b'Electronics | Engineer InnovationAvoiding the next major Power Black-Out with Digital TwinsBy Hilary Chisepo, MSc(Eng), PhD(Eng) candidate,University of Cape Town, South AfricaFigure 1. Bench-scale transformers single-phase shell type (left) three-phase five limb (centre) and three-phase three limb (right).laboratory. Numerous physical resultsrespectively. The same protocol was involving power, reactive power, andapplied using the other two multi-limb harmonics were recorded under GIC-likecore structures and showed that while conditions. Figure 1 shows the bench- the core loss calculation (time-averaged scale transformers used in thehysteresis loss) was fairly accurate laboratory testing. With future studiesagainst the measured data, the for power transformers in mind, themagnetizing currents were consistently same laboratory testing protocol waslower in the FEM. applied, as far as the FEM environment allowed, in Simcenter MAGNETTMValidation of simulationsSoftware to test the validity of theThe key parameters used to validate the simulation modeling. Figure 2 shows aFEM models with simultaneous AC and FEM model of the single-phase shell- GIC/DC components were reactive type transformer. Simulations werepower, waveform distortion (harmonics), performed in the transient domain thatand time response. Figure 3 shows good allows for multiple excitation andcorrelation between measurement andFigure 2. Full 3D FEM model of the single-generation of voltage and currents withsimulation data for reactive power andphase shell type bench transformerdifferent frequency components in theincreasing levels of DC. The generation same signal (harmonics). A coupledof even and odd harmonics in power field-circuit approach was deployedsystems is a typical marker for a partially whereby all the simultaneous AC and GIC excitations were triggered in the circuit domain. Analysis of the flux distributions to verify the effect of saturating the core was performed in the FEM domain and the instantaneous voltage and current waveforms were recorded and post-processed in RMS voltage and current calculations, and average power and reactive power calculations. Initially, the FEM model was tested under open circuit conditions with AC excitation only at the nominal voltage for comparison with the measured data. For the single-phase shell type model, the measured magnetizing current and open circuit power (core loss) were 55 mA and 3.2 W, respectively. The 3D solution yielded 41 mA and 3.4 W for theFigure 3. Measured and simulated reactive rower Q in per unit (p.u.) of the VA rating vs. DC in magnetizing current and core loss,p.u. of the transformer magnetizing current (Imag) for 1p3L laboratory transformer.37'