Adjustment method of hybrid power distribution network, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides an adjusting method of a hybrid power distribution network, electronic equipment, a storage medium and a program product. The method comprises the following steps: inputting real-time working condition data into a pre-trained minimum regulation quantity prediction model by acquiring real-time operation data such as the type, the operation condition and the real-time working condition data of a target hybrid power distribution network to obtain a corresponding prediction resource regulation quantity; and determining a resource configuration scheme according to the type of the target hybrid power distribution network, the operation condition and the predicted resource adjustment amount. The method is used for achieving the dual effects of improving the stability and the operation efficiency of the hybrid power distribution network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power systems and information technology, and particularly relates to a mixed power distribution network regulation method, an electronic device, a storage medium and a program product. BACKGROUND

[0002] In the field of power systems, with the transformation of global energy structure towards clean and low-carbon, the intelligentization and flexibility upgrading of power distribution networks have become the core direction of industry development. At present, the power distribution network technology field is facing the revolution of large-scale heterogeneous energy resources (heterogeneous resources) access. The heterogeneous resources include wind energy, photovoltaic, energy storage devices, electric vehicles and various flexible loads. The access of these heterogeneous resources will increase the operation cost and fluctuation risk of the power distribution network.

[0003] In the prior art, by setting threshold parameters (such as upper and lower limits of voltage, power fluctuation range), the charging and discharging control of the energy storage device or the simple removal of the flexible load is implemented to maintain the basic stability of the power distribution network.

[0004] However, the existing regulation means cannot realize the deep cooperation of multiple types of heterogeneous resources, cannot dynamically adapt to the complex coupling characteristics of multiple types of heterogeneous resources, cannot fully exert the potential of heterogeneous resources, and leads to the dual loss of stability and operation efficiency of the power distribution network when dealing with large-scale heterogeneous resource access. SUMMARY

[0005] The embodiments of the present application provide a mixed power distribution network regulation method, an electronic device, a storage medium and a program product to achieve the dual effects of improving the stability and operation efficiency of the mixed power distribution network.

[0006] In a first aspect, the embodiments of the present application provide a mixed power distribution network regulation method, comprising:

[0007] Obtaining real-time operation data of a target mixed power distribution network, wherein the real-time operation data at least includes the type, operation condition and real-time working condition data of the target mixed power distribution network.

[0008] Inputting the real-time working condition data into a pre-trained minimum regulation amount prediction model to obtain a predicted resource regulation amount corresponding to the real-time working condition data.

[0009] According to the type, operation condition and predicted resource regulation amount of the target mixed power distribution network, a resource configuration scheme of the target mixed power distribution network is determined.

[0010] In a possible implementation, in combination with the first aspect, the training process of the minimum regulation amount prediction model comprises:

[0011] obtain a training sample set in a preset time period, the training sample set including a plurality of sample working condition data of the hybrid power distribution network and a sample resource configuration scheme corresponding to each sample working condition data.

[0012] input the sample working condition data into the minimum adjustment amount prediction model to obtain a predicted resource adjustment amount.

[0013] calculate a loss value between the predicted resource adjustment amount and the sample resource configuration scheme according to a preset loss function.

[0014] update parameters of the minimum adjustment amount prediction model according to the loss value, and repeatedly perform the process of inputting the sample working condition data into the minimum adjustment amount prediction model to obtain the predicted resource adjustment amount based on the updated parameters until the preset loss function converges, to obtain a trained minimum adjustment amount prediction model.

[0015] In a possible implementation, in combination with the first aspect, obtaining the training sample set in the preset time period includes:

[0016] obtain sample working condition data of the hybrid power distribution network in a plurality of preset time periods.

[0017] input the sample working condition data into the pre-trained optimal response model to obtain a sample resource configuration scheme corresponding to each sample working condition data.

[0018] The sample resource configuration scheme at least includes optimized output power, node voltage, system frequency, translatable load distribution, curable load power reduction amount and transferable load power transfer amount of each resource in the hybrid power distribution network.

[0019] In a possible implementation, in combination with the first aspect, updating the parameters of the minimum adjustment amount prediction model according to the loss value includes:

[0020] calculate a gradient of the preset loss function based on the gradient descent method and the loss value.

[0021] update the parameters of the minimum adjustment amount prediction model according to the gradient of the preset loss function and the predicted resource adjustment amount.

[0022] In a possible implementation, in combination with the first aspect, after updating the parameters of the minimum adjustment amount prediction model, the method further includes:

[0023] determine the predicted resource adjustment amount output by the minimum adjustment amount prediction model when the preset loss function converges as the predicted resource adjustment amount corresponding to the sample working condition data.

[0024] In a possible implementation, in combination with the first aspect, after obtaining the trained minimum adjustment amount prediction model, the method further includes:

[0025] According to the sample working condition data, the sample resource configuration scheme and the predicted resource adjustment amount, an optimal response data set of the hybrid distribution network under different operating conditions is calculated; the optimal response data set includes optimal working condition parameters of the hybrid distribution network under different operating conditions; the optimal working condition parameters can minimize the resource adjustment cost while meeting the operating constraints.

[0026] In a possible implementation, in combination with the first aspect, according to the type, operating condition and predicted resource adjustment amount of the target hybrid distribution network, the resource configuration scheme of the target hybrid distribution network is determined, including:

[0027] According to the type, operating condition and predicted resource adjustment amount of the target hybrid distribution network, a matching is performed in the optimal response data set, to obtain a sample resource configuration scheme corresponding to the type, operating condition and predicted resource adjustment amount of the target hybrid distribution network.

[0028] The sample resource configuration scheme is determined as the resource configuration scheme of the target hybrid distribution network.

[0029] The second aspect, the embodiment of the application provides a hybrid distribution network adjustment device, including:

[0030] The acquisition module is used to acquire real-time operating data of the target hybrid distribution network, and the real-time operating data at least includes the type, operating condition and real-time working condition data of the target hybrid distribution network.

[0031] The processing module is used to input the real-time working condition data into the pre-trained minimum adjustment amount prediction model, to obtain a predicted resource adjustment amount corresponding to the real-time working condition data.

[0032] The determination module is used to determine the resource configuration scheme of the target hybrid distribution network according to the type, operating condition and predicted resource adjustment amount of the target hybrid distribution network.

[0033] In a possible implementation, in combination with the second aspect, the device further includes a training module for the training process of the minimum adjustment amount prediction model, and the training module includes:

[0034] The acquisition unit is used to acquire a training sample set in a preset time period, and the training sample set includes a plurality of sample working condition data of the hybrid distribution network and a sample resource configuration scheme corresponding to each sample working condition data.

[0035] The obtaining unit is used to input the sample working condition data into the minimum adjustment amount prediction model, to obtain a predicted resource adjustment amount.

[0036] The calculation unit is used to calculate a loss value between the predicted resource adjustment amount and the sample resource configuration scheme according to a preset loss function.

[0037] The convergence unit is configured to update parameters of the minimum adjustment amount prediction model according to the loss value, and repeatedly perform the processing of inputting the sample working condition data into the minimum adjustment amount prediction model to obtain the predicted resource adjustment amount based on the updated parameters until a preset loss function converges, so as to obtain the trained minimum adjustment amount prediction model.

[0038] In a possible implementation, in combination with the second aspect, the acquisition unit of the training module is specifically configured to:

[0039] The sample working condition data of the hybrid power distribution network in a plurality of preset time periods are acquired.

