Triphase (Belgium)

A comprehensive procedure for the derivation of optimal, full-operating-range zero voltage switching (ZVS) modulation schemes for single-phase, single-stage, bidirectional and isolated dual active bridge (DAB) ac-dc converters is presented. The converter topology consists of a DAB dc-dc converter, receiving a rectified ac line voltage via a synchronous rectifier. The DAB comprises primary and secondary side full bridges, linked by a high-frequency isolation transformer and a series inductor. ZVS modulation schemes previously proposed in the literature are either based on current-based or energy-based ZVS analyses. The procedure outlined in this paper for the calculation of optimal DAB modulation schemes (i.e., combined phase-shift, duty-cycle, and switching frequency modulation) relies on a novel, more accurate, current-dependent charge-based ZVS analysis, taking into account the amount of charge that is required to charge the nonlinear parasitic output capacitances of the switches during commutation. Thereby, the concept of “commutation inductance(s)” is shown to be an essential element in achieving full-operating-range ZVS. The proposed methods are applied to a 3.7 kW, bidirectional, and unity power factor electric vehicle battery charger which interfaces a 400 V dc-bus with the 230 Vac, 50-Hz utility grid. Experimental results obtained from a high-power-density, high-efficiency converter prototype are given to validate the theoretical analysis and practical feasibility of the proposed strategy.

Peak torque and power density requirements for traction motor drives continue to increase, while demands on reliability are getting increasingly stringent as well. With the knowledge that most of the failure mechanisms are related to excessive temperature (cycling), thermal management is a key for increasing performance, without jeopardizing reliability. This paper proposes a control strategy for active thermal management of permanent magnet synchronous motor (PMSM) drives, based on real-time estimation and feedback of switching device and motor temperatures. By regulating the switching frequency and current control limit, critical components can be safeguarded from excessive temperature rise. Furthermore, optimal dq-current control vectors are calculated within the temperature and voltage constraints, to maximize the drive's efficiency and speed-torque envelope. Hence, the control strategy enables the drivetrain to operate safely at maximum attainable performance limits. The strategy is experimentally validated on an 11-kW PMSM drive for a number of representative vehicle loads, including a maximum standstill torque test, a maximum acceleration test, and a driving cycle test.

This paper proposes an algorithm for optimal current trajectory control of (interior) permanent magnet synchronous motor (PMSM) drives, considering the motor/inverter current and voltage limitations. In contrast to common voltage feedback flux-weakening strategies, it applies a differential rather than a conventional integral approach to regulate the current vector. Linearized motor equations are used to express torque and voltage amplitude changes relative to the actual operating point as a function of dq-current increments. Optimal dq-current step adjustments are calculated from these approximate voltage/torque increment lines according to machine state feedback. This allows for the constraint optimization problem to be solved in real time. Each iteration, the dq-current setpoint vector is incrementally shifted in an optimal direction, rapidly converging to a solution within the voltage/current limits, which minimizes error on torque request with maximum efficiency. Extensive experimental results on an interior PMSM setup demonstrate the strategy's dynamic performance and robustness.

This paper discusses novel communication network topologies and components and describes an evolutionary path of bringing Ethernet into automotive applications with focus on electric mobility. For next generation in-vehicle networking, the automotive industry identified Ethernet as a promising candidate besides CAN and FlexRay. Ethernet is an IEEE standard and is broadly used in consumer and industry domains. It will bring a number of changes for the design and management of in-vehicle networks and provides significant re-use of components, software, and tools. Ethernet is intended to connect inside the vehicle high-speed communication requiring sub-systems like Advanced Driver Assistant Systems (ADAS), navigation and positioning, multimedia, and connectivity systems. For hybrid (HEVs) or electric vehicles (EVs), Ethernet will be a powerful part of the communication architecture layer that enables the link between the vehicle electronics and the Internet where the vehicle is a part of a typical Internet of Things (IoT) application. Using Ethernet for vehicle connectivity will effectively manage the huge amount of data to be transferred between the outside world and the vehicle through vehicle-to-x (V2V and V2I or V2I+I) communication systems and cloud-based services for advanced energy management solutions. Ethernet is an enabling technology for introducing advanced features into the automotive domain and needs further optimizations in terms of scalability, cost, power, and electrical robustness in order to be adopted and widely used by the industry.

