Batteries - The First Year

Egyptian sun god Ra, via dreamlike.art

The batteries for the home solar setup were installed in late August of 2022, and I’ve had a full 12 months of data collecting and experiences to go over. And I’ve discovered a few interesting things.

Quick background

An important item on the checklist for electricity had always been to take advantage of the unique power options in Texas, specifically one of the many electric companies’ plans to save as much money as possible. Green Mountain Energy, my electric provider, had a deal where they would buy excess energy from my solar panels, and if the amount of export (excess power going to the grid) exceeded the import (power pulled from the grid), there would be a credit on the bill. This resulted in such a large credit that it covered not just the electric bill in the summer but through the winter as well. In 2022 the various electric companies in Texas implemented caps on export credit. It could not exceed import usage, it was essentially “capped”.

Therefore the large amount of credit I’ve built up starting in 2019 has been slowly eaten away until I will actually have to start paying an electric bill next month. Most months the export covers the import, but there is still a charge from the grid operator Oncor for the amount imported, and a small flat fee. I’m expecting the average bill to be in the $25-35 range for the most part once that credit is gone.

While one of the reasons to have the batteries is to keep electrical systems up and operating during a grid outage, my primary purpose is to provide a means to lower import of power from the grid, especially during traditional high usage times. The charging of the batteries starts after the sun rises, and there is enough energy to meet the electric demands of the house, so depending on the time of year this means the batteries start charging by 9am (earlier in the summer). A typical sunny day it will hit 100% battery charge by roughly mid-afternoon with excess being uploaded to the grid, and starting at 5pm if the house load is enough to start pulling from the grid it will pull from the batteries instead. This usually goes on until 10:30-11pm (even later in the summer) until 50% battery levels and then it’s back to grid power. The 50% level is maintained in case of an overnight extended power outage to get the house through the night until sunrise the next day.

Obviously if it’s a cloudy day this doesn’t hold true. Oh they will still generate power during daylight hours, just not nearly as much but most days the batteries get fully charged.

Reality

With the battery system came a new piece of hardware, the Sol-Ark 15K. This handles the batteries, monitors the grid, and monitors the load. While this device mainly functions as an inverter and can handle direct solar panel input, that is not how my system is configured. As detailed in other posts in the solar series, I have two groupings of solar panels so to the Sol-Ark it just sees this as “load”. Load is typically what the house is pulling in electricity-wise, but when the solar panels produce excess electricity over the demands of the house, this excess is exported to the grid, but simply shows up as negative load.

Now since I can see the levels involving grid usage in near real time, this allows me to better monitor what causes peaks in power usage. The peaks are interesting, and have led to a few changes over the past year, as well as a few purchases.

Outages

There have been a few outages, the longest lasting several hours. None of these outages were linked to major grid problems (like the 2021 winter storm) and were all tied to equipment failures and upgrades that would impact my immediate neighborhood. In fact the main reason I was aware of them was I’d get an alert on my phone that the grid had gone down - once I was on a Zoom call for work and only discovered this halfway through the meeting when I finally got around to checking my phone.

As all of my computer gear are on critical load circuits which are protected by the batteries, I also removed the old UPS systems. The Sol-Ark 15K can better handle a power surge and even EMPs, so I am well protected from various spikes and the UPSes are redundant and simply not needed. For more info on EMPs, read this previous blog post.

The only disadvantage is that if there is an outage that only lasts a few nanoseconds (these happen maybe once a month or so) I get an alert from the Sol-Ark informing me the grid is down. During its next polling cycle 5 minutes later it reports the grid is back up, but to give you an idea how short of an outage that is, remember that I have items such as my television and a number of lights that are on non-critical circuits that would get turned off in the advent of a power outage, and these don’t even flicker. I verified this behavior with the vendor, and all I can say is that this is the only disadvantage to the system I can think of, and considering the benefits I can certainly live with it.

Changes

For starters, I got an EV in April. It is a 2023 Chevrolet Bolt EUV, and I’ve installed a level 2 charger. I had to experiment with charging schedules, amperage levels, and monitoring the electrical usage to try to reach the point where I am not pulling down from the grid while charging. As it stands right now, I only charge on sunny days between 9:30am until 1pm at 8 to 16amps, which does not pull down power from the grid (8 amp) or very little (16 amps) as the solar panels typically cover this. Once I hit 80% (to help lengthen the Bolt’s battery life) I stop charging regardless of whether it is sunny or not, and will resume my sunny charging a few days later after it drops below 45% or so. As the charger goes all the way up to 48 amps, in a quick pinch I can get to 100% fairly quickly (like in case of a pending road trip). Be that as it may, it does slightly impact my ability to export to the grid. But if you include my savings from no longer spending money on gas for my old car, the Bolt overall costs less.

I’ve also noted peaks during other electrical usages in the house. The dryer is a huge spike, so as a rule I do laundry and try to time the dryer usage to peak sunshine to lessen the demand on the grid. The electric kettle is another one (I am a huge tea drinker), although not nearly as much draw.

The computers can draw a lot of power, averaging around 325 watts continuously. Yes that is a decent chunk of power, but there are a lot of computers, including multiple public servers that are up and operational 24x7. When I got my UniFi PDU which can monitor power usage to everything plugged into it, I could see that these tower servers used 4 to 6 times the amount of power as the Intel NUC-based servers, and I opted to migrate all servers to these smaller units. Various home automations via my Home Assistant setup have also helped with the load, as many lights are now on schedules which are optimized to ensure that they are only powered on when needed.

I had needed to redo the insulation in the attic. This was done in July. While it might seem like the benefit would help keep the house cooler in the summer, I expect the main benefit to happen this coming winter. Like most houses built in the southern part of the United States in the 1970s, there are gable vents as well as roof turbine vents so the slightest breeze clears out the attic of extremely hot air, making it easier to keep the house cool. For the south, this makes sense as the warmer temperatures last from spring to late fall, as opposed to the winter months. This setup does mean the furnace is running constantly in the winter as the attic is freezing and doesn’t help with the heating needs, so redoing the attic insulation should help with the energy demands for the winter.

The Numbers

Now on to the numbers. Exactly how much positive difference have the batteries made since acquisition? Here’s the breakdown.

As I have records going back a few years (again, solarpunk nerd), you can see that the previous year’s total of 8171 kWh - with solar panels mind you - has dropped to 4826 kWh. The raw price I am paying per kWh is $0.13, however when you add in the standard monthly fee of $9.95 for the power company and the fees (rated per kWh) for power pulled from the grid, this works out to roughly $0.20 per kWh. This means that the amount of savings per year with the batteries alone is roughly $1778.00, and doesn’t include the savings from solar.

Looking at the numbers above, note that the highest average daily usage has occurred in the winter months. Like I stated earlier, I am hoping that redoing the attic insulation will help with this. I am also considering a heat pump dryer instead of the normal dryer I currently have to help reduce energy consumption.

Summary

Pre-solar panels, I was paying roughly $2800.00 annually for electricity. The entire cost of solar panels and batteries is expected to have paid for itself by 2032, assuming I do nothing else to reduce energy usage. Just bear in mind my goal is not necessarily to save money, it is to reduce grid usage and my dependence on it. The savings is a side effect. My idea of reduction of grid usage also includes water savings, reduction in natural gas usage, more energy efficient household products as old ones age and new ones are purchased. So I am far from over. But considering the reduction of grid usage, grid dependence, and how it helps save me money - totally worth it.