An Algebraic Approach to Fault Detection for Surge Avoidance in Turbo Compressor

Abstract

This paper presents an innovative algebraic sensor fault detection approach for surge avoidance in turbo compressors (TC) in the natural gas compressor stations (GCS). The main objective is surge avoidance in the presence of sensor faults in TC. In this way, the robust parity space approach for fault detection is extended to highly nonlinear dynamic of TC based on Groebner basis and elimination technique. No work has been previously reported on the use of this technique for nonlinear dynamic systems with parametric uncertainties. This algebraic approach is simulated on the Moore–Greitzer control oriented model in the presence of parametric uncertainties, disturbances, and sensor faults. Simulation results are presented to demonstrate the effectiveness of the proposed fault detection approach.