Reinforcement learning with formation energy feedback for material diffusion models.

Multi-cylinder synchronous control is critical for heavy machinery lifting operations, yet struggles to maintain precision under dynamic and asymmetric loads. Traditional strategies neglect the dynamic coupling between structural strain energy and hydraulic actuation, leading to energy accumulation and synchronization errors. This paper proposes a coordinated coupling control strategy based on dynamic strain energy balance. By decoupling lifting processes from synchronization control and integrating real-time strain energy feedback, the method dynamically compensates for deformation-induced deviations. A multi-physics co-simulation platform (MATLAB/AMESim-Adams) validates the approach, demonstrating significant improvements in synchronization accuracy and stability over conventional methods. Experimental results show the strategy reduces maximum synchronization errors by nearly half under variable loads while suppressing structural fatigue risks. This work advances high-precision multi-cylinder system design, with broader applications in heavy equipment requiring robust cooperative control.

Under the background of the dual carbon strategy, green office practices have become pivotal to sustainable development. However, relying solely on technological upgrades proves insufficient for effectively reducing energy consumption; stimulating and guiding employees' energy-saving behaviours is particularly crucial. Grounded in human-computer interaction perspectives and integrating cognitive ergonomics with dual-system theory, this study investigates how three types of energy consumption feedback (numerical, social comparison, and metaphorical) influence energy-saving intentions through emotional arousal (System 1) and cognitive load (System 2). A single-factor, three-level online scenario experiment (N=299) using the Hayes PROCESS macro tested dual mediation effects. Results indicate no significant mean difference in energy-saving intentions across feedback types. However, mediation analysis strongly supports the dual-system mechanism: emotional arousal significantly positively influences energy-saving intentions (b = 0.3431, p < .001), while cognitive load exhibits a significant negative inhibitory effect (b = −0.1156, p = 0.0177). Regarding specific mechanism activation, analysis revealed that only social comparison feedback significantly elevated participants' emotional arousal levels (b = 0.2923, p = 0.0224). This arousal indirectly promoted energy-saving intentions via an "intuition-emotion pathway" (Indirect Effect = 0.1003, 95% CI [0.0147, 0.1921]). This indicates that emotional motivation proves more effective in energy-saving behavior interventions, whereas merely increasing information volume does not further promote rational thinking. This study enriches the human-centered mechanism theory of energy feedback design, validates the applicability of the dual-system decision-making model in office energy-saving scenarios, and provides practical insights for green office system interface design. It suggests fully leveraging emotional motivation to activate users' intuitive engagement while controlling information complexity to avoid cognitive overload, thereby more effectively guiding energy-saving behaviors.

Towards establishing functional nitrogenase activities within plants.

An advanced beam energy control strategy for medical cyclotron via ADRC with model-based feedforward controller.

Development and implementation of a user-centric real-time energy monitoring and feedback system for energy conservation

Abstract This paper delves into the dynamic characteristics and behaviors of a fractional -order financial risk contagion (FOFRC) system. Through systematic dynamic analysis incorporating stability, Lyapunov exponents spectrum, bifurcation diagrams, 0–1 test and complexity algorithms, the intrinsic nonlinear characteristics of the proposed model are analyzed. To improve system stability, a Hamilton energy feedback control strategy is further developed, which demonstrates effective chaotic behavior suppression through energy regulation principles. Furthermore, a PSR-ESN framework, which combines Echo State Networks with Phase Space Reconstruction, is employed to forecast chaotic time series. Comparative evaluations reveal that the PSR-ESN framework is superior to conventional ESN, RNN and LSTM architectures’ prediction accuracy and computational efficiency. These theoretical and computational advancements substantiate the unique advantages of fractional calculus in financial system modeling and chaotic behavior control, offering new perspectives for advancing risk management strategies.

