novunex-logo

PVSwarm – Self-Adaptive Cross-Agent Intelligence for Photovoltaic Systems
Demo

PVSwarm – Self-Adaptive Cross-Agent Intelligence for Photovoltaic Systems

As the construction and use of photovoltaic (PV) systems grow due to environmental concerns and energy crises, effective monitoring of their efficiency becomes increasingly crucial. Light Induced Degradation and soiling can significantly reduce a PV system’s efficiency, prompting the need for sophisticated control mechanisms. PVSwarm proposes an innovative solution that integrates machine learning and swarm intelligence to model PV plants as self-organizing systems. By simulating interactions among various agents, including both physical components and external factors, PVSwarm aims to create a self-adaptive system capable of precise monitoring and maintenance, ultimately leading to improved energy yield and aiding in energy independence.
Partner
Lakeside Labs (Coordinator)
Pasquale Grippa, Lakeside Labs B04b, 9020 Klagenfurt
https://www.lakeside-labs.com/

Kärnten Solar (W.I.R. Sonnen Contracting)
Villacher Straße 222, 9020 Klagenfurt am Wörthersee
https://www.kaernten-solar.at/

JOANNEUM RESEARCH Forschungsgesellschaft mbH
Leonhardstraße 59, A-8010 Graz
https://www.joanneum.at/en/

CAMPUS 02 Fachhochschule der Wirtschaft GmbH
Körblergasse 126, 8010 Graz
https://www.campus02.at/

Novunex GmbH
RTD, Völkermarkter Ring 7/1, 9020 Klagenfurt
Improving PV Energy Production: Navigating Challenges with Advanced Monitoring and AI Solutions
The expansion in photovoltaic (PV) energy production and the development of PV installations highlight an active response to the ongoing energy crisis and a commitment to environmental sustainability. This growth necessitates an improvement in the level of monitoring for these systems, as increased solar plant capacities have not corresponded with advancements in surveillance technologies. Presently, most PV systems undergo basic monitoring, insufficient for identifying and addressing issues swiftly. The efficiency of PV systems is compromised by several factors, notably Light Induced Degradation (LID), which can reduce efficiency by approximately 0.5% to 1.5%, and soiling losses from snow, ice, or debris, impacting efficiency by about 1.5% to 6.2% as indicated in [KIMBER2006]. These inefficiencies result in reduced energy output and financial losses. The detection of anomalies within PV systems remains complicated due to the need to consider various factors, including temperature changes, solar irradiance, hardware degradation, and component defects. Establishing a detailed monitoring system encompasses modeling every contributing component and their interactions, a feat currently unachievable with existing technology. However, experience suggests that machine learning and algorithms designed for self-organization can effectively manage the complexities involved in such NP-hard problems.

Project Details

Launched on January 1, 2024, and scheduled to reach completion on December 31, 2026, the project, titled “PVSwarm – Self-Adaptive Cross-Agent Intelligence for Photovoltaic Systems” embarks on a 36-month journey to bolster energy efficiency in photovoltaic systems. Spanning from 2024 to 2026, PVSwarm is set to explore the intersection of photovoltaic technology, swarm intelligence, and machine learning to tackle the unique challenges associated with solar energy generation. By capitalizing on these cutting-edge AI techniques, the project aspires to refine the functionality and performance of solar power platforms, promising substantial advances in energy efficiency within the realm of sustainable energy solutions.

Project Goals

Upon the completion of PVSwarm, the project aims to deliver an array of sophisticated algorithms and software tools designed to substantially elevate the monitoring and management capabilities of photovoltaic (PV) systems. By modeling PV plants as dynamic, self-organizing systems comprising diverse interacting agents, the project introduces a new approach to precise and adaptive energy generation oversight. The insights and technological advancements stemming from this project are expected to be disseminated across leading academic journals and conferences, marking a notable contribution to the fields of renewable energy technology and smart grid management.
PVSwarm: Pioneering the Future of Renewable Energy with AI-Driven Swarm Intelligence
The PVSwarm project is at the forefront of technological innovation, offering a sophisticated system to model photovoltaic (PV) plants as networks of synergistic agents. With a focus on integrating the latest in machine learning with the principles of swarm intelligence, PVSwarm is engineered to foster an environment where the system can adapt and improve autonomously. Utilizing machine learning techniques, the PVSwarm project aims for monitoring various parameters of individual PV strings, such as current, voltage, irradiance, and temperature, to anticipate failures before they occur across both the array and string levels. To refine its analysis, PVSwarm organizes these agents into clusters or ‘swarms’ based on common attributes, facilitating a deeper and more insightful comparative analysis. This advanced analytical framework is crucial in distinguishing overarching environmental conditions from unique, string-specific discrepancies—including short circuits, module deterioration, or partial shading—which can significantly affect performance. This is instrumental in creating finely-tuned maintenance schedules and providing granular performance insights. Leveraging the robust analytical capabilities of machine learning beside the dynamic problem-solving provided by swarm intelligence, the PVSwarm project is designed to significantly enhance the operational efficiency and dependability of PV systems. Such cutting-edge, technology-driven strategies promoted by PVSwarm stand to advance renewable energy management, offering a route towards greater sustainability and contributing to the energy independence imperative.

Program: Digital Technologies 2022

“Digital Technologies 2022” is a program of the Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK) and is dedicated to funding forward-looking projects. It aims at securing Europe’s technological sovereignty in selected topics, fostering innovative solutions that address the interconnected realms of digital advancements and societal challenges.

A Comprehensive Perspective

This call aims to bolster research and development in the domain of digital technologies as a pivotal contribution towards establishing a European ecosystem of technological sovereignty. This initiative seeks to marshal the innovative capacities of companies, research and technology organizations, universities, universities of applied sciences, and secondary technical colleges located in Austria.