Metering electric energy data learning practice dataset generation method

Abstract

The application discloses a metering electric energy data learning practice dataset generation method, including the following steps: running a power system metering simulation model based on a constraint condition model; obtaining power data information and performing feature classification processing; performing label classification processing on the feature classification processing result; performing regular screening processing on the label processing result to determine a metering electric energy dataset; and adopting an electric energy data regular screening technology based on the metering electric energy data to form an accurate and standard metering electric energy data learning practice dataset.

Description

Technical Field

[0001] This invention relates to the field of electricity metering technology, and in particular to a method for generating a learning practice dataset of metered electricity energy data. Background Technology

[0002] A significant characteristic of the new power system is its digital empowerment. Digital technology is increasingly becoming a basic skill, and power system metering professionals have an increasingly urgent need to master metering energy data mining and artificial intelligence technologies such as machine learning. However, there is a significant lack of datasets derived from production practice and closely aligned with power work scenarios, severely hindering power system employees from mastering advanced digital technologies. This disconnect from production practice and the failure to fully reflect the impact of various parameters, time, environmental, and other conditions in production practice are problematic. Summary of the Invention

[0003] To overcome the shortcomings of existing technologies, this invention provides a method for generating a learning practice dataset of metered electrical energy data, which realizes the formation of an accurate and standardized learning practice dataset of metered electrical energy data based on the regular filtering technology of electrical energy data.

[0004] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0005] The method for generating a dataset for learning from metered electrical energy data includes the following steps:

[0006] S101. Run the power system metering simulation model based on the constraint condition model;

[0007] S102. Acquire power data information and perform feature classification processing;

[0008] S103. Perform labeling and classification processing on the feature classification results;

[0009] S104. Perform regular expression filtering on the labeling results to determine the metered electrical energy dataset.

[0010] Furthermore, the constraint model includes equipment parameter factors, equipment operating time factors, and environmental temperature and humidity factors.

[0011] Furthermore, the power data information includes: phase A voltage, phase B voltage, phase C voltage, phase A current, phase B current, phase C current, total active power, phase A active power, phase B active power, phase C active power, total reactive power, phase A reactive power, phase B reactive power, phase C reactive power, total power factor, phase A power factor, phase B power factor, phase C power factor, UAB magnitude, UAB phase angle, UBC magnitude, UBC phase angle, UCA magnitude, UCA phase angle, UCB magnitude, UCB phase angle, etc.

[0012] Further feature classification processing includes:

[0013] Voltage feature classification processing involves identifying features in power data information. If voltage features are identified, the data is classified as a voltage feature classification set.

[0014] Current feature classification processing involves identifying features in power data information. If current features are identified, the data is classified as a current feature classification set.

[0015] Active power feature classification processing involves identifying features in power data information. If active power features are identified, the data is determined to be an active power feature classification set.

[0016] Reactive power feature classification processing involves identifying features in power data information. If reactive power features are identified, the data is determined to be a reactive power feature classification set.

[0017] Power factor feature classification processing involves identifying features in power data information. If power factor features are identified, the data is classified as a power factor feature classification set.

[0018] Phase angle feature classification processing is used to identify features in power data information. If the identified features are phase angle features, it is determined to be a phase angle feature classification set.

[0019] Furthermore, the feature classification results include: voltage feature classification set, current feature classification set, active power feature classification set, reactive power feature classification set, power factor feature classification set, and phase angle feature classification set.

[0020] Further, the labeling and classification process includes:

[0021] The feature classification results are classified into difference level labels, which include positive difference level and negative difference level. The positive difference level includes the first positive difference level, the second positive difference level and the third positive difference level. The negative difference level includes the first negative difference level, the second negative difference level and the third negative difference level.

[0022] Furthermore, the tagging results include subsets of voltage difference levels, current difference levels, active power difference levels, reactive power difference levels, power factor difference levels, and phase angle difference levels.

[0023] Furthermore, the regular expression filtering process includes: using the first positive threshold level subset and the first negative threshold level subset as the first-level metered energy dataset, using the second positive threshold level subset and the second negative threshold level subset as the second-level metered energy dataset, and using the third positive threshold level subset and the third negative threshold level subset as the third-level metered energy dataset.

[0024] The beneficial effects of this invention are as follows: The method for generating a learning practice dataset of metered electrical energy data enables the running of a power system metering simulation model based on a constraint condition model; it acquires power data information and performs feature classification processing; it then performs label classification processing on the feature classification results; and finally, it performs regular expression filtering processing on the label processing results to determine the metered electrical energy dataset, thereby deepening the understanding of the combination of digital technology and power grid business application scenarios.

