Related methods and apparatuses for power distribution network digital platform, and device, medium and product

Abstract

Disclosed in the embodiments of the present application are related methods and apparatuses for a power distribution network digital platform, and a device, a medium and a product. Among the related methods, a training method for a simulation model of a power distribution network comprises: acquiring real-time data or historical data of a power distribution network collected by sensors, instruments and devices; performing pre-processing operations on the real-time data or historical data of the power distribution network to obtain a training sample set, wherein the pre-processing operations comprise data cleaning, denoising and format conversion; on the basis of using actual data during an actual operation process of the power distribution network as labeling information, and on the basis of training samples, training a simulation model to be trained, so as to obtain a trained simulation model; on the basis of different device components in the power distribution network having different response speeds when a system operation state changes, determining a multi-rate parallel simulation technique for the power distribution network; and using the multi-rate parallel simulation technique for the power distribution network to adjust the simulation step size of the trained simulation model, so as to obtain a final optimized model.

Description

Related methods, devices, equipment, media and products of distribution network digital platform Technical Field

[0001] The present application relates to the field of distribution network operation analysis, and in particular to methods, devices, equipment, media and products related to a distribution network digital platform. Background Art

[0002] The distribution network is a key link connecting the transmission network and the load terminal. Against the background of large-scale access of distributed resources such as distributed power sources and electric vehicles, the distribution network has gradually developed into a hub platform for the access and coordinated operation of power electronic power sources / equipment / loads such as distributed new energy, charging stations, and electric vehicles, and has the dual functions of "clean substitution" on the power supply side and "electric energy substitution" on the load side.

[0003] The access of distributed resources at numerous points, in large quantities, and over a wide area has had a serious negative impact on the safe and reliable operation of the distribution network. At the equipment level, the quality of distributed photovoltaic inverters varies greatly. After long-term operation, technical indicators decline, affecting the quality of exported power. The majority of photovoltaic power stations in townships and rural areas are in power supply mode, leading to prominent problems with reverse overload of transformers in these stations and significantly shortening transformer service life. Distributed photovoltaic power generation causes changes in the direction of power flow in the distribution network, and existing protection devices in the distribution network may malfunction or fail to operate. At the system level, the output of distributed power sources is highly uncertain. When superimposed on the load demand curve, this causes the net power of the system to exhibit a "duck" curve, and the system voltage fluctuation is prominent. The malfunction and failure of protection devices seriously affect the safe operation of the system and reduce the reliability of the system power supply. Problems such as reverse power flow and three-phase imbalance lead to increased line losses in the system, resulting in poor economic operation.

[0004] To fully address these issues, physical or digital simulation is often required to analyze the distribution network, deriving detailed strategies for different modes. These strategies can then be used as a basis for planning, transformation, or operational regulation. However, physical simulations are difficult to scale and operate, so digital simulation is often used to analyze and resolve these issues.

[0005] Distribution network digital simulation mainly includes offline simulation, online simulation, and digital-analog hybrid simulation. The main problems are as follows:

[0006] (1) Model construction: There are many non-deterministic factors affecting the distribution network under distributed resource access. The dynamic matching ability between the mechanism model and the data model of various types of equipment and complex network structures is relatively weak, making it difficult to achieve accurate mapping between the digital model and the physical entity when the operating status of the distribution network changes.

[0007] (2) Digital simulation method: Since the transient response speeds of different components in the distribution network are different, the existing digital simulation serial simulation method cannot select different step sizes according to the component type, and it is difficult to meet the speed requirements of digital simulation of active distribution networks.

[0008] (3) Digital simulation platform: In scenarios with a high proportion of distributed power sources, the measured data of the distribution network are diverse and complex. The surge in grid operation data seriously affects the simulation efficiency. Existing digital simulations are difficult to meet the requirements of high performance and high concurrency of distribution network simulations, and cannot meet the requirements of fast virtual-real interaction between physical entities and digital models.

[0009] Summary of the Invention

[0010] In view of this, the embodiments of the present application at least provide methods, devices, equipment, media and products related to a distribution network digital platform.

[0011] The technical solution of the embodiment of the present application is implemented as follows:

[0012] In a first aspect, an embodiment of the present application provides a method for training a simulation model of a distribution network, the method comprising: obtaining real-time data or historical data of the distribution network collected by sensors, instruments, and equipment; performing preprocessing operations on the real-time data or historical data of the distribution network to obtain a training sample set; the preprocessing operations include data cleaning, denoising, and format conversion; training the simulation model to be trained based on actual data during the actual operation of the distribution network as the annotation information and the training sample set to obtain a trained simulation model; determining a multi-rate parallel simulation technology for the distribution network based on the different response speeds of different equipment components in the distribution network to changes in the system operating state; using the multi-rate parallel simulation technology for the distribution network, adjusting the simulation step size of the trained simulation model to obtain a final optimized model.

[0013] In an embodiment of the present application, a training sample set is obtained by preprocessing the real-time data or historical data of the distribution network collected by the acquired sensors, instruments, and equipment; based on the actual data in the actual operation process of the distribution network as the annotation information and the training sample set, the simulation model to be trained is trained to obtain a trained simulation model; based on the different response speeds of different equipment components in the distribution network to changes in the system operation state, a multi-rate parallel simulation technology for the distribution network is determined; using the multi-rate parallel simulation technology of the distribution network, the simulation step size of the trained simulation model is adjusted to obtain a final optimized model. In this way, based on the actual data in the actual operation process of the distribution network as the annotation information and the training sample set, the simulation model to be trained is trained; this method of training the model can continuously update the model, improve the accuracy of the digital simulation of the distribution network, and can better adapt to changes in the training sample set; in order to realize parallel calculation of the model simulation, the multi-rate parallel simulation technology of the distribution network is used to deal with the situation where different components in the model have different response speeds to changes in the system operation state, so that the components with slow response speed can update the status of the components with fast response speed in real time, thereby improving the speed and accuracy of the digital simulation of large-scale active distribution networks.

[0014] In a second aspect, an embodiment of the present application provides a method for constructing a digital platform for a distribution network, the method comprising: constructing a multi-resolution physical space model of typical equipment in the distribution network based on typical equipment in different simulation scenarios in the distribution network; constructing a data space model of typical equipment in the distribution network based on typical equipment in the distribution network whose model is to be constructed; the typical equipment in the distribution network includes lines, loads, converters, switchgears and distributed power sources; constructing a knowledge space model of typical equipment in the distribution network based on a knowledge graph of knowledge related to the model to be modeled in the distribution network; constructing a data space model of typical equipment in the distribution network based on the multi-resolution physical space model, the data space model, and the knowledge space model. A virtual space model of typical equipment in the distribution network is constructed; wherein the virtual space model includes attribute information of typical equipment in the distribution network; by constructing the multi-resolution physical space model, the data space model, the knowledge space model, and an index architecture of each typical equipment in the virtual space model, calls are made between different simulation models in the multi-dimensional digital simulation model of the distribution network; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space modeling, the knowledge space model, and the virtual space model; based on the multi-dimensional digital simulation model of the distribution network, a distribution network digital platform based on a cloud platform is constructed.

[0015] In the embodiment of the present application, a multi-resolution physical space model, a data space model, a knowledge space model and a virtual space model are constructed; by constructing an index architecture of each typical device in the multi-resolution physical space model, the data space model, the knowledge space model and the virtual space model, the calls and mappings between different simulation models in the multi-dimensional digital simulation model of the distribution network are performed, wherein the multi-dimensional digital simulation model of the distribution network includes a multi-resolution physical space model, a data space model, a knowledge space model and a virtual space model; based on the multi-dimensional digital simulation model of the distribution network, an overall architecture of the distribution network digital platform based on the cloud platform is constructed, and the overall architecture of the distribution network digital platform includes a data layer, a model layer, a service layer, an application layer, a display layer and a platform control. In this way, the establishment of the multi-dimensional digital simulation model of the distribution network realizes the dynamic matching of the distribution network entity and the digital model; the data layer, model layer, service layer and application layer of the overall architecture of the distribution network digital platform provide technical means and analysis basis for the verification and testing of the control strategies of various scenarios of the distribution network; the display layer allows users to intuitively understand the operating status and various indicators of the distribution network; the platform control provides auxiliary decision support for the safe, reliable and economical operation of the distribution network.

[0016] In a third aspect, an embodiment of the present application provides a processing method for a distribution network digital platform, the method comprising: processing the real-time data or historical data of the distribution network through the data layer; the processing method comprises data storage, data sharing, data analysis, and data processing; based on the simulation model of the model layer, taking the data of the data layer as input, simulating at the service layer to obtain the simulated result; wherein the simulated result includes data under different operating scenarios in the distribution network; based on the simulated result, applying the simulated result through the application layer; the application includes resource management, analysis and evaluation, business control, and decision-making and command; displaying the simulated result and the data processed by the application layer through the display layer; realizing data sharing between the various modules of the distribution network digital platform through the API interface; protecting the system stability and security of the platform through platform control, and the platform control includes system permissions, system configuration, and security management.

[0017] In the embodiments of this application, a cloud-based digital distribution network platform architecture is constructed based on a multi-dimensional digital simulation model of the distribution network. The overall architecture comprises a data layer, a model layer, a service layer, an application layer, a presentation layer, and a platform management and control layer. The data, model, service, and application layers of the overall architecture provide technical means and analytical basis for verifying and testing distribution network control strategies in various scenarios. The presentation layer allows users to intuitively understand the operating status and various indicators of the distribution network. The platform management and control provides auxiliary decision-making support for the safe, reliable, and economical operation of the distribution network.

[0018] In a fourth aspect, an embodiment of the present application provides a training device for a simulation model of a distribution network, the device comprising: a first acquisition module for acquiring real-time data or historical data of the distribution network collected by sensors, instruments, and equipment; a preprocessing module for performing preprocessing operations on the real-time data or historical data of the distribution network to obtain a training sample set; the preprocessing operations include data cleaning, denoising, and format conversion; a training module for training the simulation model to be trained based on actual data during the actual operation of the distribution network as the annotation information and the training sample set to obtain a trained simulation model; a determination module for determining a multi-rate parallel simulation technology for the distribution network based on the different response speeds of different equipment components in the distribution network to changes in the system operating state; an adjustment module for adjusting the simulation step size of the trained simulation model using the multi-rate parallel simulation technology of the distribution network to obtain a final optimized model.

[0019] In a fifth aspect, an embodiment of the present application provides a device for constructing a distribution network digital platform, the device comprising: a first construction module for constructing a multi-resolution physical space model of typical equipment in the distribution network based on typical equipment in different simulation scenarios in the distribution network; a second construction module for constructing a data space model of typical equipment in the distribution network based on typical equipment in the distribution network whose model is to be constructed; the typical equipment in the distribution network includes lines, loads, converters, switchgear and distributed power sources; a third construction module for constructing a knowledge space model of typical equipment in the distribution network based on a knowledge graph of knowledge related to the model to be modeled in the distribution network; a fourth construction module for constructing a knowledge space model of typical equipment in the distribution network based on the multi-resolution physical space model and the data space model. and the knowledge space model, constructing a virtual space model of typical equipment in the distribution network; wherein the virtual space model includes attribute information of typical equipment in the distribution network; a calling module, used to call different simulation models in the multi-dimensional digital simulation model of the distribution network by constructing the multi-resolution physical space model, the data space model, the knowledge space model, and the index architecture of each typical equipment in the virtual space model; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space modeling, the knowledge space model, and the virtual space model; a building module, used to build a distribution network digital platform based on a cloud platform based on the multi-dimensional digital simulation model of the distribution network.

[0020] In a sixth aspect, an embodiment of the present application provides a processing device for a distribution network digital platform, the device comprising: a processing module for processing real-time data or historical data of the distribution network through the data layer; the processing method includes data storage, data sharing, data analysis, and data processing; a simulation module for using the simulation model of the model layer, taking the data of the data layer as input, to perform simulation at the service layer to obtain the simulated result; wherein the simulated result includes data under different operating scenarios in the distribution network; an application module for applying the simulated result through the application layer based on the simulated result; the application includes resource management, analysis and evaluation, business management, and decision-making and command; a display module for displaying the simulated result and the data processed by the application layer through the display layer; a data sharing module for realizing data sharing between the various modules of the distribution network digital platform through an API interface; a protection module for protecting the system stability and security of the platform through platform management, and the platform management includes system permissions, system configuration, and security management.

[0021] In a seventh aspect, an embodiment of the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the program, it implements some or all of the steps in the above method.

[0022] In an eighth aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, some or all of the steps in the above method are implemented.

[0023] In a ninth aspect, an embodiment of the present application provides a computer program comprising a computer-readable code. When the computer-readable code is executed in a computer device, a processor in the computer device executes some or all of the steps for implementing the above method.

[0024] In a tenth aspect, an embodiment of the present application provides a computer program product, comprising a computer program or instructions, which, when executed by a processor, implement the steps in the above method.

[0025] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. Those skilled in the art can also derive other drawings based on these drawings without inventive work, among which:

[0027] FIG1 is a schematic diagram of an implementation flow of a training method for a simulation model of a distribution network provided in an embodiment of the present application;

[0028] FIG2 is a schematic diagram of an implementation flow of a method for constructing a distribution network digital platform provided in an embodiment of the present application;

[0029] FIG3 is a schematic diagram of an implementation flow of a processing method for a distribution network digital platform provided in an embodiment of the present application;

[0030] FIG4 is a schematic diagram of a process for constructing a multi-dimensional digital simulation model of a distribution network according to an embodiment of the present application;

[0031] FIG5 is a schematic diagram of a process for constructing a multi-resolution physical space model in a distribution network according to an embodiment of the present application;

[0032] FIG6 is a schematic diagram of a process for constructing a data space model in a distribution network according to an embodiment of the present application;

[0033] FIG7 is a schematic diagram of an example of a neural network based on an extreme learning machine (ELM) according to an embodiment of the present application;

[0034] FIG8 is a schematic diagram of a process for constructing a knowledge space model in a distribution network according to an embodiment of the present application;

[0035] FIG9 is a knowledge graph of transformer faults in a distribution network provided by an embodiment of the present application;

[0036] FIG10 is a schematic diagram of a process for constructing a virtual space model in a distribution network according to an embodiment of the present application;

[0037] FIG11 is an index architecture of a digital simulation model of a distribution network in an embodiment of the present application;

[0038] FIG12 is a logic diagram of data interaction between the distribution network digital simulation system and other systems provided in an embodiment of the present application;

[0039] FIG13 is a multi-rate parallel simulation data interaction mode provided in an embodiment of the present application;

[0040] Figure 14 is a diagram showing the overall architecture of the cloud-based distribution network digital simulation platform provided in this embodiment of the application.

