An autonomous optimal seeking method system of power distribution network voltage control strategy and storage medium

Abstract

The application discloses a kind of power distribution network voltage control strategy's autonomous optimization method system and storage medium.The method real-time acquisition voltage amplitude data of each node of power distribution network, is based on the probability distribution of voltage amplitude and calculates voltage information entropy value;Strategy set containing multiple candidate voltage control strategies is maintained, each candidate strategy is associated with independent priority weight;According to current voltage information entropy value, target control strategy is selected and issued for execution;After execution, voltage information entropy value is calculated again, and the priority weight of target strategy is updated according to the change of entropy value before and after execution;When the entropy value after execution is not less than the entropy value before execution, the candidate strategy with the highest priority weight is selected from the strategy set to replace the current target strategy.The application quantifies voltage uncertainty by information entropy, and realizes the dynamic optimization of control strategy by using entropy value change as feedback signal, without relying on accurate prediction model and fixed control rule, with strong adaptability and robustness.

Description

Technical Field

[0001] This invention belongs to the field of power system voltage control technology, specifically relating to an autonomous optimization method, system, and storage medium for distribution network voltage control strategy. Background Technology

[0002] The basic principle of reactive power and voltage control in distribution networks is "hierarchical and zoned, local balancing"—maintaining voltage levels at each voltage level and in each power supply area within permissible limits through hierarchical reactive power balancing and local reactive power compensation within zones. With the large-scale grid connection of distributed power sources such as photovoltaic and wind power, the voltage at distribution network nodes fluctuates frequently due to the randomness and volatility of new energy output, and the problem of voltage exceeding limits is becoming increasingly prominent.

[0003] Existing voltage control methods for distribution networks are mainly divided into two categories: those based on mathematical optimization models and those based on fixed rule bases. The former relies on accurate grid topology parameters and load forecasting models, requiring online solution of complex optimization problems, resulting in high computational load, slow response speed, and a significant decrease in control effectiveness when the grid topology changes. The latter maps voltage deviations to control quantities through a pre-set rule base, but the rule base remains fixed and cannot be adaptively adjusted according to dynamic changes in the grid's operating status.

[0004] Furthermore, existing methods lack effective quantitative evaluation tools to determine the effectiveness of control strategies after execution, and are unable to autonomously optimize and iterate control strategies based on evaluation results, preventing the control strategies from continuously learning and improving from historical experience. Simultaneously, existing methods typically perform open-loop control only for a single voltage amplitude or reactive power target, making it difficult to consider the coordination needs between multiple voltage levels and power supply areas under the principle of "hierarchical zoning and local balance," and also lacking mechanisms for online quantitative evaluation and closed-loop iteration of the control strategy's effectiveness. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an autonomous optimization method, system, and storage medium for voltage control strategies in distribution networks. By constructing a candidate strategy set containing multiple control parameters such as reactive power compensation, and quantifying the voltage / reactive power control effect with information entropy, the priority weight update of the candidate strategies is driven by the change of entropy value as a feedback signal of the control effect. This achieves dynamic selection and autonomous optimization of control strategies, thereby realizing closed-loop autonomous optimization control without relying on precise prediction models and fixed control rules, while adhering to the basic principle of "layered and zoned, local balance" for reactive power and voltage control in distribution networks.

[0006] The technical problem solved by this invention is achieved through the following technical solution:

[0007] An autonomous optimization method for distribution network voltage control strategy, comprising:

[0008] The voltage amplitude data of each node in the distribution network is collected in real time through the data acquisition interface.

[0009] Construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and calculate the voltage information entropy value of each node based on the probability distribution.

[0010] Maintain a policy set containing multiple candidate voltage control policies, each of which is associated with an independent priority weight.

[0011] The candidate voltage control strategy set is constructed as follows: From historical operating data, voltage regulation results obtained through optimization algorithms or manual regulation experience of maintenance personnel under different voltage conditions are extracted and discretized into a control parameter vector containing reactive power compensation from distributed power sources, tap changer settings of on-load tap changers, and charging / discharging power of energy storage systems, thus forming an initial candidate strategy set. During execution, if N consecutive global policy reselections fail and the current policy set size is less than a preset upper limit, Then, by applying random perturbation to the control parameter vector of the candidate policy with the highest current weight, a new candidate policy is generated and added to the policy set, thereby realizing online expansion of the policy space.

[0012] Based on the current voltage information entropy value, a target voltage control strategy is selected from the strategy set, and the target voltage control strategy is sent to the voltage regulation equipment of the distribution network for execution.

[0013] After executing the target voltage control strategy, node voltage amplitude data is collected again, and the updated voltage information entropy value is calculated.

