Enhancing monitoring of unobserved low voltage grids: an application case study of neural network based state estimation

It is not uncommon nowadays to find portions of low voltage distribution networks which are unobserved, meaning not monitored by the system operator. To support the growing penetration of renewable energy, proper monitoring and control of low voltage grids is essential. However, deploying a large quantity of measurement devices is not economically viable. Therefore, innovative cost-effective methods are needed to address this challenge. This paper shares the experience gained from applying a three-phase neural network-based state estimator in a real low voltage setting with limited input measurements and topology data.