Electric and Power

Prepping is more than just staying alive, it’s about thriving and preserving your life to the greatest extent possible. Many of the other things that we rely on for survival rely upon energy. We use it for a lot of things- heating, cooling, light, communications, all sorts of the things that we use rely upon energy. In some parts of the country, heat is important. Here in Florida, not so much. What we need is energy for cooking, light, air conditioning (summer heat will kill you more than our mild winters), communications, and other things. It’s on the prepping pyramid.

If you have read this blog for long, you know that two years ago, I added a backup power supply for the house. That supply consists of solar panels and a pair of Powerwall3’s to run the house when we aren’t making enough solar energy. One of the biggest limitations is battery capacity. Now that the tax refunds are no longer available for solar and battery installations, we need to make the ones we have last as long as possible. Adding more battery power in the form of another Powerwall would be great, but it’s a fairly expensive option. So we need to stretch what we have. That is, we need to match the loads being drawn to the energy provided, and do what we can to stretch battery charge to last as long as possible.

That’s where load shedding comes in. My early attempts at load shedding involved turning off breakers, and I even tried an Aquanta water heater switch, but it was nonfunctional garbage and I returned it.

I have a Synology server that’s already running Home Assistant, and this seemed like the way to go. I tasked Home Assistant with shedding loads. A home battery can keep the lights on during an outage, but its useful runtime depends heavily on what the house asks it to power. Instead of treating every circuit equally, this project gives Home Assistant a clear set of priorities: preserve essential comfort, shed discretionary loads in stages, and put daytime solar energy to work whenever it is available.

The system uses live Tesla Powerwall data and watches whether the house is connected to the grid or operating as an island. It also monitors battery charge, solar production, household demand, and real-time battery flow. The guiding principle is reliability: each Home Assistant automation has a focused job: important decisions require stable sensor readings, and loads are restored gradually rather than all at once.

At 80% battery, the upstairs air conditioner turns off. At 70%, the electric water heater is shed. At 50%, the main-floor air conditioner moves to a 80°F setpoint, and at 35% it shuts down. This preserves the 10 kWh of power for network and security services, lighting, and refrigeration. A critical warning is sent out as a push notification at 15%.

The most interesting part happens when the sun is shining during an outage. Loads are not restored merely because the battery percentage rises. The system first confirms sustained battery charging, then adds one load at a time with a stabilization period between each step. The main-floor HVAC returns first, initially at a conservative setting. Normal main-floor comfort follows when more energy is available, then the water heater, and finally the upstairs HVAC.

The 4,500-watt water heater stores hot water when a storm likely to cause a power outage is approaching. When the Tesla Powerwalls activate Storm Watch and the home is occupied, grid energy preheats the water in the tank even if the normal schedule would have it off. During an outage, strong solar surplus can also heat water instead of allowing energy to go unused. This function is controlled by a Aeotec smart switch that allows both switching and monitoring of the energy delivered to the water heater.

Near a full battery, Tesla may curtail rooftop solar because the Powerwalls cannot accept additional charge. That makes ordinary surplus measurements misleading—the panels may be capable of producing more, but they have been told not to. To capture that energy, the automation can briefly test the water heater at 95% charge. It keeps the heater running only if measured battery flow shows that solar can support it; otherwise, it shuts the heater down and waits before trying again.

Grid restoration receives the same cautious treatment. Utility power and normal Powerwall status must remain stable for 10 minutes. The main-floor HVAC returns first, the water heater is reconciled with its schedule, and the upstairs HVAC follows later. Staggering the sequence avoids a sudden surge and reduces the chance that unstable utility service will cause equipment to cycle repeatedly.

I also have a “vacation mode” that is initiated by a simple virtual switch on the HA dashboard. Activating that turns off the water heater and sets both HVAC units to 79 degF. That level allows the HVAC units to maintain a relative humidity of less than 55%, which is a good way to minimize mold growth.

We recently spent 5 days in Maine shutting down the BOL, and the house only averaged 31 kwh per day while we were gone. Normal power usage in the summer is about 70kwh per day. Since we generate about 47 kwh per day from solar, I think this is a doable system, but I will have to add some panels in the future so we can have a little more breathing room for cloudy days.

When we are in routine operation, the system turns the water heater off from 11pm to 5am, then Monday through Friday, again from 9am to 2pm.

The result of all of this is not simply a collection of “if battery is low, turn something off” rules. It is a small energy-management system built around priorities, measured power flow, and graceful recovery. In this way, I maximize the power and battery capacity I have available.