[0014] The priority weight of the target voltage control strategy is updated based on the comparison between the updated voltage information entropy value and the voltage information entropy value before execution.

[0015] When the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution, the candidate voltage control strategy with the highest priority weight is selected from the strategy set to replace the target voltage control strategy.

[0016] In the above technical solution, the uncertainty of node voltage state is quantified by information entropy value. The smaller the information entropy value, the more concentrated and stable the node voltage distribution; the larger the information entropy value, the more dispersed and volatile the node voltage distribution. The effectiveness of the control strategy is evaluated based on the change in information entropy value before and after the execution of the control strategy. When the information entropy value decreases after the execution of the strategy, it indicates that the strategy effectively improves voltage stability, and its priority weight is increased; when the information entropy value does not decrease after the execution of the strategy, it indicates that the strategy is ineffective, triggering strategy replacement. Its beneficial effects are: it forms a complete closed loop of "execution—quantitative evaluation—feedback—update—optimization," realizing the autonomous optimization iteration of the control strategy without relying on accurate prediction models and fixed control rules. At the same time, since the candidate strategies already include control parameters such as the reactive power compensation of distributed power sources and the tap adjustment position of on-load tap changers, the voltage amplitude is closed-loop controlled by adjusting the reactive power output when the strategy is executed, thus inherently following the basic physical law of "layered and zoned, local balance" of reactive power and voltage control in distribution networks.

[0017] Preferably, the voltage information entropy value is calculated according to the following formula:

[0018]

[0019] in, This represents the voltage information entropy value. This represents the total number of voltage amplitude ranges. This indicates that the voltage amplitude falls within the first... The probability of a voltage amplitude range.

[0020] The aforementioned preferred technical solution calculates the information entropy value by discretizing the continuous voltage amplitude into a finite number of equally spaced intervals and statistically analyzing the probability distribution of each interval. Its beneficial effect is that the information entropy value can objectively and quantitatively reflect the degree of uncertainty in the node voltage state, providing a unified quantitative benchmark for evaluating the effectiveness of the control strategy.

[0021] Preferably, the priority weight is updated according to the following formula:

[0022]

[0023] in, Indicates the first The candidate voltage control strategies in the first The priority weight after the next update Indicates the first The candidate voltage control strategies in the first The priority weight after the next update This represents the voltage information entropy value before the target voltage control strategy is executed. This represents the voltage information entropy value after executing the target voltage control strategy. This represents the learning rate coefficient.

[0024] The above-mentioned preferred technical solution introduces a priority weight incremental update formula based on the change in entropy value. Only when the execution of the control strategy causes a decrease in the voltage information entropy value (i.e., ...) Only when the entropy value decreases will positive incentives be generated to update the weights; when the entropy value does not decrease, the incentive is zero. Its beneficial effect is that it makes the priority weight of the control strategy strictly reflect its historical control effect, and the strategy with better control effect will have its weight increase faster, realizing the quantitative accumulation of control experience and the survival of the fittest.

[0025] Preferably, the step of selecting the candidate voltage control strategy with the highest priority weight from the strategy set to replace the target voltage control strategy specifically includes:

[0026] Iterate through all candidate voltage control strategies in the strategy set except for the target voltage control strategy.

[0027] The candidate voltage control strategy with the highest priority weight is selected as the new target voltage control strategy.

[0028] If multiple candidate voltage control strategies have the same highest priority weight, then one of them is randomly selected as the new target voltage control strategy.

[0029] The aforementioned preferred technical solution selects the candidate strategy with the highest weight through traversing the strategy set for replacement. Its advantages are: ensuring a clear direction for strategy replacement, while avoiding deterministic bias in strategy selection through a random tie-breaking mechanism, thus improving the efficiency of strategy space exploration.

[0030] Preferably, the target voltage control strategy includes at least one of the following control parameters: reactive power compensation of the distributed power source, tap adjustment position of the on-load tap changer, and charging / discharging power of the energy storage system. The range of the reactive power compensation of the distributed power source is determined according to the hierarchical and zoning principle of reactive power voltage control in the distribution network.

[0031] The above-mentioned preferred technical solution clarifies the specific control parameter types included in the target voltage control strategy and determines the basis for selecting the reactive power compensation value. Its beneficial effects are: it enables the technical solution of this invention to be compatible with various types of voltage regulation equipment in distribution networks, possessing good versatility and scalability. Simultaneously, it ensures that the value of the reactive power compensation value conforms to the engineering practice requirements of "layered and zoned, local balance" in distribution networks.

[0032] Preferably, it further includes:

[0033] Set the information entropy convergence threshold .

[0034] When the updated voltage information entropy value is less than or equal to the information entropy convergence threshold At that time, the current target voltage control strategy will continue to operate.

