Incremental learning-based power supply quality evaluation method and system

Abstract

The invention discloses a power supply quality evaluation method and system based on incremental learning, and the method comprises the steps: carrying out the preprocessing and feature extraction of collected power supply data, and obtaining a power supply quality data set; adjusting environmental parameters of the power supply quality evaluation model by using power supply quality features, and training the power supply quality evaluation model by using the power supply quality data set to obtain an optimal power supply quality evaluation model; when the newly-added power supply data is monitored, extracting newly-added power supply quality features from the newly-added power supply data; and inputting the newly added power supply quality characteristics, the newly added power supply data and the historical power supply data into the optimal power supply quality evaluation model to obtain a power supply quality evaluation result of the current power grid. According to the method, the model parameters and the structure are automatically adjusted in an incremental learning mode to adapt to continuously increasing data volume, so that the power supply quality is quickly evaluated according to the power supply data monitored in real time, and the reliability and the stability of a power system are improved.

Description

Technical Field

[0001] The present invention belongs to the field of power quality evaluation, and in particular relates to a power quality evaluation method and system based on incremental learning. Background Art

[0002] Power quality is the quality of AC power, which is mainly provided by the power supply network to users in need. It can be transmitted, distributed and used at a sufficiently stable frequency, and is an important factor in the safety and stability of the power system. It includes four indicators: voltage quality, frequency quality, power supply reliability and serviceability. The main indicators used to evaluate power quality are voltage, frequency and waveform. Power quality can be defined as the quality of power provided to users, and users measure the reliability and safety of the power system. Generally, under ideal conditions, users are provided with a standard voltage with a constant frequency and a sinusoidal wave. In a three-phase AC system, it is required that the amplitude of the voltage and current of each phase should be equal, the phase should be symmetrical and the difference should be 120°. However, in actual situations, there is basically no ideal power state. This is because the system includes many nonlinear or asymmetric electrical equipment, or due to some special operations, external interference and faults, the power quality is affected to a certain extent.

[0003] The existing technology mainly evaluates the power quality by measuring and calculating the power indicators including voltage deviation, frequency deviation, three-phase imbalance, harmonics and interharmonics. However, these indicators are actually changing dynamically. Because of the uncertainty of data flow, it is difficult for the model to adapt to the change of data when calculating the power indicators. The model parameters and structure cannot continue to grow with the increase of data volume. Finally, the power quality evaluation result calculated is not accurate. Summary of the invention

[0004] The present invention proposes a power supply quality assessment method and system based on incremental learning, which automatically adjusts model parameters and structure through incremental learning to adapt to the growing amount of data, thereby quickly assessing the power supply quality based on real-time monitored power supply data, obtaining a comprehensive and accurate grid status and existing problems, and improving the reliability and stability of the power system.

[0005] A first aspect of the present invention provides a power supply quality assessment method based on incremental learning, the method comprising:

[0006] Preprocessing and feature extraction are performed on the collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes the preprocessed power supply data and power supply quality features;

[0007] Inputting the power quality data set into a preset power quality assessment model, first adjusting the environmental parameters of the power quality assessment model using the power quality characteristics, and then training the power quality assessment model using the power quality data set to obtain an optimal power quality assessment model;

[0008] When newly added power supply data is monitored, newly added power supply quality features are extracted from the newly added power supply data; wherein the newly added power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model;

[0009] The newly added power supply quality characteristics, the newly added power supply data and the historical power supply data are input into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

[0010] The above scheme first extracts the power supply quality characteristics of the power supply data to obtain a power supply quality data set, and then inputs the power supply quality data set into the power supply quality assessment model for learning, so as to obtain an optimal power supply quality assessment model that meets the power supply quality characteristics and has high precision; then, feature extraction is performed on the newly added power supply data to obtain newly added power supply quality characteristics that do not exist in the optimal power supply quality assessment model; finally, the newly added power supply quality characteristics, the newly added power supply data and the existing power supply data are input into the optimal power supply quality assessment model for incremental learning and evaluation calculation; by adding new power supply quality characteristics, the model parameters and structure can be automatically adjusted to adapt to the growing amount of data, so that the optimal power supply quality assessment model can better adapt to the changes and uncertainties of the data, and obtain a more accurate power supply quality assessment result of the current power grid, thereby obtaining a comprehensive and accurate status of the power grid and existing problems, and improving the reliability and stability of the power system.

