Method for generating operation scene of network construction type SVG (Static Var Generator) accessed power system based on diffusion model

Abstract

The invention discloses a diffusion model-based method for generating an operation scene of a network-constructing SVG (Static Var Generator) accessed to a power system, and relates to the technical field of power systems. The method comprises the steps that a power system operation scene data set covering reactive power output and power flow information of a network building type static var generator (SVG) is built by changing a reactive power reference value of the network building type SVG, and the operation scene data set comprises a power flow steady-state solution set and a condition label set; defining a loss function including a power flow balance loss item and a power conservation loss item, and constructing a conditional diffusion model with the minimum loss function as a target; training a conditional diffusion model based on the power flow steady-state solution set, the conditional label set and a loss function; and inputting the target condition label demand and the conditional scene generation quantity into the trained condition diffusion model to generate a power flow operation scene. According to the method, diversified high-quality stable generation of the power flow scene of the power system can be realized, and the flexibility and controllability of scene generation are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power systems, in particular to a method for generating a power system operation scenario with grid-connected SVG based on a diffusion model. BACKGROUND

[0002] The current global energy structure is accelerating the transition to new energy as the main body, and the installed capacity of renewable energy such as wind power and photovoltaic power continues to rise. High penetration of renewable energy leads to a lack of system inertia and weak voltage support, further exacerbating the risk of system voltage instability. In this context, grid-connected static var generator (SVG) has become a core equipment to support high penetration of new energy grid connection. Grid-connected SVG is a power electronic converter that can quickly adjust reactive power, and provides voltage support, reactive power support and damping support to the system through grid-connected control, thereby enhancing the active support capability of the power grid. Compared with the traditional follow-the-grid SVG, the grid-connected SVG has the advantages of faster response speed, higher regulation accuracy and wider dynamic compensation range.

[0003] In order to plan and design the grid-connected SVG access power system and evaluate the risk, it is necessary to generate diversified high-quality power flow operation scenarios. The existing statistical scenario generation methods based on Monte Carlo and hyper-Latin cube sampling need to assume the probability distribution assumption of the scenario data. However, the power system operation scenario under actual multi-working conditions presents complex nonlinear characteristics, and the scenario obtained based on a specific probability distribution cannot reflect some details of the actual scenario. Therefore, scenario generation models based on deep learning are gradually applied to the field of scenario generation. However, most of the existing scenario generation methods focus on the generation of new energy output and load scenarios, and lack of conditional generation methods for grid-connected SVG access power system operation scenarios. The existing power system power flow scenario method has deficiencies in physical consistency and condition controllability, and the generated scenario may not satisfy the physical law of power conservation and it is difficult to control the trend of the generated scenario. SUMMARY

[0004] The present application provides a method for generating a power system operation scenario with grid-connected SVG based on a diffusion model, which can realize diversified and high-quality stable generation of power system power flow scenarios and improve the flexibility and controllability of scenario generation.

[0005] The present application is implemented through the following technical solutions.

[0006] In a first aspect, the present application provides a method for generating a power system operation scenario with grid-connected SVG based on a diffusion model, which comprises:

[0007] A power system operation scenario dataset covering the reactive power output of the grid-forming SVG and the power flow information is constructed by changing the reactive power reference value of the grid-forming SVG, wherein the operation scenario dataset includes a power flow steady-state solution set and a conditional label set;

[0008] A loss function including a power flow balance loss term and a power conservation loss term is defined, and a conditional diffusion model with the minimum loss function as the target is constructed;

[0009] The conditional diffusion model is trained based on the power flow steady-state solution set and the conditional label set, and the loss function;

[0010] The target conditional label requirement and the number of conditional scenarios are input into the trained conditional diffusion model to generate a power flow operation scenario.

[0011] In some embodiments, the power flow steady-state solution set is represented as:

[0012] ;

[0013] Wherein, represents the bus voltage amplitude; represents the bus voltage phase angle; represents the generator injected active power; represents the generator injected reactive power; represents the active power absorbed by the load; represents the reactive power absorbed by the load; represents the grid bus number; represents the grid bus set; represents the reactive power output of the grid-forming SVG; represents the active power loss of the network; represents the reactive power loss of the network.

