Optimal sensor placement for state estimation of distribution networks using a genetic algorithm

The integration of renewable energy power plants has introduced high voltage issues into electrical distribution networks, complicating their voltage management. Therefore, new transformer voltage regulation solutions are required to maintain voltage profiles within the contractual range. To address this challenge, EDF R&D and Enedis - the French DSO - have developed Advanced Network Functions (FAR) for voltage regulation thanks to a dynamic change of the tap-changer setpoint and to an accurate state estimation of the network with the help of sensors. This paper presents a specific genetic algorithm for the optimal placement of voltage sensors along a feeder, to minimize the error of the associated state estimator. The genetic algorithm has proven to be robust on several networks and significantly enhances voltage estimation accuracy at each node. The optimal sensor placement allows a precise and reliable state estimation, crucial for effective voltage regulation and overall network stability. This robust and efficient solution for the optimal placement of voltage sensors in distribution networks supports the integra tion of renewable energy sources by ensuring compliant voltage profiles. This algorithm, which allows to use a technically and economically optimal number of sensors, is being used to place new sensors on the network of Enedis.