Low-voltage prediction model training method and device, storage medium and electronic equipment

Abstract

The invention provides a low-voltage prediction model training method and device, a storage medium and electronic equipment. The electronic equipment obtains a to-be-analyzed data set; wherein the to-be-analyzed data set comprises multiple groups of original records before low voltage occurs in the power distribution network, and each group of original records comprises observation values of multiple impact factors; according to the to-be-analyzed data set, screening out a plurality of significant influence factors having significant influence on the low voltage from the plurality of influence factors; constructing a training data set according to the various significant impact factors; wherein the training data set comprises a plurality of training samples, and each training sample comprises observation values of a plurality of significant influence factors and voltage state labels corresponding to the prediction time points; and training the to-be-trained model through the training data set to obtain a low-voltage prediction model. In this way, interference of redundant information on model training can be effectively reduced by extracting the significant influence factors; and the prediction performance and the stability of the trained model can be obviously improved.

Description

Technical Field

[0001] The present application relates to the field of power grids, and specifically to a low voltage prediction model training method, device, storage medium and electronic equipment. Background Art

[0002] As electricity demand continues to grow, the pace of distribution network transformation is significantly out of sync with the growth in electricity consumption. This has directly led to frequent low-voltage problems in distribution networks. To effectively address this challenge, it is particularly urgent and necessary to conduct research on the prediction of low-voltage trends in distribution networks.

[0003] However, training an efficient and accurate prediction model requires high-quality and appropriate data. Distribution network data is inherently complex and susceptible to a variety of internal and external factors, such as meteorological conditions, equipment operating status, and user electricity usage. Therefore, when building a prediction model, clearly selecting the right data types as input features becomes a key factor in determining the model's effectiveness. Summary of the Invention

[0004] In order to overcome at least one deficiency in the prior art, the present application provides a low voltage prediction model training method, device, storage medium, and electronic device, specifically including: In a first aspect, the present application provides a low voltage prediction model training method, the method comprising: Acquire a data set to be analyzed, wherein the data set to be analyzed includes multiple groups of original records before low voltage occurs in the distribution network, and each group of original records includes observation values ​​of multiple influencing factors; Screening out, from the plurality of influencing factors, a plurality of significant influencing factors having a significant impact on the low voltage according to the data set to be analyzed; Constructing a training data set based on the multiple significant influencing factors, wherein the training data set includes multiple training samples, each of the training samples includes observed values ​​of the multiple significant influencing factors and a voltage state label corresponding to a prediction time point; The model to be trained is trained using the training data set to obtain a low voltage prediction model.

[0005] In a second aspect, the present application provides a low voltage prediction model training device, the device comprising: A data preparation module is used to obtain a data set to be analyzed, wherein the data set to be analyzed includes multiple groups of original records before low voltage occurs in the distribution network, and each group of original records includes observation values ​​of multiple influencing factors; The data preparation module is further configured to screen out, from the plurality of influencing factors, a plurality of significant influencing factors having a significant impact on the low voltage based on the data set to be analyzed; The data preparation module is further configured to construct a training data set based on the multiple significant influencing factors, wherein the training data set includes multiple training samples, each of which includes observed values ​​of the multiple significant influencing factors and a voltage state label corresponding to a prediction time point; The model training module is used to train the model to be trained using the training data set to obtain a low voltage prediction model.

[0006] In a third aspect, the present application provides a storage medium storing a computer program, which implements the low voltage prediction model training method when executed by a processor.

[0007] In a fourth aspect, the present application provides an electronic device, comprising a processor and a memory, wherein the memory stores a computer program, and the computer program implements the low voltage prediction model training method when executed by the processor.