Abstract During the lifting process of cranes, the lifting mechanism usually experiences significant energy loss, which is detrimental to the lifespan of the lifting mechanism. Therefore, it is particularly necessary to study energy-saving methods for the lifting mechanism of cranes. This article takes the 90 t mobile overhead crane as an example to establish a model of its lifting mechanism hydraulic system. An analysis was conducted on the working energy of the lifting mechanism of the crane, and the energy flow and loss situation of the lifting mechanism of the crane were identified. Then the energy loss process was analyzed, and the energy-saving potential of the crane lifting mechanism was demonstrated through data calculation. According to the results, during the descent stage of the crane lifting mechanism, a large amount of gravitational potential energy is converted into thermal energy consumption of the hydraulic transmission system. Subsequent energy-saving designs for the crane lifting mechanism can add energy feedback devices to this link, with an energy-saving effect of over 40%.

Aiming at the problems of low simulation accuracy, slow dynamic response speed and poor grid-connected current quality in the current single-phase AC electronic load, this design deeply studies the efficient main power topology and load simulation strategy with fast dynamic response and low steady-state error. Based on the STM32F103C8T6 single-chip microcomputer platform, the system adopts the active power factor correction (APFC) dual-loop control strategy and unipolar modulation method to significantly improve the response speed of the system and reduce the steady-state error. In order to achieve adjustable power factor (PFC), this design adopts AC-DC-AC conversion topology. Among them, the front-end circuit is responsible for simulating the load characteristics, and a one-cycle control technology is used to ensure high-precision load simulation; the energy feedback unit of the latter stage uses the hysteresis current control technology, which effectively reduces the harmonic components in the input current, thereby improving the power factor. Through simulation analysis and actual circuit test verification, this design scheme shows significant improvement in simulation accuracy and dynamic response compared with traditional methods. Specifically, the 60 W single-phase AC electronic load constructed according to the above scheme can be flexibly adjusted in the power factor range of 0.70 to 0.94 under the condition that the input voltage U1 is 30 V and the current I1 is 2A, and the efficient feedback of electric energy to the power grid is realized. This design not only improves the overall performance of the single-phase AC electronic load, but also provides a feasible technical path for the effective utilization of energy, which has a good engineering application prospect.

An Efficient Single-Phase Full Bridge Passive SiC-Based Soft-Switching Inverter with Energy Feedback Function

Non-Intrusive Load Monitoring (NILM) technology, enabled by high-precision electrical data acquisition sensors at household entry points, facilitates real-time monitoring of electricity consumption, enhancing user interaction with smart home systems and reducing electrical safety risks. However, the growing diversity of household appliances and limitations in NILM accuracy and robustness necessitate innovative solutions. Additionally, outdated public datasets fail to capture the rapid evolution of modern appliances. To address these challenges, we constructed a high-sampling-rate voltage–current dataset, measuring 15 common household appliances across diverse scenarios in a controlled laboratory environment tailored to regional grid standards (220 V/50 Hz). We propose an AI-driven NILM method that integrates power-mapped, color-coded voltage–current (V–I) trajectories with frequency-domain features to significantly improve load recognition accuracy and robustness. By leveraging deep learning frameworks, this approach enriches temporal feature representation through chromatic mapping of instantaneous power and incorporates frequency-domain spectrograms to capture dynamic load behaviors. A novel channel-wise attention mechanism optimizes multi-dimensional feature fusion, dynamically prioritizing critical information while suppressing noise. Comparative experiments on the custom dataset demonstrate superior performance, particularly in distinguishing appliances with similar load profiles, underscoring the method’s potential for advancing smart home energy management, user-centric energy feedback, and social informatics applications in complex electrical environments.

