Model training method, false measurement interval prediction method and device based on fluctuation error time variation, equipment and medium

Abstract

The invention discloses a model training method, a false measurement interval prediction method and device based on fluctuation error time-varying characteristics, equipment and a medium. The method comprises the following steps: acquiring measurement sample data of a to-be-predicted node in a power system, wherein the measurement sample data comprises a plurality of sample historical periods and the sample data of a measurement true value of a reference sample period; for the measurement sample data of the plurality of sample historical periods, determining a fluctuation error of the measurement sample data of each sample historical moment in the current sample historical period based on a difference value of the measurement sample data of any two adjacent sample historical moments in the current sample historical period; determining a Gaussian process sample set based on fluctuation errors of the measurement sample data of the plurality of sample historical periods and the measurement sample data of the same sample historical moment corresponding to the plurality of sample historical periods; and training a pseudo measurement interval prediction model based on the Gaussian process sample set and the measurement true value of the reference sample period. And the precision and credibility of pseudo measurement interval prediction performed by the pseudo measurement interval prediction model are effectively improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of power system state estimation, and in particular to a model training method, a pseudo-measurement interval prediction method based on time-varying fluctuation errors, a device, equipment and a medium. Background Art

[0002] As the scale and complexity of power systems continue to expand, the importance of accurately estimating the state of power systems has become increasingly prominent. Power system state estimation is crucial to ensuring the stable operation of the power grid, optimizing the allocation of power resources, preventing potential failures, and improving overall energy efficiency. However, in actual power systems, due to economic cost constraints, the measurement equipment of the distribution network may not be sufficient or there may be deficiencies in the layout, resulting in the measurement data may not be completely redundant. This measurement limitation may cause data loss at certain key nodes, affecting the accuracy and reliability of state estimation.

[0003] When the measurement data of some nodes in the system become unobservable due to equipment failure, communication interruption or other reasons, the traditional state estimation program may not be able to operate normally. In this case, pseudo-measurement prediction technology provides an effective remedy. Pseudo-measurement prediction technology generates missing measurement data through simulation or prediction to supplement the deficiency of actual measurement data, so that the state estimation program can continue to execute based on incomplete data.

[0004] However, the current pseudo-measurement interval prediction methods based on the time-varying nature of fluctuation errors suffer from low accuracy and reliability. Summary of the invention

[0005] The present invention provides a model training method, a pseudo-measurement interval prediction method based on time-varying fluctuation errors, a device, equipment and a medium to solve the problems of low accuracy and credibility in the existing pseudo-measurement interval prediction method based on time-varying fluctuation errors.

[0006] According to one aspect of the present invention, a pseudo-measurement interval prediction model training method is provided, comprising:

[0007] Acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of a plurality of sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the plurality of sample history cycles;

[0008] For the measured sample data of the plurality of sample history periods, based on the difference between the measured sample data of any two adjacent sample history moments in the current sample history period, determine the fluctuation error of the measured sample data at each sample history moment in the current sample history period;

[0009] Determine a Gaussian process sample set based on the measured sample data of the multiple sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods;

[0010] Based on the Gaussian process sample set and the true measurement value of the reference sample period, a pseudo measurement interval prediction model is trained, wherein the pseudo measurement interval prediction model is determined based on a mean function and a covariance function.

[0011] According to another aspect of the present invention, a pseudo measurement interval prediction method based on time-varying fluctuation error is provided, comprising:

[0012] Obtain measurement data of multiple historical periods of the nodes to be predicted in the power system;

[0013] For the measurement data of multiple historical periods, based on the difference between the measurement data of any two adjacent historical moments in the current historical period, determine the fluctuation error of the measurement data at each historical moment in the current historical period;

[0014] Determine a Gaussian process data set based on measurement data of multiple historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the multiple historical periods;

[0015] Inputting the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period;

[0016] The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, there is provided a pseudo-measurement interval prediction model training device, comprising:

[0018] A sample data acquisition module, used to acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of a plurality of sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the plurality of sample history cycles;

[0019] A first error determination module is used to determine the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period based on the difference between the measured sample data at any two adjacent sample historical moments in the current sample historical period for the measured sample data of the multiple sample historical periods;

[0020] A sample set determination module, configured to determine a Gaussian process sample set based on the measured sample data of the plurality of sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the plurality of sample history periods;

[0021] A model training module is used to train a pseudo-measurement interval prediction model based on the Gaussian process sample set and the true measurement value of the reference sample period, wherein the pseudo-measurement interval prediction model is determined based on a mean function and a covariance function.

[0022] According to another aspect of the present invention, there is provided a pseudo measurement interval prediction device based on the time-varying property of fluctuation error, comprising:

[0023] A data acquisition module, used to acquire measurement data of multiple historical periods of a node to be predicted in the power system;

[0024] A second error acquisition module is used to determine the fluctuation error of the measurement data at each historical moment in the current historical period based on the difference between the measurement data at any two adjacent historical moments in the current historical period for the measurement data of multiple historical periods;

[0025] A data set determination module, configured to determine a Gaussian process data set based on measurement data of a plurality of historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the plurality of historical periods;

[0026] An interval prediction module, used for inputting the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period;

[0027] The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method described in any embodiment of the present invention.