[0008] Compared with the prior art, this application has the following beneficial effects: The present application provides a low voltage prediction model training method, device, storage medium and electronic device. The electronic device obtains a data set to be analyzed; wherein, the data set to be analyzed includes multiple groups of original records before the low voltage occurs in the distribution network, and each group of original records includes observations of multiple influencing factors; according to the data set to be analyzed, multiple significant influencing factors that have a significant impact on low voltage are screened out from the multiple influencing factors; based on the multiple significant influencing factors, a training data set is constructed; wherein, the training data set includes multiple training samples, and each training sample includes observations of multiple significant influencing factors and voltage state labels corresponding to the prediction time point; the training model is trained with the training data set to obtain a low voltage prediction model. In this way, by obtaining a data set to be analyzed containing observations of multiple influencing factors, and extracting significant influencing factors from it based on statistical analysis or feature selection algorithms, the interference of redundant information on model training can be effectively reduced; therefore, when the training model is trained based on the optimized training data set, the model can more accurately capture the core laws of low voltage occurrence, thereby significantly improving prediction performance and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of a low voltage prediction model training method provided in an embodiment of the present application; Figure 2 This is a schematic diagram of the implementation details of the low voltage prediction model training method provided in an embodiment of the present application; Figure 3 A second schematic diagram of implementation details of the low voltage prediction model training method provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of a low voltage prediction model training device provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present application (hereinafter referred to as the embodiments) more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of them. Generally, the components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0012] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in the present application without creative work are within the scope of protection of the present application.

[0013] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0014] In the description of this application, it should be noted that the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be understood as indicating or implying relative importance. In addition, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0015] Based on the above statement, as introduced in the background technology, distribution network data itself is highly complex and easily affected by various internal and external factors. Therefore, in the process of building a prediction model, clearly selecting which types of data to use as input features becomes one of the key links in determining the model's effectiveness.

[0016] For example, these external factors may include the low voltage conditions of the day, voltage, line radius, number of low voltage occurrences on the day, duration of low voltage, load rate, whether it is an urban area or a rural area, three-phase voltage value, whether it is a holiday, whether the equipment has a fault, equipment operating time, daily maximum temperature, minimum temperature, average temperature in the substation, whether there is lightning, whether there is ice cover, rainfall, wind speed, etc.

[0017] The influencing factors listed above cover multiple dimensions such as power system operation, equipment status, and external environment. Some of these factors are explained below.

[0018] The low voltage condition of the day is often characterized by inertia. If a substation experiences low voltage on a given day, the probability of it occurring again the next day increases significantly. Therefore, this factor may become an important reference for predicting future low voltage events.

[0019] The load factor directly measures the ratio between current power demand and supply capacity. A high load factor typically means greater current flow, which increases the resistive voltage drop in the line and can lead to low voltage issues. Therefore, this factor can be an important reference for predicting future low voltage events.

[0020] Geographical attributes are also crucial for low voltage prediction. Electricity usage patterns differ significantly between urban and rural areas. For example, rural areas may rely more heavily on agricultural loads, while urban areas are primarily focused on residential and commercial electricity consumption. This difference directly influences the shape of the load curve and its corresponding voltage behavior. Therefore, this factor can be a crucial reference for predicting future low voltage events.

[0021] Meteorological factors (daily maximum and minimum temperatures, average temperatures, lightning, ice cover, rainfall, and wind speed) can also directly or indirectly affect voltage levels. For example, user electricity usage behavior can change significantly under low and high temperatures, leading to changes in power demand. Therefore, these factors can be an important reference for predicting future low voltage levels.

[0022] Based on the discovery of the above technical problems, the following technical solutions are proposed after creative work to solve or improve the above problems. It should be noted that the defects existing in the solutions in the above prior art are the results obtained after practice and careful study. Therefore, the discovery process of the above problems and the solutions proposed in the embodiments of this application for the above problems below should be regarded as contributions to this application in the process of invention and creation, and should not be understood as technical contents known to those skilled in the art.

[0023] In view of the above technical problems, this embodiment provides a low voltage prediction model training method. Figure 1 , the method comprising: S1, obtain the data set to be analyzed.

[0024] The data set to be analyzed includes multiple sets of original records before low voltage occurs in the distribution network, and each set of original records includes observation values ​​of multiple influencing factors.

