Compared with traditional power systems, the structure of an AC/DC hybrid system with new energy sources is more complex, its stability evaluation is more difficult, and the identification and interpretation of stability influencing factors are poor. Given the above problems, this study first selects new energy and DC features as the input of the stability evaluation model and obtains the relationship between the sample prediction value and the stability result using the sigmoid function. A transient power-angle stability evaluation method for an AC/DC hybrid system based on extreme gradient Boosting (XGBoost) was proposed. To further analyze the influence of features on the transient power angle stability of the system, an interpretable analysis method of features based on SHAP is proposed, which explains the importance of new energy and DC features from a global perspective, which reflects the relationship between the size of each feature itself and the promotion and inhibition of stability results from the perspective of all samples, and then obtains the influence of features on the stability results of a single sample from a local perspective. Finally, simulation verification was performed on a 500 kV actual AC/DC hybrid system, which proves that the accuracy of the evaluation method is high and that SHAP can effectively explain the influence of new energy and DC features on the transient power angle stability of the AC/DC hybrid system.
Positive- and negative-sequence oscillation phenomena close to the fundamental frequency (40-60 Hz) occur in a practical modular multilevel converter-based high-voltage DC (MMC-HVDC) transmission system for new energy integration, which has led to a decrease in the power output of new energy sources. The mechanism of near-fundamental-frequency oscillations is more complex and has more influencing factors than oscillations in other frequency bands. This study focuses on the near-fundamental-frequency oscillation stability of an MMC-HVDC-connected direct-drive wind farm. Refined impedance models of the direct-drive wind turbine and sending-end MMC are established by considering positive and negative sequence controls. Based on the established impedance models, the mechanisms of the near-fundamental-frequency positive- and negative-sequence oscillations between the direct-drive wind farm and the sending-end MMC are revealed. In addition, the parameter phase-margin sensitivity is defined, and the key influencing factors of the near-fundamental-frequency oscillation stability of the interconnected system were analyzed quantitatively. Finally, an electromagnetic transient simulation model of the MMC-HVDC-connected direct-drive wind farm is developed. The near-fundamental-frequency positive- and negative-sequence oscillation phenomena in the actual project are reproduced, and the correctness of the near-fundamental-frequency oscillation mechanism analysis is validated.
To address the issue of high-frequency resonance in multiterminal flexible direct current (DC) transmission systems caused by unloaded line input, a simplified impedance model is established for MMC in the high-frequency range. Based on the impedance analysis, the electrical mechanism of high-frequency resonance in a single MMC caused by an unloaded line input was analyzed. The analysis shows that owing to the unloaded line input, the MMC AC-side equivalent impedance exhibits negative damping in the high-frequency range, leading to high-frequency resonance. Joint modeling was then conducted for the receiving converter stations under different control methods. By analyzing the propagation path of high-frequency resonance energy between stations and its impact on the operation of receiving MMCs, research has shown that high-frequency resonance on the AC side of MMC can affect other MMCs through DC lines. The degree of the impact is related to the external control method. To solve the problem of high-frequency resonance, a suppression method involving the addition of a bandstop filter to the voltage feedforward was adopted to eliminate the negative damping of the MMC at the high-frequency resonance point, complete the impedance reshaping of the oscillating MMC, effectively suppress the high-frequency resonance of the MMC, and reduce the negative impact of high-frequency resonance energy propagation on other MMCs. Finally, the accuracy of the MMC high-frequency resonance analysis and the effectiveness of the corresponding suppression method were verified using the RT-LAB5600 real-time online simulation platform.
The new power system driven by the "dual carbon" goal cannot adapt to the top-down energy/power balance mode at different temporal and spatial scales.Therefore, it is urgent to conduct relevant research on the construction path of the future power grid, particularly the distribution network. The micro-energy network, a comprehensive energy network with self-balance regulation ability, can leverage the pivotal role of power grids in promoting zero-carbon/low-carbon energy production and consumption and will play a pivotal role in future power grid transformation and upgrading. First, the evolution stages and modes of traditional distribution networks, active distribution networks, and future low-carbon distribution networks are summarized and compared. Second, based on the characteristics of "source-network-load" in the future, a type of future distribution network construction idea of bottom-up evolution layer by layer is proposed by constructing micro-energy network and utilizing the interconnection and interaction between micro-energy network units and distribution network. Additionally, a type of atomic future distribution network is introduced. Finally, future research directions for the evolution of distribution networks are discussed from different perspectives.
Distributed generation (DG) connected to a power network has changed the traditional single-radiation power supply network structure, while the two-way power flow and load current affect the safe and stable operation of power systems. In this study, the line protection adaptability under the scenario of DG connected to a 10 kV distribution network is analyzed. Based on the different locations of distributed generation, the effects of the infeed current, outflow current, and reverse current on the existing protection of the distribution network are analyzed theoretically. The effect of the power capacity of DG on the protection performance is clarified. Moreover, the protection performance boundary conditions of a 10 kV distribution network are quantified. This study provides theoretical support for subsequent research on protection configuration schemes and new protection principles. The simulation results verify the accuracy of the protection adaptability analysis.
