[Objective] With large-scale access to distributed power sources, new energy storage, charging facilities, etc., the physical, digital, and commercial forms of distribution networks have undergone profound changes. The traditional planning method based on manual decision-making hinders distribution network optimization due to massive factors, complex structures, and numerous pieces of equipment. Artificial intelligence technology provides a feasible solution for overcoming the technical bottlenecks of distribution network planning. [Methods] In this context, this study analyzes the challenges faced by the distribution network planning process under new circumstances, including the precise spatiotemporal prediction of source-load, probabilistic balance of power and energy, coordinated planning of source-grid-load-storage, and empowerment of digitalization and intelligence. It elaborates on the current research status of artificial intelligence-based distribution network planning, focusing on key aspects such as knowledge graph construction, source-load scenario generation, power-energy balance, planning demand reduction, and intelligent network planning. [Results] This study summarizes and analyzes the issues in artificial intelligence-based distribution network planning technologies, including difficulties in processing unstructured and semi-structured data, limited scenario applicability, low accuracy in demand deduction, lack of interpretability, and high-dimensional solution spaces for planning schemes. It proposes potential solutions in technical research, such as graph learning, transfer learning, multimodal fusion, enhanced interpretability, and human-machine hybrid intelligence enhancement. [Conclusions] Compared with traditional distribution network planning methods, artificial intelligence-based distribution network planning demonstrates significant advantages of strong generalization, applicability, and scalability. However, it still faces critical issues, such as insufficient model accuracy and poor quality of generated solutions. In future work, we will continue to investigate artificial intelligence-based distribution network planning methods based on technical prospects, aiming to address the key challenges involved. This will provide references and insights for the development and digital-intelligent transformation of distribution network planning technology systems under the new power system framework.
[Objective] To address the problems of weak turn-to-turn short-circuit faults in dry-type air-core series reactors, which are difficult to recognize, and the lack of an early warning mechanism in traditional methods, this study proposes a multi-dimensional feature and intelligent algorithm fusion of an early fault diagnosis method. This method can overcome the lack of sensitivity of a single fault feature as it is easily interfered with by the noise of the fault leakage judgment. [Methods] First, the unbalance degree, power factor, zero sequence voltage, and characteristic impedance of the shunt capacitor bank are extracted as fault feature quantities, and their respective evolution laws after the fault are analyzed. Second, principal component analysis (PCA) is used to reduce the dimension and denoise the original data to eliminate interfering information. Subsequently, the denoised features with high saturation are input into the k-nearest neighbors (KNN) algorithm to construct a fault identification and classification model. Based on Maxwell, a field-circuit coupling model is established to generate single-turn, slight, and multi-turn short-circuit datasets; noise-free and 5% noise conditions are considered to verify the robustness of the algorithm. [Results] Simulation results show that the proposed method can achieve 100% recognition accuracy for minor turn-to-turn short circuits under both no noise and 5% noise, and manually adjusting the action threshold is not required. [Conclusions] This study realized high-precision early identification of weak faults through the three-stage architecture of “feature extraction-data noise reduction-intelligent classification.” The innovations include four-dimensional feature synergy to improve fault sensitivity, a PCA-KNN joint anti-noise mechanism; and an adaptive non-threshold discrimination system. The results provide a new idea for power-equipment condition monitoring, and the generalization ability of the model can be optimized by incorporating field data.
