Voltage disturbance detection method and device, readable storage medium and electronic equipment

Abstract

The application provides a voltage disturbance detection method and device, a readable storage medium and an electronic device. The method comprises the following steps: obtaining a pre-adjustment strategy of a first time period, the first time period being a preceding time period of a real-time time period, the pre-adjustment strategy being a strategy for pre-adjusting the real-time time period in the first time period, so as to reduce the degree of voltage disturbance in the real-time time period; obtaining first voltage information of a second time period, the second time period comprising the real-time time period and being a subsequent time period of the first time period; determining a first identification result based on the first voltage information and a pre-trained disturbance identification model, the first identification result representing a disturbance occurrence identified in the second time period; and determining a disturbance response method based on the first identification result and the pre-adjustment strategy, the disturbance response method being used for adjusting the pre-adjustment strategy, so that the voltage remains stable. In this way, by combining prediction and real-time monitoring and using the disturbance identification model, the detection efficiency and accuracy can be improved, and the response speed can be accelerated.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of voltage detection, and in particular to a voltage disturbance detection method and device, a readable storage medium and an electronic device. BACKGROUND

[0002] With the rapid development of precision manufacturing industry, the requirement for power quality is getting higher and higher. Power quality problems, especially voltage sag, have a significant impact on the production process and equipment operation of these industries. Voltage sag refers to the phenomenon that the power supply voltage suddenly drops in a short time and then quickly recovers. Such voltage fluctuation can cause the precision equipment to run unstably, and even be damaged, thereby affecting the production efficiency and product quality. However, there are still problems of low detection efficiency, low accuracy and slow response in the detection of voltage disturbance. SUMMARY

[0003] The present application provides a voltage disturbance detection method, device, readable storage medium and electronic device, which can improve the detection efficiency and accuracy and speed up the response speed.

[0004] In a first aspect, the present application provides a voltage disturbance detection method applied to an electronic device, and the method comprises:

[0005] obtaining a pre-adjustment strategy for a first time period, the first time period being a preceding time period of a real-time period, and the pre-adjustment strategy being a strategy for pre-adjusting the subsequent time period in the first time period to reduce the degree of voltage disturbance in the real-time period;

[0006] obtaining first voltage information of a second time period, the second time period including the real-time period and being the subsequent time period of the first time period;

[0007] determining a first identification result based on the first voltage information and a pre-trained disturbance identification model, the first identification result representing a disturbance occurrence identified in the second time period;

[0008] determining a disturbance coping method based on the first identification result and the pre-adjustment strategy, the disturbance coping method being used to adjust the pre-adjustment strategy to keep the voltage stable.

[0009] In a second aspect, the present application provides a voltage disturbance detection device applied to an electronic device, and the device comprises:

[0010] a first obtaining module, configured to obtain a pre-adjustment strategy for a first time period, the first time period being a preceding time period of a real-time period, and the pre-adjustment strategy being a strategy for pre-adjusting the subsequent time period in the first time period to reduce the degree of voltage disturbance in the real-time period;

[0011] a second obtaining module, configured to obtain first voltage information of a second time period, the second time period including the real-time time period, the first time period being the preceding time period of the real-time time period;

[0012] a first determining module, configured to determine a first identification result based on the first voltage information and a pre-trained disturbance identification model, the first identification result representing a disturbance occurrence identified in the second time period;

[0013] a second determining module, configured to determine a disturbance coping method based on the first identification result and the pre-adjustment strategy, the disturbance coping method being used to adjust the pre-adjustment strategy so as to keep voltage stable.

[0014] In a third aspect, an embodiment of the present application provides a computer readable storage medium, having stored thereon a voltage disturbance detection program, the voltage disturbance detection program comprising program instructions, when executed by a processor of an electronic device, the processor executes part or all of the steps described in the first aspect.

[0015] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, a communication interface, and one or more programs, the memory storing the one or more programs and being configured to be executed by the processor, when the processor executes the one or more programs stored in the memory, the processor executes the instructions of part or all of the steps described in the first aspect of the present application.

[0016] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a computer readable storage medium storing a voltage disturbance detection program, the voltage disturbance detection program being operable to cause a computer to execute part or all of the steps described in the first aspect of the present application. The computer program product can be a software installation package.

[0017] By implementing the embodiments of the present application, the pre-adjustment strategy of the first time period is obtained, the first time period being the preceding time period of the real-time time period, the pre-adjustment strategy being a strategy for pre-adjusting the real-time time period in the first time period, for reducing the degree of voltage disturbance in the real-time time period; the first voltage information of the second time period is obtained, the second time period including the real-time time period, the first time period being the preceding time period of the real-time time period; the first identification result is determined based on the first voltage information and the pre-trained disturbance identification model, the first identification result representing the disturbance occurrence identified in the second time period; and the disturbance coping method is determined based on the first identification result and the pre-adjustment strategy, the disturbance coping method being used to adjust the pre-adjustment strategy so as to keep voltage stable. In this way, by combining prediction and real-time monitoring and using the disturbance identification model, the detection efficiency and accuracy can be improved, and the response speed can be accelerated. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0019] Figure 1 This is a schematic diagram of the architecture of a voltage disturbance detection system provided in an embodiment of this application;

[0020] Figure 2 This is a flowchart of a voltage disturbance detection method provided in an embodiment of this application;

[0021] Figure 3 This is a model architecture diagram of a disturbance identification model provided in an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of the structure of a voltage disturbance detection device provided in an embodiment of this application;

[0023] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0025] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or electronic device that includes a series of steps or units is not limited to the listed steps or units, but in an alternative example also includes steps or units not listed, or in an alternative example also includes other steps or units inherent to these processes, methods, products, or electronic devices.

