Summary: | 碩士 === 國立成功大學 === 電機工程研究所 === 82 === In this thesis, a state space approach is investigated to
determine the exact upper bound of singularly perturbed dis-
crete systems. In this method, the singular pertuebed para-
meter is treated as structure uncertainty. The stability
bounds for the two-type discrete singular perturbed systems (
slow sampling rate and fast sampling rate systems ) have been
determined via the Kronecker product and the Bialternate
product techniques. Compared with present literature, the
main difference is that no Nyquist plot is needed. The second
topic of this thesis is to examine the robust stability of
regular perturbation in the discrete singular perturbation
systems. A sufficient condition is derived and the so-called "
inverse " problem is solved in these systems when singular
perturbation parameter is supposed to be known in advance.
Next, an algorithm is proposed to acquire the stability range
of the pencils matrix when the nominal system is unstable.
The algorithm is based on the fact that the set of nonzeros
eigenvalues of Kronecker sum contains the information of
critical values on stability. Several application such as the
the bilinear system, the high gain system and the output
feedbak system is utilized to demonstrate the effectiveness of
the proposed scheme. Finally, the singular perturbation
methodology is employed to test the stability of the passive
suspension system. The upper bound of the resulting closed-
system is also caculated.
|