Proton exchange membrane (PEM) water electrolysis is adaptable to power fluctuation owing to its advantages of fast response and wide power adjustment range. It can be used as a flexible and adjustable resource to match the dynamic power grid supply and demand, and improve the penetration of renewable energy. This study summarizes and analyzes the modeling methods and research progress of the PEM water electrolysis system, expounds the basis for selecting parameters, and prospects the improvement direction of the PEM water electrolysis system from four aspects of water, heat, electricity, and gas. The model is an essential tool in the model development of the electrolysis hydrogen production system. In this study, the static performances and dynamic behaviors of the PEM water electrolysis system can be clarified by modeling, which can effectively support the design, operation, and control optimization of the system. This research review has certain theoretical value and guiding significance for the model establishment, development, and simulation analysis of the PEM electrolysis hydrogen production system.
An optimal allocation model for the planning of energy storage systems considering power interconnections is established for multi-storage system planning in multi-area grids to accommodate the gradual increase in the scale of renewable energy. First, an equivalent model of a multi-region power system is proposed to clarify the cooperation of interconnected regions; second, the unit start-stop planning method in a multi-region system is proposed to calculate the intraday start-stop scheme of units subordinated to different regions. Subsequently, a two-layer planning model of energy storage systems for multi-area grids considering power interconnections is proposed to allocate storage systems in interconnected grids. The optimal renewable energy consumption of the entire grid is considered as the outer layer objective to calculate the configuration of the energy storage systems. The inner model is developed to minimize the sum of the fluctuations of the intra-regional transmission power and the injection power in each region for the optimal operation (including thermal units, storage systems, and transmission lines) of multiple regions and scenarios. Subsequently, the approximate network loss value and utilization index of the transmission lines are proposed to quantify the benefits of the energy storage planning scheme. Finally, we verify the effectiveness of the proposed model and method through a case analysis. The results show that the proposed model can fully consider the requirements for renewable energy consumption in each region and effectively improve the consumption index of the entire grid, which can be applied to the collaborative planning of energy storage systems in a wide-area interconnected grid.
As new energy generation incurs more uncertainty, it leads to a more severe shift in net load data distribution. The data distribution shift means that the feature information learned by the model in historical data is no longer fully applicable to future data, thus posing a significant challenge to net load forecasting (NLF). Therefore, considering the data distribution shift problem in net load, this study proposes a short-term residential net load forecasting method based on IRM-UW-LSTM to improve net load forecasting accuracy. First, a dual-objective problem was established using invariant risk minimization (IRM), which includes accurate forecasting and learning of invariant features across different data distributions. Second, a long short-term memory neural network (LSTM) was used to deal with the nonlinear features of the time series data. Subsequently, an uncertainty weighting (UW)-based objective balancing mechanism was used to avoid overachieving either objective. In addition, a quantile regression method was introduced to extend this study to probabilistic forecasting. Finally, the effectiveness of the proposed method was verified using multiple dimensions of deterministic and probabilistic prediction results, different data distribution shift levels, and different PV penetration rates based on real residential meter data provided by Ausgrid, Australia.
To further reduce the carbon emissions of integrated energy systems (IES) and improve their energy utilization, an IES optimization scheduling strategy considering demand response (DR) under a stepped carbon trading mechanism was proposed. First, from the perspective of demand response (DR), considering the synergistic complementarity and flexible conversion ability of multiple energy sources, lateral time-shifting and vertical complementary alternative strategies for electricity, gas, and heat were introduced, and a DR model was constructed. Second, from the perspective of life-cycle assessment, the initial quota model of carbon emissions allowances was elaborated and revised. Subsequently, we introduced a tiered carbon trading mechanism that imposes a certain degree of constraint on the carbon emissions of IES. Finally, the sum of the energy purchase, carbon emission transaction, equipment maintenance, and demand response costs was minimized, and a low-carbon optimal scheduling model was constructed considering the safety constraints. This model transforms the original problem into a mixed-integer linear problem using Matlab software and optimizes the model using the CPLEX solver. The example results show that considering the carbon trading cost and demand response under the tiered carbon trading mechanism, the total operating cost of the IES is reduced by 5.69%, and the carbon emissions are reduced by 17.06%, which significantly improves the reliability, economy, and low-carbon performance of the IES.
Distributed generation using renewable energy is prevalent in cities, and energy storage is an effective method for addressing the uncertainty of its output. However, problems such as high costs and low utilization rates may be encountered in buildings that use energy storage. In the self-owned energy-storage sharing mode, an alliance mechanism based on a cooperative game was designed for urban buildings. Furthermore, a cooperative surplus distribution method, which considers the difference in energy storage investment, was developed based on an improved Raiffa solution algorithm. The simulation results demonstrate the applicability of the proposed alliance mechanism in an urban building cluster with different types of buildings, such as residential, office, and commercial buildings. Compared with the noncooperative mechanism, the cost was reduced by 10.22%, and the energy storage utilization rate increased by 18.12%. Moreover, the reduced amplitude of the cost of buildings with energy storage was much higher than that of buildings without energy storage. Therefore, the proposed cooperative surplus distribution method is reasonable, and the calculation volume is significantly reduced compared with that of the Shapley value method.
