Surge avoidance in gas compressor via fault diagnosis

In this paper, sensor fault detection technique based on support vector regression (SVR) method is used for unreal surge suppressing in the centrifugal gas compressor. No work has previously been reported on the use of fault diagnosis (FD) within compressor surge suppressing system. The main contribution of this paper is to present an innovative technique based on analytical redundancy approach for unreal surge avoidance studies via sensor FD. The implementation of this technique is useful to prevent performance deterioration, major collapses in the centrifugal gas compressordue to undesirable surge. Using the experimental data and simulation studies, the effectiveness of the proposed FD scheme is verified.