Artificial Intelligence in the Solar PV value chain : Current applications and future prospects

Abstract: With the increase in computational power, tools and data generation, the use of AI is increasing in various sectors. Currently used methods in the solar photovoltaic industry related to solar irradiance forecasting, system optimization, solar tracking etc. has been found to deliver relatively inaccurate results. By using artificial intelligence to perform these tasks, a higher degree of accuracy and precision can be achieved and is now a highly interesting topic. This thesis will investigate how artificial intelligence will impact the solar photovoltaic value chain. The investigation consists of mapping the current available artificial intelligence technologies, identifying possible future uses of artificial intelligence and also quantifying cost and social impact. This was accomplished through literature reviews, survey-style interviews, cost analysis and a case study. As the technology is fairly immature, the results found are purely theoretical, but a decrease in levelized cost of electricity could be as high as 36%. Further, given more accuracy in forecasting, a more precise estimate of cost can be made which could attract more investors and increase the penetration of solar photovoltaic in the global energy mix. Although artificial intelligence is a powerful tool it still requires large amounts of data and in order to further drive this development, data sharing agreements between actors would have to be in place.