Analysis of load energy feedback in pulsed coils power supply for the DIII-D tokamak

ABSTRACTTo enhance the operation efficiency of single‐phase full bridge inverter, a novel single‐phase full bridge passive SiC‐based soft‐switching inverter topology is proposed. The passive auxiliary network (PAN) with low energy consumption is used to make the main switch achieve wide range pseudo zero‐current switching (ZCS) switch‐on and pseudo zero‐voltage switching (ZVS) turn‐off to implement economic power saving. The absence of auxiliary switches in the PAN can render the control of the system simpler. The distinguishing feature of this soft‐switching inverter is that each bridge arm's switches share a resonant cell (comprising an absorption inductor and an absorption capacitor), significantly reducing the additional components attached to the main switches. It utilizes a large‐capacity electrolytic capacitor to act as a constant voltage source, constructing a passive and lossless energy feedback loop. This enables the reset of the absorption components, thereby improving the efficiency of the inverter. During the dead‐zone mode, the load current can continue to flow through the PAN, which can reduce the negative impact of the dead‐time state on the output current waveform of the proposed inverter. The experimental results on a 3‐kW prototype with rated output power show that the switching device can realize soft‐switching, and the distortion rate of the output current waveform of the inverter is improved.

Abstract Supergiant shells (SGSs) are the largest interstellar structures in galaxies and inject hot enriched gas into galactic halos. We have studied SGS LMC 1 to determine quantitatively whether stellar energy feedback is adequate to power the formation of an SGS. The Gaia EDR3 photometric data of the OB association LH15 inside SGS LMC 1 are used to construct color–magnitude diagrams, and stellar evolutionary tracks and isochrones are used to assess stellar masses and ages. The observed present-day mass function is compared with the Salpter initial mass function to estimate the number of massive stars that have exploded as supernovae. Their total stellar wind mechanical energy and supernova explosion energy input amounts to (57 ± 12) × 1051 erg. For the gas components of SGS LMC 1, ATCA+Parkes H i data are used to determine the total mass and kinetic energy in the H i shell, MCELS Hα image is used to determine the ionized gas mass and kinetic energy, adopting the H i expansion velocity, and ROSAT X-ray observations are used to estimate the thermal energy in the SGS interior. The sum of the kinetic and thermal energies in the three layers is estimated to be (59 ± 5) × 1051 erg. Thus, the stellar energy feedback from LH15 appears adequate to power the formation of SGS LMC 1. The radial age gradient in LH15 and the young stellar objects along the outer periphery indicate a progression of star formation, which might be a crucial factor for an SGS to grow to its large size.

ABSTRACT To solve the problems of high cost, large additional component size, large circulating loss, and complex control in existing auxiliary resonant soft‐switching inverters, a three‐phase passive auxiliary resonant pole inverter (ARPI) with symmetrical auxiliary networks and electric energy feedback function is proposed. On each phase bridge arm of the inverter, a set of passive auxiliary networks (PANs) without large electrolytic capacitors, transformers, and other devices is provided, which is capable of enhancing the power density of the inverter. Additionally, as the auxiliary commutation circuit does not employ auxiliary switches, it not only lowers the hardware cost but also simplifies the control complexity and enhances the reliability of the soft‐switching inverter operation. The proposed topology is capable of enhancing the utilization ratio of DC voltage and reducing the harmonic content of the output waveform. According to the equivalent circuit diagrams in different operating modes, the soft‐switching operating principle and realization conditions of soft‐switching for the proposed ARPI are analyzed. The parameters of passive components are designed with the aim of minimizing the PAN loss. A 3‐kW experimental prototype is constructed. The experimental results indicate that the switching tubes can realize the soft‐switching action and suppress the rate of change of voltage and current. The efficiency of the prototype can reach 98.30% at the rated output power, which is a remarkable improvement compared with the same type of ARPIs.

Urban rail transit plays an important role in the rapid economic and social development of China. The integrated storage and feedback urban rail traction power supply system is one of the strategies for the green and low-carbon development of urban rail transit. This paper establishes a real-time operation optimization model based on the Markov decision process in the context of the integrated storage and feedback urban rail traction power supply system. It uses offline training of a deep reinforcement learning agent to optimize the control parameters of the energy feedback systems and energy storage systems in real-time to reduce the traction energy consumption of the power supply system, providing a reference for energy-saving operation of subways.