[0026] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0027] The voltage disturbance detection method, apparatus, readable storage medium, and electronic device provided in this application can be applied to, for example... Figure 1 Please refer to the voltage disturbance detection system shown. Figure 1 , Figure 1 This is a schematic diagram of the architecture of a voltage disturbance detection system provided in an embodiment of this application. The voltage disturbance detection system 100 includes an electronic device 110 and a data acquisition device 120.

[0028] In this design, the data acquisition device 120 can communicate with the electronic device 110 via a network. The data acquisition device 120 is a device used to acquire voltage data, such as a waveform recorder or sensor. The electronic device 110 is a remote computer used to process large amounts of computational tasks and store data; it can be a processor, a server, or a regular computer. In this solution, a disturbance identification model and a disturbance prediction model are deployed on the electronic device 110 to identify and predict disturbances in the data acquired by the data acquisition device 120. The electronic device 110 can also be used to collect data during model usage, facilitating subsequent optimization of the disturbance identification and prediction models.

[0029] Based on this, this application provides a voltage disturbance detection method, apparatus, readable storage medium, and electronic device, which will be described in detail below with reference to the accompanying drawings.

[0030] Please see Figure 2 , Figure 2 This is a flowchart of a voltage disturbance detection method provided in an embodiment of this application. This method is applied to electronic devices, such as... Figure 2 As shown, the method includes the following steps:

[0031] S210, Obtain the pre-adjustment strategy for the first time period, where the first time period is the preceding time period of the real-time time period, and the pre-adjustment strategy is a strategy for performing pre-adjustment operations on the preceding time period during the first time period to reduce the degree of voltage disturbance in the real-time time period.

[0032] The first time period is the preceding time period of the real-time time period. The disturbance prediction result is obtained by performing disturbance prediction on the next successive time period in the first time period. When time flows from the first time period to the next successive time period, that is, when it is in the next successive time period, the time period is the time period. When it is in the real-time time period, the previous time period has already performed disturbance prediction on the current time period and obtained the disturbance prediction result, and the pre-adjustment strategy has been determined based on the disturbance prediction result.

[0033] The first time period can be of any length and can be understood as the detection cycle.

[0034] In one possible embodiment, before the step of obtaining the pre-adjustment strategy for the first time period, the method further includes: obtaining a disturbance prediction result for the first time period, the disturbance prediction result representing disturbance occurrence information of the predicted successive time period during the first time period; determining the pre-adjustment strategy based on the disturbance prediction result; and performing a pre-adjustment operation based on the pre-adjustment strategy.

[0035] The disturbance prediction result is the disturbance occurrence information predicted in the first time period, i.e., the current real-time time period. This disturbance prediction result can be obtained from a prediction model, specifically a trained neural network model. Then, a pre-adjustment strategy is determined based on the obtained disturbance prediction result. This pre-adjustment strategy is used to perform pre-adjustment operations to promptly overcome and respond to disturbances occurring in the second time period. Specifically, the disturbance prediction result may include, but is not limited to, the predicted disturbance duration, the predicted disturbance severity, and the disturbance factors. Disturbance factors are those that can cause voltage disturbances, such as severe weather, equipment failure, or increases or decreases in load.

[0036] As can be seen, in this embodiment, disturbance prediction is performed in the time period before each current real-time time period, and a pre-adjustment strategy is determined, so that a rapid response can be made according to the pre-adjustment strategy before a disturbance may occur, thereby reducing the possibility of disturbances.

[0037] In one possible embodiment, obtaining the disturbance prediction result for the first time period includes: obtaining second voltage information and influencing factor information for the first time period, wherein the influencing factor information is a possible factor affecting the voltage; determining a second identification result based on the second voltage information and the pre-trained disturbance identification model, wherein the second identification result characterizes the disturbance occurrence identified in the first time period; and determining the disturbance prediction result based on the second identification result, the influencing factor information, and the pre-trained disturbance prediction model.

[0038] The second voltage information refers to the voltage information obtained in the first time period. This second voltage information can be a voltage curve, specifically a time-voltage curve or a time-voltage graph, etc., without limitation. The influencing factor information refers to information affecting voltage disturbances obtained simultaneously when acquiring the second voltage information, and may include, but is not limited to, weather conditions, equipment malfunctions, and increases or decreases in load.

[0039] The pre-trained disturbance recognition model can be a convolutional neural network-based model, specifically a model built on a VGG neural network. This model is used to perform a recognition task based on voltage information to obtain a recognition result indicating whether a voltage disturbance has occurred. In this embodiment, the second recognition result is obtained by recognizing the second voltage information, which represents the disturbance occurrence status identified in the first time period. The pre-trained disturbance prediction model can be a neural network-based prediction model, a random forest model, or other similar models. Specifically, it can include, but is not limited to, the following types: Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM), Residual Network (ResNet), etc.

[0040] Specifically, the training method of the pre-trained disturbance prediction model may include the following steps: acquiring multiple historical voltage information and historical influencing factor information with temporal relationships, each of the historical voltage information corresponding to one of the historical influencing factor information, the temporal relationship including time periods with sequential relationships, and the historical influencing factor information including historical weather information and / or historical equipment status information and / or historical load level information; determining multiple training identification results corresponding to each historical voltage information based on the multiple historical voltage information and the pre-trained disturbance identification model; performing horizontal analysis on the historical influencing factor information to determine multiple historical influencing factor levels, the multiple influencing factor levels including weather levels corresponding to the weather information, equipment status levels corresponding to the equipment status information, and load level information corresponding to the load level information; inputting the historical influencing factor levels and the multiple training identification results into the disturbance prediction model to be trained for multiple training iterations to obtain multiple disturbance prediction results; and obtaining the pre-trained disturbance prediction model based on the disturbance prediction results.