[0025] Based on the actual business application scenarios of the power grid and the specific requirements of digital technology, the power energy data subset technology is applied to generate a data model that fits the business application scenario from the full set of power energy data attributes.

[0026] The electrical energy data source extracts massive samples and high-dimensional feature data, and uses data labeling technology to classify the original data.

[0027] Based on metered electrical energy data, a regularized screening technique for electrical energy data is used to form an accurate and standardized dataset for machine learning practice in metered electrical energy data mining. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram illustrating the steps of the method for generating a learning practice dataset of metered electrical energy data according to the present invention. Detailed Implementation

[0030] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0031] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0032] Example 1:

[0033] The method for generating a dataset for learning from metered electrical energy data includes the following steps:

[0034] S101. Running the power system metering simulation model based on the constraint condition model.

[0035] The power system metering simulation model is run based on the constraint model, which includes equipment parameter factors, equipment operating time factors, and environmental temperature and humidity factors.

[0036] S102. Acquire power data information and perform feature classification processing;

[0037] The power system metering simulation model is operated based on the constraint condition model to obtain power data information, and feature classification processing is performed. The power data information includes: phase A voltage, phase B voltage, phase C voltage, phase A current, phase B current, phase C current, total active power, phase A active power, phase B active power, phase C active power efficiency, total reactive power, phase A reactive power, phase B reactive power, phase C reactive power, total power factor, phase A power factor, phase B power factor, phase C power factor, UAB magnitude, UAB phase angle, UBC magnitude, UBC phase angle, UCA magnitude, UCA phase angle, UCB magnitude, UCB phase angle, etc.

[0038] The power data information undergoes feature classification processing, which includes:

[0039] Voltage feature classification processing involves identifying features in power data information. If voltage features are identified, the data is classified as a voltage feature classification set.

[0040] Current feature classification processing involves identifying features in power data information. If current features are identified, the data is classified as a current feature classification set.

[0041] Active power feature classification processing involves identifying features in power data information. If active power features are identified, the data is determined to be an active power feature classification set.

[0042] Reactive power feature classification processing involves identifying features in power data information. If reactive power features are identified, the data is determined to be a reactive power feature classification set.

[0043] Power factor feature classification processing involves identifying features in power data information. If power factor features are identified, the data is classified as a power factor feature classification set.

[0044] Phase angle feature classification processing is used to identify features in power data information. If the identified features are phase angle features, it is determined to be a phase angle feature classification set.

[0045] S103. Perform labeling and classification processing on the feature classification results;

[0046] The feature classification results are then labeled and categorized. These results include: voltage feature classification sets, current feature classification sets, active power feature classification sets, reactive power feature classification sets, power factor feature classification sets, and phase angle feature classification sets. The labeling and categorization process includes:

[0047] The feature classification results are classified into difference level labels, which include positive difference level and negative difference level. The positive difference level includes the first positive difference level, the second positive difference level and the third positive difference level. The negative difference level includes the first negative difference level, the second negative difference level and the third negative difference level.

[0048] For example, the voltage feature classification set is classified by difference level labeling. The difference between each voltage in the voltage feature classification set and the voltage threshold is calculated, and the voltage difference level is determined based on the difference and saved to the voltage difference level subset.

[0049] For example, the current feature classification set is classified by difference level labeling. The difference between each current in the current feature classification set and the current threshold is calculated, and the current difference level is determined based on the difference and saved to the current difference level subset.

[0050] For example, the active power feature classification set is classified by difference level labeling. The difference between each active power in the active power feature classification set and the active power threshold is calculated, and the active power difference level is determined based on the difference and saved to the active power difference level subset.

[0051] For example, the reactive power feature classification set is classified by difference level labeling. The difference between each reactive power in the reactive power feature classification set and the reactive power threshold is calculated, and the reactive power difference level is determined based on the difference and saved to the reactive power difference level subset.

[0052] For example, the power factor feature classification set is classified by difference level labeling. The difference between each power factor in the power factor feature classification set and the power factor threshold is calculated, and the power factor difference level is determined based on the difference and saved to the power factor difference level subset.

[0053] For example, the phase angle feature classification set is classified by difference level labeling. The difference between each phase angle in the phase angle feature classification set and the phase angle threshold is calculated, and the phase angle difference level is determined based on the difference and saved to the phase angle difference level subset.

[0054] S104. Perform regular expression filtering on the labeling results to determine the metered electrical energy dataset;

[0055] The tagged results are subjected to regular expression filtering. The tagged results include subsets of voltage difference level, current difference level, active power difference level, reactive power difference level, power factor difference level, and phase angle difference level.