[0041] FIG15A is a schematic diagram of the structure of a training device for a simulation model of a power distribution network provided in an embodiment of the present application;

[0042] FIG15B is a schematic diagram of the composition structure of a device for constructing a distribution network digital platform provided in an embodiment of the present application;

[0043] FIG15C is a schematic diagram of the composition structure of a processing device of a distribution network digital platform provided in an embodiment of the present application;

[0044] FIG16 is a schematic diagram of a hardware entity of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions of this application are further elaborated in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0046] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0047] It should be pointed out that the terms "first\second\third" involved in the embodiments of the present application are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.

[0048] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art in the art to which the embodiments of the present application belong. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0049] Distribution network: It is composed of overhead lines, cables, poles, distribution transformers, disconnectors, reactive power compensators and some ancillary facilities, and plays an important role in distributing electric energy in the power grid.

[0050] System order: It is the partial derivative order of the highest-order partial derivative that appears in the system, belonging to the partial differential equation system.

[0051] Admittance matrix: It is a matrix composed of the conductance values ​​(or resistance and admittance values) between each node in the power grid.

[0052] Network topology refers to the physical layout of interconnected devices using transmission media. It refers to the specific physical (real) or logical (virtual) arrangement of network components. If two networks have the same connection structure, we say they have the same network topology, even though their internal physical wiring and inter-node distances may differ.

[0053] China's voltage levels: safety voltage (usually below 36V), low voltage (divided into 220V and 380V), high voltage (10KV-220KV), extra high voltage (330KV-750KV), and ultra-high voltage (1000KV AC, ±800KV DC and above).

[0054] Short-circuit capacity: also known as maximum short-circuit current, maximum short-circuit capability, etc., refers to the maximum current value that various electrical equipment in the power system (such as cables, transformers, circuit breakers, etc.) can withstand when a short-circuit fault occurs.

[0055] A transient process is the process a circuit undergoes as it transitions from one stable state to another. This change in stable state is typically achieved by switching the circuit on or off. The nature of transients is also determined by parameters such as resistance, capacitance, and inductance in the circuit. The changes in voltage and current are non-periodic.

[0056] Simulation step: It is the logical time length of the simulation system calculation advancement. This time is actually the sampling time of the simulation model for discretizing the continuous variables in the objective world. The smaller the simulation step, the greater the sampling density, and the more guaranteed the realism of the model.

[0057] Power network security zoning: Based on the characteristics of the power secondary system, it is divided into the production control zone and the management information zone. The production control zone is divided into the control zone (Safety Zone I) and the non-control zone (Safety Zone II). The management information zone is divided into the production management zone (Safety Zone III) and the management information zone (Safety Zone IV). Different security zones have different security protection requirements, with Safety Zone I having the highest security level, followed by Safety Zone II, and so on.

[0058] Power flow calculations are the calculation of active power, reactive power, and voltage distribution within a power grid, given the power system topology, component parameters, and generation and load parameters. Power flow calculations determine the steady-state operating parameters of each component of the power system, based on the given grid structure and parameters and the operating conditions of components such as generators and loads.

[0059] Distribution network digital simulation mainly includes offline simulation, online simulation, and digital-analog hybrid simulation. The main problems are as follows:

[0060] (1) Model construction: There are many non-deterministic factors affecting the distribution network under distributed resource access. The dynamic matching ability between the mechanism model and the data model of various types of equipment and complex network structures is relatively weak, making it difficult to achieve accurate mapping between the digital model and the physical entity when the operating status of the distribution network changes.

[0061] (2) Digital simulation method: Since the transient response speeds of different components in the distribution network are different, the existing digital simulation serial simulation method cannot select different step sizes according to the component type, and it is difficult to meet the speed requirements of digital simulation of active distribution networks.

[0062] (3) Digital simulation platform: In scenarios with a high proportion of distributed power sources, the measured data of the distribution network are diverse and complex. The surge in grid operation data seriously affects the simulation efficiency. Existing digital simulations are difficult to meet the requirements of high performance and high concurrency of distribution network simulations, and cannot meet the requirements of fast virtual-real interaction between physical entities and digital models.

[0063] In view of the above reasons, this application proposes a digital simulation method and system for distribution network access for distributed resource access, establishes a multi-dimensional space digital simulation model of the distribution network, proposes a simulation driving mode of distribution network digital data-model dual-driven and a multi-rate parallel online simulation method, and designs a high-performance and high-concurrency digital simulation platform for distribution network to provide support for the intelligent and efficient interaction between the physical distribution network and the digital distribution network under distributed power access.

[0064] The present application provides a method for digital simulation of a distribution network. As shown in FIG1 , the method may include steps S110 to S150:

[0065] Step S110: Acquire real-time data or historical data of the distribution network collected by sensors, instruments, and equipment;

[0066] Here, the real-time data or historical data of the distribution network contains key information of the distribution network and provides a comprehensive and diverse data foundation; different simulation models in the distribution network respectively obtain the real-time data or historical data of the distribution network collected by sensors, instruments, and equipment corresponding to the simulation models.

[0067] In some embodiments, the simulation model includes one or more of a multi-resolution physical space model, a data space model, a knowledge space model, and a virtual space model; the above step S110, obtaining real-time data or historical data of the distribution network collected by sensors, instruments, and equipment, may include:

[0068] Acquire typical devices in the distribution network under different scenarios, and topological relationships of the typical devices in the distribution network, as real-time data or historical data corresponding to the multi-resolution physical space model;

[0069] Acquire inherent parameters of typical equipment of the distribution network under different scenarios, inherent attribute parameters of the distribution network operation process, inherent attribute parameters of the distribution network association relationship, and influencing variables of the data space model as real-time data or historical data corresponding to the data space model;

[0070] Acquire relevant knowledge of typical equipment in the distribution network as real-time data or historical data corresponding to the knowledge space model;

[0071] Information about typical equipment in the power distribution network is obtained as real-time data or historical data corresponding to the virtual space model.

[0072] Step S120: performing a preprocessing operation on the real-time data or historical data of the distribution network to obtain a training sample set; the preprocessing operation includes data cleaning, denoising, and format conversion;

[0073] Here, the training sample set is sample data used to train the model; the preprocessing operation on the real-time data or historical data of the distribution network is to improve the accuracy of subsequent model prediction and optimization through effective data preprocessing.

[0074] Here, multiple simulation models are constructed based on different operating scenarios in the distribution network; each simulation model can serve as a simulation model to be trained; these simulation models to be trained can obtain prediction values ​​through calculation and simulation based on the physical characteristics, operating parameters, and connection relationships of the distribution network equipment. The simulation models include one or more of a multi-resolution physical space model, a data space model, a knowledge space model, and a virtual space model.

[0075] Step S130: training the simulation model to be trained based on actual data during the actual operation of the distribution network as the annotation information and the training sample set to obtain a trained simulation model;

[0076] Here, actual data from the distribution network's actual operation serves as supervisory information for training the simulation model. The predicted values ​​output by the simulation model are compared with the actual data from the distribution network's actual operation to obtain discrepancy information. This discrepancy information is then passed to the simulation model for correction. This ensures more accurate simulation results and continuously improves the model's precision and generalization capabilities. By continuously adjusting the simulation model, it can better adapt to changes in input data and actual conditions.

[0077] Step S140: determining a multi-rate parallel simulation technology for the distribution network based on the different response speeds of different equipment components in the distribution network to changes in the system operating state;

[0078] Different components in a distribution network typically have different dynamic response speeds and differing abilities to track changes in system states. When the system's operating state changes, components with faster responses can quickly track system disturbances through proportional-integral-derivative control (PID) and achieve a new stable operating state. However, components with slower responses experience a longer transient process before reaching a new stable state. Therefore, it is inappropriate to iteratively solve the same simulation step size for different components. A simulation model is needed that can determine different simulation step sizes based on the varying response speeds of different components in the distribution network to changes in system operating state.

[0079] Step S150: using the multi-rate parallel simulation technology of the power distribution network, adjusting the simulation step size of the trained simulation model to obtain a final optimized model.

[0080] In an embodiment of the present application, a training sample set is obtained by preprocessing the real-time data or historical data of the distribution network collected by the acquired sensors, instruments, and equipment; based on the actual data in the actual operation process of the distribution network as the annotation information and the training sample set, the simulation model to be trained is trained to obtain a trained simulation model; based on the different response speeds of different equipment components in the distribution network to changes in the system operation state, a multi-rate parallel simulation technology for the distribution network is determined; using the multi-rate parallel simulation technology of the distribution network, the simulation step size of the trained simulation model is adjusted to obtain a final optimized model. In this way, based on the actual data in the actual operation process of the distribution network as the annotation information and the training sample set, the simulation model to be trained is trained; this method of training the model can continuously update the model, improve the accuracy of the digital simulation of the distribution network, and can better adapt to changes in the training sample set; in order to realize parallel calculation of the model simulation, the multi-rate parallel simulation technology of the distribution network is used to deal with the situation where different components in the model have different response speeds to changes in the system operation state, so that the components with slow response speed can update the status of the components with fast response speed in real time, thereby improving the speed and accuracy of the digital simulation of large-scale active distribution networks.

[0081] In some embodiments, a simulation task is a simulation process based on the trained simulation model, and the simulation task includes at least two subtasks corresponding to at least two device elements in the distribution network, wherein the at least two subtasks include at least a first subtask and a second subtask; the above-mentioned step S140, determining the distribution network multi-rate parallel simulation technology based on the different response speeds of different device elements in the distribution network to changes in the system operating state, may include steps S141 and S142:

[0082] Step S141: determining a simulation step size for the first subtask and a simulation step size for the second subtask based on different response speeds of device components of different subtasks in the distribution network to changes in system operating states; the simulation step size for the first subtask is a positive integer multiple of the step size for the second subtask;

[0083] Step S142: Based on the current simulation moment, the simulation step of the first subtask, and the sampling coefficient determined by the simulation step of the second subtask, the simulation results within each simulation step of the first subtask are sampled through a triangular wave generator, so that the second subtask can update the state of the first subtask at the starting moment of each simulation step cycle, and determine the multi-rate parallel simulation technology of the distribution network.

[0084] The embodiment of the present application proposes a method for constructing a distribution network digital platform. As shown in FIG2 , the method may include steps S210 to S260:

[0085] Step S210: constructing a multi-resolution physical space model of the typical equipment in the distribution network based on the typical equipment in different simulation scenarios in the distribution network;

[0086] Here, different multi-resolution physical space models are constructed for typical devices in different simulation scenarios;

[0087] Here, the multi-resolution physical space model includes two types: high-resolution model and low-resolution model:

[0088] High-resolution models are often used for short-timescale simulation analysis, including device-level analysis and system-level transient analysis. They primarily use methods such as the equation of state method to build refined simulation models based on the physical laws of object operation.

[0089] Low-resolution modeling is often used in long-time scale simulation analysis, mostly for system-level steady-state analysis. It mainly adopts methods such as average modeling to establish an equivalent PQ model for the modeling object.

[0090] Step S220: constructing a data space model of the typical equipment in the distribution network based on the typical equipment in the distribution network to be modeled; the typical equipment in the distribution network includes lines, loads, converters, switchgear, and distributed power sources;

[0091] Here, the data space model of typical equipment in the distribution network includes a unified data model in the distribution network and a data-driven model in the distribution network.

[0092] Step S230: constructing a knowledge space model of typical equipment in the distribution network based on the knowledge graph of the knowledge related to the model to be modeled in the distribution network;

[0093] Here, the knowledge graph of the knowledge related to the model to be modeled in the distribution network is composed of the knowledge related to the model to be modeled in the distribution network; and a knowledge space model of typical equipment in the distribution network is constructed based on the knowledge graph of the knowledge related to the model to be modeled.

[0094] Step S240: constructing a virtual space model of typical equipment in the distribution network based on the multi-resolution physical space model, the data space model, and the knowledge space model; wherein the virtual space model includes attribute information of the typical equipment in the distribution network;

[0095] Here, the virtual space model is mapped by associating the simulation model in the management module with the virtual space digital model created based on the multi-resolution physical space model, the data space model and the knowledge space model.

[0096] Taking the transformer as an example, the management module searches for the transformer model and creates a virtual space digital model of the transformer based on its multi-resolution physical space model, data space model, and knowledge space model. The transformer model and the virtual space digital model are then mapped and associated, and the successfully created digital simulation model is returned. This model contains all the information and status corresponding to the actual transformer.

[0097] Step S250: Calling between different simulation models in the multi-dimensional digital simulation model of the distribution network by constructing an index architecture of each typical device in the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model;

[0098] In this embodiment, a model indexing architecture is constructed to achieve precise mapping and call access between multi-resolution physical space models, data space modeling, knowledge space models, and virtual space models. The models of each typical device in the multi-resolution physical space model, data space modeling, knowledge space model, and virtual space model are all managed and controlled according to this model indexing architecture.

[0099] Step S260: Building a cloud-based digital platform for the distribution network based on the multi-dimensional digital simulation model of the distribution network.

[0100] In this case, the cloud platform uses distributed computing to divide the system into multiple independent components, each of which can run independently on different computing nodes. This allows for horizontal scaling to increase the number of computing nodes. Distributed tracing technology is used to track application performance and availability, thereby improving computing capabilities and enabling better monitoring and management.

[0101] Here, the cloud platform's distribution network digital platform architecture is based on the cloud platform, and proposes a high-performance and high-concurrency technical architecture based on cloud computing, microservices, and the Internet. It also builds an open and scalable integrated distribution network digital platform architecture system, and builds a cloud platform-based distribution network digital simulation platform overall architecture covering the data layer, model layer, service layer, application layer, and display layer.

[0102] In the embodiment of the present application, a multi-resolution physical space model, a data space model, a knowledge space model and a virtual space model are constructed; by constructing an index architecture of each typical device in the multi-resolution physical space model, the data space model, the knowledge space model and the virtual space model, the calls and mappings between different simulation models in the multi-dimensional digital simulation model of the distribution network are performed, wherein the multi-dimensional digital simulation model of the distribution network includes a multi-resolution physical space model, a data space model, a knowledge space model and a virtual space model; based on the multi-dimensional digital simulation model of the distribution network, an overall architecture of the distribution network digital platform based on the cloud platform is constructed, and the overall architecture of the distribution network digital platform includes a data layer, a model layer, a service layer, an application layer, a display layer and a platform control. In this way, the establishment of the multi-dimensional digital simulation model of the distribution network realizes the dynamic matching of the distribution network entity and the digital model; the data layer, model layer, service layer and application layer of the overall architecture of the distribution network digital platform provide technical means and analysis basis for the verification and testing of the control strategies of various scenarios of the distribution network; the display layer allows users to intuitively understand the operating status and various indicators of the distribution network; the platform control provides auxiliary decision support for the safe, reliable and economical operation of the distribution network.

