Method and device for generating key parameter prediction model and method and device for generating stable key parameters

Abstract

The invention discloses a method and device for generating a key parameter prediction model and a method and device for generating stable key parameters, and the method for generating the key parameter prediction model comprises the steps: obtaining the stable key parameters and a stable time sequence based on a transient stability file sample; and training an initialized key parameter prediction model by using the stable time sequence as input and the stable key parameter as a label to obtain a final key parameter prediction model. According to the method and the device provided by the embodiment of the invention, the real-time fault and the power grid operation mode change can be quickly responded, the key parameters can be modified in real time, and the latest and most reliable stable key parameters can be provided for dispatching personnel.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method and device for generating a key parameter prediction model and generating stable key parameters. Background Art

[0002] With the development of the power system, from traditional systems and operation goals to today's clean energy, new energy, and integrated energy layout energy forms, from mechanical, electromagnetic, electronic, and power electronics hybrid systems and controls, from simple power quality goals to new energy efficiency goals of integrated energy, the requirements for power construction, operation, and management are becoming increasingly higher, requiring high-precision supervision, precise control, and effective prevention and protection technologies and methods.

[0003] With increasingly complex power systems and increasingly high operational control requirements, it is difficult to control the complex characteristics of power system operation by relying solely on professional technical experts. Although there are some large-scale intelligent stability identification technologies, they all have various problems: they only contain end-to-end algorithms, with poor interpretability and difficulty in application; the training samples cannot fully reflect the situation of real data under extreme boundary conditions, and the effect is poor under boundary conditions; the degree of versatility is poor, and different structures need to be redesigned for different situations. Summary of the invention

[0004] In view of this, the present invention proposes a method and device for generating a key parameter prediction model and generating stable key parameters, aiming to solve the above problems.

[0005] In a first aspect, an embodiment of the present invention provides a method for generating a key parameter prediction model, the method comprising: obtaining stable key parameters and stable time series based on a temporarily stable file sample; using the stable time series as input and the stable key parameters as labels, training an initialized key parameter prediction model to obtain a final key parameter prediction model; wherein the key parameter prediction model comprises an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer comprises a double-layer LSTM model.

[0006] Furthermore, the method of obtaining stable key parameters and stable time series based on the temporary stable file samples includes: batch generating samples in a stable boundary state by dynamically modifying the fault settings in the temporary stable file samples, and obtaining stable key parameter indicators in the stable boundary state, as well as obtaining the power grid characterization at each moment based on each key feature of the temporary stable file samples, and selecting the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed to combine into a stable time series.

[0007] Furthermore, the method dynamically modifies the fault settings in the temporarily stable file samples, generates batches of samples in a stable boundary state, and obtains stable key parameter indicators in the stable boundary state, including: changing the severity of the fault by adjusting the fault duration, and using a binary search for the fault duration in a binary manner, each time modifying the fault duration of the temporarily stable file sample to the current binary value and re-performing a stability judgment to determine whether it is currently unstable, obtaining samples in a stable boundary state, and obtaining stable key parameter indicators based on the samples in the stable boundary state.

[0008] Furthermore, the power grid representation at each moment is obtained based on each key feature of the transient stability file sample, including: based on each key feature of the transient stability sample, key information is extracted step by step, and feature vectors are distilled to obtain a power grid representation that can represent key information of the power grid stability form at each moment.

[0009] Furthermore, based on each key feature of the transient stability sample, key information is extracted step by step, and the feature vector is distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment, including: vectorizing each key feature of the power grid in the transient stability sample to obtain a feature representation; aggregating the feature representations of the same branch to obtain a branch representation; aggregating all branch representations of the entire power grid to obtain a power grid representation.

[0010] Furthermore, the parameter output layer includes an attention network and a multi-layer fully connected layer.

[0011] In the second aspect, an embodiment of the present invention further provides a method for generating key stability parameters, which are key stability parameters used for stability judgment and control in the power system. The method comprises: based on the grid characterization of the real-time measurement or imaging system at each moment, judging the time when the fault occurs and the time when the fault is removed and obtaining a stable time series; inputting the stable time series into a key parameter prediction model to obtain a prediction result of the key stability parameters, wherein the key parameter prediction model is obtained using the methods provided in the above-mentioned embodiments.

