CN-116108979-B - Cross-period multi-source heterogeneous power data processing system and application method

Abstract

A multi-source heterogeneous power data processing system with a cycle comprises a data acquisition module, a data analysis module, a data modeling module, a multi-source heterogeneous information module, a prediction module, a switching module and a data pushing module, wherein the data acquisition module, the data analysis module, the data modeling module, the multi-source heterogeneous information module, the prediction module, the switching module and the data pushing module are application software installed in a PC, and the application method of the multi-source heterogeneous power data processing system with the cycle comprises seven steps. The invention can obtain the electricity load data model of the electricity consumption area, can predict electricity consumption according to the obtained model, can be adjusted in advance before the electricity consumption power of the area changes, ensures the stable on-site power supply, can prompt the power supply end and the electricity consumption end in time when the electricity consumption data is abnormal, reduces the probability of electricity consumption accidents, can give reasonable electricity saving suggestions to users, and can actively prompt the electricity consumption party to overhaul when the electricity consumption data and the obvious deviation occur in parallel.

Inventors

• KANG LI
• GUO YIHUA
• XU YANJUN
• YAO WENYING

Assignees

• 上海恒能泰企业管理有限公司璞能电力科技工程分公司

Dates

Publication Date

20260505

Application Date

20230103

Claims (5)

1. 1. The application method of the cross-period multi-source heterogeneous power data processing system is characterized by comprising a data acquisition module, a data analysis module, a data modeling module, a multi-source heterogeneous information module, a prediction module, a switching module and a data pushing module; the data acquisition module acquires power utilization area data, outputs the data after preliminary classification processing to the data analysis module, the step B is based on an artificial intelligence technology, performs calculation and analysis on the data acquired by the data acquisition module by using one or more of weak supervision/semi supervision/non-supervision machine learning methods, performs calculation and analysis on the data acquired by the data acquisition module based on a power utilization time sequence behavior mode of a user, and is used for constructing a time sequence behavior-based user classification algorithm model, wherein specific used user data comprises 96-point curve, daily electric quantity and month electric quantity data, the step C is used for performing characteristic attribution analysis on the data classification algorithm result obtained in the step B, and extracting corresponding characteristics of a user power utilization face according to the classification algorithm model of weak supervision learning by developing the attribution analysis on the classification result, and further performing calculation and analysis on the service result of the user classification and the prediction result based on the power utilization time sequence behavior mode of the user, the step C is used for constructing a multi-source heterogeneous electric quantity data model based on the data of the time sequence behavior, and the data of the step C is used for constructing the multi-source electric quantity and the multi-heterogeneous electric quantity cycle prediction model of the data of the power utilization time sequence behavior model, the method also comprises meteorological, economic, policy and electricity consumption change flow factors, wherein the change flow comprises capacity increase and decrease, class change and family pass of an electricity consumption area, and the data obtained in the step serves the electricity consumption prediction of a user; the method comprises the steps of E, calculating and establishing a time sequence learning model for cross-period accurate load prediction based on multi-source heterogeneous characteristic characterization by a prediction module, wherein model data comprise 96-point curves for predicting a whole month and 96-point curves for predicting a T+2 day in advance of a power consumption area for one year, one quarter and one month, F, pushing prompt data for a power consumer and a management party when the user load of the power consumption area is abnormal by a data pushing module, and giving specific abnormal types, G, adjusting the power supply of a power supply end to the power supply area at a power consumption area peak or low peak time period according to power consumption prediction data of the power consumption area by a switching module, and in the step F, judging that the power consumption area user load abnormal data comprise excessive current abnormality, excessive voltage abnormality and excessive temperature, and particularly, pushing information when the data exceeds a threshold value by a database sub-module of the data pushing module, wherein the switching module meets the requirements of the power consumption area on the power supply end to the power supply area according to the power consumption area when the power consumption end is adjusted at the peak or low peak time period, and the power consumption area is required to be adjusted.
2. 2. The method for applying a cross-cycle heterogeneous power data processing system according to claim 1, wherein in the step a, the data acquisition module acquires data of a power utilization area, including power utilization data in unit time, current and voltage data in each time period, and on-site temperature data obtained by a meter, an ammeter, a voltmeter and a thermometer.
3. 3. The method for applying a cross-cycle heterogeneous power data processing system according to claim 1, wherein in step C, the power and load prediction data obtained by the classification of the user can be displayed through a display interface and output to the switching module.
4. 4. The method for applying the cross-period multi-source heterogeneous power data processing system according to claim 1, wherein the data pushing module can give out reasonable power saving suggestions according to the power consumption data, so as to achieve the purposes of power saving and emission reduction.
5. 5. The application method of the cross-period multi-source heterogeneous power data processing system is characterized in that in the step E, the method comprises the steps of determining a unified characteristic characterization method of multi-source heterogeneous information and determining a time sequence learning model for cross-period accurate load prediction, the method comprises the steps of constructing a basic model by adopting a random forest algorithm and completing multi-source heterogeneous data fusion based on model integration, and the time sequence learning model for cross-period accurate load prediction is a cross-period accurate load prediction model established on the basis of the characteristic characterization of the multi-source heterogeneous information and mainly utilizes a gray prediction model to complete load prediction.

