Optimization Design Method Study For Integrated Flight/Propulsion Control System

Integrated flight/propulsion control is the tendency to development of futureflight vehicle control. This paper studied the optimization design of centralized anddecentralized controllers for integrated flight/propulsion control system by means ofgenetic algorithm. Centralized control considered the airframe and propulsion systemsas one integrated system and designs a global controller. Centralized control promisesoptimal system performance, however, it resulted in a complex centralized controllerwhich would have difficulties in implementation. Decentralized control designedflight and engine controllers separately and resulted in simpler, easy to implementcontrollers. However, this approach has not accounted the problem of how tocoordinate flight and engine controllers would reach optimal performance. In order tomeet the design requirements and simplify the controllers simultaneously, centralizedcontrol and decentralized control approaches were combined to design integratedcontrol system in this paper.First, the optimization design for centralized controller of integratedflight/propulsion system was studied. An optimization approach was proposed todesign centralized controller of integrated flight/propulsion system. This approachuses genetic algorithm to search the optimal weighting matrixes parameters. Thecentralized controller was solved by these weighting matrixes that satisfy therequirements of US air force fixed wing aircraft handling qualities1and as theperformance reference for decentralized control. The genetic algorithm would promisethe optimality of centralized controller and raise the design efficiency.Then, the optimization design for decentralized controller of integratedflight/propulsion system was studied. A hierarchical decentralized control structurewas provided to handle the interaction from engine to airframe. The selection ofinterface variable was discussed and the most needed control variable was chosethrough analysis. This selection simplified the interface variable controller. A criticalelement affecting the control performance was analyzed in this decentralized structure.The decentralized control performance expression which use frequency weighted matrix as design parameter was deduced. The problem of designing decentralizedcontroller was converted into optimization problem solved by using genetic algorithm.The optimized decentralized controller has achieved the performance of centralizedH∞control system. Optimization design also avoids the iteration during the designprocess.Last, the stability robustness of centralized and decentralized integratedflight/propulsion control system was discussed. Considering the model uncertainty,Small Gain Theorem approach, singular value approach and structured singular valueapproach were used to analysis and evaluate the stability robustness of centralizedcontrol system. Comparing the results from these approaches, structured singularvalue approach showed the nonconservative measure of stability robustness to actualcontrol system. The simulation and structured singular value analysis showed that theproposed optimization design methodology was so effective that the decentralizedcontrol system approximately reaches the same performance and robustness achievedby centralized control system.