Power grid operation state risk assessment method and device, equipment and medium

Abstract

The application discloses a power grid operation state risk assessment method, device, equipment and medium, and relates to the technical field of power grid risk assessment, wherein the method comprises the following steps: a first deep neural network model is structurally optimized to obtain a second deep neural network model; a full-information element data model is constructed according to feature data corresponding to different dimensions of the power grid operation state; the second deep neural network model is trained based on a database to obtain a target deep neural network model; and the target risk assessment level of the current power grid operation state is predicted based on the target deep neural network model. The application is suitable for the risk assessment of the power grid operation state in scenes such as extreme weather and large new energy generation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid risk assessment, in particular to a power grid operation state risk assessment method, device, equipment and medium. BACKGROUND

[0002] With the large-scale access of existing new energy, the demand for power system balance regulation capacity needs to be enhanced. At the same time, the large-scale grid connection of intermittent power generation resources such as wind power and solar power makes the power balance present obvious spatial and temporal imbalance, which aggravates the burden of system regulation. On the one hand, a large number of conventional units are in deep peak regulation or even shutdown standby state, but frequent regulation output brings problems such as cost increase and service life shortening. The replacement of conventional units by large-scale units leads to insufficient system regulation capacity, prominent peak regulation, frequency regulation and voltage regulation contradictions, and significantly increased difficulty of power grid operation control. On the other hand, the form of power grid changes, and the safe operation presents new characteristics. That is, with the large-scale production of new energy units at the sending end and the large-scale feeding of direct current at the receiving end, the coupling relationship between future AC / DC, sending / receiving end and source / grid / load is closer, which further leads to the increasing risk of cascading failures and affects the overall safety of the power grid.

[0003] Especially in the scene of extreme weather and large-scale new energy generation, the operation characteristics of the power grid are complex, the power grid operation state risk system cannot comprehensively and comprehensively perceive the power grid operation state risk, and thus cannot realize advanced perception of the power grid tight balance state. SUMMARY

[0004] Therefore, the present application provides a power grid operation state risk assessment method, device, equipment and medium, which mainly aims to solve the technical problems that a large number of conventional units are in deep peak regulation or even shutdown standby state in the prior art, and the cascading failure risk caused by the coupling relationship under the new form of power grid, which leads to low safety of the power grid and cannot predict the power grid operation risk state in advance.

[0005] According to one aspect of the present application, a power grid operation state risk assessment method is provided, which comprises:

[0006] The first deep neural network model is structurally optimized to obtain a second deep neural network model;

[0007] A full information element data model is constructed for the feature data corresponding to different dimensions of the power grid operation state;

[0008] The second deep neural network model is trained based on the full information element data model to obtain a target deep neural network model;

[0009] The target risk assessment level of the current power grid operation state is predicted based on the target deep neural network model.

[0010] According to another aspect of the present application, there is provided a risk assessment device for power grid operation state, comprising:

[0011] a first construction module configured to construct a second deep neural network model by performing structural optimization on a first deep neural network model;

[0012] a second module configured to construct a full-information element data model for feature data corresponding to different dimensions of the power grid operation state;

[0013] a training module configured to train the second deep neural network model to obtain a target deep neural network model based on the full-information element data model;

[0014] a prediction module configured to predict a target risk assessment level of the current power grid operation state based on the target deep neural network model.

[0015] According to yet another aspect of the present application, there is provided a computer storage medium having a computer program stored thereon, the program being executed by a processor to implement the above-mentioned risk assessment method for power grid operation state.

[0016] According to still another aspect of the present application, there is provided a computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, the processor implementing the above-mentioned risk assessment method for power grid operation state when executing the program.

[0017] By means of the above technical solutions, the risk assessment method, device, equipment and medium for power grid operation state provided by the present application can predict the risk state of the power grid operation in advance compared with the prior technical solutions which cannot predict the risk state of the power grid operation in advance. The present application constructs a second deep neural network model by performing structural optimization on a first deep neural network model, constructs a full-information element data model for feature data corresponding to different dimensions of the power grid operation state, trains the second deep neural network model to obtain a target deep neural network model based on the full-information element data model, and predicts a target risk assessment level of the current power grid operation state based on the target deep neural network model. It can be seen that the improved deep neural network model is trained based on the feature data corresponding to different dimensions of the power grid operation state to realize online identification of the risk state of the power grid operation, achieve the purpose of early perception of the power grid balance state, and further provide a warning tool for the risk state of the power grid operation possibly caused by extreme weather or rapid load climbing.

