Hydrogen energy, as a core area in global energy transition and low-carbon development, plays a critical role in supporting industrial innovation and talent cultivation through discipline construction. Universities and research institutions worldwide are actively exploring pathways for establishing a hydrogen energy discipline system. Based on a systematic study of hydrogen energy discipline development, this paper reviews the current status in China and conducts practical explorations focusing on talent cultivation and curriculum system design. From the perspectives of disciplinary layout, curriculum structure, research platforms, and faculty development, and by integrating goal-oriented analysis, pedagogical innovation, and strategic resource allocation, the paper presents achievements in cultivating specialized talent, advancing technological innovation, and serving industrial development. Furthermore, it analyzes existing challenges and proposes targeted optimization strategies, aiming to provide theoretical references and practical insights for the high-quality development of hydrogen energy disciplines in China. The findings indicate that current disciplinary construction faces challenges, including insufficient interdisciplinary integration, a shortage of practical resources, and a need for enhanced internationalization. Accordingly, recommendations for subsequent construction and exploration are proposed to facilitate the high-quality development of the hydrogen energy industry.
This study investigates the ash deposition behavior of anthracite during combustion, which significantly affects boiler safety and efficiency due to slagging tendencies. A pilot-scale one-dimensional settling furnace system was employed to conduct combustion experiments under varied operating conditions, including different loads, primary/secondary air ratios, and excess air coefficients. Ash samples were analyzed by fusion tests, scanning electron microscope(SEM), X-ray diffraction(XRD), X-ray fluorescence spectroscop(XRF), and laser sizing. Results show ash composition (C, O, Si, Al) and crystalline phases (quartz, mullite, hematite) remain stable. Increasing the load from 0.2 MW to 0.3 MW raises the ash deformation temperature from
To address the systematic limitations of traditional optimization methods for coal-fired boilers and the inability to accurately perceive in-furnace combustion states, this paper proposes a comprehensive boiler optimization system integrating intelligent sensing and control. First, an infrared temperature measurement array is deployed, combined with a gradient positioning algorithm, to achieve online reconstruction and visualization of the three-dimensional temperature field within the furnace. Second, a dynamic model of the boiler combustion process is established based on a continuous-time Bayesian network. Finally, a multi-objective particle swarm optimization algorithm with dynamically adjusted inertia weights is employed for online optimization, thereby constructing a real-time closed-loop adaptive intelligent combustion control system. Engineering application results demonstrate that the proposed system can effectively perceive the combustion state and accurately identify and provide early warnings for abnormal conditions, such as slagging and uneven combustion. After the system was put into operation, the boiler efficiency increased by no less than 0.3%, and NOx emissions were reduced by no less than 12%. In conclusion, this system provides robust technical support for resolving operational optimization challenges in utility boilers and achieving the synergistic development of safety, economic efficiency, and environmental protection.
To address the issues of dispatch failure and economic losses caused by multi-source uncertainties—including the randomness of wind and solar power generation, load fluctuations, and parameter deviations—during the aggregation of distributed energy resources in virtual power plants (VPP), this paper proposes a multi-time scale adaptive dispatching framework embedded with multi-source uncertainty modeling and an online parameter correction mechanism. Based on two-stage robust optimization and an improved quantum genetic algorithm (QGA), a pre-dispatch scheme is generated via robust optimization during the day-ahead stage. During the intraday stage, a state feedback mechanism is introduced to rolling-correct key parameters using the improved QGA, thereby establishing a closed-loop dispatching structure. Simulation results demonstrate that under significant prediction deviations in wind/solar generation and electric/thermal loads, the actual operational revenue of the proposed method increases by approximately 3.2% compared to traditional deterministic dispatching. Furthermore, the online parameter correction strategy significantly reduces the system balancing cost in most periods, with a reduction margin approaching 90%. The proposed method effectively coordinates the robustness, economics, and adaptability of the dispatching scheme, providing a technical pathway for the secure and economic operation of VPP in highly uncertain environments.
A virtual power plant (VPP) serves as a key platform for integrating distributed energy resources, enhancing energy efficiency, and promoting the absorption of renewable energy. However, a single VPP has limited regulation capacity and competitiveness, whereas the coordinated operation of multiple VPPs offers greater economic and low-carbon advantages. Based on this, this paper proposes an optimal dispatch method for multiple VPPs that considers integrated demand response and multi-energy interactions, and employs an improved Shapley value for benefit distribution. Firstly, a collaborative dispatch model for multiple VPPs incorporating integrated demand response and energy interactions is established. By guiding load changes through demand response, the model achieves peak shaving and valley filling. Through energy interactions among multiple VPPs, energy resources on the supply side are fully utilized, improving system economics. Secondly, in the benefit distribution process, a three-dimensional evaluation index system—covering economic, energy transaction, and environmental aspects—is introduced to enhance the traditional Shapley value method, providing a more comprehensive basis for benefit allocation. Simulation results demonstrate that the proposed model not only reduces operating costs and carbon emissions but also quantifies the contributions of each VPP member through multi-dimensional indicators, thereby more comprehensively reflecting their actual input.
