ABSTRACT
One challenge in designing a photovoltaic (PV) system is to predict its generation, given parameters such as location, meteorological conditions, and layout. A greater challenge is to predict the generation of such a system under snow-cover condition. Publicly available snowfall data provide records for horizontal surfaces. However, the effect of snow accumulated on a tilted PV module remains unknown. Hence, irradiance is insufficient for predicting the output of PV systems having any given layout configuration. The research in this paper aims to predict the daily generation of PV systems through the development of a predictive model flexible enough to accommodate different layout configurations based on long-term monitoring data collected from 85 sites. Snow coverage loss factors are derived empirically to enhance the performance of the model. A feed-forward artificial neural network model is developed and implemented with snow adjustments (snowfall data and snow coverage loss factors). Promising results are obtained and validated.
Acknowledgement
This research has been conducted at the University of Alberta within the scope of a project entitled, ‘Integrating Solar PV Systems into Residential Buildings in Cold-climate Regions’.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Hadia Awad http://orcid.org/0000-0002-6452-3973