A novel uncertainty quantification method for electricity performance of building photovoltaic systems from multiple weather data sources

Distributed photovoltaic (PV) systems on buildings offer a promising solution for local renewable energy integration. As interest in sustainable energy grows, the demand for building PV systems is increasing. However, a significant challenge lies in accurately quantifying the uncertainty associated with building PV performance, particularly when considering multiple weather data sources. This study proposes a novel uncertainty quantification method to address this challenge. A case study of a building PV system with a PV wall and roof in Tianjin, China, is utilized to demonstrate the proposed method. Three types of weather data from different sources are considered: typical meteorological year data, multi-year historical data, and climate change projections. Uncertainty is quantified using a weighted-mean approach, incorporating weights based on recency, period, and accuracy of the data sources. Results indicate that the proposed method effectively integrates uncertainties from various data sources, providing reliable long-term energy yield estimates for PV systems. The variation coefficient of annual yield for the PV wall and roof system across ten typical meteorological year files is approximately 10 %. This study contributes to a comprehensive understanding of uncertainty in building PV performance, enabling informed decision-making in sustainable building design and energy management.