[0041] The historical voltage information and historical influencing factor information mentioned above are included in the training set. Multiple historical voltage information sets are sequential in time, while the historical influencing factor information corresponds to each historical voltage information set and represents the historical influencing factors for each time period. Each historical influencing factor information set includes one or more of historical weather information, historical equipment status information, and historical load level information. The pre-trained disturbance recognition model is not detailed here; the training and recognition results are obtained based on the disturbance recognition model and the historical voltage information. Historical equipment status information can be the fault status or operating status of electrical equipment, and historical load level information can be the increase or decrease in electrical load. Historical influencing factors are categorized into levels. For example, historical weather information is divided into multiple levels based on weather factors that may affect voltage, with higher levels indicating greater potential impact. Similarly, historical equipment status information and historical load level information are also categorized based on their potential impact on voltage. These levels are then further divided to obtain multiple historical influencing factor levels. Then, historical image factor levels and multiple training recognition results are input into the trained disturbance prediction model for iterative training. Each input yields a disturbance prediction result, including the disturbance occurrence time and degree. Based on the disturbance prediction results and the subsequent actual results, model accuracy analysis is performed, specifically analyzing the accuracy of the historical voltage information and the prediction results. Based on this accuracy analysis, the trained disturbance prediction model is finally obtained. The specific methods for the accuracy analysis can include various methods such as mean square error, root mean square error, mean absolute error, and coefficient of determination, which are not limited here.

[0042] For details, please refer to Figure 3 , Figure 3 This is a model architecture diagram of a perturbation recognition model provided in an embodiment of this application. The pre-trained perturbation recognition model can be a RepVGG model. The architecture of the perturbation recognition model can include convolutional layers, pooling layers, fully connected layers, and inputs and outputs. Please refer to... Figure 3The specific process of determining the recognition result based on the pre-trained perturbation recognition model can include: inputting voltage information, specifically the input voltage perturbation waveform; then inputting the voltage information into a convolutional layer, where the convolutional layer performs convolution calculations on the input voltage information. Specifically, the convolutional layer performs convolution calculations on the voltage information based on convolutional kernels, and then stacks the convolutional layers, with each convolutional layer following an activation function. The number of convolutional layers included in the convolutional layer can be multiple, and the specific number is not limited here; then inputting the calculation result obtained from the convolutional layer into a pooling layer for max pooling; finally, reducing the risk of overfitting in a fully connected layer, which can use the Dropout function to reduce the risk of overfitting; and finally outputting the perturbation recognition result, the perturbation time-specific result including the perturbation degree and perturbation time.

[0043] As can be seen, in this embodiment, the possible disturbances in the next period are predicted based on the pre-trained disturbance prediction model and the period before the next period. This allows us to know that a disturbance may occur before it happens in the next stage, making the response to the disturbance in the next period more timely.

[0044] In one possible embodiment, the influencing factor information includes weather information and / or equipment status information and / or load level information; the disturbance prediction result includes predicted disturbance duration, predicted disturbance degree, and a first disturbance factor, wherein the first disturbance factor characterizes the actual influencing factor, the first disturbance factor is determined based on the influencing factor information, the disturbance degree characterizes the voltage change amplitude, and the first disturbance factor includes weather factors and / or equipment status factors and / or load level factors; determining the pre-adjustment strategy based on the disturbance prediction result includes: determining a pre-adjustment time based on the predicted disturbance duration; and determining a pre-adjustment voltage based on the predicted disturbance degree; determining a coordination scheme based on the first disturbance factor, the coordination scheme including a first coordination scheme for the weather factor and / or a second coordination scheme for the equipment status factor and / or a third coordination scheme for the load level factor; and determining the pre-adjustment strategy based on the pre-adjustment time, the pre-adjustment voltage, the first coordination scheme and / or the second coordination scheme for the equipment status factor and / or the third coordination scheme for the load level factor.

[0045] The influencing factor information refers to the information affecting voltage disturbances acquired simultaneously with the second voltage information. This information may include, but is not limited to, weather conditions, equipment malfunctions, and load increases or decreases; specifically, it corresponds to weather information, equipment status information, and load level information, respectively. The first disturbance factor is the actual factor affecting voltage disturbances, obtained based on the disturbance prediction model and the influencing factor information. It may be one or more of the influencing factors corresponding to the aforementioned influencing factor information. The predicted disturbance duration is the predicted duration of the disturbance and its start and end times, and the predicted disturbance degree is the predicted voltage fluctuation degree, which may specifically be the voltage change.

[0046] Specifically, the process begins with analyzing the predicted disturbance duration. Based on this duration, a suitable pre-regulation time window is set to ensure sufficient time for adjustment before the disturbance occurs. The predicted disturbance severity is then assessed, and the required pre-regulation voltage value is calculated to offset or reduce the disturbance's impact on the system. Weather factors are analyzed, and a first coordination plan is developed; equipment status factors are analyzed, and a second coordination plan is developed; load level factors are analyzed, and a third coordination plan is developed. The pre-regulation time, pre-regulation voltage, and the aforementioned coordination plans are then comprehensively considered. A comprehensive pre-regulation strategy is formulated, taking all relevant factors into account to ensure stable system operation during disturbances. The first coordination plan includes implementing weather contingency plans to mitigate environmental impacts; specific plans are not limited here. The second coordination plan includes rapid equipment repair or troubleshooting to prevent further line failures due to equipment malfunctions; specific plans are not limited here. The third coordination plan includes coordinating load or power generation to ensure stable power output on the line; specific plans are not limited here. These specific coordination schemes can better address different weather conditions, equipment status, and load levels, thereby improving the reliability and efficiency of the entire system. Then, the pre-regulation strategy is determined based on the pre-regulation time and pre-regulation voltage, or based on the pre-regulation time, pre-regulation voltage, and any one of the above-mentioned first, second, and third coordination schemes.

