engleski

Some improvements of the state-transition method in real-time simulation of linear systems

Real-time simulation of linear time invariant systems requires special numerical techniques which are more efficient than other standrad discertization/integration methods. Therfore, it is shown that performance of the well-known state-transition method can be significanly increased by introducing ceratin improvements to the solution of definite integral. By means of compensation filter, distrortion introduced by samplping and reconstruction processes can be significantly reduced accross the overall frequency band of interest. The values of compensation filter parameters can be tuned using nonlinear programming techniques, either in frequency or in time domain, according to design requirements. Furthermore, it is shown that one alternative modification of the standard step invariant method based on the input signal u(k+1/2) presents a robust and accurate discretization technique in comparison to many other well-known discretization/integration methods.

real-time simulation ; state-transition method ; discretization techniques ; integration methdos ; nonlinear programming

nije evidentirano