Identification of the electrical load by C-means from non-intrusive monitoring of electrical signals in non-residential buildings

The combined action of different equipment connected to an electrical installation is capable of causing unexpected changes in the load kinds inside the installation; these load variations are responsible for some electrical failures. In this paper a methodology to classify and to identify the load kinds in industrial environments is presented. Energy power quantities (EPQ) and current values are used to establish indexes in order to use them as features for a C-means algorithm and perform the load classification. The experimentation is done in a healthcare facility gathering electrical data in different electric distribution boards. The results obtained from the classification method show variations in the load behavior along the day. Furthermore, some classes can be used to recognize equipment in the electrical installation for further inspection or failure detection.