Published in Discover Electronics, their double-integral super-twisting direct power control strategy eliminates the need for complex mathematical linearization while delivering rapid transient recovery and stable steady-state regulation. In rigorous comparative evaluations, the proposed method reduced voltage overshoot and undershoot by 56 to 74 percent compared to conventional proportional-integral-derivative controllers and by 6 to 9 percent relative to standard double-integral sliding mode control. It also achieved substantial reductions in integral square error of up to 90 percent versus proportional-integral-derivative controllers and up to 13 percent against standard sliding mode benchmarks, while maintaining settling times as low as nine milliseconds. Even under extreme load shedding, the observed steady-state voltage deviation remained a negligible 0.05 volts, representing just 0.1 percent of the nominal 50-volt direct-current bus.
The operational challenge addressed by this research arises from the rapid growth of photovoltaic direct-current microgrids, which integrate diverse renewable energy sources to improve overall energy efficiency and system resilience. However, direct-current microgrids frequently suffer from severe voltage oscillations across their distribution buses, driven by weather-induced fluctuations in solar irradiance, sharp variations in electrical demand, and transient power grid faults. Left unchecked, these voltage oscillations cause system degradation, shorten battery life, and risk damaging sensitive connected electrical loads. While conventional proportional-integral controllers are widely utilized due to implementation simplicity, they struggle to preserve stability under sudden load shifts and severe grid disturbances. Furthermore, existing transient-aware strategies often suffer from parameter sensitivity and severe chattering.
To overcome these limitations, the authors focused on the bidirectional dual active bridge converter, a critical interface topology for connecting energy storage units within photovoltaic systems. The dual active bridge consists of two actively controlled switching bridges linked across a high-frequency transformer, with each bridge comprising four controllable semiconductor switches such as metal-oxide-semiconductor field-effect transistors or insulated-gate bipolar transistors in an H-bridge configuration. Unlike traditional converters that rely on passive diodes on the secondary side, controllable switches on both sides allow seamless bidirectional power flow in both buck and boost modes. The high-frequency transformer provides essential galvanic isolation while dramatically minimizing core weight and copper volume compared to line-frequency transformers.
Power transfer through the converter is governed using single-phase-shift modulation, the most widely adopted modulation technique for dual active bridge converters due to its simplicity and bidirectional capability. Under this modulation scheme, both primary and secondary H-bridges generate square-wave voltages at a fixed 50 percent duty cycle, with power magnitude and flow direction regulated by an angular phase shift between the two bridges. This angle defines a normalized phase shift ratio ranging from negative 0.5 to positive 0.5, where positive values dictate buck-mode power delivery from the primary input to the secondary output, and negative values dictate boost-mode reverse power flow. The transferred power relates quadratically to this phase shift ratio, reaching maximum transmission at values of positive or negative 0.5. By analytically inverting this relationship, the proposed controller computes the required phase shift ratio directly from demanded power, completely avoiding nested inner current loops or iterative tuning.
The foundational structure of the controller relies on double-integral sliding mode principles formulated from the converter output energy dynamics. The time rate of change of energy stored in the output capacitor equals the difference between transmitted input power and load consumption. Expressing the output capacitor voltage derivative in terms of circuit parameters yields a dynamic state equation that relates voltage variation to load current, input voltage, transformer turns ratio, switching frequency, and leakage inductance. To eliminate steady-state error without sacrificing dynamic responsiveness, the authors constructed a double-integral sliding surface incorporating the instantaneous voltage tracking error, its single time integral, and its double time integral, weighted by strictly positive coefficients. Setting the time derivative of this sliding surface to zero yielded an equivalent control law that keeps system trajectories aligned with the sliding manifold, guaranteeing that voltage tracking errors decay asymptotically to zero under steady-state operating conditions.
To eliminate the destructive chattering inherent in standard sliding mode implementations, the researchers integrated a second-order super-twisting algorithm to handle switching dynamics. The super-twisting formulation replaces discontinuous switching with a continuous hyperbolic tangent function combined with an integrated auxiliary state variable, producing smooth control actions while preserving finite-time convergence. Within the unified control framework, the equivalent control expression is combined with the super-twisting term to generate the reference power demanded by direct power control. This calculated power is immediately mapped into the required phase shift ratio and passed to the pulse-width modulation generator to trigger the bridge switches, ensuring robust disturbance rejection and rapid dynamic response while maintaining chattering-free operation across volatile load transients and severe parameter uncertainties.