[0103] In some embodiments, the above step S210, constructing a multi-resolution physical space model of the typical equipment in the distribution network based on the typical equipment in different simulation scenarios in the distribution network, may include steps S211 to S219:

[0104] Step S211: Initializing a multi-resolution physical model of a model to be constructed and setting initial values ​​of power and voltage based on typical devices in different simulation scenarios in the distribution network; the multi-resolution physical model includes two physical models with different resolutions;

[0105] Here, the multi-resolution physical model includes a high-resolution physical model and a low-resolution physical model.

[0106] Step S212: constructing a dynamic phasor network admittance matrix based on the multi-resolution physical model;

[0107] Here, the network admittance matrix is ​​a matrix composed of conductance values ​​between nodes in the multi-resolution physical model of the model to be constructed.

[0108] Step S213: Detecting whether the network topology of typical devices in the distribution network has changed;

[0109] Here, network topology refers to the physical layout of the various devices interconnected by transmission media within the multi-resolution physical model being constructed. It refers to the specific physical (real) or logical (virtual) arrangement of the network's components. If two networks have the same connectivity structure (i.e., their network topologies are the same), their internal physical wiring and inter-node distances may differ.

[0110] Step S214: when the network topology changes, modify the dynamic phasor network admittance matrix;

[0111] Step S215: When the network topology does not change, the voltage value at the current moment is calculated based on the dynamic phasor network admittance matrix and a conversion is performed from the time domain to the dynamic phasor domain;

[0112] Here, if the network topology does not change, the simulation can be ended directly by using the results of the previous power flow calculation.

[0113] Step S216: determining a dynamic phasor value of the current based on the initial values ​​of the power and voltage;

[0114] Step S217: Based on the dynamic phasor value of the current and the current voltage value, determine the dynamic phasor values ​​of each order of the node voltage respectively through a reduced-order dynamic phasor equation;

[0115] Step S218: Based on the dynamic phasor values ​​of each order of the node voltage and the multi-resolution physical model, determine the currents of all branches and perform conversion from the dynamic phasor domain to the time domain;

[0116] Step S219: When it is determined that the simulation end condition is met, a multi-resolution physical space model of a typical device in the distribution network is determined.

[0117] In some embodiments, if the simulation end condition is not met, steps S261 to S266 may be included:

[0118] Step S261: Shift the simulation time forward by one time step;

[0119] Here, the simulation for the next simulation time needs to be performed.

[0120] Step S262: when the network topology changes, modify the dynamic phasor network admittance matrix;

[0121] Step S263: When the network topology does not change, the voltage value at the current moment is calculated based on the dynamic phasor network admittance matrix and a conversion is performed from the time domain to the dynamic phasor domain.

[0122] Step S264: Based on the dynamic phasor value of the current at the previous moment and the voltage value at the current moment, determine the dynamic phasor values ​​of each order of the node voltage respectively through a reduced-order dynamic phasor equation;

[0123] Step S265: determining the current values ​​of all branches based on the dynamic phasor values ​​of each order of the node voltage and the multi-resolution physical model, and performing conversion from the dynamic phasor domain to the time domain;

[0124] In step S266, when it is determined that the simulation end condition is met, a multi-resolution physical space model of a typical device in the distribution network is determined.

[0125] In some embodiments, the above step S220, constructing a data space model of the typical equipment in the distribution network based on the typical equipment in the distribution network to be modeled, may include steps S221 to S223:

[0126] Step S221: constructing a unified data model in the distribution network based on typical devices in the distribution network for which the model is to be constructed;

[0127] Typical equipment in the unified data model for the distribution network includes lines, loads, converters, switchgear (ring main units), and distributed power sources (PV systems, for example). The unified data model for the distribution network includes a unified data model for typical equipment, a unified data model for the distribution network's operational processes, and a unified data model for relationships within the distribution network.

[0128] Step S222: constructing a data-driven model in the distribution network based on the models to be constructed under different operating states and business scenarios in the distribution network;

[0129] Here, the model to be constructed takes the user electricity load prediction model as an example, and constructs a data-driven model of the user electricity load prediction model in the distribution network.

[0130] Step S223: The data space model of the typical equipment in the distribution network includes a unified data model in the distribution network and a data-driven model in the distribution network.

[0131] In some embodiments, the above step S221, constructing a unified data model in the distribution network based on typical devices in the distribution network to be modeled, may include steps S2211 to S2214:

[0132] Step S2211: determining a unified data model of the typical devices in the distribution network based on the typical devices in the distribution network for which the model is to be constructed and inherent parameters of the typical devices in the distribution network;

[0133] The unified data model for distribution network lines includes parameters for each phase conductor, as well as positive-sequence and zero-sequence impedance and admittance parameters. It can describe three-phase balanced, unbalanced, and single-phase lines in the distribution system. Its inherent attributes include line number, feeder number, voltage level, conductor model, and conductor type.

[0134] Here, the unified data model of loads in the distribution network can be built into single-phase, two-phase or three-phase loads, and its inherent attribute parameters include the active component of a fixed number of loads, the percentage of the fixed megawatt value in the fixed megawatt value of the entire load group, the phase difference, the fixed reactive component of the load, the percentage of the fixed megavar value in the total fixed megavar value of the load group, the proportion of constant power active power loads, the proportion of constant current active power loads, the proportion of constant impedance active power loads, the proportion of constant power reactive power loads, the proportion of constant current reactive power loads and the proportion of constant impedance reactive power loads.

[0135] Here, the inherent parameters of the unified data model of the converter in the distribution network include converter type, primary reference voltage, minimum input voltage, maximum input voltage, frequency, rated power, power factor, number of output phases, number of bridges, rectifier reactance, rectifier resistance, composite resistance, minimum composite DC voltage, etc.

[0136] Here, the switchgear in the distribution network includes circuit breakers, load switches, current transformers, voltage transformers, busbars, relay protection devices and auxiliary facilities. The inherent parameters of the unified data model are based on the composite switch type, and expand the busbar wiring method and the number of outgoing lines.

[0137] The unified data model for distributed photovoltaic systems in the distribution network includes a unified data model for photovoltaic power plants and a unified data model for photovoltaic cell modules. Attribute parameters for photovoltaic power plants primarily include photovoltaic power generation type and capacity, while attribute parameters for photovoltaic cell modules primarily include model, maximum energy conversion rate, maximum output power, maximum power fluctuation range, operating voltage at maximum power, operating current at maximum power, open-circuit voltage, short-circuit current, rated operating temperature, battery type, weight, and dimensions.

[0138] Step S2212: determining a unified data model of the distribution network operation process based on typical equipment in the distribution network to be modeled and inherent attribute parameters of the distribution network operation process to be modeled;

[0139] Here, the inherent attribute parameters of the unified data model of the distribution network operation process include the remote operation unit type, monitoring point ID, monitoring point name, substation ID, region code, detection device IP address, reference voltage (not voltage level), reference frequency, identification of the monitoring point location in the grid topology model, wiring method, maximum number of harmonics supported by this monitoring point, steady-state standard name, monitoring point status, connection status, monitoring point location type, user agreement power supply capacity, maximum short-circuit capacity of the device installation point, minimum short-circuit capacity of the device installation point, rated power generation capacity, whether the monitoring point is installed at the PCC point, power factor standard value, positive voltage deviation limit, negative voltage deviation limit, monitoring point user load type, load characteristics, single-phase / three-phase information, etc.

[0140] Step S2213: Determine a unified data model of association relationships in the distribution network based on typical devices in the distribution network to be modeled and inherent attribute parameters of the distribution network association relationships in the distribution network to be modeled; the inherent attribute parameters of the distribution network association relationships include: equivalent power sources, conductors, transformer windings, and nodes connected between equivalent load endpoints;

[0141] The unified data model for distribution network relationships includes topology-related unified data modeling and data-related unified data modeling. Because each conductive device in a distribution network has two endpoints, each connected to a node, the inherent attribute parameters of the unified data model primarily include the nodes connected to endpoints such as equivalent power sources, conductors, transformer windings, and equivalent loads.

[0142] Step S2214: The unified data model in the distribution network includes a unified data model of typical equipment in the distribution network, a unified data model of the distribution network operation process, and a unified data model of association relationships in the distribution network.

[0143] In some embodiments, the above step S222, which constructs a data-driven model in the distribution network based on the to-be-constructed models under different operating states and business scenarios in the distribution network, may include steps S2221 to S2224:

[0144] Step S2221: Based on the models to be constructed under different operating states and business scenarios in the distribution network, determining influencing variables corresponding to the models to be constructed under different operating states and business scenarios;

[0145] Here, taking the user electricity load forecasting model as an example, the influencing variables of the distribution network load forecasting model include not only key factors such as power load and electricity price, but also weather factors such as air pressure, rainfall, visibility, temperature, humidity and wind speed. With reference to the above factors, the influencing variable set of the power load forecasting model is formed: {power load, electricity price, air pressure, rainfall, visibility, temperature, humidity, wind speed}.

[0146] Step S2222: Based on a double recursive analysis method, the influencing variables corresponding to the model to be constructed are used as input variables, the predicted values ​​of the model to be constructed are used as output variables, and the influencing variables corresponding to the model to be constructed are divided into a basic set and a candidate set;

[0147] Based on LASSO regression and ReliefF regression, the first 50% of the common variables in the analysis results are taken as the basic set, and the remaining influencing factors are taken as the candidate set.

[0148] Taking the user electricity load forecasting model as an example, through the double recursive analysis method, the output variable is the load forecast result, the basic set is {power load, electricity price}, and the candidate set is {air pressure, rainfall, visibility, temperature, humidity, wind speed}.

[0149] Step S2223: Based on the neural network algorithm, the basic set is used as an input variable to output the predicted value of the model to be constructed;

[0150] Here, the correlation between the output variables of typical equipment / systems and the influencing factor variables is constructed through the neural network algorithm, forming a data-driven model.

[0151] Step S2224: When there is a deviation between the actual operating data of the distribution network and the predicted value of the model to be constructed, the model is retrained by adjusting the influencing variables in the basic set and the candidate set; and a verified distribution network data-driven model is obtained.

[0152] In some embodiments, the above step S230, constructing a knowledge space model of typical equipment in the distribution network based on the knowledge graph of knowledge related to the model to be modeled in the distribution network, may include steps S231 to S238:

[0153] Step S231: Based on the relevant knowledge of the model to be constructed in the distribution network, determine the attributes and attribute value ranges of each basic concept and sub-concept of the model to be constructed, as well as the data source of the relevant knowledge of the model to be constructed;

[0154] Here, the data sources of relevant knowledge include the import of internal business data of the power system, the import of external system data, the import of general power knowledge graphs, and the crawling of power transformer fault knowledge such as relevant field encyclopedias on the Internet and relevant field technical standards in literature libraries.

[0155] Step S232: Building a basic framework of relevant knowledge of the model to be constructed in the distribution network based on the attributes and attribute value ranges of each basic concept and sub-concept;

[0156] Step S233: Based on the basic structure of the relevant knowledge of the model to be constructed in the distribution network, determining relevant words in the knowledge field of the model to be constructed in the distribution network and basic semantic associations of the relevant words in the knowledge field;

[0157] Here, the relevant words in the knowledge field include the subject words of the knowledge field and high-quality words, synonyms, abbreviations and related sentiment words related to the subject words.

[0158] Step S234: Based on the relevant words in the knowledge domain of the distribution network model to be constructed, identifying the entities to which the relevant words in the knowledge domain of the distribution network model to be constructed belong and classifying the entities;

[0159] Here, relevant terms in a knowledge domain are not necessarily domain entities. Common entities within the knowledge domain must be identified based on these terms. After entity identification, they must be categorized. Assigning entities to appropriate categories (or, in other words, associating entities with domain categories or concepts) is the fundamental goal of entity conceptualization and a key step in understanding entities.

[0160] Step S235: determining the relationships of the relevant knowledge of the model to be constructed in the distribution network based on the entities to which the relevant words in the knowledge domain of the model to be constructed in the distribution network belong; the relationships include inclusion relationships, application relationships, existence relationships, and discovery relationships;

[0161] Relationship discovery, or populating the knowledge base with relationship instances, is the core step in building the domain knowledge graph. Depending on the problem model, relationship discovery can be categorized into relationship classification, relationship extraction, and open relationship extraction. Relationship classification aims to classify a given entity pair into a known relationship; relationship extraction aims to extract the specific relationship between an entity pair from text; and open relationship extraction extracts descriptions of the relationships between entity pairs from text.

[0162] Step S236: Based on the basic structure of the relevant knowledge of the model to be constructed in the distribution network, the data source of the relevant knowledge of the model to be constructed, the basic semantic association of relevant words in the knowledge field of the model to be constructed in the distribution network, and the relationship between the relevant knowledge of the model to be constructed in the distribution network, the relevant knowledge of the model to be constructed in the distribution network is integrated;

[0163] Knowledge fusion requires entity alignment, attribute fusion, and value normalization. Entity alignment identifies the same entity across different sources; attribute fusion identifies different descriptions of the same attribute. Data values ​​from different sources often have different formats, units, or descriptions. For example, dates can be expressed in dozens of different ways, and these need to be normalized to a unified format.

[0164] Step S237: determining a knowledge graph of typical devices in the distribution network based on the integrated knowledge of the model to be constructed in the distribution network;

[0165] Step S238: Based on the knowledge graph of typical equipment in the distribution network, a knowledge space model of typical equipment in the distribution network is constructed.

[0166] The embodiment of the present application proposes a processing method for a distribution network digital platform. As shown in FIG3 , the method may include steps S310 to S360:

[0167] Step S310: Processing the real-time data or historical data of the distribution network through the data layer; the processing method includes data storage, data sharing, data analysis, and data processing;

[0168] Here, the data layer is the foundation of the digital simulation platform. The quality and efficiency of the data layer are crucial to the performance and reliability of the digital simulation platform. The main technologies include data storage, data sharing, data analysis and data processing.

[0169] Step S320: Based on the simulation model of the model layer, the data of the data layer is used as input to perform simulation at the service layer to obtain the simulation results; wherein the simulation results include data under different operating scenarios in the distribution network;

[0170] Here, the model layer includes a multi-resolution physical space model, a data space model, a knowledge space model and a virtual space model.

