Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Early diagnostics of state of dynamic objects on basis of analysis of stationary time series G.S. VERESNIKOV The article discusses approach to solving the problem of identify the dynamic object state based on the analysis of stationary time series. Solution of this problem required to a system of early dynamic object diagnosis. The proposed approach is the integration of spectral analysis (fast Fourier transform) and neural networks. The article describes the experimental data for solving two arising in practice task. The first task is to identify the mode of operation of the gear unit helicopter based on vibration data. The second task - identification speed helicopter flight based on vibration data in three directions of vibration. For both tasks received high identification accuracy dynamic object state. And also draw conclusions about the effectiveness and benefits of applying the proposed approach. Keywords: Diagnostics, dynamic objects, identification, spectral analysis, fast Fourier transform, neural networks. Contacts: E-mail: veresnikov@mail.ru Pp. 51-58.