New energy base cluster voltage stability optimization control method based on flexible interconnection
By constructing a time series of voltage stability parameters and performing normalization and averaging, combined with a multi-strategy collaborative optimization algorithm based on reinforcement learning, the reactive power interaction capacity is predicted in real time, solving the voltage stability problem of the new energy base cluster and achieving safe and efficient voltage control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-03-20
AI Technical Summary
Voltage stability issues are prominent in new energy base clusters. Traditional control methods have slow response speeds, high operation and maintenance costs, and lack rapid, real-time information exchange and coordination, leading to control conflicts and affecting the safe and efficient operation of new energy sources.
By constructing a time series of voltage stability parameters, normalizing and averaging them, and combining them with a multi-strategy collaborative optimization algorithm based on reinforcement learning, reactive power interaction capacity is predicted in real time, and voltage stability control is achieved through flexible interconnection devices.
The voltage stability of the new energy base cluster has been optimized, enhancing the overall robustness of the system in the face of new energy fluctuations and ensuring safe and efficient operation.
Smart Images
Figure CN121710285A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of flexible interconnection technology, and in particular relates to a method for optimizing and controlling the voltage stability of new energy base clusters based on flexible interconnection. Background Technology
[0002] With the significant growth of new energy sources, represented by wind power and photovoltaics, large-scale centralized development bases for new energy have been formed in areas rich in wind and solar resources. These bases typically consist of dozens or even hundreds of wind farms and photovoltaic power stations, which are connected to the main grid through connecting lines and high-voltage transmission channels, forming new energy base clusters. However, the large-scale and high-proportion integration of new energy has also brought unprecedented technical challenges, among which voltage stability issues are particularly prominent. These challenges manifest as randomness and volatility, weak inertia and low short-circuit capacity, and insufficient fault ride-through and dynamic support capabilities. Therefore, how to effectively improve the voltage stability of new energy base clusters and ensure their safe and efficient operation has become a key technical bottleneck restricting the further absorption and development of new energy.
[0003] Faced with these challenges, traditional voltage control methods are proving inadequate. For example, reactive power compensation based on synchronous condensers has a slow response speed, is a rotating device with high maintenance costs, and struggles to track rapid fluctuations in renewable energy sources. Mechanically switched capacitors, with their tiered switching, cannot achieve smooth, continuous reactive power regulation and are prone to switching oscillations when dealing with rapidly changing voltage conditions. Furthermore, the lack of rapid, real-time information exchange and coordination between different power stations can easily lead to a situation of "each operating independently" or even control conflicts.
[0004] Flexible interconnection represents a deepening and development of advanced power electronics technologies such as flexible DC and AC transmission systems. Its advantages over traditional rigid interconnection lie in its rapid and flexible decoupling control and enhanced system voltage stability. Therefore, research on voltage stability optimization control methods for new energy base clusters based on flexible interconnection has emerged. Summary of the Invention
[0005] The voltage stability optimization control method for new energy base clusters based on flexible interconnection provided in this application can accurately determine the reactive power interaction capacity between new energy base clusters at the next moment, thereby ensuring the voltage stability optimization control of the new energy base clusters.
[0006] In a first aspect, embodiments of this application provide a voltage stability optimization control method for new energy base clusters based on flexible interconnection, including: Voltage stability parameters between target new energy base clusters are obtained according to a preset sampling time interval, and a voltage stability parameter time series is constructed. Based on the time series of voltage stability parameters, the voltage stability parameters are normalized to obtain normalized values of the voltage stability parameters. The normalized values of the voltage stability parameters are averaged to obtain the average value of the voltage stability parameters. Based on the average value of voltage stability parameters, a multi-strategy collaborative optimization algorithm based on reinforcement learning is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment. Based on the reactive power interaction capacity between the target new energy base clusters at the next moment, control commands are sent to the corresponding new energy bases to achieve voltage stability control between the target new energy base clusters.
[0007] In one optional implementation, the voltage stability parameters include the maximum node voltage under different voltage levels of transmission lines, the total reactive power output of each new energy base in the target new energy base cluster, and the reactive power interaction of flexible interconnection between any two new energy bases.
[0008] In one alternative implementation, the normalized values of the voltage stability parameters are averaged, including: for each voltage stability parameter, calculating the arithmetic mean of the normalized values of the voltage stability parameter at all sampling times in the voltage stability parameter time series.