[0028] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0029] at least one processor; and

[0030] a memory communicatively connected to the at least one processor; wherein,

[0031] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the pseudo-measurement interval prediction model training method described in any embodiment of the present invention or execute the pseudo-measurement interval prediction method based on the time-varying fluctuation error described in any embodiment of the present invention.

[0032] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the pseudo-measurement interval prediction model training method described in any embodiment of the present invention or the pseudo-measurement interval prediction method based on the time-varying fluctuation error described in any embodiment of the present invention when executed.

[0033] The technical solution provided by the embodiment of the present invention obtains sample data of a node to be predicted in an electric power system, wherein the sample data includes measured sample data of multiple sample history periods and measured true values ​​of a reference sample period, and the reference sample period is located after the multiple sample history periods; for the measured sample data of the multiple sample history periods, based on the difference between the measured sample data of any two adjacent sample history moments in the current sample history period, the fluctuation error of the measured sample data at each sample history moment in the current sample history period is determined; based on the fluctuation errors of the measured sample data of the multiple sample history periods and the measured sample data at the same sample history moments corresponding to the multiple sample history periods, a Gaussian process sample set is determined; based on the Gaussian process sample set and the measured true values ​​of the reference sample period, a pseudo-measurement interval prediction model is trained, wherein the pseudo-measurement interval prediction model is determined based on a mean function and a covariance function. Through the above technical scheme, based on the difference of the measured sample data at any two adjacent sample historical moments in the current sample historical period, the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period is determined, and then based on the measured sample data and fluctuation errors of multiple sample historical periods, the Gaussian process sample set is determined, so that the Gaussian process sample set contains the distribution characteristics of the fluctuation error of the measurement sample data and its law of change over time, and then based on the Gaussian process sample set and the true measurement value of the reference sample period, the pseudo-measurement interval prediction model is trained, which effectively improves the accuracy and credibility of the pseudo-measurement interval prediction model in predicting the pseudo-measurement interval.

[0034] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0036] Figure 1is a flowchart of a pseudo-measurement interval prediction model training method provided in Embodiment 1 of the present invention;

[0037] Figure 2 It is a schematic diagram of an interval prediction result provided by the present invention;

[0038] Figure 3 This is a flow chart of a pseudo-measurement interval prediction method based on the time-varying nature of fluctuation errors provided in the second embodiment of the present invention;

[0039] Figure 4 It is a structural schematic diagram of a pseudo-measurement interval prediction model training device provided in Embodiment 3 of the present invention;

[0040] Figure 5 It is a structural schematic diagram of a pseudo-measurement interval prediction device based on the time-varying property of fluctuation error provided by the third embodiment of the present invention;

[0041] Figure 6 It is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0043] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] Embodiment 1

[0045] Figure 1This is a flowchart of a pseudo measurement interval prediction model training method provided by the first embodiment of the present invention. This embodiment is applicable to the case of training a pseudo measurement interval prediction model. The method can be executed by a pseudo measurement interval prediction model training device. The pseudo measurement interval prediction model training device can be implemented in the form of hardware and / or software. The pseudo measurement interval prediction model training device can be configured in an electronic device. Figure 1 As shown, the method includes:

[0046] S110. Acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of multiple sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the multiple sample history cycles.

[0047] In this embodiment, the power system is a complex system composed of multiple links and components, which is responsible for the production, transmission, distribution and use of electric energy. The power system includes multiple subsystems, such as a distribution subsystem and a transmission subsystem. The nodes in the power system can be understood as the connection points between electrical equipment in the power system. The sample data includes the measured sample data of multiple sample historical cycles and the measured true value of the reference sample cycle, and the reference sample cycle is located after the multiple sample historical cycles. Multiple sample historical cycles can be understood as a preset number of historical cycles. Among them, the length of the historical cycle can be set to different time units as needed, including but not limited to days, hours or minutes. The measured sample data can be understood as the power or load power of the node that needs to be predicted in the power system at the observable moment, and the measured sample data is obtained by real-time collection of sensing equipment. In this embodiment, the measured sample data is explained as power. The measured true value at the sample moment to be predicted can be understood as the measured true value of the node at the moment to be predicted.

[0048] Specifically, in order to effectively train the pseudo-measurement interval prediction model, sample data of the node to be predicted in the power system is obtained, and the sample data includes measurement sample data of multiple sample history cycles and the true value of the measurement at the time to be predicted. Wherein, the power system includes at least one node to be predicted.

[0049] For example, assume that the sensor equipment of a node to be predicted in the power system collects the power of the node at a frequency of once every 5 minutes, and obtains sample data from the sensor equipment of the node to be predicted, and the sample data includes the power collected by the sensor equipment in real time within 5 days. In order to effectively train the pseudo-measurement interval prediction model, assume that the node fails at time 0 on the 5th day, and it is necessary to use the sample data of the previous 4 days to predict the interval of the measurement data of the node at each moment on the 5th day, where the sample data of the previous 4 days is the measurement sample data of 4 sample history cycles. In this example, the sample history cycle is one day.

