Project start: 1.10.2025.

The increasing integration of renewable energy sources, battery storage systems, and new types of loads such as electric vehicle charging stations has led to significant changes in the operation of power systems. Traditional security analysis methods, based on the N-1 criterion, are becoming less suitable for addressing real operational challenges, particularly in the context of voltage stability. The causes of voltage disturbances are often linked to inaccurate modeling of electrical load, which remains one of the least understood and explored elements of the power system.

The aim of this project is to develop and validate new methods for advanced power system security analysis and for determining electrical load parameters using neural networks. The research includes the generation of a synthetic database of system responses to voltage disturbances, the development and training of neural networks for predicting the shares of static and dynamic load components (ZIP+AM model), and the creation of an algorithm for advanced power system security assessment based on modal voltage stability analysis. The developed models will enable non-invasive and reliable estimation of load parameters using real historical data on system response during faults and forced tap changer operations on power transformers.

The project is based on an interdisciplinary approach that combines power system engineering, signal processing, and machine learning. Scientific contributions are expected through the publication of research papers, development of software tools and applicable methods for transmission system operators, and the completion of one doctoral dissertation. The results will contribute to enhancing the resilience and reliability of power systems and open opportunities for collaboration in national and international research projects. The project is aligned with the goals of sustainable development and contributes to the strategic priorities of the digital and green transition.