Abstract:In the new power systems dominated by renewable energy resources, the regulation capacity on the power generation side is significantly reduced. It is urgent to unlock the flexibility of the demand-side resources and make up for the system regulation shortage caused by integration of renewable energy resources. First, it is pointed out that the demand response mechanism for the new power systems should have the four characteristics of “scalable, permanent, lightweight, and attractive”. Then, in view of the problem that the current demand response classification method is rough and difficult to reflect its essential characteristics, according to the classification principles of customer contribution evaluation methods in the demand response, the existing mechanisms are reclassified into three types of the price-based demand response, the customer baseline load-based demand response, and the customer directrix load-based demand response. The research status of various mechanisms is introduced in detail. Moreover, based on the proposed four characterized features, the analysis points out that it is difficult for the price-based and customer baseline load-based mechanisms to adapt to the development of new power systems. Different from these two types of mechanisms, which try to indirectly change the user’s electricity consumption behavior by using a certain signal, the customer directrix load-based demand response directly depicts the profiles of the users’ load curves expected by the demand response implementers, which can better meet the four characterized features. It fundamentally overcomes the problems of other mechanisms such as small scale, unsustainable, opaque, and easy to rebound, and can be well adapted to the development needs of the new power systems. Finally, the necessity, feasibility and importance of the customer directrix load-based demand response mechanism in the development of new power systems are prospected.

Abstract:Vigorously developing renewable energy and building a clean and efficient energy system has become an inevitable choice for implementing the energy revolution. In the electricity market environment, as the penetration of renewable energy keeps increasing, the share of traditional thermal power is constantly occupied. Since a reasonable market mechanism has not been established to guarantee the basic benefits of thermal power supply capacity, thermal power enterprises lack the incentive to provide auxiliary services for high proportion of renewable energy. The core idea of market value distribution lies in quantifying the contribution and value of market members and distributing benefits according to “Who creates the value who shares, who produces the cost who bears”. Therefore, a benefit allocation mechanism of electricity markets with penetration of high proportion of renewable energy is proposed. Based on the Vickrey-Clarke-Groves (VCG) mechanism, the value of a thermal power unit is defined as the substitution benefit of the unit to other power generation resources, and the true cost of thermal power for renewable energy grid-connected accommodation is accurately identified, thereby stimulating thermal power units to make truthful bids and actively reserve auxiliary capacity, forming a pattern of mutual benefit for the electricity markets with high proportion of renewable energy.

Abstract:As an important new type of flexible load resource, electric vehicles (EVs) have the potential to provide regulatory services for the power grid. However, the controllability of EVs is affected by their traveling arrangements, which shows the regularity and uncertainty. Firstly, this paper proposes a flexibility model of EVs, which can consider the difference in the data structure of public charging piles and specific charging piles. Besides, the uncertainty of the flexibility is described using the data-driven distributionally robust chance constraint (DRCC). Furthermore, considering a charging operator aggregating different types of charging stations participating in the peak regulation ancillary service market, the paper proposes a bi-layer bidding and allocating model, which enables the charging operator to manage the risk outside and to flexibly dispatch resources inside. Simulation results based on the actual data indicate that the flexibility described by the DRCC can adapt to the random changes of traveling law of EVs, and the proposed bidding and allocating strategy can effectively manage the revenue of the charging operator and the default risk.

Abstract:Electricity retailers face multiple uncertainties such as load, spot price and demand response when participating in market transactions. Comprehensively considering multiple uncertainties and modeling according to their characteristics are conducive to improve the effectiveness of electricity procurement and sale decision-making. First, a response optimization model of transferable load under time of use tariff is proposed, which takes into account the satisfaction with electricity cost and electricity consumption comfort. Then, considering the uncertainty of user response behavior, a reducible load response model based on fuzzy membership function and consumer psychology is established. Next, the probability density function and scenario method are used to simulate the load and spot price respectively. Aiming at maximizing the comprehensive utility of expected profit and conditional risk loss, a comprehensive decision-making and risk assessment model of electricity retailers including power procurement portfolio, power sale pricing, reducible load deployment and compensation is constructed. The model is transformed into a deterministic optimization problem for solution by equivalent transformation of fuzzy chance constraints. The influences of risk preference and uncertainties on the decision-making of electricity retailers are simulated and analyzed, and the effectiveness of the model is verified.

