This paper proposes a hybrid digital twin framework that couples a real-time physics-based digital twin model with a data-driven diagnostic layer implemented through cloud-based data acquisition and analysis. This framework generates synthetic datasets across multiple speed levels and fault severities for bearing fault detection and classification in industrial spindle systems, where real fault recordings are costly, risky, and difficult to reproduce. Once the system is validated, a two-stage classifier is trained and used for online fault detection and fault-type identification, whereas the deep-sequence model provides offline verification. To improve robustness, training data are enhanced with multi-domain feature enrichment and targeted data augmentation techniques that simulate measurement noise and small operating variations. The resulting models achieved strong performance under previously unseen operating conditions within the validated digital twin envelope. Overall, the proposed approach reduces the dependence on real fault experiments by enabling the risk-reduced development and evaluation of data-driven bearing fault diagnosis.

This paper proposes an innovative model-free deep reinforcement learning-based controller (RL-C) for a grid-connected 5-level packed-U-cell (PUC5) multilevel inverter (MLI). The controller is designed to deliver a high-quality grid current while maintaining the PUC5 floating capacitor voltage at its reference level. In addition, the proposed controller supports both active and reactive power exchanges, adapts to variations in voltage and current references, and remains robust under grid voltage variations. The RL agent learns optimal switching actions through direct interaction with the PUC5 system, eliminating the need for data collection or reliance on existing control models. An Actor-Critic architecture is adopted, and the Proximal Policy Optimization (PPO) algorithm is applied for training (offline) using MATLAB/Simulink, where the RL-C is evaluated under diverse PUC5 configurations and operating conditions in the testing phase. The trained agent has been implemented on an Opal-RT real-time system and validated experimentally using a laboratory-made PUC5 prototype. The performance of the proposed RL-C approach is compared to both traditional approaches including finite control set model predictive control, sliding mode control, and PI control, and other state-of-the-art RL algorithms, demonstrating superior generalization and training efficiency. Moreover, a sensitivity analysis quantifying the impact of reward design, state space, network size, and key hyperparameters on convergence and performance is carried out.

This paper assesses a Hybrid Energy Storage System (HESS) at The Green Village (TGV) of Delft University of Technology (TU Delft), designed and developed as a combination of a lithium-ion battery and hydrogen storage systems to provide a residential energy supply. This paper will evaluate the combination of producing solar-powered green hydrogen through electrolysis, as well as daily and seasonal combinations of battery and hydrogen storage, and electricity generation through a fuel cell. Through an analysis of various sensor data that contained power and hydrogen flow, and control signals, this case study reports on the overall efficiency of the HESS, encourages user energy balancing strategies, and assesses its capability to store sustainable energy over long periods. Based on the same dataset, preliminary machine learning models have been developed and evaluated to predict hydrogen production from weather and PV inverter measurements, supporting future EMS optimization. Therefore, this case study indicates improvements that led to key observations.

Multilevel inverters (MLIs) have become key enablers in renewable energy (RE) integration and electric vehicle (EV) systems, where high-quality power conversion and robustness are critical. Among the different topologies, the Crossover Switches Cell (CSC) converter has recently gained attention due to its superior voltage-boosting capability and reduced component count. While most existing studies on CSC control strategies have been limited to simulations, this work advances the field by providing comprehensive real-time experimental validation under varying operating conditions and parameter mismatches. Finite Control Set Model Predictive Control (FCS-MPC), Sliding Mode Control (SMC), and Lyapunov-based MPC (LMPC) are comparatively assessed in terms of dynamic response, voltage regulation, harmonic minimization, and robustness. Real-time implementation on an Opal-RT platform demonstrates that MPC achieves superior current control with minimal harmonics, SMC offers strong disturbance rejection and effective capacitor voltage balancing, while LMPC guaranties stability with a reduced computational burden. The presented results highlight the trade-offs between these advanced control strategies while providing practical guidelines for selecting robust control techniques for grid-connected MLIs in RE and EV applications.

