arXiv subjects

Dionysios Aliprantis

Publications and source records attributed to Dionysios Aliprantis.

4 recordsLinked to original sources

arXiv2026

Frequency-Domain Characterization of Load Demand from Electrified Highways

Electrified roadways (ER) equipped with dynamic wireless power transfer (DWPT) capabilities can patently extend the driving range and reduce the battery size of electric vehicles (EVs). However, due to the spatial arrangement of the transmitter coils in the ER, the DWPT load exhibits frequency content that could excite power system frequency dynamics. In this context, this work aims to study the spectrum of DWPT loads under different traffic conditions. Under simplifying assumptions, we develop statistical models to identify the location and relative magnitude of DWPT load harmonics. Our analysis reveals that the fundamental frequency depends on ER coil spacing and average EV speed. In the worst-case yet unlikely scenario that EVs move in a synchronized fashion, the amplitude of harmonics scales with the EV count. On the contrary, when EVs move freely, harmonics scale with the square root of the EV count. Platoon formations can accentuate harmonics. The spectral content around harmonics decreases in magnitude and increases in bandwidth with the harmonic index. The load of a single EV moving at a time-varying speed can be modeled as a frequency-modulated (FM) signal. Despite the simplifying assumptions, the derived models offer valuable insights for ER planners and grid operators. Dynamic simulations of a modified WECC model with DWPT loads synthesized from realistic EV trajectories and ER specifications corroborate some of these insights.

eess.SY↗

arXiv2025

Dynamic Modeling of Load Demand in Electrified Highways Based on the EV Composition

Electrified roadways (ERs) equipped with the dynamic wireless power transfer (DWPT) technology can achieve longer driving range and reduce on-board battery requirements for electric vehicles (EVs). Due to the spatial arrangement of transmitter (Tx) coils embedded into the ER pavement, the power drawn by the EV's receiver (Rx) coil is oscillatory in nature. Therefore, understanding the dynamic behavior of the total DWPT load is important for power system dynamic studies. To this end, we model the load of individual EVs in the time and frequency domains for constant EV speed. We establish that a nonlinear control scheme implemented in existing DWPT-enabled EVs exhibits milder frequency harmonics compared to its linear alternative. According to this model, the harmonics of an EV load decrease in amplitude with the Rx coil length. We further propose and analyze stochastic models for the total DWPT load served by an ER segment. Our models explain how the EV composition on the ER affects its frequency spectrum. Interestingly, we show that serving more EVs with longer Rx coils (trucks) does not necessarily entail milder harmonics. Our analytical findings are corroborated using realistic flows from a traffic simulator and offer valuable insights to grid operators and ER designers.

eess.SY↗

arXiv2023

Toward More Accurate and Robust Optimal Power Flow for Distribution Systems

The objective of this paper is to improve the accuracy and robustness of optimal power flow (OPF) formulations for distribution systems modeled down to the low-voltage point of connection of individual buildings. An approach for addressing the uncertain switching behavior of building loads(e.g., air conditioners, water heaters, or pool pumps) and variable renewable generation (e.g., rooftop solar) in the OPF is proposed. Rather than using time-averaged forecasts to determine voltage magnitude constraints, we leverage worst-case minimum and maximum forecasts of loads and distributed energy resource generation. Sensitivities of the power flow equations are used to predict how these deviations in load and renewable generation will impact system voltages, and the voltage constraints in the OPF are dynamically adjusted to mitigate voltage violations due to this uncertainty. A methodology for incorporating models of split-phase components and transformer core losses in the OPF formulation is also proposed. The proposed approach is validated through numerical case studies on a realistic distribution feeder using GridLAB-D, a distribution system simulation software.

eess.SY↗

arXiv2022

Physics-informed neural networks for solving parametric magnetostatic problems

The objective of this paper is to investigate the ability of physics-informed neural networks to learn the magnetic field response as a function of design parameters in the context of a two-dimensional (2-D) magnetostatic problem. Our approach is as follows. First, we present a functional whose minimization is equivalent to solving parametric magnetostatic problems. Subsequently, we use a deep neural network (DNN) to represent the magnetic field as a function of space and parameters that describe geometric features and operating points. We train the DNN by minimizing the physics-informed functional using stochastic gradient descent. Lastly, we demonstrate our approach on a \mbox{ten-dimensional} EI-core electromagnet problem with parameterized geometry. We evaluate the accuracy of the DNN by comparing its predictions to those of finite element analysis.

cs.CE↗