[0012] In the third aspect, an embodiment of the present invention also provides a device for generating a key parameter prediction model, the device comprising: a processing unit, used to obtain stable key parameters and stable time series based on temporary stable file samples; a training unit, used to use the stable time series as input and the stable key parameters as labels to train the initialized key parameter prediction model to obtain the final key parameter prediction model; wherein the key parameter prediction model comprises an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer comprises a double-layer LSTM model.

[0013] Furthermore, the processing unit is also used to: batch generate samples in a stable boundary state by dynamically modifying the fault settings in the temporary stable file samples, and obtain stable key parameter indicators in the stable boundary state, as well as various key features of the temporary stable file samples to obtain the power grid characterization at each moment, and select the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed to combine into a stable time series.

[0014] Furthermore, the method dynamically modifies the fault settings in the temporarily stable file samples, generates batches of samples in a stable boundary state, and obtains stable key parameter indicators in the stable boundary state, including: changing the severity of the fault by adjusting the fault duration, and using a binary search for the fault duration in a binary manner, each time modifying the fault duration of the temporarily stable file sample to the current binary value and re-performing a stability judgment to determine whether it is currently unstable, obtaining samples in a stable boundary state, and obtaining stable key parameter indicators based on the samples in the stable boundary state.

[0015] Furthermore, the power grid representation at each moment is obtained based on each key feature of the transient stability file sample, including: based on each key feature of the transient stability sample, key information is extracted step by step, and feature vectors are distilled to obtain a power grid representation that can represent key information of the power grid stability form at each moment.

[0016] Furthermore, based on each key feature of the transient stability sample, key information is extracted step by step, and the feature vector is distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment, including: vectorizing each key feature of the power grid in the transient stability sample to obtain a feature representation; aggregating the feature representations of the same branch to obtain a branch representation; aggregating all branch representations of the entire power grid to obtain a power grid representation.

[0017] Furthermore, the parameter output layer includes an attention network and a multi-layer fully connected layer.

[0018] In a fourth aspect, an embodiment of the present invention further provides a device for generating key stability parameters, wherein the key stability parameters are key parameters used for stability judgment and control in a power system, and the device comprises: an acquisition unit, used to judge the time when a fault occurs and the time when a fault is removed based on the grid characterization of the real-time measurement or imaging system at each moment, and to obtain a stable time series; a prediction unit, used to input the stable time series into a key parameter prediction model to obtain a prediction result of the key stability parameters, wherein the key parameter prediction model is obtained using the devices provided in the above-mentioned embodiments.

[0019] In a fifth aspect, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the methods provided in the above embodiments are implemented.

[0020] In a sixth aspect, an embodiment of the present invention further provides an electronic device, comprising: a processor; a memory for storing executable instructions of the processor; the processor is used to read the executable instructions from the memory and execute the instructions to implement the methods provided in the above embodiments.

[0021] The method and device for generating a key parameter prediction model provided by an embodiment of the present invention obtain stable key parameters and stable time series based on temporary stable file samples, and use the stable time series as input and the stable key parameters as labels to train the initialized key parameter prediction model to obtain the final key parameter prediction model. The key parameter prediction model can quickly respond to real-time faults and changes in power grid operation modes and modify key parameters in real time, thereby providing dispatchers with the latest and most reliable stable key parameters.

[0022] The method and device for generating stable key parameters provided in the embodiments of the present invention judge the time when a fault occurs and the time when a fault is removed and obtain a stable time series through the grid characterization at each moment based on the real-time measurement or imaging system, and input the stable time series into a key parameter prediction model to obtain a prediction result of the stable key parameters. The method and device can quickly respond to real-time faults and changes in grid operation modes and modify key parameters in real time, providing dispatchers with the latest and most reliable stable key parameters, and providing an effective solution for ensuring the continued safe and stable operation of my country's power grid, which is an important improvement to the existing power grid operation status. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 An exemplary flow chart of a method for generating a key parameter prediction model according to an embodiment of the present invention is shown;

[0024] Figure 2 A schematic diagram of the structure of a multi-level power grid representation according to an embodiment of the present invention is shown;

[0025] Figure 3 A schematic diagram of the structure of a key parameter prediction model according to an embodiment of the present invention is shown;

[0026] Figure 4 An exemplary flow chart of a method for generating stable key parameters according to an embodiment of the present invention is shown;

[0027] Figure 5 A schematic diagram showing the structure of a device for generating a key parameter prediction model according to an embodiment of the present invention is shown;

[0028] Figure 6 A schematic structural diagram of a device for generating stable key parameters according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0029] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0030] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0031] Figure 1 An exemplary flow chart of a method for generating a key parameter prediction model according to an embodiment of the present invention is shown.

