Power grid regulation system based on big data

Abstract

The invention discloses a power grid regulation system based on big data, relates to the technical field of power systems, and solves the problem that it is difficult to analyze acquired data through various technical methods and extract corresponding features to obtain feature data. The system comprises the following modules: a power grid information acquisition module, which is used for acquiring load storage data of a source grid; the power grid constraint condition construction module is used for constructing a power output limitation constraint condition; the power grid big data analysis module is used for extracting a power consumption deviation degree predicted value, a comprehensive power floating degree and energy storage equipment charging and discharging fluctuation degree data through various technical methods; the power prediction model module is used for obtaining a power grid state comprehensive evaluation value according to the extracted data and the power value; and the power grid early warning adjustment module is used for comparing the power grid state comprehensive evaluation value with an early warning threshold value.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and in particular relates to a power grid regulation system based on big data. Background Art

[0002] With the continuous development of technologies such as the Internet of Things and cloud computing, big data analysis technology has been widely used in various fields. As one of the important infrastructures for national economic development, the power system has gradually introduced big data analysis technology, which can obtain power grid information and analyze data in real time, realize modeling prediction and high-level early warning. For this reason, a power grid regulation system based on big data has come into being. Through the acquired source, grid, load and storage data, the corresponding feature data is analyzed and extracted through a variety of technical methods. According to the Nyquist sampling theorem, the power output limit constraint condition is constructed, and the feature data and the power value that determines whether the constraint condition is met are combined to obtain the comprehensive evaluation value of the power grid state, and the comprehensive evaluation value of the power grid state is compared with the early warning threshold, realizing real-time monitoring, intelligent evaluation and early warning of the power grid operation status.

[0003] Although the existing power regulation system has achieved power regulation to a certain extent, the existing power regulation often relies on manual experience and simple data analysis, and it is difficult to cope with the complex and multi-piece power grid operation status. It is difficult to analyze the acquired source, grid, load and storage data through a variety of technical methods and extract the corresponding feature data. It is also difficult to construct power output limit constraints based on the Nyquist sampling theorem, and combine the feature data with the power value to judge whether the constraint conditions are met to obtain a comprehensive evaluation value of the power grid status. There is a lack of comparison between the comprehensive evaluation value of the power grid status and the warning threshold, and analysis of the comparison results. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a power grid regulation system based on big data, which is used to solve the following technical problems:

[0005] Although the existing power regulation system has achieved power regulation to a certain extent, the existing power regulation often relies on manual experience and simple data analysis, and it is difficult to cope with the complex and multi-piece power grid operation status. It is difficult to analyze the acquired source, grid, load and storage data through a variety of technical methods and extract the corresponding feature data. It is also difficult to construct power output limit constraints based on the Nyquist sampling theorem, and combine the feature data with the power value to judge whether the constraint conditions are met to obtain a comprehensive evaluation value of the power grid status. There is a lack of comparison between the comprehensive evaluation value of the power grid status and the warning threshold, and analysis of the comparison results.

[0006] To solve the above problems, the first aspect of the present invention provides a power grid regulation system based on big data, comprising the following modules:

[0007] Grid information acquisition module: obtains the power supply side, grid side, load side and energy storage measurement data in the grid information, pre-processes the data, and uploads the pre-processed data to the cloud storage service;

[0008] Grid constraint condition building module: The sampling frequency is set by the Nyquist sampling theorem, and the power output limit constraint condition is built according to the signal and the number of sampling points to obtain the maximum and minimum power values;

[0009] Power grid big data analysis module: Receives data collected by the power grid information acquisition module, and uses the time domain analysis method to extract features from the data. A model is built based on the periodic power consumption deviation and seasonal power consumption deviation to obtain the power consumption deviation prediction value. Different weights are assigned to the voltage fluctuation, current fluctuation and frequency fluctuation to obtain the comprehensive power fluctuation. The sliding window technology is used to extract the number of charge and discharge times and the peak and valley values ​​extracted from the hourly data to obtain the charging and discharging fluctuation of the energy storage device.

[0010] Power forecast model module: By assigning corresponding weights to the power consumption deviation prediction value, the comprehensive power fluctuation degree and the energy storage equipment charging and discharging fluctuation degree data, weighted addition is performed, and the power value is combined to obtain the comprehensive evaluation value of the power grid status;

[0011] Power grid early warning and adjustment module: compares the comprehensive evaluation value of the power grid status with the early warning threshold and performs analysis.