[0056] The voltage difference level subset includes a first positive voltage difference level, a second positive voltage difference level, a third positive voltage difference level, a first negative voltage difference level, a second negative voltage difference level, and a third negative voltage difference level;

[0057] The current difference level subset includes the first forward current difference level, the second forward current difference level, the third forward current difference level, the first reverse current difference level, the second reverse current difference level, and the third reverse current difference level;

[0058] The active power difference level subset includes the first active power forward difference level, the second active power forward difference level, the third active power forward difference level, the first active power reverse difference level, the second active power reverse difference level, and the third active power reverse difference level;

[0059] The reactive power difference level subset includes the first reactive power forward difference level, the second reactive power forward difference level, the third reactive power forward difference level, the first reactive power reverse difference level, the second reactive power reverse difference level, and the third reactive power reverse difference level;

[0060] The power factor difference level subset includes the first power factor positive difference level, the second power factor positive difference level, the third power factor positive difference level, the first power factor negative difference level, the second power factor negative difference level, and the third power factor negative difference level;

[0061] The subset of phase angle difference levels includes the first positive phase angle difference level, the second positive phase angle difference level, the third positive phase angle difference level, the first negative phase angle difference level, the second negative phase angle difference level, and the third negative phase angle difference level.

[0062] The labeled results are then subjected to regular expression filtering to determine the metered energy dataset. This filtering process includes: selecting the first positive threshold level subset and the first negative threshold level subset as the first-level metered energy dataset; selecting the second positive threshold level subset and the second negative threshold level subset as the second-level metered energy dataset; and selecting the third positive threshold level subset and the third negative threshold level subset as the tertiary-level metered energy dataset. It should be noted that the first-level, second-level, and third-level metered energy datasets are used for metering energy, and their respective weights can be set according to actual circumstances.

[0063] In the description of this invention, it should be noted that the terms "first," "second," and "third" are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0064] The above description is merely illustrative of the embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, any modifications, equivalent substitutions, improvements, etc., made without creative effort within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for generating a learning practice dataset of metered electrical energy data, characterized in that, Includes the following steps: S101. Run the power system metering simulation model based on the constraint condition model; S102. Acquire power data information and perform feature classification processing; S103. Perform labeling and classification processing on the feature classification results; S104. Perform regular expression filtering on the labeling results to determine the metered electrical energy dataset; The labeling and classification process includes: performing difference level labeling and classification processing on the feature classification results, whereby the difference levels include positive difference level and negative difference level; the positive difference level includes a first positive difference level, a second positive difference level, and a third positive difference level; and the negative difference level includes a first negative difference level, a second negative difference level, and a third negative difference level. The regularization filtering process includes: using the first positive threshold level subset and the first negative threshold level subset as a first-level metered energy dataset, using the second positive threshold level subset and the second negative threshold level subset as a second-level metered energy dataset, and using the third positive threshold level subset and the third negative threshold level subset as a third-level metered energy dataset.

2. The method for generating a learning practice dataset of metered electrical energy data according to claim 1, characterized in that, The constraint model includes equipment parameter factors, equipment operating time factors, and environmental temperature and humidity factors.

3. The method for generating a learning practice dataset of metered electrical energy data according to claim 1, characterized in that, The feature classification results include: voltage feature classification set, current feature classification set, active power feature classification set, reactive power feature classification set, power factor feature classification set, and phase angle feature classification set.

4. The method for generating a learning practice dataset of metered electrical energy data according to claim 1, characterized in that, The tagging results include subsets of voltage difference levels, current difference levels, active power difference levels, reactive power difference levels, power factor difference levels, and phase angle difference levels.

5. The method for generating a learning practice dataset of metered electrical energy data according to claim 1, characterized in that, The feature classification process includes: Voltage feature classification processing involves identifying features in power data information. If voltage features are identified, the data is classified as a voltage feature classification set. Current feature classification processing involves identifying features in power data information. If current features are identified, the data is classified as a current feature classification set. Active power feature classification processing involves identifying features in power data information. If active power features are identified, the data is determined to be an active power feature classification set. Reactive power feature classification processing involves identifying features in power data information. If reactive power features are identified, the data is determined to be a reactive power feature classification set. Power factor feature classification processing involves identifying features in power data information. If power factor features are identified, the data is classified as a power factor feature classification set. Phase angle feature classification processing is used to identify features in power data information. If the identified features are phase angle features, it is determined to be a phase angle feature classification set.

Images