[0018] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented in accordance with the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0020] Figure 1 A flowchart illustrating a risk assessment method for power grid operating status provided in an embodiment of this application is shown.

[0021] Figure 2 A flowchart illustrating another risk assessment method for power grid operating status provided in an embodiment of this application is shown.

[0022] Figure 3 This illustration shows a schematic diagram of the structural improvement of the second deep neural network model provided in an embodiment of this application;

[0023] Figure 4 A flowchart illustrating the target deep neural network model provided in an embodiment of this application is shown;

[0024] Figure 5 This paper shows a schematic diagram of the structure of a risk assessment device for power grid operation status provided in an embodiment of this application;

[0025] Figure 6 A schematic diagram of another power grid operation status risk assessment device provided in an embodiment of this application is shown. Detailed Implementation

[0026] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0027] Addressing the technical challenges of low grid security and the inability to predict grid operational risks due to the large number of conventional generating units operating under deep peak-shaving or even standby conditions, and the cascading failure risks arising from coupling relationships in new grid configurations, this embodiment provides a risk assessment method for grid operation status. It trains an improved deep neural network model based on feature data corresponding to different dimensions of grid operation status to achieve online identification of grid operational risks, enabling proactive perception of tight grid balance and providing an early warning tool for potential grid operational risks caused by extreme weather or rapid load increases. Figure 1 As shown, the above method includes the following steps:

[0028] Step 101: By optimizing the structure of the first deep neural network model, a second deep neural network model is constructed.

[0029] In the embodiment, the first deep neural network model is a TransFormer model, which can consider the correlation between features at different time points based on the multi-head self-attention mechanism of the TransFormer model, realize multi-level and multi-subspace attention on input sequence data, and thus can capture rich feature information in the input sequence data. In order to further improve the prediction accuracy of the model, the TransFormer model is structurally optimized and constructed to obtain a second deep neural network model based on the data features related to the power grid operation state. The input sequence data is the feature data corresponding to different dimensions of the power grid operation state, mainly including power grid feature data and environmental feature data.

[0030] Step 102, constructing a full information element data model for the feature data corresponding to different dimensions of the power grid operation state.

[0031] In the embodiment, the full information element data model is constructed, that is, a corresponding database is constructed for the feature data corresponding to different dimensions of the power grid operation state. The feature data corresponding to different dimensions of the power grid operation state includes power grid feature data and environmental feature data. The power grid feature data includes full-network positive and negative backup state data and key section operation state data. The first database is a database constructed based on the full-network positive and negative backup state data. The second database is a database constructed based on the key section operation state data. The third database is a database constructed based on the environmental feature data.

[0032] Step 103, training the second deep neural network model based on the full information element data model to obtain a target deep neural network model.

[0033] In the embodiment, based on the feature data in the full information element data model, simulation data of a preset time period is generated by simulation of a dispatching automation system test verification platform, and is used as sample data for training the second deep neural network model, so as to realize training of the second deep neural network model. Further, the cross-entropy loss function and the Adam optimizer are used to realize iteration of the second deep neural network model, so as to obtain a target deep neural network model which can effectively improve the accuracy of power grid operation risk state early warning.

[0034] Step 104, predicting a target risk assessment level of the current power grid operation state based on the target deep neural network model.

[0035] In the embodiment, by selecting the power grid operation state data of a preset time period, the power grid operation state data is from the full information element data model and is input as input sequence data into the target deep neural network model, so as to extract effective features from the power grid operation state data in multiple dimensions and accurately predict the current tight balance state of the power grid (i.e., the target risk assessment level). The risk assessment level can be divided into four levels: extremely tight, serious tight, tight, and nervous. According to the needs of actual application scenarios, it can be further determined as a non-risk state and a risk state. Here, the risk assessment level is not specifically limited.

[0036] For the embodiment, the above scheme can be followed to construct a second deep neural network model by optimizing the structure of the first deep neural network model; a full information element data model is constructed for the feature data corresponding to different dimensions of the power grid operation state; the second deep neural network model is trained based on the full information element data model to obtain a target deep neural network model; and the target risk assessment level of the current power grid operation state is predicted based on the target deep neural network model. Compared with the technical solution that cannot predict the power grid operation risk state in advance, the embodiment comprehensively considers the interweaving complexity of the power grid operation state data in multiple dimensions under scenarios such as large-scale new energy generation, uses the improved deep neural network model to perform risk intelligent assessment on the power grid operation state data from different dimensions such as time and space, performs online identification of the power grid operation risk state through deep neural network learning, achieves the purpose of advanced perception of the tight balance state of the power grid, and further provides a warning tool for the power grid operation risk state that may be caused by extreme weather or rapid load climbing.