[0009] In one alternative implementation, the reinforcement learning-based multi-policy collaborative optimization algorithm includes: Train multiple policy networks, each policy network corresponding to the flexible interconnection reactive power interaction between any two new energy bases in the target new energy base cluster, and construct a reactive power interaction objective function based on each policy network. With the goal of maximizing the reactive interaction objective function, the parameters of each policy network are iteratively updated using the gradient ascent method; Each policy network is trained using the mean squared error loss function to minimize the mean squared error loss function and obtain the reactive power interaction error reference value. The reactive power interaction capacity between the target new energy base clusters at the next moment is calculated based on the reactive power interaction objective function, the reactive power interaction error reference quantity, and the reactive power interaction error correction coefficient.
[0010] In one optional implementation, the reactive power interaction objective function is the average value of the reactive power interaction expected function under each policy network, wherein each policy network is trained based on the average value of the historical flexible interconnection reactive power interaction between the corresponding two new energy bases.
[0011] In one alternative implementation, the expression for the reactive power interaction objective function is as follows: ; Where N represents the total number of policy networks, and i represents the i-th policy network. For the i-th policy network The effectiveness of reactive power interaction adjustment, o i Represents the i-th policy network The input state value, π i This represents the i-th policy function; The objective function for reactive power interaction among target new energy clusters; For the i-th policy network The expected reactive power interaction function between the target new energy clusters.
[0012] In one alternative implementation, the reactive power interaction capacity between the target new energy base clusters at the next moment is represented by positive and negative values to indicate the reactive power transmission direction between the two new energy bases.
[0013] Secondly, embodiments of this application provide a voltage stability optimization and control system for a new energy base cluster based on flexible interconnection, including: The data acquisition module is used to acquire voltage stability parameters between target new energy base clusters according to a preset sampling time interval, and to construct a time series of voltage stability parameters. The normalization module is used to normalize the voltage stability parameters based on the time series of the voltage stability parameters, and obtain the normalized values of the voltage stability parameters. The mean processing module is used to perform mean processing on the normalized values of the voltage stability parameters to obtain the average value of the voltage stability parameters; The prediction module is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment, based on the average value of voltage stability parameters and using a multi-strategy collaborative optimization algorithm based on reinforcement learning. The control module is used to send control commands to the corresponding new energy bases based on the reactive power interaction capacity between the target new energy base clusters at the next moment, so as to achieve voltage stability control between the target new energy base clusters.
[0014] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method provided in embodiments of this application.
[0015] Fourthly, embodiments of this application provide a computer-readable storage medium, characterized in that it stores a computer program thereon, which, when executed in a computer, causes the computer to perform the method provided in embodiments of this application.
[0016] The technical solution provided in this application has the following beneficial effects: This invention monitors multiple parameters in real time, including the voltage of key nodes at different voltage levels, the total reactive power output of each base, and the reactive power interaction between flexible interconnections, constructing a time series. Subsequently, the collected parameters are normalized and averaged to eliminate the influence of dimensions and smooth data fluctuations, providing a standardized and stable input that characterizes the overall system operation trend for subsequent analysis. Based on the processed data, this invention employs a multi-strategy collaborative optimization algorithm based on reinforcement learning. This algorithm establishes multiple strategy functions corresponding to different reactive power interaction channels, using the historical average reactive power interaction value as input, and aims to maximize the expected voltage stability effect. The system adaptively evaluates and optimizes the reactive power interaction adjustment of each channel, continuously iterates and corrects the strategy parameters through the gradient ascent method, and minimizes the deviation between the prediction and the target using the discriminant calculation module. Finally, it accurately outputs the optimal reactive power interaction capacity and direction between each base at the next moment, thereby enabling the adjustment of reactive power resource distribution in advance. By utilizing the fast and flexible control capabilities of flexible interconnection equipment, it can proactively intervene before voltage problems occur. At the same time, by coordinating reactive power support between multiple bases, a cluster-coordinated voltage support force is formed, which significantly enhances the overall robustness of the system in response to the volatility of new energy sources, thereby ensuring the safe and efficient operation of the new energy base cluster. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating the voltage stability optimization control method for a new energy base cluster based on flexible interconnection provided in this application embodiment. Detailed Implementation
[0018] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0019] This application provides a voltage stability optimization control method for new energy base clusters based on flexible interconnection. Figure 1 This is a flowchart illustrating the voltage stability optimization control method for a new energy base cluster based on flexible interconnection provided in this application embodiment. The method can be executed by a voltage stability optimization control system for a new energy base cluster based on flexible interconnection. The system can be implemented by software and / or hardware and can be configured in electronic devices such as computers.