It’s been a bit since I updated the news on our Solar system. December went well. We generated 1121 kWh, and consumed 912 kWh, meaning that we sent an excess of 204 kWh back to the grid for credit. Our lowest power generated was December 6, with 12.6 kWh. Our best day was 54.1 kWh on December 2. That makes December the month where we generated the least amount of power of any month so far. Not surprising, since December has less daylight than any other month.

Temperatures for December were mild, so we spent most days with the windows open and the HVAC turned off, so we didn’t use a lot of power, which contributed to an excess. At the end of December, we had a total credit of 1121 kWh in case we ever use more than we generate, and it turns out that we needed it for January.

January was rough. There was a lot more cloudy days than there have been, and this month had the lowest average temperatures for any month in more than a decade. All of this contributed to our shortfall. There was a stretch of days from the 13th to the 23rd where 6 of those 10 days saw us generate less than 10kWh each day. We generated 1190 kWh for the month, still more than December, but we used 1238 kWh. Our highest use day was January 22 at 63 kWh. That was the day that the big snowstorm hit the state, and our temperatures dropped below freezing here, although it didn’t snow.

According to my weather station, the average temperature was only 54 degrees for the month. Since we maintain an average inside temperature of 70 degrees during the winter (68 at night, 72 during the day), the HVAC was warming the house against a 16 degree gradient.

Our bill didn’t change, thanks to our banked power that we have sold to the grid.

One small problem with our system has required a call to our installer, who put in a work ticket to Tesla. Our control system has been exercising the batteries by running them down to zero. It’s supposed to do that every 90 days, but it’s been doing it every 20 days. We are waiting for more word on why it is doing that.

I haven’t done a solar update since September. Our system for November generated 1378 kWh. We used 817 kWh and sent a total of 516 kWh back to the grid. Total for the system since August: we have sent 849 kWh more power back to the grid than we have used, with most of that being in November because of a combination of two things:

• We are making an average of 46 kWh per day because it’s been pretty sunny. There were only a couple of days where it was cloudy for the month, and we don’t get thunderstorms every afternoon like we do in the summer
• we just aren’t using much power now that the weather is cool enough for the air conditioning to not be running.

For 15 days in November, we generated more than 50 kWh, despite the fact that hours of daylight are getting shorter.

We only used power than we consumed for 5 days in November. The worst of these was a 48 hour period (November 6-7) where we used 18 kWh more than we generated. During those two days, we generated 20 and 25 kWh, but used 32 and 31 kWh.

So far, it appears as though the solar power system that we installed is working as well as I had hoped. Things do get a bit tight in the heat of the summer, as air conditioning uses so much power that, if the grid were to go down for an extended time, we might not have enough power to run everything without making some conservation moves when it’s a bit cloudy or if it’s unusually hot. Perhaps by turning off the water heater or by not doing laundry on days when it is cloudy or unusually hot.

The problem isn’t generation capacity- it appears to be storage. If I add even one more Powerwall to the system, this will correct the issue.

So for us, twenty four PV panels of 420 watts each gives us more than enough generation capacity. We currently have 27 kWh of Powerwall storage capacity, and could really use about 8 or 10 more, which would mean true independence and the capability of being off grin indefinitely.

As the hurricane passed by, we lost power for a little bit and thanks to the powerwalls, we weren’t even aware that it went out for a few minutes. The lights flickered, and we lost Internet. The Internet came back up within 30 seconds or so. I think it was the small blip of the batteries taking over that caused the router to reboot. I have to see about a small UPS that will fit in the QI panel. When the power went out, the app told me that we had a total of 15 hours of backup power available, and that was without taking steps to reduce consumption. I am doing research to see how to extend that time.

It does create a bit of a security concern, being one of the minority of houses with power while everyone else is without. We aren’t the only ones, though. A quarter of the houses within a half mile of the house have solar, but not all of them have battery backups, which means that they paid about half as much as we did for their systems, but they don’t work when the grid is down.

The month for power was good: we generated 1352 kWh and used 1272 kWh. That means we exported 80 kWh more power to the grid than we used, and that will be banked as a credit on our power bill for the next year.

Average power generated was 43.6 kWh per day, with a high of 58.0 kWh and a low of 18.3 kWh in any given day. Our power bill was the $30 minimum bill, plus the ever present taxes, for a total of $42. I view the entire bill as being a tax, since we are required by law to be hooked up to the grid, and the power company requires us to pay a minimum fee of $30. Without solar, our power bill would have been about $225.

Overall, I think it’s a good system and was a good buy. I have backup power and the system offsets my power bills.

It’s been two weeks since the hurricane passed by and screwed up my analysis of the new solar system. Let’s look at the numbers from the past two weeks’ production:

Overview

The house used 621.4 kwh during the period, or an average of 44 kwh per day. The lowest demand day was 36.2 kwh, and the highest was 60.9 kwh. The system generated a total of 663 kwh, which works out to an average of 47 kwh per day. The low was 36.6 kwh, the high was 57.6 kwh.