[0047] As can be seen, in this embodiment, a pre-adjustment strategy is proposed to determine the possible disturbance time and disturbance degree based on different influencing factors, so as to make the response to voltage disturbance more timely and effective.

[0048] In one possible embodiment, the training process of the pre-trained perturbation recognition model includes the following steps: acquiring multiple training voltage information; inputting the multiple training voltage information into the perturbation recognition model to be trained, the perturbation recognition model to be trained including a multi-branch convolutional layer, a first pooling layer, and a first fully connected layer, the multi-branch convolutional layer including multiple different types of branch convolutional layers; performing multi-branch convolution operations on the multiple different types of branch convolutional layers to obtain multiple first convolution output results; performing residual fusion based on the multiple first convolution output results to obtain second convolution output results; the first pooling layer performing pooling operations based on the first output results to obtain pooling output results; the first fully connected layer integrating based on the pooling results to obtain multiple training recognition results; performing loss analysis based on the training recognition results to obtain loss results; performing repeated training and branch fusion based on the loss results to obtain the pre-trained perturbation recognition model, the pre-trained perturbation recognition model including a reconstructed single-branch convolutional layer, a second pooling layer, and a second fully connected layer.

[0049] Multiple training voltage information points are used as training data for the model. Specifically, the training voltage information can be the time-voltage curve. The architecture of the perturbation recognition model to be trained includes a multi-branch convolutional layer, a first pooling layer, and a first fully connected layer. The difference between the first pooling layer and the first fully connected layer and the second pooling layer and the second fully connected layer included in the pre-trained perturbation recognition model lies in the difference before and after training, which may involve differences in specific parameters. The multi-branch convolutional layer includes convolutional branches with convolutional kernels of different sizes, such as 1×1 branches, 3×3 branches, and identity mapping branches. During the training phase, these branches work independently, each with its own convolutional kernel and parameters. The output of the first convolution includes the convolution results of at least one branch. Then, residual fusion is performed based on the first convolution output to obtain the second convolution output. The second convolution output is then passed through the first pooling layer to reduce the feature size and computational complexity, resulting in the pooled output. The pooled training result is then passed through the fully connected layer for dimensionality reduction and to reduce the risk of overfitting, resulting in the final output training recognition result. Then, based on the training recognition results, repeated iterations of training are performed to obtain a pre-trained perturbation recognition model. The trained perturbation recognition model then undergoes reparameterized fusion of multi-branch convolutional layers to obtain a reconstructed single-branch convolutional layer.

[0050] Specifically, during training, the multi-branch convolutional layer includes 1×1 branches, 3×3 branches, and identity mapping branches. The 1×1 convolution is equivalent to a special 3×3 convolution (with many zeros in the kernel), while the identity mapping is a special 1×1 convolution (with the identity matrix as the kernel). During training, the perturbation recognition model also features a residual network structure, which effectively addresses the vanishing and exploding gradient problems common in deep neural networks. Specifically, the residual network performs residual fusion of features from the multi-branch convolutional layer and the input convolutional layer. These features can be features of the first voltage information, etc. The calculation formula for the residual network is: output = Conv main (x)+Conv residual (x)+x, where Conv main (x) represents the branch other than the identity mapping, Conv residual (x) represents the identity mapping branch, where x is a feature of the input convolutional layer. For example, Conv... main (x) includes the convolution results of 1×1 branches and 3×3 branches, Conv residual (x) includes the convolution result of the identity mapping branch.

[0051] The perturbation recognition model during training includes Batch Normalization (BN) layers within its convolutional layers, which are fused with each other. Each type of convolutional kernel in the convolutional layer has a set of weights. For example, the multi-branch convolutional layer in this scheme has multiple convolutional branches with different kernels, each with different weights. Each convolutional layer also includes an activation function; each convolutional layer is followed by an activation function. Specifically, this scheme uses the ReLU activation function, which can be ReLU(x) = max(0, x), where x is the value of the input activation function. The ReLU activation function compares the input value x with 0; if x is greater than 0, it outputs x; if x is less than or equal to 0, it outputs 0. This allows ReLU to truncate negative values ​​to 0, introducing sparsity and further improving computational efficiency.

[0052] The process of reconstructing a single-branch convolutional layer involves fusing each branch of a multi-branch convolutional layer into a single type of convolutional kernel. In this embodiment, the reconstructed single-branch convolutional layer is a convolution with a 3×3 kernel. The fused reconstructed single-branch convolutional layer also has reconstructed convolutional weights. The process of determining the reconstructed convolutional weights includes: determining the weights of multiple branches of the multi-branch convolutional layer; and fusing the weights based on the multiple branch weights to obtain the reconstructed convolutional weights of the reconstructed single-branch convolutional layer. The fusion process of fusing multiple branches includes: converting the identity mapping branch into a 1x1 convolution, then into a 3×3 convolution, and finally converting the 1×1 convolution into a 3×3 convolution. Specifically, the method of equivalently converting a 1x1 convolution into a 3x3 convolution includes padding with zeros.

