MACHINE LEARNING BASED IDENTIFICATION AND PRIORITIZATION OF ELECTRICAL CONSUMERS FOR ENERGY MONITORING

This paper presents a machine learning based tool for the automated analysis of circuit diagrams, identifying electrical consumers through computer vision. Detected technical information is extracted and summarized in a report. In a web-based interactive dashboard the identified consumers are prioritized for further actions. Based on their nominal power an ABC-analysis classifies the consumers into three groups. Within an energy portfolio they are divided into four distinctive categories. In both approaches the consumers’ classification leads to specific strategies for energy consumption measurements in the subsequent detailed analysis.