution you have in a graph of the loads throughout the day.

But if you simply count the number of pulses in an hour for example, you will have the average kW load for the hour, which = the kWh usage for that hour. But you wouldn't have recorded any upward or downward trends for that hourly period.

So the only extra info you have by measuring shorter time periods, is higher resolution of real time loads.

I gathered he was talking about the time between pulses rather than the integration time.

Commercial loads are integrated over 15 mins as that's what electricity is charged in . 5 minutes or more should suffice for domestic. You don't want to be picking up heating elements etc kicking in and out.

zagmund

Yep, the time between pulses.

I'm currently monitoring solar temperature gains at a 3 minute interval, so I will probably shoot for that on the electricity also. I don't want to be too granular and pick up every little peak, but I do want to see some detail. I already have weekly granularity (with the pencil & paper route), so it's a question of how granular to go.

If I understand the time issue as originally mentioned, it would allow me to see the peak load, but since I'm not loading anything huge and I have a standard domestic supply which is never overloaded or trips I think I don't need to pursue that route.

Thanks for the info,

z

ted1

For three minute intervals just mulitple by 20 to get a KW value.

I till that 3 minutes is still to frequent, unlike a 4-20mA signal a pulse is a quantified signal and doesn't require such a short integration period