[0035] When the updated voltage information entropy value is greater than the information entropy convergence threshold for a consecutive preset number of sampling periods. When this occurs, a global policy reselection is triggered.

[0036] The aforementioned preferred technical solution uses an information entropy convergence threshold as the criterion for determining the voltage stable state. Its advantages are: it provides clear termination conditions and global reselection trigger conditions for the iterative optimization of the control strategy, ensuring that the control process converges to the target state while avoiding stagnation at local optima.

[0037] Preferably, the initial priority weights of the candidate voltage control strategies are initialized based on historical operating data of the distribution network, specifically by: calculating the historical execution success rate of each candidate control strategy within the entropy value range of similar voltage information over the past month. Map the success rate to the initial priority weight. ,in The initial weights are based on this (typically 0.5). The mapping coefficient (typically 0.5) is used to give candidate strategies with historically good control performance a higher initial selection probability.

[0038] The aforementioned preferred technical solution initializes the priority weights of candidate strategies using historical operating data. Its beneficial effect is that it enables the control system to fully utilize historical experience, prioritizing control strategies with good historical performance from the initial stage, thus accelerating convergence.

[0039] Preferably, it further includes:

[0040] Calculate the global information entropy value of the distribution network, which is the weighted average of the voltage information entropy values ​​of all nodes.

[0041] When the global information entropy value continues to rise within a continuous sampling period, a global policy reselection is triggered.

[0042] The aforementioned preferred technical solution monitors the overall voltage stability of the distribution network through global information entropy values. Its advantages include: achieving seamless integration from local node control to global system control; timely triggering of global strategy reselection when local control fails to suppress the decline in global stability; and improving the reliability of the control system under large-scale disturbances.

[0043] An autonomous optimization control system for distribution network voltage control strategy, comprising:

[0044] The data acquisition module is used to collect voltage amplitude data of each node in the distribution network in real time through the data acquisition interface.

[0045] An information entropy calculation module, the input of which is connected to the output of the data acquisition module, is used to construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and to calculate the voltage information entropy value of each node based on the probability distribution.

[0046] The strategy storage module is used to store a set of strategies containing multiple candidate voltage control strategies, each of which is associated with an independent priority weight.

[0047] The strategy selection and execution module has its input terminals connected to the output terminals of the information entropy calculation module and the strategy storage module, respectively. It is used to select a target voltage control strategy from the strategy set based on the voltage information entropy value at the current moment and issue it for execution.

[0048] An iterative update module, whose input is connected to the output of the data acquisition module and whose output is connected to the update input of the strategy storage module, is used to collect node voltage amplitude data again through the data acquisition module after executing the target voltage control strategy, calculate the updated voltage information entropy value, update the priority weight of the target voltage control strategy based on the comparison result between the updated voltage information entropy value and the voltage information entropy value before execution, and trigger strategy replacement when the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution.

[0049] The aforementioned system technical solution establishes a data acquisition module, an information entropy calculation module, a strategy storage module, a strategy selection and execution module, and an iterative update module, clearly defining the connections and data flow between these modules. Its beneficial effects lie in the clear division of labor and collaborative cooperation among the modules, forming a complete autonomous optimization closed-loop control system of "acquisition—calculation—decision-execution—evaluation—update—optimization," demonstrating promising engineering application prospects.

[0050] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the autonomous optimization method for the distribution network voltage control strategy described in any of the preceding claims.

[0051] The above-described storage medium technology solution stores the method as a computer program in a computer-readable storage medium. Its advantages include improved portability and reusability, facilitating deployment and operation on different computing devices. Attached Figure Description

[0052] Figure 1 This is a flowchart illustrating the autonomous optimization method of the distribution network voltage control strategy of the present invention.

[0053] Figure 2 This is a block diagram of the autonomous optimization control system of the power distribution network voltage control strategy of the present invention. Detailed Implementation

[0054] The present invention will be further described in detail below through specific embodiments. The following embodiments are merely descriptive and not limiting, and should not be used to limit the scope of protection of the present invention.

[0055] Example 1

[0056] This embodiment provides an autonomous optimization method for voltage control strategies in power distribution networks.

[0057] like Figure 1 As shown, the method includes the following steps.

[0058] The voltage amplitude data of each node in the distribution network is collected in real time through the data acquisition interface.

[0059] Specifically, voltage transformers or phasor measurement units are installed at each node of the distribution network. Analog-to-digital converters (ADCs) convert the analog voltage signals output by the voltage transformers into digital voltage signals. The voltage amplitude data acquisition frequency is set to 100 to 1000 times per second. The acquired digital voltage signals are transmitted to the information entropy calculation module via a data bus. The voltage amplitude data acquired at each sampling moment is stored in a data buffer in floating-point format, forming a data buffer with a length of [missing information]. sliding time window data sequence, The value is determined based on the time scale of voltage fluctuations in the distribution network, with typical values ​​ranging from 50 to 200.