[0014] In some embodiments, the conditional label set is represented as:

[0015] ;

[0016] Wherein, represents the reactive power output reference of the grid-forming SVG; represents the small signal stability margin of the grid-forming SVG; represents the transient voltage stability margin of the grid-forming SVG connected system.

[0017] In some embodiments, the loss function is represented as:

[0018] ;

[0019] in, This represents the noise matching loss in the training objective of the diffusion model; This represents a real data sample free of noise; Represents a set of conditional tags; Indicates a time step; Indicates noise; Indicates Gaussian noise; This represents the predicted noise value by the conditional denoising network. Indicates at time step System sample data; This represents the power flow balance loss term; This represents the power conservation loss term; The weight hyperparameters representing the power flow balance loss term; This represents the weighting hyperparameter of the power conservation loss term.

[0020] In some embodiments, the power flow balance loss term is represented as:

[0021] ;

[0022] in, Indicates busbar Active power loss term:

[0023] ;

[0024] in, This indicates the active power injected into the generator; This represents the active power absorbed by the load. Indicates and The busbar numbers adjacent to the busbar; Indicates the power grid bus number; Represents the set of power grid buses; Indicates busbar The voltage amplitude; The conductance elements represent the nodal admittance matrix; Represents the susceptance element of the nodal admittance matrix; Represents a node With nodes The voltage phase angle difference between them;

[0025] Indicates busbar Reactive power loss item:

[0026] ;

[0027] in, This indicates the reactive power injected into the generator; This represents the reactive power output of a grid-type SVG; This represents the reactive power absorbed by the load.

[0028] The power conservation loss term is expressed as:

[0029] ;

[0030] in, This represents the reactive power loss of the network.

[0031] In some embodiments, the conditional diffusion model is trained based on the power flow steady-state solution set, the conditional label set, and the loss function, including:

[0032] The steady-state solution set and condition label set are used as training samples for the conditional diffusion model, and the loss function is used as the training objective function. The conditional diffusion model is trained by forward noise addition and backward noise reduction.

[0033] Secondly, the present invention provides a device for generating operation scenarios of a grid-type SVG access power system based on a diffusion model, characterized in that the device comprises:

[0034] The data acquisition module is used to: construct a power system operation scenario dataset covering the reactive power output and power flow information of the grid-type static var generator (SVG) by changing the reactive power reference value of the SVG, wherein the operation scenario dataset includes a power flow steady-state solution set and a condition label set;

[0035] The loss function definition module is used to: define a loss function including a power flow balance loss term and a power conservation loss term, and construct a conditional diffusion model with the goal of minimizing the loss function;

[0036] The diffusion model training module is used to train the conditional diffusion model based on the power flow steady-state solution set, the conditional label set, and the loss function.

[0037] The power flow operation scenario generation module is used to input the target condition label requirements and the number of conditional scenario generation into the trained conditional diffusion model to generate power flow operation scenarios.

[0038] Thirdly, the present invention provides a device for generating operation scenarios of a grid-based SVG access power system based on a diffusion model, characterized in that the device comprises:

[0039] At least one processor;

[0040] At least one memory is coupled to the at least one processor and stores instructions for execution by the at least one processor, which, when executed by the at least one processor, implement the method described above.

[0041] Fourthly, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores instructions that, when executed by a computer, cause the computer to perform the methods described above.

[0042] Fifthly, the present invention provides a computer program product, characterized in that the computer program product includes instructions, which, when executed by a computer, cause the computer to perform the methods described above.

[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0044] (1) By introducing the power flow balance loss term and the power conservation loss term, the power flow scenario of the grid-type SVG access power system generated by the model can simultaneously satisfy the node power conservation and system power conservation, thereby ensuring the physical consistency between the generated scenario and the real scenario and improving the availability of scenario data in subsequent applications.

[0045] (2) In the conditional diffusion model, reactive power output reference and stability margin index are added as conditional labels, so that the scenario generation model can quickly generate the corresponding operation scenario according to the given network-type SVG scenario generation target requirements. At this time, the physical meaning of generating power flow scenario is extended from the single probability distribution consistency to the power system demand consistency. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0047] Figure 1 This is a flowchart of a method for generating operation scenarios of a grid-type SVG access power system based on a diffusion model, according to an embodiment of the present invention.

