Presentation Details
| Sensitivity of models for snow-induced energy losses in fixed-tilt PV systems due to different ground snow depth data sources (yes) Shelbie Wickett, Ayush Chutani, Isobel Bowker, Ana Dyreson. Michigan Technological University, Houghton, MI, USA |
Abstract
With the increase of utility-scale PV installations in snowy climates, modeling generation losses from snow cover blocking light to PV panels (snow loss) has become an area of interest for grid planners and PV developers. PV snow loss estimations require site-specific snow depth data, but data availability is limited, especially across wide geographic areas. However, ground snow depth has been modeled for hydrological research into gridded points using satellite data, numerical weather modeling, and in-situ measurements, resulting in more accessible snow depth data. Our research investigates the sensitivity of simulated snow-induced PV losses to two gridded snow depth datasets using industry-standard PV system simulation software. We compare the simulation snow losses from gridded snow depth inputs to simulated snow losses from site-measured snow depth data to investigate the impacts of snow depth dataset choice in PV snow loss modeling. During March and April 2023, the two gridded datasets resulted in opposite snow loss estimation behavior. When compared to the site-measured simulation, one dataset overestimated the snow loss duration by 41 hours and the total snow loss by 52.4 kWh-DC. The dataset introduced more false snow loss hours than it missed, with 49.1% of its snow loss hours not occurring in the site-measured simulation, while 31.7% of site-measured simulation snow loss hours were not captured. In contrast, the other gridded dataset underestimated the snow loss duration by 24 hours and total snow loss by 31.7 kWh-DC. It missed more snow loss hours than it falsely introduced, missing 43.3% of the site-measured simulation snow loss hours and falsely introducing 29.2% of its snow loss hours. For this dataset, the underestimation trend continued during the snow months of 2024 and 2025. Overall, these results highlight substantial differences in snow loss timing and magnitude depending on the snow depth dataset used, motivating further study of these differences across geographical regions and longer analysis periods.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.