[0060] Construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and calculate the voltage information entropy value of each node based on the probability distribution.

[0061] For nodes Extract from its sliding time window voltage amplitude sample value The range of 0.85 to 1.15 times the rated voltage of the distribution network is evenly divided into... Equally spaced intervals The typical value ranges from 10 to 20. Statistics fall within the range of... Number of sampling points in each voltage range Then the probability of this interval is .

[0062] The voltage information entropy value is calculated according to the following formula:

[0063]

[0064] in, This represents the voltage information entropy value. This represents the total number of voltage amplitude ranges. This indicates that the voltage amplitude falls within the first... The probability of a voltage amplitude range.

[0065] Information entropy value The value range is 0 to When all voltage sample values ​​are concentrated in a single interval, This indicates that the node voltage is highly stable and has extremely low uncertainty; when the voltage sample values ​​are uniformly distributed across all nodes... When there are intervals, This indicates that the node voltage distribution is the most dispersed and has the highest uncertainty.

[0066] Maintain a policy set containing multiple candidate voltage control policies, each of which is associated with an independent priority weight.

[0067] The candidate voltage control strategy set is constructed as follows: From historical operating data, voltage regulation results obtained through optimization algorithms or manual regulation experience of maintenance personnel under different voltage conditions are extracted and discretized into a control parameter vector containing reactive power compensation from distributed power sources, tap changer settings of on-load tap changers, and charging / discharging power of energy storage systems, thus forming an initial candidate strategy set. During execution, if N consecutive global policy reselections fail and the current policy set size is less than a preset upper limit, Then, by applying random perturbation to the control parameter vector of the candidate policy with the highest current weight, a new candidate policy is generated and added to the policy set, thereby realizing online expansion of the policy space.

[0068] In this embodiment, the control parameter vector of the candidate voltage control strategy includes the reactive power compensation amount of the distributed power source. The range of values ​​for this compensation amount and the selection of candidate strategies are determined based on the principle of "hierarchical and zonal, local balance" in the reactive power voltage control of the distribution network. When the voltage exceeds the limit, the reactive power compensation capacity of the local distributed power source is given priority to achieve local reactive power balance. When the local reactive power regulation capacity is insufficient, the reactive power resources of the upper-level power grid or adjacent areas are coordinated. In the constructed strategy set, each candidate strategy is distinguished from the others based on the different combinations of its reactive power compensation amount, the tap adjustment position of the on-load tap changer, and the charging and discharging power of the energy storage system, covering a variety of control modes from "local reactive power compensation only" to "multi-level coordinated regulation".

[0069] The initial priority weights of the candidate voltage control strategies are initialized based on historical operating data of the distribution network. Specifically, this involves calculating the historical execution success rate of each candidate control strategy within the entropy value range of similar voltage information over the past month. Map the success rate to the initial priority weight. ,in The initial weights are based on this (typically 0.5). The mapping coefficient (typically 0.5) is used to give candidate strategies with historically good control performance a higher initial selection probability.

[0070] Based on the current voltage information entropy value, a target voltage control strategy is selected from the strategy set, and the target voltage control strategy is sent to the voltage regulation equipment of the distribution network for execution.

[0071] Specifically, based on the currently calculated voltage information entropy value Query strategy selection rules (e.g., ...) The interval is mapped to the corresponding candidate strategy number (or selected by weight sorting), and the target voltage control strategy is selected from the strategy set. The specific control parameter values ​​in the target voltage control strategy are determined by a linear mapping function: let the reference voltage be... The current node average voltage amplitude is The reactive power compensation amount of distributed power sources According to the formula The target tap position of the on-load tap-changing transformer is calculated by obtaining the current voltage deviation value according to the tap position lookup table and incorporating it into the voltage information entropy value. When the power is high, prioritize the setting with the larger adjustment range; the charging and discharging power of the energy storage system... According to the formula Calculation, where , For adjustment coefficients, , The gain function is defined with voltage information entropy as the independent variable; the larger the entropy value, the greater the gain, thus applying a stronger regulation effect. The reactive power compensation amount... The value range is determined according to the hierarchical and zoning principle of reactive power and voltage control in the distribution network—prioritizing ensuring the reactive power output of local distributed power sources within their capacity range to compensate for the reactive power gap required for local voltage deviation; only when the local reactive power capacity is insufficient to compensate for the voltage deviation will the deficit be treated as a coordination requirement of the upper-level power grid or adjacent areas. The reactive power compensation amount is distributed to the reactive power controller of each distributed power source, the tap adjustment position of the on-load tap-changing transformer is distributed to the on-load tap-changing transformer controller, and the charging and discharging power of the energy storage system is distributed to the energy storage system controller.