Traditional evaluation methods predominantly rely on the center of inertia frequency index, which only reflects the overall average frequency dynamics of the network, ignoring the significant characteristics of time-varying inertia and uneven spatial distribution under high penetration of renewable energy. To address the issue of varying system inertia levels caused by large-scale wind power participating in inertia support and frequency regulation, this paper proposes an inertia demand evaluation method for wind power-integrated power systems considering the spatiotemporal characteristics of inertia. Firstly, a system frequency response model incorporating wind power comprehensive control is constructed by integrating wind power virtual inertia response and pitch angle primary frequency regulation control. The system transfer function and the calculation formula for the equivalent inertia time constant are derived, clarifying the support mechanism of wind power in delaying the rate of change of frequency through rapid response to active power disturbances. Secondly, a characterization method considering the spatiotemporal characteristics of inertia is proposed. Based on the analysis of the nodal power-frequency mechanism, a node inertia matrix and a temporal dynamic model are constructed to quantify the temporal evolution laws of inertia at the same node and the spatial distribution differences among different nodes, thereby overcoming the limitations of traditional center of inertia frequency evaluation. Furthermore, by obtaining the system power-frequency equation and frequency response model, an inertia time constant model is constructed to form an inertia demand evaluation model considering spatiotemporal characteristics. Finally, simulation verification is conducted based on the modified IEEE 10-machine 39-bus system. The results demonstrate that the participation of wind power can effectively enhance the system's inertia support capability. The proposed method accurately captures inertia differences across different nodes at different times and realizes the visualization of spatiotemporal characteristics through inertia heat maps, providing a quantitative basis for inertia configuration in the planning stage and frequency stability control in the operation stage of new power systems.
To investigate the effects of blade icing and anti-icing modifications on the vibration characteristics of 5 MW wind turbines in the winter icing environment of the Yunnan-Guizhou Plateau, and to evaluate their operational safety under complex climatic conditions, this study utilizes three anti-icing test turbines retrofitted with an active aerothermal method. Field vibration monitoring data spanning four months were collected. Nine vibration state variables at critical locations, such as the nacelle, were selected. By calculating characteristic parameters and applying threshold criteria, longitudinal and comparative analyses were conducted to comprehensively evaluate the vibration status and evolution trends of the turbines. The results indicate that the overall vibration severity of the three test turbines remains at a low level with negligible individual variations, and no abnormal fluctuations were observed. The mass imbalance and aerodynamic profile alterations induced by blade icing, as well as the load variations caused by the newly installed anti-icing equipment, did not lead to a significant exacerbation of turbine vibrations. It is concluded that, under icing conditions, neither blade icing nor the anti-icing modifications have a significant impact on the overall vibration levels of the 5 MW wind turbines. The structural dynamic performance of the turbines remains stable, demonstrating the capability for long-term safe operation.
To address the issues of data noise interference, feature scale discrepancy, and insufficient multiscale meteorological pattern modeling in photovoltaic power forecasting, a prediction method based on dynamic data preprocessing and gated dense multiscale convolutional neural network (GDMS-CNN) is proposed. Firstly, an anomaly detection mechanism based on dynamic sliding-window Z-score is established, and missing values are processed via covariance-weighted multivariate interpolation. Secondly, an adaptive piecewise normalization algorithm is adopted to eliminate feature dimensional differences, and cloud-cover correction factor and atmospheric attenuation factor are constructed to enhance physical feature representation. Finally, a GDMS-CNN is designed, wherein the feature extraction efficiency is optimized by depthwise separable convolution modules, densely connected dilated convolution blocks are constructed to capture multiscale spatiotemporal correlation features, and an asymmetric gated channel attention mechanism is embedded to dynamically recalibrate feature weights. Experimental results demonstrate that the proposed method reduces the root mean square error (RMSE) by 16.4% compared with the optimal baseline model genetic algorithm-variational mode decomposition-echo state network (GA-VMD-ESN), and by 43.4% compared with the traditional random forest. The proposed method provides a novel solution for photovoltaic output forecasting and effectively enhances the reliability of power grid dispatching.