A switching control strategy to enable Zero-Voltage-Switching (ZVS) over the entire input-voltage interval and the full power range of a single-stage Dual Active Bridge (DAB) AC/DC converter is proposed. The converter topology consists of a DAB DC/DC converter, receiving a rectified AC line voltage via a synchronous rectifier. The DAB comprises primary and secondary side full bridges, linked by a high-frequency isolation transformer and inductor. Using conventional control strategies, the soft-switching boundary conditions are exceeded at the higher voltage conversion ratios of the AC input interval. Recently we presented a novel pulse-width-modulation strategy to fully eliminate these boundaries, using a half bridge — full bridge DAB configuration. In this papers the analysis is extended towards a full bridge — full bridge DAB setup, providing more flexibility to minimize the component RMS currents and allowing increased performance (in terms of efficiency and volume). Experimental results are given to validate the theoretical analysis and practical feasibility of the proposed strategy.

Power density and reliability specifications for motor drives in traction applications are getting increasingly stringent. The main challenge in meeting these conflicting requirements, is managing heat dissipation. A drive's peak torque rating is limited by switching device temperatures which must be kept below critical values at all times for the sake of reliability, preferably without major hardware adaptations. In this challenge lies a large potential for advanced control algorithms. This paper proposes a PMSM drive control strategy which combines active thermal management with dynamic DC-link voltage adaptation. The bus voltage level is adjusted to the required PMSM terminal voltage in each operating point. Doing so, switching losses can be reduced at low speed by lowering the bus voltage. At high speed, the voltage level is boosted and field-weakening operation and the associated additional losses are avoided. An 11 kW PMSM drive, with an active front-end controlling the bus voltage, is used as a test setup to mimic a series-hybrid drivetrain. Compared to a fixed DC-link voltage, efficiency maps show a significant inverter loss reduction at low speed. This results in lower switching device temperatures which in turn allows a higher peak torque rating.

III-Nitride materials are very promising to be used in next-generation high-frequency power switching applications. In this letter, we demonstrate the performance of normally off AlGaN/GaN/AlGaN double-heterostructure FETs (DHFETs) using a boost-converter circuit. The figures of merit of our large (57.6-mm gate width) GaN transistor are presented: <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$R_{\rm ON} \ast Q_{G}$</tex></formula> of 2.5 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$\Omega\cdot\hbox{nC}$</tex></formula> is obtained at <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$V_{\rm DS} = \hbox{140\ V}$</tex></formula> . The switching performance of the GaN DHFET is studied in a dedicated high-frequency boost converter: both the switching times and power losses are characterized. We show converter efficiency values up to 96.1% at 500 kHz and 93.9% at 850 kHz at output power of 100 W.

A semi-analytical modulation scheme to operate a single-phase, single-stage dual active bridge (DAB) AC-DC converter under full-operating-range zero voltage switching (ZVS) is proposed. The converter topology consists of a DAB DC-DC converter, receiving a rectified AC line voltage via a synchronous rectifier. ZVS modulation strategies previously proposed in literature are either based on current-based (CB) or energy-based (EB) ZVS analyses. The combined phase-shift, duty-cycle, and switching frequency modulation proposed in this paper relies on a novel, current-dependent charge-based (CDCB) ZVS analysis, taking into account the commutation charge of the (parasitic) switch capacitances as well as the time dependency of the commutation currents. Thereby, commutation inductance is shown to be an essential element in achieving full-operating-range ZVS. Experimental results obtained from a 3.7 kW bidirectional electric vehicle battery charger which interfaces a 400 V DC-bus with the 230 Vac, 50 Hz utility grid are given to validate the analysis and practical feasibility of the proposed strategy.

This paper presents the power electronic platform used for the VSYNC project. In this project, inverters are controlled in such a way as to exhibit a virtual rotational inertia towards the grid, in order to limit grid frequency variations in grids containing a high share of inverter-connected DER. First the layout and operation of the platform are described in detail, showing its versatility for research purposes. Next the performance of the platform is illustrated using experimental results obtaining using a grid-connected inverter in a laboratory setup.

