ESPRR: Expected Solar Performance and Ramp Rate tool

ESPRR is an interactive web application to model expected power and ramp rates for new solar PV plants in the Southwest while properly accounting for location, size, orientation, and geographic diversity. The tool consists of a Vue web frontend backed by a FastAPI backend to perform the calculations. ESPRR was initially developed with support from Salt River Project.

Development Setup

ESPRR has a few dependencies that are required. MySQL and Redis are used for storing data and queueing jobs respectively. It is easiest to run these in containers to avoid having to adjust versions on your local system. Node and Python are also required, and systems for managing Node and Python versions are shown below. It's advisable not to install ESPRR on your system python installation.

Non-library dependencies

• Docker (https://docs.docker.com/engine/install/) Docker allows us to run containers of MySQL and Redis locally and connect have ESPRR connect to them so we do not accidentally alter production data.

• DB Mate (https://github.com/amacneil/dbmate) Helps in setting up the MySQL database using migrations. Migrations can be found in the db directory. Migrations are a sequence of SQL scripts that are applied to set up a database. As the system is updated, we can add new migrations to update the database as necessary instead of having to start over from scratch.

• Node The Javascript runtime that allows us to run the dashboard. It is recommended to use NVM (https://github.com/nvm-sh/nvm) to install and manage Node versions. The dashboard is known to work on Node v15.8.

• Python Some Python interpretter. ESPRR is known to run on 3.9.2. It is recommended to use something like Anaconda/Miniconda (https://docs.conda.io/en/latest/miniconda.html#latest-miniconda-installer-links) or virtualenvs. This documentation will provide examples using conda.

MySQL and Redis container setup

Docker containers can be run or started using the esprr_docker_init.sh script. The script will create an esprr-mysql and esprr-redis container and export the DATABASE_URL environment variable required by dbmate. After the first run, these containers will persist on your machine. You can run the script again to start the containers if they have stopped. This script should be run using the command source esprr_docker_init.sh so that it can export the required environment variables.

Setting up the MySQL database

The first time you set up MySQL with the script from the previous section, you will need to set up the database migrations by running the dbmate up command from the root of this repository. This only needs to be done once unless you have deleted and recreated your MySQL containers.

Python installation and running a dev server

The api directory contains the python code for the API. Using conda, you can set up an environment,install the python dependencies, and run a development server with the commands below.

conda create -n esprr python=3.9.2

pip install -r api/requirements.txt

pip install -e api

uvicorn esprr_api.main:app --reload

The development server should start on port 8000. You can test that it is working by viewing the documentation served at http://127.0.0.1:8000/docs.

Node installation and running a dev server

The dashboard directory contains all of the javascript code for the front end. Using nvm, you can install and run a development server with the commands below.

nvm install 15.8

nvm use 15.8

# make sure you're in the dashboard directory
cd dashboard

npm install .

npm run serve -- --port 8080

The development server should be available at (http://127.0.0.1:8080)[http://127.0.0.1:8080]

Running a worker

A Redis Queue worker can be run from a python environment with the ERPRR API dependencies installed. Use the command rq worker jobs --with-scheduler to start a worker. Note that the worker will be unable to produce data until it has access to the NSRDB data discussed in the "Background Dataset" section.

Background Dataset

ESPRR uses the NSRDB dataset to produce modeled plant output. The reformat_nsrdb.py script is used to subset the raw data files to an area covering Arizona and convert them to the .zarr format. The script can be modified to output smaller files, but the raw NSRDB files are required as input and are approximately 2 TB in size each.

The rq workers should be pointed to a directory containing your processed .zarr files by setting the ESPRR_NSRDB_DATA_PATH environment variable to the path of the appropriate .zarr file.

NOTE A very small subset of NSRDB data with information around lat, lon 32.03, -110.9 can be found in the api/esprr_api/data directory. It can be decompressed using the command tar -xvf nsrdb.zarr.tar. The appropriate environment variable can then be set with export ESPRR_NSRDB_DATA_PATH="<repo_path>/esprr_api/data/nsrdb.zarr. This data is really only helpful in testing that the platform is functioning and will fail for sites not at the previously mentioned latitude and longitude.