Optimal location of voltage sensors in low voltage networks for on-load tap changer application

This study contemplates the possibility of estimating in near real-time the extreme voltages (minimum and maximum phase-to-neutral voltages) in a low voltage (LV) network using some, but as little as possible, advanced metering infrastructure (AMI) data. The main target application is to control an on-load tap changer at a secondary substation (medium voltage/LV substations) so as to maintain adequate voltage in the entire downstream LV network. The authors focus on the practical problem that consists in choosing the customers whose (AMI-based) voltage measurements will be used as inputs to the voltage control. They propose a method for the selection of customers that is based solely on statistical load profiles (without the need to resort to past AMI data) and they assess it using real-world data from 38 different French urban and semi-urban LV networks. The authors’ results show that working with simple load profiles instead of large amounts of past AMI data is acceptable, and that in most cases, choosing only a set of about ten customers at most is sufficient to guarantee an almost negligible level of error in the estimation of the minimum and maximum voltages in the LV network.