An application of autoregressive hidden Markov models for identifying machine operations
conference contributionposted on 08.12.2016, 09:56 by Dimitrios Pantazis, Adrian Ayastuy Rodriguez, Paul ConwayPaul Conway, Andrew WestAndrew West
Due to increasing energy costs there is a need for accurate management and planning of shop floor machine processes. This would entail identifying the different operation modes of production machines. The goal for industry is to provide energy monitors for all machines in factories. In addition, where they have been deployed, analysis is limited to aggregating data for subsequent processing later. In this paper, an Autoregressive Hidden Markov Model (ARHMM)-based algorithm is introduced, which can determine the operation mode of the machine in real-time and find direct application in intrusive load monitoring cases. Compared with other load monitoring techniques, such as transient analysis, no prior knowledge of the system to be monitored is required.
This work is supported financially by the Engineering and Physical Sciences Research Council (EPSRC) under the project titled “Adaptive Informatics for Intelligent Manufacturing (AI2M)” and the Centre for Doctoral Training in Embedded Intelligence (CDT-EI).
- Mechanical, Electrical and Manufacturing Engineering