[0037] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to completely describe the specific implementation process of the embodiment, another risk assessment method for the power grid operation state is provided. The embodiment can solve the technical problem that the power grid operation state risk system in the prior art cannot comprehensively and comprehensively perceive the power grid operation state risk, and thus cannot achieve advanced perception of the tight balance state of the power grid. The embodiment can comprehensively consider uncertain factors such as new energy power generation fluctuation and external environment, train the improved deep neural network model based on the feature data corresponding to different dimensions of the power grid operation state, to realize online identification of the power grid operation risk state, i.e., advanced warning of the power grid operation risk state, so that the dispatching and operation personnel can adjust the operation mode, reduce the power grid construction investment for low-risk fault prevention, and adjust the corresponding margin control in the operation process, thereby effectively reducing the probability of cascading failures, improving the ability of the power grid to resist possible risks, improving the economy of system operation, and ensuring the safe operation of the power grid. As shown in Figure 2 The risk assessment method for the power grid operation state based on the TransFormer model includes:

[0038] Step 201, a second deep neural network model is constructed by optimizing the structure of a first deep neural network model.

[0039] To illustrate the specific implementation of step 201, as a preferred embodiment, the step 201 includes: adding a multi-scale hollow convolution layer in the multi-head self-attention mechanism of the first deep neural network model to obtain a multi-scale hollow attention mechanism; and replacing the multi-layer perceptron in the first deep neural network model with a convolution layer to obtain the second deep neural network model. It should be noted that the above steps are not limited in order.

[0040] In implementation, the first deep neural network model is a TransFormer model, and the second deep neural network model fuses a plurality of encoding feature vectors obtained by encoding different stages of the multi-head self-attention mechanism as supplementary feature vectors with an initial input vector of a decoding stage to obtain a target input vector of the decoding stage.

[0041] In implementation, a basic TransFormer model suitable for power grid operation state risk assessment is constructed. The TransFormer model not only has high sensitivity and recognition ability to subtle changes in the actual power grid operation state by capturing complex patterns and dynamic changes in time series, but also has strong classification ability.

[0042] Considering that the power grid operation state risk assessment is an important part of power system dispatching operation, in order to improve the accuracy of model prediction, the basic TransFormer model (the first deep neural network model) is optimized in structure. Specifically, as shown in Figure 3 First, a multi-scale hollow attention mechanism is obtained by adding a multi-scale hollow convolution layer in the multi-head self-attention mechanism, thereby processing sequence data in parallel, correlating more factor information, reducing the number of parameters, improving the calculation efficiency of the network, and reducing the deployment requirements for hardware. Secondly, the traditional multi-layer perceptron is replaced by a 1x1 convolution layer and a 3x3 convolution layer (for example, one 1x1 convolution layer and three 3x3 convolution layers) with fewer parameters, so as to fully consider the requirement of information recovery to retain part of the redundant information, while the calculation amount is less. Finally, the encoding features of different stages are used as supplementary feature vectors of the decoding stage to realize the completion of data features and improve the comprehensiveness and accuracy of model prediction. It can be seen that the optimized TransFormer model (the second deep neural network model) can handle long-distance dependence problems and greatly improve the learning ability and efficiency of the TransFormer model.

[0043] Step 202, for the feature data corresponding to different dimensions of the power grid operation state, a full information element data model is constructed.

[0044] In the implementation, the feature data corresponding to different dimensions of the power grid operating state includes power grid feature data and environment feature data; the power grid feature data includes all-grid positive and negative backup state data and key section operating state data; correspondingly, the full information element data model includes an all-grid positive and negative backup state data model, a key section operating state data model, and an external environment feature data model, the all-grid positive and negative backup state data in the all-grid positive and negative backup state data model includes one or more of all-grid power generation data, all-grid load data, all-grid network loss data, all-grid positive and negative backup capacity, all-grid positive and negative backup margin, maximum external power transmission power, and external power load ratio; the key section operating state data in the key section operating state data model includes one or more of key section load rate of a target partition, DC near-zone first and second section power flow load rate, key power transmission channel total power, key power transmission channel maximum transmission power, and key power transmission channel margin; the environment feature data (meteorological factors) in the external environment feature data model includes one or more of temperature, humidity, rainfall, wind speed, and solar radiation intensity. The maximum external power transmission power is determined by a cross-region and cross-province plan.

