As promised I’m sharing the results of our first full winter month of solar microgrid operations in case it will help others better understand how these systems work in real life and plan for the design/deployment of their own microgrids going forward. The chart for Feb 2020 is posted here:
The data was collected daily from the reporting systems provided by our new Tesla Powerwalls, Tesla Energy Gateway, and SolarEdge HD inverter. We certainly don’t plan to keep track of our system performance at such a granular level going forward, but the collection exercise was helpful for a new system by verifying that all components are working as expected, the data from different reporting sources match-up reasonably well, our data matches that shown on the monthly electric utility bill, etc.
Fortunately the outcome of the exercise was good news! The microgrid is performing quite well and so far has either met or exceeded expectations.
Here is some background info and our observations/comments on the February data:
- Our system includes a 10 kW solar generator with 30 panels and a single 10 kW inverter, along with roughly 40 kWh of energy storage provided by 3 Tesla Powerwalls.
- The electric utility company serving this site does not support net-metering. Instead they only support dual-metering, which restricts customers from sending ANY energy back to the grid unless done via a separate meter connected to a solar generator not connected to the home in any way. Very backward situation but something we had to deal with. Thus our solar microgrid had to be setup in a “zero-export” configuration.
- In this location, the normal Jan/Feb winter weather patterns create a worst-case scenario for solar microgrid performance. The sun is lower in the sky, the days are relatively short. Rarely did we see more than one day of full sun and full-strength solar generation per week. Mostly just clouds, rain, snow, yuck. We only saw about 5 partly sunny/cloudy days during the month. Mostly just clouds, rain, snow, yuck.
- Despite the unfavorable winter weather conditions we were pleased to see that our home was about 59% self-powered by the microgrid! Not bad at all. It was also interesting to see at least some solar generation in clouds and rain. Not so much when the panels were covered in snow… Seeing these results bodes very well for performance much closer to 100% self-powered during spring, summer and fall.
- Regarding the % self-powered measurements - while we saw a good match between the technical data provided by Tesla and SolarEdge with only a 1% difference in their metrics, it seems that a better way to determine a true % self-powered metric is to measure the actual cost/value of net energy provided by the microgrid vs. the cost we actually had to pay for grid energy as shown on our electric bill. This approach yielded the 59% self-powered number vs. the 47-48% number from the equipment vendors.
- For the most part the Tesla/SolarEdge data matched-up fairly well. We saw almost no difference on energy drawn from the grid, only a 3% difference on home consumption, and only a 4% difference on their self-powered calculations. For some reason we saw a larger 7% difference on the solar production numbers. Not a huge issue, but will be researching this further with our installers.
- Also very encouraging was the fact that our total grid consumption shown in the data matched the electric bill within about 5% - close enough.
- One of the challenges designing a system for our new construction project was trying to estimate the amount of energy storage (number of PWs) needed to get as close as possible to a fully self-powered configuration given the large number of low-solar days during the winter months in this location. During Jan/Feb we learned that after getting fully-charged during a decent solar day, our 3 PWs will run the home for 1-2 subsequent days of poor generation, with no change in consumption behavior. Looks like we'll be able to run even longer with some selective load shedding during grid failures.
- We noticed a few days with aberrations where the Tesla data varied substantially from the SolarEdge data - three days on solar production, once on energy imported from the grid, and once on home consumption. Will discuss the potential causes with our installer.
So there you have it - hope it’s helpful. I may post another month in the spring/summer timeframe to share how close we get 100% self-powered at those times.
Watch out utilities - here come the microgrids!