Solar-TK: A Data-driven Toolkit for Solar Performance Modeling and Forecasting
Solar-TK is a data-driven toolkit for solar performance modeling and forecasting that is simple, extensible,and publicly accessible. Solar-TK’s simple approach models and forecasts a site’s solar output given only its location and a small amount of historical generation data. Solar- TK’s extensible design includes a small collection of independent modules that connect together to implement basic modeling and forecasting, while also enabling users to implement new energy analytics.
A key goal of Solar-TK is to be simple to use by researchers that require realistic and accurate solar performance models and forecasts, but are not experts in these areas.
Documentation
Why do we need a toolkit for solar performance modeling and forecasting?
We quote our Solar-TK paper explaining the need for a toolkit:
Much of the research work on solar performance modeling and forecasting is not accessible to researchers outside the area, either because it has not been implemented and released as open source, is too complex and time-consuming to re-implement, or requires access to proprietary data.
What does Solar-TK offer?
To address the problem, we present Solar-TK, an open data-driven toolkit for solar performance modeling and forecasting that is simple, extensible, and publicly accessible. A key goal of Solar-TK is to be simple to use by researchers that require realistic and accurate solar performance models and forecasts, but are not experts in these areas. Solar-TK includes:
- a dataset of solar power generation and energy consumption across hudreds of homes
- a module to estimate the physical specifications of a solar site, i.e. capacity, tilt, orientation
- a module to estimate the maximum generation potential of a solar site
- a module to estimate the weather-adjusted generation for a solar site
- a module to adjust the estimated output for shading by nearby buildings and trees (will be added in future)
- a module providing metrics used in solar modeling and forecasting (will be added in future)
Publications
Find our Solar-TK paper. Please consider citing our paper if you use Solar-TK in an academic work.
The BibTex citation is given below.
@inproceedings{bashir2019solar,
title={Solar-TK: A Data-driven Toolkit for Solar PV Performance Modeling and Forecasting},
author={Bashir, Noman and Chen, Dong and Irwin, David and Shenoy, Prashant},
booktitle={Proceedings of the 16th International Conference on Mobile Ad-hoc and Smart Systems (MASS’19)},
year={2019}
}
Solar-TK combines a number of insights from prior research listed below.
Staring at the Sun: A Physical Black-box Solar Performance Model
Dong Chen and David Irwin
BuildSys 2018
Link: http://www.ecs.umass.edu/~irwin/staring.pdf
SunDance: Black-box Behind-the-Meter Solar Disaggregation
Dong Chen and David Irwin
e-Energy 2017
Link: http://www.ecs.umass.edu/~irwin/e-energy17.pdf
Black-box Solar Performance Modeling: Comparing Physical, Machine Learning, and Hybrid Approaches
Dong Chen and David Irwin
Greenmetrics 2017
Link: http://www.ecs.umass.edu/~irwin/greenmetrics17.pdf
Please note that Solar-TK will evolve based on community feedback! Please use the online docs instead of the paper.
Solar-TK Help and Resources
- The users are encouraged to use stackoverflow with #solar-tk to get community help.
- You can also email to Noman Bashir at nbashir@umass.edu for help. Please be patient as it may take a few hours to a day to respond.