This study presents a data-driven offline digital twin model of an operational residential hydrogen hub equipped with more than 100 sensors. The model enables analysis and scaling of hydrogen-based hybrid energy hubs from residential to larger systems. The hub integrates photovoltaic generation, battery storage, hydrogen production via electrolysis, compressed hydrogen storage, and fuel cell electricity generation. Using year-long field data, the model reproduces the current configuration (5.34 kWp PV, 15 kWh battery, ~ 45 kg H₂) and quantifies annual performance: 5,102 kWh PV generation, hydrogen production and consumption efficiencies of 48.0% and 39.6%, 25.3 kg H₂ produced versus 49.3 kg consumed, and net grid exchange of +114 kWh. Multi-scenario sizing shows that an optimized configuration (8.46 kWp PV, 30 kWh battery, 70 kg H₂) reduces grid import to ~ 30 kWh yr^-1 while exporting ~ 380 kWh, with the H₂ buffer ending the year near its initial state under a rule-based energy management strategy. The results demonstrate the capability of a sensor-validated framework for designing integrated PV-battery-hydrogen energy hubs.

This article introduces a parallel differential power processing (PDPP) architecture for photovoltaic (PV)/battery applications. The PV to Virtual Bus (PV2VB) architecture enables the integration of a battery and manages its power while performing maximum power point tracking on the PV strings. In the proposed PV2VB PDPP architecture, the battery is positioned at the virtual bus, acting as the input for all string-level converters (SLCs). By selecting a lower voltage for the battery at the virtual bus compared to the PV string or the main bus voltages, component voltage ratings can be reduced. The architecture employs dual active bridge converters connected to bridgeless (BL) converters as SLCs to generate both positive and negative output voltages while providing isolation. These SLCs track the maximum power point of each PV string, while the central converter manages battery charging and discharging. Experimental results confirm the performance and effectiveness of the proposed PV2VB PDPP architecture, achieving efficiencies between 95.5% and 99%.

This paper proposes a hybrid model predictive control (HMPC) strategy for a three-level neutral-point-clamped (3L NPC) rectifier to optimize the operation of electrolyzers for green hydrogen production. Unlike conventional approaches that focus primarily on inverter applications, the proposed method directly addresses the critical issue of neutral-point voltage imbalance in NPC rectifiers. The control framework combines discrete switching-state selection with adaptive objective prioritization, utilizing real-time capacitor voltage measurements to exploit switching-state redundancy. A multi-objective cost function balances current tracking, DC-link voltage regulation, neutral-point voltage balancing, and switching-loss minimization. An adaptive weighting mechanism dynamically prioritizes control objectives in response to system conditions such as renewable intermittency and grid disturbances. To improve long-term reliability, the strategy introduces state diversity enforcement and inactivity penalties, reducing capacitor stress without compromising harmonic performance. The effectiveness of the proposed method is evaluated using a dynamic equivalent model of a proton exchange membrane (PEM) electrolyzer, enabling a realistic assessment of green hydrogen production. Validation through hardware in the loop (HIL) test demonstrates robust current control and stable DC-link voltages in varying operating scenarios.