ABSTRACT The early growth of black holes (BHs) in atomic-cooling haloes is likely influenced by feedback on the surrounding gas. While the effects of radiative feedback are well-documented, mechanical feedback, particularly from active galactic nucleus (AGN) jets, has been comparatively less explored. Building on our previous work that examined the growth of a 100 ${\rm M_\odot }$ BH in a constant density environment regulated by AGN jets, we expand the initial BH mass range from 1 to $10^4\, {\rm M_\odot }$ and adopt a more realistic density profile for atomic-cooling haloes. We reaffirm the validity of our analytic models for jet cocoon propagation and feedback regulation. We identify several critical radii – namely, the terminal radius of jet cocoon propagation, the isotropization radius of the jet cocoon, and the core radius of the atomic-cooling halo – that are crucial in determining BH growth given specific gas properties and jet feedback parameters. In a significant portion of the parameter space, our findings show that jet feedback substantially disrupts the halo’s core during the initial feedback episode, preventing BH growth beyond $10^4 \, {\rm M_\odot }$. Conversely, conditions characterized by low jet velocities and high gas densities enable sustained BH growth over extended periods. We provide a prediction for the BH mass growth as a function of time and feedback parameters. We found that, to form a supermassive BH ($\gt 10^6 \, {\rm M_\odot }$) within 1 Gyr entirely by accreting gas from an atomic-cooling halo, the jet energy feedback efficiency must be $\lesssim 10^{-4} \dot{M}_{\rm BH} c^2$ even if the seed BH mass is $10^4 \, {\rm M_\odot }$.

Context. In the context of the hierarchical formation of galaxies, we investigated the role played by mergers in shaping the scaling relations of galaxies, that is the projections of their fundamental plane onto the Ie − Re, Ie − σ, Re − Ms, and L − σ planes. To this end, based on the scalar virial theorem, we developed a simple theory of multiple dry mergers to read both the large-scale simulations and the companion scaling relations. Aims. The aim was to compare the results of this approach with the observational data and with two of the most recent and detailed numerical cosmo-hydro-dynamical simulations: Illustris-TNG and EAGLE (Evolution and Assembly of GaLaxies and their Environments). Methods. We derived these scaling relations for the galaxies of the Mapping Nearby Galaxies at APO (MaNGA) and Wide-field Imaging of Nearby Galaxy-Clusters Survey (WINGS) databases and compared them with the observational data, the numerical simulations, and the results of our simple theory of dry mergers. Results. The multiple dry merging mechanism is able to explain all the main characteristics of the observed scaling relations of galaxies, such as slopes, scatters, curvatures, and zones of exclusion. The distribution of galaxies in these planes is continuously changing across time because of the merging activity and other physical processes, such as star formation, quenching, and energy feedback. Conclusions. The simple merger theory presented here yields the correct distribution of galaxies in the main scaling relations at all cosmic epochs. The precision is comparable with that obtained by the modern cosmo-hydro-dynamical simulations, with the advantage of providing a rapid exploratory response on the consequences engendered by different physical effects.

In rehabilitation robotics, optimizing energy consumption and high interaction forces is essential to prevent unnecessary muscle fatigue and excessive joint loading as they often cause an inefficient trajectory planning and disrupt natural movement patterns. Stroke patients frequently exhibit asymmetrical muscle activation and impaired neuromuscular coordination, making it necessary to design a system that adapts to their specific motor limitations with energy-efficient and excessive torque control. This study presents a reinforcement learning-based trajectory optimization framework for a 3-DOF ankle rehabilitation robot, integrating musculoskeletal modeling, transactive energy and real-time physiological feedback to generate adaptive rehabilitation trajectories. The methodology utilizes electromyography (EMG) signals from key ankle muscles and joint reaction forces to refine movement patterns to ensure biomechanical efficiency. The methodology is validated using data from ten stroke patients, demonstrating its potential to enhance rehabilitation effectiveness by promoting more natural, efficient, and physiologically accurate movement trajectories.

Simulation and Verification of Beam Combination Using a Stochastic Parallel Gradient Descent Algorithm with Energy Feedback Adaptation