[0011] In a possible implementation method of the first aspect, preprocessing and feature extraction are performed on the collected power supply data to obtain a power supply quality data set, specifically:

[0012] Deleting duplicate values, deleting abnormal values, and filling missing values ​​from the power supply data in turn to obtain preprocessed power supply data;

[0013] Extracting power supply quality features from the preprocessed power supply data and performing standardization processing to obtain power supply quality features;

[0014] Among them, deleting duplicate values ​​is to identify and delete duplicate data in the power supply data according to the device number and timestamp of the power supply data; deleting outliers is to identify and process outliers in the power supply data by constructing box plots and scatter plots; filling missing values ​​is to fill missing values ​​in the power supply data using interpolation and regression methods.

[0015] The above solution improves the accuracy of the power supply data by successively deleting duplicate values, deleting abnormal values ​​and filling missing values ​​in the power supply data.

[0016] In a possible implementation method of the first aspect, power supply quality features are extracted from the preprocessed power supply data and standardized to obtain power supply quality features, specifically:

[0017] Performing keyword extraction and format conversion on the text data in the preprocessed power supply data to obtain power supply keywords of numerical type;

[0018] Encoding the time series data in the preprocessed power supply data to obtain a power supply feature vector;

[0019] A power supply quality feature is obtained according to the power supply keyword and the power supply feature vector.

[0020] The above scheme first converts the data type of the preprocessed power supply data, converts the text type data into a numerical type to facilitate subsequent processing; then encodes the time series data, converts the time series data into a feature vector suitable for subsequent processing, and provides data support for subsequent model training.

[0021] In a possible implementation method of the first aspect, before inputting the power supply quality dataset into a preset power supply quality assessment model, the power supply quality dataset needs to be divided into a training set, a validation set, and a test set according to a preset ratio.

[0022] In order to improve the efficiency of model training and prevent overfitting, the above scheme divides the power supply quality data set into a training set, a validation set and a test set. The training set is used to train the model, the validation set is used to adjust the parameters of the model, and the test set is used to evaluate the performance of the model.

[0023] In a possible implementation method of the first aspect, the environmental parameters of the power quality assessment model are first adjusted using the power quality characteristics, and then the power quality assessment model is trained using the power quality data set to obtain an optimal power quality assessment model, specifically:

[0024] First, adjusting the values ​​of the environmental parameters of the power quality assessment model according to the power quality characteristics; wherein the environmental parameters include ambient temperature, ambient humidity and fault events;

[0025] Determining the type of the power supply quality assessment model according to the environmental parameters;

[0026] Training the power supply quality assessment model according to a training set of the power supply quality data set;

[0027] After each training cycle, the calculation parameters of the power supply quality assessment model are adjusted according to the validation set of the power supply quality data set, and then the power supply quality assessment model is evaluated using the test set of the power supply quality data set until an optimal power supply quality assessment model that meets the preset accuracy is obtained.

[0028] The above scheme first adjusts the values ​​of the environmental parameters of the power supply quality assessment model according to the power supply quality characteristics, so that the power supply quality assessment model meets the calculation requirements of the power supply quality data set and improves the calculation accuracy of the model; then the power supply quality assessment model is trained according to the training set of the power supply quality data set, and the calculation parameters of the power supply quality assessment model are adjusted using the verification set of the power supply quality data set to optimize the performance of the model, so as to obtain the optimal power supply quality assessment model that meets the preset accuracy.

[0029] In a possible implementation method of the first aspect, when newly added power supply data is monitored, newly added power supply quality features are extracted from the newly added power supply data, specifically:

[0030] When new power supply data is detected, the new power supply data is first cleaned to remove invalid and erroneous data;

[0031] The contribution of the newly added power supply data is calculated, and the newly added power supply quality feature is determined according to the newly added power supply data whose contribution and importance reach a second threshold.

