Voltage monitoring of low-voltage feeders with sparse sensor deployment

Low-voltage (LV) distribution networks increasingly face voltage regulation challenges due to high photovoltaic (PV) penetration and sparse measurement infrastructure. This paper presents a simple, graph-based distribution system state estimation method tailored for radial LV feeders with high resistance- to-reactance ratios. Exploiting the radial topology, the feeder is partitioned into measurement-bounded sections, and section-level net power imbalances are inferred from sparse boundary power-flow measurements. Internal nodal injections are estimated using a uniform allocation rule, and voltage magnitudes are reconstructed through an iterative power flow calculation assuming constant-power loads. The method is validated on a real LV overhead-line feeder in Hungary under a summer daytime scenario with high PV production. Using a full power flow solution as ground truth and assuming ideal sensors, the proposed approach achieves a voltage magnitude root mean square error of approximately 0.46% of nominal voltage. A sensor placement study further shows that branch-level measurements provide the largest improvement in estimation accuracy, with diminishing returns from additional end-of-line sensing.