Power electronic converters are essential components in modern electrical systems, with dc–dc converters being particularly crucial for their extensive variety and widespread applications, notably in renewable energy systems and electric vehicles. For applications such as Photovoltaic (PV) arrays, fuel cells, and microgrids, a critical requirement is the need for non-isolated, high step-up dc–dc converters capable of providing a high-voltage gain while mitigating significant practical challenges, including high semiconductor stress and degraded efficiency at extreme duty cycles. This study proposes a comprehensive, structural-based review of over 100 non-isolated high step-up dc–dc converters, specifically excluding topologies that rely on coupled inductors. Our primary contribution is a novel categorization and comparative analysis framework that delves into the fundamental topological details and features, going beyond the focus of recently suggested reviews. The methodology begins with a systematic analysis of classical non-isolated converters (Boost, Buck-Boost, Ćuk, SEPIC, and Zeta) to establish a baseline and highlight their limitations, such as the semiconductor voltage stress being greater than the output voltage in most cases, which reduces their suitability for high-gain applications. The paper then systematically classifies advanced topologies into 23 distinct groups based on their unique structural characteristics. The comparison is rigorously quantitative and systematic, focusing on structural details and key performance metrics such as voltage gain and density, semiconductor stress, and current continuity and component count. The comprehensive analysis is conducted by deconstructing each topology into its constituent sub-converters to reveal how structural combinations influence key performance metrics. Finally, the findings facilitate a discussion on the practical applications of each topology, matching their inherent characteristics to specific real-...

ABSTRACT The reliable performance of multilevel inverters (MLIs) is critical to the advancement of power electronic systems used in renewable energy conversion, microgrid operation, and electric vehicle technologies. However, these systems are often susceptible to open-circuit faults in power switches, which can adversely affect output waveform quality and overall system stability if not detected early. This study presents an intelligent fault detection and localization strategy based on a fuzzy inference system (FIS) for a reduced device count cross-connected-source MLI. The proposed diagnostic framework employs output voltage and current signals as diagnostic parameters, enabling precise identification of fault conditions. A comprehensive set of fuzzy rules is developed to differentiate between single and double switch faults under diverse fault inception angles. Real-time validation is carried out using the dSPACE DS1202 controller in a hardware-in-the-loop environment. The findings confirm that the proposed FIS-based approach achieves high fault classification accuracy, rapid response time, and strong adaptability across varying operating conditions, highlighting its suitability for real-time condition monitoring in modern power conversion systems.

This article provides a comprehensive review of power electronics converter control and energy management for hydrogen production systems through water electrolysis. Hydrogen production from renewable energy sources is a key pathway toward decarbonizing energy systems and enabling large-scale energy storage. Efficient and dependable operation necessitates addressing the dynamic properties of electrolyzers, the intermittent nature of renewable sources, and the coordination among numerous power electronic interfaces. Unlike earlier studies that addressed these aspects separately, this review systematically connects electrolyzer modeling, converter design, control architectures, and energy management to reveal their critical interdependence. By examining these connections, the analysis reveals critical research gaps in real-time coordination, parameter adaptation, and scalable architectures, outlining pathways toward intelligent and grid-independent hydrogen production systems. This review integrates electrolyzer modeling, power converter control algorithms, AC and DC energy hub architectures, hierarchical control schemes, and energy management systems from classical to advanced methods.

In recent years, several common-ground switched-capacitor transformerless (CGSC-TL) dc–ac multilevel power converters have been introduced, providing advantages such as multilevel output voltage, voltage boosting, and mitigated leakage current. However, these structures mostly suffer from drawbacks, such as limited output voltage levels (like only five levels), lack of voltage-boosting capability, and high charging current spikes of the capacitors. This article proposes a new single-stage CGSC-TL nine-level (9L) multilevel inverter (MLI) with voltage-boosting capability and limited spikes of charging current of the capacitor, designed to be employed as a single-stage power-electronics-based interface device between renewable energy sources, such as photovoltaic (PV) systems and power grid and/or load. The proposed MLI provides several merits, such as a common-ground structure that suppresses PV-to-ground leakage current associated with PV parasitic capacitances, active and reactive power support, a wide input voltage range, and higher output voltage levels (9L) compared with other structures in the same class. Comprehensive comparative analyses, as well as simulation and experimental results, are presented to verify the performance of the proposed inverter.