[0050] S120. For the measured sample data of the multiple sample history periods, determine the fluctuation error of the measured sample data at each sample history moment in the current sample history period based on the difference between the measured sample data at any two adjacent sample history moments in the current sample history period.

[0051] In this embodiment, the fluctuation error can be used to measure the fluctuation between adjacent measured sample data. The fluctuation error has the characteristic of changing over time. During periods of drastic power and load fluctuations, the fluctuation error is usually larger; during periods of gentle power and load fluctuations, the fluctuation error is usually smaller. A sample history cycle includes multiple sample history moments. The fluctuation error of the measured sample data of the current sample history cycle includes the fluctuation error of the measured sample data at each sample history moment in the current sample history cycle.

[0052] Specifically, the difference between the measured sample data at any two adjacent sample historical moments in the current sample historical period can be calculated by the following formula:

[0053]

[0054] Among them, N represents the total number of sample history periods, T represents the set of sample history moments, It represents the difference between the measured sample data at the sample history time tk+1 and the measured sample data at the sample history time tk within the current sample history period n. The difference can represent the fluctuation error of the measured sample data at the sample history time tk. represents the measured sample data at the sample historical time tk+1, Represents the measured sample data at the sample historical moment tk. Thus, the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period n can be calculated.

[0055] Then, by repeating the above process, the fluctuation error of the measured sample data at each sample historical moment in each sample historical period can be obtained.

[0056] S130. Determine a Gaussian process sample set based on the measured sample data of the multiple sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods.

[0057] In this embodiment, the Gaussian process sample set includes a plurality of Gaussian process samples, wherein one sample history period corresponds to one Gaussian process sample.

[0058] Specifically, in order to reflect the dynamic characteristics of fluctuation errors changing over time, the fluctuation errors of the measured sample data at the same sample historical moments corresponding to multiple sample historical periods can be combined to calculate the distribution characteristics of the fluctuation errors at the sample historical moments. Then, based on the distribution characteristics of the fluctuation errors of the measured sample data and the sample historical moments of multiple sample historical periods, the Gaussian process sample set can be determined.

[0059] Exemplarily, assuming that the power fluctuation errors at 10 o'clock every day for 4 days are w1, w2, w3, and w4 respectively, the distribution characteristics of the fluctuation errors at 10 o'clock can be calculated based on w1, w2, w3, and w4.

[0060] S140. Based on the Gaussian process sample set and the true measurement value of the reference sample period, a pseudo measurement interval prediction model is trained, wherein the pseudo measurement interval prediction model is determined based on a mean function and a covariance function.

[0061] In this embodiment, the mean function is used to describe the overall trend of the Gaussian process output. The mean function can be a constant function, a linear function, a polynomial function, etc. The covariance function is used to describe the similarity or correlation of the output values ​​between different input points. The covariance function can be a Gaussian kernel function, a linear kernel function, a polynomial kernel function, etc. The selection of the mean function and the covariance function can be determined according to the needs of the specific problem. The pseudo-measurement interval prediction model is a Gaussian process regression model, which is used to predict the pseudo-measurement interval of the node to be predicted. Among them, pseudo-measurement can be understood as virtual measurement data generated by calculation or prediction when the actual measurement data is insufficient.

[0062] Specifically, the Gaussian process sample set and the measured true values ​​of the reference sample period constitute the training data.

[0063] Define the pseudo-measurement interval prediction model, including the selection of mean function and covariance function. The pseudo-measurement interval prediction model can be expressed as:

[0064] f(x)~GP(m(x),κ(x,x))

[0065] Where m(x) represents the mean function and κ(x,x) represents the covariance function.

[0066] Furthermore, the hyperparameters of the pseudo-measurement interval prediction model are initialized using prior knowledge or random initialization methods. These hyperparameters mainly refer to the parameters of the covariance function. The prior model is generated using probability distribution through the input training data and the initialized hyperparameters. The prior model describes the potential relationship between the Gaussian process sample set and the true value of the measurement of the reference sample period. The pseudo-measurement interval prediction model is then trained by minimizing the negative log marginal likelihood (NLML) to find the covariance function hyperparameters that make the pseudo-measurement interval prediction model perform best on the training data, so that the model can achieve the best prediction effect.

[0067] Optionally, the performance of the trained pseudo-measurement interval prediction model can be evaluated by obtaining a test set. The test set contains the same variables and features as the training data, but consists of a set of observations or samples that do not overlap with the training data. It is used to evaluate the generalization performance of the model. The test set is input into the pseudo-measurement interval prediction model, and the mean and variance of the prediction results are output. The prediction result is a Gaussian distribution, including two indicators: mean and variance. The mean indicates the center position of the prediction result, and the variance indicates the degree of uncertainty of the prediction result.