[0032] like Figure 1 As shown, the method includes:

[0033] Step S101: Based on the temporarily stable file sample, obtain stable key parameters and stable time series.

[0034] Furthermore, step S101 includes:

[0035] By dynamically modifying the fault settings in the temporary stability file samples, batches of samples in the stable boundary state are generated, and the key stability parameter indicators in the stable boundary state are obtained. Based on the key features of the temporary stability file samples, the power grid characterization at each moment is obtained, and the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed is combined into a stable time series.

[0036] Furthermore, by dynamically modifying the fault settings in the temporary stable file samples, batches of samples in the stable boundary state are generated, and stable key parameter indicators in the stable boundary state are obtained, including:

[0037] The severity of the fault is changed by adjusting the fault duration, and the fault duration is binary searched using the binary search method. Each time, the fault duration of the temporary stable file sample is modified to the current binary value and the stability judgment is re-performed to determine whether it is currently unstable. Samples in the stable boundary state are obtained, and stable key parameter indicators are obtained based on the samples in the stable boundary state.

[0038] First, batch generate stable boundary samples from temporary stable file samples. By modifying the PSASP51 format file content, dynamically modify the fault settings in the temporary stable file samples, such as fault location, fault duration, etc., batch generate samples in the stable boundary state, obtain the key thresholds, key coefficients and other stable key parameters of the key parameter prediction model under the stable boundary state, and organize these samples into training samples.

[0039] The principle of generating samples in a stable boundary state is mainly to change the severity of the fault by adjusting the fault duration. Since the severity of the fault will increase continuously with the increase of the fault duration, the fault duration can be searched in binary by using the binary search method. Each time, the fault duration of the sample is modified to the current binary value and the stability judgment is re-run to determine whether it is currently unstable. Finally, samples in a stable boundary state (currently stable, but a little longer fault duration will cause instability) are obtained, and at the same time, based on this state, the stability key parameters of the corresponding mechanism model in this state are obtained.

[0040] Furthermore, based on the key features of the temporary stability file samples, the grid representation at each moment is obtained, including:

[0041] Based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment.

[0042] Furthermore, based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain the grid representation that can represent the key information of the grid stability form at each moment, including:

[0043] Each key feature of the power grid in the transient stability sample is vectorized to obtain feature representation;

[0044] Aggregate the feature representations of the same branch to obtain the branch representation;

[0045] All branch representations of the entire power grid are aggregated to obtain the power grid representation.

[0046] Figure 2 FIG. 2 shows a schematic diagram of a multi-level power grid representation according to an embodiment of the present invention. Figure 2As shown, the embodiment of the present invention uses a multi-level power grid characterization structure to characterize power grid information. The multi-level power grid characterization adopts a three-level characterization method. The first level of characterization is the power grid feature level characterization, including voltage, frequency, power angle, etc. The second level of characterization is the branch level characterization. Each branch collects all the power grid feature quantities related to itself and both ends, and merges them to generate a branch level characterization; then the information representations of multiple key branches are integrated with each other to finally form a power grid characterization. Specifically, the multi-level power grid characterization is as follows:

[0047] Characteristic representation: The characterization obtained by quantizing the key characteristics of the power grid, such as voltage amplitude and voltage phase angle, represents the impact of a key characteristic at a certain moment;

[0048] Branch representation: A branch representation formed by the aggregation of feature representations of the same branch, indicating the influence of the overall feature quantity of the branch at a certain moment;

[0049] Grid representation: The grid representation is aggregated from the representations of all key branches of the entire grid, indicating the impact of the overall characteristic quantity of the entire grid at a certain moment.

[0050] Furthermore, the grid representations in the neighborhood time before and after the fault occurs and before and after the fault is removed are selected and combined into a stable time series, including:

[0051] The key node time series is used as the input of the key parameter prediction model, where the key node time series is composed of the neighborhood sequence of multiple time nodes: first, the fault occurrence time and the fault removal time are determined, and then the power grid characterization data within Δt before and after the fault occurrence and within Δt before and after the fault removal are taken out and integrated to form a new time series.

