I want to build a solar park in Copenhagen. I want to maximize my profit and I have already decided to invest in fixed solar panels that can’t change angle. I also got a great deal for panel mounts fixed at a 45° tilt. But I am having a hard time choosing which orientation (azimuth) I should place my panels at.
Choosing the azimuth is interesting because it allows me to change when my solar panels produce electricity. Pointing the panels due south (180° azimuth) produces the most electricity. But electricity prices are typically higher in the morning and evening, so maybe I can increase my profit by angling my panels slightly east or slightly west.
Another factor to consider is that solar production in Denmark is increasing.
Solar parks drive down prices when the sun shines by flooding the market with cheap electricity. Maybe I can increase my profit by changing the panels’ azimuth to be different from the other solar panels on the grid, thereby producing electricity during times of less production competition.
I decided to simulate solar production with different panel angles using NREL’s PVWatts model. I used a Bayesian optimization to try different angles and a sequential domain reduction transformer to zero in on the best angle (check out my code here). Based on this analysis, the azimuth to maximize solar production in Copenhagen is 177.3° (South and a little bit east, farther from 180° than I expected).
But I don’t care about maximizing production, I care about maximizing profit! So I set the optimization’s target function to production * price. I used historical spot price data from Denmark’s DK2 bidding zone. When I ran it for the last 4 years this is what I got:
| Year | Optimal Azimuth For Profit |
| 2019 | 177.3° |
| 2020 | 176.8° |
| 2021 | 178.3° |
| 2022 | 178.2° |
There is a significant angle difference between the optimal azimuth for profit and the optimal azimuth for total production (as high as 1°). However, in 2022 I would only have made 0.003% more profit using the optimal angle for profit (178.2°) than I would have made using the optimal angle for total production (177.3°).
Even if one degree of azimuth difference doesn’t change profit significantly, it would still be good to get ahead of any trends in optimal azimuth. Unfortunately we can’t even do that because I don’t see much of a trend in optimal azimuth over the last four years. I guess I will stick with the trusty production optimized 177.3° azimuth angle!