This article attempts to design an improved robust controller for the DC-DC buck converters subject to the load variations. The design mechanism is given by adopting the radial basis function neural networks for identifying unknown terms in which the neural weights are updated adaptively. To improve the robust behavior of the overall system, two disturbance observers are designed on the basis of cascade control. Furthermore, the control parameters are selected optimally using the complex-order particle swarm optimization algorithm. The stability of the overall system is proven by the Lyapunov theorem. The results of the proposed idea are presented and compared with a fuzzy backstepping control approach and a PID controller in the presence of supplied voltage fluctuations as well as the load variations. In addition, the experimental study is also conducted to assess the practicality of the proposed control framework.

This paper presents a reinforcement learning controller (RLC) for a single-phase full-bridge rectifier as an interface for a battery energy storage system (BESS). A novel solution is presented that combines the traditional proportional-integral (PI) regulator with an RL-based control strategy using a proximal policy optimization (PPO) agent. In a high-fidelity Simulink-based digital twin setup, the agent learns to perform optimal switching actions for a single-phase full-bridge rectifier to achieve accurate current tracking and improved power quality. Simulation results show stable DC voltage regulation at 200V, tracking response under 0.1s, and harmonic compliance with THD equal to 2.38%. The hybrid control strategy guarantees robust dynamic performance and adaptability in the context of renewable energy and storage systems’ varying source and load conditions. The findings demonstrate the potential of coupling AI-driven control with digital twins to empower the autonomy and resilience of future smart energy systems.

This paper discusses the analysis and design of a multi-port DC-DC converter using Gallium Nitride transistors for a 350V bipolar DC grid application, which could be used as the first stage to interconnect a 350V bipolar DC grid and two electric vehicle batteries. The multi-port DC-DC converter is designed with a three-level neutral-point-clamped triple-activebridge topology. The converter's parameters are selected on the basis of its performance characteristic and system specifications. Moreover, a simulation model is built to analyze the design. In the end, a prototype converter is built and the preliminary experimental results of it are shown and discussed.

This article presents a new transformerless switched-capacitor (SC) based five-level grid-connected inverter with inherent voltage-boosting capability. The proposed topology achieves a voltage gain factor of two without requiring an additional dc-dc boost converter or transformer, resulting in a more compact, cost-effective, and efficient design. A single SC cell is utilized to perform bidirectional capacitor charging during both positive and negative grid half cycles, thereby improving energy transfer efficiency and significantly reducing capacitor size and volume compared with the conventional topologies. The inverter employs a minimal number of components - only nine switches and one flying capacitor - while maintaining high performance. Only five switches operate at high frequency, which reduces switching losses, gate driver complexity, and electromagnetic interference. A straightforward control strategy ensures that the inverter delivers a high-quality sinusoidal current waveform to the grid and supports both active and reactive-power flow under various power factor conditions. The reliability of the proposed inverter is analyzed, and its performance is validated through detailed simulations and experimental results. A comparative study with the existing solutions highlights the advantages of the proposed topology in terms of efficiency, voltage gain, component count, and waveform quality.

This research presents a neural-adaptive control technique for DC microgrids in renewable hydrogen production systems. The proposed approach tackles voltage stability issues arising from variable solar production and fluctuating electrolyzer loads with an adaptive neural network-based proportional-integral (ANN-PI) controller with online system identification. The control architecture utilizes two concurrent multilayer perceptron (MLP) networks: one for real-time system identification to estimate the Jacobian matrix, and another for adaptive proportional-integral (PI) parameter adjustment. The decentralized architecture removes communication dependencies among converters, hence improving reliability and scalability. Simulation results indicate a better dynamic response with a 50% decrease in settling time, increased voltage stability retaining the DC bus voltage within ±2% of the nominal 400 V, and resilient performance under diverse situations, including load transitions and changes in solar irradiation. The neural-adaptive method effectively facilitates intelligent, model-free regulation for electric-hydrogen DC microgrids.

