Gas Turbine Control Systems Incorporating Condition Monitoring And Fault Tolerance.

Abstract

Novel concepts in gas turbine control technology incorporating condition monitoring and fault tolerance are described. Concerns with fault tolerance, system reliability, and users functionalities are discussed. The importance of incorporating mechanical and aerothermal analysis for optimum system reliability is discussed. The thesis is presented that system reliability must be defined in terms that include the process controlled (i.e., the gas turbine) and not simply the controlling components. Fault tolerant considerations for both the instrumentation points and the control system can be addressed by means of an expert system coupled with a modelling system. In the described system, an online, background running hybrid expert system compares real time input data from instruments to calculated thermodynamic and mechanical parameters. These calculations are then used to determine data deviations and component performance degradation. Further, a diagnostic system is able to make fault tolerant decisions based on prioritized rules. The paper describes how the integration of the hardware systems can be accomplished to maximize reliability. Concerns with fault tolerance of components are addressed with a practical approach to redundancy and hardware diagnostics on all subsystems. Ways used to maximize the "usability and friendliness" of interface stations are discussed as it relates to system reliability. This is of special, concern since the scientific functionalities of the system could potentially overwhelm an operator. Separate operator and engineer work stations are recommended to maximize the usefulness of the system.