Abstract: To solve the problems existing in the intelligent maintenance of intelligent equipment such as low-level intelligentization, networking and the difficulty of establishing physical model, the research is made on the framework, key technologies and system development methods of the data-driven remote prognostics and health management system (PHM) for intelligent equipment. The operating mode of the data-driven PHM system is specifically described. Based on it, the software architecture and key technologies of the PHM system are analyzed. First, the EEMD is used to denoise and reconstruct the original signal, and the reconstructed signal is applied as the input to establish the diagnostic model based on RBF neural network. Then the fault prediction model based on time series is established by dynamic neural network, and the fault alarm mechanism based on the fault threshold is set up. Finally, the hybrid programming and networking are employed to develop the data-driven remote PHM system. The practical application results show that the system, with a good practicability, can efficiently perform the core functions of fault diagnosis and prediction.

Key words: intelligent equipment, PHM, data driven, intelligentization, networking