[0052] Step S102: using the stable time series as input and the stable key parameters as labels, training the initialized key parameter prediction model to obtain the final key parameter prediction model;

[0053] The key parameter prediction model includes an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer includes a double-layer LSTM model.

[0054] Furthermore, the parameter output layer includes an attention network and multiple fully connected layers.

[0055] Figure 3 FIG. 4 is a schematic diagram showing the structure of a key parameter prediction model according to an embodiment of the present invention. Figure 3As shown, the embodiment of the present invention adopts an improved LSTM as the long-term and short-term information capture module of the key parameter time series of the parameter prediction model. LSTM is a deep learning model for processing sequence data, which can effectively capture the long-term time series feature correlation and the short-term neighborhood feature correlation at the same time. A two-layer LSTM model is used to process the power grid characterization time series generated in the previous article. Its hidden layer effectively captures the long-term data correlation in the time series, while its output layer can capture the data correlation in the short-term neighborhood, and finally feeds back to the next model for key feature generation. The embodiment of the present invention adopts an attention mechanism network plus a multi-layer fully connected network as the parameter generation module of the parameter prediction model: first, a layer of attention mechanism network is used to capture the key information of the LSTM hidden layer and output layer, and then a multi-layer fully connected network is used to map the key information to the stable key parameters of the output, and finally complete the key parameter generation.

[0056] For the training of the key parameter prediction model, Adam optimization can be used, with the loss function being MSE and the learning rate exponentially decaying to 0.001 to eventually reach model convergence. The epoch with the best effect on the validation set is taken as the final model parameter.

[0057] In the above-mentioned embodiment, stable key parameters and stable time series are obtained based on the temporary stable file samples, and the stable time series is used as input and the stable key parameters are used as labels to train the initialized key parameter prediction model to obtain the final key parameter prediction model. The key parameter prediction model can quickly respond to real-time faults and changes in power grid operation modes and modify key parameters in real time, thereby providing dispatchers with the latest and most reliable stable key parameters.

[0058] Figure 4 An exemplary flow chart of a method for generating stable key parameters according to an embodiment of the present invention is shown.

[0059] like Figure 4 As shown, the key stability parameter is a key parameter used for stability identification and control in the power system. The method includes:

[0060] Step S401: based on the real-time measurement or the grid representation of the imaging system at each moment, determine the fault occurrence moment and the fault removal moment and obtain a stable time series;

[0061] Step S402: inputting the stable time series into the key parameter prediction model to obtain the prediction result of the stable key parameter, wherein the key parameter prediction model is obtained by using the method for generating the key parameter prediction model provided in the above-mentioned embodiments.

[0062] The above-mentioned embodiment judges the time when a fault occurs and the time when a fault is removed and obtains a stable time series through the grid characterization at each moment based on the real-time measurement or imaging system, and inputs the stable time series into the key parameter prediction model to obtain the prediction results of stable key parameters. It can quickly respond to real-time faults and changes in grid operation modes and modify key parameters in real time, providing dispatchers with the latest and most reliable stable key parameters, and providing an effective solution for ensuring the continuous safe and stable operation of my country's power grid, which is an important improvement to the existing power grid operation status.

[0063] Figure 5 A schematic structural diagram of an apparatus for generating a key parameter prediction model according to an embodiment of the present invention is shown.

[0064] like Figure 5 As shown, the device comprises:

[0065] The processing unit 501 is used to obtain stable key parameters and stable time series based on the temporary stable file sample;

[0066] A training unit 502 is used to train the initialized key parameter prediction model using the stable time series as input and the stable key parameters as labels to obtain a final key parameter prediction model;

[0067] The key parameter prediction model includes an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer includes a double-layer LSTM model.

[0068] Furthermore, the processing unit 501 is further configured to:

[0069] By dynamically modifying the fault settings in the temporary stability file samples, batches of samples in the stable boundary state are generated, and the key stability parameter indicators in the stable boundary state are obtained. Based on the key features of the temporary stability file samples, the power grid characterization at each moment is obtained, and the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed is combined into a stable time series.

[0070] Furthermore, by dynamically modifying the fault settings in the temporary stable file samples, batches of samples in the stable boundary state are generated, and stable key parameter indicators in the stable boundary state are obtained, including:

[0071] The severity of the fault is changed by adjusting the fault duration, and the fault duration is binary searched using the binary search method. Each time, the fault duration of the temporary stable file sample is modified to the current binary value and the stability judgment is re-performed to determine whether it is currently unstable. Samples in the stable boundary state are obtained, and stable key parameter indicators are obtained based on the samples in the stable boundary state.

