Active voltage management method in substations based on multi-agent self-optimization control strategy

Abstract

The present invention relates to the technical field of power supply and distribution systems, and specifically to a method for active voltage management of a substation based on a multi-agent self-optimization control strategy. If the current agent meets the conditions for starting control management, the cost contribution access point and the management cost of the cost contribution access point in each access point are determined according to the different types of electrical parameters of each access point within the jurisdiction of the current agent, and then the adjusted management budget cost of the current agent and the collaborative management evaluation parameters between the current agent and other agents are determined; the voltage scheduling plan is determined according to the current voltage and adjusted management budget cost of the current agent, as well as the collaborative management evaluation parameters between the current agent and other agents. The present invention improves the accuracy of the determined voltage scheduling plan by accurately evaluating the management cost of the current agent and the collaborative management situation between the current agent and other agents.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply and distribution systems, and in particular to a method for actively managing substation voltage based on a multi-agent self-optimizing control strategy. Background Art

[0002] Substation voltage management refers to addressing unstable voltage at the end of the distribution system (the area supplied by the distribution transformer) by implementing various technical measures and management strategies to ensure voltage remains within a reasonable range and guarantee the stability and quality of the power supply. Low voltage can cause household appliances to malfunction, disrupt industrial production, and increase grid losses, while high voltage can damage electrical equipment. Therefore, substation voltage management is crucial for ensuring power supply security and improving customer satisfaction.

[0003] Multi-agent technology, an emerging research direction, is gradually being applied to substation voltage management by enabling multiple autonomous decision-making units to work together to complete specific tasks. However, existing technologies for multi-agent voltage management often suffer from issues such as poor matching between the agent voltage adjustments and the surrounding environment, and the resulting drastic adjustments to short-term fluctuations during the collaborative process. This results in inaccurate substation voltage adjustments and an inability to guarantee power supply stability and quality. Summary of the Invention

[0004] In order to solve the technical problem of inaccurate voltage adjustment in the above-mentioned substation, the purpose of the present invention is to provide a method for active voltage management in the substation based on a multi-agent self-optimization control strategy. The technical solution adopted is as follows:

[0005] In a first aspect, the present invention provides a method for active voltage management in a substation based on a multi-agent self-optimization control strategy, the method comprising:

[0006] If the current intelligent body meets the conditions for starting regulation and control, the cost contribution access point and the governance cost of the cost contribution access point among the access points are determined according to the different types of electrical parameters of the access points within the jurisdiction of the current intelligent body;

[0007] Based on the governance cost of the cost contribution access point, the governance budget cost of the current agent is determined. Based on the governance cost and the governance budget cost, the optimization cycle required for the current agent is determined. Then, based on the optimization cycle required for all agents in the level where the current agent is located, the optimization planning cycle of the level where the current agent is located is determined.

[0008] According to the governance budget costs and optimal planning cycles of the current agent and other agents, the collaborative governance evaluation parameters between the current agent and other agents are determined, and the governance budget costs of the current agent are adjusted according to the collaborative governance evaluation parameters and the governance budget costs of other agents to obtain the adjusted governance budget costs;

[0009] Determine the voltage scheduling plan based on the current voltage of the intelligent body, the adjustment governance budget cost and the collaborative governance evaluation parameters.

[0010] In conjunction with the foregoing first aspect, in some possible implementations, determining a cost-contributing access point among the access points includes:

[0011] According to the different types of electrical parameters of each access point within the current intelligent agent's jurisdiction, the system state vector of each access point is constructed;

[0012] Determine a state transfer matrix according to the system state vector, wherein the dimension of the state transfer matrix is ​​determined according to the type of the electrical parameter, and each element in the state transfer matrix corresponds to two types of the electrical parameter;

[0013] Determining abnormal fluctuation elements in the state transfer matrix according to the distribution of elements in the state transfer matrix;

[0014] Marking the type of the electrical parameter of each access point according to the abnormal fluctuation element to obtain the number of sub-marks of each type of the electrical parameter of each access point;

[0015] An accumulated value of all sub-marking times of each access point is determined as the marking time of each access point, and an access point with a marking time greater than 0 is determined as a cost-contributing access point.

[0016] In conjunction with the foregoing first aspect, in some possible implementations, determining the abnormal fluctuation element in the state transition matrix includes:

[0017] Obtaining a histogram corresponding to all elements in the state transition matrix;

[0018] Determine the height sequence corresponding to the height value of each peak column in the histogram, and determine the first-order difference sequence of the height sequence;

[0019] Determine the two target peak cylinders corresponding to the maximum difference value in the first-order difference sequence, determine the average value of all elements contained in each target peak cylinder as the element mean, and determine the maximum element mean of the two element means;

[0020] The peak column corresponding to the maximum element mean in the histogram and all elements contained in the peak column where the maximum element in the state transfer matrix is ​​located are determined as abnormal fluctuation elements in the state transfer matrix.

[0021] In conjunction with the first aspect above, in some possible implementations, determining the governance cost of the cost contribution access point includes:

[0022] Determine the difference between the real-time voltage of the cost contribution access point and the prior adaptation voltage as the voltage adjustment target;

[0023] Determine the range value of the number of sub-marks of each type of electrical parameters of the cost contribution access point;

[0024] The ratio of the number of times the cost contribution access point is marked to the range value is determined, and the product of the ratio and the voltage adjustment target is determined as the management cost of the cost contribution access point.

[0025] In conjunction with the first aspect above, in some possible implementations, determining the governance budget cost of the current intelligent entity includes:

[0026] Determine the superposition of the governance costs of all cost-contributing access points as the access point cost loss;

[0027] Determine the standard deviation of the governance costs of all cost contribution access points as the governance cost standard deviation;

[0028] Determine the difference between the governance cost of each cost-contributing access point and the cost loss of the access point as the governance cost deviation value of each access point;

[0029] The governance budget cost of the current intelligent entity is determined based on the access point cost loss and the standard deviation of the governance cost, combined with the governance cost deviation values ​​of all cost-contributing access points.

