Harmonic Retrieval Of The Dynamic Responses Of Regenerative Machining System

This thesis derives the mutual-four cumulants expressions of the complex value random harmonic signal and real random harmonic signal under the background of mixed noise, and basing on the mutual-four cumulants, the thesis presents harmonic retrieval TLS-ESPRIT method, which enables joint estimation of harmonic frequency, mutual power and phase difference under the background of mixed noise. Considering the multiple regenerative cutting effect, this thesis develops lag dynamics model of the single freedom machining system under the stimulation of random noise. Basing on the two adjacent cutting chatter displacement, builds diagonal slices of mutual-four cumulants, then presents the harmonic retrieval method of the dynamic response in the regenerative machining system. Inspects the dynamic instable process in the regenerative machining system, reveals the evolutional characteristics of the dynamic response frequency, the mutual power and the phase difference of cutting chatter between the tow adjacent revolution in the regenerative system, meanwhile it also reveals the multiple frequency chatter does also exist even in the single freedom machining system. The theoretical analysis and numeral emulation demonstrate that harmonic retrieval TLS-ESPRIT method which bases on the mutual-four cumulants can restrain the Gauss noise and/or the non-Gauss noise which were unrelated in the channels, and this method possesses both the resolution with high frequency and the stability of harmonic parameters. This method can realize the precise track of the dynamic response harmonic parameters in the regenerative machining system under the background of mixed noise, and this method can be utilized in the earlier diagnosis of the unstability of the regenerative machining.