Assessing Solar PV System Performance at Utrecht Science Park

Abstract

This research examines the performance of the Utrecht University Science Park (USP) solar PV system over a five-year period (2018-2022). USP system's energy output consistently surpassed the target of 1 million kWh annually, reaffirming its effectiveness in sustainable energy generation. KNN imputation was used to address data gaps, enhancing analysis accuracy. Assessment of performance metrics, encompassing energy output, final yield, reference yield, and performance ratio, illuminated system efficiency and variability. South-oriented, 30-degree tilt subsystems outperformed, while W-E-S orientation yielded the least. Total in-plane irradiation yield was modeled with Python pvlib and Soalrgis Prospect Model, with the former generated higher reference yields due to its clear sky modeling, while Solargis introduced more nuanced predictions based on historical weather patterns. Both pvlib and System Advisor Model (SAM) were used to simulate expected energy production, with pvlib’s results significantly higher than SAM; Energy Performance Index (EDI) is above 1 for the 2018 and 2019, with SAM’s results, and below 1 during 2020 and 2022 with pvlib’s results. These differences emphasize the importance of considering real-world conditions, including shading and atmospheric effects, to refine performance assessments. Regression analysis highlighted panel orientation and tilt angle's significant impact on final yield. Although limitations exist, future research should incorporate on-site measurements and advanced modeling techniques to account for local weather conditions' influence on system performance.