[0053] For example, if a multi-branch convolutional layer includes 1×1 branches, 3×3 branches, and identity mapping branches, these branches need to be merged into a single-branch convolutional layer with a 3x3 convolutional kernel, and the reconstructed convolutional weights of the merged 3x3 convolutional kernel branches need to be determined. The process of determining the reconstructed convolutional weights of the merged 3x3 convolutional kernel branches includes: first, determining the branch weights of each branch, specifically: first, fusing the original weights of the 3×3 branches with the BN layer to obtain the first weight, which is W' = γW + βB, where W is the original weight of the 3×3 branch, γ is the scaling factor of the BN layer, β is the offset of the BN layer, and B is the original bias term. Then, the second weight W of the 1×1 branch is obtained. 1×1 The third weight W of the identity mapping branch identity The reconstructed convolution weights are obtained by fusing the first, second, and third weights: W” = W’ + W 1×1+ W identity .

[0054] For example, a voltage waveform with time (seconds) on the horizontal axis and voltage (volts) on the vertical axis is obtained and fed as input to the convolutional layer of the perturbation detection model for analysis. The convolutional layer uses a fused 3x3 convolutional kernel with a stride of 1 and padding of 1 to extract features from the input waveform. A nonlinearity is introduced through the ReLU activation function to obtain the output feature map of the convolutional layer. Next, the output feature map is fed into a max-pooling layer with a pooling window size of 2x2 and a stride of 2, resulting in a feature map F2 with its size halved. Then, feature map F2 is flattened into a one-dimensional vector and fed into a fully connected layer. Dropout is used to prevent overfitting, and the final output is the voltage sag perturbation detection result R, indicating whether a voltage perturbation exists.

[0055] The process of obtaining loss results by loss analysis based on training recognition results can be based on comparing the training recognition results obtained in each round of training with the actual situation to obtain the training and actual loss values. The loss results are then subjected to trend analysis to determine if the model's loss value is stable and within a preset threshold, indicating that the model training is complete. The actual situation mentioned above refers to the actual disturbance occurrence.

[0056] Specifically, by monitoring the loss curve, one can intuitively understand the model's learning progress during training. If the loss value gradually decreases, it indicates that the model is gradually learning the features of the data, and the fitting effect is improving. The loss curve can help determine whether the model has overfitting or underfitting issues. Overfitting is characterized by a continuous decrease in training loss, but an increase in the loss on the validation set (or test set), meaning the model performs well on the training data but poorly on unseen data. Underfitting is characterized by high training and validation losses, indicating that the model has not learned the features of the data well. Furthermore, analyzing the loss curve can guide the adjustment of hyperparameters, such as learning rate, batch size, and regularization parameters, to optimize the model's training process.

[0057] As can be seen, in this embodiment, a disturbance identification model is proposed, which can accurately and quickly identify voltage disturbances. It captures the subtle features of voltage sags through a deep learning model, achieving high-precision identification.

[0058] S220, Obtain the first voltage information of the second time period, the second time period including the real-time time period, which is the connection time period of the first time period.

[0059] The second time period is the current real-time time period, which is the time period to which the first time period is sequentially transferred. The first voltage information is the voltage information obtained in the real-time time period.

[0060] S230, a first identification result is determined based on the first voltage information and the pre-trained disturbance identification model, wherein the first identification result characterizes the disturbance occurrence identified in the second time period.

[0061] The training process of the pre-trained disturbance identification model has been described in detail in step S210 and will not be repeated here. The disturbance identification model uses the same method for the disturbance time of the first voltage information and the second voltage information, which will not be repeated here either. Through the disturbance identification model, the occurrence of disturbances in the second time period can be identified. Specifically, the occurrence of disturbances includes at least one of disturbance occurrence time and disturbance occurrence degree.

[0062] S240, Based on the first identification result and the pre-adjustment strategy, a disturbance response method is determined, wherein the disturbance response method is used to adjust the pre-adjustment strategy to keep the voltage stable.

[0063] The pre-adjustment strategy is described in detail in step S210 above and will not be repeated here. The disturbance response method is a modified method based on the pre-adjustment strategy, which is implemented during the current real-time period, i.e., the second time period.

[0064] In one possible embodiment, the disturbance response method includes a first response method and a second response method. Determining the disturbance response method based on the first identification result and the pre-adjustment strategy includes: determining the presence of a disturbance based on the first identification result, where the presence of a disturbance includes the presence of a disturbance and the absence of a disturbance; if the presence of a disturbance is that a disturbance exists, then determining the disturbance response method as the first disturbance response method, where the first disturbance response method includes modifying the pre-adjustment strategy based on the first identification result to obtain a real-time adjustment strategy; if the presence of a disturbance is that a disturbance does not exist, then determining the disturbance response method as the second disturbance response method, where the second disturbance response method includes not modifying the pre-adjustment strategy.

[0065] This involves proactively adjusting parameters or operational strategies to prevent or mitigate the impact of predicted disturbances. Real-time monitoring of status and performance indicators determines the effectiveness of the pre-adjustment strategy and identifies any unforeseen disturbances. If a disturbance is identified, the pre-adjustment strategy is dynamically adjusted based on real-time data and the disturbance's characteristics, forming a real-time adjustment strategy. The revised real-time adjustment strategy is then executed to eliminate or mitigate the current disturbance.

[0066] The presence or absence of a disturbance can be determined by a combination of the degree and duration of the disturbance, or by one of these factors. Specifically, a disturbance degree level and a disturbance duration level can be preset. Based on the presence and duration of the disturbance, a level can be defined, and a preset threshold can be used to determine whether a disturbance exists or not.

[0067] As can be seen, in this embodiment, by combining real-time conditions and pre-adjustment strategies to obtain a more in-depth and detailed response strategy, voltage disturbances can be effectively addressed, voltage stability can be ensured, and the speed of response to disturbances occurring in the current period can be improved.