[0072] Furthermore, based on the key features of the temporary stability file samples, the grid representation at each moment is obtained, including:

[0073] Based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment.

[0074] Furthermore, based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain the grid representation that can represent the key information of the grid stability form at each moment, including:

[0075] Each key feature of the power grid in the transient stability sample is vectorized to obtain feature representation;

[0076] Aggregate the feature representations of the same branch to obtain the branch representation;

[0077] All branch representations of the entire power grid are aggregated to obtain the power grid representation.

[0078] Furthermore, the parameter output layer includes an attention network and multiple fully connected layers.

[0079] In the above-mentioned embodiment, stable key parameters and stable time series are obtained based on the temporary stable file samples, and the stable time series is used as input and the stable key parameters are used as labels to train the initialized key parameter prediction model to obtain the final key parameter prediction model. The key parameter prediction model can quickly respond to real-time faults and changes in power grid operation modes and modify key parameters in real time, thereby providing dispatchers with the latest and most reliable stable key parameters.

[0080] Figure 6 A schematic structural diagram of a device for generating stable key parameters according to an embodiment of the present invention is shown.

[0081] like Figure 6 As shown, the key stability parameter is a key parameter used for stability identification and control in the power system. The device includes:

[0082] An acquisition unit 601 is used to determine the fault occurrence time and the fault removal time and acquire a stable time series based on the power grid representation of each moment of the real-time measurement or imaging system;

[0083] The prediction unit 602 is used to input the stable time series into the key parameter prediction model to obtain the prediction result of the stable key parameter, wherein the key parameter prediction model is obtained by using the device for generating the key parameter prediction model provided by the above-mentioned embodiments.

[0084] The above-mentioned embodiment judges the time when a fault occurs and the time when a fault is removed and obtains a stable time series through the grid characterization at each moment based on the real-time measurement or imaging system, and inputs the stable time series into the key parameter prediction model to obtain the prediction results of stable key parameters. It can quickly respond to real-time faults and changes in grid operation modes and modify key parameters in real time, providing dispatchers with the latest and most reliable stable key parameters, and providing an effective solution for ensuring the continuous safe and stable operation of my country's power grid, which is an important improvement to the existing power grid operation status.

[0085] It should be noted that the device provided in the above embodiment only uses the division of the above functional modules as an example to implement its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and their specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0086] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for generating a key parameter prediction model or the method for generating stable key parameters provided in the above-mentioned embodiments is implemented.

[0087] An embodiment of the present invention also provides an electronic device, comprising: a processor; a memory for storing processor executable instructions; the processor is used to read the executable instructions from the memory and execute the instructions to implement the method for generating a key parameter prediction model or the method for generating stable key parameters provided in the above-mentioned embodiments.

[0088] The invention has been described above with reference to a few embodiments. However, it is readily apparent to a person skilled in the art that other embodiments than the ones disclosed above are equally within the scope of the invention, as defined by the appended patent claims.

[0089] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a / said / the [means, components, etc.]" are to be openly interpreted as at least one instance of the means, components, etc., unless explicitly stated otherwise. The steps of any method disclosed herein do not necessarily have to be performed in the exact order disclosed, unless explicitly stated otherwise.

[0090] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0092] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0093] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for generating a key parameter prediction model, characterized in that: The method comprises: Based on the temporary stable file samples, stable key parameters and stable time series are obtained; Using the stable time series as input and the stable key parameters as labels, training an initialized key parameter prediction model to obtain a final key parameter prediction model; The key parameter prediction model includes an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer includes a double-layer LSTM model.

2. The method according to claim 1, characterized in that The stable key parameters and stable time series are obtained based on the temporary stable file sample, including: By dynamically modifying the fault settings in the temporary stability file samples, batches of samples in the stable boundary state are generated, and the key stability parameter indicators in the stable boundary state are obtained. Based on the key features of the temporary stability file samples, the power grid characterization at each moment is obtained, and the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed is combined into a stable time series.