[0020] like Figure 1 As shown, the technical solution provided in this application includes the following steps: S110. Obtain voltage stability parameters between target new energy base clusters according to preset sampling time intervals, and construct a voltage stability parameter time series.
[0021] Among them, the various new energy bases in the target new energy base cluster are electrically connected through flexible interconnection equipment.
[0022] In some embodiments, the sampling time is denoted as , where n is a natural number representing the total number of samples, s is the index of the s-th sampling time (s∈{1,2,…,n}), and the time interval between each time is fixed (e.g., sampling once every few seconds or minutes).
[0023] Voltage stability parameters include the maximum node voltage under transmission lines of different voltage levels within the target renewable energy base cluster, the reactive power output of each renewable energy base, and the reactive power interaction between renewable energy bases. These parameters are obtained in real time through sensors deployed on transmission lines and renewable energy bases.
[0024] For example, taking a target new energy base cluster including new energy base A, new energy base B, and new energy base C as an example, the specific parameters include: Maximum node voltage of transmission lines: : indicates at time T s At that time, the maximum node voltage value (unit: kV) under the 110kV transmission line reflects the stability of the 110kV voltage level.
[0025] : indicates at time T s At that time, the maximum node voltage value (unit: kV) under the 330kV transmission line reflects the stability of the 330kV voltage level.
[0026] Total reactive power output of the new energy base: : indicates at time T s At that time, the total reactive power output value (unit: Mvar) of the new energy base A reflects the reactive power support capacity of the new energy base A.
[0027] : indicates at time T s At that time, the total reactive power output value (unit: Mvar) of the new energy base B reflects the reactive power support capability of the new energy base B.
[0028] : indicates at time T s At that time, the total reactive power output value (unit: Mvar) of the new energy base C reflects the reactive power support capability of the new energy base C.
[0029] Flexible interconnection and reactive power interaction between new energy bases: : indicates at time T sAt that time, the reactive power value (unit: Mvar) exchanged between new energy base A and new energy base B through flexible interconnection equipment. Symbol definition: positive value when reactive power flows from new energy base A to new energy base B, and negative value when reactive power flows from new energy base B to new energy base A.
[0030] : indicates at time T s At that time, the reactive power interaction between the flexible interconnection between new energy base A and new energy base C follows the same sign rule as above: the signal from new energy base A to new energy base C is positive, and the signal from new energy base C to new energy base A is negative.
[0031] : indicates at time T s At that time, the reactive power interaction of the flexible interconnection between new energy base B and new energy base C follows the same sign rule as above: positive from new energy base B to new energy base C, and negative from new energy base C to new energy base B.
[0032] The above parameters are applied at each sampling time T. s The measured values are arranged in chronological order to form a multivariate voltage stability parameter time series dataset. The voltage stability parameter time series is represented in matrix form, with each row corresponding to one parameter and each column corresponding to one sampling time. The specific matrix structure is as follows: ; The matrix has a dimension of 8×n, where 8 represents the number of parameters (including 2 voltage parameters, 3 reactive power output parameters, and 3 reactive power interaction parameters), and n represents the number of time points. The time series data captures information about the dynamic changes of voltage stability parameters over time, providing raw input for subsequent steps.
[0033] By constructing time series data, the system can track the historical changes in voltage stability parameters, laying a data foundation for subsequent normalization and averaging processes, which helps improve the accuracy and real-time performance of control.
[0034] S120. Based on the time series of voltage stability parameters, normalize the voltage stability parameters to obtain normalized values of the voltage stability parameters.
[0035] In some embodiments, the physical units and numerical ranges of various parameters (such as voltage values and reactive power values) in the time series data obtained in step S110 differ significantly (e.g., voltage is in kV; reactive power is in Mvar). Directly using these raw data for calculations leads to inconsistencies in dimensions, causing parameters with larger numerical ranges to dominate in subsequent analyses, while the role of parameters with smaller numerical ranges is weakened, affecting the accuracy and fairness of the analysis. Therefore, it is necessary to normalize the voltage stability parameters, preserving the shape of the data distribution while eliminating the influence of dimensions, making all parameters comparable in subsequent processing.