[Objective] The operational characteristics of the grid-forming modular multilevel converter (MMC) in offshore wind power transmission systems under short-circuit faults in the receiving-end grid are unclear, and the interaction mechanism between fault ride-through and grid-forming active support is not well understood, which poses a significant threat to the reliable operation of offshore wind power flexible direct current systems. In addition, traditional fault ride-through strategies face the issues of large transient impact currents and slow reactive power support responses during the switching of strategies before and after faults, in conjunction with grid control. [Methods] This study proposes a low-voltage fault internal potential reconstruction ride-through control strategy for grid-forming MMCs. Specifically, it analyzes the short-circuit current characteristics of MMCs under grid voltage dips, constructs a fault ride-through current-limiting boundary based on the voltage equation space vector of the grid-forming MMC, and reconstructs the internal potential using current optimization commands. This study aimed to demonstrate the capability of active voltage support in grid control under different grid voltage dips. [Results] The proposed fault ride-through strategy was validated using a semi-physical experimental platform, and the results indicated that low-voltage ride-through could be achieved under varying degrees of faults. As the grid-voltage dip deepens, the reactive power generated increases accordingly; however, the grid-forming current is limited at deeper dip levels. Compared with traditional fault ride-through strategies, this approach can reduce transient impact currents while quickly providing reactive power support and improving the recovery rate of the grid connection point voltage. The analysis revealed the operational characteristics of active voltage support in a grid-forming MMC and provided insights based on the vector angle correction of currents. [Conclusions] This study lays the foundation for the study of fault ride-through technology under severe voltage drop conditions in deep-sea environments, promoting the safe and reliable operation of large offshore wind power bases and effectively advancing innovation in the energy industry.
[Objective] A bidirectional converter integrates traction, feedback, and reactive power compensation and is the primary choice for upgrading the main equipment of an urban rail DC traction power supply system. In the field of low-voltage DC traction power supply, converters have the characteristics of low voltage and high current. The T-type three-level multi-unit parallel connection combined with carrier phase-shifting technology can meet the special requirements of low cost, low harmonic, high power density, high efficiency, and wide DC voltage operation range of urban rail DC traction power supply systems for the main equipment. However, this approach has problems such as excessive zero-sequence circulating currents and unbalanced midpoint potentials. [Methods] First, a zero-sequence circulating current model is established to address the circulating current problem in bidirectional converters. The circulating current is classified and analyzed, and a transformer low-voltage winding splitting, shared neutral line, L-shaped filter, and software control algorithm are proposed to suppress the zero-sequence circulating current. Second, in response to the problem of midpoint potential imbalance and based on the implementation of the equivalent space vector pulse width modulation, the relation of the midpoint potential to the phase voltage, phase current polarity, and space vector type is established by analyzing the T-shaped three-level single-phase topology structure. Accordingly, a feedback-compensation-based midpoint potential balance control method is proposed. [Results] Based on PSCAD/EMTDC simulation testing, it was found that various zero-sequence circulating currents could be effectively suppressed, achieving a midpoint potential balance in three-level converters during four-quadrant operation. [Conclusions] The proposed circulating current suppression method can not only suppress various zero sequence circulating currents, but also help reduce the cost of measuring the midpoint potential of three-level systems. The proposed midpoint potential control method does not require complex space vector modulation processes, has low computational complexity, is easy to implement, and can effectively control the midpoint potential.
[Objective] With the development of power systems and continuous expansion of energy systems, massive load power data have been generated. However, missing data are inevitable in the collection and transmission of power data, which greatly restricts the development of system-coordination optimization and advanced data applications. [Methods] To this end, this paper proposes a new power load missing data completion model based on an integrated graph convolutional variational transformer (IGCVT) network. The IGCVT model aggregates an improved graph convolutional network (GCN) and Transformer model using the variational auto-encoder (VAE) architecture. The raw data are processed by the GCN to learn spatial features and deeply mine spatial dependencies; the hidden layer data are reconstructed by the VAE to more effectively restore data distribution characteristics; and the temporal autocorrelation information of the sequence is mined based on the Transformer model. In addition, an improved whale optimization algorithm (WOA) is introduced to optimize the network model hyperparameters and improve the completion accuracy and applicability of the model. Simultaneously, to solve the problem of large errors in the completion of extreme change points of power load data, a two-way data completion method is adopted to make full use of the data information before and after the missing points. [Results] Experimental results show that, compared with the baseline model, the RMSE index is improved by 24.3%, 44.0%, and 47.9%, which verifies the superiority of the proposed method. [Conclusions] The results show that the proposed method provides a feasible solution to the problem of missing power load data and is expected to further expand the application scope of the model.