[0032] When the above scheme detects the generation of new power supply data, it first cleans the new power supply data to improve the accuracy of the data, and then calculates the contribution of the new power supply data to extract the new power supply quality features of the new power supply data with high contribution, and uses them as the input of the optimal power supply quality assessment model for incremental learning, so that the model can continuously improve its accuracy and real-time performance over time.

[0033] In a possible implementation method of the first aspect, calculating the contribution of the newly added power supply data is specifically as follows:

[0034] The specific calculation formula for the contribution of the newly added power supply data is:

[0035]

[0036] In the formula, Score(f) is the score of the power supply quality feature of the newly added power supply data, IG(f) is the increment of a certain target variable brought by the score of the power supply quality feature, Corr(f,i) is the correlation between the power supply quality feature f and the power supply quality feature i, the power supply quality feature i is the power supply quality feature that has been input into the optimal power supply quality assessment model, Std(f) is the standard deviation of the power supply quality feature f, and ΔAcc(f) is the change of the optimal power supply quality assessment model after the power supply quality feature f is input.

[0037] A second aspect of the present invention provides a power supply quality assessment system based on incremental learning, the system comprising: a feature extraction module, a model training module, an incremental learning module and a power supply quality assessment module;

[0038] The feature extraction module is used to preprocess and extract features from the collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes the preprocessed power supply data and power supply quality features;

[0039] The model training module is used to input the power quality data set into a preset power quality assessment model, first use the power quality characteristics to adjust the environmental parameters of the power quality assessment model, and then use the power quality data set to train the power quality assessment model to obtain an optimal power quality assessment model;

[0040] The incremental learning module is used to clean the newly added power supply data and extract the newly added power supply quality features when monitoring the generation of newly added power supply data; wherein the newly added power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model;

[0041] The power supply quality assessment module is used to input the newly added power supply quality characteristics, the newly added power supply data and the existing power supply data into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid; wherein the existing power supply data is the first threshold data extracted from the historical power supply data.

[0042] The third aspect of the present invention provides a terminal device, which includes: a terminal device, including a processor and a memory, the memory storing a computer program, and the processor implementing the steps of a power supply quality assessment method based on incremental learning as described in any one of the embodiments of the present invention when executing the computer program.

[0043] A fourth aspect of the present invention provides a storage medium, which stores computer-readable program code. When the computer-readable program code is executed, the steps of a power supply quality assessment method based on incremental learning described in any one of the embodiments of the present invention are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 A specific flow chart of a power supply quality assessment method based on incremental learning is provided in a certain embodiment of the present invention;

[0046] Figure 2 A structural diagram of a power supply quality assessment system based on incremental learning is provided in a certain embodiment of the present invention;

[0047] Figure 3 A structural diagram of a terminal device is provided for a certain embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] It should be understood that the step numbers used in this article are only for the convenience of description and are not intended to limit the order in which the steps are executed.

[0050] like Figure 1 As shown, Figure 1 A specific flow chart of a power supply quality assessment method based on incremental learning is provided for a certain embodiment of the present invention. The power supply quality assessment method based on incremental learning of this embodiment includes steps S1 to S4, which are described in detail as follows:

[0051] Step S1, preprocessing and feature extraction of the collected power supply data to obtain a power supply quality data set;

[0052] In this step, power supply data is first collected on the power supply side and the power consumption side of the power supply link; wherein the power supply data includes voltage, current, power, frequency, etc.

[0053] Then, the power supply data is preprocessed by deleting duplicate values, deleting abnormal values, and filling missing values ​​in turn to obtain the preprocessed power supply data.

[0054] Among them, since there may be duplicate records in the power supply data, for example, the same device records the same voltage fluctuation in the same time period, it is necessary to set a unique identifier: device number and timestamp to identify and delete duplicate data; and deleting outliers is to remove data in the power supply data that obviously deviates from the normal range, such as values ​​that are too high or too low in voltage, through visualization methods, to establish box plots and scatter plots to identify and process outliers; filling missing values ​​is because there may be missing values ​​in the power supply data, such as a device does not record voltage values ​​in a certain time period, so it is necessary to use interpolation, regression and other methods to predict these missing values ​​and fill them in the power supply data to ensure data integrity.

