ECE 590, Fall 1996, Due 12/12/96
The third-order nonlinear model for a turbojet engine developed by Brennan and Leake consists of 2 algebraic and 3 differential equations, which are given by equations 7-58 through 7-62 on page 212 of the text. The input signal for the system is the fuel mass rate W_fd. The time history for this input signal is:
The purpose of the project is to study the stability and accuracy properties of various explicit Runge-Kutta integration algorithms when simulating a moderately stiff system. The algorithms to be investigated will be: the 2-stage, second-order accurate modified Euler method; several 2-stage first-order accurate methods; and the 1-stage, first-order accurate (Euler) method. For each of the 2-stage methods, the value of the Runge-Kutta coefficient a_2 is to be 1/2.
The initial condition for the state vector is x_0 = [0.5; 1.8; 0.55]. The state response of the 2-stage, second-order accurate method with a timestep of 0.001 seconds can be used as the basis for evaluating the accuracy of the other methods and timesteps.
All of the 2-stage methods to be considered in this project, with the coefficient a_2 constrained to be 1/2, can be parametrized by the design parameter g. When g = 2, the modified Euler method results. For g > 0 and not equal to 2, the order of accuracy of the 2-stage method is 1.
For the modified Euler method (g = 2), for several 2-stage, first-order accurate methods (as a minimum, g = 1, 4, 6, 8), and for the 1-stage, first-order accurate method (normal Euler), investigate the stability and accuracy properties of the methods in simulating the given system as functions of T and g.
Report your results, discussing how the stability and accuracy of the methods compare, providing simulation and stability region plots as necessary to support your discussion and conclusions.
Click the icon to return to Dr. Beale's home page
Latest revision was made on 05/10/01 08:08 PM