This paper studies the characteristics of subsynchronous oscillation (SSO) power propagation in power systems with a large-scale wind farm integration. Based on the instantaneous power theory, a novel definition of SSO power is proposed to characterize its propagation. The SSO power contains both DC power components and AC power components. Utilizing SSO power, SSO propagation is quantitatively studied in single oscillation source systems, and further studied in multiple oscillation sources systems. Theoretical analysis reveals that in addition to the recognized impact of line impedance, power flow also affects SSO power propagation significantly. Hence, propagation impact factor is proposed for the determination of the dominant influencing element. To reveal SSO power propagation paths in a network, shunt coefficients of DC power components and AC power components are expressed respectively. Test cases under different operating conditions and a practical case are carried out to demonstrate analysis and conclusions in this paper.

Due to the low inertia and weak damping of the dc microgrid, its stability becomes even more serious and has attracted many investigations. However, as its scale and complexity continue to grow, the effectiveness, reliability, and comprehensive-ness of the stability analysis cannot be guaranteed simultaneously. Therefore, this paper proposes a passivity-based stability analysis approach by developing a physically decoupled model. Specifically, for the first time presenting the concept of virtual capacitance in modeling, the dc microgrid is physically decoupled into several low-dimensional subsystems. These subsystems have feedback structures, constituting the physically decoupled model, and the physical meanings of parameters are retained explicitly and completely. Subsequently, a sufficient condition for the stability of the physically decoupled model is derived based on strict passivity and LaSalle's Theorem. Then, by combining optimization theory, the stability domain calculation of multiple subsystems is transformed into a multi-objective optimization problem. The stability impacts of internal circuit and power parameters are evaluated, and the stability boundaries are identified to determine parameter limita-tions. Finally, numerical and time-domain simulations are carried out to validate the effectiveness and accuracy of the theoretical analysis and the superiority of the proposed method than Lyapunov direct method and TS fuzzy model method.

The interaction between the gird and wind farms has significant impact on the power grid, therefore prediction of the interaction between gird and wind farms is of great significance. In this paper, a wind turbine-gird interaction prediction model based on long short term memory (LSTM) network under the TensorFlow framework is presented. First, the multivariate time series was screened by principal component analysis (PCA) to reduce the data dimensionality. Secondly, the LSTM network is used to model the nonlinear relationship between the selected sequence of wind turbine network interactions and the actual output sequence of the wind farms, it is proved that it has higher accuracy and applicability by comparison with single LSTM model, Autoregressive Integrated Moving Average (ARIMA) model and Back Propagation Neural Network (BPNN) model, the Mean Absolute Percentage Error (MAPE) is 0.617%, 0.703%, 1.397% and 3.127%, respectively. Finally, the Prony algorithm was used to analyze the predicted data of the wind turbine-grid interactions. Based on the actual data, it is found that the oscillation frequencies of the predicted data from PCA-LSTM model are basically the same as the oscillation frequencies of the actual data, thus the feasibility of the model proposed for analyzing interaction between grid and wind turbines is verified.

This paper investigates the dispatch process of electric vehicles (EVs) when virtual power plants (VPPs) respond to electricity market demand and proposes an EV dispatch model that considers willingness to participate. The model considers the range anxiety and battery aging anxiety of EVs in the response scheduling process, which effectively improves the willingness of EVs to respond to scheduling. The model mainly involves three aspects: charging station siting, determination of EV incentive electricity price and realization of distributed dispatching. First, this paper proposes a three-step weighted mean clustering method to plan charging station siting, which reduces the driving distance of EVs when responding to grid dispatching, thus reducing range anxiety. At the same time, we studied the effect of different modifications in practical applications. Subsequently, we established an incentive price model for EVs to achieve reasonable quantitative compensation for the battery aging problem that occurs during the dispatch process. Finally, we propose a distributed scheduling strategy to improve the willingness of EVs to respond. This study is validated using real-world data from the city of Chicago. The proposed method reduces the average response distance of EVs in Chicago by 0.188 km and increases the participation willingness of EVs by 0.3243. These results demonstrate the advantages of the proposed approach in alleviating range anxiety and enhancing EVs’ willingness to participate in grid dispatch, offering valuable insights for largescale EV integration into grid interactions.

Electric vehicles (EVs) exhibit significant uncertainty in responding to grid dispatch due to the influence of EV users’ judgments and other subjective factors, which affect grid scheduling. These subjective factors are difficult to model using conventional methods. Therefore, this article proposes a flexible EV scheduling strategy based on the ride-hailing dispatch model. This strategy adopts an “order-grabbing” mechanism, where the dispatch center broadcasts demand and EVs respond based on their current status. By dynamically adjusting the dispatch price, the strategy incentivizes EV participation, enhancing flexibility in EV-grid interactions. First, the initial dispatch price is set according to grid demand and EV battery capacity, and dispatch information is sent to EV users. A credit compensation mechanism is introduced to link EVs’ historical response with their rewards, thereby standardizing response behavior. Subsequently, a response willingness model is developed to associate the dispatch price and other influencing factors with response capacity, providing a basis for price adjustments. MATLAB simulations demonstrate that the proposed method successfully completed scheduling in four different scenarios after 1–3 rounds of dynamic price adjustments, demonstrating the feasibility of the strategy. The final average response error using the proposed method was 25 kWh, representing a 61.53% reduction compared to the benchmark method, while scheduling costs decreased by 10.81%, validating both the accuracy and economic efficiency of the strategy. Compared to existing studies, this article innovatively applies the ride-hailing model’s concepts of broadcast demand and order-grabbing response to effectively address the subjective uncertainty of EV users in the response process.

The uncertainty of distributed resource (DR) aggregation responses has become a key bottleneck restricting the reliable scheduling of virtual power plants (VPPs). To mitigate this uncertainty, this paper explores the strong coupling between residential behavior and the adjustable capability of distributed resources, and proposes a dynamic aggregation strategy that accounts for such coupling. First, the concept of Behavior-sensitive Distributed Resources (BDRs) is proposed and defined, where resources whose adjustable capability is significantly influenced by the number of residential users are categorized as BDRs. By applying the Discrete Fourier Transform (DFT), this study, for the first time, reveals the daily periodic characteristics of BDRs. Subsequently, a segmented confidence interval estimation method is introduced, which, in conjunction with the daily periodicity of BDRs, enables precise characterization of adjustable capability across different time periods, thus improving the accuracy of response capability modeling. An improved two-stage greedy aggregation algorithm is then designed to simultaneously reduce the uncertainty of VPP adjustable capability after aggregation and optimize economic performance. Finally, the strategy is validated using real-world data from a resource set that includes three representative types of BDRs: electric vehicle (EV) charging stations, commercial lighting (CL), and office building heating, ventilation, and air conditioning (HVAC) systems. The results demonstrate that, compared to existing approaches, the proposed aggregation strategy reduces VPP response uncertainty by over 22.63% and lowers aggregation cost by an average of 8.44%, thereby confirming its advantages in enhancing both grid reliability and economic efficiency.