[0045] (1) The calculation formulas of the all-grid positive and negative backup capacity and the all-grid positive and negative backup margin in the all-grid positive and negative backup state data are as follows:

[0046] 1. All-grid positive and negative backup capacity

[0047] The all-grid positive and negative backup capacity refers to the capacity that can be adjusted upward (downward) under the condition of safe operation of the power grid. The calculation formulas of the all-grid positive backup capacity P Tup and the all-grid negative backup capacity P Tdown are as follows:

[0048] (1)

[0049] (2)

[0050] Wherein, Pi,max is the maximum output of all synchronous units in the system, Pi is the output of the i th synchronous unit in the system, and Pi,min is the minimum output of all synchronous units in the system.

[0051] 2. All-grid positive and negative backup margin

[0052] The all-grid positive and negative backup margin refers to the range that can be adjusted upward (downward) under the condition of safe operation of the power grid. The calculation formulas of the all-grid positive backup margin P Tup-magrin and the all-grid negative backup margin P Tdown-magrin are as follows:

[0053] (3)

[0054] (4)

[0055] (2) The target (500kV) zone grid critical section load rate, DC near-area primary and secondary section power flow load rate, and maximum transmission power of critical transmission channels in the critical section operation status data, and the calculation formula for critical transmission channel margin are as follows:

[0056] 1. Load rate of key sections of 500kV zoned grid

[0057] The critical sections of the 500kV regional grid are the focus of the control system's monitoring. These critical sections are constrained by the thermal stability transmission limits of the primary transmission equipment. Therefore, the load factor P of the critical sections of the 500kV regional grid is... Srw The calculation formula is:

[0058] (5)

[0059] Among them, P Slimit This considers the thermal stability transmission limit capacity of the N-1 interrupted section, P Sf It refers to the transmission power of key sections in the 500kV zoned grid.

[0060] 2. DC near-field primary and secondary section power flow load factor

[0061] The power output of the external transmission channel is arranged according to the superior dispatch plan. The power flow of the primary and secondary sections of the DC near-area is limited by the section stability limit. Therefore, the power flow load factor P of the primary and secondary sections of the DC near-area is... S-DC The calculation formula is:

[0062] (6)

[0063] Among them, P Sdcjq It is the first and second stage DC near-field section power flow, P Sdcjqlimit It is the stability limit for the first and second stage sections of the DC near-field.

[0064] 3. Maximum transmission capacity of key transmission channels

[0065] Maximum transmission capacity P of key transmission channels Ch-ability The calculation formula is:

[0066] (7)

[0067] Where K is a coefficient, P Ch,i For the transmission power of critical transmission channel i, For the stability limit of critical transmission channel i.

[0068] 4. Margin of critical transmission channels

[0069] Critical power path margin P Ch-magrin The calculation formula is:

[0070] (8)

[0071] Step 203, generating simulation data of a preset time period according to the feature data in the full-information element data model, and taking the simulation data as sample data for training the second deep neural network model.

[0072] Step 204, training the second deep neural network model according to the sample data to obtain a target deep neural network model.

[0073] In order to illustrate the specific implementation of step 204, as a preferred embodiment, the step 204 includes: presetting a plurality of output results of the second deep neural network model, each output result corresponding to a different risk assessment level; and iteratively training the second deep neural network model based on the sample data until the output result of the second deep neural network model is consistent with the preset risk assessment label of the power grid operation state, to obtain the target deep neural network model.

[0074] In implementation, the feature data in the full-information element data model is historical data. Based on the historical data (for example, selecting one week of historical data of the power grid operation state in a certain provincial power grid), one year of simulation data is generated by a dispatching automation system test verification platform simulation to serve as sample data, so as to construct a sample data set for training the improved TransFormer model. The power grid operation risk state label in the sample data set is determined by expert rules. Further, the second deep neural network model is trained by using a cross-entropy loss function and an Adam optimizer to obtain a target deep neural network model, i.e., an intelligent agent for predicting the power grid operation risk state. It can be seen that, in the intelligent agent training process, by improving the robustness of the model under different time periods and different load conditions, reducing the error of the loss function, and improving the accuracy of the intelligent agent, not only the effectiveness of the TransFormer model is verified, but also the feature understanding of the TransFormer model for the power grid operation state is deepened.