3. The method according to claim 2, characterized in that The method of dynamically modifying the fault settings in the temporary stable file sample, batch-generating samples in a stable boundary state, and obtaining stable key parameter indicators in the stable boundary state includes: The severity of the fault is changed by adjusting the fault duration, and the fault duration is binary searched using the binary search method. Each time, the fault duration of the temporarily stable file sample is modified to the current binary value and a new stability judgment is performed to determine whether it is currently unstable, and samples in a stable boundary state are obtained. Based on the samples in the stable boundary state, stability key parameter indicators are obtained.

4. The method according to claim 2, characterized in that: The power grid representation at each moment is obtained based on the key features of the temporary stability file sample, including: Based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment.

5. The method according to claim 4, characterized in that The key information is extracted step by step based on each key feature of the transient stability sample, and the feature vector is distilled to obtain the power grid representation that can represent the key information of the power grid stability form at each moment, including: Each key feature of the power grid in the transient stability sample is vectorized to obtain feature representation; Aggregate the feature representations of the same branch to obtain the branch representation; All branch representations of the entire power grid are aggregated to obtain the power grid representation.

6. The method according to claim 1, characterized in that The parameter output layer includes an attention network and a multi-layer fully connected layer.

7. A method for generating stable key parameters, characterized in that: The key stability parameter is a key parameter used for stability identification and control in the power system, and the method includes: Based on the grid characterization at each moment of the real-time measurement or imaging system, the fault occurrence and fault removal moments are determined and a stable time series is obtained; The stable time series is input into a key parameter prediction model to obtain a prediction result of a stable key parameter, wherein the key parameter prediction model is obtained by using the method described in any one of claims 1-6.

8. A device for generating a key parameter prediction model, characterized in that: The device comprises: A processing unit, used for obtaining stable key parameters and stable time series based on the temporary stable file samples; A training unit, used to train an initialized key parameter prediction model using the stable time series as input and the stable key parameters as labels to obtain a final key parameter prediction model; The key parameter prediction model includes an input layer, a sequence processing layer and a parameter output layer, wherein the sequence processing layer includes a double-layer LSTM model.

9. The device according to claim 8, characterized in that The processing unit is further used for: By dynamically modifying the fault settings in the temporary stability file samples, batches of samples in the stable boundary state are generated, and the key stability parameter indicators in the stable boundary state are obtained. Based on the key features of the temporary stability file samples, the power grid characterization at each moment is obtained, and the power grid characterization in the neighborhood time before and after the fault occurs and before and after the fault is removed is combined into a stable time series.

10. The device according to claim 9, characterized in that The method of dynamically modifying the fault settings in the temporary stable file sample, batch-generating samples in a stable boundary state, and obtaining stable key parameter indicators in the stable boundary state includes: The severity of the fault is changed by adjusting the fault duration, and the fault duration is binary searched using the binary search method. Each time, the fault duration of the temporarily stable file sample is modified to the current binary value and a new stability judgment is performed to determine whether it is currently unstable, and samples in a stable boundary state are obtained. Based on the samples in the stable boundary state, stability key parameter indicators are obtained.

11. The device according to claim 9, characterized in that The power grid representation at each moment is obtained based on the key features of the temporary stability file sample, including: Based on the key features of transient stability samples, key information is extracted step by step, and the feature vectors are distilled to obtain a power grid representation that can represent the key information of the power grid stability form at each moment.

12. The device according to claim 11, characterized in that The key information is extracted step by step based on each key feature of the transient stability sample, and the feature vector is distilled to obtain the power grid representation that can represent the key information of the power grid stability form at each moment, including: Each key feature of the power grid in the transient stability sample is vectorized to obtain feature representation; Aggregate the feature representations of the same branch to obtain the branch representation; All branch representations of the entire power grid are aggregated to obtain the power grid representation.

13. The device according to claim 8, characterized in that The parameter output layer includes an attention network and a multi-layer fully connected layer.

14. A device for generating stable key parameters, characterized in that: The key stability parameter is a key parameter used for stability identification and control in the power system. The device includes: An acquisition unit, used to determine the time of fault occurrence and the time of fault removal and acquire a stable time series based on the power grid representation of each moment of the real-time measurement or imaging system; A prediction unit is used to input the stable time series into a key parameter prediction model to obtain a prediction result of a stable key parameter, wherein the key parameter prediction model is obtained using the device described in any one of claims 8-13.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method described in any one of claims 1 to 6 or 7 is implemented.

16. An electronic device, comprising: processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1-6 or 7.

Images