This work investigates the modeling of lithium-ion batteries (LIBs) with a recurrent neural network (RNN), rather than with an equivalent circuit or similar type model as is typically used. The RNN is trained with dynamic battery data, such as vehicle drive cycle test results. Specialized characterization tests and model parameterization is not necessary, simplifying the process of battery modeling. A compact unified methodology consisting of an RNN with gated recurrent unit (GRU) and deep feature selection (DFS) structures is utilized. Two RNNs are evaluated, one with current as the input and another with power as the input. Both RNN forms accurately model LIB dynamic responses including battery nonlinear behavior at different temperatures. The models are compact in size, require fewer characterization tests compared to conventional equivalent circuit models (ECMs), and can be further used as an LIB simulator in model-based design and hardware-in-loop (HIL) applications to test battery management systems (BMS) and other electronic components.
9 July 2021 in IEEE Transactions on Industry Applications
Four-quadrant (FQ) switches that have both bidirectional current and reverse-voltage blocking (RVB) capabilities are appealing candidates for use in current-source inverters (CSI) to significantly reduce the switch conduction losses. However, the FQ characteristics of these switches can create problems during switching events by causing transient overcurrent pulses that damage the inverter. Commutation methods for FQ switches in CSI topologies that address these challenges are presented in this article. Several key aspects of the performance and implementation of these commutation methods are explored for a high-power density 3 kW integrated motor drive incorporating both a current-source active rectifier and inverter stage that has been developed using these FQ switch methods. Experimental results demonstrate the benefits of implementing these commutation methods in CSIs that use FQ switches. The benefits include substantial efficiency improvement compared to the conventional CSI implemented with series power diodes as well as reduced mass and volume of passive components due to the high switching frequency (100 kHz).
13 June 2018 in IEEE Transactions on Industry Applications
This paper presents an improved analytical model for estimating the high-frequency current ripple of interior permanent magnet (IPM) synchronous machines due to PWM switching. The proposed model accounts for the impact of slotting effect, magnetic saturation, and cross-coupling between the d- and q-axis. The model is subsequently used to investigate several factors that influence the PWM-induced current ripple. These include the PWM switching frequency, fundamental frequency (i.e., machine speed), dc-bus voltage, current control angle (i.e., γ angle), and the excitation current amplitude (i.e., saturation level). Experiments have been conducted to verify the analytical prediction results. These results show that the analytical model can predict the PWM-induced current ripple waveshape very well for many operating conditions and accurately estimate its rms value over a fundamental period.
A new magnetic circuit model is presented for an interior permanent magnet (IPM) synchronous machine, using a machine with three-phase distributed stator windings and three layers of flux barriers in the rotor as an example topology. The model accounts for: i) the effects of cross-coupled magnetic saturation caused by the salient rotor; ii) variation of magnetic saturation levels in the iron rotor bridges that are key elements of the unitary rotor laminations; iii) the effects of stator lamination slots on the airgap mmf distribution; and iv) the local variation of airgap permeance due to the stator slotting and the relative position of the rotor with respect to the stator. As a result of these features, the new model is capable of significantly improving the accuracy of electromagnetic performance predictions for aggressively-designed IPM machines compared to previously-available magnetic circuit models. Comparisons with finite-element analysis and measurement results are provided showing that the new model is much faster while delivering appealing accuracy compared to the FE method.
7 June 2023 in IEEE Journal of Emerging and Selected Topics in Power Electronics
The fight against global climate change demands the electrification of aircraft propulsion. One of the most significant and distinct challenges that future multi-megawatt electrified propulsion will face is high voltage at high altitudes, which could cause significant partial discharge and reduce the reliability of the propulsion system. Recognizing this challenge, a multi-disciplinary research team has proposed, designed, built, and preliminarily tested a 2-kV 1-MW 20,000 RPM integrated modular motor drive (IMMD) with an unprecedented weight power density of 9 kW/kg. This paper presents the architecture, major design approaches, test results of associated power electronics systems at the rated voltage and current, and preliminary test results of the full IMMD prototype. Full power and system-level altitude tests of the IMMD are already planned at the National Aeronautics and Space Administration (NASA’s) test facility.
8 February 2022 in IEEE Transactions on Industry Applications
Conventional current-source inverters (CSIs) using silicon-based reverse-blocking (RB) switches typically suffer from high conduction loss, low switching frequency, and bulky size. New bidirectional (BD) switches (also called four-quadrant switches) built from wide bandgap semiconductor materials such as gallium nitride and silicon carbide offer both reverse-voltage-blocking capability and low conduction loss that make them appealing candidates for increasing the CSIs efficiency. This article proposes a new modulation scheme for an emerging three-phase CSI topology (H7-CSI) to enable it to operate safely with BD switches to achieve higher efficiency (as much as 3%) compared to the conventional H6-CSI topology or the emerging H7-CSI using RB switch. In addition, this modulation strategy combined with the proposed H7-CSI-BD topology is capable of significantly reducing the conducted common-mode (CM) EMI up to 20 dB⋅μV. Based on this improvement, a promising CM EMI filter network consisting of a CM choke and Y-capacitor has been designed for the H7-CSI-BD topology to meet EMI standards. The required filter component values for the H7-CSI-BD topology are significantly lower than those required in conventional CSIs. Analysis, simulation, and experimental results are provided that confirm the advantages of the proposed BD switch-enabled H7-CSI over conventional H6-CSI and emerging H7-CSI topologies using RB switches.