Multi-level power grid voltage coordination control method and system considering distributed power generation access

Abstract

The present invention discloses a multi-level grid voltage coordinated control method and system considering the access of distributed power sources. It relates to the technical field of grid voltage coordinated control, and includes the following steps: obtaining grid information data, constructing an ultra-short-term prediction module based on TimeUer, and obtaining the power prediction value of the grid node; obtaining the basic reactive power demand based on the power prediction value of the grid node; considering the dynamic impact of electric vehicle charging piles on the reactive power demand of the grid, calculating the increase in reactive power demand caused by charging, and obtaining the second reactive power demand; obtaining the maximum reactive power output, and comparing it with the second reactive power demand, and performing multi-level coordination of the grid voltage according to the comparison result. The present invention achieves the effect of coordinated regulation of the voltage of each node of the grid to which distributed power sources are connected, and solves the problem of limited accuracy and efficiency in predicting power changes of grid nodes in the prior art.

Description

Technical Field

[0001] The present invention relates to the fields of time series prediction, blockchain technology, reactive power compensation and resource optimization scheduling, and in particular to a multi-level power grid voltage coordination control method and system considering the access of distributed power sources. Background Art

[0002] With the profound transformation of the global energy structure and the large-scale grid connection of renewable energy, the proportion of distributed power sources such as solar photovoltaic, wind power, and gas-fired power generation in the grid is increasing, bringing new challenges to the stable operation of the power system. The output of distributed power sources is affected by many factors such as weather conditions and geographical location, and has significant volatility and uncertainty, which poses a severe test for the voltage control and power balance of the power grid.

[0003] In response to the challenge of coordinated control of grid voltage after the access of distributed generation, the current control methods, such as adjusting the tap of on-load tap-changing transformers, adding reactive power compensation devices, and using grid-connected inverters for reactive power regulation, are effective to a certain extent, but the methods are relatively simple and fail to fully consider the urgent need for voltage coordination between grids at different levels after the access of distributed generation.

[0004] With the increasing proportion of distributed power sources and the increasing diversification of load demands, traditional voltage control methods have been unable to meet the needs of stable grid operation due to the lack of accurate mathematical models and optimized decision-making processes, resulting in limited accuracy and efficiency of voltage control. Due to the access of distributed power sources, the power changes at grid nodes have become more complex and difficult to predict, which not only affects local voltage fluctuations, but may also cause reactive power changes at different nodes, thereby posing a threat to the voltage quality of the entire grid. Therefore, accurately and efficiently predicting the power changes at grid nodes is crucial to achieving overall voltage coordination control of multi-level grids.

[0005] However, most of the existing voltage coordination control methods fail to fully pay attention to the importance of this prediction, resulting in the lack of solid data support and accuracy guarantee for the formulation of voltage control methods. Traditional prediction methods based on statistical methods and deep learning methods, such as ARIMAX and SARIMAX, have limitations in dealing with the irregularity and heterogeneity of input sequences, thus limiting the accuracy of distributed generation node power prediction.

[0006] In addition, with the popularization of electric vehicles, a large number of electric vehicle charging piles connected to the power grid have also caused a certain degree of fluctuation in the grid voltage. At the same time, the storage efficiency, integrity and reliability of the grid's historical information are also crucial to the grid voltage stabilization process.

[0007] For example, the invention with publication number CN109120011B proposes a method for congestion scheduling of a distributed distribution network considering distributed power sources, which belongs to the field of power system operation and control technology. The method first establishes a distributed distribution network congestion scheduling optimization model considering distributed power sources, which is composed of an objective function and constraints; relaxes the model; transforms the model by defining new variables, and then performs distributed iterative solution on the model to obtain a distributed distribution network congestion scheduling scheme considering distributed power sources. The method of the present invention takes into account the impact and role of distributed power sources, gives full play to the flexible advantages of distributed power sources, reduces the congestion level of the network, and uses a distributed method to solve and calculate the model, which can quickly reduce the local congestion level.

[0008] For example, the invention with the publication number CN108134394B discloses an optimized load shedding method considering the influence of distributed power sources. Each node of the distribution network uploads the interruptible load at the current moment and the load priority to the centralized controller of the root node. The centralized controller establishes an optimized load shedding model considering the equivalent value of the external network based on the uploaded information and the flow constraint; the centralized controller converts the nonlinear flow constraint into a convex second-order cone form by the second-order cone relaxation method, so as to quickly calculate the optimized load shedding amount considering the influence of the output of the distributed power source, and sends the information to each node control unit; after each node performs the load shedding action, the measurement unit monitors whether the node voltage returns to the rated range and determines whether the current round of load shedding process is completed. This method can realize fast and accurate load shedding control in a distribution network containing a large number of distributed power sources, restore the distribution network voltage to the maximum extent, and at the same time ensure that the voltage of the distributed power source does not exceed the limit, avoiding the overvoltage tripping of the distributed power source caused by inappropriate load shedding.

