Non-Intrusive Hybrid Energy Monitoring System

Abstract

When the optimization of the manufacturing units is considered, the management usually focus on equipment costs and technology parameters. However, being the rising price of the energy, the energy efficiency has become a sensitive issue. Detailed energy related information should be collected in order to understand the consumption profile of each relevant machine/process and identify potential energy savings. Data gathering is time-consuming and very expensive since individual metering devices have to be used for each consumer. To overcome this inconvenient, several solution have been reported. All of them are making use of a single metering device (monitoring the overall energy consumption) and some Load Signature Identification algorithms, used to disaggregate the overall energy between the identified consumers. Their relatively low detection rate (about 80-90%) hindered the spread of LSI-based architectures. In this context, a new hybrid architecture is proposed in this paper, together with it's hybrid LSI algorithm. Several test performed in an electronic prototyping facility result in a detection rate close to 100%.