[Objective] The synchronization of information in intelligent substations is of great significance for the intelligent diagnosis and protection of power system faults, of which the satellite clock synchronization method is widely used. However, the satellite clock synchronization method is dependent on dedicated communication channels or satellite synchronization clocks, which can easily lead to the loss of synchronization signals. In addition, several synchronization signal transmission channels must be installed, and damage to transmission channels will cause synchronization errors between the master and slave clocks. Information synchronization technology based on the mutation detection algorithm that does not rely on clocks and communication channels, and is not limited by synchronous communication, can effectively avoid the problem of information synchronization caused by communication. [Methods] A multiplant information synchronization scheme based on the mutation detection algorithm is proposed. It compares the performance of typical outlier detection algorithms, such as the wavelet transform and singular value decomposition, in terms of pulse width, boundary effects, and phase-shift effects. [Results] The results show that the Hankel matrix singular value decomposition algorithm has the advantages of a narrow pulse width, no influence of boundary effects, zero phase shift, and has significant advantages in fast signal singularity detection compared with wavelet transform. Accordingly, a signal synchronization scheme based on the Hankel matrix singular value decomposition algorithm was proposed. The feasibility and advantages of the proposed information synchronization scheme were verified through an analysis using a large amount of on-site recorded data. [Conclusions] The results indicate that the proposed information synchronization scheme can achieve the precise synchronization of information between various protection devices and recorders within and between plant stations.
[Objective] As the energy consumption of data centers continues to grow and renewable electricity rapidly penetrates the market, the synergy between computing power, electricity, and heat can break the siloed development of these sectors, thereby promoting the high-quality and sustainable development of data centers and novel energy systems. However, the construction of a coordinated system and the formulation of standards for computing power, electricity, and heat lack regulation and guidance, which hinders the enhancement of energy utilization efficiency and coordinated optimization levels in data centers. [Methods] This study analyzes the current research and development status of computing power-electricity-heat collaborative systems in data centers. It summarizes the deficiencies of existing terminologies and provides key terminology examples related to the synergy of computing power, electricity, and heat from aspects such as general terms, integrated optimization, and evaluation terminology. [Results] The results indicate that the coordination among computing power, electricity, and heat in data centers is still in its preliminary exploratory stage, and normative work on cross-industry collaborative terminologies should be strengthened. [Conclusions] This study establishes a standard system for the coordination of computing power, electricity, and heat, which primarily includes basic, business, and service standards, with the aim of effectively guiding the deep collaboration between data centers, power grids, and district heating networks, promoting the diversified and intensive utilization of energy, and offering references and suggestions for advancing the standardization efforts in this field.
[Objective] With the rapid development of cloud computing and "Internet+", internet data centers (IDCs), as the core infrastructure underlying cloud computing, are in a rapid expansion phase. However, because both IDC and integrated energy systems (IES) possess underlying user information, data leakage may lead to various risks. Therefore, when designing collaborative optimization solutions for IDCs and IES, it is essential to consider the privacy preservation of both systems. [Methods] First, the flexible regulation characteristics of data centers were analyzed, and a flexible demand response model for data centers based on the graph theory with M/M/1 queuing theory was constructed. Then, a spatial and temporal joint planning model for an IES incorporating IDCs was established. Based on the Karush-Kuhn-Tucker (KKT) conditions of the IDC and IES operational models, the operational models were transformed into additional constraints for the planning model, which were linearized using the big-M method. Considering the privacy preservation requirements between the IDC and IES, an enhanced Benders decomposition algorithm for mixed-integer linear programming subproblems was improved, and a distributed solution framework was designed to solve the spatio-temporal joint planning model. [Results] The results show that under the example scenarios adopted in this study, after the implementation of the IES and IDC demand response models established in this study, the annualized total cost of the system decreased by 26.79%. The enhanced Benders decomposition algorithm shows that its distributed solution speed is 1.11 times faster than the alternating direction method of multipliers. [Conclusions] This study analyzed the flexible regulation methods of IDC and IES, and constructed a feasible distributed optimization scheme for IES containing IDC that considers privacy preservation. The study provides corresponding solutions and methodological references for similar multistakeholder collaborative optimization scenarios.