[0055] After completing the preprocessing of the data, it is necessary to perform data conversion on the power supply data and extract relevant features required by the preset power supply quality assessment model. In some embodiments, the power supply quality assessment model requires power supply quality features such as voltage deviation rate, frequency deviation rate, and power factor. Therefore, the text data in the preprocessed power supply data is subjected to keyword extraction and format conversion, and the non-numeric power supply data is converted into a numerical type, such as describing the equipment operation status in the power supply data, and then extracting technical keywords based on text features, converting the text into a numerical value, and obtaining a power supply keyword of the numerical type. Then, the time series data in the power supply data is encoded, such as a voltage fluctuation curve, and the time series data can be converted into a feature vector suitable for machine learning by using methods such as symbolic coding and trend coding to obtain a power supply feature vector. Among them, the feature vector obtained by the time series data conversion is generally obtained by using a long short-term memory network or a gated cyclic unit timing model, and the feature vector can be used to effectively predict future power supply quality changes.

[0056] Finally, a power supply quality feature is obtained according to the power supply keyword and the power supply feature vector, and then a power supply quality data set is obtained according to the preprocessed power supply data and the power supply quality feature.

[0057] In addition, before inputting the power supply quality data set into the preset power supply quality assessment model, in order to improve training efficiency and prevent overfitting, the power supply quality data set needs to be divided into a training set, a validation set and a test set according to a preset ratio; wherein the training set is used to train the model, the validation set is used to adjust model parameters and select the best model, and the test set is used to evaluate the generalization performance of the model.

[0058] Step S2, inputting the power quality data set into a preset power quality assessment model, first using the power quality characteristics to adjust the environmental parameters of the power quality assessment model, and then using the power quality data set to train the power quality assessment model to obtain an optimal power quality assessment model;

[0059] In this step, the power quality data set is first input into a preset power quality assessment model, wherein the power quality assessment model is constructed based on historical data, with the power quality assessment result as the output variable and other related factors such as environmental parameters: ambient temperature, ambient humidity and fault events as input variables.

[0060] Firstly, the values ​​of the environmental parameters of the power supply quality assessment model are adjusted according to the power supply quality characteristics, and the type of the power supply quality assessment model is determined.

[0061] After the model is established, the power supply quality assessment model is trained using the training set, and the calculation parameters of the model are gradually adjusted during the training process, so that the error between the predicted results of the power supply quality assessment model and the actual results is gradually reduced. After each training cycle, the power supply quality assessment model is evaluated using the validation set, and relevant indicators such as accuracy, loss value, etc. are calculated. Then, the calculation parameters are adjusted or different training strategies are selected according to the evaluation results to optimize the performance of the model. The trained power supply quality assessment model is evaluated using the test set to understand the performance and generalization ability of the current power supply quality assessment model in practical applications, and the optimal calculation parameters or model type are determined according to the evaluation results to obtain a high-precision power supply quality assessment model as the optimal power supply quality assessment model.

[0062] Preferably, the prediction result of the power supply quality assessment model mainly describes the reliability and safety of the power supply quality.

[0063] The reliability mentioned above refers to the ability of the power system to provide power uninterruptedly. Reliability is mainly determined by factors such as the reliability of power generation equipment, transmission equipment, distribution equipment, substation equipment and user equipment. Efficiency: Efficiency refers to the ratio of energy consumed by the power system during operation to the energy output. A high-efficiency power system means more efficient energy utilization, which can reduce energy waste and costs;

[0064] The safety mentioned above refers to the ability of the power system to operate without failure or accidents. Safety is one of the most important indicators of the power system. Ensuring the safe operation of the power system can avoid major accidents and protect people's lives and property.

[0065] Step S3, when newly added power supply data is monitored, extracting newly added power supply quality features from the newly added power supply data;

[0066] In this step, the power supply link is first monitored regularly in real time. When new power supply data is detected, it is necessary to add the new power supply data to the optimal power supply quality assessment model so that the model can continuously improve its accuracy and real-time performance over time through incremental learning.

