The integration of distributed generation, heat pumps, and electric vehicles makes monitoring low-voltage networks increasingly challenging, while the existing PLC-based metering system cannot provide real-time insight due to communication delays. The aim is therefore to determine optimal locations for smart meters with GSM communication and one-minute resolution, minimizing state estimation uncertainty with as few meters as possible. Three placement methods are compared: D-optimality, eigenvector analysis of the admittance matrix, and metering point ranking. State estimation is performed using weighted least squares in Python/pandapower, and the methods are tested on real transformer substations of Elektro Ljubljana.
|