[0171] The service layer provides various algorithmic services, including power flow calculation, network loss calculation, and distributed power optimization and control calculation. The service layer receives data and information from the data and model layers and performs real-time simulation calculations. This helps users better understand and grasp the actual distribution network situation, enabling them to make better decisions.

[0172] Step S330: Based on the simulation results, the simulation results are applied through the application layer, and the applications include resource management, analysis and evaluation, business control, and decision-making and command;

[0173] Resource management: Through intelligent sensing and data collection technologies, comprehensive management and monitoring of distribution network equipment, facilities, and resources are implemented, enabling asset visualization and digital management, improving asset utilization and management efficiency. Analysis and evaluation: Through intelligent algorithms, analysis and evaluation of distribution network operating data, equipment status, and environmental factors are conducted to promptly identify problems and risks, improve fault diagnosis and prediction capabilities, and provide a basis for operation and maintenance and decision-making. Business management and control: Through distribution network automation and intelligent technologies, remote control, intelligent regulation, and intelligent operation and maintenance of the distribution network are implemented, improving operational efficiency and reliability and reducing operation and maintenance costs and energy consumption. Command and decision-making: Through digital models and scenario applications, the distribution network is visualized, predictable, controllable, and decision-making, providing users with intelligent and customized services and experiences.

[0174] Step S340: displaying the simulation results and the data processed by the application layer through the display layer;

[0175] The presentation layer is the interface for the digital simulation platform, primarily displaying a three-dimensional scene of the distribution network, along with various data and reports. Through this layer, users can intuitively understand the distribution network's operating status and various indicators, providing a basis for decision-making. This layer offers a variety of display formats, including large-screen displays, desktop displays, and mobile applications, all built on a B / S architecture.

[0176] Step S350: realizing data sharing among various modules of the distribution network digital platform through the API interface;

[0177] Here, the API interface defines the way in which the modules of the digital simulation platform interact with each other. The API interface is used to call the functions and data in the system to achieve interaction and data sharing between systems.

[0178] Step S360: Protect the system stability and security of the platform through platform management and control, which includes system permissions, system configuration, and security management.

[0179] Here, platform control is an important component of the distribution network digital simulation platform, mainly including system permissions, system configuration, security management, etc. It can protect the system stability and security of the platform, facilitate the configuration of relevant parameters, and avoid malicious attacks and data leakage.

[0180] In the embodiments of this application, a cloud-based digital distribution network platform architecture is constructed based on a multi-dimensional digital simulation model of the distribution network. The overall architecture comprises a data layer, a model layer, a service layer, an application layer, a presentation layer, and a platform management and control layer. The data, model, service, and application layers of the overall architecture provide technical means and analytical basis for verifying and testing distribution network control strategies in various scenarios. The presentation layer allows users to intuitively understand the operating status and various indicators of the distribution network. The platform management and control provides auxiliary decision-making support for the safe, reliable, and economical operation of the distribution network.

[0181] The following is a specific example of a method for constructing a multi-dimensional digital simulation model of a distribution network, a method for constructing a digital simulation model of a distribution network, and a cloud-based digital simulation platform for a distribution network. However, it is worth noting that this specific example is only for the purpose of better illustrating the present application and does not constitute an improper limitation on the present application.

[0182] The embodiment of the present application provides an overall technical solution for a method for constructing a multi-dimensional digital simulation model of a distribution network. The present application aims to construct a multi-dimensional digital simulation model of a distribution network from four dimensions: physical space, data space, knowledge space, and virtual space. The specific process is shown in FIG4 . The method may include steps S410 to S450:

[0183] Step S410: multi-resolution physical space modeling of distribution network digital simulation;

[0184] Step S420: spatial modeling of distribution network digital simulation data;

[0185] Step S430: Modeling the distribution network digital simulation knowledge space;

[0186] Step S440: digital simulation virtual space modeling of distribution network;

[0187] Step S450: Construct an index architecture for accurate mapping of models between different spaces.

[0188] The above-mentioned steps S410 to S450 are respectively described below with reference to FIG. 5 to FIG. 11 .

[0189] In some embodiments, in step S410, a multi-resolution physical space model of typical equipment in the distribution network is constructed to meet the needs of different simulation scenarios in the distribution network. Multi-resolution mechanism models of typical equipment in the distribution network mainly include two categories: high-resolution models and low-resolution models. High-resolution models are commonly used for short-timescale simulation analysis, including device-level analysis and system-level transient analysis. They mainly use state equation methods and other methods to build a refined simulation model based on the physical laws of the object's operation. Low-resolution modeling is often used for long-timescale simulation analysis, mostly system-level steady-state analysis. They mainly use average modeling methods and other methods to establish an equivalent PQ model for the modeled object.

[0190] Considering the contradiction between simulation accuracy and solution scale, a distribution network equivalent modeling method based on the dynamic phasor method is proposed, as shown in FIG5 . The specific steps may include steps S510 to S580:

[0191] Step S510: Select the system order and initialize;

[0192] Select the system topology and multi-resolution physical models of typical devices, and set initial values ​​for calculation quantities such as power and voltage.

[0193] Step S520: constructing a dynamic phasor network admittance matrix;

[0194] According to the multi-resolution physical model, the dynamic phasor network admittance matrix is ​​constructed using the dynamic vector method.

[0195] Step S530: whether the network topology has changed;

[0196] Check whether the network topology has changed; if so, return to step S520 to modify the network admittance matrix; if not, proceed to step S540.

[0197] Here, if the network topology is not changed, the simulation can be ended directly by using the results of the previous power flow calculation.

[0198] Step S540: Calculate the voltage value and convert it from the time domain to the dynamic phasor domain;

[0199] According to the dynamic phasor network admittance matrix, the voltage value of each node at the current moment is calculated, and the time domain is converted to the dynamic phasor domain.

[0200] Step S550: network solution;

[0201] Given the current dynamic phasor value of the power flow equation at the previous moment and the known voltage at this moment, the dynamic phasor values ​​of each order of the node voltage are calculated respectively through the reduced-order dynamic phasor equation.

[0202] Step S560: Update the dynamic phasor value and calculate the current of all branches from the node voltage;

[0203] According to the dynamic phasor values ​​of each node voltage, the current of all branches is calculated.

[0204] Step S570: Converting the dynamic phasor domain of each order to the time domain;

[0205] Step S580: Check whether the simulation end time has been reached.

[0206] Determine whether the convergence threshold or the set number of simulation iterations has been reached. If so, the simulation time has been reached, and the simulation ends; if not, the simulation time has not been reached, and the process moves forward one time step and enters step S530.

[0207] The model in the digital simulation of this application is updated dynamically in real time. When the model changes and is updated under the demand scenario, the updated model is substituted into step S510 of the process shown in Figure 5.

[0208] In some embodiments, step S420, the distribution network digital simulation data space modeling includes two parts, one is the distribution network unified data model, and the other is the distribution network data driven model. The specific modeling process is shown in Figure 6. First, building the distribution network unified data model may include steps S601 to S604:

[0209] Step S601: Determine the device type;

[0210] It is necessary to identify typical distribution network equipment for which a unified data model is to be constructed. The full information model of typical distribution network equipment mainly includes lines, loads, converters, switchgear (ring main units), distributed power sources (taking photovoltaics as an example), etc.

[0211] Step S602: Analyze the inherent properties of the device and determine unified modeling data for the device;

[0212] For various typical devices in the distribution network, determine their inherent attribute parameters. The unified data model of typical devices in the distribution network can include the following data models:

[0213] 1) The unified line data model includes the parameters of each phase conductor, as well as the positive-sequence and zero-sequence impedances and admittances. It can describe three-phase balanced, unbalanced, and single-phase lines in the distribution system. Its inherent attributes include the line number, feeder number, voltage level, conductor model, and conductor type.

[0214] 2) The mathematical relationship between load and voltage can be expressed by formula (1), with coefficient a p (a q ), b p (b q ) and c p (cq ) are the proportions of the constant impedance, constant current and constant power models respectively.

[0215] In formula (1), P0 and Q0 are reference powers; P and Q are actual powers; v0 is the reference voltage; and v is the actual voltage.

[0216] Therefore, the unified load data model can be built for single-phase, two-phase or three-phase loads, and its inherent attribute parameters include the active component of a fixed number of loads, the percentage of the fixed megawatt value in the fixed megawatt value of the entire load group, the phase, the fixed reactive component of the load, the percentage of the fixed megavar value in the total fixed megavar value of the load group, the proportion of constant power active power loads, the proportion of constant current active power loads, the proportion of constant impedance active power loads, the proportion of constant power reactive power loads, the proportion of constant current reactive power loads and the proportion of constant impedance reactive power loads.

[0217] 3) The inherent parameters of the unified converter data model include converter type, primary reference voltage, minimum input voltage, maximum input voltage, frequency, rated power, power factor, number of output phases, number of bridges, rectifier reactance, rectifier resistance, composite resistance, minimum composite DC voltage, etc.

[0218] 4) The switchgear station contains circuit breakers, load switches, current transformers, voltage transformers, busbars, relay protection devices and auxiliary facilities. The inherent parameters of its unified data model are based on the composite switch type, expanding the busbar wiring method and the number of outgoing lines.

[0219] 5) The unified data model for distributed photovoltaics includes a unified data model for photovoltaic power plants and a unified data model for photovoltaic cell modules. The attribute parameters for photovoltaic power plants primarily include photovoltaic power generation type and photovoltaic power generation capacity, while the attribute parameters for photovoltaic cell modules primarily include model, maximum energy conversion rate, maximum output power, maximum power fluctuation range, operating voltage at maximum power, operating current at maximum power, open-circuit voltage, short-circuit current, rated operating temperature, cell type, weight, and dimensions.

[0220] Step S603: Unifying the data model of the distribution network operation process;

[0221] The inherent attribute parameters of the unified data model of the distribution network operation process include remote operation unit type, monitoring point ID, monitoring point name, substation ID, region code, detection device IP address, reference voltage (not voltage level), reference frequency, identification of the monitoring point location in the grid topology model, wiring method, maximum number of harmonics supported by this monitoring point, steady-state standard name, monitoring point status, connection status, monitoring point location type, user agreement power supply capacity, maximum short-circuit capacity of the device installation point, minimum short-circuit capacity of the device installation point, rated power generation capacity, whether the monitoring point is installed at the PCC point, power factor standard value, positive voltage deviation limit, negative voltage deviation limit, monitoring point user load type, load characteristics, single-phase / three-phase information, etc.

[0222] Step S604: Associating the distribution network with a unified data model.

[0223] The unified data model for distribution network relationships includes topology-related unified data modeling and data-related unified data modeling. Because each conductive device in a distribution network has two endpoints, each connected to a node, the inherent attribute parameters of the unified data model primarily include the nodes connected to endpoints such as equivalent power sources, conductors, transformer windings, and equivalent loads.

[0224] Secondly, constructing a distribution network simulation data-driven model may include steps S611 to S617:

[0225] Step S611: Analyzing variables affecting the operation of distribution network equipment / system in typical scenarios;

[0226] Analyze the influencing variables of the distribution network simulation modeling object in typical scenarios. There are many devices in the distribution network, and the distribution network operating status and business scenarios are diverse. Therefore, there are many factors affecting the data-driven modeling of the distribution network, and the model complexity is too high. This application first analyzes the influencing variables, and then selects the basic variables for data-driven modeling.

[0227] Taking the user electricity load forecasting model as an example, the influencing variables of the distribution network load forecasting model include not only key factors such as power load and electricity price, but also weather factors such as air pressure, rainfall, visibility, temperature, humidity, and wind speed. With reference to these factors, the influencing variable set of the power load forecasting model is formed: {power load, electricity price, air pressure, rainfall, visibility, temperature, humidity, wind speed}.

[0228] Step S612: Divide the data into a basic set and a candidate set;

[0229] Based on a double recursive analysis method, the modeling object's influencing variables are divided into a base set and a candidate set. The purpose of variable division is to extract and select core feature variables and reduce the dimensionality of the input variables while ensuring the accuracy of the prediction model. The base set is the set of factors that have a significant impact on the modeling object's operation, while the candidate set is the set of factors with a smaller or unknown impact on the modeling object's operation.

[0230] This application classifies influencing factors based on regression methods. The following uses LASSO regression and ReliefF regression as an example to classify influencing factors.

[0231] LASSO regression mainly constructs a penalty function to compress the coefficient λ of the input variable that has no obvious effect on the output variable (predicted value) to 0, thereby achieving the purpose of variable screening. The loss function containing the penalty function is constructed as shown in formula (2):

[0232] In formula (2), n is the number of input variables; p is the number of penalty input variables; x i is the input variable; i is the measured value of the output variable; h(x) is the predicted value of the output variable; θ is the weight of each variable; and λ is the penalty coefficient for the variable weight. By adjusting the penalty coefficient λ, models with more variables can be adjusted. The larger the value, the greater the penalty for the model and the smaller the weight of the variable.

[0233] The LASSO regression algorithm tends to miss some variables that have a significant impact on the input, so the ReliefF regression algorithm is introduced to compensate for this deficiency. ReliefF regression randomly selects a sample R from the model's training set, extracts k nearest neighbor samples H from R's sample set of the same type, and then finds k nearest neighbor samples M from other different sample sets. The weight of each feature is updated, as can be seen in formula (3).

[0234] In formula (3), p(C) is the proportion of the category; p(class(R)) is the proportion of the category of a randomly selected sample; diff(A,R1,R2) is the difference between sample R and H on feature A; m is the number of loop calculations, M j (C) is a class The jth nearest neighbor sample in .

[0235] Based on LASSO regression and ReliefF regression, the first 50% of the common variables in the analysis results are taken as the basic feature set, and the remaining influencing factors are taken as the candidate set.

[0236] Still taking the data-driven load forecasting model as an example, through the regression analysis in this step, the output variable is the load forecast result, the basic set is {power load, electricity price}, and the candidate set is {air pressure, rainfall, visibility, temperature, humidity, wind speed}, etc.

[0237] Step S613: data driven model construction;

[0238] By combining the influencing factors in the basic set, algorithms such as neural networks are used to construct data-driven models for devices or systems under typical scenario requirements. This application uses the extreme learning machine (ELM) method as an example to illustrate data-driven modeling of devices or systems under typical scenario requirements. The ELM is characterized by a single hidden layer neural network for ease of demonstration.