[0048] Figure 2 This is a schematic diagram of the generation framework for a grid-type SVG access power system operation scenario based on a diffusion model according to an embodiment of the present invention.

[0049] Figure 3This is a structural block diagram of a grid-type SVG access power system operation scenario generation device based on a diffusion model according to an embodiment of the present invention.

[0050] Figure 4 This is a schematic diagram of a grid-type SVG access power system operation scenario generation device based on a diffusion model according to an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0052] On the one hand, the present invention provides a method for generating operation scenarios of grid-type SVG access power systems based on a diffusion model. Figure 1 This is a flowchart illustrating a method for generating operation scenarios of a grid-connected SVG power system based on a diffusion model, according to an embodiment of the present invention. (Reference) Figure 1 The method for generating operation scenarios of grid-type SVG access power systems based on diffusion models includes: S10 to S40.

[0053] Figure 2 This is a schematic diagram of a generation framework for a network-based SVG access power system operation scenario based on a diffusion model, according to an embodiment of the present invention. The following references... Figure 1 and Figure 2 This paper provides a detailed description of the method for generating operation scenarios of grid-type SVG access power systems based on a diffusion model, as described in this invention.

[0054] In S10, by changing the reactive power reference value of the grid-type SVG, a power system operation scenario dataset covering the reactive power output and power flow information of the grid-type SVG is constructed. The operation scenario dataset includes a power flow steady-state solution set and a condition label set.

[0055] The operational scenario dataset includes a power flow steady-state solution set and a condition label set. To enable the diffusion model to learn the physical relationship between control conditions and system power flow distribution, high-quality training samples need to be constructed first. Each training sample needs to reflect the steady-state power flow of the system under different grid-type SVG reactive power output reference settings, including node voltage and branch power information of the power system. Therefore, in this invention, the power flow steady-state solution set of the grid-type SVG connected power system can be expressed as:

[0056]

[0057] in, Indicates the amplitude of the bus voltage; Indicates the phase angle of the bus voltage; This indicates the active power injected into the generator; This indicates the reactive power injected into the generator; This represents the active power absorbed by the load. This represents the reactive power absorbed by the load. Indicates the power grid bus number; This represents the set of power grid buses, including grid-type SVG access buses; This represents the reactive power output of a grid-type SVG; This indicates the active power loss of the network; This represents the reactive power loss of the network.

[0058] In practical applications, it is necessary to simulate power system flow scenarios under different reactive power output reference settings and stability margin requirements. Therefore, to improve the flexibility and controllability of the generated operating scenarios, when constructing the sample set, the reactive power reference of the grid-type SVG output, the small-disturbance stability margin, and the system transient voltage stability margin are selected as condition labels. The condition label set can be represented as:

[0059]

[0060] in, Indicates the reactive power output reference for a network-type SVG; This indicates the small disturbance stability margin of a meshed SVG; This indicates the transient voltage stability margin of a network-type SVG access system.

[0061] Power flow steady-state solution set of grid-connected SVG power systems and conditional tag sets Together, they constitute the power flow mapping space for conditional diffusion model learning. When generating dataset samples, different values ​​of reactive power output references for grid-type SVG can be selected. By calling the PyPower power flow toolkit in Python, the dataset can be generated in batches, and the small-disturbance stability margin of the grid-type SVG and the transient voltage stability margin of the system can be calculated.

[0062] The core of the scene generation model based on conditional diffusion is learning the probabilistic data distribution. This refers to the power flow distribution under specific grid-type SVG conditions. If variables such as node voltage and power flow are missing from the samples, the model will be unable to establish a mapping relationship between the grid-type SVG and the system power flow state, thus failing to generate a reasonable power flow scenario for grid-type SVG access to the power system under conditional control. Therefore, in S10, it is necessary to collect the power flow steady-state solution set and condition label set; otherwise, subsequent physical constraint optimization cannot be achieved.

[0063] In S20, a loss function is defined, including power flow balance loss terms and power conservation loss terms, and a conditional diffusion model is constructed with the goal of minimizing the loss function. The power flow scenarios of the grid-type SVG access power system output by the conditional diffusion model need to conform to the physical laws of the power system. To ensure the rationality of the generated power flow scenarios, the conditional diffusion model is improved, and the model learning direction is adjusted to guide the model to generate diverse scenarios that are consistent with the physical reality of the scenario.

