Power grid resource comprehensive evaluation and analysis system based on big data and machine learning

Abstract

The invention relates to the technical field of power grid systems, in particular to a power grid resource comprehensive evaluation and analysis system based on big data and machine learning, and discloses a power grid digital twin unit, an N-1 fault scene evaluation unit, an N-2 fault scene evaluation unit and a resource comprehensive evaluation unit. Real-time synchronization of digital twin bodies of a power grid system is ensured, scene performance of nodes and elements in each scene is deeply quantized through N-1 and N-2 multi-scene evaluation and in combination with a machine learning means, and power flow transfer and network splitting risks caused by faults are accurately captured through coupling calculation of super-lost absolute differential load rate and topological unbalance number. And the evaluation result is ensured to cover the potential cascading failure path of the whole network, the key degree of elements in cascading failures is revealed through difference equalization of advanced combination evaluation values, and disassembly evaluation is performed on resources of a power grid system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid system, more particularly, it relates to a power grid resource comprehensive evaluation analysis system based on big data and machine learning. BACKGROUND

[0002] With the development of smart grid and new power system, power grid resource evaluation has been transformed from the traditional experience-driven mode to data-driven and model-driven. The current industry mainstream solution constructs a power grid virtual model through digital twin technology, and analyzes the resource state in combination with fault simulation. However, the existing technology still has significant technical bottlenecks in dealing with the complexity of power grid topology, composite fault risk and multi-dimensional evaluation demand: N-1 fault evaluation limitation: the existing N-1 analysis only focuses on the over-limit of power flow after a single element outage, and does not quantify the cascading effect of node vulnerability and power transfer. N-2 composite fault analysis is missing: most systems do not cover the simulation of two-element simultaneous outage scenarios, or only conduct limited analysis through random sampling, which cannot systematically evaluate the superimposed risk of topology fracture and power imbalance. Node and element evaluation is separated: in the existing solution, node voltage over-limit analysis and element load rate evaluation are carried out independently, and the coupling model of topology connection relationship and power flow is not established.

[0003] Based on the above, the present application provides a power grid resource comprehensive evaluation analysis system based on big data and machine learning. SUMMARY

[0004] In view of the deficiencies in the prior art, the purpose of the present application is to provide a power grid resource comprehensive evaluation analysis system based on big data and machine learning.

[0005] To achieve the above purpose, the present application provides the following technical solutions: The power grid resource comprehensive evaluation analysis system based on big data and machine learning comprises a power grid digital twin unit for constructing a power grid digital twin system; an N-1 fault scenario evaluation unit for simulating the power grid digital twin system under various N-1 fault scenarios at regular intervals, and determining the primary resource evaluation value of the power grid digital twin system under various N-1 fault scenarios; an N-2 fault scenario evaluation unit for simulating the power grid digital twin system under various N-2 fault scenarios, and determining the advanced resource evaluation value of the power grid digital twin system under various N-2 fault scenarios; a resource comprehensive evaluation unit for determining the basic resource evaluation value of each element, further determining the comprehensive resource evaluation value of the power grid resource, and evaluating the power grid resource according to the comprehensive resource evaluation value.

[0006] Further, the determination manner of the primary resource evaluation value of the power grid digital twin system under an N-1 fault scenario: selecting an N-1 fault scenario, adding the fault parameters corresponding to the N-1 fault scenario in the power grid digital twin system, triggering power flow calculation in the power grid digital twin system, first, obtaining the node evaluation value, further, obtaining the absolute difference load rate of overloading and the number of topology imbalance, finally, multiplying the node evaluation value, the absolute difference load rate of overloading and the number of topology imbalance to calculate the primary resource evaluation value of the power grid digital twin system under the N-1 fault scenario.

[0007] Further, the obtaining manner of the node evaluation value: obtaining the characteristic parameters of each node in the power grid digital twin system, integrating the characteristic parameters of each node into a node feature set in a set manner, importing the node feature set into a node machine learning model, and the node machine learning model outputs the node evaluation value.

