Research

• Modal analysis

Modal analysis is a mathematical tool for studying the dynamic properties of different physical systems in frequency space. It is a method for analysing, measuring and calculating the dynamic responses of different (excited) systems. In the power system, the modal analysis procedure is nowadays mainly used to perform studies of angular stability as a result of small disturbances. The question arises whether the application can be extended to the field of determining the influence of resonance states on the propagation of harmonics through the network.

The aim of the research is to test the possibility of using modal analysis to study and limit resonance phenomena in the power grid.

• Reference models for LV and MV networks

The operation of a network with distributed resources and active users can be simulated using models implemented in different simulation environments. In order to accurately analyse the impact of different consumers on the operation of a specific part of the grid, it is necessary to model this part of the grid accurately and to know the actual operating conditions. Unfortunately, in practice, actual network data is often not available to us. It is also difficult to assess the overall impact of different technologies on the whole network, as the part of the network analysed may not be representative of the whole network. For this reason, the impact of new technologies is often simulated on reference network models and the results obtained can then be generalised to the whole network. Some machine learning methods, e.g. classification and clustering, are often used to find reference or representative networks. To perform network grouping, we need a set of networks that are described by different attributes or features. Examples of features include the total length of taps, the length of overhead lines, the rated power of transformers and the proportion of residential customers. The characteristics need to be pre-processed and the most appropriate method for classification needs to be selected. The classification methods then provide us with a result in the form of groups of networks that have similar characteristics to each other and are different from the networks in the other groups. We refer to the representatives of these groups as reference networks, which capture the properties of the networks in each group. The results of the simulations on the reference models can then be more easily generalised to the whole distribution network.

• The impact of electric vehicles on the grid

The transport sector contributes a large share of global CO2 greenhouse gas emissions, and it is in the interest of most countries in the world to decarbonise this sector. In recent years, we have seen an acceleration in the replacement of internal combustion vehicles by plug-in electric or hybrid vehicles, which require electricity from the grid to operate. The increased number of plug-in electric vehicles will have an impact on the operation of the transmission and distribution grid due to their high charging power. Uncontrolled charging of a large number of electric vehicles can lead to overloading of grid elements such as transformers and power lines, as well as unacceptable voltage drops in the grid, potentially leading to high investments for grid reinforcement. The latter can be avoided by a proper implementation of the smart charging concept, which allows to adjust charging according to the current grid conditions with the aim of reducing the number of threshold violations.
When simulating the impact of EV charging on the grid, we are confronted with unknowns such as charging power, battery size, user habits and vehicle consumption, among others. The accessibility of charging infrastructure will also play a very important role in facilitating the mass switch to electric vehicles, as it must be appropriately spatially located and allow for the carefree use of electric vehicles.

• Distribution Network Health Checker

Distribution networks have been subject to new challenges in recent years, driven by, among other things, the increased integration of distributed generation, heat pumps, the growing share of electric vehicles and the desire to increase network flexibility. This requires an increased awareness of the modern distribution network, which is a prerequisite for increasing its manageability.
System awareness is directly related to the amount of available measurements and their accuracy. For various technical and economic reasons, the condition of the existing meters in the system does not usually allow for adequate system cognisability. In such cases, an adequate level of cognisability can be achieved by performing a State Estimation.
The State Estimator algorithm calculates, or estimates, the most likely state of the system, given the currently available network measurements and a mathematical model of the network. The method, which has been known in the transmission network for decades, shows promising results when applied to distribution networks of different types.
The process of assessing the state of the distribution system is essentially the same as in the case of the transmission system. However, despite the initial similarities, its design and implementation in the distribution network is fundamentally different. This is due to the many differences between the two types of networks. The main challenge in the implementation of the assessor is therefore its robustness, which allows it to quickly and reliably assess the state of the different distribution networks.