[0064] On the one hand, a power flow balance loss term is added to the conditional diffusion model. Its expression is:

[0065]

[0066] in, Indicates busbar The active power loss term; Indicates busbar The reactive power loss term.

[0067] The calculation formula is as follows:

[0068]

[0069] in, Indicates and The busbar numbers adjacent to the busbar; Indicates busbar The voltage amplitude; The conductance elements represent the nodal admittance matrix; Represents the susceptance element of the nodal admittance matrix; Represents a node With nodes The voltage phase angle difference between them.

[0070] The calculation method is as follows:

[0071]

[0072] By introducing power flow balance constraints during the training phase, the scenario generation model not only fits the power flow sample distribution but also learns the physical laws of power flow, ensuring the physical consistency of the generated scenario results for grid-type SVG access power systems.

[0073] On the other hand, a power conservation loss term is added to the conditional diffusion model. , The expression is:

[0074]

[0075] The power conservation constraint, as a global weak constraint, ensures the conservation of total power in the grid-connected SVG power system, thus avoiding the generation of unreasonable scenario data with unbalanced energy across the entire power system by the model.

[0076] Subsequently, the total training loss of the conditional diffusion model used for generating power flow scenarios in grid-connected SVG-based power systems can be obtained. , Represented as:

[0077]

[0078] in, This represents the noise matching loss in the training objective of the diffusion model; This represents a real data sample free of noise; The weight hyperparameters representing the power flow balance loss term; The weighting hyperparameters represent the power conservation loss term; Indicates Gaussian noise; This represents the predicted noise value by the conditional denoising network.

[0079] In S30, the conditional diffusion model is trained based on the power flow steady-state solution set, the conditional label set, and the loss function. Specifically, the power flow steady-state solution set is used as the training sample for the conditional diffusion model, the loss function is used as the training objective function, and forward noise addition and backward noise reduction are employed to train the conditional diffusion model.

[0080] The learning process of the conditional diffusion model includes forward denoising and backward denoising. Backward denoising is the conditional scene generation process. During model training, multi-step denoising enables the diffusion model to learn the recovery rules of the data distribution of the grid-connected SVG power system during the backward process, thereby generating high-quality expected power flow scenarios that conform to the distribution under any conditions. During training, the total training loss of the conditional diffusion model for generating power flow scenarios of the grid-connected SVG power system, constructed using S20, is... The power flow steady-state solution set constructed using S10 is used as the training sample for the conditional diffusion model.

[0081] The forward noise addition in the conditional diffusion model is at time step This is achieved by gradually adding Gaussian noise. The forward noise addition process can be viewed as a Markov process, as follows:

[0082]

[0083] in, This represents the conditional probability distribution after adding noise; Indicates at time step System sample data; At time step System sample data; Indicates at time step The diffusion scaling factor; This represents a normal distribution.

[0084] The denoising process of the conditional diffusion model is the process of gradually restoring the original data sample data from data with sampling noise. The inverse denoising process can be represented as:

[0085]

[0086] in, This represents the conditional probability distribution after denoising. The variance coefficient represents the noise level. This represents the mean value determined by the denoising network.

[0087] In S40, by inputting the target condition label requirements and the number of conditional scenario generation quantities into the trained conditional diffusion model, the power flow operation scenario can be generated.

[0088] Figure 2 This is a schematic diagram of the generation framework for a grid-type SVG access power system operation scenario based on a diffusion model according to an embodiment of the present invention. Figure 2 This paper presents the overall architecture for generating operation scenarios of grid-connected SVG power systems based on a diffusion model, including four core processes: data input, loss function design, model training, and scenario generation. Through these processes, a scenario generation model for grid-connected SVG power systems based on conditional diffusion can be trained. Thanks to the multi-step learning process of the conditional diffusion model, this model exhibits good convergence and training stability during scenario generation.

[0089] After the model is trained, the expected grid-type SVG can be integrated into the target condition variables of the power system according to the application requirements of power system planning and risk assessment. By combining the conditional scenario generation quantity input model, a specified number of diverse current flow scenarios that meet the conditional application requirements can be generated. This represents the reactive power condition input for the expected network-type SVG; The input represents the expected small disturbance stability margin condition for the networked SVG; This represents the expected transient voltage stability margin condition input for the grid-connected SVG system. Ultimately, the output of the grid-connected SVG power system operation scenario includes the bus voltage amplitude, bus voltage phase angle, generator injected reactive power, generator injected active power, network reactive power loss, network active power loss, and reactive power output by the grid-connected SVG under conditional control.