[0067] Before adding the newly added power supply data to the optimal power supply quality assessment model, it is necessary to use an online random forest algorithm to select a part of the newly added power supply data, and then perform data cleaning to remove invalid and erroneous data, and identify abnormal data for processing, reduce data uncertainty and improve data quality and consistency; wherein, isolated point detection, outlier detection, outlier detection, or mean, standard deviation, quartile, etc. are used to identify abnormal data that is significantly different from the distribution of the newly added power supply data. Alternatively, a deep learning model is used to identify abnormal data in the newly added power supply data, and the deep learning model includes a convolutional neural network or a long short-term memory network, etc. These deep learning models extract and classify features of the data. If a data point is significantly different from the feature distribution of normal data, it can be identified as abnormal data.

[0068] Then, the contribution of the newly added power supply data is calculated, and based on the newly added power supply data whose contribution and importance reach the second threshold, the newly added power supply quality features are determined to improve the accuracy and generalization ability of the model.

[0069] The specific calculation formula for the contribution of the newly added power supply data is:

[0070]

[0071] In the formula, Score(f) is the score of the power supply quality feature of the newly added power supply data, IG(f) is the increment of a certain target variable brought by the score of the power supply quality feature, Corr(f,i) is the correlation between the power supply quality feature f and the power supply quality feature i, the power supply quality feature i is the power supply quality feature that has been input into the optimal power supply quality assessment model, Std(f) is the standard deviation of the power supply quality feature f, and ΔAcc(f) is the change of the optimal power supply quality assessment model after the power supply quality feature f is input.

[0072] Step S4, inputting the newly added power supply quality characteristics, the newly added power supply data and the historical power supply data into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

[0073] In this step, a part of the historical power supply data is extracted using stratified sampling or Bootstrap sampling methods and added to the optimal power supply quality assessment model together with the newly added power supply data. This can effectively balance the new and old data and avoid data skewness and model overfitting.

[0074] Then, the newly added power supply quality characteristics, the newly added power supply data and the historical power supply data are input into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

[0075] Based on the power supply quality assessment results of the current power grid, problems and hidden dangers in the power grid can be discovered and resolved in a timely manner, which can help operators predict possible failures in the power grid, fully understand the current status and existing problems of the power grid, and take preventive measures to reduce the probability of failures and improve the reliability and stability of the power system.

[0076] Further, in order to execute the power supply quality assessment system based on incremental learning corresponding to the above method embodiment to achieve the corresponding functions and technical effects, Figure 2 A structural diagram of a power supply quality assessment system based on incremental learning is provided. For ease of description, only the parts related to this embodiment are shown. The power supply quality assessment system based on incremental learning provided by the embodiment of the present invention includes:

[0077] The feature extraction module 201 is used to preprocess and extract features from the collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes the preprocessed power supply data and power supply quality features;

[0078] A model training module 202 is used to input the power quality data set into a preset power quality assessment model, first use the power quality characteristics to adjust the environmental parameters of the power quality assessment model, and then use the power quality data set to train the power quality assessment model to obtain an optimal power quality assessment model;

[0079] The incremental learning module 203 is used to extract new power supply quality features from the new power supply data when new power supply data is monitored; wherein the new power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model;

[0080] The power supply quality assessment module 204 is used to input the newly added power supply quality characteristics, the newly added power supply data and the historical power supply data into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

[0081] In some embodiments, the feature extraction module 201 further includes:

[0082] A data preprocessing unit is used to sequentially delete duplicate values, delete abnormal values, and fill missing values ​​in the power supply data to obtain preprocessed power supply data; extract power supply quality features from the preprocessed power supply data and perform standardization processing to obtain power supply quality features; wherein, deleting duplicate values ​​is to identify and delete duplicate data in the power supply data based on the device number and timestamp of the power supply data; deleting abnormal values ​​is to identify and process abnormal values ​​in the power supply data by constructing box plots and scatter plots; filling missing values ​​is to fill missing values ​​in the power supply data using interpolation and regression methods.

[0083] The power supply quality feature extraction unit is used to extract keywords and convert the format of the text data in the preprocessed power supply data to obtain power supply keywords of numerical type; encode the time series data in the preprocessed power supply data to obtain a power supply feature vector; and obtain power supply quality features based on the power supply keywords and the power supply feature vector.