[Objective] A robust multi-stage planning model is established for a park-level integrated energy system (PIES) to minimize the total life-cycle present-value cost while considering source-load uncertainties, optimal construction sequencing, and tiered carbon trading to enhance economic efficiency and low-carbon benefits. [Methods] First, a box-type uncertainty set is used to model the source-load uncertainty and uncertainty adjustment parameters are introduced to reduce the conservatism of PIES planning and obtain a two-stage robust optimization (TSRO) model. Both discrete and continuous variables are included in the second-stage decision variables of the TSRO model, facilitating directly solving the second-stage problem using Lagrangian duality theory. Instead, an additional nested layer is required to solve the model. Therefore, this paper employs the nested column-and-constraint generation (NC&CG) algorithm to solve the model. [Results] Simulation results show that the robust programming model can improve the robustness of the system by sacrificing the economic and low-carbon benefits of the system to a certain extent and can flexibly adjust the robustness of the PIES planning scheme by changing the value of the uncertainty parameters. [Conclusions] The proposed multi-stage planning approach accounts for load growth and source-load uncertainties from a long-term system development perspective, thereby improving the flexibility and adaptability of the planning scheme.
[Objective] With the trend of energy consumption diversification, multiload forecasting plays an increasingly important role in optimizing the scheduling and operation planning of integrated energy systems(IES). [Methods] To address the problem in which the coupling relationship between multiple loads is often ignored in current integrated energy system load forecasting research, a multiple load joint forecasting method is proposed in this study for integrated energy systems based on the multi-energy demand response and improved bidirectional long short-term memory (BiLSTM). First, by integrating user demand response behavior, the input feature variables of the multi-energy demand response is constructed, and together with multiload forecasting, strong correlation features selected by the maximum information coefficient form the input feature set of the prediction model. Second, the crested porcupine optimizer is improved based on the chaotic mapping theory and elite reverse learning strategy to optimize the model parameters of the BiLSTM neural network. Finally, based on the multihead self-attention mechanism, the input feature weight is adaptively adjusted. The simulation results show that the prediction accuracy of the proposed multiload joint forecasting method is significantly improved compared with the single-load forecasting method. [Results] Compared with the multiload forecasting method without considering the demand response, the mean absolute percentage error of the electricity, heat, and cooling loads was reduced by 6.59%, 13.04%, and 24.86%, respectively. In addition, compared with other forecasting models, the model proposed in this study is more effective in improving the prediction accuracy and can achieve more accurate multi-element load forecasting. [Conclusions] The proposed load forecasting method was combined with integrated energy system dispatching to analyze the economic benefits of load forecasting. Compared with ordinary dispatching, the total operating cost of the system using the proposed load forecasting method was reduced by 16.49%, which can improve the comprehensive benefits of integrated energy systems.