Overall, it appears as though my system is properly sized as far as generating capacity. How well will we do for the main intended use? That depends on storage capacity. As I said, I have been using the power company as a battery of sorts, in that I send power to the grid during the day, then take power from the grid at night. That begs the question: What if the grid is down? To answer that, we have to dive a bit deeper into the numbers.

A Closer Look

Our highest use days are when we are both home, and when it is hottest outside. This isn’t surprising, since the air conditioning runs more on hot days, especially when we are home. On top of that, the clothes dryer uses a HUGE amount of power, and laundry days are also heavy draw days (the dryer uses more power than both air conditioners combined).

The average solar intensity at my location for any 24 hour period is 200 watts/square meter for an average production of 47 kwh. The best day for production was 60.9 kwh, with an average solar intensity of 232 watts/ square meter. No surprises there- 16% more sun gets you 30% more power.

The good news here is that the hottest days also tend to be the sunniest days, so heavy use days caused by the cooling system are also the highest production days, with the result being the largest daily deficit was only 7.4 kwh. The largest surplus was 13.4 kwh.

What this means is that, should the grid go down, this surplus is likely wasted power. In our case, we would be wasting about 3 kwh per day. Saturday, August 10, and Sunday, August11 were our worst two days:

Saturday, August 10

39.9 kwh produced, 45.9 kwh used, 6.0 kwh deficit

The lowest outdoor temperature that day was 77 degF, with the high being 92 degF. Average solar intensity was 166 watts/square meter.

The air conditioner compressors ran for 6.5 hours

During the day on Saturday, we did 2 loads of laundry.

Sunday, August 11:

44.2 kwh produced, 51.6 kwh used, 7.4 kwh deficit

The lowest outdoor temperature that day was 77 degF, with the high being 93 degF. Average solar intensity was 190 watts/square meter.

The air conditioner compressors ran for 6.6 hours

During the day on Sunday, we did 3 loads of laundry. One of them was bedding, which takes more time and power to dry.

What about the batteries?

The day that saw us import the most from the grid was Sunday, August 18. We imported 25.9 kwh from the grid on that day, with most of that being the hours of 6 pm to 10 pm. This was while we were doing laundry.

My Analysis

We are producing plenty of power for our needs. We are keeping the batteries charged at 85% so that there is plenty of power available for a grid outage, and we are using the grid like an extra battery, storing our excess production during the day in the grid for use at night. If the grid were to go down, we would likely have enough battery power to run the house as normal, but we would likely have to restrict the activities that use the most power (laundry, cooking) for daylight hours. Additionally, there would have to be a restriction on the amount of laundry- no more than one load a day.

Other than that, we appear to have sufficient total capacity, and sufficient battery storage to run the household without interruption.

Interestingly, as a side note: The area near my house has a lot of solar power. Doing the area survey with my drone, I noted that 23% of the homes within a mile of my house have solar systems installed.

I was trying to get some solid data on the performance of my solar array, but we were interrupted by Hurricane Debbie. Still, negative data is still data.

• When it’s a typical Florida sunny day with sunshine until the arrival of our typical afternoon thunderstorms, my 10 kWDC array generates about 58 kwhAC per day.
• Two days before Debbie arrived, it was getting quite cloudy. We only generated 32 kwh.
• As the rains began to arrive on Sunday, our output dropped to 18 kwh.
• It rained all day on Monday, and we only generated 17 kwh. Still, generating that much power while simultaneously getting more than 3 inches of rain is a feat.

It’s sunny today, so we are back to full power. I want a week of solid, typical weather before I lay it all out.

We lost power once for about 15 minutes, in addition to a handful of blips, where the power was out for less than a second. The Powerwalls transferred us to battery power in about 1 or 2 seconds. Just long enough that we had to reset the clocks.

More data to come.

I previously recommended the use of a controller for your water heater as a way of controlling power consumption. Now that I have the solar up and running, the Tesla app is a great way to monitor and check up on your electrical loads.

Now that I have the ability to do this, I am convinced that the Aquanta water heater controller doesn’t actually DO anything. Whether I am home or not, using hot water or not, the amount of electrical power that the Aquanta is claiming I am using to heat water doesn’t change, at about 15kwh per day. The only time the amount of power the Aquanta was claiming went down is when we put the system in “Away” mode- telling it to turn off the hot water while we were out of town. At that point, the Aquanta reported that we were using about 1 kwh per day. The only problem? The amount of power we were using for the entire household didn’t change, according to the Tesla app.