[0030] In conjunction with the first aspect above, in some possible implementations, determining the period required for the current agent to reach optimization includes:

[0031] Determine a number of local governance point pairs based on the governance costs of all cost-contributing access points, where the local governance point pairs include two cost-contributing access points.

[0032] Determine the cumulative value of the governance costs of all local governance points for all access points with different cost contributions as the total governance cost, and determine the rounded-up result of the ratio of the total governance cost to the absolute value of the governance budget cost of the current intelligent body as the period required for the current intelligent body to reach optimization.

[0033] In conjunction with the first aspect above, in some possible implementations, determining the optimal planning period of the current agent's level includes:

[0034] Determine the target interval for the optimization cycle of all agents in the current agent's level;

[0035] Determine the target optimization period in the centralized target interval among the optimization periods of all agents in the level where the current agent is located, and determine the average value of all target optimization periods as the optimization planning period of the level where the current agent is located.

[0036] In conjunction with the first aspect above, in some possible implementations, determining the collaborative governance evaluation parameters between the current agent and other agents includes:

[0037] Determine the sum of the governance budget costs of the current agent and each other agent as the sum of the governance costs;

[0038] Determine the ratio of the optimal planning period of the current agent and each other agent in the same level as the planning period ratio;

[0039] According to the sum of the governance costs and the ratio of the planning cycles, the collaborative governance evaluation parameters between the current intelligent agent and other intelligent agents are determined.

[0040] In conjunction with the first aspect above, in some possible implementations, adjusting the governance budget cost of the current agent to obtain the adjusted governance budget cost includes:

[0041] Determine the mean and variance of the collaborative governance evaluation parameters between all agents in the current agent's level and other agents, and obtain the mean and variance of the evaluation parameters;

[0042] Determine the dynamic threshold of the evaluation parameter of the current agent based on the mean and variance of the evaluation parameter;

[0043] Determine the target other intelligent agents of the current intelligent agent, and the collaborative governance evaluation parameter between the current intelligent agent and the target other intelligent agents is less than the dynamic threshold of the evaluation parameter;

[0044] Determine the cumulative value of the collaborative governance evaluation parameter between the current intelligent agent and all other target intelligent agents as the cumulative collaborative governance evaluation parameter;

[0045] Determine the cumulative governance budget cost of all other target agents of the current agent as the cumulative governance budget cost, and determine the governance budget cost adjustment coefficient based on the ratio of the cumulative collaborative governance evaluation parameter to the cumulative governance budget cost;

[0046] Determine the product of the current agent's governance budget cost and the governance budget cost adjustment coefficient as the current agent's adjusted governance budget cost.

[0047] In conjunction with the first aspect above, in some possible implementations, determining a voltage scheduling scheme includes:

[0048] The current voltage and adjusted governance budget cost of the current agent, as well as the collaborative governance evaluation parameters between the current agent and all target other agents are input into the neural network, and the neural network outputs the target other agents that can be scheduled among all target other agents of the current agent and the voltage values ​​that need to be scheduled.

[0049] In a second aspect, the present invention further provides a system for actively managing substation voltage based on a multi-agent self-optimizing control strategy, comprising a memory and a processor. The memory is configured to store executable computer program code, and the processor is configured to retrieve and execute the executable computer program code from the memory, causing the system to execute a method for actively managing substation voltage based on a multi-agent self-optimizing control strategy according to the first aspect or any possible implementation of the first aspect.

[0050] In a third aspect, the present invention further provides a computer program product, which includes: computer program code, which, when running on a computer, enables the computer to execute the method in the above-mentioned first aspect or any possible implementation of the first aspect.

[0051] In a fourth aspect, the present invention further provides a computer-readable storage medium, which stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above-mentioned first aspect or any possible implementation of the first aspect.

[0052] The present invention has the following beneficial effects: when the current intelligent body starts the regulation and control governance, the governance budget cost of the current intelligent body is determined by determining the cost contribution access points and the governance cost of the cost contribution access points within the jurisdiction of the previous intelligent body, so as to make a preliminary budget for the governance cost of the current intelligent body; according to the governance cost of the cost contribution access points and the governance budget cost of the current intelligent body, the required period for the optimization of the current intelligent body is determined, and according to the required period for the optimization of all intelligent bodies in the level where the current intelligent body is located, the optimization planning period of the level where the current intelligent body is located is determined, and then the collaborative governance evaluation parameters between the current intelligent body and the other intelligent bodies are determined, so as to adjust the governance budget cost of the current intelligent body by using the collaborative governance evaluation parameters and the governance budget cost of other intelligent bodies, and obtain the adjusted governance budget cost of the current intelligent body; finally, according to the current voltage of the current intelligent body and the adjusted governance budget cost, as well as the collaborative governance evaluation parameters between the current intelligent body and the other intelligent bodies, the voltage scheduling plan is determined. The present invention determines the voltage scheduling plan by accurately evaluating the governance cost of the current intelligent body and the collaborative governance between the current intelligent body and other intelligent bodies, thereby improving the matching degree between the intelligent body voltage adjustment and the surrounding environment, thereby realizing precise voltage adjustment within the jurisdiction of the current intelligent body, and effectively ensuring the power supply stability and quality in the jurisdiction of the current intelligent body. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 This is a flowchart of a method for active voltage management in a transformer substation based on a multi-agent self-optimization control strategy according to an embodiment of the present invention;

[0055] Figure 2 Schematic diagram of the distribution of agents at different levels according to an embodiment of the present invention;

[0056] Figure 3 This is a structural diagram of an active substation voltage management system based on a multi-agent self-optimizing control strategy according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to clearly illustrate the technical features of this solution, the present invention is described in detail below through specific implementation methods in conjunction with the accompanying drawings.