[0084] The data division unit is used to divide the power supply quality data set into a training set, a validation set and a test set according to a preset ratio before inputting the power supply quality data set into a preset power supply quality assessment model.

[0085] In some embodiments, the model training module 202 further includes:

[0086] A model training unit is used to first adjust the values ​​of environmental parameters of the power supply quality assessment model according to power supply quality characteristics; wherein the environmental parameters include ambient temperature, ambient humidity and fault events; according to the environmental parameters, determine the type of the power supply quality assessment model; train the power supply quality assessment model according to the training set of the power supply quality data set; after each training cycle, adjust the calculation parameters of the power supply quality assessment model according to the verification set of the power supply quality data set, and then use the test set of the power supply quality data set to evaluate the power supply quality assessment model until an optimal power supply quality assessment model that meets a preset accuracy is obtained.

[0087] In some embodiments, the incremental learning module 203 further includes:

[0088] A new power supply quality feature acquisition module is added, which is used to clean the new power supply data first to remove invalid and erroneous data when new power supply data is monitored; calculate the contribution of the new power supply data, and determine the new power supply quality features based on the new power supply data whose contribution and importance reach the second threshold.

[0089] Furthermore, Figure 3 This is a structural diagram of a terminal device provided in one embodiment of the present application. Figure 3 As shown, the terminal device 3 of this embodiment includes: at least one processor 30 (in Figure 3 Only one is shown) and a memory 31 and a computer program 32 stored in the memory 31 and executable on the at least one processor. When the processor 30 executes the computer program 32, the steps of a power supply quality assessment method based on incremental learning as described in any one of the embodiments of the present invention can be implemented.

[0090] The terminal device 3 may be a computing device such as a desktop computer, a cloud server, or a laptop computer. The computing device may include but is not limited to a processor 30 and a memory 31 . Figure 3 This is merely an example of the terminal device 3 and does not constitute a limitation on the terminal device 3 , which may include more or fewer components than those shown in the figure.

[0091] An embodiment of the present invention provides a storage medium, wherein the storage medium stores computer-readable program codes, and when the computer-readable program codes are executed, the steps of the above-mentioned power supply quality assessment method based on incremental learning are implemented.

[0092] This embodiment proposes a power supply quality assessment method and system based on incremental learning: preprocessing and feature extraction of collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes preprocessed power supply data and power supply quality features; inputting the power supply quality data set into a preset power supply quality assessment model, first using the power supply quality features to adjust the environmental parameters of the power supply quality assessment model, and then using the power supply quality data set to train the power supply quality assessment model to obtain an optimal power supply quality assessment model; when newly added power supply data is monitored, extracting newly added power supply quality features from the newly added power supply data; wherein the newly added power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model; inputting the newly added power supply quality features, the newly added power supply data and the historical power supply data into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid. Its beneficial effect is: automatically adjusting the model parameters and structure through incremental learning to adapt to the growing amount of data, so as to quickly assess the power supply quality according to the real-time monitored power supply data, obtain a comprehensive and accurate status of the power grid and existing problems, and improve the reliability and stability of the power system.

[0093] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A power supply quality assessment method based on incremental learning, characterized in that: include: Preprocessing and feature extraction are performed on the collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes the preprocessed power supply data and power supply quality features; Inputting the power quality data set into a preset power quality assessment model, first adjusting the environmental parameters of the power quality assessment model using the power quality characteristics, and then training the power quality assessment model using the power quality data set to obtain an optimal power quality assessment model; When newly added power supply data is monitored, newly added power supply quality features are extracted from the newly added power supply data; wherein the newly added power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model; The newly added power supply quality characteristics, the newly added power supply data and the historical power supply data are input into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

2. The power supply quality assessment method based on incremental learning according to claim 1, characterized in that: The collected power supply data is preprocessed and feature extracted to obtain a power supply quality data set, specifically: Deleting duplicate values, deleting abnormal values, and filling missing values ​​from the power supply data in turn to obtain preprocessed power supply data; Extracting power supply quality features from the preprocessed power supply data and performing standardization processing to obtain power supply quality features; Among them, deleting duplicate values ​​is to identify and delete duplicate data in the power supply data according to the device number and timestamp of the power supply data; deleting outliers is to identify and process outliers in the power supply data by constructing box plots and scatter plots; filling missing values ​​is to fill missing values ​​in the power supply data using interpolation and regression methods.

