Researcher: Willem Wagter
Project Duration: October 2022 – June 2023
Project Partner: Liander
Research DESCRIPTION:
Electrification is an impactful consequence of to goal to reduce carbon emissions. In the household environment this entails electric cooking, heating with electric heat pumps, charging electric vehicles and the installation of solar PV installations. All these processes put a bigger claim on the electric grid houses are connected to. These new appliances have a higher peak power draw than current situations. Moreover, the new peaks are also highly likely to occur in different houses at the same time. This combined results in the power use of residential neighbourhoods becoming more volatile and with higher total peak power loads. The Low Voltage (LV) electricity grid is faced with a big challenge and must be expanded in many places in a short period of time to facilitate the above mentioned part of the energy transition Distribution System Operators, or DSO’s, are responsible for the maintenance and expansions of the LV grid. In order for the existing LV grids to be able to deal with the increase in peak loads, a DSO can choose to increase the capacity by replacing cables and transformers by new ones with a higher capacity. This is a costly and time consuming process. Furthermore, due to labour shortage it is simply not possible to do all the needed work at all the necessary places at the same time. Therefore, DSO’s are also looking into the possibilities of flexibility as a way to facilitate the higher demand and more peak loads with less need to upgrade the existing infrastructure. Batteries are a type of flexibility that have the potential to reduce peak loads. This research is focussed on the effect of battery storage on LV grids. For a DSO to be able to effectively deploy battery storage, they need to know the potential, and what possible preconditions for a successfully deployment. In this study two scenarios with different storage types are compared in a simulation to a baseline without storage. One scenario has a number of home batteries, and the other has a single neighbourhood battery. To be able to arrive at representative conclusions while dealing with limitation in computational capacity, the Netherlands is divided into several archetypical neighbourhoods based on a classification based on characteristics of the build environment combined with demographic data. Real existing LV grids samples from these different are sampled from the archetypes are then used for simulation. This makes that the differences in the effect of battery storage between different geographic areas can be discovered based on a real world examples. The results showed that loads on the grids in different geographic areas are indeed quite different. In neighbourhoods with many detached houses or terraced houses the solar production peak is much higher compared to areas with a higher share of tenements and apartment complexes. This limits the potential peak reduction of the battery storage as the solar peaks are too high for available battery sizes. The results also showed that the deployment of distributed home batteries in 30 percent of the connected houses have a much smaller potential impact on peak reduction than using a single neighbourhood battery with the same combined capacity. This is based on the home batteries being controlled by a simple self consumption optimisation algorithm. This type of control is shown insufficient for peak power reducing on a grid level. The neighbourhood battery had a higher impact on total peak reduction due to a forecast based and centralized control. However, the results show the the neighbourhood battery has an impact of 20 percent in the best case. A reduction of this magnitude still will not free up enough capacity on a single LV grid that the impact would be great. Concluding from this study, home batteries with basic controls have proven to be insufficient for peak reducing on a grid scale. More advanced controls and regulatory requirements are needed for this to start to have a significant effect. Neighbourhood batteries on the other hand can result in a reliable peak power reduction between 10 and 20 percent. However, it is doubted that this reduction is significant enough for the DSO to be able defer, cancel or reduce expansion. Additionally, the potential reduction is the most meaningful in areas without the presence of a dominant solar peak.