Evaluation method based on running reliability intelligent evaluation model and related device

Abstract

The application provides an evaluation method based on an operation reliability intelligent evaluation model and related devices. System data of a power system is obtained, and after a minimum load shedding model is established according to the system data of the power system, the inherent mode of the minimum load shedding model is determined, and a robust learning loss function is constructed based on the inherent mode of the minimum load shedding model. Finally, strong generalization learning is performed according to the minimum load shedding neural network, the robust learning loss function and load disturbance, and an operation reliability intelligent evaluation model is obtained. The system data of the power system is input into the operation reliability intelligent evaluation model, and a result capable of evaluating the operation reliability of the system can be obtained, effectively improving the efficiency and accuracy of the evaluation result.

Description

Technical Field

[0001] This application relates to the field of power systems and their automation technology, and in particular to an evaluation method and related apparatus based on an intelligent evaluation model for operational reliability. Background Technology

[0002] Operational reliability assessment is an important tool for assessing and providing early warning of operational risks in power systems. Against the backdrop of the vigorous development of renewable energy worldwide, power systems face significant uncertainties due to the intermittent and random nature of wind and solar power, necessitating real-time operational reliability assessments.

[0003] However, existing numerical reliability assessment methods require iteratively solving minimum load shedding optimization problems under numerous system states, resulting in significant computational bottlenecks and difficulty meeting current assessment time requirements. Traditional numerical reliability assessment methods can be broadly categorized into analytical and simulation methods. Analytical methods directly derive analytical formulas for reliability calculations based on probabilistic models of uncertainties and failures in critical equipment. Analytical methods perform well in small power systems with fewer system states and smaller computational scales. However, in large-scale power systems involving complex operating conditions and a large number of critical events, analytical methods can become extremely complex and may even be unable to account for certain operating conditions. Simulation methods, on the other hand, perform simulation experiments on different events and then statistically calculate reliability indicators, thus avoiding the limitations of analytical methods in adapting to complex operating conditions and large-scale power systems. However, the real-time application of simulation methods also suffers from the cumulative computational bottleneck of solving numerous system states, making it impossible to complete the assessment within a 5-minute operating time window and hindering its application in online power system assessment. Therefore, researchers have proposed a data-driven assessment method based on neural networks.

[0004] Data-driven assessment methods, leveraging the advantages of neural networks—no iteration required and high computational speed—are an effective way to overcome the computational bottleneck of online assessment using simulation methods. However, these methods require training the neural network with a large number of load shedding samples, and then using this trained network to replace the time-consuming minimum load shedding calculations of the simulation method. Real-world power systems are highly resilient, and load shedding scenarios typically occur infrequently, leading to significant imbalances in the load shedding samples. The neural network trained based on this imbalance exhibits bias, resulting in inaccuracies in the data-driven assessment method. Since the load shedding samples satisfy the minimum load shedding model, how to efficiently and directionally expand the load shedding samples using this model, avoiding large-scale random sampling, is a pressing issue that needs to be addressed. Summary of the Invention

[0005] In view of this, this application provides an evaluation method and related apparatus based on an intelligent evaluation model for operational reliability, which effectively improves the performance of the operational reliability evaluation neural network.

[0006] The first aspect of this application provides an evaluation method based on an intelligent operational reliability evaluation model, including:

[0007] Obtain system data from the power system;

[0008] A minimum load shedding model is established based on the system data of the power system.

[0009] Determine the inherent pattern of the minimum load shear model;

[0010] Based on the inherent pattern of the minimum load shedding model, a robust learning loss function is constructed.

[0011] Based on the minimum load shedding model, the robust learning loss function, and the load disturbance, a strong generalization learning method is used to obtain an intelligent assessment model for operational reliability.

[0012] The system data of the power system is input into the intelligent assessment model for operational reliability to obtain results that can be used to assess the operational reliability of the system.

[0013] Optionally, establishing the minimum load shedding model based on the system data of the power system includes:

[0014] Based on the bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum gradeability, generator set, bus node set and branch set, a minimum load shedding model is constructed.

[0015] Optionally, constructing a robust learning loss function based on the inherent patterns of the minimum shear load model includes:

[0016] Based on the output features of the minimum load shedding neural network, the corresponding labels, the sample size of the training set, and the output dimension, an initial loss function is constructed.

[0017] Based on the initial loss function, disturbance rate, load input, load reduction of output features, and weight coefficients for learning disturbance samples, a robust learning loss function is constructed.

[0018] Optionally, the step of obtaining an intelligent operational reliability assessment model by performing strong generalization learning based on the minimum load shedding model, the robust learning loss function, and load disturbances includes:

[0019] Receive model training configuration information; wherein, the model training configuration information includes at least the training set and the target number of iterations;

[0020] Initialize the parameters of the minimum shear load neural network to obtain the initial neural network for calculating the minimum shear load;

[0021] Randomly shuffle the training set;

[0022] Obtain the original sample input features from the shuffled training set;

[0023] The original sample input features are input into the initial minimum shear load neural network, and the output is the initial sample predicted minimum shear load.