[0239] From the determined basic set, the training samples are taken as The total number of samples is N, x is the element in the basic set, and y is the output result of the device or system in a typical scenario. The input vector of the tth sample is x t =[x t1 ,x t2 ,...,x tn ], the dimension is n, and the output vector is y t =[y t1 ,y t2 ,...,y tm ], the dimension is m. The neural network can be shown in Figure 7. If the number of nodes in the hidden layer 702 is L, the input layer 701 includes the input variable x t1 、x t2 、x tn , and the bias b of the hidden layer unit; the output layer 703 includes y t1 、y t2 and y tn , then the correlation between the output variables of typical equipment / systems and the influencing factor variables can be seen in formula (4):

[0240] In formula (4), f(x) is the activation function; ω i is the influence weight of the input variable; β i is the output weight; b i is the bias of the hidden layer unit.

[0241] In this way, the correlation between the output variables of typical equipment / systems and the influencing factor variables is constructed, forming a data-driven model.

[0242] The data-driven load forecasting model is also used as an example for explanation.

[0243] The basic set and candidate set determined by step S612, the input variables of the neural network in this step are power load and power price, that is, x in formula (4); ω i Obtained by LASSO regression and ReliefF regression analysis; β i , f() and b i Obtained by neural network calculation; y i It is the output of the neural network; its physical meaning is the predicted value of the load model.

[0244] Step S614: Model accuracy verification;

[0245] The accuracy of the model can be verified based on actual operation data. If there is a deviation between the two, first adjust the input and output weights ω i and β i Observe whether the error has decreased; if the error has not changed much, it is necessary to add the elements in the candidate set to the basic set one by one, retrain the model to form an updated data-driven model, and perform a model accuracy check.

[0246] If the accuracy of the updated data-driven model is improved, the element in the candidate set is added to the basic set; otherwise, the factor is discarded. This process continues until all elements in the candidate set are analyzed and the model accuracy reaches the required threshold.

[0247] Step S615: whether the accuracy meets the requirements;

[0248] When the model accuracy reaches the required threshold, determine whether the accuracy meets the requirements; if so, proceed to step S616; if not, proceed to step S617 to update the basic set, and then proceed to step S612.

[0249] Step S616: Build a data model library.

[0250] In some embodiments, step S430 is to construct a distribution network simulation knowledge space model based on the knowledge graph. Taking the knowledge related to transformer failure of key power equipment in the distribution network as an example, the construction process of the distribution network digital simulation knowledge space model is analyzed, as shown in Figure 8, which may specifically include steps S810 to S880:

[0251] Step S810: pattern design;

[0252] The goal of knowledge graph model design is to give machines the basic framework of the cognitive field. This paper designs a knowledge graph for distribution power transformer faults.

[0253] First, determine the basic concepts of power transformer fault knowledge.

[0254] The basic concepts of power transformer fault knowledge include: equipment name, fault hazards, inspection process, inspection items and inspection methods, etc.

[0255] Then, the subclass relationships between concepts in the power transformer fault knowledge category are clarified.

[0256] The logical sequence of concepts is equipment, troubleshooting, inspection process, inspection items, inspection methods, and troubleshooting, specifically including the following:

[0257] 1) Potential fault hazards include overload operation, abnormal oil temperature, abnormal noise, cooling fan failure, ungrounded, seepage, casing damage, etc.

[0258] 2) Inspection process includes line inspection process, instrument inspection process, load status inspection process, environment inspection process, appearance inspection process and function inspection process;

[0259] 3) Inspection items include grounding wire inspection, power automatic cut-off device inspection, temperature control device inspection, cooling fan inspection, overload inspection, odor inspection, abnormal sound inspection, casing inspection, connection parts inspection, etc.;

[0260] 4) Inspection methods include computer video inspection, manual inspection, electrical quantity inspection, sound inspection, temperature inspection, air temperature inspection, etc.

[0261] Finally, we build a basic framework for knowledge about power transformer faults in distribution networks, clarifying the attributes and attribute value ranges of each concept and sub-concept. For example, we need to clarify the oil temperature range during normal operation of power transformers.

[0262] Step S820: data source selection;

[0263] The data sources for knowledge on power transformer faults in distribution networks mainly include four categories:

[0264] 1) Importing internal business data of the power system, mainly historical data imported from the property management platform and operating data (including measurement and control information, protection information, and recording information) imported from the SCADA (data acquisition or monitoring system);

[0265] 2) Importing external system data, mainly including simulation data obtained by power personnel based on commercial simulation software;

[0266] 3) Import of general power knowledge graph;

[0267] 4) Crawling knowledge about power transformer faults, such as related encyclopedias on the Internet and technical standards in related fields in literature libraries.

[0268] Step S830: vocabulary mining;

[0269] For distribution network fault scenarios, starting from basic concepts and basic vocabulary, based on the knowledge space architecture constructed in step S810, the subject words of this knowledge field and high-quality vocabulary, synonyms, abbreviations and related sentiment words related to the subject words are mined, and the basic semantic associations between these words are mined.

[0270] For example, the subject terms related to transformer fault knowledge include transformer, transformer fault hazards, transformer inspection process, transformer inspection items and transformer inspection methods; high-quality vocabulary related to the subject terms includes two-winding transformer, three-winding transformer, etc.; synonyms include transformer overload and line current too high, etc.; related emotional words include "should", "can", "maybe", etc.

[0271] Step S840: Entity discovery;

[0272] The domain vocabulary involved in step S810, step S820 and step S830 only identifies important phrases and words in the corresponding domain, which is not necessarily a domain entity. Common entities in the domain need to be identified based on the domain text.

[0273] For example, when words such as "oil temperature detection" are mentioned, it is necessary to recognize that it may be a transformer in the distribution network field.

[0274] After the entity is identified, it needs to be classified. Whether the entity can be classified into the corresponding category (or whether an entity can be associated with a domain category or concept) is the basic goal of entity conceptualization and a key step in understanding the entity.

[0275] For example, classifying power transformers into categories such as power equipment and winding equipment is of great significance for understanding the meaning of transformers through knowledge graphs.

[0276] Another important task in entity mining is entity linking, which involves linking entity mentions in text to corresponding entities in the knowledge base. Entity linking is a key step in expanding entity understanding and enriching entity semantic representation.

[0277] Step S850: relationship discovery;

[0278] Relationship discovery, or populating a knowledge base with relationship instances, is a core step in building a domain knowledge graph. Depending on the problem model, relationship discovery can be categorized into relationship classification, relationship extraction, and open relationship extraction. Relationship classification aims to classify a given entity pair into a known relationship; relationship extraction aims to extract the specific relationship between an entity pair from text; and open relationship extraction extracts descriptions of the relationships between entity pairs from text.

[0279] Take the knowledge relationship classification of power transformer faults as an example.

[0280] There are four main types of relationships, including relationship, application relationship, existence relationship and discovery relationship. In practical applications, it can be seen that transformer fault knowledge is classified into the above four types of relationships.

[0281] For example, the appearance inspection process of a transformer includes bushing inspection, connection component inspection, color-changing silicone inspection, and oil leakage inspection. That is, there is an association relationship between the above key words; there is a discovery relationship between bushing inspection and bushing damage; there is an application relationship between bushing inspection, color-changing silicone inspection, etc. and computer video inspection; there is an existence relationship between the transformer and bushing damage, etc.

[0282] Referring to the above relationship classification, it can be seen that the relationship between the transformer's fault hazards, inspection process, inspection items, and inspection methods constitutes a transformer fault knowledge graph.

[0283] Step S860: knowledge fusion;

[0284] Knowledge fusion requires entity alignment, attribute fusion, and value normalization. Entity alignment identifies the same entity across different sources; attribute fusion identifies different descriptions of the same attribute. Data values ​​from different sources often have different formats, units, or descriptions. For example, dates can be expressed in dozens of ways, and these need to be normalized to a unified format.

[0285] Step S870: quality control;

[0286] Knowledge about power transformer faults in distribution networks may be missing, incorrect, or outdated. During the rolling interaction and analysis with the physical distribution network, expert knowledge needs to be continuously supplemented, corrected, and updated. This, combined with step S810, allows for quality control of the expert knowledge space within the distribution network.

[0287] Step S880: Domain knowledge graph.

[0288] Based on the construction of the knowledge space in the above steps, a knowledge graph of typical equipment or systems in the distribution network is formed. The knowledge graph of transformer faults in the distribution network is shown in Figure 9, including transformer 901, temperature check 902, line inspection process 903, instrument inspection process 904, oil-immersed transformer 905, cooling fan failure 906, temperature control device 907, grounding wire inspection 908, appearance inspection process 909, abnormal sound 910, odor inspection 911, overload 912, environment inspection process 913, load status inspection 914, overload operation 915, Abnormal oil temperature 916, oil leakage 917, ungrounded wire 918, cooling fan inspection 919, presence of boiling oil 920, temperature control device 921, manual inspection 922, automatic power switching 923, functional inspection process 924, casing damage 925, casing inspection 926, abnormal sound inspection 927, content inspection 928, air temperature inspection 929, discoloration of color-changing silicone 930, computer vision inspection 931, connection component inspection 932, and electrical quantity inspection 933.

[0289] After forming the knowledge graph of typical equipment or systems in the distribution network, as shown in Figure 8, the user can edit the knowledge space graph through user editing 81, and the knowledge graph can be updated by many users other than the user who can edit and change the knowledge graph through crowdsourcing construction 82.

[0290] In some embodiments, a specific flow chart of creating a digital simulation virtual space model of a distribution network in step S440 is shown in FIG10 . Taking a transformer as an example, the construction of the virtual space digital model thereof may include steps S1010 to S1050:

[0291] Step S1010: searching for an object model in the 3D model display and management module;

[0292] Taking the transformer as an example, the physical characteristics required for constructing the transformer virtual space model are determined.

[0293] Step S1020: creating a digital simulation model;

[0294] Taking the transformer as an example, after finding the appropriate model, the model management module begins to create a virtual space digital model of the transformer based on the physical space model, data space model and knowledge space model. This digital model contains all the information corresponding to the actual target transformer.

[0295] Step S1030: initialization state;

[0296] Taking the transformer as an example, initialize the state of the transformer model.

[0297] Step S1040: model mapping;

[0298] Here, the model in step S1020 is associated and mapped with the model in step S1010.

[0299] Step S1050: Return to the successfully created virtual space digital simulation model.

[0300] Once the digital simulation model is configured and mapped, the successfully created digital simulation model is returned. This model contains all the information and status corresponding to the actual transformer and can be displayed and used in subsequent simulation applications.

[0301] In some embodiments, regarding step S450 above, an index architecture for accurate mapping of models between different spaces is constructed. By constructing a model index architecture, accurate mapping and calling of models in different spaces are achieved. The models of each typical device in the physical space, data space, knowledge space, and virtual space are all managed and controlled according to the model index architecture. The distribution network digital simulation model index is shown in Figure 11. The distribution network digital twin model library 1100 may include the following models:

[0302] 1. The first-level directory 1101 of the model library mainly includes distribution network equipment model 1105, load model 1106, node model 1107, power electronic equipment model 1108, distributed power supply model 1109 and system model 1110, etc.

[0303] 2. The secondary directory 1102 under the distribution network equipment model 1105 mainly includes the power supply equipment model 1111 and the power consumption equipment model 1112.

[0304] 1) The third-level directory 1103 under the power supply equipment model mainly includes primary equipment models of typical power supply equipment in the distribution network, such as transformers, lines, circuit breakers, fuses, grounding switches, load switches, and section switches, as well as secondary equipment models such as mutual inductors, intelligent terminals, and reactive power compensation devices.

[0305] 2) The third-level directory 1103 under the electrical equipment model 1112 mainly includes motor and generator models.

[0306] 3. The secondary directories under load model 1106 mainly include static load model 1113, node load model 1114, dynamic load model 1115, steady-state load model 1116, transient load model 1117, random distribution load model 1118, user load model 1119, system load model 1120 and other models of commonly used loads in distribution networks.

[0307] 1) The third-level directory 1103 under the static load model 1113 mainly includes constant impedance load model, constant power load model, constant current load model and polynomial model.

[0308] 2) The third-level directory 1103 under the dynamic load model 1115 mainly includes the load model considering the mechanical transient process of the induction motor and the load model considering the electromechanical transient process of the induction motor.

[0309] 3) The third-level directory 1103 under the random distribution load model 1118 mainly includes the uniform distribution load model, the normal distribution load model and the chi-square distribution load model.

[0310] 4. The secondary directory 1102 under the node model 1107 mainly includes segment node model, branch node model, distribution transformer / substation node model, user node model, distribution station node model, ring station node model, switch station node model, distributed power supply / energy storage / microgrid node model, node voltage model, and node current model.

[0311] 5. The secondary directory 1102 under the power electronic equipment model 1108 mainly includes models of typical power electronic conversion equipment such as rectifiers, inverters, choppers, and AC-AC converters.

[0312] 6. The secondary directory 1102 under the distributed power supply model 1109 mainly includes the DC power supply model 1121 and the AC power supply model 1122.

[0313] 1) The third-level directory 1103 under the DC power supply model 1121 mainly includes photovoltaic power generation model, fuel cell model, battery model, and supercapacitor system model.

[0314] 2) The third-level directory 1103 under the AC power model 1122 mainly includes wind power generation model, gas turbine power generation model, small hydropower model, flywheel energy storage model, and biomass power generation model.

[0315] The embodiment of the present application provides an overall technical solution of a distribution network simulation method for distributed resource access. The method may include steps S1710 to S1730:

[0316] Step S1710: Determine a hybrid simulation driving mode based on model driving and data driving;

[0317] The hybrid symbiotic simulation-driven model, based on model-driven and data-driven approaches, integrates both model-driven and data-driven approaches. Through multi-source information collection and data preprocessing, predictions are made in conjunction with existing simulation models. Data feedback and model optimization are then used to update the model and optimize the results.

[0318] In this mode, the simulation system obtains information from multiple different sources and combines this information to generate simulation results. The model-driven approach refers to using existing simulation models to generate output results, while the data-driven approach uses real-time collected data to update the model and optimize the simulation results. Specifically, this hybrid symbiotic simulation-driven mode can include steps S1711 to S1716:

[0319] Step S1711: multi-source information collection;

[0320] Collect real-time or historical data from various sensors, instruments, devices, etc. These data contain key information about the distribution network and can provide a comprehensive and diverse data foundation.

[0321] Step S1712: data preprocessing;

[0322] Preprocess the collected data, including data cleaning, denoising, format conversion, etc. Effective data preprocessing can improve the accuracy of subsequent model prediction and optimization.