[0040] The sample working condition data are input into the pre-trained optimal response model to obtain a sample resource configuration scheme corresponding to each sample working condition data. The sample resource configuration scheme at least includes optimized output power, node voltage, system frequency, translatable load distribution, cuttable load power reduction amount and transferable load power transfer amount of each resource in the hybrid power distribution network.

[0041] In a possible implementation, in combination with the second aspect, the convergence unit of the training module is specifically configured to:

[0042] The gradient of the preset loss function is calculated based on the gradient descent method and the loss value.

[0043] The parameters of the minimum adjustment amount prediction model are updated according to the gradient of the preset loss function and the predicted resource adjustment amount.

[0044] In a possible implementation, in combination with the second aspect, the convergence unit of the training module is further configured to:

[0045] The predicted resource adjustment amount output by the minimum adjustment amount prediction model when the preset loss function converges is determined as the predicted resource adjustment amount corresponding to the sample working condition data.

[0046] In a possible implementation, in combination with the second aspect, after the convergence unit of the training module obtains the trained minimum adjustment amount prediction model, the calculation unit of the training module is further configured to:

[0047] The optimal response data set of the hybrid power distribution network under different operating conditions is calculated according to the sample working condition data, the sample resource configuration scheme and the predicted resource adjustment amount. The optimal response data set includes optimal working condition parameters of the hybrid power distribution network under different operating conditions. The optimal working condition parameters can minimize the resource adjustment cost while meeting the operating constraints.

[0048] In a possible implementation, in combination with the second aspect, the determination module is specifically configured to:

[0049] According to the type, operation condition and predicted resource adjustment amount of the target hybrid power distribution network, matching is performed in the optimal response data set to obtain a sample resource configuration scheme corresponding to the type, operation condition and predicted resource adjustment amount of the target hybrid power distribution network.

[0050] The sample resource configuration scheme is determined as the resource configuration scheme of the target hybrid power distribution network.

[0051] In a third aspect, an electronic device is provided, including a memory and a processor.

[0052] The memory stores computer execution instructions.

[0053] The processor executes the computer execution instructions stored in the memory, so that the processor executes the first aspect and / or various possible implementation manners of the first aspect.

[0054] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores computer execution instructions, which are executed by the processor to implement the first aspect and / or various possible implementation manners of the first aspect.

[0055] In a fifth aspect, a computer program product is provided, including a computer program, which is executed by the processor to implement the first aspect and / or various possible implementation manners of the first aspect.

[0056] The hybrid power distribution network adjustment method, the electronic device, the storage medium and the program product provided in the embodiments of the present application can improve the stability and operation efficiency of the hybrid power distribution network by obtaining the type, operation condition and real-time operation data of the target hybrid power distribution network, inputting the real-time operation data into the pre-trained minimum adjustment amount prediction model to obtain the corresponding predicted resource adjustment amount, and determining the resource configuration scheme according to the type, operation condition and predicted resource adjustment amount of the target hybrid power distribution network. BRIEF DESCRIPTION OF DRAWINGS

[0057] The accompanying drawings, which are incorporated into and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the application.

[0058] Figure 1 A scene schematic diagram of the hybrid power distribution network adjustment method provided in the present application;

[0059] Figure 2 A flowchart of the hybrid power distribution network adjustment method provided in the present application Figure 1 ;

[0060] Figure 3A flowchart of a regulating method of a hybrid power distribution network provided by the present application Figure 2 ;

[0061] Figure 4 A flowchart of a regulating method of a hybrid power distribution network provided by the present application Figure 3 ;

[0062] Figure 4 A flowchart of a regulating method of a hybrid power distribution network provided by the present application Figure 5 ;

[0063] Figure 6 A schematic diagram of a model training process of a minimum regulating amount prediction model of a regulating method of a hybrid power distribution network provided by the present application

[0064] Figure 7 A structural schematic diagram of a regulating device of a hybrid power distribution network provided by the present application

[0065] Figure 8 A structural schematic diagram of an electronic device provided by the present application.

[0066] The specific embodiments of the present application have been shown through the above-described drawings, and will be described in more detail hereinafter. These drawings and the written description are not intended to restrict the scope of the present application concept in any way, but to illustrate the present application concept to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0067] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to indicate the same or similar components. The embodiments described in the following exemplary embodiments are not meant to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present application as detailed in the appended claims.

[0068] First, the terms involved in the present application are explained:

[0069] Hybrid power distribution network: generally refers to a power distribution network composed of multiple different types of power lines, equipment or energy forms. In the power system, it refers to a power distribution network that contains both AC and DC lines, or integrates traditional power sources and distributed energy sources.

[0070] Power flow constraint: in the power system, it refers to the limit conditions set for the direction and size of active power and reactive power flowing in the network to ensure safe and stable operation of the system.

[0071] Gradient descent: Generally refers to an optimization algorithm for solving the minimum value of a function by iteratively updating the parameters in the opposite direction of the gradient. In power systems, it is often used in power system optimization problems to solve the optimal solution of the objective function.

[0072] Heterogeneous resources: Generally refers to a collection of resources with different properties, characteristics or types. In power systems, it refers to resources composed of different technical types, capacity scales, and operating characteristics of power sources (such as wind power, photovoltaic, thermal power) or energy storage devices.

[0073] Flexible resources: Generally refers to resources with flexible adjustment capabilities. In power systems, it refers to resources that can quickly respond to system dispatch instructions and flexibly adjust output or load, such as energy storage devices and adjustable loads.

[0074] Secondly, the application background of the embodiments of the present application is explained:

[0075] In the field of power systems, with the transformation of global energy structure towards clean and low-carbon, the intelligentization and flexibility upgrade of distribution networks have become the core direction of industry development. Currently, the field of distribution network technology is facing a revolution of large-scale heterogeneous energy resources (heterogeneous resources) access, including wind power, photovoltaic, energy storage devices, electric vehicles and various flexible loads. The access of these heterogeneous resources will increase the operating cost and fluctuation risk of the distribution network. In the existing technology, threshold parameters (such as upper and lower limits of voltage, power fluctuation range) are set to control the charging and discharging of energy storage devices, or to implement simple removal of flexible loads to maintain the basic stability of the distribution network. However, the existing control means cannot realize the deep cooperation of multiple types of heterogeneous resources, cannot dynamically adapt to the complex coupling characteristics of multiple types of heterogeneous resources, cannot fully utilize the potential of heterogeneous resources, and cannot effectively reduce the adjustment cost of the distribution network.

[0076] To solve the above problems, the inventors have studied whether historical data can be used as sample data to analyze past sample data through mathematical modeling and machine learning, and to calculate the optimal operating condition parameters of different types of hybrid distribution networks under different operating conditions. These parameters can minimize the resource adjustment cost while meeting the operating constraints, and then build an optimal response data set based on the optimal operating condition parameters, so as to match the corresponding optimal operating condition parameters according to real-time operating data in the future, thereby realizing the dual optimization effect of ensuring stable operation of the hybrid distribution network and effectively reducing the adjustment cost when the hybrid distribution network is faced with large-scale heterogeneous flexible resource access.

[0077] Taking the daily adjustment of a hybrid distribution network as an example, combined with Figure 1 , the specific application scenario of the adjustment method of the hybrid distribution network provided by the present application is explained. For example Figure 1As shown, the specific application scenarios of the present application include a power grid control system 101, a plurality of intelligent sensing devices 102, and a plurality of local controllers 103. The intelligent sensing devices 102 capture real-time power grid operation data, which is transmitted to the power grid control system 101 through a communication network. The power grid control system 101 calculates and determines adjustment parameters based on the data, and converts them into specific adjustment instructions. The instructions are issued to the local controllers 103 of each resource through a communication link, driving the local controllers 103 to perform adjustment actions on the corresponding devices. At the same time, relying on the real-time operation data fed back by the plurality of intelligent sensing devices 102, the adjustment effect is continuously monitored. It should be understood that the intelligent sensing devices 102 and the local controllers 103 can be integrated in the same device.

