Feeder mapping and load flow algorithms for LV distribution grids

Abstract

With the advent of the electric vehicle, distributed generation and other emerging technologies such as heat pumps, the complexity and uncertainty of LV Distribution Networks (LVDNs) is increasing. Advanced Metering Infrastructure (AMI) at both the transformer substation and consumer level is producing a significant amount of data that can be exploited for coordination and monitoring. Load-flow analysis is one key application for this data used in network studies and scenario testing to determine the voltages at all levels in the network based on expected loading and ensure that these remain at acceptable levels. Load-flow simulations require the knowledge of the connectivity model (CM), i.e. the whole set of connections between consumers and their corresponding transformer feeder and phase but this is often partially or completely unknown within typical LVDNs. This novel research proposes Feeder Mapping (FM) algorithms that have been developed to resolve the connectivity model based on AMI time series data. Classical optimisation and nature-inspired algorithms which exploit energy conservation have been developed and evaluated. A Load-Flow model using the derived CMs and based on the Backward-Forward Sweep is proposed, integrating the Fortescue transform for handling load imbalance.