Description

Cross-period multi-source heterogeneous power data processing system and application method Technical Field The invention relates to the technical field of power, in particular to a cross-period multi-source heterogeneous power data processing system and an application method. Background Along with the development of power supply technology, in order to ensure safe power supply, a power department or a user of a related power utilization terminal can install monitoring equipment such as temperature, humidity, voltage, current and the like in a corresponding area besides an electric energy meter and the like so as to monitor the working performance of electric equipment in the corresponding area and whether power supply is normal or not, and ensure safe power supply as far as possible. Although the existing related electricity consumption monitoring equipment realizes intelligent power supply management to a certain extent, certain defects still exist due to the limitation of the structure, and the intelligent power supply monitoring equipment is specifically shown as follows. The power utilization method comprises the steps that firstly, various monitored data can only reflect the current power utilization data condition, a power utilization party and a power supply party can only adjust the power utilization mode according to current data, for example, a power utilization peak appears in a certain area in a time period, so that power supply voltage is reduced, then the power supply party can allocate sufficient power for the area to meet the working requirement of electric equipment, or the power utilization party reduces some loads at the moment to meet the requirement that important electric equipment can work normally, and the mode is characterized in that no matter the power utilization party is in a passive regulation mode, no matter the power utilization party cannot be prompted in advance, accordingly, the specific corresponding treatment work is not brought, the timeliness of the corresponding treatment method cannot be achieved, and the safety power supply is not affected. Secondly, the power utilization party cannot be prompted when power utilization accidents (such as short circuits, open circuits and the like) possibly occur in the power utilization area, and in normal cases, the power utilization party can know the situation in time after observing that the corresponding monitoring equipment has abnormal data, so that a targeted treatment plan cannot be made in advance, and adverse effects are caused on safe power supply. In summary, it is particularly necessary to provide a system and an application method capable of performing analysis modeling based on daily electricity consumption behavior of a user, and ensuring safe and stable power supply to the user as much as possible. Disclosure of Invention In order to overcome the defects in the prior art, due to the defects of the technology, the invention provides a multi-source heterogeneous power data processing system and an application method which can uniformly calculate, analyze and model all data under the combined action of corresponding module units, automatically adaptively adjust power loads in different time periods of a related area through the modeled data, ensure stable on-site power supply, provide reasonable power saving suggestions for users based on the obtained data, actively prompt a power consumer to overhaul when the power supply data has obvious deviation in parallel (failure probability), and effectively ensure safe power supply. The technical scheme adopted for solving the technical problems is as follows: The cross-period multi-source heterogeneous power data processing system is characterized by comprising a data acquisition module, a data analysis module, a data modeling module, a multi-source heterogeneous information module, a prediction module, a switching module and a data pushing module; the data acquisition module acquires power utilization area data, outputs the data after preliminary classification processing to the data analysis module, the data analysis module carries out calculation analysis on data acquired by the data acquisition module based on a data-driven user classification algorithm by using one or more of weak supervision/semi-supervision/non-supervision machine learning methods based on an artificial intelligence technology, carries out calculation analysis on the data acquired by the data acquisition module based on a power utilization time sequence behavior mode of a user, is used for constructing a time sequence behavior-based user classification algorithm model, comprises 96-point curves, daily electric quantity, month data and the like, the data modeling module carries out characteristic attribution analysis on the data classification algorithm result obtained in the step B, and carries out attribution analysis on the classification result according to the weak supervision/semi-supervision ma