[0025] S2. According to the data set to be analyzed, multiple significant influencing factors having a significant impact on low voltage are screened out from multiple influencing factors.

[0026] S3, construct a training data set based on multiple significant influencing factors.

[0027] The training data set includes multiple training samples, and each training sample includes the observation values ​​of multiple significant influencing factors and the voltage state label corresponding to the prediction time point.

[0028] S4, training the to-be-trained model using the training data set to obtain a low voltage prediction model.

[0029] In this way, by obtaining a data set to be analyzed containing observations of multiple influencing factors and extracting significant influencing factors from it based on statistical analysis or feature selection algorithms, the interference of redundant information on model training can be effectively reduced; therefore, when the model to be trained is trained based on the optimized training data set, the model can more accurately capture the core laws of low voltage occurrence, thereby significantly improving the prediction performance and stability.

[0030] In this embodiment, the electronic device for implementing the low voltage prediction model training method may be, but is not limited to, a mobile terminal, a tablet computer, a laptop computer, a desktop computer, a server, and the like. As long as sufficient computing power can be provided for model training. When it is a server, the server may be a single server or a server group. The server group may be centralized or distributed (for example, the server may be a distributed system). In some embodiments, the server may be local or remote relative to the user terminal. In some embodiments, the server may be implemented on a cloud platform; as an example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud (CommunityCloud), a distributed cloud, an inter-cloud (Inter-Cloud), a multi-cloud (Multi-Cloud), etc., or any combination thereof. In some embodiments, the server may be implemented on an electronic device having one or more components.

[0031] In order to make the solution provided by this embodiment clearer, the server is used as an electronic device to implement the low voltage prediction model training method. Figure 1 Each step of the method shown is described in detail. However, it should be understood that the operations of the flowchart can be implemented in any order, and steps that have no logical contextual relationship can be reversed or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart or remove one or more operations from the flowchart under the guidance of the content of this application. Figure 1 , the method comprising: S1, obtain the data set to be analyzed.

[0032] The data set to be analyzed includes multiple sets of original records before low voltage occurs in the distribution network, and each set of original records includes observation values ​​of multiple influencing factors.

[0033] Specifically, these data come from the historical operation data and meteorological data of multiple substations. Among them, historical operation data includes basic information such as line type, line length, equipment operation time, and equipment failure information. In addition, historical operation data also includes historical load and voltage data, for example, daily load rate, substation load, heavy overload data, voltage data, and low voltage conditions from September 2023 to August 2024. Substation meteorological data This includes meteorological data for each substation every day of the year, such as temperature, rainfall, wind speed, icing, lightning and other weather conditions, as well as information on whether it is a holiday.

[0034] To ensure the validity of the data, preliminary data processing is required after obtaining the data. Specifically, the server screens the substations with missing data and eliminates those that do not meet the conditions. For example, if certain substations lack key information, such as line information, equipment operation information, or voltage data, and this missing data cannot be filled through relevant relationships, these substations will be excluded. This ensures that the data used in subsequent analysis and modeling is of high quality and integrity.

[0035] For data that can be completed by calculation or other methods, appropriate measures should be taken to fill in the gaps. For example, data such as the number of low voltage occurrences and duration in the substation every day can be calculated using the daily three-phase voltage data. The expression for calculating voltage imbalance is:

[0036] Where, Indicates the voltage imbalance. Indicates the maximum negative sequence voltage, Expressed as average voltage.

[0037] The server compares the voltage imbalance with the lowest value of the three-phase voltage and the standard voltage value. When the lowest value is lower than 80% of the standard voltage and the voltage imbalance exceeds 40%, it is considered that low voltage has occurred.

[0038] Based on the description of the data set to be analyzed in the above embodiment, the following Figure 1 Step S2 in the following is explained: S2. According to the data set to be analyzed, multiple significant influencing factors having a significant impact on low voltage are screened out from multiple influencing factors.

[0039] In order to screen a variety of influencing factors, this embodiment calculates the impact score of each influencing factor on low voltage. As an optional implementation, Figure 2 As shown, step S2 may include: S2-1, based on the data set to be analyzed, obtain the impact scores of multiple influencing factors on low voltage.