Accurate short-term forecasting of multienergy loads is the basis for the dispatch and operation of integrated energy systems. There is a strong coupling between multiple loads in an integrated energy system, and the existing single load forecasting is challenging to explore the complex internal relationship between multiple loads. Therefore, a short-term forecasting method for multienergy loads based on a multihead probabilistic sparse self-attention (MPSS) model was proposed. First, the Pearson correlation coefficient was used to analyze the correlation between multiple loads, the coupling features between multiple loads were extracted, a multihead probabilistic sparse self-attention mechanism with improved location coding was used to learn the dependencies of long-sequence inputs, and the parameter soft sharing mechanism of multivariate prediction tasks was adopted. The sharing mechanism realizes the joint prediction of multiple loads through a differentiated selection of shared features using different subtasks. Finally, the performance of the proposed model was verified using the multiple-load dataset of the Tempe Campus of Arizona State University. Compared with other forecasting models, the results show that the proposed multivariate load forecasting method can effectively improve forecasting accuracy.
The power sides of new power systems contain a large proportion of new energy sources. Owing to various types, distributed locations, and nonlinear outputs, the through-current level of the short circuit decreases, and its uncertainty increases when a fault occurs in the power grid, which makes it difficult for traditional current protection to preset and trip. Therefore, an adaptive current protection method based on real-time generated settings using local information is proposed, referring to traditional three-stage current protection. This method considers the actual fault characteristics of new energy sources to generate a preset. First, an equivalent source model is established that can distinguish the different outputs of new energy sources between generators and identify them online to obtain the unchanged parameters before and after the fault. Then, after a fault occurs, the local measured information is used to obtain the changeable parameters. Subsequently, preset values were generated in real-time based on the actual fault scenario and the actual output of new energy sources during the fault. Finally, the trip is decided based on the measured current and the preset value. Continuous identification before a fault occurs can effectively separate generators and new energy sources in the model. Online presetting based on actual fault scenarios and the output of new energy sources can effectively solve the problems of variety, distribution, and nonlinearity. The feasibility and correctness were verified by both analysis and comparison of various cases.
The emergence of a large number of new agents and power market transactions are crucial characteristics of new power systems. A planning model for source-network-load multiagent coordination and interaction based on game theory is proposed to comprehensively consider the interests of multiple subjects and market factors. First, the market relationship of each agent in the distribution network is analyzed, and the distribution locational marginal price is integrated into the optimal configuration model of the distributed power operators, distribution network operators, and power users. Second, a demand side response model is established based on the power utility function of the user. Third, combined with the game relationship between the three, a source-network-load coordinated interaction model is constructed, and an iterative search method based on the particle swarm optimization algorithm is used to solve it. Finally, the improved IEEE 33-bus system and the actual 9-bus system are used for the example analysis. The results show that considering the interest relationship of each subject in the electricity market environment can slow down the investment of lines and distributed power sources while ensuring the balance of interests among the subjects, which has practical significance.
The current subjectively fixed performance guarantee quota model lacks objectivity in the cost calculation of performance guarantees for power-selling companies, which affects transaction decisions. Therefore, this study proposes a medium- and long-term transaction optimization decision-making model for power-selling companies based on a dynamic performance guarantee quota. First, a quota adjustment coefficient was proposed, considering the credit overdraft level, market loss level, and national policy inclination of the power-selling company. Subsequently, a dynamic performance guarantee quota model was constructed. Second, to maximize the profits of power-selling companies, a medium- and long-term trading optimization decision-making model was constructed, considering the impact of changes in the cost of performance guarantee on the trading decisions. This model incorporates the dynamic adjustment of the performance guarantee quota. Finally, by analyzing the dynamic and fixed performance guarantee quota, performance guarantee costs, and transaction profits of power-selling companies participating in medium- and long-term transactions under multiple scenarios, the suitability and effectiveness of the proposed optimization decision model for the medium- and long-term transactions of power-selling companies based on the dynamic performance guarantee quota were verified.
The large-scale penetration of renewable energy sources poses significant challenges to the stable operation of power systems. Driven by double uncertainties on both the supply and demand sides, demand response resources based on terminal flexible loads need to be explored. Considering the load differentiation characteristics of different types of users, multitype load aggregators based on cooperation and win-win were introduced. Flexible dispatching of the power system was performed based on the complementary characteristics of the heterogeneous load response behaviors. Moreover, each load aggregator was assigned the dual status of a carbon trading integrator to enter the carbon trading market. A carbon trading model based on a reward-punishment ladder was constructed using the electricity load forecasting method to allocate carbon emission quotas for a system free of charge. Based on this, to minimize the sum of the operating costs of a cooperative alliance of multiple load aggregators, a pre-day optimization model of the interaction and cooperation among multiple aggregators was developed and solved. The Shapley value method was introduced for the cooperative game, and the cost was shared according to the contribution of each participant to the operation of the cooperative alliance. The results show that the overall and individual operational costs and the carbon emissions of the alliance are significantly reduced under the cooperative operation mechanism.
