My electricity usage analysedday

[Elbow]

 

I wonder if this will be interesting to others.  I installed an EmonCMS system with some current sensors to get a proper understanding of my load and here's what I found:

We're three in the house with an ordinary 200L electric geyser.  It's short showers since we are in Cape Town. I have an old Intel server system that draws 230W or so or 5.5 kWh per day.  So that's pretty expensive to run.

Total daily consumption averages 28.2 kWh/day, or about 850 kWh/month.  The 250 that is at the high CTCC rate is obviously my first target.

Daily usage:

• Existing geyser consumes an average of 10.5kWh per day - 320/month
• My NAS server consumes 5.5kWh per day - 167/month
• The main oven an average of 1.4 kWh/day, very variable of course.
• The balance of 10.8 kWh/day goes to our convection microwave (used a lot), kettle, dishwasher, washing machine, my son's gaming computer, TV, lights etc.

Geyser takes 320 kWh/month - based on the proposed CTCC tariffs it costs me R753 / month.

I modeled my usage with the replacement of my existing geyser with a heatpump, which would probably reduce usage from 320 to 106.  Switching my NAS to a more efficient server could take that down to 70-ish.

Those two changes would cut my CTCC bill like so:

As-is cost based on new CTCC tariff
Daily charge:		250.00
Consumption <=600:	600	907.20
Consumption >600	263	617.69
Total:	863	1,774.89
All consumption from the grid, after heatpump and other savings:
Daily charge:		250.00
Consumption <=600:	600	907.20
Consumption >600	19	44.48	Saving:
Total:	619	1,201.68	573.21
Saving from as-is	28%	32%

So that's a substantial saving at a capital cost of say R25k for the heatpump and R10k spent on the NAS.  Perhaps the same sort of benefit by using a Heatbox or similar at a lower cost than the heatpump.

I then modelled the benefit of a 3kW PV system - with some effort to move demand to the daytime.

I got generation predictions for a 3kW solar PV system here in Cape Town (stats from that US Department of Energy "pvwatts" site).

On the demand side I took my detailed "instant" load figures for every 10 minute interval and figured out what shortfall would still have to be drawn from the grid.

Here's how generation compares to load for the period 2017-08-17 to 2017-08-29, scaled up to a month.  The consumption is after the efficiency changes described above.  I also added pool pump running during the day:

	PV (kWh/month)	Percentage	PV kWh/day	Consumption	Percentage	Usage kWh/day	Excess/-Shortfall
Breakfast	11	3.23%	0.36	41	6.56%	1.33	(30)
Daytime	307	89.59%	10.07	318	51.45%	10.46	(12)
Suppertime	25	7.18%	0.81	115	18.55%	3.77	(90)
Evening	0	0.00%	0.00	67	10.84%	2.20	(67)
Night	0	0.00%	0.00	78	12.61%	2.56	(78)
	342		11.24	619		20.32	(276)

"Breakfast" is 6am to 9am, "Suppertime" is 4pm-8pm, "Evening" is 8pm till 11pm.

You can see the usual challenge of excess generation in the day even though the heatpump and pool pump do use some of it up.

With all the shortfall bought from the grid it comes out like so:

Add 3kWp PV, "gross" shortfall from grid:
Daily charge:		250.00
Consumption:	346	522.66
Total:	346	772.66
Saving from as-is	60%	56%
Further saving from adding PV:	44%	36%	429.02

So 619kWh drawn reduced to 346.  This is for August.

If you could grid-tie on this tariff then doing so would give this result:

Feed excess generation back to grid, ie only buy nett requirement (not SSEG tariff):
Daily charge:		250.00
Consumption:	277	418.56
Total:	277	668.56
Saving from as-is	68%	62%
Further saving from back-feeding:	20%	13%	104.10

So that does make a further saving but not too much at this time of the year.

Using the same load figures against predicted generation for January comes out like so:

	PV (kWh/month)	Percentage	PV kWh/day	Consumption	Percentage	Usage kWh/day	Excess/-Shortfall
Breakfast	35	7.05%	1.14	41	6.56%	1.33	(6)
Daytime	384	77.94%	12.62	318	51.45%	10.46	66
Suppertime	74	15.01%	2.43	115	18.55%	3.77	(41)
Evening	0	0.00%	0.00	67	10.84%	2.20	(67)
Night	0	0.00%	0.00	78	12.61%	2.56	(78)
	493		16.19	619		20.32	(126)

 

As-is cost based on new CTCC tariff
Daily charge:		250.00
Consumption <=600:	600	907.20
Consumption >600	263	617.69
Total:	863	1,774.89
All consumption from the grid, after heatpump and other savings:
Daily charge:		250.00
Consumption <=600:	600	907.20
Consumption >600	19	44.48	Saving:
Total:	619	1,201.68	573.21
Saving from as-is	28%	32%
Add 3kWp PV, "gross" shortfall from grid:
Daily charge:		250.00
Consumption:	293	442.62
Total:	293	692.62
Saving from as-is	66%	61%
Further saving from adding PV:	53%	42%	509.06
Feed excess generation back to grid, ie only buy nett requirement (not SSEG tariff):
Daily charge:		250.00
Consumption:	126	190.95
Total:	126	440.95
Saving from as-is	85%	75%
Further saving from back-feeding:	57%	36%	251.66

 

I also did some modelling of the benefit of a battery pack.  I do see benefit but it doesn't seem like a good investment since it only helps to "move" daytime excess to evening. I know @plonkster

says that often.

