Karlsruhe Institute of Technology: Solar Park 2.0- Higher Yield on the Same Area
Shade, dirt, or aging considerably reduce the yield of large photovoltaic facilities. Karlsruhe Institute of Technology (KIT) and partners from science and industry have now launched the Solar Park 2.0 project to reduce these losses. Innovative circuits, novel power electronics, and AI-supported optimization are expected to increase the yield and service life of facilities and to reduce their operation costs. The Federal Ministry for Economic Affairs and Climate Action (BMWK) funds the just started project with around EUR 2.5 million.
To achieve climate neutrality, use of renewable energy sources must be increased massively. “Large solar parks are very important in this connection,” says Nina Munzke. The researcher at KIT’s Institute of Electrical Engineering (ETI) initiated the Solar Park 2.0 project at KIT’s Battery Technical Center. “However, it is a problem to find larger areas for such facilities in the densely populated regions of the world. To nevertheless reach our climate goals, available areas must be used far more efficiently.” Within the Solar Park 2.0 project, researchers develop electronic components and methods for this purpose. “We want to increase the power output of photovoltaic facilities under unfavorable conditions, such as shade, dirt, or aging, and to optimize the efficiency and power yield,” Munzke says.
Increased Yields by Novel Power Electronics
To use a photovoltaic module with maximum efficiency, it has to work close to its maximum power point (MPP). “The output of the module results from the current multiplied by the voltage level. At the MPP, output is highest, meaning that the maximum possible yield is reached,” says Lukas Stefanski, ETI. However, the MPP changes with temperature, position of the sun, and other factors. Hence, optimum operation requires the voltage to be adjusted continuously. Specialized power optimizers are applied for this purpose. Still, maximum power point tracking (MPPT) in conventional circuits mainly takes place in the central inverter. “When several photovoltaic modules are connected in series or strings and several strings are connected in parallel, shading and failures of single modules reduce the power produced by the whole facility,” Stefanski says. “It is more advantageous to control single modules and to optimize the voltage applied to the strings depending on the specific circuitry.”
Grafik: Ebenen des MPP-Tracking in großen PV-FreiflächenanlagenGrafik: Batterietechnikum
Levels of MPP tracking in large photovoltaic facilities. (Graphics: Batterietechnikum)
For this purpose, the HiLEM (stands for High Efficiency Low Effort MPPT) circuit patented by KIT is applied in the Solar Park 2.0 project. This circuit replaces combiner boxes that are usually applied for the parallel connection of strings and enables an efficient MPPT on the string level. Combination of a HiLEM circuit with novel power optimizers developed jointly by Karlsruhe University of Applied Sciences and the companies of BRC and PREMA then allows for simultaneous MPPT on both the string and module levels. “This does not only increase the yield of the photovoltaic facility, but also the service life. At the same time, operation costs are reduced,” Stefanski says.
Planned Test Facility on Campus North
The new optimization components are planned to be evaluated in two photovoltaic test facilities of 30 kilowatt-peak (kWp) each. One plant will be used to run test scenarios for the new power optimizers. The second plant will serve as a reference. Both plants will be set up next to each other on a free area of the existing solar park of KIT’s Energy Lab 2.0. Work will also be aimed at developing an AI-supported method to predict power production of photovoltaic facilities based on operation data. This method will then be used to identify possibly shaded, damaged, or dirty modules. “This will help us find out at which point of solar parks installation of power optimizers would be worthwhile,” says Markus Becker from ETI. The AI is trained with long-term data from the existing solar park of Energy Lab 2.0 and data collected by the wireless monitoring system (WSN) developed by the Institute for Photovoltaics (ipv) of the University of Stuttgart.