Research On Turbo-shaft Engine Model Correction And Control Method

High precision turbo-shaft engine model is the basis of control system design. It is an important research subject to improve the accuracy of the model and make it match the test data. Adopting advanced control laws which can improve the response speed of the turbo-shaft engine power and enhance its disturbance rejection capability is an important and difficult problem in turbo-shaft engine control system design. Thus the matching of turbo-shaft engine model with experimental data and the control law design of the engine were researched in this dissertation.In terms of the model matching, the model accuracy on the design point was corrected first. The component characteristic correction coefficients were optimized based on the algorithm with varying fitness function to reduce the overall modeling errors and improve the modeling accuracy. A Newton – Raphson iteration algorithm with varying calculation step was used. The calculation step was adjusted according to the residuals change trends of the balance equation. The convergence ability and convergence speed of the model were improved. On the basis of the steady-state model, the differential evolution algorithm was used to modify the entraining coefficient, total pressure recovery coefficient, and the characteristics of the component. The matching precision to the test data of the model outputs after correction was met the requirements on the design point.To the non-design point model matching problem, two methods were adopted. One was multi-point matching method which adjusted two adjacent rotate speed line. The other used the correction factor function approach, and the corrected speed line after matching was taken as the reference to adjust the non-design points.Based on the model after correction, the controllers were designed for the power control system and the speed control system. The state variable model was built based on the small disturbance method. The differential evolution algorithm was used to get the parameter of the state variable model, and the accuracy of the model was improved. For power control system, an LMI(linear matrix inequality) method was designed to solve H2/H ? controllers. For speed control system, an multi-objective optimization algorithm was researched to design the H2/H? controllers. The validity of the algorithms was verified by simulation.