Voltage Pseudo-Measurement Generation for Distribution System State Estimation: A Smart Meter-Based Approach

The increasing deployment of smart meters (SMs) in low-voltage distribution networks offers promising opportunities for real-time distribution system state estimation (DSSE). This paper presents a probabilistic DSSE framework based on Linear Weighted Least Squares that leverages voltage pseudo-measurements derived from SM data. Unlike conventional approaches focused on power pseudo-measurements, we propose a novel voltage angle estimation model derived from voltage magnitudes using linear regression, enabling single-iteration state estimation that simultaneously solves for bus voltages and voltage regulator tap positions. A simple voltage pseudo-measurement model based on adjacent node voltage similarity addresses measurement loss in low-observability systems. The methodology demonstrates superior accuracy over artificial neural networks and has been validated across American and European grid topologies, including a realistic three-phase network with approximately 4000 nodes where medium-voltage states are inferred from low-voltage SM metering data. Even with significant measurement loss, the majority of true states remain within estimated confidence intervals, confirming the reliability of the proposed pseudo-measurement models. This framework supports real-time smart grid operation without costly computational infrastructure or PMU deployment.