A Partnership of UConn and Eversource

Eversource Energy Center

eversource-txt-logo

Harnessing Data Analytics to Model Solar Energy Power Outputs

Fig. 1 Illustration of some results in this project. (a) Comparison results of the Silhouette values for different k in February. (b) Absolute values of correlation coefficients in January and meteorological weights in different months. (c) Comparison results of the clustering for different k in February. (d) Comparison results of the return levels and 95% error bounds for different k in February.

 

Contact Zefan Tang for additional information.

Extreme Photovoltaic Power Analytics (EPVA), a method to obtain high-fidelity information for estimating extreme power output, is Zefan Tang’s response to the dire need to enhance the power distribution infrastructure in the United States. Under the supervision of Professor Peng Zhang at ECE-UConn, PhD Student Tang is designing this powerful tool to provide the forecasting of extreme power output essential to the real-time integration of residential solar photovoltaic (PV) systems in Connecticut’s power grid.

As PVs have increased in popularity in the state, Tang has recognized the great importance of analyzing the impact of these energy sources on the grid. His method first automatically divides utility service territories into subcategories with similar behaviors of peak power to compute a probabilistic estimate of PV output in any region and at any time interval.

The biggest challenge Tang has encountered in developing EPVA is the lack of research associated with his project. According to the literature, multiple years of data are needed to support estimations of PV output in specific regions.  Since most solar panels in Connecticut are newly installed, however, the data are scarce. Another challenge is the influence of weather on PV output. Environmental conditions such as cloud cover make it difficult to analyze the power output for any particular space and time interval.

As Tang finds ways to overcome these challenges and advance the development of EPVA, the results and insights obtained through his method will offer valuable resources to both research communities and the power industry. He has, to date, finished an algorithm for the regional frequency analysis and produced results for the distribution of extreme power. He has also built software for easily using this tool and sent it to Eversource managers, who were pleased with the results. Tang’s next step is to submit his work for publication in a top journal, IEEE Transactions on Power Systems.

Published: July 24, 2018

Categories: Recent News

Available Archives

Eversource Energy Center | Innovation Partnership Building: 159 Discovery Drive, Unit 5276, Storrs, CT 06269-5276 | E-Mail: eversourceenergycenter@uconn.edu