[0008] Further, the obtaining manner of the number of topology imbalance: obtaining the load rate of each element in the power grid digital twin system, when the load rate of an element is higher than the upper threshold load rate, the corresponding element is marked as an overloaded element, when the load rate of an element is lower than the lower threshold load rate, the corresponding element is marked as a power loss element, each overloaded element and each power loss element in the power grid digital twin system are matched one by one, and when the matched overloaded element and power loss element are in a topology connection state in the power grid digital twin system, the number of topology imbalance is increased by one.

[0009] Further, the obtaining manner of the absolute difference load rate of overloading and power loss: summing and averaging the load rates of all overloaded elements to calculate the average overload rate, summing and averaging the load rates of all power loss elements to calculate the average power loss rate, and calculating the absolute difference between the average overload rate and the average power loss rate to obtain the absolute difference load rate of overloading and power loss.

[0010] Further, the determination manner of the basic resource evaluation value of the element: selecting an element, obtaining the primary resource evaluation value of the element under the corresponding N-1 fault scenario, synchronously obtaining the advanced resource evaluation values of the element under all N-2 fault scenarios, further calculating a plurality of advanced combination evaluation values, summing and averaging all advanced combination evaluation values to calculate the basic resource evaluation value of the element.

[0011] Further, the calculation manner of the advanced combination evaluation value: matching the primary resource evaluation value with each advanced resource evaluation value one by one, calculating the difference between the matched advanced resource evaluation value and the primary resource evaluation value to obtain the advanced combination evaluation value.

[0012] Further, the determination method of the comprehensive resource evaluation value of the power grid resource: the first-order resource evaluation value of the power grid digital twin system under a kind of N-1 fault scene is summed and average value is calculated, the average first-order resource evaluation value is calculated, the advanced resource evaluation value of the power grid digital twin system under the N-2 fault scene is summed and average value is calculated, the average advanced resource evaluation value is calculated, the basic resource evaluation value of each element is summed and average value is calculated, the average basic resource evaluation value is calculated, the average first-order resource evaluation value, the average advanced resource evaluation value and the average basic resource evaluation value are calculated to calculate the comprehensive resource evaluation value of the power grid resource.

[0013] Compared with the prior art, the present application has the following beneficial effects: The system of the present application combines digital twin modeling technology to ensure real-time synchronization of the power grid system digital twin, through N-1, N-2 multi-scenario evaluation, combined with machine learning means, the scene performance of the nodes and elements is quantified in depth, through the coupling calculation of the super loss absolute difference load rate and the number of topology imbalance, the risk of power flow transfer and network splitting caused by fault is accurately captured, the evaluation result covers the potential chain failure path of the whole network, through the advanced combination evaluation value difference mean value, the key degree of the element in the chain failure is revealed, the resources of the power grid system are evaluated in a disassembled manner, and the whole chain quantitative analysis from equipment state to system risk is realized. BRIEF DESCRIPTION OF DRAWINGS

[0014] Fig. 1 It is a structural schematic diagram of the power grid resource comprehensive evaluation analysis system based on big data and machine learning. Fig. 2 It is a flow chart of the power grid resource comprehensive evaluation analysis system based on big data and machine learning. Fig. 3 It is the determination process of the first-order resource evaluation value under the N-1 fault scene. DETAILED DESCRIPTION

[0015] Reference Figs. 1 to 3 The power grid resource comprehensive evaluation analysis system based on big data and machine learning includes a power grid digital twin unit, an N-1 fault scene evaluation unit, an N-2 fault scene evaluation unit and a resource comprehensive evaluation unit.