Anyway - hope all that number crunching is interesting.  My data, YMMV etc etc.

[___]

It certainly is very interesting. Especially that you can move >30% of your consumption to daytime. That is impressive. I don't have a pool (and hence no pool pump), but one thing I always wanted to do is time the geyser to run after 10AM. It's not a difficult thing to do, just have to get around to it.

My latest math on the BlueNova batteries. They are around 10k per kwh, and they claim 7000 cycles to 70%, around 90% round trip efficiency. So 10000/(0.9*0.7 * 7000) = R2.26. That's more than the R1.93/kwh on lower end domestic, but slightly less than the R2.34 on the higher end (the >600kwh portion). There is a possible case to be made for a hybrid self-consumption setup.

There's two impediments still: Initial capital outlay, and I'm still a bit wary of the 7000 cycle number. As I said before, there are only so many people who make the cells (Winston and CALB), and though I won't pretend to know who uses what, it's interesting that the European battery makers don't quote such high numbers.

Early stats from Tesla drivers suggests that 3000 cycles is doable. But that's an NMC battery, not LiFePO4. One would expect slightly more out of the LFP.

I still don't know the reason for the discrepancy. It could be that the batteries are built with spare capacity, ie they are really around 20% bigger than advertised, so that it takes a little longer to get to the 20%-irreversibly-lost point, but at this point in time I simply don't know. At only half the cycle life, LFP makes no sense yet.

(But I do expect the price to halve in the next decade).

[Elbow]

4 hours ago, plonkster said:

It certainly is very interesting. Especially that you can move >30% of your consumption to daytime. That is impressive. I don't have a pool (and hence no pool pump), but one thing I always wanted to do is time the geyser to run after 10AM. It's not a difficult thing to do, just have to get around to it.

Hi @plonkster,

Perhaps I'm unrealistic - but what I did is took my actual load as logged by EmonCMS - 10 minutes by 10 minutes.  So that was my starting point.
I then manually "moved" dishwasher load to the daytime, and then added the pool pump for a few hours and then the heatpump - in my spreadsheet I ran the one after the other.

So I end up with quite high daytime usage specifically since the proposed heatpump runs in the day.

The actual implementation idea is to watch available power from the PV,  plus the water temperature and try to make a program that runs the heatpump at the best generation time but uses grid if I can't get the water hot by some deadline time.  My Infini will "supplement" my PV from the grid so it doesn't have to be exact.  I'd look for a consistent excess  ( PV watts - load watts > 600 ? ) from the PV for some minutes before turning the heat pump on.  If the heatpump is on then turn it off as soon as shortfall is too great - perhaps 1000W so that the algorithm has hysteresis.

Or, turn off if total load goes over some threshold that would stretch my inverter.  That should mean when the kettle is turned on the heatpump goes off.

If the water isn't warm enough by some deadline time we just turn on the heatpump and it will draw from the grid.

It seems like a heatpump heats water to 60 degrees by design, so I wondered about a further electric element in the storage tank and if I still have spare power just to continue to heat the water up as high as 90 degrees.  Its like a water battery except you can only draw out again in hot water.  I'd of course need a tempering valve to keep delivered water at a safe temperature and the plumbers may tell me why the idea is nuts?

Without the heatpump I guess the excess generation goes immediately to a traditional element in the geyser - I see that's been discussed  (but that will use the power much less efficiently), or back to the grid on the SSEG tariff, or into batteries.

SSEG tariff isn't that attractive.  Batteries are expensive and a lot of work, especially for a lot of capacity.  So this is the way my mind has been going.

Regards,
Elbow

 

[___]

5 hours ago, Elbow said:

Or, turn off if total load goes over some threshold that would stretch my inverter.  That should mean when the kettle is turned on the heatpump goes off.

I have the same idea. It's pretty typical to take a shower while doing a load of washing, and then pop the washing into the tumble dryer just as you get out of the shower. That way the washer heats water, the tumble dryer heats air, and the geyser heats water, all at the same time. On Sundays we often cook lunch as well. By dropping the geyser when the load is over 5kw and it is before noon, one can ensure that the peak comes down, and the flatter profile is easier to service with solar power.