Automating Substation Modeling Using Labeled Images

This paper presents a comprehensive workflow for automating substation modeling using labeled aerial images. Our approach streamlines the process into three key steps: generating the input dataset as labeled images, inferring an electrical substation model based on standard design principles and exemplarily importing the model into DIgSILENT PowerFactory as a simulation environment. We demonstrate that this new automated process significantly enhances speed, improves consistency, and facilitates easier updation of models, outperforming traditional manual methods.