[0036] In some embodiments, each parameter in the time series is normally calculated independently. For any time T s The original value of any measured parameter Its normalized value The calculation formula is: ; in, Indicates at time T s The original measured value of a certain voltage stability parameter, X max X represents the maximum value of this voltage stability parameter among all n sampling times. min This represents the minimum value of the voltage stability parameter among all the measurements taken at the n sampling times. This indicates that the voltage stability parameter is at time T. s The normalized value of is between 0 and 1.
[0037] For example, continuing with the target new energy base cluster including new energy base A, new energy base B, and new energy base C, the eight voltage stability parameters are normalized respectively, and the complete set of equations is as follows: ; in, T represents s Normalized value of the maximum node voltage at a given time on a 110kV transmission line; T represents s Normalized value of the maximum node voltage at a given time on a 330kV transmission line; T represents s Normalized value of total reactive power output of the new energy base A at any time; T represents s Normalized value of reactive power output of B main unit in the new energy base at all times; T represents s Normalized value of total reactive power output of the C-type new energy base at all times; T represents sNormalized value of reactive power interaction in flexible interconnection between new energy base A and new energy base B at any time; T represents s Normalized value of reactive power interaction in flexible interconnection between new energy base A and new energy base C at any time; T represents s Normalized value of reactive power interaction in flexible interconnection between new energy base B and new energy base C at any time; , They represent The maximum and minimum values of the maximum node voltage under the 110kV transmission line at these n fixed time intervals; , These represent the maximum and minimum values of the maximum node voltage under a 330kV transmission line at n fixed time intervals, respectively. , These represent the maximum and minimum values of the total reactive power output of the new energy base A at n fixed time intervals, respectively. , These represent the maximum and minimum values of the total reactive power output of the new energy base B at n fixed time intervals, respectively. , These represent the maximum and minimum values of the total reactive power output of the new energy base C at n fixed time intervals, respectively. , These represent the maximum and minimum values of the flexible interconnection reactive power interaction between new energy base A and new energy base B at n fixed time intervals, respectively. , These represent the maximum and minimum values of the flexible interconnection reactive power interaction between new energy base A and new energy base C at n fixed time intervals, respectively. , These represent the maximum and minimum values of the flexible interconnection reactive power interaction between new energy base B and new energy base C at n fixed time intervals.
[0038] S130. The normalized values of the voltage stability parameters are averaged to obtain the average value of the voltage stability parameters.
[0039] In some embodiments, step S120 yields a sequence of normalized values for each voltage stability parameter over n consecutive time points. These sequences contain detailed information about how the parameter changes over time, but may also include random fluctuations or short-term disturbances. Therefore, it is necessary to average the normalized values of the voltage stability parameter to extract the overall average level or stable state of the parameter within the most recent time window, thereby filtering out short-term noise, highlighting the long-term trend of the parameter, and enabling subsequent optimization control decisions to be based on a more stable and representative system state.
[0040] In some embodiments, for each voltage stability parameter, the arithmetic mean of the normalized values of the voltage stability parameter at all sampling times in the voltage stability parameter time series is calculated, using the following formula: ; in, It represents the average value of a certain voltage stability parameter.
[0041] For example, continuing with the target new energy base cluster including new energy base A, new energy base B, and new energy base C, the normalized values of the eight voltage stability parameters are calculated to obtain their arithmetic mean. The complete set of equations is as follows: ; in, T represents s The average value of the maximum node voltage at any given time on a 110kV transmission line; T represents s The average value of the maximum node voltage at a given time on a 330kV transmission line; T represents s The average value of the total reactive power output of the new energy base A at any given time; T represents s The average value of the total reactive power output of the new energy base B at any given time; T represents s The average value of total reactive power output at the C-type new energy base at any given time; T represents s The average value of the flexible interconnection reactive power interaction between new energy base A and new energy base B at any given time; T represents s The average value of the flexible interconnection reactive power interaction between new energy base A and new energy base C at any given time; T represents s The average value of the flexible interconnection reactive power interaction between new energy base B and new energy base C at any given time.