[0068] As can be seen, in this embodiment, a pre-adjustment strategy for a first time period is obtained. The first time period is the preceding time period of the real-time period, and the pre-adjustment strategy is a strategy to perform pre-adjustment operations on the preceding time period during the first time period to reduce the degree of voltage disturbance in the real-time period. First voltage information for a second time period is obtained. The second time period includes the real-time period and is the preceding time period of the first time period. A first identification result is determined based on the first voltage information and a pre-trained disturbance identification model. The first identification result characterizes the disturbance occurrence identified in the second time period. A disturbance response method is determined based on the first identification result and the pre-adjustment strategy. The disturbance response method is used to adjust the pre-adjustment strategy to keep the voltage stable. Thus, by combining prediction and real-time monitoring and using a disturbance identification model, detection efficiency and accuracy can be improved, and response speed can be accelerated.

[0069] Please see Figure 4 , Figure 4 This is a schematic diagram of a voltage disturbance detection device provided in an embodiment of this application. The device is applied to electronic equipment. The voltage disturbance detection device 400 includes: a first acquisition module 410, a second acquisition module 420, a first determination module 430, and a second determination module 440, wherein...

[0070] The first acquisition module 410 is used to acquire the pre-adjustment strategy for a first time period, where the first time period is the preceding time period of the real-time time period, and the pre-adjustment strategy is a strategy for performing pre-adjustment operations on the preceding time period during the first time period to reduce the degree of voltage disturbance in the real-time time period.

[0071] The second acquisition module 420 is used to acquire the first voltage information of the second time period, the second time period including the real-time time period, which is the connection time period of the first time period;

[0072] The first determining module 430 is used to determine a first identification result based on the first voltage information and a pre-trained disturbance identification model, wherein the first identification result characterizes the disturbance occurrence identified in the second time period;

[0073] The second determining module 440 is used to determine a disturbance response method based on the first identification result and the pre-adjustment strategy. The disturbance response method is used to adjust the pre-adjustment strategy so that the voltage remains stable.

[0074] In one possible embodiment, before the step of obtaining the pre-adjustment strategy for the first time period, the method further includes: obtaining a disturbance prediction result for the first time period, the disturbance prediction result representing disturbance occurrence information of the predicted successive time period during the first time period; determining the pre-adjustment strategy based on the disturbance prediction result; and performing a pre-adjustment operation based on the pre-adjustment strategy.

[0075] In one possible embodiment, obtaining the disturbance prediction result for the first time period includes: obtaining second voltage information and influencing factor information for the first time period, wherein the influencing factor information is a possible factor affecting the voltage; determining a second identification result based on the second voltage information and the pre-trained disturbance identification model, wherein the second identification result characterizes the disturbance occurrence identified in the first time period; and determining the disturbance prediction result based on the second identification result, the influencing factor information, and the pre-trained disturbance prediction model.

[0076] In one possible embodiment, the training method of the pre-trained disturbance prediction model includes the following steps: acquiring multiple historical voltage information and historical influencing factor information with temporal relationships, each of the historical voltage information corresponding to one of the historical influencing factor information, the temporal relationship including time periods with sequential relationships, and the historical influencing factor information including historical weather information and / or historical equipment status information and / or historical load level information; determining multiple training identification results corresponding to each historical voltage information based on the multiple historical voltage information and the pre-trained disturbance identification model; performing horizontal analysis on the historical influencing factor information to determine multiple historical influencing factor levels, the multiple influencing factor levels including weather levels corresponding to the weather information, equipment status levels corresponding to the equipment status information, and load level information corresponding to the load level information; inputting the historical influencing factor levels and the multiple training identification results into the disturbance prediction model to be trained for multiple training iterations to obtain multiple disturbance prediction results; and obtaining the pre-trained disturbance prediction model based on the disturbance prediction results.

[0077] In one possible embodiment, the influencing factor information includes weather information and / or equipment status information and / or load level information; the disturbance prediction result includes predicted disturbance duration, predicted disturbance degree, and a first disturbance factor, wherein the first disturbance factor characterizes the actual influencing factor, the first disturbance factor is determined based on the influencing factor information, the disturbance degree characterizes the voltage change amplitude, and the first disturbance factor includes weather factors and / or equipment status factors and / or load level factors; determining the pre-adjustment strategy based on the disturbance prediction result includes: determining a pre-adjustment time based on the predicted disturbance duration; and determining a pre-adjustment voltage based on the predicted disturbance degree; determining a coordination scheme based on the first disturbance factor, the coordination scheme including a first coordination scheme for the weather factor and / or a second coordination scheme for the equipment status factor and / or a third coordination scheme for the load level factor; and determining the pre-adjustment strategy based on the pre-adjustment time, the pre-adjustment voltage, the first coordination scheme and / or the second coordination scheme for the equipment status factor and / or the third coordination scheme for the load level factor.

[0078] In one possible embodiment, the disturbance response method includes a first response method and a second response method. Specifically, the second determining module 440, in determining the disturbance response method based on the first identification result and the pre-adjustment strategy, is configured to determine the presence of a disturbance based on the first identification result, where the presence of a disturbance includes the presence of a disturbance and the absence of a disturbance. If the presence of a disturbance is the presence of a disturbance, the disturbance response method is determined to be the first disturbance response method, which includes modifying the pre-adjustment strategy based on the first identification result to obtain a real-time adjustment strategy. If the presence of a disturbance is the absence of a disturbance, the disturbance response method is determined to be the second disturbance response method, which includes not modifying the pre-adjustment strategy.