[0078] The technical solutions of the present application and how the technical solutions solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0079] Figure 2 Flowchart of a mixed power distribution network adjustment method provided by the present application Figure 1 As shown in the figure, the method comprises: Figure 2

[0080] S201, acquiring real-time operation data of a target mixed power distribution network, the real-time operation data at least including the type, operation condition and real-time working condition data of the target mixed power distribution network.

[0081] In this step, the type of the target mixed power distribution network can be defined by its component information, specifically embodied as the composition of wind energy, photovoltaic, energy storage, electric vehicles, flexible load and other resources contained in the target mixed power distribution network, including information such as the number of each type of resource. The operation condition can be the operation constraint corresponding to the target mixed power distribution network, as well as external input parameters that will affect the power grid system, such as environmental parameters. The real-time working condition data can be the key electrical parameters of the target mixed power distribution network, such as voltage, current, system frequency, active power and reactive power, etc., as well as the core working condition parameters of each flexible resource, such as the real-time wind speed and cut-in / cut-out wind speed threshold corresponding to the wind energy resource, the solar radiation intensity and photovoltaic module operating temperature corresponding to the photovoltaic resource, etc.

[0082] It should be understood that, in order to facilitate subsequent analysis and calculation, the operation condition and real-time working condition data proposed in the present application can have overlapping parts, for example, the operation condition can include the solar radiation intensity and photovoltaic module operating temperature corresponding to the photovoltaic resource as external inputs. The difference between the two is that the operation condition does not include internally generated result indicators, such as voltage, current, system frequency, active power and reactive power. ​

[0083] S202, input the real-time working condition data to the pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data.

[0084] In this step, the real-time working condition data is input to the pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data, which can enable the target hybrid power distribution network to achieve maximum support at a minimum adjustment cost. Among them, the maximum support means that the power grid safety and reliability at this time is optimal, that is, it can resist load fluctuation, power output mutation and other disturbances to the greatest extent, and still maintain voltage and frequency stability in extreme working conditions, guarantee continuous power supply for critical loads, and all equipment operating parameters are within the safety threshold.

[0085] S203, determining a resource configuration scheme of the target hybrid power distribution network according to the type, operating condition and predicted resource adjustment amount of the target hybrid power distribution network.

[0086] In this step, according to the type, operating condition and predicted resource adjustment amount of the target hybrid power distribution network, matching is performed in the optimal response data set to obtain a sample resource configuration scheme corresponding to the type and operating condition of the target hybrid power distribution network, and then the sample resource configuration scheme is determined as the resource configuration scheme of the target hybrid power distribution network. Among them, the optimal response data set is calculated based on the predicted resource adjustment amount and contains the optimal working condition parameters of the hybrid power distribution network under different operating conditions, which can minimize the resource adjustment cost while meeting the operating constraints. The sample resource configuration scheme obtained by matching contains the optimal working condition parameters corresponding to the current situation of the target hybrid power distribution network. Therefore, the optimal working condition parameters of the target hybrid power distribution network can be obtained by matching in the optimal response data set according to the type, operating condition and predicted resource adjustment amount of the target hybrid power distribution network.

[0087] In one possible implementation, the power grid system can accurately determine the optimal working condition parameters by real-time acquisition of real-time operating data of the target hybrid power distribution network, combination of the minimum adjustment amount prediction model and the optimal response data set, and real-time adjustment of the power grid to achieve maximum support at a minimum cost.

[0088] This application provides a method for regulating a hybrid distribution network. It acquires real-time operational data, including the type, operating conditions, and operating status of the target hybrid distribution network. This real-time data is then input into a pre-trained minimum regulation prediction model to obtain the corresponding predicted resource regulation. Based on the type, operating conditions, and predicted resource regulation of the target hybrid distribution network, a resource allocation scheme incorporating optimal operating parameters is determined. These optimal operating parameters enable the hybrid distribution network to minimize resource regulation costs while meeting operational constraints. This achieves a dual optimization effect: ensuring stable grid operation and effectively reducing regulation costs when dealing with large-scale heterogeneous flexible resource access.

[0089] Figure 3 A flowchart illustrating a regulation method for a hybrid distribution network provided in this application. Figure 2 ,like Figure 3 As shown, in this embodiment... Figure 2 Based on the examples, the training process of the minimum adjustment prediction model in this method is explained, specifically including the following steps:

[0090] S301. Obtain the training sample set within a preset time period. The training sample set includes multiple sample operating condition data of the hybrid distribution network, as well as the sample resource configuration scheme corresponding to each sample operating condition data.

[0091] In this step, sample operating condition data of the hybrid distribution network are acquired over multiple preset time periods. This sample operating condition data is then input into a pre-trained optimal response model to obtain a sample resource allocation scheme corresponding to each sample operating condition data point. The sample operating condition data and the sample resource allocation schemes are then correlated to form a training sample set. The sample resource allocation scheme includes at least the optimized output power, node voltage, system frequency, shiftable load allocation, load power reduction, and load power transfer of each resource in the hybrid distribution network. The composition of the sample operating condition data is consistent with the type of real-time operating condition data in S201, only the time periods are different, which will not be elaborated further here.

[0092] In one possible implementation, the sample operating condition data may include wind speed, wind power generation, solar radiation intensity of photovoltaic modules, photovoltaic module operating temperature, photovoltaic power generation, energy storage charging and discharging power, energy storage battery state of charge, charging and discharging power of electric vehicles, state of charge of electric vehicles, charging vehicle connection and disconnection time, shiftable loads, slashable flexible loads, and transferable flexible loads.

[0093] S302. Input the sample operating condition data into the minimum adjustment amount prediction model to obtain the predicted resource adjustment amount.

[0094] In this step, the sample working condition data is input into the current version of the minimum adjustment amount prediction model, and the minimum adjustment amount prediction model calculates and outputs the predicted resource adjustment amount based on its current parameters.

[0095] In one possible implementation, the minimum adjustment amount prediction model can be adjusted using power balance, voltage stability, frequency stability, and power flow constraints as constraints.

[0096] S303, calculate the loss value between the predicted resource adjustment amount and the sample resource configuration scheme according to the preset loss function.

[0097] S304, update the parameters of the minimum adjustment amount prediction model according to the loss value.

[0098] In this step, the gradient of the preset loss function can be calculated based on the gradient descent method and the loss value, and the parameters of the minimum adjustment amount prediction model can be updated according to the gradient of the preset loss function and the predicted resource adjustment amount, so that the loss value is reduced.

[0099] S305, determine whether the preset loss function converges. If the determination result is yes, execute S306. If the determination result is no, execute S302-S305.

[0100] In this step, it is determined whether the loss function converges. If it does not converge, the process of inputting sample working condition data into the minimum adjustment amount prediction model to obtain the predicted resource adjustment amount is repeated based on the updated parameters until the preset loss function converges, and a trained minimum adjustment amount prediction model is obtained.

[0101] It should be noted that the predicted resource adjustment amount output by the minimum adjustment amount prediction model when the preset loss function converges can be determined as the predicted resource adjustment amount corresponding to the sample working condition data.

[0102] S306, obtain the trained minimum adjustment amount prediction model.