[0068] The technical solution provided in the first embodiment of the present invention obtains sample data of a node to be predicted in an electric power system, wherein the sample data includes measured sample data of multiple sample history periods and measured true values ​​of a reference sample period, and the reference sample period is located after the multiple sample history periods; for the measured sample data of the multiple sample history periods, based on the difference between the measured sample data of any two adjacent sample history moments in the current sample history period, the fluctuation error of the measured sample data at each sample history moment in the current sample history period is determined; based on the fluctuation errors of the measured sample data of the multiple sample history periods and the measured sample data at the same sample history moments corresponding to the multiple sample history periods, a Gaussian process sample set is determined; based on the Gaussian process sample set and the measured true values ​​of the reference sample period, a pseudo measurement interval prediction model is trained, wherein the pseudo measurement interval prediction model is determined based on a mean function and a covariance function. Through the above technical scheme, based on the difference of the measured sample data at any two adjacent sample historical moments in the current sample historical period, the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period is determined, and then based on the measured sample data and fluctuation errors of multiple sample historical periods, the Gaussian process sample set is determined, so that the Gaussian process sample set contains the distribution characteristics of the fluctuation error of the measurement sample data and its law of change over time, and then based on the Gaussian process sample set and the true measurement value of the reference sample period, the pseudo-measurement interval prediction model is trained, which effectively improves the accuracy and credibility of the pseudo-measurement interval prediction model in predicting the pseudo-measurement interval.

[0069] In some embodiments, based on the fluctuation errors of the measured sample data of the multiple sample history periods and the measured sample data at the same sample history moments corresponding to the multiple sample history periods, a Gaussian process sample set is determined, including: based on the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods, the error expectation and error variance of the multiple same sample history moments are determined; based on the measured sample data of the multiple sample history periods and the error expectation and error variance of the multiple same sample history moments, the Gaussian process samples of each sample history period are determined; based on the Gaussian process samples of each sample history period, the Gaussian process sample set is determined. Through the above technical solution, the constructed Gaussian process sample set not only takes into account the actual observation values ​​of the data, but also incorporates the fluctuation errors of the data, laying the foundation for further improving the prediction accuracy, robustness and generalization ability of the model.

[0070] In this embodiment, the error expectation represents the average value or central tendency of the fluctuation error, and the error variance represents the degree of dispersion or uncertainty of the fluctuation error. The error expectation and error variance reflect the distribution characteristics of the fluctuation error at the historical moment of the sample.

[0071] Specifically, the fluctuation errors of the measured sample data at the same sample historical moments corresponding to multiple sample historical periods are added together and averaged to obtain the expected errors at each same sample historical moment, and then the error variances at each same sample historical moment are calculated based on the fluctuation errors of the measured sample data at the same sample historical moments corresponding to multiple sample historical periods and the expected errors at each same sample historical moment. For example, the calculated fluctuation errors of the power at 10 o'clock each day for 4 days are w1, w2, w3, and w4 respectively, and the expected errors and error variances at 10 o'clock can be calculated based on w1, w2, w3, and w4.

[0072] Then, the expected errors and error variances of multiple identical sample historical moments are superimposed on the measured sample data of each sample historical period, and the Gaussian process samples of each sample historical period can be obtained. Optionally, the Gaussian process samples of each sample historical period are processed, and the Gaussian process sample set is determined based on the processing results. The processing operation may be data cleaning, data standardization, or clustering.

[0073] In some embodiments, based on the measurement sample data of the multiple sample history periods and the error expectation and error variance of the multiple same sample history moments, the Gaussian process samples of each sample history period are determined, including: for each sample history period in the multiple sample history periods, based on the measurement sample data of the current sample history period and the error expectation and error variance of the multiple same sample history moments, the sample interval corresponding to the measurement sample data of each sample history moment in the current sample history period is calculated, and the Gaussian process samples of the current sample history period are determined based on the sample interval corresponding to the measurement sample data of each sample history moment. Through the above technical solution, the fluctuation error and the measurement sample data are combined to construct the Gaussian process samples, which realizes the accurate quantification of the uncertainty of the data, allows the model to consider the inherent uncertainty of the data when predicting the data, and lays a solid foundation for further improving the accuracy of the prediction of the pseudo-measurement interval prediction model.

[0074] Specifically, for each sample history period in multiple sample history periods, the following method can be used to determine the Gaussian process samples of each sample history period:

[0075] Calculate the first sum of the error expectation and the error variance of the first preset multiple of each identical sample historical moment, and calculate the first difference between the error expectation and the error variance of multiple identical sample historical moments and the second preset multiple, wherein the first preset multiple and the second preset multiple can be set to be the same or different according to actual conditions.

[0076] Then, the first sum and the first difference are added to the measured sample data at the corresponding sample history moments, respectively, to obtain the sample interval corresponding to the measured sample data at each sample history moment in the current sample history period. Then, the sample interval corresponding to the measured sample data at each sample history moment is determined as the Gaussian process sample of the current sample history period.

[0077] Therefore, the Gaussian process samples of multiple historical periods can be expressed as follows:

[0078]

[0079] in, represents a collection of Gaussian process samples over multiple sample history periods, represents the Gaussian process sample of the Nth historical period, Represents the sample interval corresponding to the measured sample data at the sample historical time tk+1 in the Nth historical period.

[0080] In some embodiments, a Gaussian process sample set is determined based on Gaussian process samples in each sample history period, including: clustering the Gaussian process samples in each sample history period to obtain Gaussian process cluster samples; and determining the Gaussian process sample set based on the Gaussian process cluster samples.