To address the practical challenges of poor data quality, strong hyperparameter coupling, and significant peak positioning errors in power system carbon emission peak forecasting, a framework integrating robust data preprocessing with an improved grey-convolution hybrid model is proposed. Firstly, a dynamic quantile boundary-based outlier detection and multi-window weighted robust repair procedure is established, together with a random forest feature importance-based chained multiple imputation method, to suppress outlier disturbances and high-dimensional missingness in non-Gaussian data. Subsequently, an improved convolutional network incorporating variational mode decomposition, dilated convolution, and attention gating is constructed, with an embedded improved grey model for long-term trend extraction; hyperparameter optimization is achieved through a grey relational grade-guided whale optimization algorithm. Experimental results show that compared with genetic algorithm, particle swarm optimization, and grey wolf optimization, the proposed algorithm reduces mean absolute percentage error by 39.7%, 32.5%, and 25.4%, and peak prediction time deviation by 77.1%, 71.4%, and 60.0%, respectively. Compared with autoregressive integrated moving average (ARIMA), long short-term memory - prophet (LSTM-Prophet), time-series transformer (TST), empirical mode decomposition-LSTM (EMDE-LSTM), and variational mode decomposition - gated recurrent unit (VMD-GRU), the proposed model reduces mean absolute percentage error to 2.89% and peak prediction time deviation to 0.7 h, while improving the inflection point capture rate to 93.8%. This study provides a new technical approach for accurate carbon emission peak prediction and offers data support for power system emission reduction strategy formulation.
To address the economic, security, and computational efficiency challenges in transmission-distribution coordinated dispatch with high renewable energy penetration, a multi-period optimization model and an adaptive improved grey wolf optimization (AIGWO) are proposed. First, a three-dimensional objective function framework is established by comprehensively considering economic cost, environmental penalty, and security risk, with complete modeling of unit ramping constraints, energy storage charging-discharging time-coupled constraints, and network security boundary conditions. On this basis, three innovative mechanisms are designed: (1) a nonlinear adaptive convergence factor to dynamically balance global exploration capability, (2) knowledge-guided population initialization to improve solution quality, and (3) hybrid binary-real encoding to cooperatively optimize continuous and discrete variables, thereby overcoming the low convergence efficiency and poor discrete decision space processing of conventional algorithms. Experiments are performed on a modified IEEE 33-bus system for verification. Results show that compared with the standard grey wolf optimization, particle swarm optimization, mixed integer linear programming, and deep Q-network, AIGWO reduces the total cost by 1.72%~5.03%, decreases line overload violations by 89.2%~93.5%, lowers voltage violations by 82.4%~90.3%, and shortens the computation time by 9.32%~94.22%. The proposed algorithm provides an efficient solution for coordinated transmission-distribution resource optimization in high-fluctuation source-load scenarios.
To address voltage over-limit and instability issues caused by high penetration of distributed photovoltaic in distribution grids, this paper proposes a voltage regulation strategy based on distributed photovoltaic cluster partition. Firstly, a dual-criterion modularity function considering net load and equivalent electrical distance is constructed. A community detection algorithm based on modularity (i.e. fast-unfolding) is applied to dynamically partition photovoltaic clusters. Secondly, differentiated voltage regulation is designed according to the severity of cluster over-limit conditions: intra-cluster reactive power adjustment for mild over-limits, and multi-device hierarchical collaborative control for severe over-limits, with task allocation based on response speed and economic priority. Finally, an optimization model targeting minimization of network losses and voltage deviation is established. The improved multi-organization particle swarm optimization (MPSO) algorithm, combined with niche-based techniques, is employed to determine the optimal regulation sequence and device action levels. Simulation results demonstrate that this method effectively controls voltage fluctuations, reduces network losses, and enhances system stability in both the modified IEEE 33-node system and the IEEE 123-node system.
The uncertainty of wind and photovoltaic power generation results in carbon emissions of low-carbon scheduling methods not meeting expectations. Therefore, a multi-time-scale low-carbon scheduling method for regional integrated energy systems under chance constrained planning is proposed. Firstly, it uses the power of wind and solar power generation at different time periods as random variables, and introduces confidence level quantification constraints, an improved particle swarm algorithm is used to determine the optimal decision variables. Secondly, it introduces carbon capture power plants to capture, store, and reuse CO2, constructs a carbon cycle system, and designs a tiered carbon trading mechanism and user demand response mechanism. Finally, it designs a multi-time-scale real-time rolling control plan, constructs real-time scheduling objective functions and constraints, and achieves low-carbon scheduling of regional integrated energy systems at multiple time scales. The experimental results show that the designed scheduling method reduces carbon emissions by
TSINGHUA JOURNALS HOME