[0090] In this invention, the following technical effects can be achieved: (1) By introducing power flow balance loss term and power conservation loss term, the power flow scenario of the grid-type SVG access to the power system generated by the model can simultaneously satisfy the node power conservation and system power conservation, thereby ensuring the physical consistency between the generated scenario and the real scenario and improving the availability of scenario data in subsequent applications; (2) In the conditional diffusion model, reactive power output reference and stability margin index are added as conditional labels, so that the scenario generation model can quickly generate the corresponding operating scenario according to the given grid-type SVG scenario generation target requirements. At this time, the physical meaning of generating the power flow scenario is extended from the single probability distribution consistency to the power system demand consistency.

[0091] On the other hand, the present invention provides a device for generating operation scenarios of a grid-type SVG access power system based on a diffusion model. Figure 3 This is a structural block diagram of a grid-type SVG access power system operation scenario generation device based on a diffusion model according to an embodiment of the present invention. (Reference) Figure 3 The device for generating operation scenarios of a grid-type SVG access power system based on a diffusion model includes: a data acquisition module, a loss function definition module, a diffusion model training module, and a power flow operation scenario generation module.

[0092] The data acquisition module is used to: construct a power system operation scenario dataset covering the reactive power output and power flow information of the grid-type static var generator (SVG) by changing the reactive power reference value of the SVG. The operation scenario dataset includes a power flow steady-state solution set and a condition label set.

[0093] The loss function definition module is used to: define a loss function that includes a power flow balance loss term and a power conservation loss term, and construct a conditional diffusion model with the goal of minimizing the loss function.

[0094] The diffusion model training module is used to train the conditional diffusion model based on the power flow steady-state solution set, the conditional label set, and the loss function.

[0095] The power flow operation scenario generation module is used to input the target condition label requirements and the number of conditional scenario generation into the trained conditional diffusion model to generate power flow operation scenarios.

[0096] In implementing the functions of the integrated modules described above in hardware, this embodiment of the invention provides a structure for the network-based SVG access power system operation scenario generation device based on the diffusion model involved in the above embodiments. Figure 4 This is a schematic diagram of a grid-type SVG access power system operation scenario generation device based on a diffusion model according to an embodiment of the present invention. (Reference) Figure 4The device for generating operation scenarios of a grid-based SVG-connected power system based on a diffusion model includes: at least one processor; and at least one memory. The at least one memory is coupled to the at least one processor and stores instructions for execution by the at least one processor, which, when executed by the at least one processor, implement the methods described above.

[0097] A processor can be a set of logic blocks, modules, and circuits that implement or execute the various exemplary logic blocks, modules, and circuits described in connection with embodiments of the present invention. The processor can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in connection with embodiments of the present invention. A processor can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, etc.

[0098] The memory may be read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, electrically erasable programmable read-only memory (EEPROM), disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.

[0099] In one implementation, the memory can exist independently of the processor. The memory can be connected to the processor via a bus and used to store instructions or program code. When the processor calls and executes the instructions or program code stored in the memory, it can implement the methods provided in the embodiments of the present invention. In another implementation, the memory can also be integrated with the processor.

[0100] On the other hand, the present invention also provides a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium) storing computer program instructions that, when executed on a computer, cause the computer to perform the method as described in any of the above embodiments.

[0101] Exemplary examples show that the aforementioned computer-readable storage media may include, but are not limited to: magnetic storage devices (e.g., hard disks, floppy disks, or magnetic tapes), optical discs (e.g., compact disks (CDs), digital versatile disks (DVDs), etc.), smart cards, and flash memory devices (e.g., erasable programmable read-only memory (EPROMs), cards, sticks, or key drives, etc.). The various computer-readable storage media described in this invention may represent one or more devices and / or other machine-readable storage media for storing information. The term "machine-readable storage medium" may include, but is not limited to, wireless channels and various other media capable of storing, containing, and / or carrying instructions and / or data.

[0102] This invention provides a computer program that, when run on a computer, causes the computer to perform the method of any of the above embodiments.