[0103] In this step, the parameters of the trained minimum adjustment amount prediction model correspond to the optimal solution that minimizes the loss function. The trained minimum adjustment amount prediction model can be deployed in a real-time system, and the optimal predicted resource adjustment amount can be quickly obtained by inputting real-time working condition data without repeated optimization calculation.

[0104] The embodiment of the application provides a kind of mixed power distribution network regulation method, by obtaining in preset time period training sample set, the training sample set include multiple sample working condition data and its corresponding sample resource configuration scheme, sample working condition data is input to minimum regulation amount prediction model, obtain predicted resource regulation amount, according to preset loss function, calculate the loss value between predicted resource regulation amount and sample resource configuration scheme, according to loss value, update the parameters of minimum regulation amount prediction model, based on the parameter after updating, repeatedly execute the processing of sample working condition data is input to minimum regulation amount prediction model, obtain predicted resource regulation amount, until preset loss function converges, obtain the trained minimum regulation amount prediction model, to achieve for mixed power distribution network in dynamic change continuously efficient, stable operation provides solid support effect.

[0105] Figure 4 The flow chart of the mixed power distribution network regulation method provided by the application Figure 3 As Figure 4 The embodiment of the application provides a kind of mixed power distribution network regulation method, by obtaining in preset time period training sample set, the training sample set include multiple sample working condition data and its corresponding sample resource configuration scheme, sample working condition data is input to minimum regulation amount prediction model, obtain predicted resource regulation amount, according to preset loss function, calculate the loss value between predicted resource regulation amount and sample resource configuration scheme, according to loss value, update the parameters of minimum regulation amount prediction model, based on the parameter after updating, repeatedly execute the processing of sample working condition data is input to minimum regulation amount prediction model, obtain predicted resource regulation amount, until preset loss function converges, obtain the trained minimum regulation amount prediction model, to achieve for mixed power distribution network in dynamic change continuously efficient, stable operation provides solid support effect.

[0106] S401, obtain the sample working condition data of mixed power distribution network in multiple preset time periods.

[0107] S402, sample working condition data is input to pre-trained optimal response model, obtains the sample resource configuration scheme corresponding to each sample working condition data.

[0108] In this step, sample resource configuration scheme at least includes the output power of each resource optimization in mixed power distribution network, node voltage, system frequency, translatable load distribution, cuttable load power reduction amount, transferable load power transfer amount.

[0109] S403, according to multiple sample working condition data and its corresponding sample resource configuration scheme, combined with preset loss function, obtain the trained minimum regulation amount prediction model.

[0110] S404, multiple sample working condition data is input to trained minimum regulation amount prediction model, and corresponding predicted resource regulation amount is obtained.

[0111] S405, according to sample working condition data, sample resource configuration scheme and predicted resource regulation amount, calculate the optimal response data set of mixed power distribution network under different operating conditions.

[0112] In this step, the optimal response data set includes the optimal working condition parameters of mixed power distribution network under different operating conditions, and the optimal working condition parameters can minimize resource regulation cost while meeting operating constraints.

[0113] S406, acquire real-time operation data of the target hybrid power distribution network, the real-time operation data at least including a type, an operation condition and real-time working condition data of the target hybrid power distribution network.

[0114] S407, input the real-time working condition data into the pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data.

[0115] For detailed descriptions of S406-S407, refer to the related descriptions in the foregoing embodiments S201-S202, which are not repeated here.

[0116] S408, match in the optimal response data set according to the type, the operation condition and the predicted resource adjustment amount of the target hybrid power distribution network to obtain a corresponding sample resource configuration scheme.

[0117] In this step, the sample resource configuration scheme corresponding to the type, the operation condition and the predicted resource adjustment amount of the target hybrid power distribution network is obtained by matching in the optimal response data set according to the type, the operation condition and the predicted resource adjustment amount of the target hybrid power distribution network, and the sample resource configuration scheme includes corresponding optimal working condition parameters.

[0118] S409, determine the sample resource configuration scheme as a resource configuration scheme of the target hybrid power distribution network.

[0119] In this step, since the sample resource configuration scheme includes optimal working condition parameters, the optimal working condition parameters of the target hybrid power distribution network under the current operation condition can be obtained.

[0120] The hybrid power distribution network adjustment method provided by the embodiments of the present application can obtain a trained minimum adjustment amount prediction model by combining a preset loss function according to sample working condition data and corresponding sample resource configuration schemes, further obtain an optimal response data set of a hybrid power distribution network under different operation conditions, and then match a sample resource configuration scheme including corresponding optimal working condition parameters in the optimal response data set according to real-time operation data of a target hybrid power distribution network to determine the sample resource configuration scheme as a resource configuration scheme of the target hybrid power distribution network, so that the hybrid power distribution network can determine optimal working condition parameters based on current operation data when coping with large-scale access of heterogeneous flexible resources, so as to minimize resource adjustment cost while meeting operation constraints, thereby achieving the dual effects of improving stability and operation efficiency of the hybrid power distribution network.

[0121] On the basis of any one of the foregoing embodiments, the following describes the embodiments in combination with Figure 5This embodiment provides a detailed explanation of a regulation method for a hybrid distribution network through specific examples. First, in steps S501 to S506, the mathematical models related to the hybrid distribution network are analyzed and explained. These mathematical models can form the basis for the optimal response model and minimum regulation prediction model in the aforementioned embodiments and this embodiment. Based on this, the implementation process of a regulation method for a hybrid distribution network is then explained in detail.

[0122] S501 integrates and simulates the behavior and characteristics of various heterogeneous flexible resources within a distribution network under operating conditions.

[0123] In this step, the integration and simulation of various heterogeneous flexible resources within the distribution network requires the construction of corresponding models based on the characteristics of each type of flexible resource to accurately represent their behavior under operating conditions. These heterogeneous flexible resources include wind power, photovoltaic power, energy storage, electric vehicles, and flexible loads (which can be shifted, reduced, or transferred). It should be noted that heterogeneous flexible resources and flexible resources in this embodiment are the same concept; the use of "heterogeneous flexible resources" primarily highlights that these resources possess different technology types, capacity scales, and operational characteristics.

[0124] Specifically, the modeling of wind and solar power can be based on meteorological data and the parameters of the power generation equipment itself; the modeling of energy storage and electric vehicle charging systems can be based on charging and discharging efficiency and capacity; and the modeling of flexible loads can focus on their spatiotemporal flexibility, based on their characteristics of being able to be translated, reduced, and transferred.

[0125] In one possible implementation, a mathematical model of the behavioral characteristics of wind turbines corresponding to wind energy resources under operating conditions can be established as follows:

[0126]

[0127] E WT (t)=P WT (t)Δt

[0128] Among them, P WT (t) represents the average output power of the wind turbine in time period t; P WTr V is the rated output power of the wind turbine; V(t) is the average wind speed in time period t; V in V is the cut-in wind speed of the wind turbine; out V is the cut-out wind speed of the wind turbine; r E represents the rated wind speed of the wind turbine. WT (t) represents the output power of the wind turbine in time period t; Δt represents the length of the time period.

[0129] In a possible implementation, a mathematical model of the behavior characteristics of a photovoltaic cell corresponding to a photovoltaic resource under working conditions can be established as:

[0130]

[0131] E PV (t)=P PV (t)Δt

[0132] wherein P PV (t) is the actual output power of the photovoltaic cell in the t time period; P rate is the rated power of the photovoltaic cell; η Pv is the conversion efficiency of the photovoltaic cell; G(t) is the solar radiation intensity in the t time period; G STC is the solar radiation intensity under standard test conditions; G min is the minimum radiation intensity required for starting the photovoltaic cell; T c (t) is the working temperature of the photovoltaic module in the t time period; T max is the maximum allowable working temperature of the photovoltaic module; E PV (t) is the output electric quantity of the photovoltaic cell in the t time period; and Δt is the time period length.

