Artificially intelligent agents deployed in the real world must be able to reliably cooperate with humans (as well as other, heterogeneous AI agents). To provide formal guarantees of successful cooperation, we must make some assumptions about how these partner agents could plausibly behave. Realistic assumptions must account for the fact that other agents may be just as adaptable as our agent is. In this work, we consider the setting where an AI agent must cooperate with members of some target population of agents in a finitely repeated two-player general-sum game, where individual utilities are private. Two natural assumptions in this setting are 1) all agents in the target population are individually rational learners, and 2) when paired with another member of the population, with high-probability the agents will achieve the same expected utility as they would under some Pareto-efficient equilibrium strategy of the underlying stage game. Our theoretical results show that these assumptions alone are insufficient to select an AI strategy that achieves zero-shot cooperation with members of the target population. We therefore consider the problem of learning such a cooperation strategy using observations of members of the target population interacting with one another, and provide upper bounds on the sample complexity of learning such a cooperation strategy. Our main result shows that, under the above assumptions, these bounds can be much stronger than those arising from a “naive” reduction of the problem to one of imitation learning.

This paper aims to investigate the dynamic charging performance of an 11 kW dynamic inductive power transfer (DIPT) system. First, a multi-objective optimization (MOO) method is proposed to find the Pareto front of the DD charging pad. Then, the optimal design with a 96.82% efficiency is selected as the target design for the DIPT system. Based on the coupler mutual inductance at different misalignment, the orientation of the transmitter (Tx) and receiver (Rx) pads and the distance between Tx pads are studied and optimized. To obtain the dynamic characteristics of the DIPT system, the impact of mutual inductance variation is investigated, and a dynamic model using Laplace phasor transformation is built to solve the waveform amplitude of electrical variables. Finally, a time-variant circuit model is built. Based on the simulations, the dynamic model is proved to be accurate, and the proposed DIPT system displays a good dynamic charging performance.

Multiactive bridge (MAB) converter is a promising solution for integrating multiple renewable sources, storage, and loads for various applications. However, the MAB converter is challenging to control due to the inherent coupling between the port power flows. To that end, this article presents a decoupling control strategy based on linear active disturbance rejection control. The proposed controller observes the coupling disturbance using a linear extended state observer and subsequently rejects the observed disturbance resulting in dynamic decoupling. Experiments conducted on a 2-kW 100-kHz Si-C-based four-port MAB converter laboratory prototype illustrate the decoupling performance of the proposed control strategy. Compared to the traditional decoupling control strategy, the proposed approach is decentralized and model independent, only requiring information regarding its order.

Triple active bridge (TAB) as an isolated multiport converter is a promising integrated energy system for smart grids or electric vehicles. This article aims to derive and analyze zero voltage switching (ZVS) regions of TAB, in which both switching losses are reduced, and electromagnetic interference issues are mitigated. In the proposed closed-form solution of ZVS criteria, parameters such as the parasitic capacitance of the switches, the leakage inductance of the transformer, the switching frequency, the port voltage, the phase-shift inside and between the full-bridges are all taken into account. The analysis shows how the five degrees of freedom can be used to maintain ZVS operation in various operating points. The analysis and derived closed-form ZVS criteria are experimentally verified using a laboratory prototype. The derived analytical ZVS criteria are a powerful tool to study and optimize the operation of TAB converters.

This paper proposes a power flows decoupling controller for the triple active bridge converter. The controller is based on a full-order continuous-time model of the TAB converter derived using the generalized average modelling (GAM) technique. GAM uses the Fourier series expansion to decompose the state-space variables into two components, which represent the active power and the reactive power. The controller uses the active power components of the transformer currents to decouple the active power flows between converter ports. Additionally, the implementation of the decoupling controller in the digital domain is detailed in the paper. The decoupling performance of the proposed controller is validated in a hardware experiment.

Inductive power transfer (IPT) is becoming increasingly popular in stationary electric vehicle charging systems. In this paper, the influence of the different IPT coupler geometries on the performance factors efficiency, power density, misalignment tolerance, and stray field is studied. Five different cou-pler topologies namely the circular, rectangular, double-D (DD-DD) and the double-D transmitter with double-D-Quadrature receiver (DD-DDQ) are considered in this study. The electromagnetic behavior of the couplers is modeled using three-dimensional finite element analysis. To ensure a fair quantitative comparison, a multi-objective optimization framework is developed to analyze the Pareto trade-offs between conflicting performance metrics like power densities, efficiencies, and misalignment tolerance for all the considered coupler topologies.

With the rising popularity of the power electronic based systems with integrated energy storage, the multi-port isolated converter topologies are gaining popularity. In contemporary literature, the triple active bridge (TAB) converter is the most popular among these topologies. However, the TAB was not yet described with the continuous-time full-order model. In this paper, the continuous-time full-order model of the TAB converter is derived. The derived model is validated with the measurement of the control-to-output transfer functions. The derived model can provide useful insights into the operation of the converter and can be used for controller design.

