Applications of gain-scheduled control in power systems and V/STOL aircraft

This thesis explores the versatility of new methods for gain-scheduled control design which address the parameter varying nature of system dynamics as well as hard constraints on state and control variables. The conducted designs are as follows. (1) Gain-scheduled power system stabilizer (PSS) design using linear matrix inequality (LMI) methods for {dollar}{lcub}cal H{rcub}sp{lcub}infty{rcub}{dollar}-optimization. The scheduling variables in this PSS design are the mechanical power input and power angle. Under the formulation of a single Lyapunov function for the overall vertex linear parameter varying (LPV) power system, the performance of this gain-scheduled design is established even in the presence of fast varying mechanical power input and power angle which may be caused by severe system failures. (2) Gain-scheduled boiler-turbine controller design using set-valued methods for {dollar}ellsp1{dollar}-optimization. The nonlinear boiler-turbine dynamics are brought into LPV form which is characterized by a nonlinear dependence on the scheduling variable, the drum pressure. In the local controller design, the parameter variation constraints are not explicitly addressed since the drum pressure is a slowly varying quantity. However, hard constraints on state and control variables are addressed using set-valued methods and heuristic governing strategies. (3) Gain-scheduled V/STOL aircraft controller design using set-valued methods for {dollar}ellsp1{dollar}-optimization. The nonlinear non-minimum phase aircraft dynamics are formulated as an LPV system with the roll angle as the varying parameter, i.e., the scheduling variable. In the controller construction, the change rate of scheduling variable, i.e., derivative of the roll angle, is explicitly addressed as a system constraint so that the hazard of a fast varying scheduling variable is eliminated.