[0323] Step S1713: model prediction;

[0324] Based on existing simulation models or pre-trained models, data is predicted to generate initial simulation results. These models can generate simulation results through calculation and simulation based on the physical characteristics, operating parameters, and connection relationships of the distribution equipment.

[0325] Step S1714: data feedback;

[0326] The results generated by the model are compared with the actual data, and the difference information is passed to the model through a feedback mechanism for correction. This ensures more accurate simulation results and continuously improves the accuracy and generalization ability of the model.

[0327] Step S1715: model optimization;

[0328] Using the results of data feedback, the model is optimized and improved to further improve its accuracy and generalization ability. By continuously adjusting the model and parameters, the model can better adapt to changes in input data and changes in actual conditions.

[0329] Step S1716: Simulation execution.

[0330] After the model is optimized, a data-driven approach is used to update the model in real time, and a model-driven approach is used to generate the final simulation results.

[0331] Step S1720: reading and processing online data in the distribution network;

[0332] Online data reading in the distribution network is achieved through the data interaction interface in the distribution network system. As shown in Figure 12, the sources of online data mainly include the distribution network automation system 1201 of the safety protection zone III, the dispatching automation system 1202, the GIS system 1207 of the zone III, the distribution production management (MIS) system 1203, the marketing automation system 1204, the marketing management system 1205, the data center 1206, other automation systems in zones I and II 1208, and the distribution network online simulation interface 1209 of zone III. These data are exchanged through the middleware service 1200 that complies with IEC61968.

[0333] Online data processing in the distribution network includes online computational data splicing and dynamic intelligent device mapping. Online computational data splicing is the basis for obtaining detailed online computational data for the entire network; the device mapping table (i.e., dynamic intelligent device mapping) bridges the gap between online computational data and offline data.

[0334] 1. Online calculation data splicing;

[0335] Power grid dispatch centers typically employ a hierarchical and partitioned approach to grid monitoring. Typically, a specific energy management system only models the primary grid within its jurisdiction, while external or lower-level grids are treated as equivalent (or simplified) grids. This decentralized, independent modeling approach fragments the unified physical grid in the real world within the computational data, making it impossible to provide online data integration programs with detailed online computational data describing the entire grid.

[0336] Online calculation data splicing automatically combines dispersed online calculation data from upstream and downstream dispatch systems by analyzing the boundary information between them, creating a single set of online calculation data that describes a detailed grid model for the jurisdiction. This integrated online grid calculation data encompasses not only the main grid information within the jurisdiction but also the grid information for sub-regions, providing a unified, detailed grid model and real-time information, laying the foundation for subsequent online data integration. Furthermore, each energy management system essentially maintains its own equipment information, ensuring no additional burden. Instead, the previously dispersed and independent maintenance work can now be shared.

[0337] 2. Dynamic intelligent mapping of power grid equipment names.

[0338] The basic principle of dynamic intelligent mapping of power grid device names is to dynamically form various device mapping tables based on the device names, connection relationships, numbers, component parameters, and manual customization information in online calculation data / offline data. It supports many-to-many mapping and mapping between different types of devices.

[0339] For cross-regional distribution networks, due to the wide range of data involved, large data volumes, and complex corresponding relationships, it is difficult to solve the problem of dynamic mapping of power grid equipment using traditional manual maintenance methods. Based on the actual situation and laws of the distribution network, the following principles and methods are adopted in the development of the program to achieve dynamic intelligent mapping of distribution network equipment:

[0340] 1) Dynamic device mapping based on the plant / station as the basic mapping unit. In power grid calculation data, plants / stations are located in the middle layer of the grid hierarchy. Their number is relatively limited, making plant / station mapping between different data relatively easy to maintain. Furthermore, establishing plant / station mapping greatly simplifies mapping of power grid devices inside and outside the plant.

[0341] 2) Dynamic intelligent mapping of equipment names is based on the standard naming of plants and equipment as the core technology, while also referring to the connection relationship, numbering, parameters and other related information of the equipment.

[0342] 3) Implement complex mapping of equivalent devices and equivalent networks based on the equivalent attributes of interconnecting equipment (lines or transformers) and devices. Due to modeling inconsistencies between offline and online calculation data, in some cases, the online calculation data only includes equivalent measurements of power plant outgoing lines. In other cases, where the actual system includes power plants commissioned in advance, the online calculation data includes detailed plant modeling, while the offline data is relatively simple. In these cases of online / offline data inconsistencies, complex mapping relationships must be established to ensure that online measurements of equivalent devices and equivalent power grids are accurately and completely reflected in the integrated power flow.

[0343] Based on the above analysis, the main process of forming the mapping table may include steps S1721 to S1726:

[0344] Step S1721: Generate a plant station mapping table by combining intelligent matching and manual designation methods;

[0345] Step S1722: Generate a generator and plant power mapping table;

[0346] Step S1723: forming a transformer mapping table;

[0347] Step S1724: forming an AC line (T-connected line);

[0348] Step S1725: forming a load, load-line, and load-transformer mapping table;

[0349] Step S1736: Generate a mapping table for other electrical devices such as series / parallel capacitors / reactors, DC lines, etc.

[0350] The mapping of power grid device names also includes device name mapping between online calculation data and offline data. By mapping the power grid device names between the main zone and sub-zone online calculation data, it is possible to compare the differences in current and parameters of the same electrical equipment in different zones, promptly assess the quality of the online calculation data, and lay a good foundation for online data integration. At the same time, through the mapping table of equivalent devices and specific devices, the main zone and sub-zone online calculation data can be spliced ​​together. When the online data integration program needs to update offline data, the power grid device name mapping relationship between the old and new offline data can promptly identify devices with changed names and newly added devices, bringing convenience to the debugging and maintenance program.

[0351] Step S1730: Multi-rate parallel simulation technology.

[0352] To achieve parallel computing for digital simulation of distribution networks, on the one hand, it is necessary to decompose large-scale networks to ensure balanced computing among CPU cores and that the computing power of each core is close to the optimal computing power of a single core; on the other hand, the decomposition scheme also needs to minimize the amount of interaction between cores, thereby reducing the amount of data stored at the interface.

[0353] Different components of the distribution network usually have different dynamic response speeds, and their ability to track changes in system states is also different. When the system operating state changes, components with fast response speeds can quickly track system disturbances through proportional-integral-derivative control (PID) and reach a new stable operating state; components with slow response speeds will experience a longer transient process before reaching a new stable operating state. Therefore, it is inappropriate to use the same simulation step size for iterative solution of different components. Using long step size simulation for components with fast response speeds can not only ensure their faster response speed, but also not generate a huge amount of calculation in a short time; using short step size simulation for components with slow response speeds can accelerate convergence by reducing the iterative step size of the solver, so that it can track the state changes of the system as quickly as possible and shorten the simulation process. Accordingly, this application proposes a multi-rate parallel simulation technology that uses different simulation step sizes for different subsystems to accelerate the transient response of the system without significantly increasing resource consumption, thereby ensuring the safe and stable operation of each subsystem.

[0354] However, due to the different simulation step sizes of multiple subsystems in parallel computing, data interaction can only be performed at times that are integer multiples of the maximum simulation step size. This makes it impossible for the small-step simulation system to obtain the simulation results of the external system at the beginning of each simulation time step, making it difficult to ensure the real-time nature of information interaction. This application proposes a multi-rate parallel simulation method between CPU subtasks. The data update mode of CPU subtasks with different sampling step sizes is shown in Figure 13. Among them, CPU subtask 1 uses a large sampling step size T1, and CPU subtask 2 uses a small sampling step size T2, and T1 = mT2, where m is a positive integer.

[0355] The multi-rate parallel simulation data interaction mode between CPU subtasks is shown in FIG13 , and the specific execution steps may include steps S1310 to S1360:

[0356] Step S1310: Call the corresponding distribution network model, set the subtask sampling step, complete the initialization and n Always on time;

[0357] Step S1320: Subtask 1 passes its own [t n-2 , t n-1 ] when the simulation results x n-1 and in [t n-1 , t n ]Simulation results within the time step x n ;

[0358] Formula (5) is the simulation result of subtask 1 passing to subtask 2 at step k:

[0359] In formula (5), x n,k Indicates the simulation result of subtask 1 passing to subtask 2 at step k, x n-1 Indicates that subtask 1 is in [t n-2 , t n-1 ], x n Indicates that [t n-1 , t n ]Simulation results within the time step, Represents the interpolation coefficient, m is a positive integer.

[0360] Step S1330: Subtask 2 passes its own [t n-1 , t n ]Simulation results within the time step y n ;

[0361] Step S1340: Subtask 2 confirms that step S1320 has been completed. n , t n+1] time step, the delay module is used to record the state x of subtask 1 at the end of the previous simulation step n-1 , subtask 2 uses x n-1 and the current state x of subtask 1 n Perform interpolation to obtain the predicted state values ​​of m subtasks 1 in sequence, and perform m iterative calculations;

[0362] Step S1350: Subtask 1 confirms that step S1330 has been completed and uses the current state of subtask 2 y n Perform an iterative calculation;

[0363] Step S1360: Determine whether the simulation end condition is met; if so, end the simulation; if not, proceed to step S1320 to perform subtask 1 and subtask 2 at t n+1 Simulation of time.

[0364] During the simulation process, subtask 2 obtains the calculation results of subtask 1 in real time, and uses a triangular wave generator to complete the adaptive calculation and sampling of each large-step interpolation coefficient (k-1) / m, so that subtask 2 can update the status of subtask 1 at the beginning of each simulation cycle, and can obtain the current status of subtask 1 that is closer to the real-time measurement value at the beginning of each simulation step of subtask 2.

[0365] The embodiment of the present application provides an overall technical solution for a method for constructing a digital simulation platform for a distribution network based on a cloud platform.

[0366] This application proposes a cloud-based digital distribution network platform architecture. With cloud platform 1400 as its foundation, it proposes a high-performance, high-concurrency technology architecture based on cloudification, microservices, and the Internet. It also builds an open, scalable, and integrated distribution network digital platform architecture system. It also constructs a cloud-based 1432-based distribution network digital simulation platform overall architecture covering the data layer 1405, model layer 1404, service layer 1403, application layer 1402, and presentation layer 1401. The overall architecture of the cloud-based 1400 distribution network digital simulation platform is shown in Figure 14. It includes the following modules:

[0367] 1) Cloud platform 1400;

[0368] Cloud Platform 1400 builds a microservice-based, internet-based, high-performance, high-concurrency technology architecture. Within Cloud Platform 1400, distributed computing is used to divide the system into multiple independent components, each of which can run independently on different computing nodes. This allows for increasing the number of computing nodes through horizontal scaling. Distributed tracing technology is used to track application performance and availability, thereby improving computing capabilities and enabling better monitoring and management.

[0369] 2) Data layer 1405;

[0370] The data layer 1405 is the foundation of the digital simulation platform. The quality and efficiency of the data layer 1405 are crucial to the performance and reliability of the digital simulation platform. Key technologies include data storage 1420, data sharing 1421, data analysis 1422, and data processing 1423. Data storage 1420: Stores externally collected data, internal calculation results, model data, and other data. It consists of a graphics library and an attribute database, stored in Oracle / Access, and provides a basic support system and basic general data services, data persistence, and database access capabilities for platform layer calls. Data sharing 1421: Through network protocols (such as HTTP, TCP, etc.), digital simulation basic data is shared with each application module in a synchronous or asynchronous manner, providing channels for display such as power flow calculation, network loss calculation, and distributed energy optimization scheduling. Data analysis 1422: Extracts value from data collected by a large number of digital devices. Data-based models, predictive analysis techniques, and association rule mining algorithms are used to derive data relationships and patterns, understanding data trends. Data processing (1423): Technical approaches include data deduplication, missing value processing, outlier processing, and data standardization. Data processing (1423) transforms raw data into a format suitable for distribution network digital simulation applications.

[0371] 3) Model layer 1404;

[0372] The model layer 1404 is the core of the digital simulation platform and primarily includes the distribution network's physical model 1416, data model 1417, knowledge model 1418, and virtual model 1419. By managing and maintaining these models, the digital simulation platform enables digital representation, monitoring, prediction, and control of the distribution network, improving the reliability and efficiency of grid operations.

[0373] 4) Service layer 1403;

[0374] The service layer 1403 provides various algorithmic services, including power flow calculation 1413, loss calculation 1414, and distributed power optimization and control calculation 1415. The service layer 1403 receives data and information from the data layer and model layer to perform real-time simulation calculations, thereby helping users better understand and grasp the actual distribution network situation and make better decisions.

[0375] 5) Application layer 1402;

[0376] The distribution network digital simulation application layer 1402 specifically implements applications such as resource management 1409, analysis and evaluation 1410, business management and control 1411, and decision-making and command 1412. Resource management 1409: Through intelligent sensing and data collection technologies, comprehensive management and monitoring of distribution network equipment, facilities, and resources are achieved, enabling asset visualization and digital management, improving asset utilization and management efficiency. Analysis and evaluation 1410: Through intelligent algorithms, analysis and evaluation of distribution network operating data, equipment status, and environmental factors are conducted to promptly identify problems and risks, improve fault diagnosis and prediction capabilities, and provide a basis for operation and maintenance and decision-making. Business management and control 1411: Through distribution network automation and intelligent technologies, remote control, intelligent regulation, and intelligent operation and maintenance of the distribution network are implemented, improving operational efficiency and reliability and reducing operation and maintenance costs and energy consumption. Command and decision-making 1412: Through digital models and scenario applications, the distribution network is visualized, predictable, controllable, and decision-making, providing users with intelligent and customized services and experiences.

[0377] 6) presentation layer 1401;

[0378] Presentation layer 1401 is the digital simulation platform's interface, primarily displaying a three-dimensional scene of the distribution network, along with various data and reports. Through presentation layer 1401, users can intuitively understand the distribution network's operating status and various indicators, providing a basis for decision-making. Presentation layer 1401 includes various display formats, including large-screen 1406 displays based on a B / S architecture, desktop 1407 displays, and mobile 1408 application displays.

[0379] 7) API interface 1424;

[0380] API interface 1424 defines how the various modules of the digital simulation platform interact with each other. Through API interface 1424, functions and data in the system are called to achieve interaction and data sharing between systems. The distribution network digital simulation platform API interface 1424 mainly includes a simulation service API interface 1426, a model data API interface 1427, and a real-time calculation API interface 1428. The simulation service API interface 1426 provides computing service interfaces such as flow calculation, network loss calculation, and distributed power optimization and control; the model data API interface 1427 provides various distribution network model data, including data interfaces such as node data, network topology, equipment data, and load data; and the real-time calculation API interface 1428 provides output interfaces for various real-time data and calculated data.