[0072] This invention primarily addresses continuous voltage control under normal operating conditions in distribution networks. It is also applicable to voltage dips caused by grid faults—when a node voltage amplitude is detected to be below a preset fault threshold, the information entropy value increases sharply, and the drive strategy set automatically switches to a candidate strategy containing the reactive current setpoint for fault ride-through mode. The reactive power compensation amount of this candidate strategy is determined according to the fault ride-through requirements, thereby achieving dynamic reactive power support during voltage dips. Under normal operating conditions, autonomous optimization control of voltage / reactive power is performed according to the aforementioned method.

[0073] After executing the target voltage control strategy, node voltage amplitude data is collected again, and the updated voltage information entropy value is calculated.

[0074] After executing the target voltage control strategy, wait for a complete control cycle before collecting the node voltage entropy value. This control cycle is set to 5 to 15 minutes (typically 10 minutes). This duration is longer than the sum of the reactive power response time (seconds) of the distributed power inverter and the mechanical action time of the on-load tap changer (approximately 30 seconds to several minutes), ensuring that the effect of the voltage regulation equipment is fully reflected in the node voltage measurements. Once the effect of the voltage regulation equipment is fully reflected, collect the voltage amplitude data of each node again through the data acquisition interface, and calculate the updated voltage entropy value using the same method as described above. .

[0075] The priority weight of the target voltage control strategy is updated based on the comparison between the updated voltage information entropy value and the voltage information entropy value before execution.

[0076] The priority weights are updated according to the following formula:

[0077]

[0078] in, Indicates the first The candidate voltage control strategies in the first The priority weight after the next update Indicates the first The candidate voltage control strategies in the first The priority weight after the next update This represents the voltage information entropy value before the target voltage control strategy is executed. This represents the voltage information entropy value after executing the target voltage control strategy. This represents the learning rate coefficient, typically ranging from 0.01 to 0.1.

[0079] The meaning of this update formula is: only when the execution of the control strategy causes the voltage information entropy value to decrease (i.e., Only then can positive incentives be generated. Update the weights; when the entropy value does not decrease, the incentive is zero and the weights remain unchanged.

[0080] When the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution, the candidate voltage control strategy with the highest priority weight is selected from the strategy set to replace the target voltage control strategy.

[0081] Specifically, firstly, compare and :

[0082] like This indicates that the current target voltage control strategy has effectively improved voltage stability and will continue to operate.

[0083] like This indicates that the current target voltage control strategy has failed to improve voltage stability, triggering a strategy replacement.

[0084] Iterate through all candidate voltage control strategies in the strategy set, excluding the target voltage control strategy.

[0085] Select the candidate voltage control strategy with the highest priority weight as the new target voltage control strategy;

[0086] If multiple candidate voltage control strategies have the same highest priority weight, then one of them is randomly selected as the new target voltage control strategy.

[0087] Example 2

[0088] Based on Example 1, this embodiment further provides an information entropy convergence threshold determination and a global strategy reselection mechanism.

[0089] Set the information entropy convergence threshold . The typical value is the maximum information entropy value. 10% represents the desired voltage stability target level.

[0090] Each time the control strategy is executed, the updated voltage information entropy value is calculated. back:

[0091] like If the current voltage state has reached the target stable level, the current target voltage control strategy will continue to operate without the need for strategy iteration.

[0092] like If the current voltage state has not yet reached the target stable level, the strategy evaluation and update process will continue.

[0093] When the updated voltage information entropy value is greater than the information entropy convergence threshold for a consecutive preset number of sampling periods (e.g., 5 consecutive control periods). When this occurs, it indicates that none of the candidate policies in the current policy set can adjust the voltage state to the target level, triggering a global policy reselection—either re-initializing the policy set or selecting a new policy from a wider range of candidate policies to add to the set.

[0094] Example 3

[0095] This embodiment, based on Embodiment 1, further provides a global information entropy monitoring and global policy reselection mechanism.

[0096] Calculate the global information entropy value of the distribution network, where the global information entropy value is the weighted average of the voltage information entropy values ​​of all nodes:

[0097]

[0098] in, This represents the global information entropy value of the distribution network. Represents a node The voltage information entropy value, Represents a node The weighting coefficients, This indicates the total number of nodes in the distribution network.

[0099] Weighting coefficient The weighting can be set according to the importance and voltage sensitivity of the nodes. Nodes connecting important loads, located in voltage-deficient areas, or close to distributed generation grid connection points are assigned higher weighting coefficients. The weighting coefficients of all nodes satisfy the following conditions: .