[0079] In one possible embodiment, the training process of the pre-trained perturbation recognition model includes the following steps:

[0080] Multiple training voltage information is acquired; the multiple training voltage information is input into a perturbation recognition model to be trained, the perturbation recognition model to be trained includes a multi-branch convolutional layer, a first pooling layer, and a first fully connected layer, the multi-branch convolutional layer includes multiple different types of branch convolutional layers; multiple first convolutional output results are obtained by performing multi-branch convolution operations on the multiple different types of branch convolutional layers; residual fusion is performed on the multiple first convolutional output results to obtain a second convolutional output result; the first pooling layer performs pooling operations on the first output results to obtain a pooling output result; the first fully connected layer integrates the pooling results to obtain multiple training recognition results; loss analysis is performed on the training recognition results to obtain a loss result; the model is repeatedly trained and branch fusion is performed on the loss result to obtain the pre-trained perturbation recognition model, the pre-trained perturbation recognition model includes a reconstructed single-branch convolutional layer, a second pooling layer, and a second fully connected layer.

[0081] It is worth noting that the specific functional implementation of the voltage disturbance detection device 400 is described above. Figure 2 The description of the voltage disturbance detection method illustrates, for example, the first acquisition module 410 used to implement the relevant content of S210. Each unit or module in the voltage disturbance detection device 400 can be individually or entirely merged into one or more other units or modules, or some of the units or modules can be further divided into multiple functionally smaller units or modules. This achieves the same operation without affecting the technical effects of the embodiments of the present invention. The aforementioned units or modules are based on logical function division. In practical applications, the function of one unit (or module) is implemented by multiple units (or modules), or the function of multiple units (or modules) is implemented by one unit (or module).

[0082] As can be seen, the voltage disturbance detection device described in this application first acquires a pre-adjustment strategy for a first time period, which is the preceding time period of the real-time period. The pre-adjustment strategy is a strategy to perform pre-adjustment operations on the preceding time period during the first time period to reduce the degree of voltage disturbance in the real-time period. It then acquires first voltage information for a second time period, which includes the real-time period and is the preceding time period of the first time period. Based on the first voltage information and a pre-trained disturbance recognition model, it determines a first recognition result, which characterizes the disturbance occurrence detected in the second time period. Based on the first recognition result and the pre-adjustment strategy, it determines a disturbance response method, which adjusts the pre-adjustment strategy to maintain voltage stability. Thus, by combining prediction and real-time monitoring and utilizing a disturbance recognition model, detection efficiency and accuracy can be improved, and response speed can be accelerated.

[0083] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. As shown in the figure, the electronic device 500 includes a processor 510, a memory 520, a communication interface 530, and one or more programs 521. The one or more programs 521 are stored in the memory 520 and configured to be executed by the processor 510. The electronic device 500 corresponds to... Figure 1 Electronic device 110 in the embodiment.

[0084] The processor 510, memory 520, and communication interface 530 are interconnected and perform communication between them.

[0085] The memory 520 can be a volatile memory such as dynamic random access memory (DRAM) or a non-volatile memory such as a hard disk drive (HDD). The memory 520 is used to store a set of executable program code, and the processor 510 is used to call one or more programs 521 stored in the memory 520, which can execute some or all of the steps of any of the methods described in the above method embodiments.

[0086] Among them, electronic devices 500 may include servers, laptops, mobile computers, computers, etc. The above are just examples and not an exhaustive list, including but not limited to the above electronic devices.

[0087] As can be seen, the electronic device 500 described in this application embodiment acquires a pre-adjustment strategy for a first time period, where the first time period is the preceding time period of the real-time period. The pre-adjustment strategy is a strategy for performing pre-adjustment operations on the preceding time period during the first time period to reduce the degree of voltage disturbance during the real-time period. It then acquires first voltage information for a second time period, where the second time period includes the real-time period and is the preceding time period of the first time period. Based on the first voltage information and a pre-trained disturbance recognition model, it determines a first recognition result, which characterizes the disturbance occurrence detected in the second time period. Based on the first recognition result and the pre-adjustment strategy, it determines a disturbance response method, which adjusts the pre-adjustment strategy to maintain voltage stability. Thus, by combining prediction and real-time monitoring and utilizing a disturbance recognition model, detection efficiency and accuracy can be improved, and response speed can be accelerated.

[0088] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0089] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include an electronic device.

[0090] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0091] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0092] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0093] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0094] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0095] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer electronic device (which may be a personal computer, electronic device, or network electronic device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0096] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0097] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A voltage disturbance detection method, characterized in that, Applied to electronic devices, the method includes: Obtain the pre-adjustment strategy for the first time period, where the first time period is the preceding time period of the real-time time period. The pre-adjustment strategy is a strategy for performing pre-adjustment operations on the preceding time period during the first time period, which is used to reduce the degree of voltage disturbance in the real-time time period. Obtain the first voltage information for the second time period, the second time period including the real-time time period, which is the connection time period of the first time period; A first identification result is determined based on the first voltage information and the pre-trained disturbance identification model. The first identification result represents the occurrence of disturbances identified in the second time period. Based on the first identification result and the pre-adjustment strategy, a disturbance response method is determined. The disturbance response method is used to adjust the pre-adjustment strategy so that the voltage remains stable.

2. The voltage disturbance detection method according to claim 1, characterized in that, Prior to the step of obtaining the pre-adjustment strategy for the first time period, the method further includes: Obtain the disturbance prediction result for the first time period, wherein the disturbance prediction result represents the disturbance occurrence information of the subsequent time period predicted during the first time period; The pre-adjustment strategy is determined based on the disturbance prediction results, and the pre-adjustment operation is performed based on the pre-adjustment strategy.