[0081] Specifically, in order to avoid the variance of the prediction result being too large and thus losing the meaning of interval prediction, it is necessary to use a preset clustering algorithm to cluster the Gaussian process samples of each sample history period to obtain Gaussian process clustering samples, wherein the specific preset clustering algorithm is not limited in this embodiment, for example, it can be k-means clustering or Gaussian mixture model clustering. Then, based on the Gaussian process clustering samples, a Gaussian process sample set is determined. In this way, Gaussian process samples with excessive differences are removed, laying a solid foundation for further enhancing the accuracy and credibility of the model.

[0082] In some embodiments, clustering the Gaussian process samples of each sample history period to obtain Gaussian process cluster samples includes: determining the number of Gaussian distribution subsets; calculating the posterior probability that the Gaussian process samples of each sample history period belong to the Gaussian distribution subset; based on the posterior probability that the Gaussian process samples of each sample history period belong to the Gaussian distribution subset, clustering the Gaussian process samples of each sample history period to obtain Gaussian process cluster samples. Through the above technical solution, clustering of Gaussian process samples of each sample history period is achieved, effectively removing samples with too large differences, reducing the variance of fluctuation errors, and ensuring that the predicted confidence interval is meaningful.

[0083] In this embodiment, the number of Gaussian distribution subsets is used to determine how many groups or “clusters” the Gaussian process samples are divided into.

[0084] Specifically, the number of Gaussian distribution subsets is determined according to the number of operations such as load redistribution or topology changes in the power system in a short period of time. For the Gaussian process samples in each sample history period, the probability density function in the Gaussian mixture model is used to calculate the posterior probability that the Gaussian process samples in the current sample history period belong to each Gaussian distribution subset, and then according to the obtained posterior probability, the Gaussian process samples in each sample history period are assigned to the Gaussian distribution subset that best matches it, that is, the Gaussian distribution subset with the highest posterior probability is selected as the clustering affiliation of the Gaussian process samples, and then the assigned Gaussian distribution subset is determined as the Gaussian process cluster sample, where the index of the subset is the cluster label of the sample.

[0085] Optionally, before clustering, the Gaussian process samples of each sample history period may be normalized and dimensionally reduced to reduce the amount of calculation and improve clustering efficiency. Then, clustering is performed on the pre-processed Gaussian process samples of each sample history period.

[0086] For example, in order to verify the correctness and effectiveness of this solution, the actual load data of the New York Independent System Operator (NYISO) in November 2022 was used as the experimental data set for a detailed simulation analysis. The following are the specific implementation steps and processes:

[0087] 1. Data preparation and simulation system establishment: First, the power measurement data of the node named "MILLWD" was extracted from the NYISO database. In order to simulate the fault situation in the actual scenario, it is assumed that the node fails at 0:00 on November 30, 2022.

[0088] 2. Historical data collection and preprocessing: In order to predict the node power after the failure, the historical data of the node from November 1 to November 29, 2022 was collected.

[0089] 3. 95% confidence interval prediction: Based on the collected historical data, a 95% confidence interval prediction was made for the power of the "MILLWD" node from 0:00 to 23:55 on November 30. Since NYISO collects data every 5 minutes, there are 288 collection moments in this period, and 288 power predictions were made for these moments.

[0090] 4. Time-varying analysis of fluctuation errors: In order to evaluate the accuracy and reliability of the prediction results, special attention is paid to the time-varying nature of fluctuation errors. By statistically analyzing the fluctuation errors at each acquisition moment, the trend and law of fluctuation errors changing over time can be observed, thereby verifying whether this solution can effectively handle time-varying fluctuation errors.

[0091] 5. Interval state estimation and data independence analysis: After each prediction is completed, interval state estimation is performed to evaluate the credibility and accuracy of the prediction results. Since each prediction is based on an independent data set, the measurement vector before each prediction is complete and observable. Although the data used in the 288 predictions are continuous in time, the data sets based on each prediction are independent and do not interfere with each other.

[0092] 6. Update of mean function and covariance function: In the prediction process, mean function and covariance function have an important impact on the accuracy of prediction results. Therefore, these two functions will be updated after each prediction to reflect the latest data characteristics and trends. This dynamic update method enables the prediction model to better adapt to the time-varying data environment and improve the prediction accuracy and reliability.

[0093] 7. The hyperparameter settings for Gaussian process regression are as follows:

[0094] Table 1 Hyperparameters of Gaussian process regression

[0095]

[0096] Experimental results show that this scheme can accurately predict the power of nodes with a 95% confidence interval and can effectively handle time-varying fluctuation errors, with high prediction accuracy and reliability.

[0097] 8. Indicator Definition

[0098] For interval prediction problems, it can be evaluated from two aspects: accuracy and credibility. Among them, the accuracy evaluation uses the mean relative error (MRE) as the core indicator; while the credibility evaluation comprehensively considers the interval coverage and the average width of the interval.

[0099] MRE is an accuracy evaluation index for interval prediction. The smaller the MRE, the higher the prediction accuracy. The specific calculation method is as follows:

[0100]

[0101] In the formula, represents the measured predicted value of a node at time k, z k It represents the true value of the measurement of a node at time k. T represents the time series prediction of the measurement of a node.

