Fast tools for predicting compressor performance during surge and stall

Abstract

Two fast models for predicting compressor post-stall performance are presented and compared in this investigation. Available experimental and numerical results in the open literature have been used to validate the accuracy of the models and investigate their capabilities (the so called SK and one-dimensional models). By comparing the SK model to experimental results it is revealed that the SK model can accurately predict the speed of the stall cells over time. Furthermore, this model can predict the compressor deep surge cycles fairly accurately. However, it is shown that the SK model does not have sufficient accuracy in modeling compressor mild surge phenomena without reverse flow. It is demonstrated that the one-dimensional model, which is capable of simulating the stable as well as unstable parts of the compressor, can predict surge cycles in subsonic/transonic rotors and single/multi-stage compressors, fairly accurately. By comparing NASA Rotor-67 time-accurate numerical results at surge conditions to the one-dimensional model, it is demonstrated that the model can reliably predict the rotor surge cycles in a few minutes. The maximum error was 3.85% for surge cycle frequency, and 0.94 % and 13.21% for maximum and minimum surge cycle mass flow rate, respectively.