For realizing bidirectional and isolated AC/DC converters, soft-switching techniques/topologies seem to be a favourable choice as they enable a further loss and volume reduction of the system. Contrary to the traditional dual-stage approach, using a power factor corrector (PFC) stage in series with a DC/DC isolation stage, we showed recently that the same functionality can be achieved under full soft-switching operation using a single-stage dual active bridge (DAB) AC/DC converter. This paper investigates the performance of this single-stage approach by comparing it with a state-of-the-art conventional dual-stage concept (both soft-switching converters), where a bidirectional interleaved triangular current mode (TCM) PFC rectifier was chosen in combination with a DAB DC/DC converter. The advantages and drawbacks of each concept are discussed in detail, focusing on the impact of the utilized semiconductor technology and silicon area on the converter efficiency. Furthermore, a comprehensive comparison of power density is allowed by the analytical models that correlate the component losses with their respective volume.

A switching control strategy to extend the zero-voltage-switching (ZVS) operating range of a Dual Active Bridge (DAB) AC/DC converter to the entire input-voltage interval and the full power range is proposed. The converter topology consists of a DAB DC/DC converter, receiving a rectified AC line voltage via a synchronous rectifier. The DAB comprises a primary side half bridge and secondary side full bridge, linked by a high-frequency isolation transformer and inductor. Using conventional control strategies, the soft-switching boundary conditions are exceeded at the higher voltage conversion ratios of the AC input interval. A novel pulse-width-modulation strategy to fully eliminate these boundaries and its analysis are presented in this paper, allowing increased performance (in terms of efficiency and stresses). Additionally, by using a half bridge / full bridge configuration, the number of active components is reduced. A prototype converter was constructed and experimental results are given to validate the theoretical analyses and practical feasibility of the proposed strategy.

A boost converter was constructed using a high voltage enhancement mode (E-mode) AlGaN/GaN/AlGaN DHFET transistor grown on Si<;111>. The very low dynamic onresistance (R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dyn</sub> ≈ 0.23 Ω) and very low gate-charges (e.g. Q <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">gate</sub> ~15 nC at Vos = 200 V) result in minor transistor losses. Together with a proper design of the passive components and the use of SiC diodes, very high overall efficiencies are reached. Measurements show high conversion efficiencies of 96.1% (P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">out</sub> = 106 W, 76 to 142 V at 512.5 kHz) and 93.9% (P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">out</sub> = 97.5 W, 78 to 142 V at 845.2 kHz). These are, to our knowledge, the highest efficiencies reported for an enhancement mode GaN DHFET on Si in this frequency range. The transistor switching losses are concentrated in the turn-on interval, and dominate at high frequencies. This is due to a limited positive gate-voltage swing, as the gate-source diode restricts the positive drive voltage.

In traction applications, electrical drivetrain components are subjected to unpredictable load and temperature variations depending on the driving cycle and ambient conditions. As performance and power density requirements are getting increasingly stringent, the power electronic devices and electromagnetic actuators are stressed heavily due to temperature cycling effects and face the risk of overheating, compromising lifetime and reliability. To protect the drivetrain from thermally induced failure, a model-based thermal management strategy is proposed in this paper. Critical component temperatures are calculated online with a combined loss and thermal model and are limited progressively by applying constraints to loss-influencing operating variables. Starting from the requested torque, the dq-current setpoint calculation is formulated as a constraint optimization problem in order to protect all drivetrain components while maximizing overall efficiency over the entire torque-speed operating range, including field weakening at elevated speed. Unlike conventional approaches, which are often ad-hoc or based on de-rating, the proposed strategy allows the drivetrain to operate safely at maximum performance limits, without unnecessarily degrading performance.

This paper deals with the losses of Switched Reluctance Machine Drives (SRDs) for electric vehicle propulsion, consisting of the losses in the Switched Reluctance Machine (SRM) and the converter, across some operating regions based on different control strategies, namely Firing Angle (FA) control and Torque Sharing Function (TSF) control. The SRM model which takes into account copper losses and converter losses is obtained through MATLAB/Simulink. The losses and efficiencies in the machine, the converter and the SRD as function of speed and average torque are compared for a 30 kW peak 8/6 SRD.