[0012] As a further solution of the present invention, the acquisition of power supply side, grid side, load side and energy storage measurement data in the grid information, preprocessing the data, and uploading the preprocessed data to a cloud storage service includes the following steps:

[0013] On the power supply side, power generation data, including power generation and power generation time, are collected in real time by connecting to the power plant's automation system interface. On the grid side, grid operation data, including voltage, current and frequency, are collected by intelligent sensor devices at each node of the grid. On the load side, user electricity consumption data, including power consumption during peak, normal and off-peak hours and total daily, weekly and monthly electricity consumption, are collected by deploying smart meters. For energy storage measurement, the operation data of energy storage equipment is obtained through the energy storage management system, including the number of battery charges, discharges, charge and discharge amounts. The collected data is pre-processed, including removing outliers and filling in missing data, and converting data from different sources into a unified format. The processed data is uploaded to the cloud storage service, and the grid big data analysis module downloads the data from the cloud for analysis.

[0014] As a further solution of the present invention, the sampling frequency is set by the Nyquist sampling theorem, and the power output limit constraint condition is constructed according to the signal and the number of sampling points to obtain the maximum and minimum power values, which includes the following steps:

[0015] By sampling the signal, setting the sampling frequency according to the Nyquist sampling theorem, using sampling software to discretize the analog signal to obtain digital sample points, and calculating the square sum of the signal divided by the number of sampling points to obtain the power between each data point;

[0016] Power calculation formula:

[0017]

[0018] in, is the power corresponding to the sample point, For each sample point, is the index of the sample point, is the total number of sampling points;

[0019] Traverse all the calculated power values ​​and find the maximum power value As the maximum power in signal processing, the minimum power value found As the minimum power in signal processing;

[0020] The power output limit constraint is constructed by obtaining the maximum power value and the minimum power value:

[0021]

[0022] in, is the minimum power value, is the power value at a certain time point, is the maximum power value.

[0023] As a further solution of the present invention, data collected by the power grid information acquisition module is received, and features are extracted from the data using a time domain analysis method, and a model is constructed by the periodic power consumption deviation degree and the seasonal power consumption deviation degree to obtain a power consumption deviation degree prediction value, including the following steps:

[0024] Obtain the raw data of electricity consumption during peak hours, normal hours, and off-peak hours, as well as the total electricity consumption per day, week, and month, assign a timestamp to each raw data, sort the data points by timestamp, convert the raw data into a time series format, extract the key features of the time series through time domain analysis methods, including periodicity and seasonal fluctuations, identify the periodic components in the data by drawing a time series graph, and for each identified period, calculate the average electricity consumption within each period;

[0025] The calculation formula of average power consumption in a cycle is:

[0026]

[0027] in, is the average power consumption during the period, For the Daily electricity consumption, is the total number of days;

[0028] The degree of deviation of cycle power consumption is obtained by the ratio of power consumption on one day in the cycle to the average power consumption of the cycle;

[0029] The calculation formula of the deviation degree of cycle power consumption is:

[0030]

[0031] in, is the degree of deviation of cycle power consumption, For the Daily electricity consumption, is the average power consumption during the period;

[0032] The seasonal electricity consumption deviation is obtained by the ratio of the average electricity consumption in each season to the average electricity consumption throughout the year;

[0033] Calculation formula for seasonal electricity consumption deviation:

[0034]

[0035] in, is the seasonal electricity consumption deviation degree, when =1, is the average electricity consumption in spring, is the deviation degree of electricity consumption in spring. =2, is the average electricity consumption in summer, is the deviation degree of summer electricity consumption. =3, is the average electricity consumption in autumn, is the deviation degree of electricity consumption in autumn. =4, is the average electricity consumption in winter, is the degree of deviation of winter electricity consumption, is the average annual electricity consumption;

[0036] Historical data including the deviation degree of periodic electricity consumption and seasonal electricity consumption are collected. After preprocessing the data, the ARIMAX model is fitted using the historical data to obtain the preliminary form of the electricity consumption deviation model. The model is verified by white noise test, and the fitting and verification process is repeated to obtain a trained model. The trained ARIMAX model is used to predict new input data to obtain the predicted value of the electricity consumption deviation degree.