[0100] According to the global information entropy value The changing trend is used to determine the overall voltage stability of the distribution network. When the global information entropy value... In continuous When it continues to rise within a sampling period (i.e.) The system determines that the voltage stability of the distribution network is continuously deteriorating, and even if the local control strategies of each node are still running, it still triggers a global strategy reselection.

[0101] Global strategy reselection includes: selecting new candidate strategies from a broader candidate strategy space to add to the strategy set, reinitializing the priority weights of all candidate strategies, or triggering a joint control strategy that coordinates multiple voltage regulation devices.

[0102] Example 4

[0103] This embodiment provides an autonomous optimization control system for distribution network voltage control strategies.

[0104] like Figure 2 As shown, the system includes: a data acquisition module, an information entropy calculation module, a strategy storage module, a strategy selection and execution module, and an iterative update module.

[0105] The data acquisition module is used to collect voltage amplitude data of each node in the power distribution network in real time through the data acquisition interface.

[0106] Specifically, the data acquisition module includes an analog front-end unit, an analog-to-digital converter (ADC), a data buffer unit, and a communication interface unit. The analog front-end unit filters and amplifies the analog voltage signal output from the voltage transformer, eliminating high-frequency noise and power frequency interference. The ADC converts the processed analog voltage signal into a digital voltage signal, with a resolution of 12 to 16 bits and a sampling rate of 100 to 1000 times per second. The data buffer temporarily stores the converted digital voltage signal, forming a sliding time window data sequence. The communication interface unit transmits the voltage amplitude data from the buffer to the information entropy calculation module; the communication interface uses a parallel data bus or a high-speed serial bus.

[0107] The input end of the information entropy calculation module is connected to the output end of the data acquisition module. It is used to construct the probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and calculate the voltage information entropy value of each node based on the probability distribution.

[0108] Specifically, the information entropy calculation module includes an interval statistics unit and an entropy value calculation unit. The interval statistics unit is used to divide the voltage amplitude range into... The system divides the data into equally spaced intervals and counts the number of sampling points within each interval. The entropy calculation unit is used to calculate the information entropy formula described in Example 1 based on the probability of occurrence of each interval. Calculate the voltage entropy value of each node.

[0109] The strategy storage module is used to store a strategy set containing multiple candidate voltage control strategies, each candidate voltage control strategy being associated with an independent priority weight.

[0110] Specifically, the strategy storage module includes a strategy storage unit and a weight storage unit. The strategy storage unit stores the specific control parameters of the candidate voltage control strategies, including the reactive power compensation of the distributed power source, the tap adjustment position of the on-load tap changer, and the charging and discharging power of the energy storage system. The weight storage unit stores the priority weight values ​​corresponding to each candidate strategy, and the weight values ​​are stored in non-volatile memory in floating-point format.

[0111] The input terminal of the strategy selection and execution module is connected to the output terminal of the information entropy calculation module and the output terminal of the strategy storage module, respectively, and is used to select the target voltage control strategy from the strategy set according to the voltage information entropy value at the current time and issue it for execution.

[0112] Specifically, the strategy selection and execution module includes a strategy query unit and an instruction issuance unit. The strategy query unit is used to query the strategy storage module based on the current information entropy value and select a target strategy that matches the current entropy value range. The instruction issuance unit is used to convert the control parameter vector of the target strategy into control instructions and issue them to each voltage regulation device.

[0113] The input of the iterative update module is connected to the output of the data acquisition module, and the output of the iterative update module is connected to the update input of the strategy storage module. It is used to collect node voltage amplitude data again through the data acquisition module after executing the target voltage control strategy, calculate the updated voltage information entropy value, update the priority weight of the target voltage control strategy according to the comparison result between the updated voltage information entropy value and the voltage information entropy value before execution, and trigger strategy replacement when the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution.

[0114] Specifically, the iterative update module includes an effect evaluation unit, a priority weight update unit, and a strategy replacement triggering unit. The effect evaluation unit is used to compare the information entropy values ​​after control. Information entropy value before control Calculate the change in entropy. The priority weight update unit is used to update the priority weight according to the priority weight update formula described in Example 1, based on the control effect. Update the priority weight of the corresponding policy in the policy storage module. The policy replacement trigger unit is used to... When a policy replacement is triggered, the candidate policy with the highest priority weight is selected from the policy set to replace the current target policy.

[0115] The output of the data acquisition module is connected to the input of the information entropy calculation module. The output of the information entropy calculation module is connected to the first input of the strategy selection and execution module. The output of the strategy storage module is connected to the second input of the strategy selection and execution module. The output of the strategy selection and execution module is connected to the voltage regulation equipment of the power distribution network. The first input of the iterative update module is connected to the output of the data acquisition module. The second input of the iterative update module is connected to the output of the information entropy calculation module. The output of the iterative update module is connected to the update input of the strategy storage module.