[0103] This invention provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the method of any of the above embodiments.

[0104] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for generating operation scenarios of a grid-connected SVG power system based on a diffusion model, characterized in that, The method includes: By changing the reactive power reference value of the grid-type static var generator (SVG), a power system operation scenario dataset covering the reactive power output and power flow information of the grid-type SVG is constructed. The operation scenario dataset includes a power flow steady-state solution set and a condition label set. Define a loss function that includes a power flow balance loss term and a power conservation loss term, and construct a conditional diffusion model with the goal of minimizing the loss function; The conditional diffusion model is trained based on the power flow steady-state solution set, the conditional label set, and the loss function. Input the target condition label requirements and the number of conditional scenario generation quantities into the trained conditional diffusion model to generate power flow operation scenarios.

2. The method according to claim 1, characterized in that, The steady-state solution set of the power flow is represented as follows: ; in, Indicates the amplitude of the bus voltage; Indicates the phase angle of the bus voltage; This indicates the active power injected into the generator; This indicates the reactive power injected into the generator; This represents the active power absorbed by the load. This represents the reactive power absorbed by the load. Indicates the power grid bus number; Represents the set of power grid buses; This represents the reactive power output of a grid-type SVG; This indicates the active power loss of the network; This represents the reactive power loss of the network.

3. The method according to claim 1, characterized in that, The conditional label set is represented as follows: ; in, Indicates the reactive power output reference for a network-type SVG; This indicates the small disturbance stability margin of a meshed SVG; This indicates the transient voltage stability margin of a network-type SVG access system.

4. The method according to any one of claims 1 to 3, characterized in that, The loss function is expressed as: ; in, This represents the noise matching loss in the training objective of the diffusion model; This represents a real data sample free of noise; Represents a set of conditional tags; Indicates a time step; Indicates noise; Indicates Gaussian noise; This represents the predicted noise value by the conditional denoising network. Indicates at time step System sample data; This represents the power flow balance loss term; This represents the power conservation loss term; The weight hyperparameters representing the power flow balance loss term; This represents the weighting hyperparameter of the power conservation loss term.

5. The method according to claim 4, characterized in that, The power flow balance loss term is expressed as: ; in, Indicates busbar Active power loss term: ; in, This indicates the active power injected into the generator; This represents the active power absorbed by the load. Indicates and The busbar numbers adjacent to the busbar; Indicates the power grid bus number; Represents the set of power grid buses; Indicates busbar The voltage amplitude; The conductance elements represent the nodal admittance matrix; Represents the susceptance element of the nodal admittance matrix; Represents a node With nodes The voltage phase angle difference between them; Indicates busbar Reactive power loss item: ; in, This indicates the reactive power injected into the generator; This represents the reactive power output of a grid-type SVG; This represents the reactive power absorbed by the load. The power conservation loss term is expressed as: ; in, This represents the reactive power loss of the network.

6. The method according to claim 1, characterized in that, Based on the power flow steady-state solution set, the conditional label set, and the loss function, the conditional diffusion model is trained, including: The steady-state solution set and condition label set are used as training samples for the conditional diffusion model, and the loss function is used as the training objective function. The conditional diffusion model is trained by forward noise addition and backward noise reduction.

7. A device for generating operation scenarios of a grid-type SVG access power system based on a diffusion model, characterized in that, The device includes: The data acquisition module is used to: construct a power system operation scenario dataset covering the reactive power output and power flow information of the grid-type static var generator (SVG) by changing the reactive power reference value of the SVG, wherein the operation scenario dataset includes a power flow steady-state solution set and a condition label set; The loss function definition module is used to: define a loss function including a power flow balance loss term and a power conservation loss term, and construct a conditional diffusion model with the goal of minimizing the loss function; The diffusion model training module is used to train the conditional diffusion model based on the power flow steady-state solution set, the conditional label set, and the loss function. The power flow operation scenario generation module is used to input the target condition label requirements and the number of conditional scenario generation into the trained conditional diffusion model to generate power flow operation scenarios.

8. A device for generating operation scenarios of a grid-type SVG access power system based on a diffusion model, characterized in that, The device includes: At least one processor; At least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor, the instructions implementing the method of any one of claims 1 to 6 when executed by the at least one processor.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 6.

Images