[0016] The power grid digital twin unit constructs a power grid digital twin system: the first step of construction is to collect multi-source data and build a digital twin body model (multi-source data includes geometric data, physical parameters, geometric data includes geometric data of power transmission lines, geometric data of substations, etc., geometric data of power transmission lines: LiDAR point cloud (flight height 100-200m, point density≥100 points / m²) + unmanned aerial vehicle oblique photography (resolution≤2cm / pixel), geometric data of substation: BIM model (constructed by Revit software, accuracy up to LOD400, including bolts, gaskets and other components) + indoor laser scanning (such as Faro Focus S70, accuracy±2mm), physical parameters include equipment nameplate data (rated voltage / current, impedance parameters), design drawings (such as transformer winding connection diagram, cable layout diagram), historical test data (such as withstand voltage test report, winding deformation test results), digital twin body modeling includes spatial modeling, physical modeling and topological modeling, spatial modeling: using Unity / UE engine to build a three-dimensional scene, power transmission lines are generated by point cloud data to generate NURBS surface models of towers and conductors, and substation equipment is converted by BIM model light weight (triangle face number is compressed to 10% of the original model), physical modeling: based on finite element analysis (FEA) to establish device mechanism model: transformer: coupled electromagnetic-thermal-mechanical field model, using COMSOL Multiphysics solver, winding eddy current loss calculation error<3%; power transmission line: icing-dancing dynamics model based on ANSYS, wind speed-conductor vibration frequency fitting degree R²>0.95; spatial modeling: construct power grid node-branch correlation matrix (N×B matrix, N is the number of nodes, B is the number of branches), integrate GIS geographic information, label line length, elevation, soil resistivity and other parameters); Step 2: Sensor network deployment, data transmission, and dynamic twin updating (Multi-modal sensors are configured at key equipment (such as main transformers, 220kV and above circuit breakers): Vibration: PCB 356A16 three-axis acceleration sensor (frequency range 0.2Hz-20kHz); Partial discharge: Ultra-high frequency (UHF) sensor (band 300MHz-3GHz) + ultrasonic sensor (20kHz-100kHz); Temperature: Fiber Bragg Grating sensor (FBG), distributed temperature measurement accuracy ±0.5℃, spatial resolution 1m. Transmission line: 5G RedCap low-power module (bandwidth 1-2MHz) is used to transmit vibration / icing data, with a delay of <50ms; Substation: Real-time monitoring data is transmitted through industrial Ethernet (IEEE 802.3), and OPC UA protocol is used to realize data interconnection of different manufacturers' equipment. Real-time data stream processing: Flink framework is used to perform real-time statistics on partial discharge pulse sequence (such as updating PRPD spectrum every second). Extended Kalman filter (EKF) algorithm is used to fuse measured data and physical model prediction values to correct model parameters. For example, transformer winding thermal resistance parameters are automatically calibrated every week to ensure temperature prediction error <1.5℃.

[0017] N-1 fault scenario evaluation unit, which periodically simulates various N-1 fault scenarios for the power grid digital twin system (N-1 fault scenario is a scenario in which any single component (such as a line or a transformer) in the power grid digital twin system is shut down, and different N-1 fault scenarios do not target the same component), thereby determining the initial resource evaluation value of the power grid digital twin system under various N-1 fault scenarios.

