Validating real LV feeder models using smart meter data: a practical experience from project EDGE

The proliferation of residential distributed energy resources is driving the need for advanced and proactive approaches to ensure the integrity of distribution networks. Common techniques accessible to distribution companies often involve power flow analysis, which entail the availability of adequate electrical models (an accurate and complete set of data on customer phase grouping, network topology and line impedances). However, in low-voltage (LV) residential areas, these models are often inaccurate and/or incomplete, creating significant challenges and limitations for distribution companies when trying to apply any model-based approach in practice. Exploiting the availability of smart meter data, this paper presents a three-step, data-driven methodology to validate electrical models. The proposed methodology has been successfully demonstrated in Project EDGE (a high-profile, government-funded trial in Australia) to produce accurate, power flow-ready models of real-world feeders. This work also highlights the key insights and learnings from working with real data and LV feeders.