[Objective] As the penetration rate of photovoltaics (PVs) increases, their volatility and randomness lead to intensified peak and valley fluctuations in the user net load, resulting in an increase in electricity demand. Energy storage can reduce the demand for electricity by utilizing the characteristics of peak shaving and valley filling; however, the high initial investment in energy storage limits its large-scale application on the user side. [Methods] A photovoltaic park energy storage optimal configuration method based on Stackelberg game pricing and information gap decision theory (IGDT) with electric vehicle (EV) demand response is proposed. First, considering the uncertainty of the grid, time-of-use, demand, and purchase and sale electricity price of EVs and PV output, an energy storage configuration model based on IGDT and an optimized operation model for EV clusters were constructed. Second, with the park as the leader and EVs as followers, a Stackelberg game model is constructed to minimize the costs of the park and EVs. Then, the Stackelberg game model is transformed into a mixed-integer linear programming problem for a solution using Karush-Kuhn-Tucker (KKT) conditions and the dual theorem of linear programming. Finally, we analyzed a PV park in a certain region as the research object. [Results] The results show that the proposed strategy reduced the annual comprehensive cost of the park by 12.06% and the charging and discharging costs of EV users by 54.88%. Owing to the participation of EVs in park scheduling, the storage configuration capacity and power were reduced by 62.80%, and the on-grid power by 1.32%, which improved the local consumption rate of PV. Compared with the IGDT model proposed in this study, the park cost of the robust optimization model is 1.97% higher, which proves that the IGDT model is more economical. [Conclusions] The comparison shows that the strategy proposed in this study meets the charging demand of EVs while reducing the comprehensive cost of the park, reduces the charging cost of the owners, and realizes a mutual benefit and win-win situation on both sides of the game.
[Objective] To accurately describe the impact of the volatility and randomness of the output power of new energy sources on the operation of an electricity-gas interconnected system and improve the robustness of its optimal operation, a robust optimization method based on multiple-interval uncertainty sets is proposed. [Methods] First, the traditional interval set of a distributed photovoltaic output is constructed, and then, a single-interval correlation uncertainty set is established considering the time correlation. Second, to solve the problems of poor robustness and large conservatism in the traditional set, the single-interval set is cut and divided to obtain a multi-interval correlation set. Furthermore, considering the energy conversion device, an active-reactive power coordinated robust optimization model based on the multi-interval uncertainty set is established for electricity-gas interconnection. [Results] The simulation results for the improved IEEE 33 node power grid and Belgium 20 node gas network interconnection system showed that the proposed method could reduce the conservatism and improve the robustness of the optimization results, thus proving the effectiveness of the proposed method. [Conclusions] The proposed active-reactive coordinated robust optimization method for an electricity-gas interconnected system based on the multiple interval uncertainty set can not only effectively deal with the challenges generated by the volatility of new energy sources but also improve the system’s robustness while reducing the conservatism. This method provides a new solution for the optimal operation of an interconnected electricity-gas system and is of great significance for promoting safe and economic operation of energy systems with access to a high proportion of new energy sources.
[Objective] Given the sudden drop in photovoltaic power generation in highway service-area microgrids and the interruption of interconnection lines with the large power grid caused by extreme events such as snowstorms and heavy rains, a two-stage optimization scheduling method for highway service-area microgrids, considering emergency power-supply vehicles, is proposed to ensure the stable energy supply of microgrids in service areas. [Methods] In the day-ahead stage, the flexibility of the dispatchable energy storage capacity in the integrated energy microgrid in the service area is considered. The optimization goal is to minimize the operation cost of the microgrid. A day-ahead economic scheduling model of the microgrid is established, and the lower limit of the charge state of the energy storage resource is obtained as a parameter of the intraday optimization scheduling stage. In this stage, based on model predictive control theory, an intraday rolling optimization scheduling model is established. Intraday rolling optimization considers the participation of emergency power-supply vehicles in the scheduling of service-area microgrids, and the various energy loads in the service area are graded according to their importance. The lowest operating cost of the emergency power-supply vehicle and lowest load-shedding penalty are taken as objective functions to obtain the final intraday optimization scheduling. [Results] By constructing simulation examples of photovoltaic output drop and grid interruption under extreme scenarios, the proposed method is compared and analyzed with the state of charge (SOC) optimization constraint scheme without energy storage and the scheduling scheme without considering an emergency power-supply vehicle. The simulation results indicate that the proposed method significantly reduces the curtailment of electrical, thermal, and cooling loads while concurrently achieving a reduction in system operational costs. [Conclusions] Case analysis verifies the effectiveness of the proposed two-stage optimization scheduling method for the service-area microgrid in extreme scenarios. The coordinated scheduling mechanism of the energy storage system and the emergency power-supply vehicle not only enhances the external shock resistance of the service-area microgrid but also achieves the dual goals of improving energy-supply stability and optimizing operating costs, ensuring a stable energy supply for important loads in the service-area microgrid.