Abstract:The stochastic volatility of renewable energy generation and the time series coupling characteristics of energy storage operation control bring many challenges to the energy management and optimal operation of microgrid, which becomes a hot issue for academic research. This paper proposes an energy management and optimization strategy for microgrid based on improved dueling deep Q network (DQN) algorithm. The strategy adopts a multi-parameter operation exploration mechanism and an optimally designed neural network structure to learn the environmental information such as the power output of distributed renewable energy, the electricity price of energy trading market and the state of electric load, and applies the learned strategy to microgrid energy management and optimization. The simulation results show that the performance of the energy management and optimization strategy for microgrid based on the improved dueling DQN algorithm is better than the scenario-based stochastic programming algorithm, the DQN algorithm and the dueling DQN algorithm.

Abstract:Uncertainties of loads and distributed generators are the difficulties in the microgrid operation optimization. This paper proposes a day-ahead distributionally robust optimization method for grid-connected microgrid considering extreme scenarios to ensure the robustness and efficiency of microgrid energy management. Firstly, considering the uncertainty of sources and loads, a fuzzy set of probability distribution based on Wasserstein distance is constructed, and the fuzzy set is modified by the extreme scenario method to improve the robustness of the fuzzy set. Secondly, considering that the microgrid and the distribution network are different stakeholders in the grid-connected mode, the day-ahead operation optimization models of the microgrid and the distribution network are established respectively, and the improved analytical target cascading is used to carry out decoupling and iterative solution. Finally, effectiveness of the proposed method is verified by the simulation comparison.

Abstract:China is one of the countries that suffered the most from typhoon disasters. Strong typhoons can lead to mass disconnection and tower collapse, which can then evolve into electric disasters, and the risks cannot be ignored. In response to typhoon disasters, the PageRank algorithm is used to establish a distribution network failure-rate model that takes into account the relevance between disconnection and tower collapse. Meanwhile, for the more complex part of meteorological and geographic information in the model, the data-driven method is used to introduce the information disturbance factor μ into the relevant distribution network failure-rate model. The mapping relationship between meteorological and geographic information and μ is built through the radial basis function (RBF) neural network model, and the driving results are verified and corrected. Finally, a new quantitative index of resilience is proposed, which takes the optimum index and maximized economy as the upper and lower objective functions, and a two-level planning model for the siting and sizing of flexible resources in typhoon disasters is constructed considering the relevance impact. By coordinating the planning of flexible resources, the distribution network achieves the economic optimum while increasing its resilience. The accuracy and effectiveness of the model are verified by calculation examples.

Abstract:Wind disaster may lead to large-scale power failure in distribution networks. Effective prediction of user number in power outage for distribution networks can provide auxiliary guidance for emergency repair. This paper proposes a prediction model for user number in power outage for distribution networks based on high-efficient data dimensionality reduction. First, based on the data-driven method and considering all characteristic variables, i.e., global variables, a prediction model for user number in power outage for distribution networks based on random forest algorithm is constructed. Then, the relationship between various characteristic variables and response variables affecting power outage users of distribution networks during typhoon is analyzed in detail by using partial dependence plots. The characteristic dimension is reduced and the secondary modeling forms a more efficient prediction model for user number in power outage for distribution networks after dimensionality reduction. Compared with the prediction model considering global variables, the prediction model for user number in power outage for distribution networks after characteristic dimensionality reduction reduces the burden of data collection to a certain extent, improves the calculation efficiency, and provides an effective basis for disaster prevention and mitigation of power grid.

Abstract:The data-driven dynamic security assessment under the concept of security region urgently needs a systematic fault location feature representation method. Based on the concept of the electrical coordinate system, the identification of blind areas for fault location feature representation and an expansion method of electrical coordinate system are proposed, which further improve the representation accuracy of fault location in partial areas. First, the reasons why the representation accuracy of the electrical coordinate system is insufficient in partial areas are analyzed. The blind point, blind line and blind area of fault location feature representation are defined. Then, by analyzing the locational relationship between two points, the judgment principle of blind lines is determined, and the rapid searching method of blind lines considering adjacent lines is proposed. Finally, the identification algorithm of initial blind area based on adjacency matrix is proposed, and the blind area is fine-tuned based on the model assessment results. The fault location representation accuracy in the blind area is improved by the expansion of reference nodes of the electrical coordinate system. Taking the transient stability assessment as an example, the applicability of the proposed method in the dynamic security assessment is verified by the example of a provincial power grid in China.