[0040] There are many ways to evaluate the impact score of each influencing factor on low voltage. In this embodiment, the server maliciously uses the PCA algorithm to analyze the data set to be analyzed to obtain the impact scores of various influencing factors on low voltage.

[0041] S2-2, based on the impact scores of multiple influencing factors on low voltage, select multiple significant influencing factors from the multiple influencing factors.

[0042] Among them, multiple significant impact factors are the top N impact factors with the largest impact scores.

[0043] It should be understood that in the process of building a low voltage prediction model, screening significant influencing factors is of great significance for subsequent model training. That is, it is necessary to screen out a few key factors that have the most significant impact on the low voltage phenomenon from a large number of possible influencing factors.

[0044] For example, the data set to be analyzed includes various information accumulated during the operation of the distribution network, such as the low voltage conditions of the day, voltage value, line radius, number of low voltage occurrences on the day, low voltage duration, load rate, urban or rural area attributes, three-phase voltage value, holiday status, equipment failure condition, equipment operation time, and meteorological conditions (such as maximum temperature, minimum temperature, average temperature, lightning, icing, rainfall and wind speed).

[0045] To quantify the importance of each influencing factor, this embodiment introduces principal component analysis (PCA). Through PCA analysis, the server can assess the specific contribution of each influencing factor to the low voltage condition and select the top N factors with the highest impact scores as significant influencing factors. After this analysis, the 10 factors with the greatest impact on low voltage conditions were ultimately selected, including the current day's low voltage condition, voltage value, line condition, urban or rural area attributes, holiday status, equipment failure, equipment operating hours, daily maximum temperature, daily rainfall, and ice cover.

[0046] S3, construct a training data set based on multiple significant influencing factors.

[0047] The training data set includes multiple training samples, and each training sample includes the observation values ​​of multiple significant influencing factors and the voltage state label corresponding to the prediction time point.

[0048] Further research has revealed that traditional non-time-series models often fail to capture the dynamic changes in influencing factors when constructing low-voltage prediction models. However, in actual power system operation, many significant influencing factors (such as temperature, equipment load, and meteorological conditions) are not static but change over time. These dynamic changes can be key indicators of low voltage occurrence.

[0049] Specifically, low voltage in distribution networks is often the result of a combination of factors. For example, a sudden rise in temperature at a given moment can cause a surge in electricity demand, leading to a voltage drop in a local area. Alternatively, if a device fails after prolonged high-load operation, it can quickly lead to low voltage in the region. This dynamic information is crucial for accurately predicting low voltage, but traditional non-sequential models lack attention to the time dimension.

[0050] To address this issue, each training sample in this embodiment also includes information on changes in some significant influencing factors from the time point of collection of its own observation value to the time point of prediction.

[0051] It should be understood that when constructing the training dataset, for some significantly influencing factors, their actual observed values ​​at the prediction time point and the observed values ​​at the sampling time point are clearly known. Therefore, by comparing the actual observed values ​​at the prediction time point with the observed values ​​at the previous sampling time point, the change information of the significantly influencing factor can be calculated.

[0052] However, in the actual reasoning process, for the same significant influencing factor, considering that it is impossible to know in advance the actual observation value at a certain point in the future, but in order to still be able to obtain the change information of the significant influencing factor, the time series-based prediction model can be used to obtain the predicted value of the significant influencing factor at a certain point in the future.

[0053] For example, suppose it is necessary to predict whether a certain substation will experience low voltage tomorrow, and the temperature, load factor, and wind speed are selected as significant influencing factors. During the training phase, historical data can be used to show that the temperature at 8 o'clock this morning is 25°C, the load factor is 70%, and the wind speed is 5m / s. At the same time, historical data can also be used to show that the actual temperature at 4 o'clock tomorrow afternoon is 30°C, the load factor is 80%, and the wind speed is 6m / s. By comparing the data at these two time points, we can calculate the temperature change information as "+5°C", the load factor change information as "+10%", and the wind speed change information as "+1m / s". These change information is carried in the training samples to train the low voltage prediction model.

