An Expert System to Identify the Steady State Performance Characteristics of Journal Bearings through Non-Intrusive Image Processing Techniques

Dynamic and steady state characteristics of rotating shafts housed in journal bearings are partially defined by the flow of lubricating fluids within a small gap between the shaft and the housing. This paper identifies the steady state operation and generates qualitative, as well as quantitative results with regard to flow velocities and accelerations within the journal bearing. The small gap between the shaft and the housing is illuminated by the laser beam, the resulting scene is magnified, digitized and stored for further processing. This non-intrusive measurement process combined with image processing techniques creates a complete quantitative flow trajectory profile. The image processing has two steps. A statistical model-based technique first identifies small special particles within the fluid. This information is then conveyed to a rule-based expert system in the second step to determine the flow characteristics. Additional circumferential pressure measurements between the shaft and the casing complement the results of image processing for a complete non-intrusive identification of the steady state performance of the journal bearing.