[0024] A load disturbance is randomly generated within the target interval;

[0025] The load disturbance is added to the training set to obtain the disturbance training set;

[0026] Obtain input features of perturbation samples from the perturbation training set;

[0027] The perturbation sample input features are input into the initial minimum load shedding neural network, and the output is the predicted minimum load shedding amount of the perturbation sample;

[0028] The loss value of the minimum shear load for the initial sample and the minimum shear load for the perturbed sample is calculated using the robust learning loss function;

[0029] The gradient of the neural network parameters is calculated using backpropagation based on the loss value.

[0030] The optimizer is used to update the neural network parameters to obtain the trained minimum shear load neural network;

[0031] If the current iteration count reaches the target iteration count, then the trained minimum load shedding neural network will be used as the intelligent evaluation model for operational reliability.

[0032] If the current iteration count has not reached the target iteration count, the trained minimum load shedding neural network is used as the new initial minimum load shedding neural network, and the random shuffling training set step is returned to be executed.

[0033] A second aspect of this application provides an evaluation device based on an intelligent operational reliability evaluation model, comprising:

[0034] The system data acquisition unit is used to acquire system data of the power system;

[0035] The model building unit is used to build a minimum load shedding model based on the system data of the power system.

[0036] An inherent mode determination unit is used to determine the inherent mode of the minimum load shearing model;

[0037] The loss function construction unit is used to construct a robust learning loss function based on the inherent patterns of the minimum load shedding model.

[0038] A strong generalization learning unit is used to perform strong generalization learning based on the minimum load shedding model, the robust learning loss function, and the load disturbance to obtain an intelligent evaluation model for operational reliability.

[0039] The evaluation unit is used to input system data of the power system into the intelligent evaluation model for operational reliability, and obtain results that can be used to evaluate the operational reliability of the system.

[0040] Optionally, the model building unit includes:

[0041] The model establishes sub-units to construct a minimum load shedding model based on bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum ramp rate, generator set, bus node set, and branch set.

[0042] Optionally, the loss function construction unit includes:

[0043] The initial sample loss function construction sub-unit is used to construct the initial loss function based on the output features of the minimum load neural network and the corresponding labels, the sample size of the training set, and the output dimension.

[0044] The perturbation sample loss function construction sub-unit is used to construct a robust learning loss function based on the initial loss function, perturbation rate, load input, load reduction amount of output features, and weight coefficients of perturbation sample learning.

[0045] Optionally, the strong generalization learning unit includes:

[0046] A receiving unit is used to receive model training configuration information; wherein, the model training configuration information includes at least the training set and the target number of iterations;

[0047] An initialization unit is used to initialize the parameters of the minimum shear load neural network, thereby obtaining the initial neural network for calculating the minimum shear load.

[0048] Shuffled units are used to randomly shuffle the training set.

[0049] The first acquisition unit is used to acquire the original sample input features from the shuffled training set;

[0050] The first input unit is used to input the original sample input features into the initial minimum shear load neural network and output the initial sample predicted minimum shear load.

[0051] The load disturbance generation unit is used to randomly generate a load disturbance within the target interval;

[0052] An addition unit is used to add the load disturbance to the training set to obtain a disturbance training set;

[0053] The second acquisition unit is used to acquire perturbation sample input features from the perturbation training set;

[0054] The second input unit is used to input the perturbation sample input features into the initial minimum shear load neural network and output the predicted minimum shear load of the perturbation sample.

[0055] The loss unit is used to calculate the loss value of the minimum shear load of the initial sample and the minimum shear load of the perturbed sample using the robust learning loss function;

[0056] A backpropagation unit is used to calculate the gradient of the neural network parameters using the loss value through backpropagation.

[0057] The update unit is used to update the neural network parameters using the optimizer to obtain the trained minimum shear load neural network;

[0058] A determining unit is used to use the trained minimum load shedding neural network as an intelligent evaluation model for operational reliability if the current iteration number reaches the target iteration number.

[0059] An activation unit is configured to, if the current iteration count has not reached the target iteration count, use the trained minimum load neural network as a new initial minimum load neural network and activate the shuffling unit to execute the random shuffling training set.

[0060] A third aspect of this application provides an electronic device, comprising:

[0061] One or more processors;

[0062] A storage device on which one or more programs are stored;

[0063] When the one or more programs are executed by the one or more processors, the one or more processors implement the evaluation method based on the intelligent evaluation model for operational reliability as described in any one of the first aspects.

[0064] The fourth aspect of this application provides a computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the evaluation method based on the intelligent evaluation model for operational reliability as described in any one of the first aspects.