[0102] According to interval estimation theory, credibility evaluation indicators include interval coverage and interval average width. Interval coverage evaluates the credibility of interval estimation (interval prediction), while interval average width evaluates the uncertainty of interval estimation (interval prediction). In order to comprehensively consider interval coverage and interval average width, the definition is: true value average width as a simple method for credibility evaluation. The smaller the true value average width, the smaller the uncertainty of the interval estimation (interval prediction) and the higher the credibility. The calculation formula is as follows:

[0103]

[0104]

[0105]

[0106] In the formula, ξ(z k ) is a judgment function for whether the measured true value falls within the prediction interval; if it falls within the interval, it is 1, otherwise it is 0. k Indicates the true value of the measurement; Indicates the upper limit of the prediction interval; It represents the lower limit of the prediction interval; T represents the time series prediction number of a certain measurement.

[0107] 9. Experimental results

[0108] In order to analyze the impact of the time-varying nature of the fluctuation error on the prediction of the pseudo-measurement interval, the following steps are adopted:

[0109] (1) Cluster the Gaussian process samples and determine the Gaussian process sample set.

[0110] (2) The pseudo-measurement interval prediction model is trained using the Gaussian process sample set and the true measurement value, and the prediction interval is constructed based on the training results.

[0111] (3) Compare the prediction interval of the node to be predicted with the actual measured value and analyze the time-varying nature of the fluctuation error.

[0112] Figure 2 is a schematic diagram of an interval prediction result provided by the present invention, such as Figure 2 As shown in the figure, overall, the prediction interval has covered the measured true value as much as possible, which fully verifies the effectiveness of this scheme. The high consistency between the prediction interval and the true value shows that this scheme can provide accurate predictions in most time periods. As the number of predictions increases, the estimation of the mean and covariance function characteristics becomes more accurate. This is because over time, the model can accumulate more historical data and better capture the dynamic changes of the data. Therefore, the prediction interval will gradually cover the true value, and the prediction performance will also be improved.

[0113] Embodiment 2

[0114] Figure 3 This is a flowchart of a pseudo measurement interval prediction method based on the time-varying fluctuation error provided by the second embodiment of the present invention. This embodiment is applicable to the case of predicting the pseudo measurement interval. The method can be executed by a pseudo measurement interval prediction device based on the time-varying fluctuation error. The pseudo measurement interval prediction device based on the time-varying fluctuation error can be implemented in the form of hardware and / or software. The pseudo measurement interval prediction device based on the time-varying fluctuation error can be configured in an electronic device. Figure 3 As shown, the method includes:

[0115] S210: Acquire measurement data of multiple historical periods of a node to be predicted in the power system.

[0116] Specifically, in order to realize the prediction of the measurement data of the node to be predicted, it is necessary to obtain the measurement data of multiple historical periods of the node to be predicted in the power system. In this embodiment, the measurement data is the power of the node to be predicted at historical time. If the pseudo-measurement interval prediction model is trained by the load power of the node, the measurement data is the load power of the node to be predicted at historical time.

[0117] S220. For the measurement data of multiple historical periods, based on the difference between the measurement data of any two adjacent historical moments in the current historical period, determine the fluctuation error of the measurement data at each historical moment in the current historical period.

[0118] Specifically, the difference between the measurement data of any two adjacent historical moments in the current historical period is calculated, and the fluctuation error of the measurement data of each historical moment in the current historical period can be obtained based on the difference. For example, the difference between the measurement data at historical moment t1 and the measurement data at historical moment t1+1 in the current historical period is calculated: t1+1 -z t1 =v, and take the difference as the fluctuation error of the measurement data at the historical moment t1. By analogy, the fluctuation error of the measurement data at each historical moment in the current historical period can be obtained.

[0119] Then, by repeating the above process, the fluctuation error of the measured sample data at each historical moment in each sample historical period can be obtained.

[0120] S230 , determining a Gaussian process data set based on measurement data of multiple historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the multiple historical periods.

[0121] Specifically, in order to reflect the dynamic characteristics of the fluctuation error changing over time, the fluctuation errors of the measurement sample data at the same historical moment corresponding to multiple historical periods are combined to calculate the distribution characteristics of the fluctuation error at each historical moment. Then, based on the measurement data of multiple historical periods and the distribution characteristics of the fluctuation error at each historical moment, the Gaussian process sample set is determined.

[0122] S240: Input the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period.

[0123] The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method described in any embodiment of the present invention.

[0124] Specifically, by inputting the Gaussian process data set into the pseudo-measurement interval prediction model, the expectation and variance of the pseudo-measurement corresponding to the current period can be obtained, and then the pseudo-measurement confidence interval can be predicted based on the expectation and variance of the pseudo-measurement.

[0125] The technical solution provided in the second embodiment of the present invention determines the fluctuation error of the measurement data at each historical moment in the current historical period based on the difference between the measurement data at any two adjacent historical moments in the current historical period for the measurement data of multiple historical periods, and then determines the Gaussian process data set based on the measurement data of multiple historical periods and the fluctuation errors of the measurement data at the same historical moments corresponding to the multiple historical periods, so that the Gaussian process data set contains the distribution characteristics of the fluctuation errors of the measurement data and the laws of their changes over time, and then inputs the Gaussian process data set into the trained pseudo-measurement interval prediction model to obtain the pseudo-measurement interval corresponding to the current period, thereby effectively improving the accuracy and credibility of the pseudo-measurement interval prediction.