[0075] In implementation, as Figure 4As shown, the improved TransFormer model can extract data features from multiple feature heads in more detail based on multi-head attention, and fuse and decode the data features of different feature heads via multiple linear layers and two 1x1 convolution layers, respectively, to output the risk assessment result of the power grid operation state, which includes a prediction probability vector between 0 and 1 for each risk assessment level (extremely tight, serious tight, tight, and nervous) set. The maximum value is selected from the obtained four prediction probability vectors, and the risk assessment level corresponding to the maximum value is taken as the target risk assessment level. For example, if the prediction probability vector obtained by softmax is (0.8, 0.5, 0.4, 0.2), the maximum value is 0.8, and it is determined that the power grid reserve is in an extremely tight state. It can be seen that the target risk assessment level of the current power grid operation state can be predicted by the target deep neural network model, and the classification and early warning of the tight balance state of the power grid can be realized.

[0076] Step 205, obtaining current feature data corresponding to different dimensions of the current power grid operation state.

[0077] Step 206, outputting multiple prediction probability values for different risk assessment levels by using the target deep neural network model according to the current feature data.

[0078] Step 207, taking the risk assessment level corresponding to the maximum prediction probability value in the multiple prediction probability values as the target risk assessment level.

[0079] By applying the technical solution of the embodiment, taking a provincial power grid as an example, the target deep neural network model can extract effective features from multi-dimensional power grid feature data, accurately predict the tight balance state of the power grid, and accurately distinguish the risk state from the non-risk state in the power grid operation state, with an accuracy rate reaching the expected target. It can be seen that the improved target deep neural network model not only performs well in sample construction and feature selection, but also achieves satisfactory results in prediction accuracy.

[0080] It can be seen that by optimizing the structure of the first deep neural network model, a second deep neural network model is constructed; a full information element data model is constructed for feature data corresponding to different dimensions of the power grid operation state; based on the full information element data model, the second deep neural network model is trained to obtain a target deep neural network model; and based on the target deep neural network model, a target risk assessment level of the current power grid operation state is predicted. Compared with the prior art that cannot predict the power grid operation risk state in advance, the embodiment constructs a TransForm model for predicting the power grid operation risk state through structural optimization, and constructs a data model containing full information elements, so as to fully consider the influence of power grid feature data and environmental feature data on the power grid operation state, and provide a data basis for subsequent sample data construction and model training. It can be seen that accurate prediction of the power grid operation state can help power grid operators to timely adjust the allocation of power resources to optimize the operation strategy of the power grid.

[0081] Further, as Figure 1 A specific implementation of the method, the embodiment of the present application provides a power grid operation state risk assessment device, as Figure 5 shown, the device comprises a first construction module 51, a second construction module 52, a training module 53, and a prediction module 54.

[0082] The first construction module 51 is configured to optimize the structure of the first deep neural network model to construct a second deep neural network model.

[0083] The second construction module 52 is configured to construct a full information element data model for feature data corresponding to different dimensions of the power grid operation state.

[0084] The training module 53 is configured to train the second deep neural network model based on the full information element data model to obtain a target deep neural network model.

[0085] The prediction module 54 is configured to predict a target risk assessment level of the current power grid operation state based on the target deep neural network model.

[0086] In a specific application scenario, as Figure 6 shown, the first construction module 51 comprises a first replacement sub-module 511 and a second replacement sub-module 512.

[0087] The first replacement sub-module 511 is configured to add a multi-scale hollow convolution layer in the multi-head self-attention mechanism of the first deep neural network model to obtain a multi-scale hollow attention mechanism.

[0088] The second replacement sub-module 512 is configured to replace the multi-layer perception in the first deep neural network model with a convolution layer to obtain a second deep neural network model.

[0089] In a specific application scenario, the second deep neural network model fuses a plurality of encoding characteristic vectors obtained by encoding different stages of the multi-head self-attention mechanism as supplementary feature vectors with an initial input vector of a decoding stage to obtain a target input vector of the decoding stage.