[0381] 8) Platform control 1425.

[0382] Platform management and control 1425 is an important component of the distribution network digital simulation platform, which mainly includes permission management 1429, configuration management 1430, security management 1431 and other contents. It can protect the system stability and security of the platform, facilitate the configuration of relevant parameters, and avoid malicious attacks and data leakage.

[0383] In response to the problems of passive acceptance, insufficient management and control, low degree of information digitization, and limited observable and controllable capabilities of distribution networks under the current background of distributed resource access, and considering the shortcomings of existing simulation of distribution networks, this application provides a distribution network digital simulation method and system for distributed resource access: from the dimensions of physical space, data space, knowledge space and virtual space, a method for constructing a multi-dimensional digital simulation model of a distribution network is proposed, and an index architecture for simulation models of different dimensions is established to support the synchronization and interaction between digital models and physical entities; a multi-rate parallel simulation algorithm and simulation platform for distribution networks are proposed, which improves the speed and accuracy of digital simulation of large-scale active distribution networks, provides necessary technical means and analysis basis for the verification and testing of distribution network control strategies with a high proportion of distributed power sources, breaks through the limitations of insufficient verification means and single decision-making basis of existing distribution network collaborative operation strategies, and provides auxiliary decision-making support for the safe, reliable and economical operation of distribution networks, thereby effectively tapping the power supply potential of distribution networks and improving the quality and level of power supply services.

[0384] Based on the foregoing embodiments, an embodiment of the present application provides a related device of a distribution network digital platform, which includes the modules included and the units included in each module, etc., and can be implemented by a processor in a computer device; of course, it can also be implemented by a specific logic circuit; in the implementation process, the processor can be a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP) or a field programmable gate array (FPGA), etc.

[0385] An embodiment of the present application provides a training device for a simulation model of a power distribution network. As shown in FIG15A , the device 1500 includes:

[0386] A first acquisition module 1501 is configured to acquire real-time data or historical data of the distribution network collected by sensors, instruments, or equipment;

[0387] A preprocessing module 1502 is configured to perform preprocessing operations on the real-time data or historical data of the distribution network to obtain a training sample set; the preprocessing operations include data cleaning, denoising, and format conversion;

[0388] A training module 1503 is configured to train the simulation model to be trained based on actual data during the actual operation of the distribution network as the annotation information and the training sample set to obtain a trained simulation model;

[0389] A determination module 1504 is configured to determine a multi-rate parallel simulation technology for a distribution network based on the different response speeds of different equipment components in the distribution network to changes in the system operating state;

[0390] The adjustment module 1505 is used to use the multi-rate parallel simulation technology of the power distribution network to adjust the simulation step size of the trained simulation model to obtain a final optimized model.

[0391] In some embodiments, the first acquisition module includes:

[0392] a first acquisition unit, configured to acquire typical devices in the distribution network under different scenarios, and topological relationships of the typical devices in the distribution network, as real-time data or historical data corresponding to the multi-resolution physical space model;

[0393] a second acquisition unit, configured to acquire, in different scenarios, inherent parameters of typical equipment of the distribution network, inherent attribute parameters of the distribution network operation process, inherent attribute parameters of the distribution network association relationship, and influencing variables of the data space model, as real-time data or historical data corresponding to the data space model;

[0394] a third acquisition unit, configured to acquire relevant knowledge of typical equipment in the distribution network as real-time data or historical data corresponding to the knowledge space model;

[0395] The fourth acquisition unit is configured to acquire information of typical equipment in the power distribution network as real-time data or historical data corresponding to the virtual space model.

[0396] In some embodiments, a simulation task is a simulation process based on the trained simulation model, and the simulation task includes at least two subtasks corresponding to at least two device elements in the distribution network, wherein the at least two subtasks include at least a first subtask and a second subtask; the determination module includes:

[0397] a first determining unit, configured to determine a simulation step size of the first subtask and a simulation step size of the second subtask based on different response speeds of device components of different subtasks in the distribution network to changes in system operating states; the simulation step size of the first subtask is a positive integer multiple of the step size of the second subtask;

[0398] The second determination unit is used to sample the simulation results within each simulation step of the first subtask through a triangular wave generator based on the sampling coefficient determined based on the current simulation moment, the simulation step of the first subtask, and the simulation step of the second subtask, so that the second subtask can update the state of the first subtask at the starting moment of each simulation step cycle, and determine the multi-rate parallel simulation technology of the distribution network.

[0399] The embodiment of the present application provides a device for constructing a distribution network digital platform, as shown in FIG15B , wherein the device 1510 includes:

[0400] A first construction module 1511 is configured to construct a multi-resolution physical space model of typical devices in the distribution network based on typical devices in different simulation scenarios in the distribution network;

[0401] A second construction module 1512 is configured to construct a data space model of typical devices in the distribution network based on the typical devices in the distribution network for which a model is to be constructed; the typical devices in the distribution network include lines, loads, converters, switchgear, and distributed power sources;

[0402] The third construction module 1513 is used to construct a knowledge space model of typical equipment in the distribution network based on the knowledge graph of the knowledge related to the model to be modeled in the distribution network;

[0403] A fourth construction module 1514 is configured to construct a virtual space model of typical devices in the distribution network based on the multi-resolution physical space model, the data space model, and the knowledge space model; wherein the virtual space model includes attribute information of the typical devices in the distribution network;

[0404] A calling module 1515 is configured to call different simulation models in the multi-dimensional digital simulation model of the distribution network by constructing an indexing architecture for each typical device in the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model;

[0405] The building module 1516 is used to build a cloud-based distribution network digital platform based on the multi-dimensional digital simulation model of the distribution network.

[0406] In some embodiments, the first building block includes:

[0407] An initialization unit, configured to initialize a multi-resolution physical model of a model to be constructed and set initial values ​​of power and voltage based on typical devices in different simulation scenarios in the distribution network; the multi-resolution physical model includes two physical models with different resolutions;

[0408] A construction unit, configured to construct a dynamic phasor network admittance matrix based on the multi-resolution physical model;

[0409] a detection unit, configured to detect whether a network topology of typical devices in the distribution network has changed;

[0410] a first modifying unit, configured to modify the dynamic phasor network admittance matrix when the network topology changes;

[0411] A first calculation unit is configured to calculate a voltage value at a current moment and perform a conversion from a time domain to a dynamic phasor domain based on the dynamic phasor network admittance matrix when the network topology does not change;

[0412] a third determining unit, configured to determine a dynamic phasor value of the current based on the initial values ​​of the power and voltage;

[0413] a fourth determining unit, configured to determine, based on the dynamic phasor value of the current and the current voltage value, respectively the dynamic phasor values ​​of the node voltages by using a reduced-order dynamic phasor equation;

[0414] a fifth determining unit, configured to determine the currents of all branches and perform conversion from the dynamic phasor domain of each order to the time domain based on the dynamic phasor values ​​of each order of the node voltage and the multi-resolution physical model;

[0415] The sixth determining unit is configured to determine a multi-resolution physical space model of a typical device in the distribution network when it is determined that the simulation end condition is met.

[0416] In some embodiments, the first building block further includes:

[0417] The shift unit is used to shift the simulation time forward by one time step;

[0418] a second modifying unit, configured to modify the dynamic phasor network admittance matrix when the network topology changes;

[0419] A second calculation unit is configured to calculate the voltage value at the current moment and perform conversion from the time domain to the dynamic phasor domain based on the dynamic phasor network admittance matrix when the network topology does not change;

[0420] a seventh determining unit, configured to determine, based on the dynamic phasor value of the current at a previous moment and the voltage value at the current moment, the dynamic phasor values ​​of each order of the node voltage by using a reduced-order dynamic phasor equation;

[0421] an eighth determining unit, configured to determine current values ​​of all branches based on the various-order dynamic phasor values ​​of the node voltage and the multi-resolution physical model, and perform conversion from the various-order dynamic phasor domain to the time domain;

[0422] The ninth determining unit is configured to determine a multi-resolution physical space model of a typical device in the distribution network when it is determined that the simulation end condition is met.

[0423] In some embodiments, the second building block includes:

[0424] A first construction unit is configured to construct a unified data model in the distribution network based on typical devices in the distribution network of the to-be-constructed model;

[0425] The second construction unit is used to construct a data-driven model in the distribution network based on the models to be constructed under different operating states and business scenarios in the distribution network; the data space model of typical equipment in the distribution network includes a unified data model in the distribution network and a data-driven model in the distribution network.

[0426] In some embodiments, the first building block comprises:

[0427] A first determining subunit is configured to determine a unified data model of typical devices in the distribution network based on the typical devices in the distribution network for which the model is to be constructed and inherent parameters of the typical devices in the distribution network;

[0428] a second determining subunit, configured to determine a unified data model of a distribution network operation process based on typical equipment in the distribution network of the to-be-built model and inherent attribute parameters of an operation process of the distribution network of the to-be-built model;

[0429] The third determination subunit is used to determine a unified data model of the association relationship in the distribution network based on the typical equipment in the distribution network of the model to be constructed and the inherent attribute parameters of the distribution network association relationship in the model to be constructed; the inherent attribute parameters of the distribution network association relationship include: equivalent power sources, conductors, transformer windings, and nodes connected between equivalent load endpoints; the unified data model in the distribution network includes a unified data model of the typical equipment in the distribution network, a unified data model of the distribution network operation process, and a unified data model of the association relationship in the distribution network.

[0430] In some embodiments, the second building block comprises:

[0431] a fourth determining subunit, configured to determine, based on the models to be constructed in different operating states and business scenarios in the distribution network, influencing variables corresponding to the models to be constructed in different operating states and business scenarios;

[0432] a partitioning subunit, configured to divide the influencing variables corresponding to the model to be constructed into a basic set and a candidate set based on a double recursive analysis method, taking the influencing variables corresponding to the model to be constructed as input variables and the predicted values ​​of the model to be constructed as output variables;

[0433] an output subunit, configured to output a predicted value of a model to be constructed based on a neural network algorithm, taking the basic set as an input variable;

[0434] The adjustment subunit is used to retrain the model by adjusting the influencing variables in the basic set and the candidate set when there is a deviation between the actual operation data of the distribution network and the predicted value of the model to be constructed; thus obtaining a verified distribution network data-driven model.

[0435] In some embodiments, the third building block includes:

[0436] a tenth determining unit, configured to determine, based on the relevant knowledge of the model to be constructed in the distribution network, the attributes and attribute value ranges of the basic concepts and sub-concepts of the model to be constructed, and the data source of the relevant knowledge of the model to be constructed;

[0437] A building unit, configured to build a basic framework of relevant knowledge of a model to be constructed in a distribution network based on the attributes and attribute value ranges of each basic concept and sub-concept;

[0438] an eleventh determining unit, configured to determine, based on a basic architecture of relevant knowledge of the model to be constructed in the distribution network, relevant words in the knowledge field of the model to be constructed in the distribution network and basic semantic associations of the relevant words in the knowledge field;

[0439] an identification unit, configured to identify entities to which the relevant vocabulary in the knowledge field of the model to be constructed in the distribution network belongs and classify the entities based on the relevant vocabulary in the knowledge field of the model to be constructed in the distribution network;

[0440] a twelfth determining unit, configured to determine, based on entities to which relevant vocabulary in the knowledge domain of the model to be constructed in the distribution network belongs, relationships of relevant knowledge of the model to be constructed in the distribution network; the relationships include inclusion relationships, application relationships, existence relationships, and discovery relationships;

[0441] a fusion unit, configured to perform knowledge fusion on the relevant knowledge of the model to be constructed in the distribution network based on a basic architecture of the relevant knowledge of the model to be constructed in the distribution network, a data source of the relevant knowledge of the model to be constructed, basic semantic associations of relevant words in the knowledge field of the model to be constructed in the distribution network, and relationships between the relevant knowledge of the model to be constructed in the distribution network;

[0442] a thirteenth determining unit, configured to determine a knowledge graph of typical devices in the distribution network based on the integrated relevant knowledge of the model to be constructed in the distribution network;

[0443] The third construction unit is used to construct a knowledge space model of typical equipment in the distribution network based on the knowledge graph of typical equipment in the distribution network.

[0444] The embodiment of the present application provides a processing device for a distribution network digital platform, as shown in FIG15C , wherein the device 1520 includes:

[0445] The processing module 1521 is used to process the real-time data or historical data of the distribution network through the data layer; the processing method includes data storage, data sharing, data analysis, and data processing;

[0446] A simulation module 1522 is configured to perform simulation at the service layer based on the simulation model of the model layer and the data of the data layer as input to obtain the simulation results; wherein the simulation results include data under different operating scenarios in the distribution network;

[0447] Application module 1523, configured to apply the simulation results through an application layer based on the simulation results; the applications include resource management, analysis and evaluation, business management and control, and decision-making and command;

[0448] A presentation module 1524 is configured to present the simulation results and the data processed by the application layer through a presentation layer;

[0449] The data sharing module 1525 is used for processing the distribution network digital platform and realizing data sharing between various modules of the distribution network digital platform through the API interface;

[0450] The protection module 1526 is used to protect the system stability and security of the platform through platform management and control, and the platform management and control includes system permissions, system configuration, and security management.

[0451] The description of the above device embodiment is similar to the description of the above method embodiment and has similar beneficial effects as the method embodiment. In some embodiments, the functions or modules included in the device provided in the embodiments of the present application can be used to perform the methods described in the above method embodiments. For technical details not disclosed in the device embodiments of the present application, please refer to the description of the method embodiments of the present application for understanding.

[0452] It should be noted that, in the embodiment of the present application, if the above method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for making a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk. In this way, the embodiment of the present application is not limited to any specific hardware, software or firmware, or any combination of hardware, software and firmware.

[0453] An embodiment of the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the program, some or all of the steps in the above method are implemented.

[0454] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements some or all of the steps in the above method. The computer-readable storage medium may be transient or non-transient.

[0455] An embodiment of the present application provides a computer program, including computer-readable code. When the computer-readable code is run in a computer device, a processor in the computer device executes some or all of the steps for implementing the above method.

[0456] An embodiment of the present application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and when the computer program is read and executed by a computer, implements some or all of the steps in the above method. The computer program product can be implemented specifically by hardware, software, or a combination thereof. In some embodiments, the computer program product is embodied as a computer storage medium. In other embodiments, the computer program product is embodied as a software product, such as a software development kit (SDK), etc.