[0133] In a possible implementation, a mathematical model of the behavior characteristics of a battery storage corresponding to a main grid storage resource under working conditions can be established as:

[0134]

[0135] wherein P B (t) is the power of the battery storage in the t time period; SOC B (t) is the state of charge of the battery storage in the t time period; SOC B,min and SOC B,max are the minimum state of charge and the maximum state of charge of the battery storage, respectively; E B (t) is the charge and discharge quantity of the battery storage; Δt is the time interval; η charge and η discharge are the charging efficiency and the discharging efficiency, respectively; and E B,total is the total energy capacity of the battery storage.

[0136] In a possible implementation, the charging and discharging process of an electric vehicle is similar to that of a conventional storage battery, and the charging and discharging power can be continuously adjusted by a converter at a charging pile. However, unlike the storage battery, the electric vehicle has a traffic attribute, and actual factors such as access and exit time, battery capacity, and user travel demand need to be considered in the process of participating in demand response. It is assumed that the power of each electric vehicle is constant during charging and discharging, and the rated power of all electric vehicles is the same, and the battery self-discharge rate and the charging and discharging energy loss are ignored.

[0137] First, the mathematical model of vehicle access state is established as follows:

[0138]

[0139] n(t) =∑ i δ EV,i (t)

[0140] Further, the mathematical model of the behavior characteristics of the ith vehicle under working conditions can be established as follows:

[0141] P EV,i (t) = δ EV,i (t)P EV,rate (t)

[0142]

[0143]

[0144] wherein δ EV,i (t) is the access state of the ith vehicle; t in,i and t out,i are the access time and exit time of the ith vehicle respectively; n(t) is the number of vehicles accessing the power grid in the period t; P EV,i (t) is the power of the ith vehicle; P EV,rate (t) is the rated power of the electric vehicle; E EV,i (t) is the charging and discharging amount of the ith vehicle; η EV,charge and η EV,discharge are the charging and discharging efficiencies; SOC EV,i (t) is the state of charge of the electric vehicle in the t period; SOC EV,min,i and SOC EV,max,i are the minimum and maximum state of charge of the ith vehicle respectively; E EV,i,total is the total energy capacity of the ith vehicle; Δt is the time interval; P EV (t) is the charging or discharging amount of the electric vehicle in the t period; E EV (t) is the charging or discharging amount of the electric vehicle in the t period.

[0145] In one possible implementation, the shiftable load can be scheduled to vary in the load power supply time, i.e., the total amount of power supply is unchanged, the power supply period is variable, and the compensation is small or even no compensation is needed by the power supply department. The load needs to be shifted as a whole, and the power consumption time spans multiple dispatching periods. The mathematical model of the behavior characteristics of the shiftable load under working conditions is as follows:

[0146]

[0147] P min ≤ PSL ≤P max

[0148]

[0149] t start ≤t op ≤t end -T op

[0150] wherein, P SL,t is the power of the load at time t; t op is the actual start time; T op is the running time; P flex is the constant power of the load during running; P min and P max are the minimum power and maximum power of the load, respectively; E SL is the total power demand of the load; t start and t end are the earliest start time and the latest end time allowed for the flexible load, respectively.

[0151] In one possible implementation, the curable load is equivalent to the discharging process of the virtual energy storage by reducing the power within a certain range, and the load that can withstand a certain interruption or reduced power and reduced running time is partially or fully reduced according to the supply and demand. The mathematical model of the behavior characteristics of the curable load under working conditions is as follows:

[0152] 0≤P RL,j (t)≤y RL,j (t)P RL,j,max (t)

[0153]

[0154] T j =[t RL,j,start ,t RL,j,end ]

[0155] 0≤n RL,j (t)≤n RL,j,max (t)

[0156] z RL,j,min ≤z RL,j ≤z RL,j,max

[0157] P RL (t)=∑ j P RL,j (t)

[0158] E RL (t)=P RL (t)Δt

[0159] wherein P RL (t) is the total power of the curable load; E RL (t) is the power of the curable load; P RL,j (t) is the power of the jth curable load at the t period; P RL,j,max (t) is the upper limit of the jth curable load at the t period; y RL,j (t) is the load adjustment state identifier, 1 indicates that the group of loads is allowed to be adjusted at the t moment, and 0 indicates that the group of loads is not allowed to be adjusted (maintaining the original power); δ RL,j (t) is the state variable of the jth curable load at the t period, 1 is scheduled, and 0 is not scheduled; T j represents the scheduling period of the jth curable load, t RL,j,start , t RL,j,end respectively represent the start and end times of the scheduling period; n RL,j (t) is the scheduling number of the jth curable load at the t period; n RL,j,max (t) is the upper limit of the scheduling number of the jth curable load at the t period; z RL,j is the single scheduling duration of the jth curable load; z RL,j,min , z RL,j,max are the upper and lower limits of the scheduling duration.

[0160] In one possible implementation, the transferable load model has a total power of the transferable load in a scheduling period, but can "transfer" the energy consumption time within a certain time range to realize time shifting of energy. The power consumption in each time period can be flexibly adjusted, but the total power of the load after the transfer is kept unchanged compared with the power before the transfer. The increased or decreased power consumption can be regarded as the charge / discharge power of virtual energy storage, and the behavior characteristic mathematical model of the transferable load under working conditions is as follows:

[0161]

[0162] x TL,i,in (t) x TL,i,out (t) = 0

[0163]

[0164] T i = [t TL,i,start , t TL,i,end ]

[0165]

[0166] P TL (t) =∑ i [P TL,i,base (t) + P TL,i,in (t) - P TL,i,out(t)]

[0167] E TL (t) TL (t)Δt

[0168] wherein, P TL (t) is the total power of the transferable load; E TL (t) is the transferred power at time t; P TL,i,out (t) is the equivalent discharging power of the i-th transferable load at time t; P TL,i,in (t) is the equivalent charging power of the i-th transferable load at time t; T i denotes the scheduling period of the i-th transferable load; t TL,i,start , t TL,i,end denote the start and end times of the scheduling period, respectively; P TL,i,out,max (t-1), P TL,i,in,max (t-1) is the maximum charging and discharging power of the i-th transferable load at time t-1; x TL,i,in (t), x TL,i,out (t) is the equivalent charging / discharging state variable, and the charging / discharging state is 1, and vice versa; P TL,i,base (t) is the basic power of the i-th transferable load at time t.

[0169] S502, obtain sample working condition data of the hybrid power distribution network in a plurality of preset time periods.

[0170] In this step, the sample working condition data includes node voltage U, line current I, active power P, reactive power Q, frequency f, loss parameter and the like of the whole network of the hybrid power distribution network, as well as sample working condition data of corresponding flexible resources. The sample working condition data of the flexible resources can include: wind speed corresponding to the wind energy resource, wind power generation power; solar radiation intensity of the photovoltaic module, working temperature of the photovoltaic module, photovoltaic power generation power corresponding to the photovoltaic resource; energy storage charging and discharging power, energy storage battery state of charge corresponding to the energy storage resource; charging and discharging power of the electric vehicle resource, state of charge of the electric vehicle, charging and disconnecting time of the electric vehicle corresponding to the electric vehicle resource; transferable load, reducible flexible load and transferable flexible load corresponding to the flexible load.

[0171] S503, establish a sample working condition data set including sample working condition data of the flexible resources and the whole network of the power distribution network.