[0018] The determination method of the primary resource evaluation value of the power grid digital twin system under a kind of N-1 fault scene: select a kind of N-1 fault scene, add the fault parameter corresponding to the N-1 fault scene in the power grid digital twin system (such as the line L1 fault corresponding to the N-1 fault scene, set the circuit breaker state of the line L1 to "trip" and disconnect the electrical connection of its two end nodes in the power grid digital twin system), trigger the power flow calculation of the power grid digital twin system (redistribute the power originally flowing through the line L1), obtain the characteristic parameters of each node in the power grid digital twin system (the node is the intersection point of power transmission in the power grid digital twin system, such as the connection point of the line / transformer, the generator outlet node, the load access node, the transformer high and low voltage side node), integrate the characteristic parameters of each node into a node characteristic set in the form of a set (such as, first organize the characteristic parameters of each node according to certain rules to form a structured data set, the format of the data set is | node number | voltage (kV) | current (A) | active power (MW) | reactive power (Mvar) | load level (MW) |, for example: | node 1 | 220 | 500 | 100 | 30 | 100 |, | node 2 | 110 | 300 | 60 | 20 | 60 |), import the node characteristic set into the node machine learning model, the node machine learning model exports the node evaluation value, obtains the load rate of each element (except the fault element) in the power grid digital twin system, when the load rate of an element is higher than the upper threshold load rate (the upper threshold load rate is higher than the lower threshold load rate, and the upper threshold load rate and the lower threshold load rate are set based on industry standards), the corresponding element is marked as an overloaded element, when the load rate of an element is lower than the lower threshold load rate, the corresponding element is marked as a power loss element (the remaining cases do not need to be marked, which means the load rate of the element is between the upper threshold load rate and the lower threshold load rate), each overloaded element and each power loss element in the power grid digital twin system are matched one by one, when the matched overloaded element and power loss element are in a topological connection state in the power grid digital twin system, the topological imbalance number is increased by one, the load rates of all overloaded elements are summed and averaged to calculate the average overload rate, the load rates of all power loss elements are summed and averaged to calculate the average power loss rate, the average overload rate and the average power loss rate are calculated by absolute difference value to calculate the absolute difference overload rate, finally, the node evaluation value, the absolute difference overload rate and the (statistical) topological imbalance number are multiplied (the multiplication is a dimensionless calculation, the node evaluation value*the absolute difference overload rate*the topological imbalance number), and the primary resource evaluation value of the power grid digital twin system under the N-1 fault scene is calculated.

[0019] The construction steps of the node machine learning model are as follows: a plurality of node feature sets are collected, a graph neural network model (GNN) is built, the node feature sets are used as basic data, the graph neural network model (GNN) is trained, and a node evaluation value is assigned to each node feature set during the training process (the node feature set is subjected to feature standardization processing). The value range of the node evaluation value is set to be between 1 and 40 (a physical basis of the node evaluation value is established, for example, node evaluation value = f(voltage out-of-limit degree, fault propagation probability), wherein f is a weighted summation function, and the weight can be calibrated through expert knowledge or historical fault data). The size of the node evaluation value has a clear meaning, and the larger the value is, the more vulnerable the node in the power grid digital twin system under the fault scenario is. Then, the plurality of node feature sets are divided into a training set, a validation set and a test set according to a specific proportion, and the specific division proportion is determined to be 70:15:15 (to ensure the generalization ability of the model on unseen data). The parameters are dynamically adjusted through the validation set, hyperparameter tuning (such as learning rate, layer search) and overfitting prevention (such as Dropout, L2 regularization), and finally the node machine learning model is constructed.

[0020] The N-2 fault scenario evaluation unit further simulates the power grid digital twin system under various N-2 fault scenarios (N-2 fault scenarios are scenarios in which any two elements in the power grid digital twin system are out of operation, and the elements corresponding to different N-2 fault scenarios are not repeated). The advanced resource evaluation value of the power grid digital twin system under various N-2 fault scenarios is determined.

[0021] The determination method of the advanced resource evaluation value of the power grid digital twin system under a kind of N-2 fault scenario is as follows: a kind of N-2 fault scenario is selected, and the fault parameters corresponding to the N-2 fault scenario are added to the power grid digital twin system (for example, if the N-2 fault scenario corresponds to line L1 and line L2 fault, the circuit breaker state of line L1 and line L2 in the power grid digital twin system is set to "off", and the electrical connection between the nodes at both ends of line L1 and line L2 is disconnected). The power grid digital twin system triggers power flow calculation (re-distributes the power originally flowing through line L1 and line L2), and obtains the load rate of each element in the power grid digital twin system (except the two fault elements). The subsequent method is consistent with the determination method of the primary resource evaluation value of the power grid digital twin system under a kind of N-1 fault scenario, that is, the node evaluation value, the absolute difference load rate and the number of topology imbalance are calculated first, and then the product of the node evaluation value, the absolute difference load rate and the number of topology imbalance is calculated (the product calculation is a dimensionless calculation), and the advanced resource evaluation value of the power grid digital twin system under the N-2 fault scenario is calculated.