[0054] During the actual inference phase, assume that the temperature at 8:00 AM today is 26°C, the load factor is 80%, and the wind speed is 7 m / s. Since the temperature, load factor, and wind speed at 4:00 PM tomorrow are unknown in advance, a time series prediction model (e.g., LSTM) can be used to predict the temperature at 4:00 PM tomorrow to be 31°C, the load factor to be 82%, and the wind speed to be 7 m / s. These predicted values ​​are then compared with the actual observed values ​​at 8:00 AM today to calculate the temperature change as "+5°C," the load factor change as "+2%," and the wind speed change as "+0 m / s." This change information, along with the observed values ​​of other significant influencing factors, is input into the trained low voltage prediction model to obtain a clear prediction of whether a low voltage will occur.

[0055] In this way, the model can more comprehensively capture the dynamic change characteristics of the influencing factors, thereby enhancing the training effect of the model.

[0056] Based on the explanation of the training data set in step S3 in the above embodiment, we will continue to Figure 1 Step S4 in the following is explained: S4, training the to-be-trained model using the training data set to obtain a low voltage prediction model.

[0057] The study found that during the training process of the low-voltage prediction model, different hyperparameter settings can have a significant impact on model performance. For example, when the model to be trained is a random forest model, the number of trees, maximum depth, and minimum number of samples for leaf nodes are all hyperparameters. The choice of these hyperparameters directly affects the structure and behavior of the model. If the number of trees is too small, the model may be too simple and unable to fully capture the complex patterns in the data; if the number of trees is too large, the model may become too complex and overfit. Similarly, the maximum depth setting will also affect how well the model fits the data: too small a depth may lead to underfitting, while too large a depth may introduce too much noise.

[0058] In view of the above findings, if Figure 3 As shown, in Figure 2 Based on this, this embodiment provides the following optional implementation of step S4: S4-1, initialize the to-be-trained model through multiple sets of hyperparameters to obtain multiple initialized to-be-trained models.

[0059] Exemplarily, let's continue to take random forest as an example. Random forest relies on the settings of several key hyperparameters, such as the number of trees, the maximum depth of the tree, the minimum number of samples of leaf nodes, etc. The selection of these hyperparameters will directly affect the complexity, generalization ability and prediction accuracy of the model. However, in practical applications, due to the complex interaction between hyperparameters, it is difficult to directly determine the optimal hyperparameter combination through intuition or experience. In this regard, the present embodiment can use the grid method to traverse all possible parameter combinations within the preset parameter range to initialize the random forest, thereby obtaining the initialized random forest. The parameter range of each hyperparameter of the random forest tree model can refer to the following range: The range of the number of trees (1-300), the maximum depth of the tree (31), the minimum number of samples required (2-11), the minimum number of samples for leaf nodes (1-4), and whether to replace samples (yes). Set the maximum number of iterations to 1000 and the target error to 0.05%.

[0060] S4-2, training the multiple initialized models to be trained respectively using the training data set to obtain multiple candidate models.

[0061] S4-3, selecting multiple preferred models from multiple candidate models.

[0062] In order to select multiple optimal models from multiple candidate models, it is necessary to quantitatively evaluate the training effects of the multiple candidate models. In contrast, this embodiment provides the following optional implementation of step S4-3: S4-3-1, obtaining the training results of multiple candidate models after training with the training data set.

[0063] Optionally, this embodiment evaluates the training effect of the candidate model by integrating multiple dimensions. For each candidate model, the server determines the samples to be counted that are predicted by the candidate model as low voltage from multiple training samples; determines the number of positive samples from all the samples to be counted, where positive samples represent training samples with a voltage state label of low voltage; obtains a first ratio between the number of samples and the total number of samples to be counted, and a second ratio between the number of samples and the total number of positive samples in the multiple training samples; calculates the ratio product between the first ratio and the second ratio, and the ratio sum between the first ratio and the second ratio; and obtains the training effect of the candidate model based on the ratio between the ratio product and the ratio sum.

