Rough Set Neural Network Feature Extraction and Pattern Recognition of Shaft Orbits Based on the Zernike Moment

In the shaft axis monitoring of hydrogenerating unit condition and fault diagnosis, orbit is intuitive comprehensively reflects operation state, different orbits correspond to types, which can accurately indicate a system vibration fault. Shaft identification has important significance for diagnosis. getting feature extraction pattern recognition orbit, Zernike moment better than Hu moment; it advantages small calculation error high rate. A rough set neural network (RS-BP hybrid model) established, uses just 13 eigenvalues reserved by selection algorithm as input variables; same rate reduces time step. The simulation shows that model achieved good results in orbits.