Generalization and Application of Matrix Integrators

AbstractThis paper presents the generalization of recently developed Matrix Integrators (MI). The technique utilizes the pole-placement from control theory to compute the integration coefficients for any arbitrary sampling time. The regression coefficients of the integrator under investigation are generalized with the aid of Stirling numbers of the first kind. A heuristic argument is given for the equivalent of this method with a special class of th order hold techniques, which are formulated by means of the Gregory-Newton interpolation function. The MI method is used for simulating a small perturbation model of the NASA Lewis 40-60 mixed compression inlet. It is shown that the MI can provide a significant speed-up compared to conventional linear multistep methods.