[0022] The resource comprehensive evaluation unit then determines a basic resource evaluation value of each element, further determines a comprehensive resource evaluation value of the power grid resource, and evaluates the power grid resource according to the comprehensive resource evaluation value (specifically, a plurality of levels can be divided according to the range of the comprehensive resource evaluation value, and the specific levels are not limited in the specific embodiment, but are only used for illustration, for example, five levels are divided according to the range of the comprehensive resource evaluation value, and the higher the level, the weaker the comprehensive state of the power grid resource, and the more the power grid resource needs to be managed).

[0023] The determination method of the basic resource evaluation value of the element: select an element, obtain the primary resource evaluation value of the element under the N-1 fault scene (for example, select the line L1, and obtain the primary resource evaluation value of the line L1 fault under the N-1 fault scene), and synchronously obtain the advanced resource evaluation value of the element under all N-2 fault scenes (for example, select the line L1, and obtain the advanced resource evaluation value of the line L1 fault under all N-2 fault scenes, for example, the N-2 fault scene of the line L1 and the line L2, the N-2 fault scene of the line L1 and the line L3, and the N-2 fault scene of the line L1 and the line L4), match the primary resource evaluation value with each advanced resource evaluation value one by one, calculate the difference between the matched advanced resource evaluation value and the primary resource evaluation value, calculate the advanced combination evaluation value, and calculate the average of the sum of all advanced combination evaluation values to obtain the basic resource evaluation value of the element.

[0024] The determination method of the comprehensive resource evaluation value of the power grid resource: perform average sum calculation on the primary resource evaluation value of the power grid digital twin system under a kind of N-1 fault scene to calculate the average primary resource evaluation value EA, perform average sum calculation on the advanced resource evaluation value of the power grid digital twin system under the N-2 fault scene to calculate the average advanced resource evaluation value EB, perform average sum calculation on the basic resource evaluation value of each element to calculate the average basic resource evaluation value EC, and use the formula to calculate the comprehensive resource evaluation value of the power grid resource; dy1 is the primary correction coefficient, dy2 is the advanced correction coefficient, and dy3 is the basic correction coefficient.

[0025] The above system combines digital twin modeling technology to ensure real-time synchronization of the power grid system digital twin, performs N-1 and N-2 multi-scene evaluation, combines machine learning means, quantifies the scene performance of nodes and elements in each scene in depth, accurately captures the risk of power flow transfer and network splitting caused by faults through coupling calculation of the super loss absolute difference rate and the number of topology imbalance, ensures that the evaluation result covers the potential cascading failure path of the whole network, reveals the key degree of the element in the cascading failure through average value calculation of the difference of the advanced combination evaluation value, and performs disassembled evaluation on the resources of the power grid system, thereby realizing whole-chain quantitative analysis from the equipment state to the system risk.

[0026] The above formulas are all dimensionless values calculated, and the preset parameters in the formulas are set by a person skilled in the art according to actual conditions.

[0027] The above embodiments can be implemented wholly or partially by software, hardware, firmware, or any other combination. When implemented by software, the above embodiments can be implemented wholly or partially in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0028] It should be understood that the size of the sequence number of each process described above in various embodiments of the present application does not mean the order of execution, and the execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0029] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0030] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device, and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0031] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the division of the above-described device embodiments is only a logical function division, and there can be another division manner for actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, or the among different units, can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0032] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the parts of the technical solutions that make contributions to the prior art, or the parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and various other media that can store program codes.