[0009] Although some grid scheduling and load reduction methods that take into account the impact of distributed power sources have been proposed in the prior art, these methods have alleviated the grid pressure caused by the access of distributed power sources to a certain extent, but there are still limitations in the prediction accuracy and efficiency of grid node power changes. Summary of the invention

[0010] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a multi-level power grid voltage coordination control method and system taking into account the access of distributed power sources. By introducing advanced time series prediction models, blockchain technology, voltage stabilization calculation methods and resource coordination scheduling methods, coordinated regulation of voltages at various nodes of the power grid to which distributed power sources are connected is achieved, thereby improving the stability and reliability of power grid operation, so as to solve the problem in the prior art that the accuracy and efficiency of predicting power changes at power grid nodes are limited.

[0011] To achieve the above object, the present invention provides the following technical solutions:

[0012] A multi-level grid voltage coordination control method considering the access of distributed power sources includes the following steps: acquiring grid information data, constructing an ultra-short-term prediction module based on TimeUer, and obtaining power prediction values ​​of grid nodes; obtaining basic reactive power requirements based on the power prediction values ​​of grid nodes; considering the dynamic impact of electric vehicle charging piles on the reactive power demand of the grid, calculating the increase in reactive power demand caused by charging, and obtaining a second reactive power demand; obtaining the maximum reactive power output, and comparing it with the second reactive power demand, and performing multi-level coordination of the grid voltage according to the comparison result.

[0013] In a preferred embodiment, the step of acquiring power grid information data specifically includes:

[0014] Combined with the chain storage structure, several smart contracts are preset to execute the data access and management operations corresponding to the data tables in each storage structure respectively; the real-time acquired power grid information data is transmitted in batches to the blockchain network through the edge nodes of the blockchain platform according to the preset data format and sending frequency; the monitoring terminal of the power grid obtains the power grid information data through the blockchain network, and the blockchain network and the monitoring terminal transmit data through encryption methods.

[0015] In a preferred embodiment, the power prediction value of the power grid node is specifically as follows: the power grid information data is cut and processed to obtain a number of non-overlapping time blocks, and the number of non-overlapping time blocks are embedded to obtain a first embedding vector and a second auxiliary information data; the self-attention mechanism is combined to obtain the internal time dependency between the first embedding vectors to obtain a first feature map; based on the spatial attention mechanism, the first feature map is subjected to feature extraction in the spatial dimension to obtain a second feature map; based on the cross attention mechanism, the correlation between the first embedding vector and the second auxiliary information data is captured to obtain the correlation between the variables; based on the TAU time attention mechanism, the second feature map is combined to obtain feature information containing time changes; the feature information is subjected to linear projection processing to obtain the power prediction value of the power grid node.

[0016] In a preferred embodiment, the embedding operation is performed on several non-overlapping time blocks to obtain the first embedding vector and the second auxiliary information data, specifically: the predicted target data in the power grid information data is segmented to obtain several non-overlapping time blocks; the embedding operation is performed according to the segmented time blocks and the complete input sequence respectively to obtain several first embedding vectors, wherein the first embedding vector includes a time block embedding vector and an input sequence embedding vector; the embedding operation is performed on the non-target data according to each complete input sequence to obtain the second auxiliary information data.

[0017] In a preferred embodiment, the TAU temporal attention mechanism is combined with the second feature map to obtain feature information containing time changes, specifically: based on the TAU temporal attention mechanism, the temporal attention is decomposed into intra-frame static attention and inter-frame dynamic attention; the intra-frame static attention and the inter-frame dynamic attention are inner-producted, and the Hadamard product is performed with the original feature map to obtain the feature information containing time changes.

[0018] In a preferred embodiment, the second reactive power demand is obtained by: obtaining the charging data of the charging pile during peak power consumption, and calculating the increase in reactive power demand based on the power characteristic curve of the charging pile and the number of charges; superimposing the increase in reactive power demand with the basic reactive power demand to obtain the second reactive power demand.

[0019] In a preferred embodiment, the maximum reactive power output is obtained and compared with the second reactive power demand, and the grid voltage is coordinated at multiple levels according to the comparison result, specifically: if the reactive power demand of the node is less than the adjustable maximum reactive power, a distributed power source is selected to perform reactive power compensation on the grid node; if the reactive power demand of the node is greater than the adjustable maximum reactive power, a distributed power source and a reactive compensation device are used for reactive power compensation; if the reactive power output by the reactive compensation device reaches the maximum reactive power regulation capacity, the reactive resources of the distributed power sources and the reactive compensation device in the same level grid are dispatched.

[0020] In a preferred embodiment, the scheduling of the reactive resources of the distributed power sources and reactive compensation equipment in the power grid of the same level is specifically as follows:

[0021] Obtain the information of available distributed power sources and reactive power compensation equipment, and include the information of available distributed power sources and reactive power compensation equipment in the reactive power compensation range of the node; obtain the reactive power demand of reactive power compensation equipment in the power grid of the same level, and keep the reactive power output of the compensation equipment between the minimum reactive power regulation capacity and the maximum reactive power regulation capacity.