[0090] In a specific application scenario, the feature data corresponding to different dimensions of the power grid operation state includes power grid feature data and environmental feature data; the power grid feature data includes all-network positive and negative backup state data and key section operation state data; the all-network positive and negative backup state data includes one or more of all-network power generation data, all-network load data, all-network network loss data, all-network positive and negative backup capacity, all-network positive and negative backup margin, maximum external power transmission power, and external power load ratio; the key section operation state data includes one or more of key section load rate under a target partition, DC near-zone first and second section power flow load rate, key power transmission channel total power, key power transmission channel maximum transmission power, and key power transmission channel margin; and the environmental feature data includes one or more of temperature, humidity, rainfall, wind speed, and solar radiation intensity.

[0091] In a specific application scenario, the training module 53 includes a generation sub-module 531 and a training sub-module 532.

[0092] The generation sub-module 531 is configured to generate simulation data of a preset time period according to the feature data in the all-information element data model, and use the simulation data as sample data for training the second deep neural network model.

[0093] The training sub-module 532 is configured to train the second deep neural network model according to the sample data to obtain a target deep neural network model.

[0094] In a specific application scenario, the training sub-module 532 is specifically configured to preset a plurality of output results of the second deep neural network model, each output result corresponding to a different risk assessment level; and iteratively train the second deep neural network model based on the sample data until the output result of the second deep neural network model is consistent with a preset risk assessment label of the power grid operation state, to obtain the target deep neural network model.

[0095] In a specific application scenario, the prediction module 54 includes an acquisition sub-module 541, an output sub-module 542, and a selection sub-module 543.

[0096] The acquisition sub-module 541 is configured to acquire current feature data corresponding to different dimensions of a current power grid operation state.

[0097] The output submodule 542 is configured to output, according to the current feature data, a plurality of predicted probability values for different risk assessment levels by using a target deep neural network model.

[0098] The selection submodule 543 is configured to select a risk assessment level corresponding to a maximum predicted probability value in the plurality of predicted probability values as a target risk assessment level.

[0099] It should be noted that other corresponding descriptions of the various functional units involved in the risk assessment device for power grid operation state provided by the embodiments of the present application can be referred to the corresponding descriptions in Figure 1 and Figure 2 , which will not be repeated here.

[0100] Based on the above method as shown in Figure 1 and Figure 2 , accordingly, the embodiments of the present application also provide a computer storage medium, which stores a computer program, and the program is executed by a processor to implement the above-mentioned power grid operation state risk assessment method as shown in Figure 1 and Figure 2 .

[0101] Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.) and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in various implementation scenarios of the present application.

[0102] Based on the above method as shown in Figure 1 , Figure 2 , and Figure 5 , Figure 6 the virtual device embodiment, in order to achieve the above-mentioned purpose, the embodiments of the present application also provide a computer device, which can be a personal computer, a server, a network device, etc., and the entity device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the above-mentioned power grid operation state risk assessment method as shown in Figure 1 and Figure 2 .

[0103] Optionally, the computer device can further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, and the like. The user interface can include a display, an input unit such as a keyboard, and the like. Optionally, the user interface can further include a USB interface, a card reader interface, and the like. The network interface can optionally include a standard wired interface, a wireless interface (such as a Bluetooth interface, a WI-FI interface), and the like.

[0104] Those skilled in the art can understand that the computer device structure provided by the embodiment does not constitute a limitation on the entity device, and can include more or fewer components, or combine certain components, or different component arrangements.

[0105] The storage medium can further include an operating system and a network communication module. The operating system is a program for managing hardware and software resources of the computer device, and supports the running of information processing programs and other software and / or programs. The network communication module is used to realize communication between components in the storage medium, and communication with other hardware and software in the entity device.

[0106] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and necessary general hardware platforms, or by hardware. By applying the technical solutions of the present application, compared with the technical solutions in which the power grid operation state risk system cannot comprehensively and comprehensively perceive the power grid operation state risk in the prior art, the embodiment constructs a TransForm model for predicting the power grid operation risk state through structural optimization, and constructs a data model containing full information elements, so as to fully consider the influence of power grid characteristic data and environmental characteristic data on the power grid operation state, and provide a data basis for subsequent construction of sample data and model training. It can be seen that accurate power grid operation state prediction can help power grid operators to timely adjust the allocation of power resources, so as to optimize the operation strategy of the power grid.