[0065] As can be seen from the above scheme, this application provides an evaluation method and related apparatus based on an intelligent operational reliability assessment model. The method involves acquiring system data of the power system, establishing a minimum load shedding model based on the system data, determining the inherent pattern of the minimum load shedding model, constructing a robust learning loss function based on the inherent pattern of the minimum load shedding model, and finally performing strong generalization learning based on the minimum load shedding model, the robust learning loss function, and load disturbances to obtain an intelligent operational reliability assessment model. Inputting the system data of the power system into the intelligent operational reliability assessment model yields results that can be used to assess the operational reliability of the system, effectively improving the efficiency and accuracy of the assessment results. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0067] Figure 1 A detailed flowchart of an evaluation method based on an intelligent operational reliability evaluation model provided in this application embodiment;

[0068] Figure 2 A flowchart illustrating a method for constructing a robust learning loss function, as provided in another embodiment of this application;

[0069] Figure 3 A flowchart illustrating a strong generalization learning method provided in another embodiment of this application;

[0070] Figure 4 A schematic diagram illustrating the neural network computation error under different training set sample sizes, provided for another embodiment of this application;

[0071] Figure 5 A schematic diagram of an evaluation device based on an intelligent evaluation model for operational reliability, provided as another embodiment of this application;

[0072] Figure 6 This is a schematic diagram of an electronic device that implements an evaluation method based on an intelligent evaluation model for operational reliability, as provided in another embodiment of this application. Detailed Implementation

[0073] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0074] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0075] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0076] It should be noted that the concepts of "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0077] It should be noted that the terms "a" and "a plurality of" used in this application are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0078] This application provides an evaluation method based on an intelligent operational reliability evaluation model, such as... Figure 1 As shown, the specific steps include:

[0079] S101. Obtain system data of the power system.

[0080] The system data of the power system includes, but is not limited to, information on load reduction at bus nodes, cost coefficients, load information, voltage amplitude and phase angle information, node conductance matrix, node susceptance matrix, generator status, and active power output of generators at the previous moment, etc., which are not limited here.

[0081] S102. Establish a minimum load shedding model based on the system data of the power system.

[0082] In the specific implementation of this application, a minimum load shedding model can be constructed based on the bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum ramp rate, generator set, bus node set and branch set. No limitation is imposed here.

[0083] The minimum shear load model can be represented as follows:

[0084] (1)

[0085] (2)

[0086] (3)

[0087] (4)

[0088] (5)

[0089] (6)

[0090] (7)

[0091] (8)

[0092] (9)

[0093] (10)

[0094] in, and These represent the active and reactive load reduction amounts for bus node i, respectively. The cost coefficient is * (e.g., in equation (1)). This represents the active power cost coefficient of bus node i. and These represent the active and reactive loads of bus node i, respectively. and The active and reactive power output of the generator connected to bus node i; and These represent the voltage magnitude and phase angle at bus node i, respectively. The phase angle difference between the i-th node and the j-th bus node; and These represent the elements in the i-th row and j-th column of the nodal conductance matrix and the nodal susceptance matrix, respectively. This represents the active power between the i-th node and the j-th node; Indicates the generator status; This represents the active power output of generator i at the previous moment; The maximum gradeability of generator i; , , These are collections of generators, busbar nodes, and branches, respectively. and These represent the upper and lower limits of *, respectively (e.g., in equation (10)). This represents the upper limit of active power between the i-th node and the j-th node. The lower limit of the active power between the i-th node and the j-th node.

[0095] S103. Determine the inherent mode of the minimum load shedding model.

[0096] Understandably, the minimum load shedding problem calculates the amount of load that needs to be reduced without violating grid operation constraints, where whether the load reduction is zero depends on the load size. Therefore, if it is known that a node needs to reduce its load, it means that the node is overloaded. If the load is further increased at that node, the increased load, calculated by the optimal load shedding model, also needs to be reduced. Conversely, if the load at the node is not overloaded, no reduction is needed. This rule can be expressed mathematically as:

[0097] (11)

[0098] in, Let i be the load change at node i; The minimum load shedding model is represented by a function that describes the load reduction from the input features to node i. This represents the remaining input characteristics. These remaining input characteristics include, but are not limited to, active load, reactive load, equipment status, grid topology, and renewable energy output, etc., and are not specified here.

[0099] This invention proves the optimality and feasibility of Equation (11) by demonstrating that the new sample after adding load disturbance (the disturbance sample obtained by perturbing the initial sample according to Equation (11)) does not affect the initial sample (the output feature is obtained by randomly sampling the power system state, obtaining the input feature, and solving the minimum load shedding model of the power system using the interior point method). For the initial sample ( , (where the load reduction of node i is not zero), when given (Including but not limited to input characteristics for minimum load shedding calculation such as active load, reactive load, equipment status, grid topology, and renewable energy output, which are not limited here) output characteristics It can be calculated by solving the minimum shear load model equations (1)-(10). According to equation (11), the disturbance sample based on the initial sample can be directly calculated, i.e. , ,in This represents the load disturbance. When the disturbance sample is substituted into the minimum load shedding model, the constraints (3)-(6) and (8)-(10) are independent of the load and load reduction and will not change. In equation (2), since the increase in load and load reduction is the same, the injected power at node i remains unchanged, so the power flow equation remains unchanged. For (7), adding the same disturbance to both sides of the inequality does not change the inequality. Therefore, all the constraints of the disturbance sample in the minimum load shedding model are exactly the same as those of the initial sample, and the feasibility remains unchanged. The objective function then becomes However, when solving the minimum load shedding model, the load disturbance... It is a constant and will not affect the optimality of the load reduction model. Therefore, when the disturbance is... When added to the minimum load shedding model of the initial sample, neither optimality nor feasibility changes, and all input-output characteristics are the same except for the perturbation node i. The proof is complete, demonstrating that adding the perturbation to the minimum load shedding model does not change optimality or feasibility.