[0058] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0059] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0060] As used herein, the term "including" and its variations are open-ended, i.e., "including but not limited to." The term "based on" means "based, at least in part, on." The term "one embodiment" means "at least one embodiment," the term "another embodiment" means "at least one additional embodiment," and the term "some embodiments" means "at least some embodiments." Other terms are defined in the following description.

[0061] It should be noted that the concepts of "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0062] Although operations or steps are described in a particular order in the drawings in the embodiments of the present invention, this should not be understood as requiring that these operations or steps be performed in the particular order shown or in a serial order, or that all of the operations or steps shown be performed to obtain a desired result. In the embodiments of the present invention, these operations or steps may be performed serially; they may also be performed in parallel; or a portion of these operations or steps may be performed.

[0063] At the same time, it is understood that the data involved in the technical solutions of the present invention (including but not limited to the data itself, the acquisition or use of the data) must comply with the requirements of relevant laws, regulations and relevant provisions. Unless otherwise defined, all technical and scientific terms used in this invention have the same meanings as those commonly understood by those skilled in the art to which this invention belongs, and all parameters or indicators in the formulas involved in this invention are normalized values ​​to eliminate dimension effects.

[0064] The following will introduce in detail a method for active voltage management in a transformer substation based on a multi-agent self-optimizing control strategy provided by an embodiment of the present invention in conjunction with the accompanying drawings.

[0065] Figure 1FIG. 1 shows a basic flow chart of a method for active voltage management in a transformer substation based on a multi-agent self-optimizing control strategy according to an embodiment of the present invention. Figure 1 As shown, the method specifically includes the following steps:

[0066] Step S100: If the current intelligent body meets the conditions for starting regulation and control, the cost contribution access point and the governance cost of the cost contribution access point among the access points are determined according to different types of electrical parameters of the access points within the jurisdiction of the current intelligent body.

[0067] Specifically, the purpose of the intelligent agent's voltage management in the substation is to adjust the inconsistent voltage in the jurisdiction to ensure that the voltage in the jurisdiction responds in seconds, thereby limiting the voltage fluctuation and ensuring the stability and quality of the power supply in the jurisdiction.

[0068] The intelligent agent is regarded as a coupling of physical nodes and computing units deployed at the power grid terminal, and any intelligent agent is defined as ; Among them, S represents the node data unit, which collects the node voltage, node active power, node reactive power, etc. at the terminal by deploying power sensors; A represents the node action unit, which is responsible for executing node actions, such as circuit switching, tap adjustment, inverter set point adjustment, etc.; R represents the node communication unit, which is responsible for communicating with other intelligent agents.

[0069] For the current agent Pre-set regulation voltage range (205V-225V) is the limit range of the trigger control. The real-time monitoring voltage of any entry point within the jurisdiction of the current agent z When , the current agent is judged If the conditions for starting regulation and control are met, then at the current moment Next start regulatory governance: Current intelligent body By sensing the changes in power data output within its jurisdiction, the voltage environment within the jurisdiction is judged to evaluate the voltage resources required for governance, thereby judging the ability of the jurisdiction to respond to voltage environment governance in the process of voltage optimization using the voltage resources that can be used locally.

[0070] Using the current agent The node data unit S obtains the current agent The voltage of each access point within the jurisdiction , active power , reactive power , thus obtaining the current agent Different types of electrical parameters of each access point within the jurisdiction. Different types of electrical parameters of each access point within the jurisdiction are used to screen the cost contribution access points among all access points and determine the governance cost of the cost contribution access points. The governance cost reflects the governance cost of the current intelligent body. The cost of voltage management at each access point within the jurisdiction is contributed by the cost.

[0071] Step S200: Determine the governance budget cost of the current intelligent entity based on the governance cost of the cost contribution access point, and determine the period required for the current intelligent entity to reach optimization based on the governance cost and the governance budget cost, and then determine the optimization planning period of the level where the current intelligent entity is located based on the period required for the optimization of all intelligent entities in the level where the current intelligent entity is located.

[0072] Specifically, by Analyze the governance cost of all cost contribution access points within the jurisdiction of the current intelligent entity A preliminary assessment of the governance cost of the current intelligent entity is conducted. All costs within the jurisdiction contribute to the governance cost of the access point and the current intelligent entity The governance budget cost of the current intelligent entity is determined The required period for optimization reflects the current intelligent agent The deviation between local governance costs and total governance costs within a jurisdiction.

[0073] Determine the current agent as above The same way as the optimization period required for the current agent can be determined The period required for other agents in the same layer to reach excellence. Figure 2 Give a distribution diagram of agents at different levels. Figure 2 In the above figure, voltage regulation agent, temperature control agent and distributed power supply agent belong to the same level. The deviation of the cycle required for the optimization of all agents at the same level is used to obtain the current agent The optimal planning cycle of the level. Analyze the optimal planning cycle and optimal cost balance at the current level. The optimization planning cycle of the level is large, which means that the current agent Relying on local node voltage resource balance cannot achieve local voltage management. Therefore, it is necessary to combine local voltage management costs with external ones to achieve voltage management within the range to stabilize it. The additional cost overhead caused by the node scheduling distance generated during the management process and the power environment stability of the node itself need to be used as the basis for the scheduling process.

[0074] Step S300: Determine the collaborative governance evaluation parameters between the current intelligent agent and other intelligent agents based on the governance budget costs and optimization planning cycles of the current intelligent agent and other intelligent agents, and adjust the governance budget costs of the current intelligent agent based on the collaborative governance evaluation parameters and the governance budget costs of other intelligent agents to obtain the adjusted governance budget costs.