[0042] Mean averaging integrates the dynamic time-series information of each voltage stability parameter into a representative static feature value, effectively smoothing measurement noise and short-term fluctuations, allowing the algorithm to focus on the macroscopic and continuous trends of system state changes. Reducing the dimensionality of each parameter from n-dimensional time-series data to a 1-dimensional average significantly reduces the amount of data and complexity that subsequent algorithms need to process, while retaining key information characterizing the recent operating state of the system.
[0043] S140. Based on the average value of voltage stability parameters, a multi-strategy collaborative optimization algorithm based on reinforcement learning is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment.
[0044] In some embodiments, a multi-policy collaborative optimization algorithm based on reinforcement learning includes: Train multiple policy networks, each policy network corresponding to the flexible interconnection reactive power interaction between any two new energy bases in the target new energy base cluster, and construct a reactive power interaction objective function based on each policy network. With the goal of maximizing the reactive interaction objective function, the parameters of each policy network are iteratively updated using the gradient ascent method; Each policy network is trained using the mean squared error loss function to minimize the mean squared error loss function and obtain the reactive power interaction error reference value. The reactive power interaction capacity between the target new energy base clusters at the next moment is calculated based on the reactive power interaction objective function, the reactive power interaction error reference quantity, and the reactive power interaction error correction coefficient.
[0045] For example, let's continue with the illustration using the target new energy base cluster, which includes new energy base A, new energy base B, and new energy base C: A multi-policy collaborative optimization algorithm based on reinforcement learning defines three policy functions for the reactive power interaction pairs among the three new energy bases. Each policy function is responsible for calculating the adjustment amount for a reactive interaction pair: Policy function This represents a reactive power interaction control strategy between new energy base A and new energy base B through flexible interconnection. Policy function This represents a reactive power interaction control strategy that connects new energy base A and new energy base C through flexible interconnection. Policy function This represents a reactive power interaction control strategy that connects new energy base B and new energy base C through flexible interconnection. Wherein, the input state o of the strategy i The average value of the reactive power interaction parameters calculated in step S130 is taken from: o1 represents the average value of reactive power interaction between the flexible interconnection between new energy bases A and B. .
[0046] o2 represents the average reactive power interaction between the flexible interconnection between new energy bases A and C. .
[0047] o3 represents the average reactive power interaction between the flexible interconnection between new energy bases B and C. .
[0048] It is the parameter vector of the i-th policy function, representing the effective adjustment amount of reactive power interaction, the value of which is obtained through learning and optimization.
[0049] The goal of the algorithm is to maximize the reactive interaction objective function. This function is defined as the average of the reactive power interaction expectation function under all strategies: ; Where N represents the total number of policy networks, and i represents the i-th policy network. For the i-th policy network The effectiveness of reactive power interaction adjustment, o i Represents the i-th policy network The input state value, π i This represents the i-th policy function; The reactive power interaction objective function among the target new energy clusters reflects the quality of the overall control strategy; the larger the value, the better the voltage stability control effect. For the i-th policy network The expected reactive power interaction function between target new energy clusters under state o is represented by the function in state o. i The following strategy The expected long-term cumulative reward (i.e., voltage stabilization effect) is obtained from the reactive power interaction expectation function through the formula (i.e., mean square error loss function) of the subsequent reactive power interaction discrimination calculation.
[0050] Furthermore, in order to maximize the objective function The algorithm uses the gradient ascent method to iteratively update the parameters of each strategy. This allows the strategy to be continuously improved. The update rules are as follows: ; ; in, and They represent the i-th policy network respectively. The effectiveness of reactive power interaction adjustments between the current time and the next time. It is the learning rate for effectively adjusting the quantity, controlling the step size of parameter updates, and its value is a positive real number. It is the objective function Regarding the strategy parameter θ π The gradient indicates the direction of parameter optimization, and J represents the effective adjustment amount of reactive power interaction. The expected function Q is represented. π The gradient of action 'a' represents the impact of the change in action on the expected effect. The policy function π represents the parameter θ. π The gradient represents the effect of parameter changes on the policy output.
[0051] Through repeated iterative updates, the policy parameter θ πGradually converge to the optimal value, so that the objective function J π maximize.
[0052] Furthermore, evaluate the expected function Q of the current strategy. πi The algorithm improves the accuracy of the prediction function by minimizing the prediction error. It uses mean squared error as the loss function to calculate the difference between the prediction function and the target value. ; in, This is a reference value for reactive power interaction error, representing the prediction error of the expectation function. The smaller the value, the more accurate the prediction. The goal is to minimize the reactive power interaction expectation function among renewable energy clusters. Interactive objective function with reactive power Differences, y i The expected value representing the ideal is calculated by the next step of calculating the reactive power interaction capacity between the target new energy base clusters at the next moment.
