Generating Synthetic Time Series Photovoltaic Data with Real-World Physical Challenges and Noise for Use in Algorithm Test and Validation
- National Renewable Energy Laboratory (NREL), Golden, CO (United States)
The PV Fleet Data Initiative and other projects seek the develop algorithms for automated analysis of PV time series data for extraction of statistical information and other parameters of the data such as degradation rates, soiling loss information, tracker performance, clipping or curtailment, system availability and other valuable information. While there is a vast body of PV data available for application of said extraction algorithms it is difficult to validate these algorithms because the true parameters to be extracted are not known. There has been a wide use of synthetic data in the literature for algorithm validation but this synthetic data is typically very bounded by the problem or topic at hand. The PV Fleet Data Initiative project has demonstrated that real time series PV data almost always includes a host of data quality and physical problems that, in reality, any automated PV abstraction algorithm must handle appropriately. For this reason, this work describes the development of a complex synthetic PV times series data set that includes data quality and physical problems that have been experienced in real world PV data. The various quality and physical problems are documented in the synthetic data so that users can test the validity of various PV extraction algorithms as well as develop new algorithms to solve problems this data set can support.
- Research Organization:
- National Renewable Energy Laboratory (NREL), Golden, CO (United States)
- Sponsoring Organization:
- USDOE Office of Energy Efficiency and Renewable Energy (EERE), Renewable Power Office. Solar Energy Technologies Office
- DOE Contract Number:
- AC36-08GO28308
- OSTI ID:
- 1999775
- Report Number(s):
- NREL/TP--5K00-86459; MainId:87232; UUID:01b3f574-5622-46a7-93f0-1d4b9956d996; MainAdminID:70207
- Country of Publication:
- United States
- Language:
- English
Similar Records
Generating Synthetic Time Series PV Data with Real-World Physical Challenges and Noise for Use in Algorithm Test and Validation
PVAnalytics: A Python Package for Automated Processing of Solar Time Series Data
AI and ML Applications for PV Reliability and System Performance
Dataset
·
Tue Sep 05 00:00:00 EDT 2023
·
OSTI ID:1999772
PVAnalytics: A Python Package for Automated Processing of Solar Time Series Data
Conference
·
Wed Sep 07 00:00:00 EDT 2022
·
OSTI ID:1887283
AI and ML Applications for PV Reliability and System Performance
Conference
·
Thu Dec 14 23:00:00 EST 2023
·
OSTI ID:2274802