[0064] Among them, according to the ratio between the product of the proportions and the sum of the proportions, the expression of the training effect of the candidate model is:

[0065] Where, is the training score that represents the training effect, Indicates the first proportion, Indicates the second proportion.

[0066] For example, assume that there are 1000 training samples in the training data set, of which the actual voltage state label of 800 samples is low voltage (i.e. positive samples), and the remaining 200 samples are non-low voltage (i.e. negative samples). The candidate model identifies the 1000 training samples, predicts 750 training samples as positive samples, and predicts 250 training samples as negative samples. Among the 750 training samples predicted as positive samples, only 600 training samples are actually labeled as low voltage. Therefore, the first proportion is 600 / 750=0.8; similarly, the second proportion is 600 / 800=0.75. Substituting the first proportion and the second proportion into the above expression, the training score can be obtained as:

[0067] S4-3-2, selecting multiple preferred models from the multiple candidate models based on the training results after training with the training data set.

[0068] Optionally, in this embodiment, multiple candidate models are screened using a preset score threshold, which may be 0.9, and models with a training score greater than 0.9 are selected as preferred models.

[0069] The above embodiment explains how to select the preferred model. Figure 3 , step S4 further includes: S4-4, selecting a low voltage prediction model from the multiple preferred models based on the test results of the test data set on the multiple preferred models.

[0070] It should be understood that if a preferred model performs too well on training data (i.e., has a very high training score), it may overlearn specific noise or details in the training data, resulting in poor performance on unseen data (such as test data). To further evaluate the generalization ability of the preferred model, a test dataset is introduced. The test dataset consists of data that was not used in training and therefore effectively reflects the model's performance on unknown data. Each preferred model is verified in descending order of training score, and the test results of multiple preferred models are collected; the model with the best test result is selected as the low-voltage prediction model.

[0071] Based on the same inventive concept as the low voltage prediction model training method provided in this embodiment, this embodiment also provides a low voltage prediction model training method device, which includes at least one software function module that can be stored in a memory or solidified in an electronic device in the form of software. The processor in the electronic device is used to execute the executable module stored in the memory. For example, the software function module and computer program included in the device. Please refer to Figure 4 Functionally, the device can include: The data preparation module 11 is used to obtain a data set to be analyzed, wherein the data set to be analyzed includes multiple groups of original records before the low voltage occurs in the distribution network, and each group of original records includes observation values ​​of multiple influencing factors; The data preparation module 11 is further configured to screen out multiple significant influencing factors having a significant impact on low voltage from multiple influencing factors based on the data set to be analyzed; The data preparation module 11 is further configured to construct a training data set based on the multiple significant influencing factors, wherein the training data set includes multiple training samples, each training sample includes the observed values ​​of the multiple significant influencing factors and a voltage state label corresponding to the predicted time point; The model training module 12 is used to train the model to be trained using a training data set to obtain a low voltage prediction model.

[0072] In this embodiment, the data preparation module 11 is used to implement Figure 1 In steps S1, S2, and S3, the model training module 12 is used to implement Figure 1 Therefore, for a detailed description of each of the above modules, please refer to the specific implementation of the corresponding step.

[0073] Optionally, the data preparation module 11 is further specifically configured to: According to the data set to be analyzed, the impact scores of various influencing factors on low voltage are obtained; According to the impact scores of multiple influencing factors on low voltage, multiple significant influencing factors are selected from the multiple influencing factors, among which the multiple significant influencing factors are the top N influencing factors with the largest impact scores.

[0074] Optionally, the data preparation module 11 is further specifically configured to: The PCA algorithm is used to analyze the data set to be analyzed, and the impact scores of various influencing factors on low voltage are obtained.