[0022] In a preferred embodiment, the system of the multi-level grid voltage coordination control method considering distributed power supply access is characterized in that it includes a power prediction module, a reactive power demand calculation module, a power demand adjustment module, and a grid voltage coordination and reactive power optimization module: a power prediction module, which is used to obtain grid information data, construct an ultra-short-term prediction module based on TimeUer, and obtain the power prediction value of the grid node; a reactive power demand calculation module, which is used to obtain the basic reactive power demand based on the power prediction value of the grid node; a power demand adjustment module, which is used to consider the dynamic impact of electric vehicle charging piles on the reactive power demand of the grid, calculate the increase in reactive power demand caused by charging, and obtain the second reactive power demand; a grid voltage coordination and reactive power optimization module, which is used to obtain the maximum reactive power output, compare it with the second reactive power demand, and perform multi-level coordination of the grid voltage according to the comparison result.

[0023] The technical effects and advantages of the multi-level grid voltage coordination control method and system considering the access of distributed power sources in the present invention are as follows:

[0024] 1. By designing a new time series prediction model TimeUer, the active and reactive power of distributed power sources and the reactive charging power of charging piles can be predicted in the ultra-short term, thereby significantly improving the prediction accuracy of the node power of distributed power sources and electric vehicle charging piles.

[0025] 2. Through the analysis of constraint conditions and the optimization of objective functions, the reactive resource dispatch between the current regional power grid and the power grids of the same level and the upper power grid nodes is realized, thereby coordinating the voltage of multiple power grids.

[0026] 3. By introducing advanced time series prediction models, blockchain technology, voltage stabilization calculation methods and resource coordination and scheduling methods, the coordinated regulation of voltages at each node of the power grid to which distributed power sources are connected can be achieved, thereby improving the stability and reliability of power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flowchart of a multi-level grid voltage coordination control method considering the access of distributed power sources provided in an embodiment of the present application;

[0028] Figure 2 An overall scheme diagram of a multi-level grid voltage coordination control method considering the access of distributed power sources provided in an embodiment of the present application;

[0029] Figure 3 A diagram of the power grid information data storage structure based on blockchain technology provided in an embodiment of the present application;

[0030] Figure 4 A diagram of the spatial attention mechanism module architecture provided for an embodiment of the present application;

[0031] Figure 5 The module architecture diagram of the TAU temporal attention mechanism provided in the embodiment of the present application;

[0032] Figure 6 The architecture diagram of the ultra-short-term prediction module based on TimeUer provided in the embodiment of the present application;

[0033] Figure 7 A schematic diagram of a multi-level grid voltage overall control method scheme provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0036] Embodiment 1, Figure 1 The present invention provides a multi-level grid voltage coordinated control method considering the access of distributed power sources, comprising the following steps:

[0037] Step S1, power grid information data storage based on blockchain technology.

[0038] The present invention considers a regional distribution network including multiple distributed power sources, electric vehicle charging piles and traditional power grid nodes. In the specific implementation of the present invention, the construction of a power grid information data storage system based on blockchain technology is a key prerequisite. Taking into account the security, scalability, processing speed and privacy requirements of power grid data storage, the Hyperledger Fabric blockchain platform is selected as the basic platform to effectively ensure the secure storage and efficient management of power grid data. The present invention uses a blockchain platform to store 4 tables, and the storage data structure is as follows: Figure 3 As shown, the [Distributed Power Source Table] stores the type, geographical location, historical power generation data, equipment parameters and other records of each distributed power source in the power grid; the [Grid Node Table] covers the real-time values ​​and historical data of node voltage, current, active power, reactive power and other information; the [Reactive Compensation Device Table] contains device type, rated capacity, installation location, current reactive output, adjustment capability range, operation reliability records, etc.; the [Electric Vehicle Charging Pile Table] records the location, quantity, charging power characteristics, connection status, and interaction data with the power grid of the charging piles.

[0039] The present invention uses a chain storage structure to record the historical changes of data. Each data block contains the hash value of the previous data block to ensure the integrity and non-tamperability of the data. For example, for the power generation data of distributed power sources, each time a new record is generated, the system packages it into a standardized data block. The hash value of the previous data block is embedded in the header of this data block. The hash algorithm adopts industry strong encryption standards such as SHA-256 to ensure the uniqueness and irreversibility of the hash value, thereby linking it to the previous data block to form a complete data chain, thereby tracing the power generation history of each power source, and any tampering with historical data can be easily discovered through the verification of the hash value. In order to realize the interaction between the middle layer and the data storage layer, the present invention uses DataStorageContract, DataQueryContract and DataUpdateContract smart contracts to respectively execute the storage operations corresponding to each data table, provide external data query interfaces and perform data update operations. DataStorageContract deeply embeds the storage logic of 4 data tables. When receiving various types of legal data push, the contract automatically parses the data format and accurately writes the data to the corresponding data table location to ensure that the data is accurately stored in the library. At the same time, it records the storage log to provide metadata support for data tracing. DataQueryContract supports multi-dimensional query condition combinations, such as querying historical power generation data of distributed power sources by time range, filtering charging pile status information by geographical area, etc. DataUpdateContract closely tracks the dynamic update needs of power grid operation data and monitors data change signals in real time. Once the distributed power source parameter adjustment, power grid node operation status mutation, charging pile equipment upgrade and other situations trigger data update needs, the contract quickly starts the update process, verifies the legitimacy and integrity of the updated data, accurately modifies the corresponding data records on the blockchain ledger, and synchronously updates the index information to ensure data timeliness and consistency.