At this point, I began suspecting that the Aquanta is a random number generator. It reports to you how much money it is saving you, but isn’t actually doing anything. I decided to test it and see what was going on. I am going to make a serious accusation here, so let me state that I did all voltage readings using a FLUKE 115 True RMS multimeter.

I measured the voltage across the output of the Aquanta while it was in “efficient” mode with a green LED showing and got a reading of 249 VAC. The water heater uses power consumed of 4500 watts (4.5kw), so a little application of Ohm’s law results in 18 amps and 14 Ohms of resistance. We can directly use Ohm’s law because the water heater has no capacitance or inductance, it is a purely resistive load.

Then I put it into “Away” mode. The LED on the unit turned red, so I measured the output voltage and got 70 VAC. Another application of Ohm’s law using the same 14 Ohms we calculated before tells us that the water heater is now using 5 amps and 50 watts of power. Take that 50 watts of power and multiply it by 24 hours in a day, and you get about 1 kWH per day.

So it doesn’t shut off the water heater, it merely lowers the voltage from line voltage down to about 70 VAC. I am not sure when or how it does this, other than in vacation mode, but I don’t think I am saving any power.

I just don’t think that this is a good unit to use if you want to save power, so I am withdrawing my recommendation of the Aquanta unit. My recommendation is to either use an electro-mechanical water heater timer, or a smart pool pump relay.

So we are four days into operation of the solar equipment. Right now, we are operating on a limited basis, more on that in a moment. I am not home because we are on the road, but both air conditioners, all vampire loads, and the water heater are running.

Output

We chose a system that is 10 kilowatts DC. That translates to about 8600 watts AC. The system begins generating electricity at about 9am, and reaches a peak of 8kw at around 10:30. By 1pm, the Powerwalls are fully recharged from the night before. At that point, there is no place else for the generated power to go, so system output drops to match whatever the house is using. More on that later.

We haven’t used any electricity from the grid since the sun came up on Friday morning. That indicates that I have enough panels for my house.

Storage

There are two places that I can store the power I generate. One of them is in the Powerwall, which has a total capacity of 27 KWh. Right now, I am maintaining a minimum of 30% as emergency backup, and using the rest to compensate for lower output at night or overcast conditions, or to make up for transient high loads, like when both air conditioners and the water heater are running at the same time. The advantages of using the battery are that the power stays within my home, and losing the grid means that I can still access it. The disadvantage is that the upfront cost of batteries is high.

The second place that I can store generated power is in the grid. The electric company buys my excess power in the form of credits that I can redeem when my system can’t keep up with the loads that I am placing on it- nighttime, stormy weather, or when loads simply exceed what I am producing. The advantage of this is that the upfront cost is low, but the disadvantage is that it relies on the electrical grid for redemption.

I can’t use the grid as storage because I don’t yet have permission to operate (PTO) from the power company. I should get it within two weeks after our final electrical inspection, which is supposed to be this week. So we should be fully operational by August 9.

Results So Far

Each day, we are generating between 35 and 45 kWh before panel output is reduced when the batteries are full. The solar energy being generated is directly running the house during the day, with the rest charging the Powerwalls, which run the house at night.

The water heater is using 4 kw when running, the upstairs AC is using 1.5kw, the downstairs AC uses 2.7kw, and the rest of the house uses 0.3kw. Since the ACs and water heater don’t run all of the time, the panels are more than capable at this point of keeping up by charging the batteries during the periods when the large appliances aren’t running.

Once we get our PTO, I will know more.

The solar power system was installed this week, We turned it on this morning. 24 panels, each capable of supplying 420 watts, for a total capacity of 10 kW. We can’t yet sell the power back to the power company, because they haven’t yet approved our application. Until then, we will run off of batteries and solar, with the excess being given to the power company free of charge. Hopefully, that will change within a week or two.

The install took two days, even though it was supposed to only take one. On day one, the team got the Powerwalls mounted, and 21 of the 24 panels on the roof before an incoming afternoon thunderstorm stopped work for the day. On day two, they got the final three panels up, ran all of the conduit and wiring, then shut power off to the house for about an hour so they could make all of the connections. They turned the system on, but that was at 1700, after it began raining again, so we didn’t generate any solar at all yesterday.

At 1000 this morning, we were generating 5 kW from solar while only using 0.5 kW, with the 4.5kW of excess going into the batteries. Our Powerwalls are already charged to 50% of capacity, and we have already generated 6.5 kWh purely from solar.

I will revisit the numbers within a week or so. It’s still to early to talk about how well the system is going to meet our needs.

All permits for the solar project have been issued and approved. Now we need to get on the installation schedule.