AUTOMATED ELECTRICAL BEHAVIOUR MODELLING OF ANALOGUE AND MIXED SIGNAL CIRCUITS USING DATA-BASED METHODS

摘要

This paper presents an automated approach for electricalrnblack box behaviour modelling. The introduced strategyrnis based on the transient simulation data of in- and outputrnpins of an analogue or mixed signal circuit. The presentedrnmethod extracts the magnitude and the phase informationrnby measuring in- and output currents and voltagesrnand creates a new approximated electrical pin compatiblernblack-box model. The behaviour model is made up ofrncross connected intelligent admittances blocks and their respectivernfunctional interconnections. These complex connectionsrnbetween the admittance blocks are approximatedrnusing Support Vector Machines. The electrical behaviourrnmodels have a speed-up factor up to 37x and an accuracyrngreater then 95% compared to the original circuit simulation.rnA special filter structure enables the modeling of frequency,rnamplitude and load dependent circuits. The electricalrnfunctioning of the methodology is shown by modeling arnlinear and passive analogue Butterworth filter and an activerntime variant mixed signal circuit.