[0037] As a further solution of the present invention, the voltage floating degree, the current floating degree and the frequency floating degree are assigned different weights to obtain the comprehensive power floating degree by combining them, which includes the following steps:

[0038] By collecting grid operation data, including voltage, current and frequency data, through intelligent sensor devices at each node of the grid, statistical features, including extreme values ​​and average values, are extracted from the acquired data through descriptive statistical analysis. The voltage fluctuation degree, current fluctuation degree and frequency fluctuation degree are obtained based on the statistical features and real-time data. An expert team in the power field is organized to evaluate the relative importance of voltage, current and frequency in grid stability based on the experts' experience in power system operation, detection and analysis, and determine the weight coefficients corresponding to the three fluctuation degrees.

[0039] The calculation formula of comprehensive power fluctuation degree is:

[0040] in, The comprehensive power fluctuation degree, is the number of measurements, For the The voltage value obtained by the measurement is is the maximum voltage value, is the minimum voltage value, For the The current value obtained by the measurement is is the maximum current value, is the minimum current value, For the The frequency value obtained by the measurement is is the maximum frequency value, is the minimum frequency value, , and are the weight coefficients corresponding to voltage, current and frequency.

[0041] As a further solution of the present invention, the use of sliding window technology to extract the number of charge and discharge times and the peak and valley values ​​extracted from hourly data to obtain the degree of charge and discharge fluctuation of the energy storage device includes the following steps:

[0042] Collect data of energy storage devices on different time scales in real time from the energy storage management system, including key parameters of the energy storage devices, such as the number of charges, discharges, charge capacity, and discharge capacity. After cleaning the collected data, extract time-related features, extract daily peak and valley features from hourly data, and use sliding window technology to extract the average value of the number of charges and discharges in a continuous time period and combine it with the peak and valley features extracted from hourly data to obtain the degree of fluctuation in the charge and discharge of energy storage devices.

[0043] The calculation formula for the fluctuation degree of charging and discharging of energy storage equipment is:

[0044]

[0045] in, is the charging and discharging fluctuation degree of the energy storage device. For the The peak value of electricity consumption per day, For the Peak electricity consumption per day, is the total number of days, For the Number of charges per day, For the The number of discharges per day.

[0046] As a further solution of the present invention, the method of assigning corresponding weights to the power consumption deviation degree prediction value, the comprehensive power floating degree and the energy storage device charging and discharging fluctuation degree data, performing weighted addition, and obtaining the comprehensive evaluation value of the power grid state by combining the power values ​​includes the following steps:

[0047] After standardizing the three data of power consumption deviation, comprehensive power fluctuation and energy storage equipment charging and discharging fluctuation, the covariance matrix of the data is calculated to obtain the eigenvalues ​​and eigenvectors of the covariance matrix. After determining the weights of each principal component according to the contribution rate of the eigenvalue, the weight coefficients of each indicator are obtained.

[0048] Calculation formula for comprehensive evaluation value of power grid status:

[0049]

[0050] in, is the comprehensive evaluation value of the power grid status, is the predicted value of the power consumption deviation degree, The comprehensive power fluctuation degree, It is the charging and discharging fluctuation data of energy storage equipment. , and is the corresponding weight coefficient; among them, is the power value at a certain time point, when When the constraints are met, The value of is 0, when When the constraints are not met, is assigned a value of 1.

[0051] As a further solution of the present invention, the comprehensive evaluation value of the power grid state is compared with the warning threshold value and analyzed, including the following steps:

[0052] Based on historical data and grid operation experience, the 95th percentile of the comprehensive evaluation value of the grid status in historical data is selected as the warning threshold;

[0053] If the comprehensive evaluation value of the power grid status is ≥ the warning threshold, the system will issue a warning signal;

[0054] If the comprehensive evaluation value of the power grid status is less than the warning threshold, the system maintains real-time monitoring.