[0100] The process of obtaining the perturbation sample by perturbing the initial sample according to equation (11) includes, but is not limited to: selecting samples in the initial sample set whose minimum shear load is not zero, and then adding perturbation to the output features of the selected samples. The perturbation amount satisfies the condition of equation (11). Then, according to equation (11), the corresponding perturbation amount is directly added to the input feature of the perturbation node. Finally, the input and output features after perturbation are saved to obtain the perturbation sample.

[0101] S104. Based on the inherent pattern of the minimum load shedding model, construct a robust learning loss function.

[0102] Understandably, in data-driven minimum load shedding calculations, the training of minimum load shedding neural networks requires hundreds of thousands or even more samples due to complex nonlinearities and topological combinatorial explosion characteristics. This invention constructs a loss function for robust learning of neural networks based on the inherent pattern of minimum load shedding, that is, automatically adding perturbation samples, thereby improving the neural network's learning of the minimum load shedding solution mapping.

[0103] Optionally, in another embodiment of this application, one implementation of step S104 is as follows: Figure 2 As shown, it includes:

[0104] S201. Construct the initial loss function based on the output features of the minimum load shedding neural network, the corresponding labels, the sample size of the training set, and the output dimension.

[0105] Learning a neural network involves optimizing its trainable parameters. Typically, supervised training of the neural network is performed using sample labels, i.e., minimizing the loss function L, which can be represented as follows:

[0106] (12)

[0107] in, and Let represent the i-th output feature and corresponding label of the neural network for the k-th sample, respectively; M represents the number of samples in the training set; and N represents the output dimension.

[0108] S202. Based on the initial loss function, disturbance rate, load input, load reduction of output features, and weight coefficients for learning disturbance samples, a robust learning loss function is constructed.

[0109] Continuing with the above examples, we can see the inherent pattern of the minimum load shedding model: if the initial load reduction is not zero, we can add perturbations to the load and its reduction in the original samples without changing the optimality and feasibility. Therefore, this invention adds perturbations to the non-zero load reduction of all samples to generate more training data. The loss function can be rewritten as equation (13):

[0110] (13)

[0111] in, Let be the perturbation rate of node i, with a value range of . ; To load When the input is , the load reduction amount of the i-th output feature of the neural network; The weight coefficient for learning the perturbation samples can be set to 0.1.

[0112] S105. Based on the minimum load shedding model, robust learning loss function, and load disturbance, a strong generalization learning is performed to obtain an intelligent assessment model for operational reliability.

[0113] Optionally, in the specific implementation of this application, one embodiment of step S105 is as follows: Figure 3 As shown, it includes:

[0114] S301, Receive model training configuration information.

[0115] The received model training configuration information includes at least the training set and the target number of iterations.

[0116] It should be noted that the content of the training set includes, but is not limited to, the active load, reactive load, active output, reactive output, nodal conductance matrix, nodal susceptance matrix, minimum load shedding amount, etc. of the i-th sample, etc., which are not limited here.

[0117] S302. Initialize the parameters of the minimum shear load neural network to obtain the initial neural network for calculating the minimum shear load.

[0118] S303. Randomly shuffle the training set and obtain the original sample input features from the shuffled training set.

[0119] S304. Input the original sample input features into the initial minimum shear load neural network, and output the initial sample predicted minimum shear load.

[0120] S305. Randomly generate a load disturbance within the target interval, add the load disturbance to the training set to obtain the disturbance training set, and obtain the disturbance sample input features from the disturbance training set.

[0121] The target range is pre-set by technical personnel, experts, etc., for example: No restrictions are imposed here.

[0122] S306. Input the perturbation sample input features into the initial minimum shear load neural network, and output the predicted minimum shear load amount of the perturbation sample.

[0123] S307. Use the robust learning loss function to calculate the loss values ​​of the minimum shear load for the initial sample and the minimum shear load for the perturbed sample.

[0124] S308. Calculate the gradient of the neural network parameters using backpropagation with the loss value.

[0125] S309. Use the optimizer to update the neural network parameters to obtain the trained minimum shear load neural network.

[0126] S310. Determine whether the current iteration count has reached the target iteration count.

[0127] Specifically, if the current iteration count reaches the target iteration count, then step S311 is executed; if the current iteration count does not reach the target iteration count, then step S312 is executed.

[0128] S311. Use the trained minimum load shedding neural network as an intelligent evaluation model for operational reliability.

[0129] S312. Use the trained minimum load shedding neural network as the new initial minimum load shedding neural network.

