Voltage-Based Load Recognition in Low Voltage Distribution Grids with Deep Learning

Due to the increasing penetration of renewable energies in lower voltage level, there is a need develop new control strategies stabilize grid voltage. For this, an approach using deep learning recognize electric loads profiles presented. This based on idea classify local environment inverter’s connection point provide information for adaptive strategies. The proposed concept uses power systematically generate training data. During hyper-parameter optimizations, multi-layer perceptron (MLP) and convolutional neural networks (CNN) are trained, validated, evaluated determine best task configurations. demonstrated example recognition two vehicles. Finally, influence distance test from transformer active load measurement point, respectively, onto accuracy investigated. A larger between inverter improved recognition, while decreased accuracy. developed shows promising results simulation control.