Search

The problem of voltage regulation in unknown constant resistive loads is addressed in this paper from the nonlinear control point of view for second-order DC-DC converters. The converters’ topologies analyzed are: (i) buck converter, (ii) boost converter, (iii) buck-boost converter, and (iv) non-inverting buck-boost converter. The averaging modeling method is used to model these converters, representing all these converter topologies with a generalized port-Controlled Hamiltonian (PCH) representation. The PCH representation shows that the second-order DC-DC converters exhibit a general bilinear structure which permits to design of a passivity-based controller with PI actions that ensures the asymptotic stability in the sense of Lyapunov. A linear estimator based on an integral estimator that allows reducing the number of current sensors required in the control implementation stage is used to determine the value of the unknown resistive load. The main advantage of this load estimator is that it ensures exponential convergence to the estimated variable. Numerical simulations and experimental validations show that the PI passivity-based control allows voltage regulation with first-order behavior, while the classical PI controller produces oscillations in the controlled variable, significantly when the load varies.

Load and generation variations and the random nature of harmonics in non-linear devices (NLDs) are the source of multiple uncertainties that can be handled in harmonic load flows (HLFs) by probabilistic or interval formulations. The paper presents a new combined analytical technique (CAT) for iterative HLFs in the presence of correlated input uncertainties from photovoltaic (PV) systems in radial distribution systems (RDSs). This technique merges the point-estimate method (PEM), a probabilistic formulation, and complex affine arithmetic (AA), an interval formulation. It then uses the information derived in Legendre series approximation (LGSA) to approximate harmonic voltage distributions. Unlike other methods, this CAT includes iterative harmonic penetration (IHP), which provides a way to deal with the interaction of background harmonic voltage on PV harmonic current. The CAT was examined in a real ENDE 100 RDS system. Thanks to PEM and AA, the CAT significantly reduced the computational burden, an evident improvement over the Monte-Carlo simulation (MCS). Furthermore, the simulation results showed that it accurately reconstructed the harmonic voltage distributions (magnitude and phase angle). The iterative approach also underlines the relevance of background harmonic interaction. The CAT outperformed the incomplete CAT (ICAT), which was based solely on a probabilistic HLF formulation.

Probabilistic harmonic load flow HLF is currently enjoying renewed popularity. This is not surprising since in the near future, thousands of photovoltaic PV systems will be integrated into distribution systems. However, as yet, there is no model capable of explaining PV harmonic current behaviour in probabilistic HLF studies. To fill this gap, the harmonic modelling of PV systems proposed in this paper has three key points. First, it provides an effective model of the relationship between PV harmonic current emission and background harmonic voltages. Second, it statistically characterises PV harmonic currents relative magnitude and phase angle at different fundamental-frequency current output intervals using historical time-series data. In this statistical characterisation, the first fifth moments of each PV harmonic current are used to accurately approximate the raw probability density function PDF by means of the Legendre series. Finally, the third key point of this harmonic modelling is a method capable of determining the distribution functions of PV harmonic currents absolute magnitude and phase angle, based on the statistical characterisation and a fundamental-frequency probabilistic PV model. The numerical results obtained confirm the effectiveness of this PV model. Copyright © 2014 John Wiley & Sons, Ltd.