26th International Congress on Condition Monitoring and Diagnostics Engineering Management - COMADEM 2013, Helsinki (Finland). 11-13 June 2013
Summary:
The detection of anomalies as soon as possible in a wind turbine is critical in order to avoid possible serious consequences which may cause fatal failures. Continuous monitoring of the main variables characterizing its life and the analysis of this information are the first steps for an effective incipient detection of anomalies. In the case of a wind farm, a significant number of wind turbines may be located there and the continuous monitoring and analysis of their data implies an important workload for any operation team. This paper describes a multi-agent system designed to assist in the operation of a wind farm by the automatic detection of possible anomalies and diagnosis of their causes. The main roles of the agents in the multi-agent are: data acquisition, analysis of information for detecting anomalies and the diagnosis of their possible root causes. The paper describes the objectives and roles of the agents, the architecture of the multi-agent system, the software used and the main characteristics of the prototype developed.
Keywords: Multi-agent system, anomaly detection, neural networks, diagnosis, wind turbine, expert system
Publication date: 2013-06-11.
Citation:
M.A. Sanz-Bobi, R.J. Andrade Vieira, M.A. Marañón-Astolfi, Multi-agent diagnostic system for incipient detection of anomalies on a wind farm, 26th International Congress on Condition Monitoring and Diagnostics Engineering Management - COMADEM 2013, Helsinki (Finland). 11-13 June 2013.
