This article investigates the resilient consensus problem in open multi-agent systems (OMASs), characterized by dynamic agent arrivals and departures where the total population is time-varying and unknown a priori. Existing resilient consensus algorithms typically rely on the assumption of a static network robustness condition, which is frequently violated due to the inherent openness of OMASs. To overcome this limitation, a fully distributed topology reconfiguration approach leveraging the properties of circulant digraphs is proposed, guaranteeing that the open system consistently maintains the required level of network robustness. Building upon this framework, a novel open resilient consensus algorithm is developed to ensure state agreement despite continuous agent churn. Furthermore, we establish and analyze the sufficient conditions for achieving resilient consensus under both $\boldsymbol{F}$ -local and $\boldsymbol{f}$ -fraction local attack models, explicitly deriving the lower bounds for the parameter $\boldsymbol{k}$ of the reconfigured circulant graphs. Finally, numerical simulations and hardware-in-the-loop experiments on a Raspberry Pi platform are presented to validate the theoretical correctness and practical effectiveness of the proposed approaches in dynamic environments.
This article proposes a virtual synchronous current source (VSCS) control for the LCL-type grid-connected inverter in an inductive power grid. To achieve this objective, the duality relationship between voltage and current sources is analyzed first. Regarding the external characteristics, the relationship between active power and angular frequency in a voltage source is dual to that between negative reactive power and angular frequency in a current source. Therefore, by replacing active power with negative reactive power, the synchronous control for the voltage-type inverter can be transformed into the synchronous control for the current-type inverter. Then, a VSCS control is proposed based on the above correspondence, which is dual to the virtual synchronous generator (VSG) control. The proposed control can replace the phase-locked loop (PLL) and improve the applicability of the current-type inverter to weak grids. Finally, a comparison with existing methods is conducted. The inverter with the proposed control is stable under strong, weak, and even isolated grids in the verification. Meanwhile, the inertia and damping can be set like the VSG control. The aforementioned analyses are validated by both simulation and experimental results.
The lithium-ion batteries (LIBs) applied in electric vehicles (EVs) pose hazard risks. When LIBs undergo electrochemical distortion, the battery cell near the abnormal cell is propagated, and it causes the application to sustain severe accidents. To address this safety issue, it is important to diagnose a fault in battery pack in advance. Here, we recommend a novel technique to detect battery faults by a signal processing algorithm applied to real driving profile to determine the electrochemical property from electrical signals. Novel indicators related to battery faults physically and electrochemically are selected, extracted and analyzed to be used for diagnosis. Furthermore, it expands to the module level from the cell level, particularly a real battery sub-module installed in IONIQ 5. In addition, various conditions, e.g., temperature, charge/discharge current rate (C-rate), and aging, are employed for representing the normal conditions and for abnormal, fault-induced data. A statistical method for selecting the threshold is adopted and validated for diagnosis between normal and fault case from abnormal fault to operating fault. Furthermore, an imbalanced module, overcharge module and normal module is implemented to verify the state-of-safety (SOS) diagnosis. As a result, the designed algorithm satisfies the diagnosis performance.
The second-order generalized integrator (SOGI)-based weighted grid-voltage feedforward (WGVF) scheme enables grid-connected inverters (GCIs) to balance harmonic suppression and weak-grid stability by independently weighting the selected grid-voltage harmonics. However, the popular impedance-ratio-based design cannot directly determine the weighting coefficient for a prescribed stability margin because the weighting coefficient, stability margin, and impedance-intersection frequency are implicitly related. To address this issue, this article proposes an admittance-circle stability criterion for the SOGI-based WGVF. Owing to the resonant terms in the SOGI-based WGVF, the Nyquist locus of the sum of the grid admittance and the GCI output admittance is approximately circular within the neighborhood frequency range of each selected harmonic frequency. Based on this circular-locus characteristic, the origin-encirclement condition can be assessed geometrically, and an admittance-circle stability margin is defined. A closed-form relationship between the margin and the weighting coefficient is derived, enabling direct calculation of the weighting coefficient based on the grid admittance and the prescribed margin. Experimental results from a 6-kVA prototype are provided to verify the proposed stability criterion and weighting-coefficient design method.
This article presents an interleaved resonant series–parallel switched-capacitor (RSPSC) converter that effectively utilizes the parasitic inductance of the input cable. Unlike conventional resonant switched-capacitors (ReSC), which require bulky discrete inductors for resonance, the proposed topology incorporates the inherent cable inductance into the resonant network on the input side while halving the energy-storage requirement of the internal output inductor. By employing a series–parallel structure that maximizes the energy utilization of the input inductance, the proposed converter minimizes the overall passive component volume. In the proposed topology, all capacitors achieve soft-charging through the input inductor and soft-discharging via the output inductor. The proposed architecture is theoretically evaluated through a comprehensive comparison with existing RSPSC topologies, in terms of passive component utilization, output impedance, normalized switch stress, and output voltage ripple. A hardware prototype for 4-to-1 universal serial bus power delivery (USB-PD) programmable power supply (PPS) battery charging applications was implemented for experimental verification. The prototype achieved a peak system efficiency of 98.7% and a full-load efficiency of 96.3% at 50 A (20-to-5 V conversion), both inclusive of gate driver losses. Additionally, the prototype exhibited a power density of 5,058 W/in3, considering the total volume of the power stage and gate driving circuitry.
