|Opis:||By increasing the share of photovoltaic power plants in the electricity network, the impact thereof increases as well. Photovoltaic power plants may have a negative impact, primarily on the distribution network. Rapid transitions of clouds cause rapid changes in the output power of photovoltaic power plants, which can cause fluctuations in the voltage, which is not the best for network elements. Currently, there are two effective solutions, namely network reinforcement and the installation of energy storage devices. Both have a negative feature, which is a very high price. The doctoral dissertation presents, designs and develops, up to the level of a functional prototype, a solution that does not produce rapid changes in the output power of photovoltaic power plant. The essence of the proposed solution is to prevent the emergence of rapid changes in the output power of photovoltaic systems, and therefore the mitigation of the consequences, which is typical for existing solutions, is not necessary.
The entire system consists of a cloud forecasting system, managing individual micro inverters, reactive power generation, and a car charging station to compensate for disturbances. A functional prototype of the system for cloud passing forecasting, managing individual micro inverters and generation of reactive power has been developed. The system was tested on an experimental photovoltaic power plant, equipped with a data acquisition system which provided the results of the measurements. The presented system was simulated in the scope of a distribution network model in order to evaluate the impacts of the photovoltaic power plant operation on a low voltage distribution network. Simultaneously, active participation of car charging stations in the mitigation of transients caused by photovoltaic power plants was tested.
A cloud passing forecasting system prototype that provides information on when the cloud will cover the sun was built. The camera with appropriate optical filters and with a processor card appropriately processes the captured photos and makes forecasts of cloud passing. The built-in algorithm is computationally unproblematic, which enables low-cost production. The system allows predictions of cloud passing within one minute so that it can start a specific time before the passage of the cloud gradually reduce the output power of the photovoltaic power plant. Instead of a step change in the output power, a programed change predefined time constant in output power is achieved. An algorithm for managing such a power plant is designed so that it does not cause interference to the network. The algorithm ensures that all transients have entered the desired time constant upon arrival of the cloud and when the cloud moves away, the photovoltaic power plant does not cause rapid changes in the output power. The system is low cost and does not represent a significant additional cost for the installation of a photovoltaic power plant for net self-sufficiency.
Photovoltaic power plants can additionally help the network by participating in the generation of reactive power according to the needs of the network. The dissertation presents an algorithm for the distribution of reactive power among individual micro inverters within one power plant so that the total output active power is maximized. The algorithm takes into account the current values of the working power of the individual microwaves and the efficiency characteristic.
In networks where major voltage changes occur in a short period, car filling stations can be of a further assistance. It has been simulated in the dissertation that, by temporarily adjusting charging power, the car filling station has a positive influence on the voltage profile of the low voltage distribution network. During transients it temporarily decreases and then gradually increases the charging power.|