[0075] Optionally, the model training module 12 is further specifically configured to: Initialize the to-be-trained model by using multiple sets of hyperparameters to obtain multiple initialized to-be-trained models; Using the training data set, multiple initialized models to be trained are trained separately to obtain multiple candidate models; Selecting multiple preferred models from multiple candidate models; According to the test results of the test data set on the multiple preferred models, a low voltage prediction model is selected from the multiple preferred models.

[0076] Optionally, the model training module 12 is further specifically configured to: Obtain the training results of multiple candidate models after training with the training data set; Based on the training results of multiple candidate models after training with the training data set, multiple preferred models are selected.

[0077] Optionally, the model training module 12 is further specifically configured to: For each candidate model, determine samples to be counted that are predicted by the candidate model to be low voltage from multiple training samples; Determine the number of positive samples from all samples to be counted, where the positive samples represent training samples whose voltage state label is low voltage; Obtaining a first ratio between the number of samples and the total number of samples to be counted, and a second ratio between the number of samples and the total number of positive samples in the plurality of training samples; Calculate the product of the first proportion and the second proportion, and the sum of the first proportion and the second proportion; The training effect of the candidate model is obtained according to the ratio between the product of the proportions and the sum of the proportions.

[0078] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0079] It should also be understood that if the above embodiments are implemented in the form of software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.

[0080] Therefore, this embodiment further provides a storage medium that is computer-readable. The storage medium stores a computer program that, when executed by a processor, implements the low-voltage prediction model training method provided in this embodiment. The storage medium can be any medium capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0081] This embodiment provides an electronic device for implementing a low voltage prediction model training method. Figure 5 As shown, the electronic device may include a processor 22 and a memory 21. In addition, the memory 21 stores a computer program, and the processor implements the low voltage prediction model training method provided in this embodiment by reading and executing the computer program corresponding to the above embodiment in the memory 21.

[0082] Continue to see Figure 5 The electronic device further includes a communication unit 23. The memory 21, the processor 22 and the communication unit 23 are electrically connected to each other directly or indirectly via a system bus 24 to achieve data transmission or interaction.

[0083] The memory 21 may be an information recording device based on any electronic, magnetic, optical or other physical principles, for recording execution instructions, data, etc. In some embodiments, the memory 21 may be, but is not limited to, a volatile memory, a non-volatile memory, a storage drive, etc.

[0084] In some embodiments, the volatile memory may be a random access memory (RAM); in some embodiments, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, etc.; in some embodiments, the storage drive may be a magnetic disk drive, a solid-state drive, any type of storage disk (such as a CD, DVD, etc.), or a similar storage medium, or a combination thereof.

[0085] The communication unit 23 is configured to transmit and receive data via a network. In some embodiments, the network may include a wired network, a wireless network, a fiber optic network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a wide area network (WAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, or a near field communication (NFC) network, or any combination thereof. In some embodiments, the network may include one or more network access points. For example, the network may include a wired or wireless network access point, such as a base station and / or a network switching node, through which one or more components of the service request processing system may connect to the network to exchange data and / or information.

[0086] The processor 22 may be an integrated circuit chip having signal processing capabilities, and the processor may include one or more processing cores (e.g., a single-core processor or a multi-core processor). By way of example only, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), an application-specific instruction-set processor (ASIP), a graphics processing unit (GPU), a physical processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a controller, a microcontroller unit, a reduced instruction set computer (RISC), or a microprocessor, or any combination thereof.

[0087] I understand. Figure 5 The structure shown is for illustration only. The electronic device 100 may also have Figure 5 More or fewer components than shown, or with Figure 5 Different configurations shown. Figure 5 The components shown may be implemented in hardware, software, or a combination thereof.

[0088] It should be understood that the devices and methods disclosed in the above embodiments may also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes may actually be executed substantially in parallel, or they may sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, may be implemented using a dedicated hardware-based system that performs a specified function or action, or may be implemented using a combination of dedicated hardware and computer instructions.