[0107] Those skilled in the art can understand that the drawings are only a schematic diagram of a preferred implementation scenario, and the modules or flows in the drawings are not necessarily required for implementing the present application. Those skilled in the art can understand that the modules in the device in the implementation scenario can be distributed in the device in the implementation scenario according to the description of the implementation scenario, or can be changed and located in one or more devices different from the implementation scenario. The modules of the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0108] The above application number is only for description, and does not represent the advantages and disadvantages of the implementation scene. The above disclosure is only some specific implementation scenes of the application, but the application is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the application.

Claims

1. A method for risk assessment of power grid operating status, characterized in that, include: By optimizing the structure of the first deep neural network model, a second deep neural network model is constructed. A comprehensive information element data model is constructed based on the characteristic data corresponding to different dimensions of power grid operation status. Based on the full information element data model, the second deep neural network model is trained to obtain the target deep neural network model; Based on the target deep neural network model, predict the target risk assessment level of the current power grid operation status; The first deep neural network model is a TransFormer model. The steps to construct the second deep neural network model by optimizing the structure of the first deep neural network model include: By adding a multi-scale dilated convolutional layer to the multi-head self-attention mechanism of the first deep neural network model, a multi-scale dilated attention mechanism is obtained. The second deep neural network model is obtained by replacing the multilayer perceptron in the first deep neural network model with convolutional layers.

2. The method according to claim 1, characterized in that, The second deep neural network model uses multiple encoded feature vectors obtained from different stages of the multi-head self-attention mechanism as supplementary feature vectors and fuses them with the initial input vector of the decoding stage to obtain the target input vector of the decoding stage.

3. The method according to claim 1, characterized in that, The characteristic data corresponding to different dimensions of the power grid operation status include power grid characteristic data and environmental characteristic data; the power grid characteristic data includes the overall grid positive and negative reserve status data and the key section operation status data; the overall grid positive and negative reserve status data includes one or more of the following: overall grid power generation data, overall grid load data, overall grid loss data, overall grid positive and negative reserve capacity, overall grid positive and negative reserve margin, maximum external power transmission power, and the proportion of external power to load; the key section operation status data includes one or more of the following: target zone offline key section load rate, DC near-area primary and secondary section power flow load rate, total power of key transmission channels, maximum transmission power of key transmission channels, and key transmission channel margin; the environmental characteristic data includes one or more of the following: temperature, humidity, rainfall, wind speed, and solar irradiance.

4. The method according to claim 1 or 3, characterized in that, The step of training the second deep neural network model based on the full-information element data model to obtain the target deep neural network model includes: Based on the feature data in the full information element data model, simulation data with a preset time period is generated and used as sample data for training the second deep neural network model. Based on the sample data, the second deep neural network model is trained to obtain the target deep neural network model.

5. The method according to claim 4, characterized in that, The step of training the second deep neural network model to obtain the target deep neural network model based on the sample data includes: Multiple outputs of the second deep neural network model are preset, and each output corresponds to a different risk assessment level; The second deep neural network model is iteratively trained based on the sample data until the output of the second deep neural network model is consistent with the preset risk assessment label of the power grid operation status, thus obtaining the target deep neural network model.

6. The method according to claim 1, characterized in that, The step of predicting the target risk assessment level of the current power grid operating state based on the target deep neural network model includes: Obtain current feature data corresponding to different dimensions of the current power grid operation status; Based on the current feature data, the target deep neural network model is used to output multiple predicted probability values ​​for different risk assessment levels; The risk assessment level corresponding to the highest predicted probability value among the multiple predicted probability values ​​is taken as the target risk assessment level.

7. A risk assessment device for power grid operating status, characterized in that, include: The first building module is used to construct a second deep neural network model by optimizing the structure of the first deep neural network model; The second construction module is used to build a full-information element data model for the feature data corresponding to different dimensions of the power grid operation status; The training module is used to train the second deep neural network model to obtain the target deep neural network model based on the constructed full-information element data model. The prediction module is used to predict the target risk assessment level of the current power grid operation status based on the target deep neural network model. Wherein, the first deep neural network model is the TransFormer model, and the first building module includes a first replacement submodule and a second replacement submodule; The first replacement submodule is used to add a multi-scale dilated convolutional layer to the multi-head self-attention mechanism of the first deep neural network model to obtain a multi-scale dilated attention mechanism. The second replacement submodule is used to replace the multilayer perceptron in the first deep neural network model with a convolutional layer to obtain a second deep neural network model.

8. A computer storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the risk assessment method for the power grid operating status as described in any one of claims 1 to 6.

9. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the risk assessment method for the power grid operating status as described in any one of claims 1 to 6.

Images