Learning-Based Customer Voltage Visibility With Sparse High-Reporting-Rate Smart Meters

The conventional passive distribution network (DN) has evolved into an active DN enriched with distributed energy resources (DERs). The network’s voltage profile becomes volatile and fluctuating, necessitating real-time high-granular voltage estimation for prompt decision-making in the network. This paper introduces an innovative real-time voltage estimation scheme on the customer side using sparsely but strategically deployed high-reporting-rate smart meters (HRRSMs). With a deep learning model capturing network and temporal correlations, the real-time voltage data reported by sparse HRRSMs can be leveraged to estimate the voltages at the terminals of all the unmeasured customers in high resolution and in real time. A clustering-based sparse HRRSM installation technique adhering to budgetary constraints and alleviates communication network burdens is also developed.