[0033] The above describes only the specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A comprehensive power grid resource assessment and analysis system based on big data and machine learning, characterized in that: include A power grid digital twin unit is used to construct a power grid digital twin system. The N-1 fault scenario assessment unit periodically simulates various N-1 fault scenarios of the power grid digital twin system, thereby determining the initial resource assessment values ​​of the power grid digital twin system under various N-1 fault scenarios. The N-2 fault scenario assessment unit simulates various N-2 fault scenarios for the power grid digital twin system and determines the advanced resource assessment values ​​of the power grid digital twin system under various N-2 fault scenarios. The resource comprehensive assessment unit determines the basic resource assessment value of each component, further determines the comprehensive resource assessment value of the power grid resources, and assesses the power grid resources based on the comprehensive resource assessment value.

2. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 1, characterized in that, The method for determining the initial resource assessment value of the power grid digital twin system under an N-1 fault scenario is as follows: Select an N-1 fault scenario, add the corresponding fault parameters to the power grid digital twin system, trigger power flow calculation in the power grid digital twin system, first obtain the node assessment value, then obtain the absolute differential load rate and the number of topological imbalances, and finally multiply the node assessment value, the absolute differential load rate, and the number of topological imbalances to calculate the initial resource assessment value of the power grid digital twin system under the N-1 fault scenario.

3. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 2, characterized in that, Method for obtaining node evaluation values: Obtain the feature parameters of each node in the power grid digital twin system, integrate the feature parameters of each node into a node feature set, import the node feature set into the node machine learning model, and derive the node evaluation value from the node machine learning model.

4. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 2, characterized in that, The method for obtaining the number of topology imbalances is as follows: Obtain the load rate of each other component in the power grid digital twin system. When the load rate of a component is higher than the upper threshold load rate, the corresponding component is marked as an overloaded component. When the load rate of a component is lower than the lower threshold load rate, the corresponding component is marked as a power-off component. Match each overloaded component and each power-off component in the power grid digital twin system one by one. When the matched overloaded component and power-off component are in a topological connection state in the power grid digital twin system, the number of topology imbalances is increased by one.

5. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 4, characterized in that, The method for obtaining the absolute differential load rate of overload and underload is as follows: the average overload rate is calculated by summing and averaging the load rates of all overloaded components, the average underload rate is calculated by summing and averaging the load rates of all underloaded components, and the absolute differential load rate of overload and underload is calculated by calculating the absolute difference between the average overload rate and the average underload rate.

6. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 1, characterized in that, The method for determining the basic resource evaluation value of a component is as follows: Select a component, obtain the initial resource evaluation value of the component under the N-1 fault scenario, simultaneously obtain the advanced resource evaluation values ​​of the component under all N-2 fault scenarios, further calculate multiple advanced combination evaluation values, sum and average all advanced combination evaluation values, and calculate the basic resource evaluation value of the component.

7. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 6, characterized in that, The calculation method for the advanced combination evaluation value is as follows: the initial resource evaluation value is matched with each advanced resource evaluation value one by one, and the difference between the matched advanced resource evaluation value and the initial resource evaluation value is calculated to obtain the advanced combination evaluation value.

8. The power grid resource comprehensive assessment and analysis system based on big data and machine learning according to claim 1, characterized in that, The method for determining the comprehensive resource assessment value of power grid resources is as follows: the average initial resource assessment value is calculated by summing and averaging the initial resource assessment values ​​of the power grid digital twin system under an N-1 fault scenario; the average advanced resource assessment value is calculated by summing and averaging the advanced resource assessment values ​​of the power grid digital twin system under an N-2 fault scenario; the average basic resource assessment value is calculated by summing and averaging the basic resource assessment values ​​of each component; and the comprehensive resource assessment value of the power grid resources is calculated based on the average initial resource assessment value, the average advanced resource assessment value, and the average basic resource assessment value.

Images