[0075] Specifically, based on the governance budget costs and optimal planning cycles of different intelligent entities, the collaborative governance evaluation between different intelligent entities is judged, and the collaborative governance evaluation parameters between different intelligent entities are determined to carry out collaborative governance scheduling at the same level and between different levels. When the collaborative governance evaluation parameters between different intelligent entities are small, it means that the collaborative efficiency of the two intelligent entities is insufficient. Therefore, according to the current intelligent entity Evaluation parameters for collaborative governance with other intelligent entities, as well as the governance budget costs of other intelligent entities. Other intelligent entities refer to the level of the current intelligent entity. Or an agent at any other level outside the current level, The governance budget cost is punitively adjusted to obtain the current agent The adjusted governance budget cost will reallocate the cut budget to intelligent entities with high collaborative efficiency to accelerate the global optimization process.

[0076] Step S400: Determine a voltage scheduling plan based on the current voltage of the intelligent body, the adjusted governance budget cost, and the collaborative governance evaluation parameters.

[0077] Specifically, according to the current agent The current voltage and adjusted governance budget cost, as well as the current agent Collaborative governance evaluation parameters with other intelligent agents determine the adaptive voltage dispatch scheme, thereby achieving the current intelligent agent Accurate voltage adjustment within the jurisdiction ensures the stability and quality of power supply in the substation.

[0078] In the above-mentioned active voltage management method for the substation based on the multi-agent self-optimization control strategy, when the current agent meets the conditions for starting the control management, the management budget cost of the current agent is determined according to the management cost of all cost contribution access points within the jurisdiction of the current agent, and then the optimization planning cycle of the level of the current agent is determined, and the collaborative management evaluation parameters between the current agent and other agents are determined according to the management budget cost and optimization planning cycle of the current agent and other agents, so as to make a punitive adjustment to the management budget cost of the current agent according to the collaborative management evaluation parameters between the current agent and other agents, as well as the management budget cost of other agents, and finally, according to the current voltage of the current agent and the adjusted management budget cost, as well as the collaborative management evaluation parameters between the current agent and other agents, the voltage scheduling plan is adaptively determined, thereby realizing precise voltage adjustment within the jurisdiction of the current agent, and effectively ensuring the power supply stability and quality in the jurisdiction of the current agent.

[0079] Furthermore, in a possible implementation, determining the cost-contributing access point among the access points in step S100 includes:

[0080] Step S101: constructing a system state vector of each access point according to different types of electrical parameters of each access point within the jurisdiction of the current intelligent agent.

[0081] Specifically, according to the current agent Any access point within the jurisdiction at the current moment and the voltage over a period of time (such as 1 minute), and arrange these voltages in chronological order to obtain the voltage observation vector , Represents the voltage observation vector Similarly, we can get the active power observation vector , reactive power observation vector .in, Represents the active power observation vector The nth active power in Represents the reactive power observation vector The nth reactive power in , active power observation vector and reactive power observation vector Superposition is performed to obtain the current agent The system state vector of any access point within the jurisdiction of .

[0082] Step S102: determining a state transfer matrix according to the system state vector, wherein the dimension of the state transfer matrix is ​​determined according to the type of the electrical parameter, and each element in the state transfer matrix corresponds to two types of electrical parameters.

[0083] Specifically, the current agent The system state vector of each access point within the jurisdiction Enter the current agent in sequence In the completed EKF (Extended Kalman Filter), the EKF outputs the state transfer matrix corresponding to each access point , since the predicted value is not output, the computational overhead is reduced and the computational real-time performance is increased to meet the needs of local agents for priority processing and scheduling. It should be understood that the state transfer matrix The dimension is determined according to the type of electrical parameters. Since the types of electrical parameters in this embodiment are voltage, active power and reactive power, the state transfer matrix determined is Is a 3*3 square matrix, the state transfer matrix Each element in corresponds to two types of voltage, active power and reactive power.

[0084] Step S103: determining abnormal fluctuation elements in the state transfer matrix according to the distribution of elements in the state transfer matrix.

[0085] Specifically, for the current agent The state transition matrix of each access point in the jurisdiction The influence of each element on the power supply status is judged to determine the state transfer matrix The various abnormal fluctuation elements in.

[0086] More specifically, in one implementation, determining abnormal fluctuation elements in the state transition matrix includes:

[0087] Step S1031: Obtain a histogram corresponding to all elements in the state transition matrix.

[0088] Specifically, statistics of the current agent The state transition matrix of each access point in the jurisdiction For each element r in the histogram, take the arithmetic square root of the number of elements r nr to get the histogram bar spacing , with spacing State transfer matrix After the value range of all elements r is segmented evenly, the state transfer matrix is ​​established according to the number of elements in the segment The histogram of all elements r in .

[0089] Step S1032: determining a height sequence corresponding to the height value of each peak column in the histogram, and determining a first-order difference sequence of the height sequence.

[0090] Specifically, in the histogram, along the direction of increasing element r, the height values ​​of each peak column in the histogram are arranged to obtain a height sequence, and the difference between each height value in the height sequence and its previous height value is calculated, and these differences are arranged according to the arrangement order of the height values ​​in the height sequence, thereby obtaining a first-order difference sequence of the height sequence.

[0091] Step S1033: Determine two target peak cylinders corresponding to the maximum difference value in the first-order difference sequence, determine the average value of all elements contained in each target peak cylinder as the element mean, and determine the maximum element mean of the two element means.

[0092] Specifically, in the current agent The state transition matrix of each access point in the jurisdiction In the first-order difference sequence of the corresponding height sequence, the maximum difference value is determined, and the two height values ​​of this maximum difference value in the height sequence are determined. Then, the two peak columns of these two height values ​​in the histogram are determined, and these two peak columns are used as target peak columns. The average value of all elements contained in each target peak column is determined, which is called the element mean. From this, two element means can be obtained. The larger value of the two element means is determined as the maximum element mean.