3. The power supply quality assessment method based on incremental learning according to claim 2, characterized in that: The power supply quality feature is extracted from the pre-processed power supply data and standardized to obtain the power supply quality feature, specifically: Performing keyword extraction and format conversion on the text data in the preprocessed power supply data to obtain power supply keywords of numerical type; Encoding the time series data in the preprocessed power supply data to obtain a power supply feature vector; A power supply quality feature is obtained according to the power supply keyword and the power supply feature vector.

4. The power supply quality assessment method based on incremental learning according to claim 1, characterized in that: Before inputting the power quality data set into a preset power quality assessment model, the power quality data set needs to be divided into a training set, a validation set and a test set according to a preset ratio.

5. The power supply quality assessment method based on incremental learning according to claim 1, characterized in that: The first step of using the power supply quality characteristics to adjust the environmental parameters of the power supply quality assessment model, and then using the power supply quality data set to train the power supply quality assessment model to obtain the optimal power supply quality assessment model is as follows: First, adjusting the values ​​of the environmental parameters of the power quality assessment model according to the power quality characteristics; wherein the environmental parameters include ambient temperature, ambient humidity and fault events; Determining the type of the power supply quality assessment model according to the environmental parameters; Training the power supply quality assessment model according to a training set of the power supply quality data set; After each training cycle, the calculation parameters of the power supply quality assessment model are adjusted according to the validation set of the power supply quality data set, and then the power supply quality assessment model is evaluated using the test set of the power supply quality data set until an optimal power supply quality assessment model that meets the preset accuracy is obtained.

6. The power supply quality assessment method based on incremental learning according to claim 1, characterized in that: When the newly added power supply data is monitored, newly added power supply quality features are extracted from the newly added power supply data, specifically: When new power supply data is detected, the new power supply data is first cleaned to remove invalid and erroneous data; The contribution of the newly added power supply data is calculated, and the newly added power supply quality feature is determined according to the newly added power supply data whose contribution and importance reach a second threshold.

7. The power supply quality assessment method based on incremental learning according to claim 6, characterized in that: The calculation of the contribution of the newly added power supply data is specifically as follows: The specific calculation formula for the contribution of the newly added power supply data is: In the formula, Score(f) is the score of the power supply quality feature of the newly added power supply data, IG(f) is the increment of a certain target variable brought by the score of the power supply quality feature, Corr(f,i) is the correlation between the power supply quality feature f and the power supply quality feature i, the power supply quality feature i is the power supply quality feature that has been input into the optimal power supply quality assessment model, Std(f) is the standard deviation of the power supply quality feature f, and ΔAcc(f) is the change of the optimal power supply quality assessment model after the power supply quality feature f is input.

8. A power supply quality assessment system based on incremental learning, characterized in that: include: Feature extraction module, model training module, incremental learning module and power supply quality assessment module; The feature extraction module is used to preprocess and extract features from the collected power supply data to obtain a power supply quality data set; wherein the power supply quality data set includes the preprocessed power supply data and power supply quality features; The model training module is used to input the power quality data set into a preset power quality assessment model, first use the power quality characteristics to adjust the environmental parameters of the power quality assessment model, and then use the power quality data set to train the power quality assessment model to obtain an optimal power quality assessment model; The incremental learning module is used to extract new power supply quality features from the new power supply data when new power supply data is monitored; wherein the new power supply quality features are power supply quality features that have never been input into the optimal power supply quality assessment model; The power supply quality assessment module is used to input the newly added power supply quality characteristics, the newly added power supply data and the historical power supply data into the optimal power supply quality assessment model to obtain the power supply quality assessment result of the current power grid.

9. A terminal device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the method implements the steps of a power supply quality assessment method based on incremental learning as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores computer-readable program codes, and when the computer-readable program codes are executed, the steps of a power supply quality assessment method based on incremental learning according to any one of claims 1 to 7 are implemented.

Images