[0055] Beneficial effects of the present invention:

[0056] The present invention sets the sampling frequency through the Nyquist sampling theorem, constructs the power output limit constraint condition according to the signal and the number of sampling points, obtains the maximum and minimum power values, and determines the upper and lower limits of the constraint conditions; and uses the time domain analysis method, descriptive statistical analysis and sliding window technology to analyze and extract features of the acquired source grid load storage data to obtain a variety of feature data, including the predicted value of the degree of deviation of power consumption, the comprehensive power floating degree and the degree of charging and discharging fluctuation of the energy storage equipment, and combines the feature data and the power value to obtain the comprehensive evaluation value of the power grid state, thereby realizing comprehensive monitoring and comprehensive evaluation of the power grid state. According to the statistical analysis results of historical data and combined with the power grid operation experience, the 95% quantile of the comprehensive evaluation value of the power grid state in the historical data is selected as the early warning threshold, and the currently calculated comprehensive evaluation value of the power grid state is compared with the preset early warning threshold, thereby realizing the monitoring and early warning mechanism, so that when an abnormal situation occurs in the power grid, the early warning system can respond quickly and adjust the system in time. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.

[0058] Figure 1 It is a module flow chart of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 As shown, the present invention is a power grid regulation system based on big data, comprising the following modules:

[0061] Grid information acquisition module: obtains the power supply side, grid side, load side and energy storage measurement data in the grid information, pre-processes the data, and uploads the pre-processed data to the cloud storage service;

[0062] Grid constraint condition building module: The sampling frequency is set by the Nyquist sampling theorem, and the power output limit constraint condition is built according to the signal and the number of sampling points to obtain the maximum and minimum power values;

[0063] Power grid big data analysis module: Receives data collected by the power grid information acquisition module, and uses the time domain analysis method to extract features from the data. A model is built based on the periodic power consumption deviation and seasonal power consumption deviation to obtain the power consumption deviation prediction value. Different weights are assigned to the voltage fluctuation, current fluctuation and frequency fluctuation to obtain the comprehensive power fluctuation. The sliding window technology is used to extract the number of charge and discharge times and the peak and valley values ​​extracted from the hourly data to obtain the charging and discharging fluctuation of the energy storage device.

[0064] Power forecast model module: By assigning corresponding weights to the power consumption deviation prediction value, the comprehensive power fluctuation degree and the energy storage equipment charging and discharging fluctuation degree data, weighted addition is performed, and the power value is combined to obtain the comprehensive evaluation value of the power grid status;

[0065] Power grid early warning and adjustment module: compares the comprehensive evaluation value of the power grid status with the early warning threshold and performs analysis.

[0066] Specifically, by obtaining corresponding data from the power supply side, grid side, load side and energy storage installation equipment respectively, pre-processing the collected data, uploading the processed data to the cloud storage service, the grid big data analysis module needs to regularly download the latest data from the cloud for analysis; the sampling frequency is set according to the Nyquist sampling theorem, the power between each data point is obtained by calculating the square sum of the signal divided by the number of sampling points, the maximum power value and the minimum power value are determined, and the power output limit constraint condition is constructed; the ARIMAX model is fitted with the historical periodic electricity consumption deviation degree and the seasonal electricity consumption deviation degree data to obtain the preliminary form of the electricity consumption deviation model, and then verified, the fitting and verification process is repeated to obtain the trained model, and the trained ARIMAX model is fitted to the new The input data is predicted to obtain the predicted value of the deviation of electricity consumption, and the peak-to-valley value characteristics and the average value characteristics of the number of charge and discharge times are combined to obtain a multi-dimensional feature data, and the voltage fluctuation degree, current fluctuation degree and frequency fluctuation degree are obtained respectively. An expert team in the power field is organized to evaluate the relative importance of voltage, current and frequency in the stability of the power grid based on the experts' experience in power system operation, detection and analysis, and determine the weight coefficients corresponding to the three fluctuation degrees to describe the charging and discharging fluctuation degree of the energy storage device; the predicted value of the deviation of electricity consumption, the comprehensive power fluctuation degree and the charging and discharging fluctuation degree of the energy storage device are combined with the power value to obtain the comprehensive evaluation value of the power grid state; the comprehensive evaluation value of the power grid state is compared with the warning threshold and analyzed.

