Multi-scale and multi-structure element edge detection of parameter images for rotating machinery

The vibration signal parameter images for rotating machinery were studied as the object. The 3D vibration parameter images were obtained from rotor's normal state, misalignment and bearing pedestal looseness which were examined on the modeling of 600MW turbine rotor experimental bench. These 3D images of different faults were pretreated and transformed to 2D gray-scale images. According to the multi-scale filtering enhancement processing method, the noise was filtered out and the available information was enhanced of the parameter images. The edges of the parameter images were detected using the multi-structure edge detection operator. The results show that the multi-scale and multi-structure element edge detection method has strong anti-noise ability, this method can filter out the noises and obtain the edge features of vibration signal parameter images for rotating machinery effectively. This method is suitable for the condition monitoring whose environmental noise is comparatively complicated for rotating machinery.