[0457] It should be noted that the descriptions of the various embodiments above tend to emphasize the differences between the various embodiments, and their similarities or similarities can be referenced to each other. The descriptions of the above device, storage medium, computer program, and computer program product embodiments are similar to the descriptions of the above method embodiments and have similar beneficial effects as the method embodiments. For technical details not disclosed in the embodiments of the device, storage medium, computer program, and computer program product of this application, please refer to the description of the method embodiments of this application for understanding.

[0458] An embodiment of the present application provides a computer device, as shown in Figure 16, the hardware entity of the computer device 1600 includes: a processor 1601, a communication interface 1602 and a memory 1603, wherein: the processor 1601 generally controls the overall operation of the computer device 1600. The communication interface 1602 enables the computer device to communicate with other terminals or servers via a network. The memory 1603 is configured to store instructions and applications executable by the processor 1601, and can also cache data to be processed or processed by each module in the processor 1601 and the computer device 1600 (for example, image data, audio data, voice communication data and video communication data), which can be implemented by flash memory (FLASH) or random access memory (Random Access Memory, RAM). Data transmission between the processor 1601, the communication interface 1602 and the memory 1603 can be carried out through a bus 1604.

[0459] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned steps / processes does not mean the order of execution, and the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments.

[0460] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0461] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0462] The units described above as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, the functional units in the various embodiments of the present application may all be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0463] Those skilled in the art will understand that all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, and other media that can store program codes.

[0464] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology, 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 enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0465] The above is only an implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A method for training a simulation model of a distribution network, wherein: The method comprises: Acquiring real-time or historical data of the distribution network collected by sensors, instruments, and equipment; Performing preprocessing operations on the real-time data or historical data of the distribution network to obtain a training sample set; the preprocessing operations include data cleaning, denoising, and format conversion; Based on the actual data during the actual operation of the distribution network as the annotation information and the training samples, the simulation model to be trained is trained to obtain a trained simulation model; Determine a multi-rate parallel simulation technology for the distribution network based on the different response speeds of different equipment components in the distribution network to changes in the system operating state; The multi-rate parallel simulation technology of the distribution network is used to adjust the simulation step size of the trained simulation model to obtain a final optimized model.

2. The method according to claim 1, wherein: The simulation model includes one or more of a multi-resolution physical space model, a data space model, a knowledge space model, and a virtual space model; The real-time data or historical data of the distribution network collected by sensors, instruments, and equipment may be obtained by: Acquire typical devices in the distribution network under different scenarios, and topological relationships of the typical devices in the distribution network, as real-time data or historical data corresponding to the multi-resolution physical space model; Acquire inherent parameters of typical equipment of the distribution network under different scenarios, inherent attribute parameters of the distribution network operation process, inherent attribute parameters of the distribution network association relationship, and influencing variables of the data space model as real-time data or historical data corresponding to the data space model; Acquire relevant knowledge of typical equipment in the distribution network as real-time data or historical data corresponding to the knowledge space model; Information about typical equipment in the power distribution network is obtained as real-time data or historical data corresponding to the virtual space model.

3. The method according to claim 1 or 2, wherein: A simulation task is a simulation process based on the trained simulation model, wherein the simulation task includes at least two subtasks corresponding one-to-one to at least two device components in the distribution network, wherein the at least two subtasks include at least a first subtask and a second subtask; The method of determining a multi-rate parallel simulation technology for the distribution network based on different response speeds of different equipment components in the distribution network to changes in the system operating state includes: Determining a simulation step size for the first subtask and a simulation step size for the second subtask based on different response speeds of device components of different subtasks in the distribution network to changes in system operating states; the simulation step size for the first subtask is a positive integer multiple of the step size for the second subtask; Based on the current simulation moment, the simulation step of the first subtask, and the sampling coefficient determined by the simulation step of the second subtask, the simulation results within each simulation step of the first subtask are sampled through a triangular wave generator, so that the second subtask can update the state of the first subtask at the starting moment of each simulation step cycle, and determine the multi-rate parallel simulation technology of the distribution network.

4. A method for constructing a distribution network digital platform, wherein: The method comprises: Based on typical devices in different simulation scenarios in the distribution network, construct a multi-resolution physical space model of the typical devices in the distribution network; Based on the typical equipment in the distribution network to be modeled, a data space model of the typical equipment in the distribution network is constructed; the typical equipment in the distribution network includes lines, loads, converters, switchgears, and distributed power sources; Based on the knowledge graph of the knowledge related to the model to be modeled in the distribution network, a knowledge space model of typical equipment in the distribution network is constructed; Based on the multi-resolution physical space model, the data space model, and the knowledge space model, a virtual space model of typical equipment in the distribution network is constructed; wherein the virtual space model includes attribute information of the typical equipment in the distribution network; By constructing an indexing architecture for each typical device in the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model, calls are made between different simulation models in the multi-dimensional digital simulation model of the distribution network; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space modeling, the knowledge space model, and the virtual space model; Based on the multi-dimensional digital simulation model of the distribution network, a cloud-based distribution network digital platform is built.

5. The method according to claim 4, wherein: The multi-resolution physical space model of the typical equipment in the distribution network is constructed based on the typical equipment in different simulation scenarios in the distribution network, including: Initializing a multi-resolution physical model of the model to be constructed and setting initial values ​​of power and voltage based on typical devices in different simulation scenarios in the distribution network; the multi-resolution physical model includes two physical models with different resolutions; constructing a dynamic phasor network admittance matrix based on the multi-resolution physical model; Detecting whether a network topology of typical devices in the distribution network has changed; modifying the dynamic phasor network admittance matrix in the event of a change in the network topology; When the network topology does not change, the voltage value at the current moment is calculated based on the dynamic phasor network admittance matrix and the voltage is converted from the time domain to the dynamic phasor domain. Determining a dynamic phasor value of current based on the initial values ​​of power and voltage; Based on the dynamic phasor value of the current and the voltage value at the current moment, respectively determine the dynamic phasor values ​​of each order of the node voltage through a reduced-order dynamic phasor equation; Based on the various order dynamic phasor values ​​of the node voltage and the multi-resolution physical model, determining the currents of all branches and performing conversion from the various order dynamic phasor domains to the time domain; When it is determined that the simulation end condition is reached, a multi-resolution physical space model of typical equipment in the distribution network is determined.

6. The method according to claim 5, wherein: When the simulation end conditions are not met, including: Move the simulation forward by one time step; In the event that the network topology changes, modifying the dynamic phasor network admittance matrix; When the network topology does not change, the voltage value at the current moment is calculated based on the dynamic phasor network admittance matrix and the voltage is converted from the time domain to the dynamic phasor domain. Based on the dynamic phasor value of the current at the previous moment and the voltage value at the current moment, respectively determine the dynamic phasor values ​​of each order of the node voltage through a reduced-order dynamic phasor equation; Based on the dynamic phasor values ​​of each order of the node voltage and the multi-resolution physical model, the current values ​​of all branches are determined, and the dynamic phasor domain of each order is converted to the time domain; When it is determined that the simulation end condition is met, a multi-resolution physical space model of typical equipment in the distribution network is determined.

7. The method according to claim 4, wherein: The step of constructing a data space model of the typical equipment in the distribution network based on the typical equipment in the distribution network to be modeled comprises: Constructing a unified data model in the distribution network based on typical devices in the distribution network of the to-be-constructed model; Constructing a data-driven model in the distribution network based on the models to be constructed under different operating states and business scenarios in the distribution network; The data space model of typical equipment in the distribution network includes a unified data model in the distribution network and a data-driven model in the distribution network.

8. The method according to claim 7, wherein: The step of constructing a unified data model in the distribution network based on typical devices in the distribution network to be modeled includes: Determining a unified data model of the typical devices in the distribution network based on the typical devices in the distribution network for which the model is to be constructed and inherent parameters of the typical devices in the distribution network; Determining a unified data model for the distribution network operation process based on typical equipment in the distribution network to be modeled and inherent attribute parameters of the distribution network operation process to be modeled; Determine a unified data model of association relationships in the distribution network based on typical equipment in the distribution network to be modeled and inherent attribute parameters of the distribution network association relationships in the distribution network to be modeled; the inherent attribute parameters of the distribution network association relationships include: equivalent power sources, conductors, transformer windings, and nodes connected between equivalent load endpoints; The unified data model in the distribution network includes a unified data model of typical equipment in the distribution network, a unified data model of the distribution network operation process, and a unified data model of association relationships in the distribution network.

9. The method according to claim 7, wherein: The data-driven model in the distribution network is constructed based on the models to be constructed under different operating states and business scenarios in the distribution network, including: Based on the models to be constructed under different operating states and business scenarios in the distribution network, determining influencing variables corresponding to the models to be constructed under different operating states and business scenarios; Based on the double recursive analysis method, the influencing variables corresponding to the model to be constructed are used as input variables, the predicted values ​​of the model to be constructed are used as output variables, and the influencing variables corresponding to the model to be constructed are divided into a basic set and a candidate set; Based on the neural network algorithm, the basic set is used as an input variable and the predicted value of the model to be constructed is output; In the case that there is a deviation between the actual operating data of the distribution network and the predicted value of the model to be constructed, the model is retrained by adjusting the influencing variables in the basic set and the candidate set to obtain a verified distribution network data-driven model.

10. The method according to any one of claims 4 to 9, wherein: The method of constructing a knowledge space model of typical equipment in the distribution network based on relevant knowledge of the model to be constructed in the distribution network includes: Based on the relevant knowledge of the model to be constructed in the distribution network, determine the attributes and attribute value ranges of each basic concept and sub-concept of the model to be constructed, as well as the data source of the relevant knowledge of the model to be constructed; Based on the attributes and attribute value ranges of each basic concept and sub-concept, a basic framework of relevant knowledge of the model to be constructed in the distribution network is established; Determining, based on a basic architecture of relevant knowledge of the model to be constructed in the distribution network, relevant words in the knowledge domain of the model to be constructed in the distribution network and basic semantic associations of the relevant words in the knowledge domain; Based on relevant words in the knowledge field of the model to be constructed in the distribution network, identifying entities to which the relevant words in the knowledge field of the model to be constructed in the distribution network belong and classifying the entities; Determining relationships among relevant knowledge of the model to be constructed in the distribution network based on entities to which relevant vocabulary in the knowledge domain of the model to be constructed in the distribution network belongs; the relationships include inclusion relationships, application relationships, existence relationships, and discovery relationships; Based on the basic structure of the knowledge related to the model to be constructed in the distribution network, the data source of the knowledge related to the model to be constructed, the basic language of the relevant vocabulary in the knowledge field of the model to be constructed in the distribution network and performing knowledge fusion on the relevant knowledge of the model to be constructed in the distribution network; Determine a knowledge graph of typical devices in the distribution network based on the integrated knowledge of the model to be constructed in the distribution network; Based on the knowledge graph of typical equipment in the distribution network, a knowledge space model of typical equipment in the distribution network is constructed.

11. A method for processing a distribution network digital platform, the method comprising: Processing the real-time data or historical data of the distribution network through the data layer; The processing methods include data storage, data sharing, data analysis, and data processing; Based on the simulation model of the model layer, the data of the data layer is used as input to perform simulation at the service layer to obtain the simulation results; wherein the simulation results include data under different operating scenarios in the distribution network; Based on the simulation results, the simulation results are applied through the application layer; the applications include resource management, analysis and evaluation, business control, and decision-making and command; The simulation results and the data processed by the application layer are displayed through the display layer; Data sharing between various modules of the distribution network digital platform is achieved through the API interface; Through platform management and control, the system stability and security of the platform are protected. The platform management and control includes system permissions, system configuration, and security management.

12. A training device for a simulation model of a distribution network, wherein: include: A first acquisition module is used to acquire real-time data or historical data of the distribution network collected by sensors, instruments, and equipment; A preprocessing module, configured to perform preprocessing operations on the real-time data or historical data of the distribution network to obtain a training sample set; The preprocessing operations include data cleaning, denoising, and format conversion; A training module, configured to train the simulation model to be trained based on actual data during the actual operation of the distribution network as the annotation information and the training sample set, to obtain a trained simulation model; a determination module for determining a multi-rate parallel simulation technology for a distribution network based on different response speeds of different equipment components in the distribution network to changes in system operating states; The adjustment module is used to use the multi-rate parallel simulation technology of the distribution network to adjust the simulation step size of the trained simulation model to obtain a final optimized model.

13. A device for constructing a distribution network digital platform, wherein: include: A first construction module is configured to construct a multi-resolution physical space model of typical equipment in the distribution network based on typical equipment in different simulation scenarios in the distribution network; A second construction module is configured to construct a data space model of typical equipment in the distribution network based on typical equipment in the distribution network for which a model is to be constructed; the typical equipment in the distribution network includes lines, loads, converters, switchgear, and distributed power sources; A third construction module is configured to construct a knowledge space model of typical equipment in the distribution network based on a knowledge graph of knowledge related to the model to be modeled in the distribution network; a fourth construction module, configured to construct a virtual space model of typical equipment in the distribution network based on the multi-resolution physical space model, the data space model, and the knowledge space model; wherein the virtual space model includes attribute information of the typical equipment in the distribution network; a calling module, configured to call different simulation models in the multi-dimensional digital simulation model of the distribution network by constructing an index architecture for each typical device in the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model; wherein the multi-dimensional digital simulation model of the distribution network includes the multi-resolution physical space model, the data space model, the knowledge space model, and the virtual space model; A building module is used to build a cloud-based distribution network digital platform based on the multi-dimensional digital simulation model of the distribution network.

14. A processing device for a distribution network digital platform, wherein: include: A processing module, configured to process the real-time data or historical data of the distribution network through the data layer; The processing methods include data storage, data sharing, data analysis, and data processing; The simulation module is used to simulate the service layer based on the simulation model of the model layer and take the data of the data layer as input to obtain the simulation result; wherein the simulation result The results include data under different operating scenarios in the distribution network; An application module, configured to apply the simulation results through an application layer based on the simulation results; the applications include resource management, analysis and evaluation, business control, and decision-making and command; A display module is used to display the simulation results and the data processed by the application layer through the display layer; A data sharing module is used to realize data sharing between various modules of the distribution network digital platform through an API interface; The protection module is used to protect the system stability and security of the platform through platform management and control. The platform management and control includes system permissions, system configuration, and security management.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, wherein: When the processor executes the program, the steps of the method according to any one of claims 1 to 11 are implemented.

16. A computer-readable storage medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

17. A computer program product comprising a computer program or instructions, wherein: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

Images