Abstract:The tight coupling between different electric-heat-gas energy supply systems makes it difficult to identify the weak nodes in the integrated energy system (IES). From a data-driven perspective, a weak node identification method of IES based on the big data from the unified power flow model is proposed. Firstly, a unified power flow model for the IES is established based on the physical characteristics of the power, heat, and natural gas systems and their coupling links. Secondly, the high-dimensional state matrix is constructed using historical and real-time data of unbalanced initial value of the power flow model. Thirdly, the M-P law and ring law are used to qualitatively analyze the operation states of the system. Finally, the average spectral radius (MSR), the maximum eigenvalue of the state matrix, and the node weakness index are calculated with the combination of the entropy theory. The proposed method maps the node weakness information to data changes, and the identification of the relative weakness of nodes in IES are realized by sorting the index values of the node weakness. The simulation results verify the effectiveness of the proposed method.

Abstract:In recent years, dynamic state estimation for generators based on synchrophasor measurement unit (PMU) measurement data has attracted more and more attention. However, the PMU operation in the field is affected by many factors, which may cause the generator terminal voltage or current measurement phasors which are the input of the state estimation to have bad data and affect the state estimation results. To solve this problem, a dynamic state estimation method for generators considering the bad data in input is proposed. After analyzing the influence of input bad data on dynamic state estimation, this method uses the characteristic that the generator states are restricted by the dynamic equation and cannot change suddenly, and a state prediction method based on exponential smoothing is proposed. The predicted states are compared with the forecasting states computed by the generator dynamic equation, and a method for detecting bad input data is proposed. Furthermore, when detecting bad data, predicted states by exponential smoothing are used to replace the forecasting states of the dynamic equation, and an input correction method based on the least square method is proposed. The results of simulation examples show that the method can effectively suppress the influence of input bad data on dynamic state estimation for generators and improve the robustness of the dynamic state estimation algorithm.

Abstract:The share of intermittent renewable energy is continuously increasing in modern power systems, which significantly declines the efficiency of traditional methods for power system reliability evaluation. Therefore, a non-parametric stratified importance sampling method for reliability evaluation of the power system is proposed, which solves the problem that it is difficult to control sample correlation when Latin hypercube sampling is combined with non-parametric importance sampling, and realizes the organic combination of these two methods. This method can concentrate the system state samples in important areas that have a large contribution to the reliability index, and reduce the repetition of samples, so it can significantly improve the efficiency of reliability evaluation of modern power systems with multiple renewable energy plants. The IEEE RTS-79 standard test system is modified and simulation examples are carried out by using the historical output data of actual wind farms, photovoltaic power plants and hydropower plants. The results show that the calculation speed of the proposed non-parametric stratified importance sampling method is significantly faster than that of individual importance sampling method. The effectiveness of the method has been proven.

Abstract:Because the dynamic gain of excitation systems has a significant impact on the dynamic stability of salient-pole generators, it is often necessary to adjust the dynamic gain and then configure the power system stabilizer (PSS) in order to improve the damping of the power system. However, there are some unclear understandings about the impact of dynamic gain on the dynamic stability of generators. Firstly, the impact of dynamic gain of excitation systems on the dynamic damping is analyzed. Then, the impact of dynamic gain of excitation systems on the damping torque, damping ratio of salient-pole generators and the frequency response characteristics of excitation systems is calculated and analyzed with typical unit parameters and operation conditions. The simulation is carried out in the single-machine infinite system, two-machine system and real system. The calculated results are consistent with the simulation results. Finally, the experiments are carried out in the laboratory and the engineering field, and the correctness of the simulation and calculation results is verified, namely the larger the dynamic gain of excitation system is, the weaker the dynamic damping of salient-pole generator is. In order to improve the dynamic damping of salient-pole generators (such as hydrogenerator), the dynamic excitation gain should be smaller on the premise of meeting the requirements of relevant standards, and the system damping should be further improved by configuring corresponding PSS.