[0130] The following is a strong generalization learning algorithm (Algorithm 1) provided in the embodiments of this application:

[0131] Algorithm 1

[0132] enter The learning rate is r, the number of iterations is E, and the training dataset is... Initialize the neural network parameters H, e=0. Output The parameters H of the trained neural network. 1 While e<E: 2 Randomly shuffle the training sample set. 3 For all : 4 When the input is At that time, the output of the neural network with parameter H is calculated. . 5 Generate numerical values ​​in Load disturbance in the area ; 6 Determine to explore new samples And calculate the neural network output ; 7 Use equation (13) to calculate the loss value and perform backpropagation to calculate the gradient of the neural network parameters; 8 Update the neural network parameters H using the optimizer; 9 End for 10 e = e + 1 11 End while 12 Output H

[0133] in, The data in the table represent the active load, reactive load, active power output, reactive power output, node conductance matrix, node susceptance matrix, and minimum load shedding for the i-th sample, respectively.

[0134] It should be noted that, The minimum shear load predicted by the neural network for the initial sample corresponds to the value in formula (13). ; The minimum shear load predicted by the neural network for the perturbation sample corresponds to formula (13) .

[0135] S106. Input the system data of the power system into the intelligent assessment model for operational reliability to obtain results that can be used to assess the operational reliability of the system.

[0136] Continuing with the above embodiments, after obtaining the intelligent assessment model for operational reliability through strong generalization learning based on the minimum load shedding model, robust learning loss function, and load disturbance, the system data of the power system can be input into the intelligent assessment model for operational reliability to obtain results that can be used to assess the operational reliability of the system, thereby effectively improving the efficiency and accuracy of the assessment results.

[0137] The present invention will be further described below with reference to specific implementation schemes.

[0138] Example 1:

[0139] Operational reliability assessments in power systems typically involve load fluctuations, stochastic renewable energy generation, and outages of critical electrical equipment. For load fluctuations, this embodiment assumes they follow a normal distribution, with a default mean of 0.3 and a standard deviation of 0.3. This embodiment considers two renewable energy sources: wind power and solar irradiance. It is assumed that wind speed and solar irradiance follow Weibull and Beta distributions, respectively, with wind speed following a Weibull distribution. Solar irradiance follows Distribution, among which This embodiment connects multiple wind farms and photovoltaic power plants to different nodes, enabling a renewable energy penetration rate exceeding 20%. In the IEEE 39-node system, four 200MW wind farms and four 200MW photovoltaic power plants are randomly connected to different nodes in the power system. The renewable energy penetration rate is 20.26%. To reduce training data imbalance, this embodiment samples N-1 and N-2 scenarios to cover all training cases. This embodiment generates 20K and 10K samples for training and testing, respectively.

[0140] To demonstrate the effectiveness of the proposed method, this embodiment compares traditional operational reliability assessment methods with those using and without the proposed learning method, denoted as M0, M1, and M2, respectively. All neural networks were built and trained using the PyTorch framework on a desktop computer with an Intel(R) Core(TM) i7-10700K CPU @ 3.80 GHz, 16 GB RAM, and an NVIDIA GeForce RTX 2080 Ti. The Adam optimizer was used to train the neural networks at a learning rate of 0.001. The training iterations were 2000.

[0141] M0: Traditional operational reliability assessment methods use the interior point method to determine the minimum load shedding. This method serves as the benchmark method for data analysis.

[0142] M1: This method utilizes an embedded graph convolutional neural network architecture and trains a minimum shear load computation neural network using traditional supervised learning methods.

[0143] M2: This method utilizes an embedded graph convolutional neural network architecture and trains a minimum shear load computation neural network using the strong generalization learning method proposed in this invention.

[0144] This embodiment uses initial training sample sets with different data volumes to train the neural network, and then statistically analyzes the prediction error of the trained neural network on the test set. The results are as follows: Figure 4 As shown.

[0145] Depend on Figure 4It can be seen that, under different training set sample sizes, the computational error of M2 is consistently smaller than that of M1. Furthermore, as the number of training samples decreases, the computational error of M1 increases sharply, especially when the number of training data is less than 6000. Conversely, the computational error of M2 increases slowly, and only begins to rise when the number of data is less than 1000. The main reason for this phenomenon is that the inherent pattern-guided strong generalization learning method proposed in this invention is not limited to the given training data and can generate more load-reducing samples using known samples. This not only overcomes the problem of load imbalance but also reduces the neural network's dependence on a large number of training samples, verifying the effectiveness of the minimum load-reducing inherent pattern-guided strong generalization learning algorithm proposed in this invention.

[0146] Example 2:

[0147] Operational reliability assessments in power systems typically involve load fluctuations, stochastic renewable energy generation, and outages of critical electrical equipment. For load fluctuations, this embodiment assumes they follow a normal distribution, with a default mean of 0.3 and a standard deviation of 0.3. This embodiment considers two renewable energy sources: wind power and solar irradiance. It is assumed that wind speed and solar irradiance follow Weibull and Beta distributions, respectively, with wind speed following a Weibull distribution. Solar irradiance follows Distribution, among which This embodiment connects multiple wind farms and photovoltaic power plants to different nodes, achieving a renewable energy penetration rate exceeding 20%. In the IEEE 39-node system, four 200MW wind farms and four 200MW photovoltaic power plants are randomly connected to different nodes in the power system. The renewable energy penetration rate is 20.26%. To reduce training data imbalance, this embodiment samples N-1 and N-2 scenarios to cover all training cases. During the reliability assessment phase, this embodiment simulates a 1% failure probability for reliability evaluation. This embodiment generates 20K and 10K samples for training and testing, respectively.

