Daniel Nagy and Stephan Staudacher
Gas turbine monitoring, detection, diagnosis, observability
Gas turbine maintenance is heading towards full condition-based monitoring since the pressure on life cycle cost is ever increasing. Nevertheless, in currently operating machines, a minimum number of measurements is used because of the associated costs and issues of reliability which are linked to any sensor. Therefore, it is an ongoing quest to define an integrated gas turbine health monitoring system which supports the engineers in their difficult task of finding possible faulty modules or engine parts during service. This paper presents a novel health monitoring process approach, dealing simultaneously with single fault events and gradual deteriorations of the gas tur- bine. The modular buildup of the presented process is capable to use purpose-built algorithms provided by an integrated detection algorithm for an autonomous decision between the deterioration types to analyse the health of a gas turbine. The health monitoring process is also capable of robust fault identification due to mea- surement losses out of the in-service instrumentation. In that case, decision guidance is provided for the engineer to isolate the most likely faulty component. Integrated information fusion capabilities between thermodynamically and non-thermodynamically measure- ments (if available) increase the successful fault identification and fault diagnosis. As a design feature for new in-service instrumen- tations an optimal measurement selection algorithm is capable to vote the needed measurements for a satisfactory health monitoring of a gas turbine. The health monitoring system is designed as a user friendly software package with a graphical user interface. An adaption of the health monitoring system to a given gas turbine is guaranteed through an interface to a thermodynamic model of the monitored gas turbine.
Important Links:
Go Back