[0172] In this step, after obtaining the sample operating condition data, it is preprocessed to create a sample operating condition dataset. Each element in the sample operating condition dataset includes at least the distribution network sample operating condition data and the corresponding flexible resource sample operating condition data. The preprocessing process includes: checking for missing and outlier values; using interpolation or regression methods to complete missing data; removing potentially erroneous data; using filtering algorithms to remove random noise; smoothing data curves; preserving main trend characteristics; normalizing values ​​of different dimensions; unifying the range; and standardizing the sampling frequency for different resources to ensure all data have the same timestamp.

[0173] S504. Calculate the equivalent model of the hybrid distribution network containing various types of flexible resources.

[0174] In this step, based on the sample operating condition dataset, equivalent parameters of each flexible resource in the hybrid distribution network are extracted through variable regression analysis or machine learning methods to determine the equivalent model of the hybrid distribution network containing various types of flexible resources.

[0175] In one possible implementation, an equivalent model P is constructed using variable regression analysis. eq The mathematical model is as follows:

[0176] P eq =α0+α1P WT +α2P PV +α3P b +α4P EV +α5P SL +α6P RL +α7P TL +e

[0177] It can be represented in matrix form as follows:

[0178] Y = Xα + e

[0179] in: m represents the number of historical operating condition data; n represents the quantity of each type of flexible resource; α is the equivalent parameter. e represents the residual term of the regression model; the other parameters are the same as those described above and will not be repeated here.

[0180] Solve for α = (X) by minimizing the sum of squared errors. T X) -1 X T Y, the solution is... Substitute the solved value into the equivalent model P eq Then, equivalent models of various types of flexible resource hybrid distribution networks can be obtained.

[0181] S505, analyze the overall behavior and characteristics of the hybrid distribution network.

[0182] In this step, the overall behavior and characteristics of the hybrid distribution network can be analyzed using the following parameters: total power generation P gen and total load power P load ; node voltage V m and voltage deviation ΔV; system frequency f and system frequency deviation Δf, and can also include the purchase cost C grid , energy storage loss C storage , scheduling cost, etc. C dispatch . Among them, if the total cost of the distribution network needs to be considered, the economy of the hybrid distribution network is measured. In addition to the working condition parameters of the distribution network, the purchase cost C grid , energy storage loss C storage , scheduling cost, etc. C dispatch .

[0183] In one possible implementation, the mathematical model of the purchase cost is:

[0184] C grid = N WT C WT + N PV C PV + N b C b

[0185] Where C WT , C PV , C b are the unit prices of wind turbines, photovoltaic panels, and batteries, respectively; N WT , N PV , N b are the number of wind turbines, photovoltaic panels, and batteries, respectively.

[0186] The mathematical model of the battery energy storage loss is:

[0187]

[0188] C storage = C B,loss + C EV,loss

[0189] Where E B,loss (t) is the amount of power lost during the charging and discharging of the battery; C B,price is the electricity price of the battery; C B,loss is the loss cost of the battery energy storage; E EV,loss,i (t) is the amount of power lost during the charging and discharging of the electric vehicle; C EV,price is the electricity price of the electric vehicle; C EV,lossThe loss cost of the energy storage of the electric vehicle; the rest of the parameters are consistent with the foregoing, which will not be repeated here.

[0190] The mathematical model of the scheduling cost is:

[0191] C SL =C SL,price P SL

[0192]

[0193]

[0194] C dispatch =C SL +C RL,j +C TL,i

[0195] Wherein, C SL is the total scheduling cost of the shiftable load; C SL,price is the incentive price of the shiftable load; C RL,price,j is the incentive price of the jth curtaileble load; C RL,capacity,j is the capacity cost of the jth curtaileble load; C RL,j is the total scheduling cost of the jth curtaileble load; C TL,price,i is the incentive price of the ith transferable load; C TL,i is the total incentive cost of the ith transferable load. The rest of the parameters are consistent with the foregoing, which will not be repeated here.

[0196] S506, determining the optimal configuration criterion of the flexible resource.

[0197] In this step, the determination of the optimal configuration criterion of the flexible resource can be the determination of the limiting condition of the optimal configuration of the flexible resource.

[0198] In one possible implementation, the power balance, voltage stability, frequency stability and power flow constraint are used to determine whether the flexible resource configuration is optimal.

[0199] Wherein, the mathematical model of the power balance is:

[0200]

[0201] In the formula, when ΔP≈0, the system power is balanced.

[0202] The mathematical model of the voltage stability is:

[0203]

[0204] In the formula, S V is the voltage stability. When S V≈0, the voltage is stable.

[0205] The mathematical model of frequency stability is:

[0206]

[0207] In the formula, S f is the frequency stability. When S f ≈0, the voltage is stable.

[0208] The mathematical model of power flow constraint is:

[0209]

[0210] In the formula, P m and Q m are the active power and reactive power of node m respectively; U m and U n are the voltages between nodes m and n respectively; N is the total number of nodes; G mn and B mn are the conductance and susceptance of the line between node m and node n; and θ mn is the phase angle of the voltage between node m and node n.

[0211] The above remaining parameters are consistent with the foregoing, and will not be described here again.

[0212] S507, the sample resource configuration scheme of the multi-type flexible resource is corresponded to the sample working condition data through the optimal response model.

[0213] In this step, the optimal response model is established, and the sample resource configuration scheme of the multi-type flexible resource is corresponded to the sample working condition data based on the optimization algorithm of the neural network. The input layer of the optimal response model based on the neural network includes the sample working condition data of each type of flexible resource; and the output layer is the sample resource configuration scheme of the multi-type flexible resource, including: the optimized output power of each resource, node voltage, system frequency, translatable load distribution, reducible load power reduction amount, transferable load power transfer amount, and the total number of all target configuration parameters. The optimization target is to maximize the performance of the power distribution network.

[0214] In one possible implementation, the input layer includes the operation data of each type of flexible resource, including wind energy: wind speed V(t), wind power P WT (t), solar radiation intensity G(t), working temperature T c (t), photovoltaic power P PV (t), energy storage charge and discharge power P B (t), energy storage battery state of charge SOC B , automobile charge and discharge power P EV(t), state of charge of the car SOC EV , access and disconnection time t in and t out , flexible load P SL , P RL and P TL . The input layer size is n, representing the total number of all flexible resource adjustable parameters.

[0215] The model hidden layer is used to capture the complex nonlinear relationship between the input parameters, and the expression capacity of the model is improved through the nonlinear activation function. The hidden layer is designed as follows: the hidden layer 1 extracts the preliminary features between the input parameters, sets the activation function RE LU, and the number of neurons is 2n; the hidden layer 2 models the coupling relationship between the flexible resources, sets the activation function RE LU, and the number of neurons is n; the hidden layer 3 further extracts high-order features and maps to the output space, sets the activation function Sigmoid, and the number of neurons is n / 2.

[0216] The output layer is the sample resource configuration scheme of multiple types of flexible resources, including: the output power of each resource after optimization, node voltage, system frequency, translatable load allocation, reducible load power reduction amount, and transferable load power transfer amount. The output layer size is m, representing the total number of all target configuration parameters. The optimization target is to maximize the performance of the power distribution network, and the loss function of the comprehensive target is defined as:

[0217]

[0218] wherein, is the overall optimization target of the model, and the model is balanced in multiple aspects by minimizing the value; σ1 and σ2 are weight coefficients; is the mean square error of the node voltage; is the mean square error of the system frequency.

[0219] S508, a set of training samples of multiple types of flexible resources is established.