[0093] Step S1034: Determine the peak column corresponding to the maximum element mean of the histogram and all elements contained in the peak column where the maximum element in the state transfer matrix is ​​located as abnormal fluctuation elements in the state transfer matrix.

[0094] Specifically, determine the state transfer matrix The maximum value of all elements r in the histogram is in the peak column, and all elements contained in the peak column and all elements contained in the peak column corresponding to the mean of the maximum elements in the histogram are used as the state transfer matrix Abnormal fluctuation elements in.

[0095] At this point, the current agent can be accurately obtained The state transition matrix of any access point within the jurisdiction Abnormal fluctuation elements in.

[0096] Step S104: marking the type of the electrical parameter of each access point according to the abnormal fluctuation element, and obtaining the number of sub-marks of each type of the electrical parameter of each access point.

[0097] Specifically, according to the state transfer matrix The abnormal fluctuation elements in the state transfer matrix The types of electrical parameters are marked according to the identifiers in the rows and columns in , thereby obtaining the marking times of each type of electrical parameter of the corresponding access point, and the marking times are recorded as sub-marking times.

[0098] Step S105: determining the accumulated value of all sub-marking times of each access point as the marking time of each access point, and determining the access point with a marking time greater than 0 as a cost-contributing access point.

[0099] Specifically, determine the state transfer matrix The accumulated value of the number of sub-marks of each type corresponds to the state transfer matrix The number of times the corresponding access point is marked, from which the current agent can be determined The number of times each access point is marked in the jurisdiction of the current agent. The access point with a mark, that is, the number of marks is greater than 0, is determined as the current agent. The cost contribution of the access point.

[0100] By constructing the system state vector of each access point and determining the state transfer matrix, and then identifying the abnormal fluctuation elements in the state transfer matrix, the cost-contributing access points within the current intelligent body's jurisdiction can be accurately identified.

[0101] Furthermore, in a possible implementation, determining the governance cost of the cost-contributing access point in step S100 includes:

[0102] Step S106: determining a difference between the real-time voltage of the cost contribution access point and the a priori adapted voltage as a voltage adjustment target.

[0103] Specifically, determine the current cost contribution access point Real-time voltage A priori adaptation voltage The difference ( ) as the voltage adjustment target. Among them, the prior adaptation voltage Refers to the current cost contribution access point rated voltage.

[0104] Step S107: determining the range value of the number of sub-marks of each type of electrical parameter of the cost-contributing access point.

[0105] Specifically, the access point that contributes to the current cost The number of sub-tags of each type of electrical parameter is counted to calculate the current cost contribution access point Differentiate the power supply parameter anomaly types, that is, determine the maximum value of the sub-mark times of each type and minimum value , get the range value of the number of sub-markers , and the range value As an assessment of the specificity of abnormal power supply parameters.

[0106] Step S108: determining a ratio of the number of times the cost-contributing access point is marked to the extreme value, and determining a product of the ratio and the voltage adjustment target as the governance cost of the cost-contributing access point.

[0107] Specifically, determine the current cost contribution access point Number of markings and extreme value The ratio of the current cost contribution access point Abnormal contribution, combined with the current cost contribution access point The voltage adjustment target ( ), and thus evaluate the current cost contribution access point by the following formula Cost of voltage management:

[0108] ;

[0109] Where: Represents the current agent Current cost contribution access points within the jurisdiction governance costs; and Represents the current cost contribution access point Real-time voltage and prior adaptation voltage; Indicates the current cost contribution access point The number of markings; and Represents the current cost contribution access point The maximum and minimum values ​​of the number of sub-tags of each type.

[0110] By combining the voltage adjustment target of the cost contribution access point and the ratio of the number of markings of the cost contribution access point to the extreme value of the number of sub-markings of each type of electrical parameter, the cost of voltage management of the cost contribution access point can be accurately evaluated.

[0111] Furthermore, in a possible implementation, determining the governance budget cost of the current intelligent entity in step S200 includes:

[0112] Step S201: determining the superposition value of the governance costs of all cost-contributing access points as the access point cost loss.

[0113] Specifically, for the current agent The governance costs of all cost contribution access points within the jurisdiction of the current intelligent entity are superimposed to obtain Access point cost loss . Current Agent Access point cost loss The larger the value of The higher the governance budget cost.

[0114] Step S202: determining the standard deviation of the governance costs of all cost-contributing access points as the governance cost standard deviation.

[0115] Step S203: determining the difference between the management cost of each cost-contributing access point and the cost loss of the access point as a management cost deviation value of each access point.

[0116] Specifically, since the current agent Access point cost loss Represents the current agent Abnormal voltage management expenses in the jurisdiction, so the access point cost loss Comparing the governance cost of each cost contribution access point can reflect the deviation of the current intelligent agent. The abnormal contribution of cost contribution access points within the jurisdiction. Thus, the governance cost and access point cost loss of each cost contribution access point are determined. The difference is taken as the governance cost deviation value of each access point.

[0117] Step S204: Determine the governance budget cost of the current intelligent entity based on the access point cost loss and the governance cost standard deviation, and in combination with the governance cost deviation values ​​of all cost-contributing access points.

[0118] Specifically, according to the current agent The governance cost of all cost contribution access points within the jurisdiction is relative to the current intelligent entity Access point cost loss ratio, and combined with the current agent Corresponding access point cost loss , the current agent is determined by the following formula Governance budget costs:

[0119] ;

[0120] Where: Represents the current agent governance budget costs; Represents the current agent The corresponding access point cost loss; Represents the current agent The total number of all cost-contributing access points within the jurisdiction; Represents the current agent Cost contribution access points within the jurisdiction governance costs; Represents the current agent The corresponding standard deviation of governance costs.

[0121] By calculating the access point cost loss and governance cost standard deviation, as well as the governance cost deviation value of all cost-contributing access points, the governance budget cost of the current intelligent entity can be accurately determined.

