Flying and handling qualities oriented longitudinal robust control of a fighter aircraft in a large flight envelope

Abstract

In the scope of this thesis study, robust control design apprach has been applied to F-16 aircraft which is aimed to satisfy Level 1 FHQ within the specified flight envelope. First, a brief information about the histoy of flight mentioned in the introduction chapter. This historical storyline starts from the early sketches of Leonardo da Vinci and extends along to Wright Brothers who had achieved the first sustainable, controlled heavier than air flight. Then the innovations in aerospace industry is mentioned along with the advances in technology at the same time. The milestone successes are explained which has brought us to realize the design of fly-by-wire flight control algoritms. Then a literature review of the documents about the F-16 aircraft, FHQ criteria, multivariable robust control applications and mathematical backgroud of this approach. The structure of the thesis has outlined. Then, F-16 aircraft has been presented along with the aerodynamic data and how the force and moment of the aircraft is related with the aerodynamic and thrust data. The presented data of the F-16 aircraft was obtained from the researches of NASA Langley Research Center which is based on wind tunnel test results of the F-16 aircraft. The mathematical model of the F-16 aircraft is introduced. This mathematical model includes the airframe spacifications, the mass data, the systems that represent actuator and sensor and the environmental data which gives the atmospheric properties with respect to the flight condition of the aircraft. Then, the trim and linearization algorithms are introduced for the steady-state wings level flight condition. The inputs, states and outputs related to the longitudinal motion of the aircraft has been identified and the resultant state space linear system which represents the characteristics of the aircraft is obtained. The longitudinal modes which are phugoid and the short-period mode are mentioned. Next, the flying and handling qualities to evaluate the performance of the aircraft are emphasized. The reason for the use of flying and handling qualities are determined and related with the pilot evaluations Cooper-Harper ratings. The suggested flying and handling qualities are explained for the use of both design guidance and evaluation criteria. It is mentioned that the CAP criterion is used as design guideline whereas the Bandwidth and Dropback criteria are used as evaluation criteria for the aircraft in the both frequency and time domain. The corresponding flying and handling qualitieslevels are detailed for the criteria and related intervals for the properties are supported with the graphical representations. The robust control approach is introduced while mentioning the background of the method. The norm definitions are done and the feedback properties are given in the related chapter of this thesis in order to associating the design purposes with the feedback properties. The relationships between the open-loop characteristics and closed loop results are identified and the loop-shaping aproach is emphasized. Yhen the uncertainty definitions are identified. The classes of uncertainty and where they are reasoned for is explained. An uncertainty definition which is suitable for the use in this thesis is mentioned. Then the H_∞ Loop Shaping approach is expressed. The normalized coprime factorization method is explained and the design of both one degree of freedom and two degrees of freedom H_∞ Loop Shaping approaches are detailed with the design steps. Then, the control structure used in this thesis is explained. It is aimed to design a pitch rate controller which will results in Level 1 flying and handling qualities within a specified flight envelope. The design has been made for one design point and then the resulted parameters are used for the whole flight envelope. This enables to overcome the complexity of gain scheduling manner and provides robustness against any probable loss of air data such as angle of attack. The controller architecture of NASA research was presented for the longitudinal axis. Then the optimization structure to find the design parameters which ensures that the pitch rate demand flight control law results in Level 1 flying and handling qualities within a specified flight envelope. The root mean square approach has been applied in optimization phase. It is purposed that the time responses of 5 different design points after step input should follow a desired transfer function response specified during the design of the two degrees of freedom H_∞ Loop Shaping algorithm as close as possible. Moreover, in order to satisfy the specified flying and handling qualities, time delay parameter is included in the optimization cost which makes the optimization multiobjective optimization with a weigthed sum cost function. The resultant optimized parameters for the design of two degrees of freedom H_∞ Loop Shaping architecture is given. The results of both nominal design point and the responses of 5 different design points along the flight envelope are presented. The flying and handling qualities evaluations are shown. Then the performance and stability robustness results are associated with the results. The comparison study between the two degrees of freedom H_∞ Loop Shaping algorithm and the NASA control structure which emphasizes a classical PI controller has been presented. The results are satisfactory as all the design points resulted in Level 1 flying and handling qualities responses in both frequency and time domain. It is seen that the control architecture is successful for performance and stability robustness as all uncertain plants are following the nominal response and no frequency response has crossed a nichols exclusion zone defined. The two degrees of freedom H_∞ Loop Shaping algorithm outperformed the NASA PI controller as Level 2 results are seen for NASA PI controller responses. The use of two degrees of freedom H_∞ Loop Shaping structure lowered the time delays as it was purposed in the optimization goals as the effective time delay results are less than the NASA PI controller.