This article proposes the use of a proportional–integral–derivative (PID) controller as an adaptive mechanism within the framework of the model reference adaptive system-based rotor flux (MRASF) for accurate rotor speed estimation in induction motors. The controller is designed to impose specific dynamic behaviors and performance criteria, addressing the sensitivity of MRASF dynamics to slip speed. To achieve this, a full-order transfer function of the MRASF is employed to systematically derive the PID parameters using compensation and pole placement techniques. The proposed control strategy ensures that the estimated rotor speed closely follows the desired performance across various operating conditions. The effectiveness of the MRASF–PID design is validated through both simulation and experimental testing under open-loop voltage/frequency control of the induction machine, confirming the successful realization of the targeted dynamic and steady-state performance.

The power versus voltage curve of a photovoltaic (PV) panel exhibits several maximum power points (MPPs) in a partial shading (PS) condition. Thus, it remains an optimization challenge to ensure that PV systems operate at their global MPP (GMPP). Scanning the output characteristics of the PV panels seems a general solution for this issue. However, applying a short circuit to the terminal of PV panels where there exists an electrolytic capacitor, has a detrimental effect on the lifetime of the system. To this end, in this article, a GMPP estimator is proposed as a global solution for conventional maximum power point tracking (MPPT) algorithms under PS conditions. The proposed technique improves existing simple MPPT algorithms with original approaches as follows: first, an accurate microscopic analysis of a PV characteristic in PS conditions is considered, second, an original definition of the dominant cells and modules in a PV panel is proposed that allows to reduce the PS patterns to a finite number, and third, the search area for the MPPT operation is reduced to find the accurate GMPP by proposing two voltage boundaries. The lower boundary corresponds to the GMPP under uniform shading condition that can be determined using a closed form formula, while the upper one refers to the GMPP of a dominant cell in a PV module that can be determined using an artificial intelligence technique. This can also help set the initial duty cycle in a convex area around the GMPP. The functionality of the proposed GMPP estimator is experimentally validated.

The modular multilevel converter (MMC) uses many power electronic components in the high voltage direct current (HVDC) application. One of the major concerns in half-bridge MMC is the fault in the converter submodules. It raises the question of whether the reliability and high-quality performance of the MMC can be increased significantly as the active device controls the power flow between the AC- and DC-sides. During the SM fault within the MMC leg, the unbalance is introduced in-side the MMC converter. The unbalanced voltage within the leg of the MMC will continuously introduce an AC-current component on the DC-side of the converter. Thus, the hybrid proportional-internal (PI) control and proportional-resonant control (PR) is introduced in controlling the power flow within the internal MMC to eliminate the AC-current component and ensure pure DC-current in the internal MMC. This study investigates the internal power flow control of a three-phase rectifier MMC with symmetric and asymmetric SM fault conditions. Compared with conventional control methods, the proposed control can tolerate SM faults and eliminate the AC-current component within the converter, increasing the converter's performance. Simulation results are included and discussed to verify the proposed control.

This paper presents a simple yet robust dq-frame current control strategy for a single-phase 5-level Packed U-Cell (PUC5) inverter, targeting efficient and reliable integration of renewable energy (RE) sources into the grid. The proposed approach exploits the inherent self-balancing characteristics of the PUC5 topology, thereby eliminating the need for capacitor voltage sensors or active balancing circuits. This leads to a simplified hardware structure and reduced overall system complexity and cost. A conventional PI-based dq controller is employed to regulate the injected grid current, ensuring decoupled control of active and reactive powers. Extensive simulations in MATLAB/Simulink demonstrate that the system maintains a unity power factor and achieves low total harmonic distortion (THD) using a basic L-type filter. Moreover, one of the key outcomes is the natural and stable self-balancing of the flying capacitor voltage, which consistently maintains its voltage around half the DC link voltage with low ripples, even under dynamic changes in current reference and DC input voltage. These results confirm the PUC5-controller solution’s reliability and suitability (compelling alternative to more complex multilevel converter architectures) for grid-connected applications.