Search

The Soft-Switching Solid-State Transformer (S4T) is a flexible, multi-port, and isolated ZVS soft-switching power converter with emerging applications in multi-port data center power delivery, renewable energy integration in EV fast charging stations, and storage integration in next generation solar farms. As a current source converter, the S4T requires reverse blocking (RB) semiconductor devices with bidirectional voltage blocking capability and unidirectional current flow. While these switches had been implemented from the series connection of discrete devices in early S4T prototypes, larger scale projects require a more manufacturable, cost-effective, robust, and compact approach. This work presents the first device-level analysis of the soft-switching principles of the S4T in a commercial, high current 1.7 kV RB switch phase-leg module in which each switch position consists of the series connection of a Si IGBT and a SiC Schottky diode. The hybrid RB module was extensively tested in a soft-switching current source inverter bridge, replicating the unique ZVS soft-switching environment of the S4T, and enabling measurement of conduction loss, residual switching loss, and dI/dt and dV/dt stress. With Si IGBTs, residual switching loss from IGBT tail current and commutation loop parasitic inductance limited the achievable turn-off energy reduction under soft-switching. Nevertheless, the ZVS soft-switching environment of the S4T enabled a > 88.8% reduction in turn-on loss and a > 55.2% reduction in turn-off loss at 800 V and 90°c cold plate temperature.

This paper presents two oversampling based control methods for a multiport low-inertia current-source soft-switching solid-state transformer (S4T). The traditional duty cycle or model predictive control calculates the reference currents and the duty cycle for each port, requiring accurate model prediction and sophisticated control design to balance the input and output power on a per cycle basis. To reduce the computational burden and enable the use of a simpler and more cost-effective microcontroller, two oversampling based control methods are proposed. This type of control exhibits unusual properties, such as unequal switching frequencies at individual ports and across the high frequency transformer. The proposed methods can be extended to cover more ports, simply by adding states in the state machine, thus demonstrating multi-variable control. These methods show promise in the implementation of control for very low-cost multiport converter systems.

While conventional reverse blocking (RB) switch structures consisting of an active switch (MOSFET or IGBT) and a series diode can enable high efficiency operation in high-voltage (>600 VAC and >1000 VDC), high-power (25-50 kVA) current source converters (CSCs), the high conduction losses associated with conventional RB structures have prevented widescale adoption of CSCs in low-voltage applications. This work presents the design and experimental validation of the patent-pending Synchronous Reverse Blocking Switch, a method of integrating a bidirectional dual-active-switch structure into the Soft-Switching Solid-State Transformer (S4T) topology to unlock high efficiency operation in low-voltage, high-current applications. By leveraging the operating principles of the S4T, whose control eliminates the possibility of shoot-through and provides zero-voltage switching (ZVS) conditions for all power devices across the entire load range, the Synchronous Reverse Blocking Switch has been demonstrated to offer significant conduction loss reductions while also mitigating the reverse recovery of the PN-junction MOSFET body diode which replaces the series diode of the conventional RB switch. Key contributions of this work include device-level validation of the reverse recovery mitigation mechanism of the Synchronous RB Switch, and system level validation through the design and performance measurement of a 48 VDC S4T bridge, which achieved 97.8% efficiency at a power delivery level of 750W and evidenced the strategy’s benign reverse recovery properties.

By 2030, electric vehicles (EVs) are estimated to represent more than 30% of new passenger vehicle sales. Most EVs today feature powertrains built around 400V lithium battery packs, with SOOV powertrains planned for new designs. Though the chemical and thermodynamic stability of lithium batteries have been extensively studied, the high voltage safety of EV batteries remains a pressing risk especially following accidents where on-board safety systems are compromised. As market penetration of EVs increases worldwide, the brunt of the high-voltage safety risk will be borne by passengers, first responders, manufacturing workers, and maintenance workers. In this work, the authors present an intrinsically safe and modular 48 volt DC powertrain architecture that meets and surpasses current EV power and performance specifications while delivering the additional benefits of low-voltage touch safety and reduced isolation requirements, providing flexibility in the design of the battery pack and maintaining high motor drive efficiency.

Optimizing complex industrial processes can help improve overall productivity, minimize down-times, and reduce operational costs. Traditionally plant managers have accomplished this by installing sub-metering devices for specific processes. However, sub-metering devices are expensive, require complex installation procedures, and need customization on the factory floor to get the data to the enterprise cloud. Furthermore, interpreting the acquired data and implementing suitable actions require significant subject matter expertise and a comprehensive understanding of the loads' energy consumption profiles and their exposure to power quality disturbances. As a result, this is often prohibitively expensive for small or medium scale facilities. This paper proposes a novel architecture for gaining actionable insights from the various processes being monitored. At the heart of the system is an ‘intelligent terminal block’ (ITB) that gathers process specific data such as voltage, current, and power consumption and performs edge computing in a decentralized fashion. From the acquired data, the ITB can help minimize downtime through fault diagnostics and predictive maintenance, provide situational awareness by recording power quality disturbances and their effect on equipment, and enable cost savings through energy efficiency insights like CVR factor estimation. The ITB itself requires minimal customization on the factory floor and uses existing communication back-hauls to report these insights to the cloud, capable of working with intermittent connectivity. This architecture allows plant managers to economically obtain significant insights into the plant's operation.

The prototype fuel cell electric vehicle (FCEV) Maxwell currently holds the Guinness World Record for "Most Fuel-Efficient Vehicle" at 6,196 km/l (14,573 MPGe). The vehicle draws less than 21W of electrical power at 25 km/h. This work outlines the powertrain architecture and presents the vehicle's sources of electrical and mechanical losses, quantified through rigorous experimentation. The individual components are modelled and combined into a total model of performance representing all energy flow in the vehicle. The total model is found to have excellent predictive power with only 4.3% error compared to measured on-track performance.