This work deals with a frequency-domain homogenization technique for litz-wire bundles in transformers. The approach consists in adopting a frequency-dependent complex reluctivity in the litz-wire bundles and a frequency-dependent complex impedance in the electrical circuit. The litz-wire bundles become homogeneous conductors which are easy to integrate into a finite element model of a transformer. The magnetic flux and the impedance of the litz-wire transformer computed from the homogenized model agree well with the reference fine model in which each litz-wire turn is finely discretized. The errors of the computed resistance and inductance values are less than 3% and 0.03% respectively with several times faster calculation.

Lipid metabolism and citric acid production by <italic>Yarrowia lipolytica</italic> are metabolically inter-dependent and depletion in citric acid production by proper media development can increase lipid accumulation in <italic>Y. lipolytica</italic> upto 72% of biomass dry weight.

This work deals with a frequency-domain homogenization method for litz-wire bundles embedded in a finite element (FE) model. The approach consists in adopting a frequency-dependent complex reluctivity in the litz-wire bundles and a frequency-dependent complex impedance in the electrical circuit in terms of dimensionless coefficients. The litz-wire bundles become homogeneous conductors which are easy to integrate into a finite element model. The homogenization methods in previous works are also reviewed. The complex permeability achieved with our model is compared to the most accurate expression found in literature. The agreement is very good despite the different elementary cell and condition setting. Moreover, the homogenization method in this paper is validated for both 2-D and 3-D FE calculations with the translational-symmetry transformer model and the axisymmetric inductor model respectively. The reference solutions of those models are computed by finely discretizing each conductor inside the litz-wire bundle. The results of the computed resistance and inductance agree well with those computed by the reference fine model, resulting in less than 3% and 0.05% maximum errors respectively in 2-D FE model, and 4% and 1.8% respectively in 3-D FE model with several times less computational cost.

This paper presents a novel structured design method for actively-damped, lossless, LC-ladder filters of general order. The method builds on the theory of ladder network synthesis and on a control algorithm actively emulating the termination of a ladder network. Given a desired reference transfer function, a normalized ladder network is synthesized and scaled to meet the physical requirements. As such, it simultaneously tackles the design and sizing of the filter, and the active damping of filter resonances. The associated digital-control architecture is also elaborated. The method is applied to designing a 3 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">rd</sup> -order LCL filter, of which the experimental results show properly damped responses. Moreover, the design of a 5 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> -order LCLCL filter is introduced and the result is promising for the applications requiring high bandwidth and/or high attenuation.

In order to mitigate economical, ecological and technological challenges aircraft manufacturers have shifted towards more electrical aircraft, bringing with it many new technologies to be implemented and a redesign of the existing electrical network. This paper details the development of an electrical distribution centre as taking place under the European Framework 7 Clean Sky Program. The developed distribution centre is to be integrated into an existing electrical test bench which recreates the entire electrical system of an aircraft and allows both aircraft and aerospace equipment manufactures to conduct verification activities over different aircraft electrical architectures. The task then of the distribution centre is to bring an element of flexibility to the existing electrical test bench in order for it to be easily reconfigured into any aircraft architecture, including both airplane and rotorcraft. In order to achieve this, the distribution centre contains both low voltage and high voltage DC bus bars. The connection of sources and loads to the bus bars, and the interconnection between bus bars themselves, are achieved by using Racks, each of which contains a controllable and sequenceable switching element together with high fidelity proprietary measurement instrumentation and embedded protection. The entire system has undergone in-house testing and successfully passed factory acceptance tests before its delivery and subsequent commissioning, with the results being noted and discussed here.

This paper presents a structured design method for half-bridge converters with higher-order lossless output filters, focusing on the popular LCL filters. The method builds on the theory of terminated ladder networks and on a control algorithm actively emulating the ladder network's termination. Given a desired transfer function, a normalized ladder network is accordingly synthesized and scaled to meet the physical requirements. The component sizing and the active damping of filter resonances are simultaneously tackled. The associated digital-control architecture for LCL filters is explained. A Kalman observer is employed to deal with the inherent sample delay. An LCL filter is designed based on the proposed method and is experimentally validated with the half-bridge converters operating as active front end. The grid reactive currents are stepped and the results show properly damped responses. The system has good disturbance rejection and is passive according to the simulated input admittance. The robustness of the system is also demonstrated by numerical sensitivity analysis. These excellent characteristics can facilitate the design of the upper-layer or incorporated outer-loop controller, and can enhance the performance of the complete system.