Machine diagnosis based multi-agent technology by autonomous sensor with energy harvesting

A machine diagnosis system based on multi-agent technology is proposed. We mainly focus on developing a multi-agent system for rotating machinery fault diagnosis by vibration sensor with energy harvesting. To estimate the inner rotation of machines in plant frequency analysis is frequently used. Our approach for diagnosis is agent-based, where vibration data is analyzed by a set of software agents coming from distributed servers to the user side. Another feature of this study is the development of autonomous vibration sensors. It earns electric power from vibration so that we are free from battery maintenance, and continuous online monitoring is enabled. Based on the implementation results of the existing multi-agent system design prototype, the harvesting sensor working process can produce total energy of 205µW with a working cycle of about 6.5 minutes, the energy harvester works, and the power accumulates continuously.