[0126] Embodiment 3

[0127] Figure 4 Schematic diagram of the structure of a pseudo-measurement interval prediction model training device provided by Embodiment 3 of the present invention. Figure 4 As shown, the device comprises:

[0128] The sample data acquisition module 31 is used to acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of a plurality of sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the plurality of sample history cycles;

[0129] A first error determination module 32 is used to determine the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period based on the difference between the measured sample data at any two adjacent sample historical moments in the current sample historical period for the measured sample data of the multiple sample historical periods;

[0130] A sample set determination module 33, configured to determine a Gaussian process sample set based on the measured sample data of the multiple sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods;

[0131] The model training module 34 is used to train the pseudo measurement interval prediction model based on the Gaussian process sample set and the true measurement value of the reference sample period, wherein the pseudo measurement interval prediction model is determined based on the mean function and the covariance function.

[0132] The technical solution provided by the third embodiment of the present invention effectively improves the accuracy and credibility of the pseudo-measurement interval prediction model in predicting the pseudo-measurement interval.

[0133] Optionally, the sample set determination module 33 includes:

[0134] A first determination submodule is used to determine the error expectation and error variance of multiple identical sample historical moments based on the fluctuation errors of the measured sample data at the identical sample historical moments corresponding to the multiple sample historical periods;

[0135] A second determination submodule is used to determine the Gaussian process samples of each sample history period based on the measured sample data of the multiple sample history periods and the error expectations and error variances of the multiple same sample history moments;

[0136] The third determination submodule is used to determine a Gaussian process sample set based on the Gaussian process samples of each sample history period.

[0137] Optionally, a second determination submodule is specifically used to calculate, for each sample history period in a plurality of sample history periods, the sample interval corresponding to the measured sample data at each sample history moment in the current sample history period based on the measured sample data of the current sample history period and the error expectation and error variance of the plurality of identical sample history moments, and determine the Gaussian process sample of the current sample history period based on the sample interval corresponding to the measured sample data at each sample history moment.

[0138] Optionally, the third determining submodule includes:

[0139] A sample determination unit, used for clustering Gaussian process samples of each sample history period to obtain Gaussian process cluster samples;

[0140] The sample set determination unit is used to determine a Gaussian process sample set based on the Gaussian process clustering samples.

[0141] Optionally, the sample determination unit includes:

[0142] A first determination subunit is used to determine the number of Gaussian distribution subsets;

[0143] A probability calculation subunit, used to calculate the posterior probability that the Gaussian process sample of each sample history period belongs to the Gaussian distribution subset;

[0144] The sample clustering subunit is used to cluster the Gaussian process samples of each sample history period based on the posterior probability that the Gaussian process samples of each sample history period belong to the Gaussian distribution subset to obtain Gaussian process cluster samples.

[0145] The pseudo-measurement interval prediction model training device provided in the embodiment of the present invention can execute the pseudo-measurement interval prediction model training method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0146] Embodiment 4

[0147] Figure 5 Schematic diagram of the structure of a pseudo-measurement interval prediction device based on the time-varying fluctuation error provided by the third embodiment of the present invention. Figure 5 As shown, the device comprises:

[0148] The data acquisition module 41 is used to acquire the measurement data of multiple historical periods of the node to be predicted in the power system;

[0149] The second error acquisition module 42 is used to determine the fluctuation error of the measurement data at each historical moment in the current historical period based on the difference between the measurement data at any two adjacent historical moments in the current historical period for the measurement data of multiple historical periods;

[0150] A data set determination module 43, configured to determine a Gaussian process data set based on measurement data of a plurality of historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the plurality of historical periods;

[0151] An interval prediction module 44, used for inputting the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period;

[0152] The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method described in any embodiment of the present invention.

[0153] The pseudo-measurement interval prediction device based on the time-varying fluctuation error provided in the embodiment of the present invention can execute the pseudo-measurement interval prediction method based on the time-varying fluctuation error provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0154] Embodiment 5

[0155] Figure 6 : is a structural schematic diagram of an electronic device provided in Embodiment 4 of the present invention. The electronic device may be an electronic device intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0156] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0157] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0158] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, such as a pseudo-measurement interval prediction model training method or a pseudo-measurement interval prediction method based on the time-varying nature of the fluctuation error.

[0159] In some embodiments, the pseudo-measurement interval prediction model training method or the pseudo-measurement interval prediction method based on the time-varying property of the fluctuation error may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the pseudo-measurement interval prediction model training method or the pseudo-measurement interval prediction method based on the time-varying property of the fluctuation error described above may be executed. Alternatively, in other embodiments, the processor 11 may be configured to execute the pseudo-measurement interval prediction model training method or the pseudo-measurement interval prediction method based on the time-varying property of the fluctuation error by any other appropriate means (e.g., by means of firmware).

