Estimating Household Loads in Low Voltage Grids: A Framework for Incomplete Grid Monitoring

This paper presents a novel framework for estimating household loads in low voltage (LV) distribution grids with incomplete grid monitoring. The increasing adoption of renewable energy sources and electrification of mobility and heating necessitates enhanced load estimation techniques to ensure safe and optimised grid operation. Traditional methods relying on standard load profiles (SLPs) are insufficient for today’s volatile LV grids. This framework clusters households based on their load profile and classifies them based on additional features. So multiple different representative load profiles are created. The framework then assigns a cluster and with that a representative load profile to a household based on its features. This enhances the observability of the LV grid and avoids congestions. The framework is evaluated using a prototypic implementation in a test network with pandapower. The results show that the clustered load profile follows the measured load profile better than the SLP-based estimation. The next stage of development is the dynamic load estimation using live measurements from smart meters and additional information to link those smart meters to the unknown households. This will further improve the estimation method.