[0116] The connections and data flows between the modules form a complete autonomous optimization closed-loop control system of "acquisition—calculation—decision-execution—evaluation—update—optimization".

[0117] Example 5

[0118] This embodiment provides a computer-readable storage medium.

[0119] The computer-readable storage medium stores a computer program that, when executed by a processor, implements an autonomous optimization method for the distribution network voltage control strategy as described in Embodiment 1 or Embodiment 2.

[0120] The computer-readable storage medium is a read-only memory, random access memory, disk storage medium, optical disk storage medium, or flash memory device.

[0121] The computer program includes computer-readable code that, when executed on a computer, causes the computer to perform the following operations:

[0122] The voltage amplitude data of each node in the distribution network is collected in real time through the data acquisition interface.

[0123] Construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and calculate the voltage information entropy value of each node based on the probability distribution.

[0124] Maintain a policy set containing multiple candidate voltage control policies, each of which is associated with an independent priority weight.

[0125] Based on the current voltage information entropy value, a target voltage control strategy is selected from the strategy set, and the target voltage control strategy is sent to the voltage regulation equipment of the distribution network for execution.

[0126] After executing the target voltage control strategy, node voltage amplitude data is collected again, and the updated voltage information entropy value is calculated.

[0127] The priority weight of the target voltage control strategy is updated based on the comparison between the updated voltage information entropy value and the voltage information entropy value before execution.

[0128] When the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution, the candidate voltage control strategy with the highest priority weight is selected from the strategy set to replace the target voltage control strategy.

[0129] By storing the method in the form of a computer program on a computer-readable storage medium, it is easy to deploy and run on different computing devices.

[0130] Example 6

[0131] This embodiment illustrates the technical effects of the present invention through a specific application example.

[0132] Taking a 33-node distribution network with a high proportion of distributed photovoltaic power as an example, the following three methods are used for comparison:

[0133] Method 1: Traditional fixed-rule voltage control method (using a fixed "if-then" rule base, the rules do not change with the system state).

[0134] Method 2: Information entropy-driven adaptive policy selection method (including information entropy calculation and policy selection, but excluding priority weight iteration update and policy replacement).

[0135] Method 3: The information entropy-driven autonomous optimization control method of the present invention (including a complete closed loop of information entropy calculation, strategy selection, priority weight iterative update and strategy replacement).

[0136] The simulation results are shown in the table below:

[0137]

[0138] From the above results, we can conclude that:

[0139] Compared with Method 1, Method 2 reduced the average voltage deviation by 25.0% and the duration of voltage over-limit by 51.2%, indicating that the information entropy-driven adaptive selection of the strategy can effectively improve the voltage control accuracy.

[0140] Compared to Method 2, Method 3 further reduced the average voltage deviation by 37.5%, reduced the duration of voltage over-limit by 69.0%, and improved the effectiveness of the control strategy to 95.3%. It only requires 8 iterations to converge to a stable state, indicating that the introduction of priority weight iteration update and strategy replacement mechanism enables the control system to continuously learn from historical control experience and autonomously optimize itself.

[0141] Furthermore, Method 3 is superior to Method 1 and Method 2 in terms of average response time of the control strategy, indicating that the technical solution of the present invention improves control accuracy while taking into account the speed of control response.

[0142] In summary, this invention quantifies the uncertainty of node voltage states through information entropy, and uses entropy changes as feedback signals to drive the priority weight updates of candidate strategies, thereby achieving dynamic selection and autonomous optimization of control strategies. While adhering to the basic principles of "layered and zoned, local balancing" reactive power and voltage control in distribution networks, this method achieves coordinated voltage / reactive power control through optimized adjustment of control parameters such as reactive power compensation in candidate strategies. It outperforms traditional methods in terms of voltage deviation suppression, voltage limit elimination, control response speed, and control strategy effectiveness, demonstrating promising prospects for engineering applications.