[0122] Furthermore, in a possible implementation, determining the period required for the current agent to reach optimization in step S200 includes:

[0123] Step S205: Determine a number of local governance point pairs based on the governance costs of all cost-contributing access points, where the local governance point pairs include two cost-contributing access points.

[0124] Specifically, when adjusting the voltage in a local area, the voltage increase and voltage decrease between the access points in the local area are used to complement each other, so as to achieve the active voltage environment management of the local location without relying on external power resources. Any cost-contributing access point within the jurisdiction , if the cost contribution access point The absolute value of governance costs Smaller than the current agent The average governance cost of all cost contribution access points in the jurisdiction of Without governance, select any cost contribution access point from the remaining cost contribution access points with the same governance cost sign in ascending order of absolute cost value, and record its governance cost as , and then select the governance cost with the same sign as the governance cost from the governance costs that do not form a local governance access point pair. The absolute value of | |The closest governance cost , the governance cost 、 The corresponding two cost contribution access points are recorded as a set of local governance access point pairs. In another possible implementation, the two groups of governance costs with the same governance cost sign among the remaining cost contribution access points can be used as the left endpoint and the right endpoint, respectively, and the negative correlation mapping result of the absolute value of the difference between the governance costs of each left endpoint and each right endpoint (such as using the opposite of the absolute value of the difference as the exponent of the exponential function to achieve negative correlation mapping) is used as the matching value. Based on the matching value, a matching algorithm such as the KM (Kuhn-Munkres) matching algorithm is used to perform a one-to-one matching of the left endpoint and the right endpoint, and the two cost contribution access points corresponding to the left endpoint and the right endpoint in each one-to-one matching pair are recorded as a set of local governance access point pairs.

[0125] Step S206: Determine the cumulative value of the governance costs of all local governance point pairs for all access points with different cost contributions as the total governance cost, and determine the rounded-up result of the ratio of the total governance cost to the absolute value of the governance budget cost of the current intelligent body as the period required for the current intelligent body to reach optimization.

[0126] Specifically, based on the current agent The sum of the governance costs of all cost contribution access points selected as local governance points within the jurisdiction of the current agent Absolute value of governance budget cost The rounded value of the ratio is recorded as the current agent The period required for optimization .

[0127] By determining several local governance point pairs and calculating the ratio of the cumulative value of the governance costs of all access points with different cost contributions contained in all local governance point pairs to the absolute value of the governance budget cost of the current intelligent entity, the cycle required for the current intelligent entity to reach optimization can be accurately determined.

[0128] Furthermore, in a possible implementation, determining the optimal planning period of the level where the current agent is located in step S200 includes:

[0129] Step S207: Determine the centralized target interval of the period required for optimization of all agents in the level where the current agent is located.

[0130] Specifically, considering that the optimization process requires multiple agents to complete the optimization in a similar time, the slow optimization of a single agent will affect the stability of the power environment. Therefore, based on the current agent The period required for the optimization of all agents in the same level to calculate the high heat range to obtain the current agent The optimal planning cycle of the current level. In a specific implementation, determine the current agent The 3sigma interval of the period required for optimization of all agents in the level where the 3sigma interval is located is used as the centralized target interval.

[0131] Step S208: Determine the target optimization period that is located in the centralized target interval among the optimization periods of all agents in the level where the current agent is located, and determine the average value of all target optimization periods as the optimization planning period of the level where the current agent is located.

[0132] Specifically, determine the current agent All the optimization cycles required for all agents in the same level are located in the centralized target interval, and these optimization cycles are called target optimization cycles. The average value of these target optimization cycles is calculated as the average value of the previous agent. The optimal planning cycle at the current level .

[0133] By determining the centralized target interval of the optimization period required for all agents in the current agent's level, and calculating the average value of the target optimization period within the centralized target interval, the optimization planning period for the current agent's level can be accurately determined.

[0134] Furthermore, in a possible implementation, determining the collaborative governance evaluation parameters between the current agent and other agents in step S300 includes:

[0135] Step S301: Determine the sum of the governance budget costs of the current agent and each other agent as the sum of the governance costs.

[0136] Step S302: Determine the ratio of the optimal planning period of the current agent and each other agent in the same level as the planning period ratio.

[0137] Step S303: Determine collaborative governance evaluation parameters between the current agent and other agents based on the sum of the governance costs and the ratio of planning cycles.

[0138] Specifically, in order to judge the collaborative governance evaluation between agents at the same level and different levels, so as to facilitate the collaborative governance scheduling between agents at the same level and different levels, according to the current agent The ratio of the governance cost of each other agent and the planning period is calculated by the following formula: Evaluation parameters for collaborative governance with each other agent:

[0139] ;

[0140] Where: Represents the current agent and other agents Evaluation parameters of collaborative governance among Represents the current agent governance budget costs; Representing other agents governance budget costs; Represents the current agent The optimal planning cycle of the current level; Representing other agents The optimal planning cycle at the level.

[0141] The above method determines the sum of the governance costs and the ratio of the planning cycles between the current agent and each other agent, and then determines the collaborative governance evaluation parameters between the current agent and each other agent, thereby achieving an accurate evaluation of the collaborative governance between the current agent and each other agent.

[0142] Furthermore, in a possible implementation, the governance budget cost of the current agent is adjusted in step S300 to obtain the adjusted governance budget cost, including:

[0143] Step S304: Determine the mean and variance of the collaborative governance evaluation parameters between all agents in the level where the current agent is located and other agents, and obtain the mean value and variance of the evaluation parameters.

[0144] Step S305: Determine the dynamic threshold of the evaluation parameter of the current agent based on the mean value and variance of the evaluation parameter.