[0148] To demonstrate the effectiveness of the proposed method, this embodiment compares traditional operational reliability assessment methods with those using and without the proposed learning method, denoted as M0, M1, and M2, respectively. All neural networks were built and trained using the PyTorch framework on a desktop computer with an Intel(R) Core(TM) i7-10700K CPU @ 3.80 GHz, 16 GB RAM, and an NVIDIA GeForce RTX 2080 Ti. The Adam optimizer was used to train the neural networks at a learning rate of 0.001. The training iterations were 2000.

[0149] M0: Traditional operational reliability assessment methods use the interior point method to determine the minimum load shedding. This method serves as the benchmark method for data analysis.

[0150] M1: This method utilizes an embedded graph convolutional neural network architecture and trains a minimum shear load computation neural network using traditional supervised learning methods.

[0151] M2: This method utilizes an embedded graph convolutional neural network architecture and trains a minimum shear load computation neural network using the strong generalization learning method proposed in this invention.

[0152] This embodiment uses operational reliability adequacy indices for analysis, including Loss of Load Probability (LOLP) and Expected Demand Not Supplied (EDNS), expressed as:

[0153] ;

[0154] ;

[0155] in, Let S be the probability of the i-th system state; S is the set of system states where load reduction occurs. This represents the load reduction amount at node j under system state i.

[0156] After training with 20,000 samples in the IEEE 39-bus system, the probability of unit and line out of operation was set to 1%, and operational reliability was assessed. Reliability indicators were statistically analyzed, and the results are shown in Table 1. It can be observed that the neural network trained using traditional learning methods, although exhibiting small computational errors, still has a relative error exceeding 10% in the operational reliability assessment. In contrast, the proposed minimum load shedding inherent pattern-guided strong generalization learning algorithm expands the load shedding samples through random perturbation, promoting the neural network's learning of these samples. This further reduces the computational error of the load shedding probability to below 10%, reaching 4.82%, thus verifying the effectiveness of the proposed learning algorithm.

[0157] Table 1

[0158] method LOLP EDNS M0 0.04479 / 9.90 / M1 0.04932 10.11% 10.158 2.61% M2 0.04263 4.82% 10.035 1.36%

[0159] in, This represents the relative error of LOLP. This indicates the relative error of EDNS.

[0160] As can be seen from the above scheme, this application provides an evaluation method based on an intelligent operational reliability assessment model. The method involves acquiring system data of the power system, establishing a minimum load shedding model based on the system data, determining the inherent pattern of the minimum load shedding model, constructing a robust learning loss function based on the inherent pattern of the minimum load shedding model, and finally performing strong generalization learning based on the minimum load shedding model, the robust learning loss function, and load disturbances to obtain an intelligent operational reliability assessment model. Inputting the system data of the power system into the intelligent operational reliability assessment model yields results that can be used to assess the operational reliability of the system, effectively improving the efficiency and accuracy of the assessment results.

[0161] Another embodiment of this application provides an evaluation device based on an intelligent operational reliability evaluation model, such as... Figure 5 As shown, it specifically includes:

[0162] The system data acquisition unit 501 is used to acquire system data of the power system.

[0163] Model building unit 502 is used to build the minimum load shear model.

[0164] Optionally, in another embodiment of this application, one implementation of the model building unit 502 includes:

[0165] The model establishes sub-units to construct a minimum load shedding model based on bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum ramp rate, generator set, bus node set, and branch set.

[0166] For details on the specific working process of the units disclosed in the above embodiments of this application, please refer to the corresponding method embodiments, which will not be repeated here.

[0167] The inherent mode determination unit 503 is used to determine the inherent mode of the minimum load shearing model.

[0168] Loss function building unit 504 is used to construct a robust learning loss function based on the inherent patterns of the minimum load shedding model.

[0169] Optionally, in another embodiment of this application, one implementation of the loss function construction unit 504 includes:

[0170] The initial sample loss function construction sub-unit is used to construct the initial loss function based on the output features of the minimum load neural network, the corresponding labels, the sample size of the training set, and the output dimension.

[0171] The perturbation sample loss function construction sub-unit is used to construct a robust learning loss function based on the initial loss function, perturbation rate, load input, load reduction amount of output features, and weight coefficients of perturbation sample learning.

[0172] For details on the specific working process of the units disclosed in the above embodiments of this application, please refer to the corresponding method embodiments, such as... Figure 2 As shown, it will not be elaborated further here.

