Combined rule-based model-based control of compression system rotating stall

Axial compressors applications require safe and efficient operation over a wide range of operating conditions. Aerodynamic instabilities such as surge and rotating stall could lead to low pressure ratios and even complete failure. Previous research has shown that surge is effectively controlled with model-based nonlinear feedback controllers. The rotating stall avoidance schemes currently adopted maintain a safe stall margin from a predetermined stall line leading to loss of achievable pressure rise during the compressor operation. This work is aimed at reducing or even eliminating the stall margin through the use of active control schemes The rotating stall phenomenon does not have a well defined model that could be used for controller design. Fuzzy logic controllers (FLC) are applicable to plants that are poorly modeled and require only qualitative information about the plant for successful application. Hence in this study, a rule-based FLC with rules synthesized using the available approximate Moore-Greitzer (MG) model is applied to the rotating stall control problem. Initially, extensive simulations were carried out using the Moore-Greitzer model and results of these simulations were used to determine heuristic control actions for the rulebased FLC. Using these rules a Sugeno fuzzy controller was synthesized and its effectiveness was verified using closedloop simulations. The FLC was then implemented on the LICCHUS axial compressor and its effectiveness was verified through real-time experiments. These experiments confirm simulation results.