Abstract:In order to visualize the complex spatio-temporal relationships and multi-source heterogeneous data in large power grids, a visualized architecture of power grid for the dispatching and control cloud is proposed. For the visualization of the full-scene varying from specific equipment to grid overview, the full-state of life cycle from history to the future and the full-information including both elements and topics, a four-layer framework is adopted, which is composed of the grid diagram, station diagram, primary equipment diagram, and secondary equipment diagram. This complex architecture allows the illustrations of full-scene grid, full-state grid and full-information grid, including the power system characteristics, operation trajectories and real-time information to become thoroughly descriptive in the areas of full-scene grid, full-state grid and full-information grid. On this basis, the key technologies to realize the visual power grid are studied. Based on the correlation relationship between electrical relationship and object-oriented model, the visualized correlation map is automatically constructed to comprehensively display the composition and operation of the power grid from multiple levels of scene, state and information, so as to support major applications including real-time operation, operation efficiency analysis, equipment failure analysis, and equipment examination analysis, etc.

Abstract:On the background of in-depth reform of the electricity market, integrated energy service providers are not only the main energy providers directly providing energy services for users, but also an important carrier of energy conservation and emission reduction. Research on the cooperative operation of integrated energy service providers will help to promote the low-carbon economy of regional integrated energy systems. Firstly, this paper proposes the regulatory framework of two-layer collaborative optimization operation, which mobilizes the enthusiasm of comprehensive energy service providers at the bottom layer to participate in the carbon emission reduction through top-layer supervision, and the stakeholders at the bottom layer with limited information to achieve collaborative optimization. Secondly, based on Nash bargaining theory, a cooperative game model of integrated energy service providers considering carbon emission limits and penalty coefficients is established in order to explore the potential of cooperative operation of integrated energy service providers to promote carbon emission reduction and improve individual interests and social benefits. Then, the non-convex model is decomposed into two convex optimization problems under linear constraints, namely the benefit maximization sub-problem and the energy transaction payment sub-problem. The adaptive alternating direction method of multipliers is further used to solve the above problems. Finally, the effectiveness of the proposed strategy is verified by examples.

Abstract:Offshore wind power is faced with the deep coupling effect of complex and changeable meteorological and sea conditions, which leads to the accuracy of output forecasting to be improved. At the same time, the “black box” structure of the forecasting model leads to the low reliability of output forecasting results in the engineering applications. To address the above problems, an ultra-short-term offshore wind power output forecasting model is proposed based on the dual-stage attention long short-term memory (DALSTM) network. Based on the long short-term memory neural network, the dual-stage attention mechanism of feature space and time series is introduced to dynamically explore the potential correlations between the offshore wind power output and the input features. The importance evaluation is carried out from feature and time to improve the interpretability of the proposed DALSTM model. Finally, simulations are conducted on the supervisory control and data acquisition data collected from the Donghai Bridge offshore wind farm of China. The results show that the proposed DALSTM model can effectively forecast the ultra-short-term offshore wind power output, and has higher forecasting accuracy and stability than the traditional forecasting models, as well as reasonable interpretability.

Abstract:In order to improve the operation characteristics of the direct drive-doubly fed hybrid wind power system under non-ideal grid voltage in a distributed wind power application scenario, a scheme is proposed to realize the flexible fault voltage ride-through in the hybrid distributed wind power system by applying the unified power quality conditioner (UPQC) with a nine-switch converter. Focusing on the control of the nine-switch converter, based on the analysis of the DC-side voltage sharing relationship of the nine-switch converter, the third harmonic injection modulation method for the nine-switch converter is studied to improve the DC voltage utilization rate. Considering the difference in the DC voltage demand of the nine-switch converter, a strategy of dynamically assigning modulation ratio limiting based on the degree of voltage failure is designed. Through the simulation analysis of the above schemes under voltage sag, rise, and harmonics and asymmetrical conditions, the feasibility of the nine-switch converter to optimize the operation characteristics of direct drive-doubly fed hybrid distributed wind power is verified.