[0173] The strong generalization learning unit 505 is used to perform strong generalization learning based on the minimum load shedding model, the robust learning loss function, and the load disturbance to obtain an intelligent assessment model for operational reliability.

[0174] The evaluation unit 506 is used to input the system data of the power system into the intelligent evaluation model for operational reliability, and obtain results that can be used to evaluate the operational reliability of the system.

[0175] For details on the specific working process of the units disclosed in the above embodiments of this application, please refer to the corresponding method embodiments, such as... Figure 1 As shown, it will not be elaborated further here.

[0176] Optionally, in another embodiment of this application, one implementation of the strong generalization learning unit 505 includes:

[0177] The receiving unit is used to receive model training configuration information.

[0178] The model training configuration information includes at least the training set and the target number of iterations.

[0179] The initialization unit is used to initialize the parameters of the minimum shear load neural network, thereby obtaining the initial neural network for calculating the minimum shear load.

[0180] Shuffled units are used to randomly shuffle the training set.

[0181] The first acquisition unit is used to acquire the original sample input features from the shuffled training set.

[0182] The first input unit is used to input the original sample input features into the initial minimum shear load neural network and output the initial sample predicted minimum shear load.

[0183] The load disturbance generation unit is used to randomly generate a load disturbance within the target interval.

[0184] Adding units is used to add load disturbances to the training set, resulting in a disturbance training set.

[0185] The second acquisition unit is used to acquire perturbation sample input features from the perturbation training set.

[0186] The second input unit is used to input the perturbation sample input features into the initial minimum shear load neural network and output the predicted minimum shear load of the perturbation sample.

[0187] The loss unit is used to calculate the loss value of the minimum shear load of the initial sample and the minimum shear load of the perturbed sample using the robust learning loss function.

[0188] The backpropagation unit is used to calculate the gradient of the neural network parameters by backpropagating the loss value.

[0189] The update unit is used to update the neural network parameters using the optimizer to obtain the trained minimum shear load neural network.

[0190] The determination unit is used to use the trained minimum load shedding neural network as an intelligent evaluation model for operational reliability if the current iteration number reaches the target iteration number.

[0191] The activation unit is used to activate the shuffling unit to randomly shuffle the training set if the current iteration number has not reached the target iteration number.

[0192] For details on the specific working process of the units disclosed in the above embodiments of this application, please refer to the corresponding method embodiments, such as... Figure 3 As shown, it will not be elaborated further here.

[0193] As can be seen from the above scheme, this application provides an evaluation device based on an intelligent operational reliability evaluation model. It acquires system data of the power system, establishes a minimum load shedding model based on the system data, determines the inherent pattern of the minimum load shedding model, and constructs a robust learning loss function based on the inherent pattern of the minimum load shedding model. Finally, it performs strong generalization learning based on the minimum load shedding model, the robust learning loss function, and load disturbances to obtain an intelligent operational reliability evaluation model. Inputting the system data of the power system into the intelligent operational reliability evaluation model yields results that can be used to evaluate the operational reliability of the system, effectively improving the efficiency and accuracy of the evaluation results.

[0194] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0195] Another embodiment of this application provides an electronic device, such as... Figure 6 As shown, it includes:

[0196] One or more processors 601.

[0197] Storage device 602, on which one or more programs are stored.

[0198] When the one or more programs are executed by the one or more processors 601, the one or more processors 601 implement the evaluation method based on the intelligent evaluation model for operational reliability as described in the above embodiments.

[0199] Another embodiment of this application provides a computer storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, it implements the evaluation method based on the intelligent evaluation model for operational reliability as described in the above embodiments.

[0200] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0201] It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0202] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0203] Another embodiment of this application provides a computer program product, which, when executed, is used to perform the above-described evaluation method based on the intelligent evaluation model for operational reliability.

[0204] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, it performs the functions defined in the methods of the embodiments of this application.

[0205] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in this application is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely exemplary forms of implementing this application.

[0206] While several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0207] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described application concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. An evaluation method based on an intelligent operational reliability assessment model, characterized in that, include: Obtain system data from the power system; A minimum load shedding model is established based on the system data of the power system. Determine the inherent pattern of the minimum load shear model; Based on the inherent pattern of the minimum load shedding model, a robust learning loss function is constructed. Based on the minimum load shedding model, the robust learning loss function, and the load disturbance, a strong generalization learning method is used to obtain an intelligent assessment model for operational reliability. The system data of the power system is input into the intelligent assessment model for operational reliability to obtain results that can be used to assess the operational reliability of the system. The step of obtaining the intelligent operational reliability assessment model by performing strong generalization learning based on the minimum load shedding model, the robust learning loss function, and load disturbances includes: Receive model training configuration information; wherein, the model training configuration information includes at least the training set and the target number of iterations; Initialize the parameters of the minimum load shedding neural network to obtain the initial minimum load shedding neural network; Randomly shuffle the training set; Obtain the original sample input features from the shuffled training set; The original sample input features are input into the initial minimum shear load neural network, and the output is the initial sample predicted minimum shear load. A load disturbance is randomly generated within the target interval; The load disturbance is added to the training set to obtain the disturbance training set; Obtain input features of perturbation samples from the perturbation training set; The perturbation sample input features are input into the initial minimum load shedding neural network, and the output is the predicted minimum load shedding amount of the perturbation sample; The loss value of the minimum shear load for the initial sample and the minimum shear load for the perturbed sample is calculated using the robust learning loss function; The gradient of the neural network parameters is calculated using backpropagation based on the loss value. The optimizer is used to update the neural network parameters to obtain the trained minimum shear load neural network; If the current iteration count reaches the target iteration count, then the trained minimum load shedding neural network will be used as the intelligent evaluation model for operational reliability. If the current iteration count has not reached the target iteration count, the trained minimum load shedding neural network is used as the new initial minimum load shedding neural network, and the random shuffling training set step is returned to be executed.

