Energy costs after a year of operating Pololu's 305 kW solar project

This is the fourth post in a series documenting the installation and first year of operation of Pololu’s 305 kW solar array. Here are the previous posts:

• Part 1: background starting in late 2022 and how we committed to the $650,000 project by the beginning of January 2023, with a target completion date of May 31, 2023.
• Part 2: installation from January 2023 through first day of operation on October 5, 2023.
• Part 3: System failures and production results during the first year of operation.

In this post, we will look at how our solar installation affected our electricity bills. Because operations were so dramatically altered by the pandemic starting in 2020, we have to go back to 2019 for the best baseline for electricity consumption and cost. To maximize the scientific rigor of our observations, we changed as many variables as possible since then, including:

• Putting insulating foam in many of our windows. We blocked out almost all of our windows in the early weeks of the covid lockdowns to minimize our air conditioning load, and many of the pieces are still up, especially in unused offices.

  Putting foam insulation in all our windows, May 2020.

• Replacing the metal halide warehouse and parking lot lights with LED lights.

  400W metal halide parking lot light bulbs (left) replaced with 95W LED versions (right).

• Adding special solar-blocking film to our windows.

  Pololu window tinting, July 2022.

• Replacing over 800 fluorescent lights with LED lights. If they went from 100W each to 50W each, that’s a 40kW reduction (if they were all on).

  Ceiling light (and/or HVAC system) replacement.

• Replacing 26 air conditioners, most of them over 20 years old.

  HVAC system replacement during solar panel installation, June 2023.

• Optimizing our HVAC heating and cooling schedules. However, we kept the building slightly cooler in the summer of 2024 than in the summer of 2019, partly because it was possible thanks to all the above building improvements.
• Changing the number of employees. We had about 80 employees at the end of 2019; we averaged around 60 in 2024.
• Having the hottest summer ever. In July 2024, we had the highest-ever daytime temperature of 120 degrees (49C) and several days over 115 (46C).
• Changing electricity prices and pricing structures.

2019 baseline electricity cost and consumption

Our total electricity cost for 2019 was $83,142, with July and August being the highest months, at $12,570 and $12,487, respectively. Here are the details for August, which had a shorter billing period and therefore had the highest cost per day:

August 2019 NV Energy bill.

Note that during this time, we had several different electricity prices based on the time of day that we used the power. A large portion of the total bill is also based on the maximum rate of use (i.e. kW, not kWh) within each period, not just the total power used. The Demand Charge portion is for the maximum rate during the billing period; the Facility Charge is based on the maximum over a year (328 kW in this case).

Our total electricity consumption for 2019 was 684,384 kWh. So just based on the total use and cost for 2019, it seemed plausible for us to be able to use up the 520,000 kWh of estimated generation and for us to save $60k or more per year on electricity bills.

2022-23 baseline electricity cost and consumption

The closest 12 months of bills I can get prior to possible effects of the solar work were July 21, 2022 through July 21, 2023. Over that year, our total consumption was 428,318 kWh, and our electricity bills added up to $66,719. I think most of the reduced consumption relative to 2019 was due to our still reduced operations because of the pandemic effects and the solar film and insulation we had put on the windows (as opposed to most of the LED light and AC replacements, which would not have had much effect until part way through 2023).

NV Energy bills after solar installation

The first few weeks of operation with the new meter got billed on a partial billing cycle, so we did not get the first real NV Energy bill until November. Here are the eagerly awaited results:

November 2023 NV Energy bill, after the first full month of operating our solar generation system.

There are now four rows at the top, showing what NV Energy delivered (kWhD), what they received from us (kWhR), our accumulated energy (kWhA), and the net we used for the period (kWhN). Since we had so far used more than we put back into the grid, the accumulated energy is zero, and it would be zero until the spring, when we started generating more than we consumed.

Here is the NV Energy power consumption view for that period:

Daily power meter results from October 21 through November 20, 2023.

As I covered in the previous post, only 2/3 of the system was functioning until November 10, so if the whole system had been functioning, we probably would have covered that 7,906 net kWh that we got from NV Energy.

Here is the next interesting bill, for August 2024. This roughly matches the reference period from 2019, and it’s also the first time we had the whole system running for the whole billing period during the summer:

August 2024 NV Energy bill.

This month was also when the extra energy we had accumulated over the spring finally ran out: we had 5,545 kWh still banked going into the billing period, and our consumption was 8,074 kWh more than we gave back, so we were left with 2,529 kWh net energy to pay for.

Here is the last energy bill I will share here, and it’s the most recent one, roughly for October 2024. For September 2024, we just barely outproduced our consumption, so we had 158 banked kWh going into October. But more interestingly, the rates changed part way through the period.

October 2024 NV Energy bill.

You might have noticed the rates also changed between our first net metering bill a year ago and the August bill, with the price per kWh going down but the demand charge going up. The rates further decreased in the most recent bill, with no other increases. The decreases are mostly because the price of natural gas has come back down after a big spike in 2022 (the chart is from the US Energy Information Administration):

Natural gas price history.

Adding up the 12 months of bills ending with that last one from October 2024, we used a total of 38,206 kWh and paid a total of $17,607.

New wrinkle in the generated power estimate

In the previous post, I talked about being reliant on SolarEdge’s reporting of the generated power and estimated our actual system generation at about 535 kWh over the first year. That was based on this report on September 28, 2024, when we crossed 500 MWh generated:

We passed 500 MWh reported all-time production on September 28, 2024.

I added 25 MWh as an estimate for the period the system wasn’t reporting in June and another 10 MWh for the week from September 28 to the anniversary on October 5. But since I published that post last week, SolarEdge has changed their website, and the lifetime energy number displayed has been jumping around. Here is what it is showing on November 13, 2024:

SolarEdge website has been showing much lower lifetime energy result in early November 2024.

The SolarEdge monitoring site has been showing higher and lower lifetime numbers over the past week, but they have all been below 500 MWh. I used a separate report feature, and I put in the October 21, 2023 to October 21, 2024 range to match the NV Energy bills, and it gave me this:

SolarEdge net energy report for October 21, 2023 through October 21, 2024.

This report has not been fluctuating, and it seems to match other numbers (e.g. the monthly bar graphs), so until I get any new information that would undermine it, I will treat that as the best estimate. If I add the 25 MWh not reported in June to the 461 MWh in this new 1-year report, I get 486 MWh as my actual estimate for our first year. And doing the same 9/8 adjustment for all the down times the first year, my estimate for full production is now around 545 MWh (down from the 600 MWh I was estimating in that post). Adding the 38 MWh NV Energy says we used to the 486 MWh production estimate, we end up with 524 MWh used by our facility over the year. This means that had the whole system worked all year, we likely would have covered all of our electricity use for the whole year.

Here is a summary of our power use and electricity bills for 2019, the year preceding the solar installation, and the first year of operation:

That last column of dollars per kWh is just the total energy divided by the total we paid NV Energy, so this final year is of course very skewed by the base facility/demand charges and such.

So, how much did we actually save? As I think you’ll agree after seeing these details, it’s complicated. But I think it’s safe to say that we saved around $60k. Check back next week for the final post of the series, where I will discuss some complications from the tax incentives, reflect on some lessons learned, and finally try to answer the question of if this whole project has been worth it.