[0040] In the existing data acquisition system and monitoring system of the power grid, by embedding the data upload module in the inverter controller of the distributed power source and adopting the standardized IEC61850 communication protocol, the real-time power generation data is stably pushed to the edge node of the blockchain platform in accordance with the predetermined JSON structured data format and a frequency of once every 5 seconds, and then the data is transferred in batches to the core blockchain network. At the same time, at the monitoring terminal of the power grid, the existing SCADA system data acquisition interface is used to seamlessly connect to the blockchain network access layer. The voltage, current and other data of the node are synchronized to the blockchain network, and the data transmission adopts redundant backup and breakpoint continuation mechanism to ensure that the data is not lost or disordered. By calling the smart contract query interface on the blockchain, the reactive output capacity of the distributed power source and the voltage data of the power grid node are obtained as the input parameters of the voltage coordination control algorithm, so as to realize accurate decision-making and control based on blockchain data and achieve the stability of the power grid voltage. At the same time, through the encryption technology of the blockchain, an asymmetric encryption algorithm is used in the data transmission link to generate public and private key pairs for the data sender and receiver respectively. The sender uses the recipient's public key to encrypt the data, ensuring that only the recipient holding the corresponding private key can decrypt and view the data, preventing the data from being stolen or monitored during network transmission, ensuring the security of the data during transmission and use, and preventing data leakage and malicious tampering.

[0041] Step S2, constructing an ultra-short-term prediction module based on TimeUer.

[0042] The present invention constructs an ultra-short-term prediction module based on TimeUer. The TimeUer model is improved based on TimeXer, and the TAU time attention mechanism is introduced, and the spatial attention mechanism is designed. With the help of TimeXer's own self-attention mechanism and cross-attention mechanism, the internal time dependency and correlation between variables in the time series data can be captured. The specific architecture and functions of the TimeXer model and the TAU time attention unit have been introduced in mature literature, and the present invention will not repeat them.

[0043] The present invention takes the date, ledger data, historical output data, numerical weather forecast data, reactive power and historical reactive charging power of electric vehicle charging piles in the past 15 days of each station as input, and takes the node active power and reactive power of the grid nodes connected by distributed power sources and charging piles and other equipment in the next 4 hours as output. Among them, the ledger data includes the installed capacity and the longitude and latitude center coordinates of the station; the historical output data includes the station active power, and records the historical power generation period and power output value according to the collected date and time; based on the type of distributed power station, key numerical weather forecast data is collected, mainly including weather data such as irradiance, temperature, humidity, air pressure, cloud cover and precipitation forecast.

[0044] As the object of model prediction, for the sequence data of active power and reactive power of nodes with length T, P={ } and Q={ }, after being input into TimeUer, it will be divided into M non-overlapping time blocks, and according to the divided time blocks and the complete input sequence, the embedding operation is performed separately through linear projection to obtain M time block embedding vectors ={ }and ={ }, and their respective embedding vectors for the complete sequence and The ledger data and weather forecast data are embedded in each complete input sequence to provide auxiliary information for the prediction of active power and reactive power. Figure 4 As shown, the TAU time attention mechanism model is as follows Figure 5 As shown in the figure, the TimeUer model architecture diagram is as follows Figure 6 As shown in Figure 2, the self-attention mechanism can capture the internal time dependency in the time series data through Formula 1 and Formula 2, and obtain a feature map with a shape of T′×C×H×W.

[0045] (Formula 1)

[0046] (Formula 2)

[0047] Among them, T′ represents the length of the feature sequence, C represents the number of feature map channels, H and W represent the height and width of the feature map respectively. Representative TimeUer blocks, represents the total number of TimeUer blocks, and [·,·] represents the concatenation operation.

[0048] (Formula 3)

[0049] The obtained feature map is then subjected to the spatial attention shown in Formula 3 to extract the input features in the spatial dimension and focus on important feature areas. represents the Hadamard product, express Activation function, represents the convolution operation, and Represent average pooling and maximum pooling respectively. Sequential block feature diagram representing active and reactive power.

[0050] Cross-attention mechanism to embed the complete sequence of active power and reactive power and Q, K and V are the complete sequence embedding vectors of the ledger data and weather forecast data information, realizing the cross attention shown in Formula 4 to capture the correlation between variables.