Abstract:To solve the difficulty of voltage regulation of the LLC resonant converter in the high-power application scenarios of DC distribution network, such as renewable energy sources, energy storage, data center, power electronic transformer, this paper proposes a voltage-regulatable resonant zero-voltage zero-current switching (ZVZCS) converter with an auxiliary transformer. The auxiliary transformer and auxiliary half bridge are added in the proposed converter based on the traditional LLC resonant converter. The duration that the auxiliary transformer participates in the boost can be adjusted by changing the phase-shifting angle between the auxiliary half bridge and the primary-side full bridge. Thus, the voltage regulation function is realized on the premise of ensuring the soft switching of the main switch, and the auxiliary switches can realize zero-voltage turn-on. The circuit structure, working principle and parameter design process of the converter are described in detail. The principle of the proposed topology and the correctness of the parameter design are verified through a 100 kW converter simulation model. The simulation results show that the proposed converter has better efficiency in the high-power scenario than the traditional LLC resonant converter. Finally, a 2.5 kW experimental prototype is built to verify the effectiveness of the proposed topology.

Abstract:Lithium-ion batteries are commonly used as loads in current wireless power supply applications. The wireless battery charger should provide an initial constant current (CC) and a subsequent constant voltage (CV) output required by the battery. The additional converter at the secondary side, the hybrid topology, or the high-order network can realize CC-to-CV transition, which have problems of complicated control and poor reliability. To solve this problem, this paper proposes a three-coil wireless battery charging system for self-adapting to the battery charging curve. By adding an auxiliary coil and a passive rectifier bridge on the primary side, the system can automatically realize CC-to-CV transition without any additional control. The auxiliary coil is connected in parallel with the input source through a specific compensation network, and the output voltage threshold is not limited by the magnetic coupling mechanism parameter. Thus the self-adapting to the battery charging curve is realized. Meanwhile, the proposed three-coil wireless battery charger operates at the same frequency in both charging modes, which ensures the soft switching of the inverter and the near zero reactive power circulation of the converter, to reduce the device stress and improve the transmission efficiency. Moreover, the proposed charger has the function of automatic open-circuit protection. In this paper, the design principles and operation principle of the charger are systematically described, and an experimental prototype with 72 V/2.5 A output is built to verify the effectiveness of the proposed charger.

Abstract:For the problems of high resistance tolerance, noise interference, and operation refusal on remote high-resistance internal fault of the existing flexible DC line protection, this paper proposes a single-end protection scheme for the flexible DC transmission line based on the similarity of energy spectrum matrix of traveling waves. By analyzing the boundary effect of current limiting inductors in the flexible DC power grid, the difference of voltage traveling waves of internal and external faults in time and frequency domains is obtained. The S-transform is used to construct the energy spectrum matrix to increase the difference of energy spectrum distribution between internal and external faults. The principal components analysis is used to reduce the dimension and the storage space of sample data. By using the normalized cross-correlation coefficient, the similarity between the fault energy spectrum matrix and the fault sample matrix is calculated, and the single-end flexible DC protection criterion is constructed to distinguish the internal and external faults. Considering the factor that overhead lines are vulnerable to lightning, the lightning interference identification criterion is constructed. Simulation results show that the proposed algorithm is not affected by the fault location, the high resistance and the noise interference.

Abstract:The development of integrated energy system plays a huge role in improving energy efficiency, reducing carbon emissions and increasing the permeability of renewable energy. Faced with the increasingly closely coupled integrated energy system of electricity-heat and electricity-gas interconnection, the existing energy management mode and scheduling means cannot give full play to their due advantages. Therefore, the realization of the integrated, efficient and accurate state estimation of the integrated energy system can provide reliable data support for the subsequent collaborative scheduling and safe operation. In view of this, this paper briefly summarizes the research on the state estimation of the integrated energy system in the background of carbon neutrality, and reviews the development history and difficulties of the theory of state estimation for the integrated energy system. The general idea of the research on the state estimation of the electricity-heat and electricity-gas integrated energy systems is analyzed from the three aspects of model, data and time scale. Finally, the future research directions of the integrated energy system state estimation are prospected.