Multi-model structured H-infinity design of a robust and aggressive turbofan jet engine controller

Abstract

Jet engine control comprises tracking either the fan speed or engine pressure ratio setpoints. Further, safe operation entails maintaining several additional parameters, such as high-pressure turbine temperature, combustor pressure, core shaft acceleration and other ones within prescribed limits. A Min-Max selector that features PI controllers is frequently used to handle these requirements. However, this arrangement is overly conservative in the limits management, which unnecessarily slows down the engine response. To overcome this shortcoming, a new controller that adopts the traditional Min-Max structure in combination with the Ndot control, the Conditionally Active and the Conditioning Technique approaches is developed. PI regulators are replaced by dynamic output feedback controllers, which are designed according to a multi-model structured H-infinity methodology. This approach makes it possible to marry robustness with performance, which are two conflicting objectives. Singular value analysis tools demonstrate the robustness of the resulting design. Linear and nonlinear simulations indicate that the proposed controller optimizes the engine response time under the constraint of keeping a set of parameters within prescribed bounds. The features of the proposed design are lucrative for actual implementation in the industry.