[Objective] Under the advancing energy transition driven by China's national "dual-carbon" strategy, the escalating penetration of renewable energy (RE) sources has not only heightened power system spinning reserve requirements due to their inherent stochasticity and volatility in generation patterns, but has also precipitated a marked surge in peak-shaving and frequency regulation expenditures necessary for maintaining power supply reliability. This dynamic further exacerbates the fundamental multi-objective conflict between economic operation costs and reliability assurance in modern power systems. Particularly, tail risks triggered by extreme weather and the dynamic mismatches between stochastic RE fluctuations and conventional unit regulation rates invalidate conventional deterministic scheduling models reliant on typical scenarios. [Methods] To address this, this paper first constructs RE generation scenarios using Latin hypercube sampling (LHS) and modified k-means clustering, verifying their reserve feasibility, while transforming reserve-infeasible scenarios into extreme scenario sets. A two-stage distributionally robust optimization (DRO) model is proposed, minimizing day-ahead operational costs and intraday costs including carbon trading, rescheduling expenses, and risk penalties. A discrete probability ambiguity set with comprehensive norm constraints is established to rigorously characterize RE uncertainty by incorporating extreme scenarios. [Results] Case studies on an improved IEEE 39-node system using the column-and-constraint generation (C&CG) algorithm demonstrate that, compared with traditional deterministic and DRO models based on typical scenarios, the proposed approach increases scheduling costs by 7.11% and 14.37% respectively, but reduces renewable curtailment rates by 8.28% and 34.65%, and load shedding rates by 8.19% and 33.32%. [Conclusions] This methodology effectively resolves the limitations of conventional approaches in coordinating economic efficiency, reliability, and low-carbon requirements while ensuring system robustness, offering a viable solution for secure operations in renewable-dominated power systems.
[Objective] Photovoltaic power forecasting is a key technology to improve the efficiency of solar energy utilization and reduce operating costs. However, the traditional model has the problem of insufficient time-trend learning ability and error accumulation in multistep photovoltaic power prediction, which limits the improvement of prediction accuracy. [Methods] This paper presents a multistep prediction model for photovoltaic power generation based on a temporal convolutional network (TCN) and DLinear combined model. First, it improves the complete ensemble empirical mode decomposition with adaptive noise and ICEEMDAN decomposes multivariate meteorological sequences to reveal their potential features and obtain multidimensional subsequences that make it easier to learn multiscale features. Second, the TCN is used to model the local time sequence information and mine short-time sequence features. Finally, DLinear decomposes the sequence into trend and residual components, learns multiscale features through linear networks, and directly outputs multistep (four-step prediction, 15 min per step) photovoltaic power prediction results. [Results] Experimental results show that each module of the proposed method can significantly improve the prediction performance of the model. Compared with ICEEMDAN-CNN-BiLSTM, Informer, and Autoformer, the normalized root mean square error (NRMSE) decreased by 22.455%, 6.139%, and 8.504% on average, respectively, with obvious advantages. [Conclusions] Through the improved ICEEMDAN decomposition and TCN-DLinear combined model, this study effectively solves the shortcomings of traditional methods in multistep prediction and significantly improves the accuracy and reliability of photovoltaic power prediction. The research results provide a new technical path for the accurate prediction of photovoltaic power generation, theoretical support for the efficient management and operation of solar power generation, and data support for the safe and stable operation of new power systems. In the future, the generalizability of this model under different meteorological conditions and geographical environments should be further explored to promote the development of solar power generation technology.