[0145] Specifically, according to the current agent The average and variance of the collaborative governance evaluation parameters between all agents at the same level and other agents are determined by the following method: Dynamic threshold of evaluation parameters:

[0146] ;

[0147] Where: Represents the current agent Dynamic threshold value of evaluation parameters; Represents the mean of the evaluation parameters, that is, the current agent Evaluation parameters for collaborative governance between all agents at the same level and other agents The average value of Represents the evaluation parameter variance, that is, the current agent Evaluation parameters for collaborative governance between all agents at the same level and other agents variance; Represents the adaptive coefficient. In a specific implementation, set .

[0148] Step S306: Determine the target other intelligent agents of the current intelligent agent, and the collaborative governance evaluation parameter between the current intelligent agent and the target other intelligent agent is less than the dynamic threshold of the evaluation parameter.

[0149] Specifically, for the current agent With other agents Collaborative governance evaluation parameters between Make a judgment, when the collaborative governance evaluation parameter is less than the dynamic threshold of the evaluation parameter , that is, When , the current agent is judged With other agents The collaborative efficiency is insufficient, and the corresponding other agents Determine as the current agent Target other agents. Thus, several target other agents of the current agent can be determined.

[0150] Step S307: Determine the cumulative value of the collaborative governance evaluation parameter between the current intelligent agent and all other target intelligent agents as the cumulative collaborative governance evaluation parameter.

[0151] Step S308: Determine the cumulative value of the governance budget costs of all target entities of the current entity as the cumulative governance budget cost, and determine the governance budget cost adjustment coefficient based on the ratio of the cumulative collaborative governance evaluation parameter to the cumulative governance budget cost.

[0152] Step S309: Determine the product of the governance budget cost of the current intelligent agent and the governance budget cost adjustment coefficient as the adjusted governance budget cost of the current intelligent agent.

[0153] According to the current agent The collaborative governance evaluation parameters between all target agents and the current agent The governance budget cost of all target other agents is calculated by the following formula for the current agent Adjusted governance budget costs:

[0154] ;

[0155] Where: Represents the current agent Adjusted governance budget costs; Represents the current agent governance budget costs; Represents the current agent With the target other agents Evaluation parameters of collaborative governance between them; Represents the current agent The target of other agents governance budget costs; Represents the current agent The target of other agents The total number of .

[0156] The above method determines the dynamic threshold of the evaluation parameters of the current intelligent agent, realizes the screening of other intelligent agents corresponding to the current intelligent agent, and thus realizes punitive adjustment of the governance budget cost of the current intelligent agent based on the collaborative governance evaluation parameters between the current intelligent agent and all target other intelligent agents and the governance budget cost of all target other intelligent agents of the current intelligent agent, and finally realizes the accurate budget of the governance expenditure of the current intelligent agent.

[0157] Furthermore, in a possible implementation, determining the voltage scheduling scheme in step S400 includes:

[0158] The current voltage and adjusted governance budget cost of the current agent, as well as the collaborative governance evaluation parameters between the current agent and all target other agents are input into the neural network, and the neural network outputs the target other agents that can be scheduled among all target other agents of the current agent and the voltage values ​​that need to be scheduled.

[0159] Specifically, by pre-training a neural network, which can be a recurrent neural network (RNN), and Current voltage and adjusted governance budget costs , and the current agent Evaluation parameters for collaborative governance with other intelligent entities of each target Input into the pre-trained neural network, and the neural network outputs the current agent The target other intelligent agents that can be scheduled among all target other intelligent agents and the voltage values ​​that need to be scheduled.

[0160] Among them, in the process of training the neural network, the real-time voltage of different intelligent agents when voltage regulation is performed on different intelligent agents can be obtained based on a large amount of historical data of voltage regulation in the substation area, and the current intelligent agent can be obtained according to the above Adjusted governance budget costs , and the current agent Evaluation parameters for collaborative governance with other intelligent entities of each target In a similar manner, the adjusted governance budget costs of different agents and the collaborative governance evaluation parameters between each agent and all target agents can be determined. The target agents that can be scheduled and the voltage values ​​they need to be scheduled are then obtained as label values ​​for voltage regulation of each agent, thereby obtaining a training dataset. This training dataset is used to train a neural network, resulting in a trained neural network. Since the specific neural network training process is known in the art, it will not be detailed here.

[0161] The above method pre-trains the neural network, learns the characteristics of the voltage regulation of the current agent through the neural network, and inputs the current voltage and adjustment governance budget cost of the current agent, as well as the collaborative governance evaluation parameters between the current agent and all other target agents into the neural network, so as to finally obtain an accurate voltage scheduling plan.

[0162] Based on the same inventive concept, the embodiment of the present invention also provides a system for active voltage management in a substation based on a multi-agent self-optimization control strategy, such as Figure 3 As shown, the system includes: a memory 301, a processor 302, and a computer program code 303 stored in the memory 301 and running on the processor 302, wherein when the processor 302 executes the computer program code 303, the system can execute any of the aforementioned methods for active voltage management of substations based on a multi-agent self-optimization control strategy.

[0163] In embodiments of the present invention, the system can be divided into functional modules based on the above-described method examples. For example, these modules can correspond to individual functional modules, or two or more functions can be integrated into a single processing module. The integrated modules can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and represents only a logical functional division. In actual implementation, other division methods may be used.

[0164] Based on the same inventive concept, an embodiment of the present invention also provides a computer program product, which includes: computer program code, which, when running on a computer, enables the computer to execute any one of the aforementioned methods for active voltage management of substations based on a multi-agent self-optimizing control strategy.

[0165] Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes any one of the aforementioned methods for active voltage management of substations based on a multi-agent self-optimization control strategy.