[0220] In this step, a set of training samples of multiple hybrid power distribution networks containing wind energy, photovoltaic, energy storage, charging cars, flexible loads and the like under different operating conditions is formed. By inputting the sample working condition data into the pre-trained optimal response model, the sample resource configuration scheme corresponding to each sample working condition data can be obtained. Each element in the training sample set at least includes the above-mentioned sample resource configuration scheme and the corresponding sample working condition data.

[0221] S509, a minimum adjustment amount prediction model of power adjustment of multiple types of flexible resources under the maximum support state of the power grid is established.

[0222] In this step, it is determined that the power grid is in the maximum support state, and the calculation parameters of the minimum adjustment amount prediction model are optimized in combination with the training sample set to adjust the minimum adjustment amount. Gradient descent method or evolutionary algorithm can be used to iteratively optimize the minimum adjustment amount prediction model, including: initializing parameters, setting initial adjustment power P adjust,init ; the optimization goal is to obtain the minimum flexible resource adjustment amount P adjust,i , and the total number of flexible resources is N; setting constraint conditions such as power balance, voltage stability, frequency stability, and flexible resource characteristics; making the loss function converge, and when the maximum number of iterations is reached, the optimal adjustment power of each flexible resource is the minimum.

[0223] In one possible implementation, the steps of establishing a minimum adjustment amount prediction model of power adjustable multi-type flexible resource under the maximum support state of the power grid are described by taking Figure 6 as an example. It should be understood that Figure 6 the method for calculating the minimum adjustment amount of power adjustable multi-type flexible resource shown in the above is only an example, and the minimum adjustment amount of power adjustable multi-type flexible resource can also be calculated in other ways.

[0224] Step 1, set the objective function and initial parameters.

[0225] Using gradient descent method for iterative optimization, first initialize parameters, set initial adjustment power P adjust,init . The optimization goal is to obtain the minimum flexible resource adjustment amount P adjust,i , also known as optimal adjustment power, at this time, the optimal adjustment power set of all flexible resources is where P is the minimum adjustment power converged after the iteration of the i-th flexible resource is completed.

[0226] Step 2, set the constraint conditions.

[0227] The following constraint conditions are in turn power balance, voltage stability, frequency stability, energy storage constraint, electric vehicle constraint, and flexible load constraint:

[0228] P gen +P adjust =P load

[0229] V min ≤V i ≤V max

[0230] |Δf|≤Δf max

[0231]

[0232] The third step is to converge the loss function, making the optimal adjustment power close to the optimal value.

[0233] Calculate the gradient of the loss function using gradient descent:

[0234]

[0235] Update the adjustment power, where η is the learning rate:

[0236]

[0237] The fourth step is to determine whether the current parameters are the optimal solution.

[0238] If the current loss function Once the convergence threshold is reached, the current solution is considered the optimal solution. The maximum number of iterations has been reached.

[0239] Fifth step: Save the current optimal solution.

[0240] At this point, the optimal adjustment power P for each flexible resource adjust,i satisfy:

[0241] Step 6: Repeat continuously within the preset time period. Figure 6 The process shown can be used to obtain a trained minimum adjustment prediction model.

[0242] S510. Input multiple sample operating condition data into the trained minimum adjustment amount prediction model to obtain the corresponding predicted resource adjustment amount.

[0243] S511. Obtain the optimal response dataset for different hybrid distribution networks.

[0244] In this step, based on sample operating condition data, sample resource allocation schemes, and predicted resource adjustment amounts, the optimal response dataset of the hybrid distribution network under different operating conditions is calculated. The optimal response dataset includes the optimal operating condition parameters of the hybrid distribution network under different operating conditions. The optimal operating condition parameters enable the hybrid distribution network to minimize resource adjustment costs while meeting operating constraints.

[0245] In one possible implementation, the optimal response dataset includes corresponding hybrid distribution network types, operating conditions, sample operating condition data, sample resource allocation schemes, and predicted resource adjustment amounts, wherein the sample resource allocation schemes include optimal operating condition parameters.

[0246] It should be noted that, in Figure 5 The processing steps S501 to S511 shown in the embodiments do not constitute a specific limitation on a method for regulating a hybrid distribution network. In other embodiments of this application, a method for regulating a hybrid distribution network may include more thanFigure 5 The embodiments can include more or fewer steps, for example, a regulating method of a hybrid power distribution network can include Figure 5 some steps in the embodiments, or, Figure 5 some steps in the embodiments can be replaced by steps with the same functions, or, Figure 5 some steps in the embodiments can be split into multiple steps, etc.

[0247] Figure 7 A structural schematic diagram of a regulating device of a hybrid power distribution network provided in the present application is shown in FIG. 7. Figure 7 As shown in FIG. 7, the regulating device 70 of a hybrid power distribution network provided in the present embodiment includes:

[0248] The acquisition module 701 is configured to acquire real-time operation data of a target hybrid power distribution network, wherein the real-time operation data at least includes the type, operation condition and real-time working condition data of the target hybrid power distribution network.

[0249] The processing module 702 is configured to input the real-time working condition data into a pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data.

[0250] The determination module 703 is configured to determine a resource configuration scheme of the target hybrid power distribution network according to the type, operation condition and predicted resource adjustment amount of the target hybrid power distribution network.

[0251] In a possible implementation, the device further includes a training module 704 configured to perform a training process of the minimum adjustment amount prediction model, and the training module 704 includes:

[0252] The acquisition unit is configured to acquire a training sample set in a preset time period, wherein the training sample set includes a plurality of sample working condition data of a hybrid power distribution network and a sample resource configuration scheme corresponding to each sample working condition data.

[0253] The obtaining unit is configured to input the sample working condition data into the minimum adjustment amount prediction model to obtain a predicted resource adjustment amount.

[0254] The calculation unit is configured to calculate a loss value between the predicted resource adjustment amount and the sample resource configuration scheme according to a preset loss function.

[0255] The convergence unit is configured to update parameters of the minimum adjustment amount prediction model according to the loss value, and repeatedly perform the process of inputting the sample working condition data into the minimum adjustment amount prediction model to obtain the predicted resource adjustment amount based on the updated parameters until the preset loss function converges, thereby obtaining the trained minimum adjustment amount prediction model.

[0256] In a possible implementation, the acquisition unit of the training module 704 is specifically configured to:

[0257] Obtain sample working condition data of the hybrid power distribution network in a plurality of preset time periods.

[0258] Input the sample working condition data into the pre-trained optimal response model to obtain a sample resource configuration scheme corresponding to each sample working condition data. The sample resource configuration scheme at least includes optimized output power, node voltage, system frequency, translatable load allocation, curable load power reduction amount, and transferable load power transfer amount of each resource in the hybrid power distribution network.

[0259] In a possible implementation, the convergence unit of the training module 704 is specifically configured to:

[0260] Based on the gradient descent method and the loss value, calculate the gradient of the preset loss function.

[0261] According to the gradient of the preset loss function and the predicted resource adjustment amount, update the parameters of the minimum adjustment amount prediction model.

[0262] In a possible implementation, the convergence unit of the training module 704 is further configured to:

[0263] When the preset loss function converges, the predicted resource adjustment amount output by the minimum adjustment amount prediction model is determined as the predicted resource adjustment amount corresponding to the sample working condition data.

[0264] In a possible implementation, after the convergence unit of the training module 704 obtains the trained minimum adjustment amount prediction model, the calculation unit of the training module 704 is further configured to:

[0265] According to the sample working condition data, the sample resource configuration scheme, and the predicted resource adjustment amount, calculate an optimal response data set of the hybrid power distribution network under different operating conditions; the optimal response data set includes optimal working condition parameters of the hybrid power distribution network under different operating conditions; the optimal working condition parameters can minimize the resource adjustment cost while meeting the operating constraints.

