Modular Load Estimation Framework for Low-Observability Distribution Networks: Crete Case Study

This paper presents a lightweight, data-sparse methodology for estimating nodal loads in low-observability distribution networks. A modular approach is designed for settings where smart metering is unavailable and only feederlevel SCADA measurements exist, while allowing for increasing levels of truth points as available. Two estimation strategies are shown: a baseline “null-hypothesis” estimator, which assumes uniform loading, and a profile-based estimator, which uses temporal variations extracted from smart meters. A novel alpharithmetic naming convention is proposed. The methodology is implemented on a real MV feeder in rural Crete, and its performance is validated against SCADA data. The framework is designed to comply with system operator workflow and allow for rapid estimator development. The results show that even under severe observability constraints, accurate voltage and power estimates can be achieved using this replicable approach.