Abstract

۱٫ Introduction

۲٫ Methods

۳٫ Results

۴٫ Conclusions

Code availability

CRediT authorship contribution statement

Declaration of competing interest

Acknowledgements

References

Abstract

     Urban sustainability is a significant factor in combating climate change. Replacing polluting by renewable energies is fundamental to reduce the emission of greenhouse gases. Photovoltaic (PV) facilities harnessing solar energy, and particularly self-consumption PV facilities, can be widely used in cities throughout most countries. Therefore, locating spaces where photovoltaic installations can be integrated into urban areas is essential to reduce climate change and improve urban sustainability. An open-source software (URSUS-PV) to aid decision-making regarding possible optimal locations for photovoltaic panel installations in cities is presented in this paper. URSUS-PV is the result of a data mining process, and it can extract the characteristics of the roofs (orientation, inclination, latitude, longitude, area) in the urban areas of interest. By combining this information with meteorological data and characteristics of the photovoltaic systems, the system can predict both the next-day hourly photovoltaic energy production and the long-term photovoltaic daily average energy production.

Introduction

     Cities have become a determining factor in climate change, as they are the place where much energy is consumed (64% of global primary energy use) and high levels of greenhouse gases emitted (70% of the global total), due to the use of fossil fuels as energy sources (International Energy Agency, 2016). There is a great opportunity for citizens to reduce these emissions. Replacing polluting energy sources by renewable energies that respect the environment and do not compromise future generations is one of the essential requirements to achieve energy-sustainable cities and favour the fight against climate change. Additionally, switching to renewable energy sources as a detriment to polluting energies could improve health and quality of life. Precisely, one of the goals proposed in the 2030 Agenda for Sustainable Development by UN is making cities inclusive, safe, resilient and sustainable (United Nations, 2015).

     Solar energy has seen a large increase among renewable energies. According to the International Energy Agency (IEA), there was a 22% growth up to 720 TWh (representing 3% of global electricity generation) in 2019 (International Energy Agency, 2020). Although large photovoltaic infrastructures are away from cities, there has been an exponential rise in distributed installations in buildings, industry and houses in Europe, the United States and Japan (International Energy Agency, 2020). This is very important since local production reduces transportation losses and enhance citizen’s responsibility because of inspiration for searching for energy self-sufficiency. In recent years, a new type of building, based on that type of installation, has been proposed as an evolution of Zero Energy Building (ZEB): the Positive Energy Building (PEB) (Magrini et al., 2020).

Conclusions

     An open-source tool, URSUS-PV, has been built to estimate potential electricity that can be generated in photovoltaic (PV) facilities in the short-term (one-day-ahead) and in the long-term (daily average) in an urban area of interest (neighbourhoods, streets, complex buildings). It could be potentially useful for multiple types of users, including municipalities, public administrations, companies in the photovoltaic sector, cooperatives or neighbourhood communities.

     One of the tools’ most significant benefits is the automation of the complex process. It initially had to be performed manually to obtain global results in urban areas of interest to produce short-term or long-term photovoltaic energy estimations. After conducting a data mining process, such manual processing has now been computerized. CRISP-DM methodology has supported the process. We have executed all the steps from the business understanding to the final deployment, which includes models discovered in intermediate phases. Therefore, following such methodology, once the data sources (meteorological and geographical data) have been identified, its acquisition and integration are automated. Transformations done in the preprocessing stage are also programmed, meaning that segmentation of the roofs available in urban areas of interest and extraction of their features can be easily computed. The calculating photovoltaic energy potential phase has also been fully automated using meteorological data of the city and the configuration of PV facility that could be integrated into each roof.