2. The evaluation method based on the intelligent assessment model for operational reliability according to claim 1, characterized in that, The step of establishing a minimum load shedding model based on the system data of the power system includes: Based on the bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum gradeability, generator set, bus node set and branch set, a minimum load shedding model is constructed.

3. The evaluation method based on the intelligent assessment model for operational reliability according to claim 1, characterized in that, The robust learning loss function is constructed based on the inherent pattern of the minimum load shedding model, including: Based on the output features of the minimum load shedding neural network, the corresponding labels, the sample size of the training set, and the output dimension, an initial loss function is constructed. Based on the initial loss function, disturbance rate, load input, load reduction of output features, and weight coefficients for learning disturbance samples, a robust learning loss function is constructed.

4. An evaluation device based on an intelligent operational reliability evaluation model, characterized in that, include: The system data acquisition unit is used to acquire system data of the power system; The model building unit is used to build a minimum load shedding model based on the system data of the power system. An inherent mode determination unit is used to determine the inherent mode of the minimum load shearing model; The loss function construction unit is used to construct a robust learning loss function based on the inherent patterns of the minimum load shedding model. A strong generalization learning unit is used to perform strong generalization learning based on the minimum load shedding model, the robust learning loss function, and the load disturbance to obtain an intelligent evaluation model for operational reliability. The evaluation unit is used to input the system data of the power system into the intelligent evaluation model for operational reliability, and obtain results that can be used to evaluate the operational reliability of the system. The strong generalization learning unit includes: A receiving unit is used to receive model training configuration information; wherein, the model training configuration information includes at least the training set and the target number of iterations; An initialization unit is used to initialize the parameters of the minimum load shearing neural network to obtain the initial minimum load shearing neural network. Shuffled units are used to randomly shuffle the training set. The first acquisition unit is used to acquire the original sample input features from the shuffled training set; The first input unit is used to input the original sample input features into the initial minimum shear load neural network and output the initial sample predicted minimum shear load. The load disturbance generation unit is used to randomly generate a load disturbance within the target interval; An addition unit is used to add the load disturbance to the training set to obtain a disturbance training set; The second acquisition unit is used to acquire perturbation sample input features from the perturbation training set; The second input unit is used to input the perturbation sample input features into the initial minimum shear load neural network and output the predicted minimum shear load of the perturbation sample. The loss unit is used to calculate the loss value of the minimum shear load of the initial sample and the minimum shear load of the perturbed sample using the robust learning loss function; A backpropagation unit is used to calculate the gradient of the neural network parameters using the loss value through backpropagation. The update unit is used to update the neural network parameters using the optimizer to obtain the trained minimum shear load neural network; A determining unit is used to use the trained minimum load shedding neural network as an intelligent evaluation model for operational reliability if the current iteration number reaches the target iteration number. An activation unit is configured to, if the current iteration count has not reached the target iteration count, use the trained minimum load neural network as a new initial minimum load neural network and activate the shuffling unit to execute the random shuffling training set.

5. The evaluation device based on the intelligent evaluation model for operational reliability according to claim 4, characterized in that, The model building unit includes: The model establishes sub-units to construct a minimum load shedding model based on bus node load reduction information, cost coefficient, load information, voltage amplitude and phase angle information, generator output information connected to the bus node, phase angle difference between bus nodes, node conductance matrix, node susceptance matrix, active power between nodes, generator status, generator active output at the previous moment, generator maximum ramp rate, generator set, bus node set, and branch set.

6. The evaluation device based on the intelligent evaluation model for operational reliability according to claim 4, characterized in that, The loss function construction unit includes: The initial sample loss function construction sub-unit is used to construct the initial loss function based on the output features of the minimum load neural network and the corresponding labels, the sample size of the training set, and the output dimension. The perturbation sample loss function construction sub-unit is used to construct a robust learning loss function based on the initial loss function, perturbation rate, load input, load reduction amount of output features, and weight coefficients of perturbation sample learning.

7. An electronic device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the evaluation method based on the intelligent evaluation model for operational reliability as described in any one of claims 1 to 3.

8. A computer storage medium, characterized in that, It stores a computer program, wherein the computer program, when executed by a processor, implements the evaluation method based on the intelligent evaluation model for operational reliability as described in any one of claims 1 to 3.

Images