A tool to model cost and utilization of solar power systems. Uses power generation data from NREL's 2006 solar plant datasets. Inspired by this tweet (reference folder).

(optimization step count decreased for demo)

See possible improvements? File an issue! Feedback is welcome!

• Prerequisites
  • Python 3 (my specific version is 3.11.13)
  • Git
• Clone repository
  • git clone https://github.com/alecchendev/solar.git
  • cd solar
• Create virtual environment
  • python3 -m venv venv
  • source venv/bin/activate or whichever works for your shell, e.g. source venv/bin/activate.fish
• Install dependencies
  • python -m pip install -r requirements.txt
• Run script
  • python solar.py
  • Note: after first installing dependencies, the script may take a while to start. Subsequent runs should be quick.
• Run tests
  • python test.py
• Type check
  • mypy solar.py test.py
• Format/lint
  • ruff format
  • ruff check

Notes on specific commands

• For any commands that take a state name as input, you can provide the abbreviation or the full name in any case, e.g. "az", "AZ", "aRizONA"
• For commands where you compute the optimal power configuration (optimize and all), the default is to run for 41 different load costs. This will take roughly 9 minutes for a plant with 5 min intervals. You can change this by modifying the constant DEFAULT_LOAD_COSTS in the code (see To Do section for making this better).
• For creating the states_power_cost_per_energy_by_load_cost.png graph, currently the only way to do this is to run the all command on the states you want to compare (see To Do section for making this better).

• Prescribe a way to do things to simplify usage, but allow user control
• Be able to run one command for installation, one command to run the script
• Script should be able to go from nothing downloaded, all the way to the final results in one command
• Script should be able to take any intermediate input along the way and produce the next intermediate output
• Script should be easy to use as a library if imported, but also easy to copy paste chunks and modify, say in a jupyter notebook
• Script should be a single file (we're not doing anything super complicated here, everything should fit comfortably in one file)
• If a chunk takes time to process, we should print something to inform the user
• Make things simple

• Plots
  • Create cost per usage by load cost graph comparing locations by inputting different state files instead of needing to run the all command
  • Plant stats (e.g. histogram of capacities for a state)
  • Utilization for each day over a year
  • Power generation overlaying days for a full year
• Optimize - allow specifying load costs, or at least a test flag to only optimize over a single load cost
• Combine east/west dataset separations (Texas, New Mexico only)
• Code - delete StrEnums for column name constants?
• Delete mean_plant_for_state, just choose one deterministically
• Pre-compute optimal results/visuals and serve them from a website so people only need to run this script if they want to verify or customize
• Distribute as a python package so you can just pip install <name> and run
• Expand to consider international data from the NSRDB