[0051] (Formula 4)

[0052] TAU decomposes temporal attention into intra-frame static attention SA and inter-frame dynamic attention DA. The calculation formulas are shown in Formula 5 and Formula 6.

[0053] (Formula 5)

[0054] (Formula 6)

[0055] in, represents a 1×1 convolution operation, represents the depth-wise dilated convolution, represents a depthwise convolution with a smaller kernel, represents a fully connected layer.

[0056] Intra-frame static attention SA is performed by sequential block feature maps of active power and reactive power. First, deep convolution is performed, and then deep dilated convolution and 1×1 convolution are used to capture the long-distance dependencies within the frame; and the dynamic attention DA between frames is learned by first averaging pooling and then fully connected layers. Channel weights between different frames to capture the temporal variation trends between frames.

[0057] (Formula 7)

[0058] Then, by using formula 7, we can perform the inner product of SA and DA and perform the Hadamard product with the original feature map to obtain the feature information containing the extracted time changes. represents the inner product operation. Finally, the next TimeUer block is obtained through the feedforward operation. and Input, as shown in Formula 8, where Represents a feed-forward operation.

[0059] (Formula 8)

[0060] Cross-attention can introduce information from sequences such as weather forecast data into the sequence data of the predicted object, effectively integrate auxiliary information valuable to the prediction process, and adapt to irregularities and heterogeneity such as missing values ​​and inconsistent sampling frequencies in the sequence data, effectively improving the overall prediction performance of the model. After three TimeUer blocks, the predicted values ​​of active power and reactive power of power grid nodes can be obtained with high accuracy through linear projection. According to the predicted results and the node position in the power grid, the predicted values ​​of node active power and reactive power are stored in the power grid node table of the blockchain storage system according to the node ID.

[0061] Step S3, considering reactive power regulation of the grid nodes to which the distributed generation is connected.

[0062] In order to stabilize the voltage of each node in the regional power grid, the present invention calculates the voltage value of each node using a method shown in Formula 9 based on the active power and reactive power of the node predicted in step S2.

[0063]

[0064]

[0065] (Formula 9)

[0066]

[0067]

[0068] in, is the voltage at node i, is the initial voltage of node i, is the reactive power prediction value of node i, is the initial reactive power of node i, is the predicted active power value of node i, is the initial active power of node i, is the reactive power balance coefficient, It is the active power balance coefficient, which is set by professional business personnel according to actual conditions and can be adjusted dynamically. is the power correction factor.

[0069] First, find the initial voltage, active power and reactive power of the current node from the blockchain power grid node data table, determine the reactive power balance coefficient, active power balance coefficient and power correction coefficient; find the active power prediction value and reactive power prediction value of the current node from the blockchain power grid node data table, and then calculate and ; When the convergence conditions are met, the voltage of node i is repeatedly iterated to obtain the final future voltage value and store it in the node table in the blockchain.

[0070] After obtaining the node voltage, the grid voltage power sensitivity can be obtained through the sensitivity matrix of formula 10.

[0071] (Formula 10)

[0072] Where i and j are any two nodes in the power grid. is the voltage of node i in the power grid, is the predicted active power of node j, is the predicted reactive power of node j, is the sensitivity of the voltage at node i to the active power at node j, is the sensitivity of the voltage at node i to the reactive power at node j.

[0073] After obtaining the voltage power sensitivity, the basic reactive power demand of each node can be calculated by formula 11:

[0074] (Formula 11)

[0075] in, is the basic reactive power demand of node i, is the voltage value of node i at time t, is the optimal reference voltage of node i.

[0076] Considering the dynamic impact of the charging state change of electric vehicle charging piles on reactive power demand, the present invention designs a method for calculating the increase in reactive power demand to improve the practical applicability of coordinated control of grid voltage. For example, if a large number of electric vehicle charging piles near a node are charged simultaneously during the peak power consumption in the evening, the increase in reactive power demand caused by charging the charging piles can be calculated by formula 12 according to the power characteristic curve and charging quantity of the charging piles, and it is superimposed on the basic reactive power demand, as shown in formula 13, and the reactive power demand of the node is stored in the node table in the blockchain.

[0077] (Formula 12)

[0078] (Formula 13)

[0079] in, is the increase in reactive power demand, M is the total number of charging piles near node i, is the reactive charging power of each charging pile found in the blockchain electric vehicle charging pile table, is the power factor, is the total reactive power demand of node i.

[0080] Taking into account the reactive power demand of node i and the impact of large-scale distributed power access on the grid voltage, the present invention evaluates the reactive power output capacity of the distributed power supply. The maximum adjustable reactive power output of each distributed power supply node can be obtained through formula 14 and stored in the distributed power supply table in the blockchain.

[0081] (Formula 14)

[0082] in, represents the maximum reactive power output that can be adjusted at node j, is the rated capacity of distributed power supply, is the current active power output of the node.