Inductive power transfer (IPT) is becoming increasingly popular in stationary electric vehicle (EV) charging systems. In this paper, the influence of the different IPT coupler geometries on the performance factors such as efficiency, power density, misalignment tolerance, and stray field is studied. Four different coupler topologies, namely the circular, rectangular, double-D (DD-DD), and the double-D transmitter with double-D-quadrature receiver (DD-DDQ) are considered in this study. The electromagnetic behavior of the couplers is modeled using three-dimensional finite-element method, which is validated by experiments on a laboratory prototype. A multi-objective optimization (MOO) framework is developed to analyze the Pareto tradeoffs between conflicting performance metrics for the couplers. Optimization results depict that the circular topology performs best among the selected topologies regarding higher coupling coefficient, and efficiency for similar active mass and coupler area. Circular and rectangular couplers perform better than the polarized couplers like DD-DD and DD-DDQ regarding stray field exposure in both vertical and lateral direction of the coupler position in the EV. However, polarized couplers show more tolerance toward misalignment compared to circular and rectangular couplers. Thus, this study provides information regarding the specific strengths and weaknesses of different coupler topologies, which can be used during the initial design phase.

Microgrid with integrated photo-voltaics (PV) and battery storage system (BSS) is a promising technology for future residential applications. Optimally sizing the PV system and BSS can maximise self-sufficiency, grid relief, and at the same time can be cost-effective by exploiting tariff incentives. To that end, this paper presents a comprehensive optimisation model for the sizing of PV, battery, and grid converter for a microgrid system considering multiple objectives like energy autonomy, power autonomy, payback period, and capital costs. The proposed approach involves developing a holistic techno-economic microgrid model based on variables like PV system power, azimuth angle, battery size, converter ratings, capital investment and electricity tariffs. The proposed method is applied to determine the optimum capacity of a PV system and BSS for two case residential load profiles in the Netherlands and Texas, US to investigate the effect of meteorological conditions on the relative size of PV and battery. Based on the optimisation results, thumb rules for optimal system sizing are derived to facilitate microgrid design engineers during the initial design phase.

Autonomous operation of the dc grids with converter interfaced renewable energy sources and energy storage with droop based control can lead to instability. This paper analyzes the stability of droop based closed loop controllers in a dc nanogrid. Linear state space modeling approach is used to model the small signal model of the droop controlled dc-dc converters. The dominant eigenvalues are analyzed, and the effect of closed-loop gains of the converters are investigated. Detailed parametric sensitivity analysis and participation factor of the system parasitics and the controller gains are also presented. Based on that, a segmented droop strategy is proposed to divide the operating ranges into segments with adaptive controller gains to ensure system stability in all of them. A particle swarm optimization algorithm is used to optimize for the converter gains of individual segments.

Low voltage dc distribution grids face issues associated with arc faults, aggravated by the absence of current zero crossing. The focus of this paper is to comprehensively develop a method of series arc fault detection at the load side power electronics, based on the electrode dependent initial voltage drop occurring at the arc initiation. The proposed arc detection algorithm is described along with the structure and time constants of the designed bandpass filter. The operational boundaries of the arc detection algorithm are defined for copper electrodes depending on the set threshold voltage and the system parameters, like grid inductance, resistance, and the load capacitance. Further, the detection time and the zone of guaranteed positive detection are depicted. These are validated through test simulations on the state space model of the system. Finally, experimental validation of the proposed scheme is carried out, wherein, a series arc is generated in the dc circuit and the programmed micro-controller provides a real time signal upon detecting the arcing event. The results on variation in detection time with set threshold voltage are also presented experimentally.

Inductive power transfer (IPT) is gaining popularity across a wide range of battery charging applications like biomedical, consumer electronics and electric vehicle (EV) charging. One of the major challenges in designing IPT charge pads is determining the optimal physical sizes of the magnetic couplers resulting in efficient power transfer and low cost of materials. In EV applications, it is especially difficult due to the variation in nominal air gap, required power levels associated with different vehicle classes, and charging locations that may be encountered. This paper aims to determine the relationship between optimal coupler sizes and the nominal air gap of an IPT system. Finite element analysis (FEA) is used to model the electromagnetic behavior of the magnetic couplers. A multi-objective optimization framework is built to reveal the Pareto fronts which show the trade-offs between the power transfer efficiencies and the coupler power densities at different air gaps. This method is applied on polarized double-D (DD) couplers for a 5 kW IPT system at different air gaps. Analyzing the power densities of the Pareto Optimal designs an approximate relation between optimal pad sizes and the air gap is derived. Results show that there is an exponential relationship between the optimal coupler sizes and the nominal air gap.

This paper deals with a generic methodology to evaluate the magnetic parameters of contactless power transfer systems. Neumann's integral has been used to create a matrix method that can model the magnetics of single coils (circle, square, rectangle). The principle of superposition has been utilized to extend the theory to multi-coil geometries, such as double circular, double rectangle and double rectangle quadrature. Numerical and experimental validation has been performed to validate the analytical models developed. A rigorous application of the analysis has been carried out to study misalignment and hence the efficacy of various geometries to misalignment tolerance. The comparison of single-coil and multi-coil inductive power transfer systems (MCIPT) considering coupling variation with misalignment, power transferred and maximum efficiency is carried out.

Misalignment during wireless charging of EVs can lead to low efficiencies (<85%) of power transfer which can lead to thermal issues and high leakage fields. This paper explores the most optimal solution towards a misalignment tolerant system. To that end, a DD-DDQ coil system with series-series compensation combined with an impedance matching control algorithm is explored. A multi-objective optimization approach is presented to visualize the trade-off between the design aspects which result in high misalignment tolerance. Finally, the results are compared with DD-DD coil designs to quantify the advantage of the proposed system.