Active Hierarchical Fuzzy Control for Gas Turbine Altitude Relighting Using Multi-Objective Optimization

The developments of high bypass turbofan engines and improvements in high temperature materials have resulted in smaller engine cores. These smaller core engines result in correspondingly lower airflows that introduce their own design problems, including failure to reach idle, stagnation, increased time to reach idle and lack of surge margin. Such problem can occur in engine relighting at low altitude and immediately following take-off. Engineers have provided solutions to such problems but they tend to be specific solutions for individual engines. In this paper a systematic approach to improve engine performance in altitude relighting is addressed which does not involve any modification to hardware, including the engine components or the fuel metering unit. The application of active control solutions has been evaluated to improve engine low altitude relighting performance. Initially an active fuzzy controller has been designed to improve the pull-away performance after relight. Further, an active hierarchical fuzzy controller has also been designed to establish if it can offer further benefits. Multi-objective optimization is used, both to optimize the parameters of the fuzzy controllers and also to produce a concise number of fuzzy rules for the fuzzy controller(s). Compared to the conventional controller, with a look-up table approach, the active fuzzy controllers can reduce the time to reach idle speed, by utilising compressor surge margin and increase combustion operability. In this way, a generic and more flexible framework for designing relighting logic emerges, that captures the knowledge incorporated within the design of the original logic. This leads to significant improvements in the design of future relighting control systems.