[0083] In order to realize the reactive power compensation of the distributed power supply and reactive power compensation equipment in the vicinity of node i, the present invention introduces the influence coefficient of the equipment on the node , and its calculation formula is shown in Formula 15.

[0084] (Formula 15)

[0085] Where S represents the influence coefficient of the device on the voltage of node i, is the distance between the device and node i, is the attenuation factor, and its value can be flexibly adjusted according to the grid topology and equipment characteristics. For each device in the grid, its influence coefficient on the node can be calculated according to the distance between it and the target node i and the grid topology information. With this method, the influence range of different devices on different nodes can be clearly obtained. Greater than the preset impact threshold , then it is determined that the device has sufficient influence on node i. At this time, the node ID within its influence range is recorded in the distributed power supply table and reactive compensation device table in the blockchain. When calculating the influence coefficient, the real-time operating status, reliability and other information of the distributed power supply and reactive compensation equipment in the blockchain technology record table are considered. If the reactive power regulation accuracy of the station is high and the failure rate is low in the past month, the influence coefficient of each power supply in the station is appropriately increased. Otherwise, the influence coefficient is reduced to achieve dynamic correction of the influence coefficient.

[0086] According to the records of the blockchain distributed power supply table and reactive power compensation device table, within the affected range, if the reactive power required to be compensated by node i is within the maximum reactive power range that can be adjusted by the distributed power supply node j, the distributed power supply is preferred to perform reactive power compensation on the grid node; if the reactive power demand of node i exceeds the maximum reactive power range that can be adjusted by the distributed power supply node j, in addition to using the distributed power supply for reactive power compensation, the reactive compensation equipment installed in the grid must also be enabled.

[0087] Such as static VAR compensator, inverter, etc., assist in reactive power compensation to achieve grid voltage stability, as shown in Formula 16 and Formula 17.

[0088]

[0089] (Formula 16)

[0090] (Formula 17)

[0091] The reactive power output of the reactive power compensation device And the reactive power output of distributed power generation The data is updated to the distributed power supply table and reactive power compensation device table in the blockchain at a frequency of once every 5 seconds, so as to be queried and obtained during the voltage stabilization process in step S4.

[0092] Step S4: coordinated control of grid voltage based on reactive power compensation.

[0093] Under the premise of step S1, step S2 and step S3, according to the influence coefficient S, the reactive power output of the distributed power supply whose node ID is the target node is searched in the distributed power supply table and reactive power compensation device table stored in the blockchain determined by step S3. , and the reactive power output by the reactive power compensation device . In order to stabilize the grid voltage within the specified range and achieve the goal of regional voltage stability.

[0094] The present invention iteratively calculates the voltage control optimization objective function under the constraint conditions of satisfying each variable in the manner shown in Formula 18.

[0095]

[0096]

[0097] (Formula 18)

[0098]

[0099]

[0100] In the objective function, M represents the prediction duration, which is 4 hours, and N represents the total number of nodes. represents the voltage of the ith node at time step t, represents the reference voltage of the node, J represents the number of distributed generation sources connected to each regional power grid, represents the operating cost of the jth distributed generation, Represents the output reactive power of the power supply; in the constraints and Respectively represent the minimum and maximum values ​​of the voltage specified for node i, and are the minimum and maximum reactive power regulation capacity of reactive power compensation device k, respectively. is the influence coefficient of the device on node i, The minimum influence coefficient threshold for setting voltage regulation.

[0101] For each node i and each time step t, calculate the square of the deviation of the node voltage from the reference voltage , and summing over all nodes and time steps, we get At the same time, considering the operating cost of distributed power generation, calculate , and add it to the objective function. Check the voltage of each node i at each time step t Whether it is within the specified range; according to the device parameters in the compensation device table of the blockchain, for each reactive compensation device k, check whether its reactive power output is within the rated parameter range; at the same time, it is necessary to ensure that the impact range S is not less than the set minimum threshold. By continuously iterating the above calculation and judgment process, the voltage control objective function is optimized, so that the grid voltage is stable within the specified range, achieving the goal of regional voltage stability, and uploading the grid data to the blockchain node table, distributed power supply table and reactive compensation device table in real time through the DataUpdateContract contract at a frequency of once every 5 seconds.

[0102] Step S5: coordinated control of multi-level power grid voltage based on resource scheduling.

[0103] In the process of stabilizing the voltage in step S4, if the reactive power output by the reactive power compensation device has reached the maximum reactive power regulation capacity , it means that the current grid has a large reactive power demand at the node during voltage stabilization, and faces the risk of insufficient reactive power output of the current regional grid itself and being unable to meet the reactive power demand. At this time, it is necessary to dispatch the reactive resources of distributed power sources and reactive compensation devices in the grid at the same level. As shown in Formula 19 and Formula 20, by adjusting the attenuation factor in the influence coefficient to ', the adjusted influence coefficient is , view the information of the available distributed power sources and reactive compensation equipment in the power grid of the same level within the influence range through the blockchain platform data table, including its capacity, current operating status, etc., update the information of the connected nodes in the distributed power source table and reactive compensation device table of the blockchain platform, and include the distributed power sources and reactive compensation devices in the adjacent power grid of the same level into the reactive compensation range of node i, so as to realize the optimal configuration of reactive resources of distributed power sources in the power grid of the same level.