3. The voltage disturbance detection method according to claim 2, characterized in that, The process of obtaining the disturbance prediction result for the first time period includes: Obtain the second voltage information and influencing factor information for the first time period, wherein the influencing factor information refers to possible factors affecting the voltage; A second identification result is determined based on the second voltage information and the pre-trained disturbance identification model, wherein the second identification result characterizes the disturbance occurrence identified in the first time period; The disturbance prediction result is determined based on the second identification result, the influencing factor information, and the pre-trained disturbance prediction model.

4. The voltage disturbance detection method according to claim 3, characterized in that, The training method for the pre-trained perturbation prediction model includes the following steps: Acquire multiple historical voltage information and historical influencing factor information with a time sequence relationship. Each historical voltage information corresponds to one historical influencing factor information. The time sequence relationship includes time periods with a sequential relationship. The historical influencing factor information includes historical weather information and / or historical equipment status information and / or historical load level information. Based on the multiple historical voltage information and the pre-trained disturbance recognition model, determine multiple training recognition results corresponding to each historical voltage information; A horizontal analysis is performed on the historical influencing factor information to determine multiple historical influencing factor levels, including weather levels corresponding to the weather information, equipment status levels corresponding to the equipment status information, and load level information corresponding to the load level information. Based on the historical influencing factor levels and the multiple training identification results, the disturbance prediction model to be trained is trained multiple times to obtain multiple disturbance prediction results. The pre-trained perturbation prediction model is obtained based on the perturbation prediction results.

5. The voltage disturbance detection method according to claim 3, characterized in that, The influencing factor information includes weather information and / or equipment status information and / or load level information. The disturbance prediction result includes the predicted disturbance duration, the predicted disturbance degree, and a first disturbance factor. The first disturbance factor characterizes the actual influencing factor. The first disturbance factor is determined based on the influencing factor information. The disturbance degree characterizes the voltage change amplitude. The first disturbance factor includes weather factors and / or equipment status factors and / or load level factors. The step of determining the pre-adjustment strategy based on the perturbation prediction result includes: The pre-adjustment time is determined based on the predicted disturbance duration; and the pre-adjustment voltage is determined based on the predicted disturbance degree. A coordination scheme is determined based on the first disturbance factor, the coordination scheme including a first coordination scheme for the weather factor and / or a second coordination scheme for the equipment status factor and / or a third coordination scheme for the load level factor; The pre-adjustment strategy is determined based on the pre-adjustment time, the pre-adjustment voltage, the first coordination scheme, and / or the second coordination scheme for the equipment status factors and / or the third coordination scheme for the load level factors.

6. The voltage disturbance detection method according to claim 5, characterized in that, The disturbance response method includes a first disturbance response method and a second disturbance response method, wherein determining the disturbance response method based on the first identification result and the pre-adjustment strategy includes: The presence of a disturbance is determined based on the first identification result, wherein the presence of a disturbance includes the presence of a disturbance and the absence of a disturbance; If the disturbance exists, then the disturbance response method is determined to be the first disturbance response method, which includes modifying the pre-adjustment strategy based on the first identification result to obtain a real-time adjustment strategy. If the disturbance exists instead of the disturbance does not exist, then the disturbance response method is determined to be the second disturbance response method, which includes not modifying the pre-adjustment strategy.

7. The voltage disturbance detection method according to claim 1 or 4, characterized in that, The training process of the pre-trained perturbation recognition model includes the following steps: Acquire multiple training voltage information; The multiple training voltage information is input into the perturbation recognition model to be trained. The perturbation recognition model to be trained includes a multi-branch convolutional layer, a first pooling layer, and a first fully connected layer. The multi-branch convolutional layer includes multiple branch convolutional layers of different types. Multiple first convolution outputs are obtained by performing multi-branch convolution operations on the multiple different types of branch convolutional layers; Residual fusion is performed based on the multiple first convolution outputs to obtain the second convolution output; The first pooling layer performs a pooling operation based on the first convolution output to obtain the pooling output result; The first fully connected layer integrates the pooling output results to obtain multiple training recognition results; Loss results are obtained by performing loss analysis based on the training and recognition results. Based on the loss result, the model is repeatedly trained and branched to obtain the pre-trained perturbation recognition model, which includes a reconstructed single-branch convolutional layer, a second pooling layer, and a second fully connected layer.

8. A voltage disturbance detection device, characterized in that, Applied to electronic devices, the device includes: The first acquisition module is used to acquire the pre-adjustment strategy for a first time period, where the first time period is the preceding time period of the real-time time period, and the pre-adjustment strategy is a strategy for performing pre-adjustment operations on the preceding time period during the first time period, which is used to reduce the degree of voltage disturbance in the real-time time period. The second acquisition module is used to acquire the first voltage information of the second time period, the second time period including the real-time time period, which is the connection time period of the first time period; The first determining module is used to determine a first identification result based on the first voltage information and a pre-trained disturbance identification model, wherein the first identification result characterizes the occurrence of disturbances identified in the second time period; The second determining module is used to determine a disturbance response method based on the first identification result and the pre-adjustment strategy. The disturbance response method is used to adjust the pre-adjustment strategy so that the voltage remains stable.

9. A computer-readable storage medium, characterized in that, The device stores a voltage disturbance detection program, including execution instructions, which, when executed by a processor of an electronic device, perform the voltage disturbance detection method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The method includes a processor and a memory storing execution instructions, the memory storing one or more programs; when the processor executes the execution instructions stored in the memory, the processor performs the voltage disturbance detection method as described in any one of claims 1 to 7.

Images