A rigorous theoretical foundation for the control strategy was established through a Lyapunov stability proof confirming finite-time convergence. The authors formulated a quadratic Lyapunov function candidate based on the sliding variable and the transformed super-twisting auxiliary state. By proving that the associated parameter matrices remain positive definite under positive controller gains and physical converter constants, the time derivative of the Lyapunov function was shown to satisfy a finite-time stability inequality bounded by the square root of the Lyapunov candidate. This condition proves that the sliding surface and auxiliary variable converge to zero within a finite settling duration. Once the system reaches the sliding manifold, its closed-loop error dynamics are governed by a characteristic differential equation with positive Hurwitz coefficients. This mathematical proof guarantees that the voltage tracking error asymptotically converges to zero, ensuring precise output voltage regulation even under unexpected parameter variations and external load disturbances across all operational states.
The proposed control strategy was evaluated through extensive simulation trials in MATLAB and Simulink based on converter parameters for a 50-volt direct-current bus. In the first scenario, the target reference voltage was subjected to abrupt step transitions, shifting successively between 60 volts, 40 volts, 50 volts, and 70 volts under a stable input voltage. The proposed controller restored output voltage to the reference values substantially faster than both proportional-integral-derivative and standard sliding mode controllers. In a second scenario, the system was subjected to sudden load resistance perturbations, stepping from 3 ohms to 1.5 ohms at 0.02 seconds and then from 1.5 ohms to 6 ohms at 0.07 seconds under a constant 50-volt setpoint. The proposed strategy exhibited remarkable robustness, achieving significantly reduced overshoot and minimal integral absolute error compared to the benchmark controllers, while matching the best settling times across both resistance transitions and outperforming benchmarks across all evaluated parameters.
The controller was further tested against severe input voltage fluctuations and simulated within a practical photovoltaic-battery charging architecture. When input voltage was dropped from 100 volts to 50 volts at 0.03 seconds and then increased to 150 volts at 0.07 seconds, the proposed controller maintained superior voltage stability and ensured a noticeably smoother inductor current profile than the conventional proportional-integral-derivative scheme. When integrated with a photovoltaic module supplying an energy storage battery under step changes in solar irradiance, the controller maintained consistent direct-current link voltage during active battery charging. Across varying solar irradiance levels, the super-twisting direct power controller governed battery voltage and current waveforms with minimal distortion, supporting smooth battery state-of-charge progression and demonstrating resilience against the nonlinear charging characteristics of electrochemical battery banks. Comparative waveforms confirmed that the proposed strategy maintained stable state-of-charge trajectories without excessive transient oscillations.
To confirm practical feasibility and validate the simulation findings, the authors deployed a real-time processor-in-the-loop experimental setup. The control algorithm executed on an external 32-bit Arduino Due microcontroller operating at an 84-megahertz clock frequency, interfaced via a universal serial bus link with a host computer simulating the converter dynamics. The host computer operated with eight gigabytes of random-access memory and an Intel i5 processor running at 2.59 gigahertz, while the 84-megahertz Arduino Due was chosen because its 32-bit architecture provides superior real-time execution. Experimental voltage waveforms closely matched simulation results, with minor transient differences traced to unmodeled hardware delays and conversion latencies. Measured communication latency was influenced by active serial debugging routines intentionally retained for testing, which can be eliminated in commercial deployments. Future research will extend the scheme to compensate for switching dead-time effects, suppress transformer direct-current bias currents, and conduct quantitative efficiency measurements across diverse operating environments.
Subject of Research: Advanced sliding mode control and direct power regulation for dual active bridge DC-DC converters in renewable energy applications
Article Title: A super-twisting enhanced double integral sliding mode direct power control strategy for dual active bridge converters
Article References: Cham, J. D., & Alombah, N. H. (2026). A super-twisting enhanced double integral sliding mode direct power control strategy for dual active bridge converters. Discover Electronics, 3(1), Article 118. https://doi.org/10.1007/s44291-026-00272-1
Image Credits: AI Generated
DOI: 10.1007/s44291-026-00272-1
Keywords: Dual Active Bridge Converter, Super-Twisting Algorithm, Double Integral Sliding Mode Control, Direct Power Control, Single-Phase-Shift Modulation, Output Voltage Regulation, DC-DC Power Conversion, Photovoltaic Microgrids, Battery Energy Storage Systems, Processor-In-The-Loop Validation, Chattering Reduction, Transient Response Optimization, Galvanic Isolation, High-Frequency Transformer, Lyapunov Stability Analysis, Integral Square Error, Hardware Latency
News Source: Faith Mcneil. (November 12, 2026). Super-Twisting Direct Power Control Stabilizes Voltage in Dual Active Bridge Converters. Scienmag.