[0089] The above descriptions are merely examples of various embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A low voltage prediction model training method, characterized in that: The method comprises: Acquire a data set to be analyzed, wherein the data set to be analyzed includes multiple groups of original records before low voltage occurs in the distribution network, and each group of original records includes observation values ​​of multiple influencing factors; Screening out, from the plurality of influencing factors, a plurality of significant influencing factors having a significant impact on the low voltage according to the data set to be analyzed; Constructing a training data set based on the multiple significant influencing factors, wherein the training data set includes multiple training samples, each of the training samples includes observed values ​​of the multiple significant influencing factors and a voltage state label corresponding to a prediction time point; The model to be trained is trained using the training data set to obtain a low voltage prediction model.

2. The low voltage prediction model training method according to claim 1, characterized in that: Each of the training samples also includes change information of some of the significant influencing factors from the collection time point of its own observation value to the prediction time point.

3. The low voltage prediction model training method according to claim 1, characterized in that: According to the data set to be analyzed, multiple significant influencing factors having a significant impact on the low voltage are screened out from the multiple influencing factors, including: Obtaining, based on the data set to be analyzed, impact scores of the multiple influencing factors on low voltage; According to the impact scores of the multiple impact factors on low voltage, the multiple significant impact factors are selected from the multiple impact factors, wherein the multiple significant impact factors are top N impact factors with the largest impact scores.

4. The low voltage prediction model training method according to claim 3, characterized in that: Obtaining, based on the data set to be analyzed, impact scores of the multiple influencing factors on low voltage, including: The data set to be analyzed is analyzed using a PCA algorithm to obtain the impact scores of the multiple influencing factors on low voltage.

5. The low voltage prediction model training method according to claim 1, characterized in that: The training model is trained using the training data set to obtain a low voltage prediction model, including: Initializing the model to be trained using multiple sets of hyperparameters to obtain multiple initialized models to be trained; The plurality of initialized models to be trained are trained respectively using the training data set to obtain a plurality of candidate models; Selecting multiple preferred models from the multiple candidate models; A low voltage prediction model is selected from the multiple preferred models according to test results of the test data set on the multiple preferred models.

6. The low voltage prediction model training method according to claim 5, characterized in that: The model to be trained is a random forest model, and multiple preferred models are selected from the multiple candidate models, including: Obtaining training results of the multiple candidate models after training with the training data set; The multiple preferred models are selected from the multiple candidate models based on the training results after training with the training data set.

7. The low voltage prediction model training method according to claim 6, characterized in that: Obtaining training results of the multiple candidate models after training with the training data set includes: For each candidate model, determining samples to be counted that are predicted by the candidate model to be low voltage from the multiple training samples; Determine the number of positive samples from all the samples to be counted, wherein the positive samples represent training samples whose voltage state label is low voltage; Obtaining a first ratio between the number of samples and the total number of samples to be counted, and a second ratio between the number of samples and the total number of positive samples in the plurality of training samples; Calculate the product of the first proportion and the second proportion, and the sum of the first proportion and the second proportion; The training effect of the candidate model is obtained according to the ratio between the product of the proportions and the sum of the proportions.

8. A low voltage prediction model training device, characterized in that: The device comprises: A data preparation module is used to obtain a data set to be analyzed, wherein the data set to be analyzed includes multiple groups of original records before low voltage occurs in the distribution network, and each group of original records includes observation values ​​of multiple influencing factors; The data preparation module is further configured to screen out, from the plurality of influencing factors, a plurality of significant influencing factors having a significant impact on the low voltage based on the data set to be analyzed; The data preparation module is further configured to construct a training data set based on the multiple significant influencing factors, wherein the training data set includes multiple training samples, each of which includes observed values ​​of the multiple significant influencing factors and a voltage state label corresponding to a prediction time point; The model training module is used to train the model to be trained using the training data set to obtain a low voltage prediction model.

9. A storage medium, characterized in that: The storage medium stores a computer program, and the computer program, when executed by a processor, implements the low voltage prediction model training method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device includes a processor and a memory, the memory stores a computer program, and the computer program, when executed by the processor, implements the low voltage prediction model training method according to any one of claims 1 to 7.

Images