Model-based method for condition monitoring and diagnosis of compressors

Abstract

In process industries, compressor plants provide process mediums with certain flows and pressures. Due to corrosion, wear and fouling, the plant behavior may change throughout the entire life cycle of several decades. This paper presents a new model-based method for condition monitoring and diagnosis of turbo compressors, in which the characteristic operating behavior is mainly influenced by fouling. The real-time capable method is implemented on a plant control unit and validated by Hardware-in-the-Loop simulation. Furthermore, the diagnostic result is used to automatically adapt the controller behavior. The presented approach increases the original system’s resource efficiency without reducing its robustness.