[0104] At the same time, the power prediction in step S2 is performed on the device m in the same level power grid, and the reactive power of the device is adjusted through the operation in step S3. Similar to step S4, the reactive power output of the compensation device is adjusted by combining the device parameters in the data table. Maintaining minimum reactive power regulation capacity and maximum reactive power regulation capacity Through continuous iterative calculation and repeated judgment, the voltage control objective function is optimized, the grid voltage in the target area is coordinated and stabilized within the specified range, and the grid data is updated to the corresponding data table in real time.

[0105] For example, in a regional power grid composed of multiple community distribution networks, when a community distribution network has a voltage problem and needs reactive power support, the blockchain can be used to query the idle capacity and availability information of distributed power sources and reactive power compensation equipment in the adjacent community distribution network. According to the real-time operating status and reactive power regulation capabilities of the participating distributed power sources, the reactive output of each power source is optimized and allocated to ensure that the overall power source can efficiently provide reactive power support for the target community distribution network.

[0106] (Formula 19)

[0107]

[0108] (Formula 20)

[0109]

[0110] Similarly, if the distributed power sources and reactive power compensation devices in the same-level power grid are included in the reactive power compensation range of node i, the reactive power output of the equipment in the same-level power grid will be The maximum reactive power regulation capacity has been reached , it means that the reactive resource coordination between the same-level power grids can no longer meet the reactive demand of the current node, and it is necessary to dispatch the distributed power sources in the upper-level power grid and the reactive resources of the reactive compensation device at the same time. As shown in formula 21 and formula 22, the attenuation factor in the influence coefficient is also adjusted to , the adjusted influence coefficient is , the distributed power sources and reactive power compensation devices in the upper power grid area are included in the reactive power compensation range of node i. At the same time, blockchain technology is used to obtain the information of the available distributed power sources and reactive power compensation devices, so as to realize the rapid query and deployment record of cross-level power grid resources and ensure the efficiency and reliability of multi-level power grid voltage coordination control. At the same time, the power prediction in step S2 is performed on the compensation device n, and the reactive power of the device is adjusted through the operation in step S3. Similarly to step S4, the reactive power output of the device is adjusted by combining the device parameters in the data table. Maintaining minimum reactive power regulation capacity and maximum reactive power regulation capacity By dispatching reactive resources between multi-level power grids, reactive compensation for node i is achieved, and the overall voltage is coordinated and stabilized. The schematic diagram of the overall voltage coordination control method is shown in Figure 7 shown.

[0111] For example, in a multi-layer power grid architecture that includes a municipal power grid and multiple district power grids, when the reactive power coordination between district power grids cannot meet the voltage requirements of a key node, the blockchain is used to apply for reactive power resource support from the upper-level municipal power grid. The municipal power grid deploys large-scale distributed power sources and reactive power compensation equipment under its jurisdiction according to the voltage conditions and resource distribution of the entire network to perform reactive power compensation for key nodes of the district power grid. During the dispatching process, the output characteristics of the distributed power source and the connection method with the district power grid are considered. By adjusting the output power and reactive power distribution of the distributed power source, it is ensured that it can effectively enhance the reactive power support capacity of the upper-level power grid to the lower-level power grid and stabilize the voltage of key nodes.

[0112] (Formula 21)

[0113]

[0114] (Formula 22)

[0115]

[0116] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0117] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0118] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0119] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0120] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0121] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-level grid voltage coordination control method considering the access of distributed power sources, characterized in that: The following steps are involved: Obtain power grid information data, build an ultra-short-term prediction module based on TimeUer, and obtain the power prediction value of the power grid node; Obtain basic reactive power demand based on power forecast values ​​of grid nodes; The power prediction value of the grid node is specifically: Performing a cutting process on the power grid information data to obtain a plurality of non-overlapping time blocks, and performing an embedding operation on the plurality of non-overlapping time blocks to obtain a first embedding vector and a second auxiliary information data; Combined with the self-attention mechanism, the internal time dependency between the first embedding vectors is obtained to obtain the first feature map; Based on the spatial attention mechanism, feature extraction is performed on the first feature map in the spatial dimension to obtain the second feature map; Based on the cross attention mechanism, the correlation between the first embedding vector and the second auxiliary information data is captured to obtain the correlation between variables; Based on the TAU time attention mechanism, the feature information containing time changes is obtained in combination with the second feature map; Performing linear projection processing on the characteristic information to obtain a power prediction value of the power grid node; Considering the dynamic impact of electric vehicle charging piles on the reactive power demand of the power grid, the increase in reactive power demand caused by charging is calculated to obtain the second reactive power demand; The maximum reactive power output is obtained and compared with the second reactive power demand, and the grid voltage is coordinated at multiple levels according to the comparison result.

2. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 1 is characterized in that: The step of obtaining the power grid information data specifically includes: Combined with the chain storage structure, several smart contracts are preset to execute the data access and management operations corresponding to the data tables in each storage structure; The real-time grid information data is transmitted in batches to the blockchain network through the edge nodes of the blockchain platform according to the preset data format and transmission frequency; The monitoring terminal of the power grid obtains power grid information data through the blockchain network, and the blockchain network and the monitoring terminal transmit data through encryption methods.

3. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 2 is characterized in that: The embedding operation is performed on a plurality of non-overlapping time blocks to obtain a first embedding vector and a second auxiliary information data, specifically: The predicted target data in the power grid information data is segmented to obtain a number of non-overlapping time blocks; According to the divided time blocks and the complete input sequence, embedding operations are performed respectively to obtain a plurality of first embedding vectors, wherein the first embedding vectors include time block embedding vectors and input sequence embedding vectors; An embedding operation is performed on the non-target data according to each complete input sequence to obtain second auxiliary information data.

4. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 3 is characterized in that: The TAU-based time attention mechanism is combined with the second feature map to obtain feature information containing time changes, specifically: Based on the TAU temporal attention mechanism, temporal attention is decomposed into intra-frame static attention and inter-frame dynamic attention; The inner product of the static attention within the frame and the dynamic attention between frames is performed, and the Hadamard product is performed with the original feature map to obtain the feature information containing time changes.

5. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 4 is characterized in that: The basic reactive power demand is specifically formulated as follows: in, is the basic reactive power demand of node i, is the voltage value of node i at time t, is the optimal reference voltage of node i, is the sensitivity of the voltage at node i to the reactive power at node j.

6. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 5, characterized in that: The second reactive power demand is obtained by: Obtain charging data of charging piles during peak hours of electricity consumption, and calculate the increase in reactive power demand based on the power characteristic curve of the charging piles and the number of charges; The reactive power demand increase is added to the basic reactive power demand to obtain a second reactive power demand.

7. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 6 is characterized in that: The maximum reactive power output is obtained and compared with the second reactive power demand, and the grid voltage is coordinated in multiple levels according to the comparison result, specifically: If the reactive power demand of the node is less than the maximum adjustable reactive power, the distributed power source is selected to perform reactive power compensation on the grid node; If the reactive power demand of the node is greater than the maximum adjustable reactive power, the distributed power source and reactive power compensation equipment are used for reactive power compensation; If the reactive power output by the reactive compensation device reaches the maximum reactive power regulation capacity, the distributed power sources and reactive power resources of the reactive compensation device in the power grid of the same level are dispatched.

8. The multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in claim 7 is characterized in that: The scheduling of the reactive resources of the distributed power sources and reactive compensation equipment in the power grid of the same level is specifically: Acquire the information of the available distributed power source and reactive power compensation equipment, and include the information of the available distributed power source and reactive power compensation equipment into the reactive power compensation scope of the node; Obtain the reactive power demand of the reactive power compensation equipment in the same level power grid, and keep the reactive power output of the compensation equipment between the minimum reactive power regulation capacity and the maximum reactive power regulation capacity.

9. A multi-level grid voltage coordination control system considering the access of distributed power sources, adopting a multi-level grid voltage coordination control method considering the access of distributed power sources as claimed in any one of claims 1 to 8, characterized in that: Including power prediction module, reactive power demand calculation module, power demand adjustment module, grid voltage coordination and reactive power optimization module: The power prediction module is used to obtain power grid information data, build an ultra-short-term prediction module based on TimeUer, and obtain the power prediction value of the power grid node; The power prediction value of the grid node is specifically: Performing a cutting process on the power grid information data to obtain a plurality of non-overlapping time blocks, and performing an embedding operation on the plurality of non-overlapping time blocks to obtain a first embedding vector and a second auxiliary information data; Combined with the self-attention mechanism, the internal time dependency between the first embedding vectors is obtained to obtain the first feature map; Based on the spatial attention mechanism, feature extraction is performed on the first feature map in the spatial dimension to obtain the second feature map; Based on the cross attention mechanism, the correlation between the first embedding vector and the second auxiliary information data is captured to obtain the correlation between variables; Based on the TAU time attention mechanism, the feature information containing time changes is obtained in combination with the second feature map; Performing linear projection processing on the characteristic information to obtain a power prediction value of the power grid node; A reactive power demand calculation module is used to obtain basic reactive power demand based on power prediction values ​​of grid nodes; The power demand adjustment module is used to consider the dynamic impact of the electric vehicle charging pile on the reactive power demand of the power grid, calculate the increase in reactive power demand caused by charging, and obtain the second reactive power demand; The grid voltage coordination and reactive power optimization module is used to obtain the maximum reactive power output and compare it with the second reactive power demand, and perform multi-level coordination of the grid voltage according to the comparison result.

Images