[0053] minimize This can make the expected function Q πi It is closer to the true value, thus providing a reliable reference for strategy optimization.
[0054] Furthermore, based on the reactive power interaction objective function, the reactive power interaction error reference value, and the reactive power interaction error correction coefficient, the reactive power interaction capacity between the target new energy base clusters at the next moment is calculated, and the final output target Q value y is obtained. i Calculate using the following formula: ; Among them, y i The target Q value, which is the final output of the i-th strategy, is directly used as the reactive power interaction capacity at the next moment. i It is a reward signal, derived from the reactive power interaction objective function obtained when the expected voltage stability effectiveness is achieved. , representing the immediate benefit of the current strategy, and γ is the reactive power interaction error correction coefficient (γ∈[0,1]), used to adjust the error reference value. The impact on the output value should be considered to avoid excessive amplification of errors.
[0055] Step S140 uses a reinforcement learning algorithm to transform the average value of historical operating data into forward-looking control commands. The algorithm employs multi-strategy collaborative optimization, fully considering the coupling relationships between various new energy bases and avoiding the conflict problem of "each fighting its own battle" in traditional control. Output result y iThis method directly guides flexible interconnected devices to adjust reactive power flow, thereby compensating for voltage fluctuations in real time and improving the overall voltage stability of the renewable energy base cluster. It possesses adaptive and learning capabilities, enabling it to cope with the randomness and volatility of renewable energy power generation and ensuring the safe and efficient operation of the system.
[0056] S150: Based on the reactive power interaction capacity between the target new energy base clusters at the next moment, send control commands to the corresponding new energy bases to achieve voltage stability control between the target new energy base clusters.
[0057] Among them, the reactive power interaction capacity between the target new energy base clusters at the next moment is represented by positive and negative values to indicate the reactive power transmission direction between the two new energy bases.
[0058] For example, let's continue with the illustration using the target new energy base cluster, which includes new energy base A, new energy base B, and new energy base C: Based on the reactive power interaction capacity y output in step S140 i The system generates specific control commands: When y1>0, a control command is generated: New energy base A transmits reactive power of magnitude |y1| to new energy base B through flexible interconnection equipment.
[0059] When y1<0, a control command is generated: New energy base B transmits reactive power of magnitude |y1| to new energy base A through flexible interconnection equipment.
[0060] When y2>0, a control command is generated: New energy base A transmits reactive power of magnitude |y2| to new energy base C through flexible interconnection equipment.
[0061] When y2<0, a control command is generated: New energy base C transmits reactive power of magnitude |y2| to new energy base A through flexible interconnection equipment.
[0062] When y3>0, a control command is generated: New energy base B transmits reactive power of magnitude |y3| to new energy base C through flexible interconnection equipment.
[0063] When y3 < 0, a control command is generated: New energy base C transmits reactive power of magnitude |y3| to new energy base B through flexible interconnection equipment.
[0064] Control commands are transmitted in real time to the flexible interconnected equipment control units at each new energy base via the communication system. The commands include a clear direction of reactive power transmission, transmission capacity, and execution time (next sampling time).
[0065] This application also provides a voltage stability optimization and control system for a new energy base cluster based on flexible interconnection. The system includes: The data acquisition module is used to acquire voltage stability parameters between target new energy base clusters according to a preset sampling time interval, and to construct a time series of voltage stability parameters. The normalization module is used to normalize the voltage stability parameters based on the time series of the voltage stability parameters, and obtain the normalized values of the voltage stability parameters. The mean processing module is used to perform mean processing on the normalized values of the voltage stability parameters to obtain the average value of the voltage stability parameters; The prediction module is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment, based on the average value of voltage stability parameters and using a multi-strategy collaborative optimization algorithm based on reinforcement learning. The control module is used to send control commands to the corresponding new energy bases based on the reactive power interaction capacity between the target new energy base clusters at the next moment, so as to achieve voltage stability control between the target new energy base clusters.
[0066] The execution process of the system part of this application embodiment is the same as that of the method part of the embodiment described above, and will not be repeated here.
[0067] This application also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods described in the above embodiments.