[0266] In a possible implementation, the determination module 703 is specifically configured to:

[0267] According to the type, operating condition, and predicted resource adjustment amount of the target hybrid power distribution network, match in the optimal response data set to obtain a sample resource configuration scheme corresponding to the type, operating condition, and predicted resource adjustment amount of the target hybrid power distribution network.

[0268] Determine the sample resource configuration scheme as the resource configuration scheme of the target hybrid power distribution network.

[0269] The adjustment device for the hybrid power distribution network provided in this embodiment can execute the method provided in the method embodiments, and has similar implementation principles and technical effects. Therefore, no further description is given here.

[0270] Figure 8 A structural schematic diagram of an electronic device is provided in the present application. As shown in the figure, the electronic device 80 provided in the present embodiment comprises at least one processor 801 and a memory 802. Optionally, the device 80 further comprises a communication component 803. Wherein, the processor 801, the memory 802 and the communication component 803 are connected through a bus 804. Figure 8

[0271] In the process of implementation, the at least one processor 801 executes the computer execution instructions stored in the memory 802, so that the at least one processor 801 executes the above-mentioned method.

[0272] The specific implementation process of the processor 801 can refer to the above-mentioned method embodiment, which has similar implementation principles and technical effects, and will not be described here in detail.

[0273] In the above-mentioned embodiments, it should be understood that the processor can be a central processing unit (English: Central Processing Unit, for short: CPU), and can also be other general-purpose processors, digital signal processors (English: Digital Signal Processor, for short: DSP), application specific integrated circuits (English: Application Specific Integrated Circuit, for short: ASIC) and the like. The general-purpose processor can be a microprocessor, or the processor can also be any conventional processor, etc. The steps of the method disclosed in combination with the application can be directly embodied as the execution of the hardware processor, or executed by the combination of hardware and software modules in the processor.

[0274] The memory can contain a random access memory (RAM), and can also include a non-volatile memory (Non-volatile Memory, NVM), for example, at least one disk memory.

[0275] The bus can be an industry standard architecture (Industry Standard Architecture, ISA) bus, a peripheral component (Peripheral Component, PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, the bus in the drawings of the present application does not limit only one bus or one type of bus.

[0276] ​The application further provides a computer program product comprising a computer program which, when executed by a processor, implements the method described above.

[0277] The application further provides a computer readable storage medium, wherein computer execution instructions are stored in the computer readable storage medium, and when a processor executes the computer execution instructions, the method described above is implemented.

[0278] The readable storage medium described above can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0279] An exemplary readable storage medium is coupled to the processor, so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in the device.

[0280] The division of units is only a logical function division, and in actual implementation, there can be another division mode, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0281] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on a plurality of network units. According to actual needs, some or all of the units can be selected to achieve the purpose of the embodiment.

[0282] In addition, the functional units in each embodiment of the application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0283] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the embodiments of the method of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0284] It can be understood by those skilled in the art that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware. The aforementioned program can be stored in a computer readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, and various media that can store program codes.

[0285] Finally, it should be noted that: those skilled in the art will easily think of other embodiments of the present application after considering the specification and practicing the application disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application that follow the general principles of the present application and include common knowledge or conventional technical means in the art that are not disclosed in the present application, and is not limited to the precise structure described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present application is only limited by the appended claims.

Claims

1. A method of regulating a hybrid distribution network, characterized in that, The method comprises: obtaining real-time operation data of a target hybrid power distribution network, wherein the real-time operation data at least includes a type, an operation condition and real-time working condition data of the target hybrid power distribution network; inputting the real-time working condition data into a pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data; determining a resource configuration scheme of the target hybrid power distribution network according to the type, the operation condition and the predicted resource adjustment amount of the target hybrid power distribution network.

2. The method of claim 1, wherein, The training process of the minimum adjustment amount prediction model comprises: obtaining a training sample set in a preset time period, wherein the training sample set comprises a plurality of sample working condition data of a hybrid power distribution network and a sample resource configuration scheme corresponding to each sample working condition data; inputting the sample working condition data into a minimum adjustment amount prediction model to obtain a predicted resource adjustment amount; calculating a loss value between the predicted resource adjustment amount and the sample resource configuration scheme according to a preset loss function; updating parameters of the minimum adjustment amount prediction model according to the loss value, and repeatedly performing the process of inputting the sample working condition data into the minimum adjustment amount prediction model to obtain a predicted resource adjustment amount based on the updated parameters until the preset loss function converges, thereby obtaining a trained minimum adjustment amount prediction model.

3. The method of claim 2, wherein, The obtaining of the training sample set in the preset time period comprises: obtaining the sample working condition data of the hybrid power distribution network in a plurality of preset time periods; inputting the sample working condition data into a pre-trained optimal response model to obtain a sample resource configuration scheme corresponding to each sample working condition data; wherein the sample resource configuration scheme at least includes an optimized output power, a node voltage, a system frequency, a translatable load distribution, a cuttable load power reduction amount and a transferable load power transfer amount of each resource in the hybrid power distribution network.

4. The method of claim 2, wherein, The updating of the parameters of the minimum adjustment amount prediction model according to the loss value comprises: calculating a gradient of the preset loss function based on a gradient descent method and the loss value; updating the parameters of the minimum adjustment amount prediction model according to the gradient of the preset loss function and the predicted resource adjustment amount.

5. The method of claim 4, wherein, After updating the parameters of the minimum adjustment amount prediction model, the method further comprises: determining the predicted resource adjustment amount output by the minimum adjustment amount prediction model when the preset loss function converges as the predicted resource adjustment amount corresponding to the sample working condition data.

6. The method of claim 2, wherein, After obtaining the trained minimum adjustment amount prediction model, the method further comprises: calculating an optimal response data set of the hybrid power distribution network under different operation conditions according to the sample working condition data, the sample resource configuration scheme and the predicted resource adjustment amount; the optimal response data set comprises optimal working condition parameters of the hybrid power distribution network under different operation conditions; the optimal working condition parameters can minimize resource adjustment cost while meeting operation constraints.

7. The method of claim 6, wherein, The determination of the resource configuration scheme of the target hybrid power distribution network according to the type, the operation condition and the predicted resource adjustment amount of the target hybrid power distribution network comprises: According to the type of the target hybrid power distribution network, the operation condition, and the predicted resource adjustment amount, matching is performed in the optimal response dataset to obtain a sample resource configuration scheme corresponding to the type of the target hybrid power distribution network, the operation condition, and the predicted resource adjustment amount; The sample resource configuration scheme is determined as the resource configuration scheme of the target hybrid power distribution network.

8. A regulating device for a hybrid distribution network, characterized in that Comprise: An acquisition module is configured to acquire real-time operation data of a target hybrid power distribution network, wherein the real-time operation data at least includes a type of the target hybrid power distribution network, an operation condition, and real-time working condition data; A processing module is configured to input the real-time working condition data into a pre-trained minimum adjustment amount prediction model to obtain a predicted resource adjustment amount corresponding to the real-time working condition data; A determination module is configured to determine a resource configuration scheme of the target hybrid power distribution network according to the type of the target hybrid power distribution network, the operation condition, and the predicted resource adjustment amount.

9. An electronic device, comprising: Comprise: A memory and a processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory, so that the processor executes the method in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to implement the method in any one of claims 1-7.

11. A computer program product, characterised in that, A computer program is executed by the processor to implement the method in any one of claims 1-7.

Images