[0160] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0161] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0162] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0163] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0164] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0165] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0166] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0167] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A pseudo-measurement interval prediction model training method, characterized in that: include: Acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of a plurality of sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the plurality of sample history cycles; For the measured sample data of the plurality of sample history periods, based on the difference between the measured sample data of any two adjacent sample history moments in the current sample history period, determine the fluctuation error of the measured sample data at each sample history moment in the current sample history period; Determine a Gaussian process sample set based on the measured sample data of the multiple sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods; Based on the Gaussian process sample set and the true measurement value of the reference sample period, a pseudo measurement interval prediction model is trained, wherein the pseudo measurement interval prediction model is determined based on a mean function and a covariance function.

2. The method according to claim 1, characterized in that: Determining a Gaussian process sample set based on the measured sample data of the multiple sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the multiple sample history periods includes: Determine the error expectation and error variance of multiple same sample historical moments based on the fluctuation errors of the measured sample data at the same sample historical moments corresponding to the multiple sample historical periods; Determine a Gaussian process sample for each sample history period based on the measured sample data of the multiple sample history periods and the error expectations and error variances of the multiple same sample history moments; Based on the Gaussian process samples of each sample history period, a Gaussian process sample set is determined.

3. The method according to claim 2, characterized in that Based on the measured sample data of the multiple sample history periods and the error expectations and error variances of the multiple same sample history moments, a Gaussian process sample of each sample history period is determined, including: For each sample history period in multiple sample history periods, the sample interval corresponding to the measured sample data at each sample history moment in the current sample history period is calculated based on the measured sample data of the current sample history period and the error expectation and error variance of the multiple same sample history moments, and the Gaussian process sample of the current sample history period is determined based on the sample interval corresponding to the measured sample data at each sample history moment.

4. The method according to claim 2, characterized in that: Based on the Gaussian process samples of each sample history period, a Gaussian process sample set is determined, including: Cluster the Gaussian process samples of each sample history period to obtain Gaussian process cluster samples; Based on the Gaussian process clustering samples, a Gaussian process sample set is determined.

5. The method according to claim 4, clustering the Gaussian process samples of each sample history period to obtain Gaussian process cluster samples, comprising: Determine the number of Gaussian distribution subsets; Calculate the posterior probability that the Gaussian process sample of each sample history period belongs to the Gaussian distribution subset; Based on the posterior probability that the Gaussian process samples of each sample history period belong to the Gaussian distribution subset, the Gaussian process samples of each sample history period are clustered to obtain Gaussian process cluster samples.

6. A pseudo-measurement interval prediction method based on the time-varying nature of fluctuation error, characterized in that: include: Obtain measurement data of multiple historical periods of the nodes to be predicted in the power system; For the measurement data of multiple historical periods, based on the difference between the measurement data of any two adjacent historical moments in the current historical period, determine the fluctuation error of the measurement data at each historical moment in the current historical period; Determine a Gaussian process data set based on measurement data of multiple historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the multiple historical periods; Inputting the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period; The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method according to any one of claims 1 to 5.

7. A pseudo-measurement interval prediction model training device, characterized in that: include: A sample data acquisition module, used to acquire sample data of a node to be predicted in the power system, wherein the sample data includes measured sample data of a plurality of sample history cycles and measured true values ​​of a reference sample cycle, and the reference sample cycle is located after the plurality of sample history cycles; A first error determination module is used to determine the fluctuation error of the measured sample data at each sample historical moment in the current sample historical period based on the difference between the measured sample data at any two adjacent sample historical moments in the current sample historical period for the measured sample data of the multiple sample historical periods; A sample set determination module, configured to determine a Gaussian process sample set based on the measured sample data of the plurality of sample history periods and the fluctuation errors of the measured sample data at the same sample history moments corresponding to the plurality of sample history periods; A model training module is used to train a pseudo-measurement interval prediction model based on the Gaussian process sample set and the true measurement value of the reference sample period, wherein the pseudo-measurement interval prediction model is determined based on a mean function and a covariance function.

8. A pseudo-measurement interval prediction device based on the time-varying nature of fluctuation error, characterized in that: include: A data acquisition module, used to acquire measurement data of multiple historical periods of a node to be predicted in the power system; A second error acquisition module is used to determine the fluctuation error of the measurement data at each historical moment in the current historical period based on the difference between the measurement data at any two adjacent historical moments in the current historical period for the measurement data of multiple historical periods; A data set determination module, configured to determine a Gaussian process data set based on measurement data of a plurality of historical periods and fluctuation errors of measurement data at the same historical moments corresponding to the plurality of historical periods; An interval prediction module, used for inputting the Gaussian process data set into a pseudo measurement interval prediction model to obtain a pseudo measurement interval corresponding to a current period; The pseudo-measurement interval prediction model is trained based on the pseudo-measurement interval prediction model training method according to any one of claims 1 to 5.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the pseudo-measurement interval prediction model training method as described in any one of claims 1-5 or execute the pseudo-measurement interval prediction method based on the time-varying fluctuation error as described in claim 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it implements the pseudo-measurement interval prediction model training method as described in any one of claims 1 to 5 or the pseudo-measurement interval prediction method based on the time-varying fluctuation error as described in claim 6.

Images