[0166] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for active voltage management in substations based on a multi-agent self-optimization control strategy, characterized by: The method comprises: If the current intelligent body meets the conditions for starting regulation and control, the cost contribution access point and the governance cost of the cost contribution access point among the access points are determined according to the different types of electrical parameters of the access points within the jurisdiction of the current intelligent body; Based on the governance cost of the cost contribution access point, the governance budget cost of the current agent is determined. Based on the governance cost and the governance budget cost, the optimization cycle required for the current agent is determined. Then, based on the optimization cycle required for all agents in the level where the current agent is located, the optimization planning cycle of the level where the current agent is located is determined. According to the governance budget costs and optimal planning cycles of the current agent and other agents, the collaborative governance evaluation parameters between the current agent and other agents are determined, and the governance budget costs of the current agent are adjusted according to the collaborative governance evaluation parameters and the governance budget costs of other agents to obtain the adjusted governance budget costs; Determine the voltage dispatch plan based on the current intelligent body voltage, adjustment governance budget cost and collaborative governance evaluation parameters; Determine the period required for the current agent to reach optimality, including: Determine a number of local governance point pairs based on the governance costs of all cost-contributing access points, where the local governance point pairs include two cost-contributing access points. Determine the cumulative value of the governance costs of all local governance point pairs for all access points with different cost contributions as the total governance cost, and determine the rounded-up result of the ratio of the total governance cost to the absolute value of the governance budget cost of the current intelligent body as the period required for the current intelligent body to reach optimization.

2. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 1 is characterized in that: Determine the cost contribution of each access point, including: According to the different types of electrical parameters of each access point within the current intelligent agent's jurisdiction, the system state vector of each access point is constructed; Determine a state transfer matrix according to the system state vector, wherein the dimension of the state transfer matrix is ​​determined according to the type of the electrical parameter, and each element in the state transfer matrix corresponds to two types of the electrical parameter; Determining abnormal fluctuation elements in the state transfer matrix according to the distribution of elements in the state transfer matrix; Marking the type of the electrical parameter of each access point according to the abnormal fluctuation element to obtain the number of sub-marks of each type of the electrical parameter of each access point; An accumulated value of all sub-marking times of each access point is determined as the marking time of each access point, and an access point with a marking time greater than 0 is determined as a cost-contributing access point.

3. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 2 is characterized in that: Determining abnormal fluctuation elements in the state transfer matrix includes: Obtaining a histogram corresponding to all elements in the state transition matrix; Determine the height sequence corresponding to the height value of each peak column in the histogram, and determine the first-order difference sequence of the height sequence; Determine the two target peak cylinders corresponding to the maximum difference value in the first-order difference sequence, determine the average value of all elements contained in each target peak cylinder as the element mean, and determine the maximum element mean of the two element means; The peak column corresponding to the maximum element mean in the histogram and all elements contained in the peak column where the maximum element in the state transfer matrix is ​​located are determined as abnormal fluctuation elements in the state transfer matrix.

4. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 2 is characterized in that: Determine governance costs for cost contribution access points, including: Determine the difference between the real-time voltage of the cost contribution access point and the prior adaptation voltage as the voltage adjustment target; Determine the range value of the number of sub-marks of each type of electrical parameters of the cost contribution access point; The ratio of the number of times the cost contribution access point is marked to the range value is determined, and the product of the ratio and the voltage adjustment target is determined as the management cost of the cost contribution access point.

5. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 1 is characterized in that: Determine the current governance budget cost of the agent, including: Determine the superposition of the governance costs of all cost-contributing access points as the access point cost loss; Determine the standard deviation of the governance costs of all cost contribution access points as the governance cost standard deviation; Determine the difference between the governance cost of each cost-contributing access point and the cost loss of the access point as the governance cost deviation value of each access point; The governance budget cost of the current intelligent entity is determined based on the access point cost loss and the standard deviation of the governance cost, combined with the governance cost deviation values ​​of all cost-contributing access points.

6. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 1 is characterized in that: Determine the optimal planning cycle for the current agent level, including: Determine the target interval for the optimization cycle of all agents in the current agent's level; Determine the target optimization period in the centralized target interval among the optimization periods of all agents in the level where the current agent is located, and determine the average value of all target optimization periods as the optimization planning period of the level where the current agent is located.

7. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 1 is characterized in that: Determine the collaborative governance evaluation parameters between the current agent and other agents, including: Determine the sum of the governance budget costs of the current agent and each other agent as the sum of the governance costs; Determine the ratio of the optimal planning period of the current agent and each other agent in the same level as the planning period ratio; According to the sum of the governance costs and the ratio of the planning cycles, the collaborative governance evaluation parameters between the current intelligent agent and other intelligent agents are determined.

8. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 1 is characterized in that: Adjust the current agent's governance budget cost to obtain the adjusted governance budget cost, including: Determine the mean and variance of the collaborative governance evaluation parameters between all agents in the current agent's level and other agents, and obtain the mean and variance of the evaluation parameters; Determine the dynamic threshold of the evaluation parameter of the current agent based on the mean and variance of the evaluation parameter; Determine the target other intelligent agents of the current intelligent agent, and the collaborative governance evaluation parameter between the current intelligent agent and the target other intelligent agents is less than the dynamic threshold of the evaluation parameter; Determine the cumulative value of the collaborative governance evaluation parameter between the current intelligent agent and all other target intelligent agents as the cumulative collaborative governance evaluation parameter; Determine the cumulative governance budget cost of all other target agents of the current agent as the cumulative governance budget cost, and determine the governance budget cost adjustment coefficient based on the ratio of the cumulative collaborative governance evaluation parameter to the cumulative governance budget cost; Determine the product of the current agent's governance budget cost and the governance budget cost adjustment coefficient as the current agent's adjusted governance budget cost.

9. The method for active voltage management in a substation based on a multi-agent self-optimization control strategy according to claim 8 is characterized in that: Determine the voltage dispatch plan, including: The current voltage and adjusted governance budget cost of the current agent, as well as the collaborative governance evaluation parameters between the current agent and all target other agents are input into the neural network, and the neural network outputs the target other agents that can be scheduled among all target other agents of the current agent and the voltage values ​​that need to be scheduled.

Images