[0143] Although embodiments and drawings of the present invention have been disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, variations and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims

1. An autonomous optimization method for voltage control strategy in a distribution network, characterized in that, include: The voltage amplitude data of each node in the distribution network is collected in real time through the data acquisition interface; Construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and calculate the voltage information entropy value of each node based on the probability distribution; Maintain a policy set containing multiple candidate voltage control policies, each candidate voltage control policy being associated with an independent priority weight; The candidate voltage control strategy set is constructed in the following way: From historical operating data, voltage regulation results obtained through optimization algorithms or manual regulation experience of maintenance personnel under different voltage conditions are extracted and discretized into control parameter vectors containing reactive power compensation of distributed power sources, tap adjustment positions of on-load tap changers, and charging and discharging power of energy storage systems, forming an initial candidate strategy set. During execution, if global strategy reselection fails N times consecutively and the current strategy set size is less than a preset upper limit, a new candidate strategy is generated and added to the strategy set by applying random perturbation to the control parameter vector of the candidate strategy with the highest current weight. Based on the current voltage information entropy value, a target voltage control strategy is selected from the strategy set, and the target voltage control strategy is sent to the voltage regulation equipment of the distribution network for execution. After executing the target voltage control strategy, node voltage amplitude data is collected again, and the updated voltage information entropy value is calculated. Based on the comparison between the updated voltage information entropy value and the voltage information entropy value before execution, the priority weight of the target voltage control strategy is updated. When the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution, the candidate voltage control strategy with the highest priority weight is selected from the strategy set to replace the target voltage control strategy.

2. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, The voltage information entropy value is calculated according to the following formula: ； in, This represents the voltage information entropy value. This represents the total number of voltage amplitude ranges. This indicates that the voltage amplitude falls within the first... The probability of a voltage amplitude range.

3. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, The priority weights are updated according to the following formula: ； in, Indicates the first The candidate voltage control strategies in the first The priority weight after the next update Indicates the first The candidate voltage control strategies in the first The priority weight after the next update This represents the voltage information entropy value before the target voltage control strategy is executed. This represents the voltage information entropy value after executing the target voltage control strategy. This represents the learning rate coefficient.

4. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, The step of selecting the candidate voltage control strategy with the highest priority weight from the strategy set to replace the target voltage control strategy specifically includes: Iterate through all candidate voltage control strategies in the strategy set, excluding the target voltage control strategy. Select the candidate voltage control strategy with the highest priority weight as the new target voltage control strategy; If multiple candidate voltage control strategies have the same highest priority weight, then one of them is randomly selected as the new target voltage control strategy.

5. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, The target voltage control strategy includes at least one of the following control parameters: reactive power compensation of distributed power sources, tap adjustment position of on-load tap changer, and charging and discharging power of energy storage system. The range of values ​​for reactive power compensation of distributed power sources is determined according to the hierarchical and zoning principle of reactive voltage control in distribution networks.

6. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, Also includes: Set the information entropy convergence threshold ; When the updated voltage information entropy value is less than or equal to the information entropy convergence threshold At the same time, the current target voltage control strategy will continue to operate; When the updated voltage information entropy value is greater than the information entropy convergence threshold for a consecutive preset number of sampling periods. When this occurs, a global policy reselection is triggered.

7. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, The initial priority weights of the candidate voltage control strategies are initialized based on historical operating data of the distribution network. Specifically, this involves calculating the historical execution success rate of each candidate control strategy within the entropy value range of similar voltage information over the past month, and mapping the success rate to the initial priority weights. ,in Based on the initial weights, The mapping coefficients are used to give candidate strategies with historically good control performance a higher initial selection probability.

8. The autonomous optimization method for distribution network voltage control strategy according to claim 1, characterized in that, Also includes: Calculate the global information entropy value of the distribution network, where the global information entropy value is the weighted average of the voltage information entropy values ​​of all nodes; When the global information entropy value continues to rise within a continuous sampling period, a global policy reselection is triggered.

9. An autonomous optimization control system for distribution network voltage control strategy, characterized in that, include: The data acquisition module is used to collect voltage amplitude data of each node in the distribution network in real time through the data acquisition interface; An information entropy calculation module, the input of which is connected to the output of the data acquisition module, is used to construct a probability distribution of the voltage amplitude of each node based on the voltage amplitude data of each node, and to calculate the voltage information entropy value of each node based on the probability distribution. The strategy storage module is used to store a strategy set containing multiple candidate voltage control strategies, each of which is associated with an independent priority weight. The strategy selection and execution module has its input terminals connected to the output terminals of the information entropy calculation module and the strategy storage module, respectively. It is used to select a target voltage control strategy from the strategy set based on the voltage information entropy value at the current moment and issue it for execution. An iterative update module, whose input is connected to the output of the data acquisition module and whose output is connected to the update input of the strategy storage module, is used to collect node voltage amplitude data again through the data acquisition module after executing the target voltage control strategy, calculate the updated voltage information entropy value, update the priority weight of the target voltage control strategy based on the comparison result between the updated voltage information entropy value and the voltage information entropy value before execution, and trigger strategy replacement when the updated voltage information entropy value is greater than or equal to the voltage information entropy value before execution.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements an autonomous optimization method for a distribution network voltage control strategy as described in any one of claims 1 to 8.

Images