37th EU PVSEC, 07 - 11 September 2020

27 June 2019

EU PVSEC 2019: Preview of Highlights. Topic 5 “PV Systems”

Before our Early Bird Offer ends on Sunday, let us continue to present the highlights and best moments at EU PVSEC – today’s preview looks at presentations of Topic 5 “PV Systems”.

Under this title a wide range of subtopics will be presented such as designing, modelling and operation of PV systems as well as their performance and maintenance. There are also sub-topics on forecasting and system failure studies but this year a special focus is given to storage of PV energy. As costs of PV electricity continue to fall, storage gains in importance which not only allows for smoother grid integration but also for modelling and designing grids with a very high share of PV electricity, including seasonal storage. Any 100% renewable energy scenario depends on efficient storage technology.

Our chairperson, Dr Florence LAMBERT of CEA Liten will deliver a keynote presentation of the PV system topic giving a comprehensive overview of the “Current status of battery energy storage” (5DP.2.1). Dr Lambert will certainly explain several technologies in the race for electricity storage. One for instance is featured in the oral presentation titled “Demonstration of a Novel HBr-Flow Battery for Grid Integration of PV” (5CO.14.3), where you will learn about the advantages of a commercially available battery system with a chemistry beyond the mainstream lithium-ion flavour.

Based on long and extensive experience with PV Systems and availability of powerful computing resources, modelling of performance of PV systems has become ever more reliable and sophisticated. This increases confidence of both investors and utility providers. But what is the best software modelling in use today? This question will be answered by the presentation “Cross-Validation of PV System Simulation Software” (5CO.13.6). The authors compare four software packages in use today, and simulate different systems in various climatic conditions, technologies and design variations (e.g. fixed tilt vs. tracking). Preliminary results show consistency regarding technologies and design. However, as one can expect it also depends on the weather files which generate source data. The presentation reveals in any case the different approaches and algorithms of the simulation tools.

Regarding PV system simulation, many papers in this conference describe the efforts to understand how much more energy production can be expected from bifacial module. The Plenary keynote presentation “Next Generation Tools for Accurate Energy Yield Estimation of Bifacial PV Systems” (5DP.2.2) is one of the many, but certainly the most comprehensive regarding “Best Practices, Improvements and Challenges”. For instance, the authors conclude in their literature review on bifacial modelling that the gains reported cover a range from between 4.3% to 40%. This presentation shall help you to better assess realistic expectations in bifacial modules and the conditions required for a significant energy gain.

On performance of PV systems, one highlight will be the keynote plenary presentation “Performance Optimization through Advanced Data Analytics - Practical Applications Covering More Than 2GWp in Europe and India” (5DP.2.3). To some of you “analytics” may sound like a trendy buzzword, but the authors demonstrate convincingly how data-analytics can yield results to improve PV systems. The presentation features practical examples to show the importance of correct data analytics. This contributes to understand what is happening with the PV system and enables conclusions for improvements. The data analytics has been based on and tested with PV systems totalling 2000 MW!

Another approach of analysing data is presented by one of our student award finalists in the talk “Automatic Fault Detection of Photovoltaic Array by Convolutional Neural Networks during Aerial Infrared Thermography” (5BO.6.4). The screening of large PV arrays using UAVs (better known as drones) is today's standard technique for detection of faults like bad cell contacts or cracks. However, manual screening of thousands of images may not be for everyone and this is how this student’s work is motivated to describe the algorithms for an automatic analysis of these images. This software similar to a “face recognition” tool is an example for one of the many presentations at EU PVSEC addressing topics such as recent data algorithms, machine learning, artificial intelligence or neural networks to forecast, design, model and analyse PV systems.

Another Student Awards Finalist Presentation delivers research on the “Simulation of Performance Differences between Off-Shore and Land-Based Photovoltaic Systems” (5DO.3.6) comparing “Floating PV” systems with ground-based systems. For the off-shore system the student takes into account not only a (lower) temperature profile due to the close proximity to water, but also varying tilt angles due to the ocean waves!

There are more papers as part of Topic 5 dealing with forecasting along different timescales. A novelty will be shown in the presentation on “Nowcasting of Irradiance Using a Network of All-Sky-Imagers” (5DO.2.1). It describes the approach of using sky-images for recognising cloud patterns that will affect the PV system output within 15 minutes.

We hope that this overview made you curious to listen to more visual, oral and plenary sessions of Topic 5. As we have already announced earlier this week, we will arrange several technical tours for conference delegates, exhibitors and visitors. You have the choice between these 4 highly interesting sites showcasing the efforts France made to progress on its way to generate carbon free energy.