Minimum Data Acquisition Time for Prediction of Periodical Variable Structure System

Abstract

The term of data acquisition can be meant as 'long-term' acquisition of integral data (maximum-, average-, and root-mean-square values) sampled in order of hundred milliseconds up to seconds. Another meaning is 'on-line' data acquisition for real time control with data sampled in hundred microseconds. The goal of this work is to obtain (determine) necessary sample of acquisited data at shortest minimum time. The chapter of the work deals with specific utilisation of data acquisition for identification and prediction of transient behaviour of power electronic systems. Transients in dynamic systems are originated by changes of energetic state of state variables of accumulation elements (in electrical systems: chokes’ currents and capacitors’ voltages). Their duration is of non-zero time in every case, as the instant change of energetic state of accumulation elements would require infinite power. Duration of transients is theoretically infinite, except the cases when the transient phenomenon does not occur in the system at all (connection of inductive load at the instant of time equal to current phase angle in steady state). Time behaviour of state variables in linear systems is given by actual values of all elements in the system. State variables time behaviour and thus also system response in inverter systems depends as well on sequence of switching elements operation. Later on, the paper analyses systems with periodically variable structure (e.g. in Fig. 1).