[0068] This application also proposes a computer storage medium storing a computer program, which, when executed by a processor, implements the methods described in the above embodiments.
[0069] Computer storage media may be simply referred to as media. Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Dual Data SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Memory Bus Dynamic RAM (RDRAM). The various embodiments described in this specification are presented in a progressive manner, with reference allowed to each other for similar or identical parts. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for apparatuses, devices, and non-volatile computer storage media are described simply because they are substantially similar to the method embodiments; relevant details can be found in the descriptions of the method embodiments.
[0070] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this application.
Claims
1. A voltage stability optimization control method for new energy base clusters based on flexible interconnection, characterized in that, include: Voltage stability parameters between target new energy base clusters are obtained according to a preset sampling time interval, and a voltage stability parameter time series is constructed. Based on the time series of the voltage stability parameters, the voltage stability parameters are normalized to obtain normalized values of the voltage stability parameters. The normalized values of the voltage stability parameters are averaged to obtain the average value of the voltage stability parameters; Based on the average value of the voltage stability parameters, a multi-strategy collaborative optimization algorithm based on reinforcement learning is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment. Based on the reactive power interaction capacity between the target new energy base clusters at the next moment, control commands are sent to the corresponding new energy bases to achieve voltage stability control between the target new energy base clusters.
2. The method according to claim 1, characterized in that, The voltage stability parameters include the maximum node voltage under different voltage levels of transmission lines, the total reactive power output of each new energy base in the target new energy base cluster, and the reactive power interaction of flexible interconnection between any two new energy bases.
3. The method according to claim 1, characterized in that, The normalized values of the voltage stability parameters are averaged, including: for each voltage stability parameter, calculating the arithmetic mean of the normalized values of the voltage stability parameter at all sampling times in the voltage stability parameter time series.
4. The method according to claim 1, characterized in that, The multi-strategy collaborative optimization algorithm based on reinforcement learning includes: Train multiple policy networks, each policy network corresponding to the flexible interconnection reactive power interaction between any two new energy bases in the target new energy base cluster, and construct a reactive power interaction objective function based on each policy network. With the goal of maximizing the reactive interaction objective function, the parameters of each policy network are iteratively updated using the gradient ascent method; Each policy network is trained using the mean squared error loss function to minimize the mean squared error loss function and obtain the reactive power interaction error reference value. The reactive power interaction capacity between the target new energy base clusters at the next moment is calculated based on the reactive power interaction objective function, the reactive power interaction error reference value, and the reactive power interaction error correction coefficient.
5. The method according to claim 4, characterized in that, The reactive power interaction objective function is the average value of the reactive power interaction expected function under each strategy network, wherein each strategy network is trained based on the average value of the historical flexible interconnection reactive power interaction between the two corresponding new energy bases.
6. The method according to claim 5, characterized in that, The expression for the reactive power interaction objective function is as follows: ； Where N represents the total number of policy networks, and i represents the i-th policy network. For the i-th policy network The effectiveness of reactive power interaction adjustment, o i Represents the i-th policy network The input state value, π i This represents the i-th policy function; The objective function for reactive power interaction among target new energy clusters; For the i-th policy network The expected reactive power interaction function between the target new energy clusters.
7. The method according to claim 1, characterized in that, The reactive power interaction capacity between the target new energy base clusters at the next moment is represented by positive and negative values to indicate the reactive power transmission direction between the two new energy bases.
8. A voltage stability optimization control system for a new energy base cluster based on flexible interconnection, characterized in that, The system includes the following modules: The data acquisition module is used to acquire voltage stability parameters between target new energy base clusters according to a preset sampling time interval, and to construct a time series of voltage stability parameters. The normalization module is used to normalize the voltage stability parameters according to the time series of the voltage stability parameters to obtain the normalized values of the voltage stability parameters. The mean processing module is used to perform mean processing on the normalized values of the voltage stability parameters to obtain the average value of the voltage stability parameters; The prediction module is used to output the reactive power interaction capacity between the target new energy base clusters at the next moment, based on the average value of the voltage stability parameters and using a multi-strategy collaborative optimization algorithm based on reinforcement learning. The control module is used to send control commands to the corresponding new energy bases according to the reactive power interaction capacity between the target new energy base clusters at the next moment, so as to realize voltage stability control between the target new energy base clusters.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program, implements the method of any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed in a computer, causes the computer to perform the method described in any one of claims 1-7.