[0067] In one embodiment of the present invention, the obtaining of power supply side, grid side, load side and energy storage measurement data in the grid information, preprocessing the data, and uploading the preprocessed data to a cloud storage service includes the following steps:

[0068] On the power supply side, power generation data, including power generation and power generation time, are collected in real time by connecting to the power plant's automation system interface. On the grid side, grid operation data, including voltage, current and frequency, are collected by intelligent sensor devices at each node of the grid. On the load side, user electricity consumption data, including power consumption during peak, normal and off-peak hours and total daily, weekly and monthly electricity consumption, are collected by deploying smart meters. For energy storage measurement, the operation data of energy storage equipment is obtained through the energy storage management system, including the number of battery charges, discharges, charge and discharge amounts. The collected data is pre-processed, including removing outliers and filling in missing data, and converting data from different sources into a unified format. The processed data is uploaded to the cloud storage service, and the grid big data analysis module downloads the data from the cloud for analysis.

[0069] Specifically, corresponding data is obtained from the power supply side, grid side, load side and energy storage installation equipment respectively, the collected data is pre-processed, outliers are identified and removed using statistical methods, and missing data is filled by median filling. Since the data comes from different sources and has different formats and units, data conversion is performed so that all data are expressed in a unified format and unit. Cloud storage space is set up, and the processed data is uploaded to the cloud storage service. The power grid big data analysis module needs to regularly download the latest data from the cloud for analysis.

[0070] In one embodiment of the present invention, the sampling frequency is set by the Nyquist sampling theorem, a power output limit constraint condition is constructed according to the signal and the number of sampling points, and the maximum and minimum power values ​​are obtained, including the following steps:

[0071] By sampling the signal, setting the sampling frequency according to the Nyquist sampling theorem, using sampling software to discretize the analog signal to obtain digital sample points, and calculating the square sum of the signal divided by the number of sampling points to obtain the power between each data point;

[0072] Power calculation formula:

[0073]

[0074] in, is the power corresponding to the sample point, For each sample point, is the index of the sample point, is the total number of sampling points;

[0075] Traverse all the calculated power values ​​and find the maximum power value As the maximum power in signal processing, the minimum power value found As the minimum power in signal processing;

[0076] The power output limit constraint is constructed by obtaining the maximum power value and the minimum power value:

[0077]

[0078] in, is the minimum power value, is the power value at a certain time point, is the maximum power value.

[0079] Specifically, the sampling frequency is set according to the Nyquist sampling theorem, and the analog signal is discretized and sampled using sampling software to obtain digital sample points. Each sample point represents the amplitude value of the signal at a specific time point. The power between each data point is obtained by calculating the square sum of the signal and dividing it by the number of sampling points. The maximum power value and the minimum power value are determined, and the power output limit constraint condition is constructed.

[0080] In one embodiment of the present invention, the data collected by the receiving power grid information acquisition module is extracted by using a time domain analysis method, and a model is constructed by the periodic power consumption deviation degree and the seasonal power consumption deviation degree to obtain a power consumption deviation degree prediction value, including the following steps:

[0081] Obtain the raw data of electricity consumption during peak hours, normal hours, and off-peak hours, as well as the total electricity consumption per day, week, and month, assign a timestamp to each raw data, sort the data points by timestamp, convert the raw data into a time series format, extract the key features of the time series through time domain analysis methods, including periodicity and seasonal fluctuations, identify the periodic components in the data by drawing a time series graph, and for each identified period, calculate the average electricity consumption within each period;

[0082] The calculation formula of average power consumption in a cycle is:

[0083]

[0084] in, is the average power consumption during the period, For the Daily electricity consumption, is the total number of days;

[0085] The degree of deviation of cycle power consumption is obtained by the ratio of power consumption on one day in the cycle to the average power consumption of the cycle;

[0086] The calculation formula of the deviation degree of cycle power consumption is:

[0087]

[0088] in, is the degree of deviation of cycle power consumption, For the Daily electricity consumption, is the average power consumption during the period;

[0089] The seasonal electricity consumption deviation is obtained by the ratio of the average electricity consumption in each season to the average electricity consumption throughout the year;

[0090] Calculation formula for seasonal electricity consumption deviation:

[0091]

[0092] in, is the seasonal electricity consumption deviation degree, when =1, is the average electricity consumption in spring, is the deviation degree of electricity consumption in spring. =2, is the average electricity consumption in summer, is the degree of deviation of electricity consumption in summer. =3, is the average electricity consumption in autumn, is the deviation degree of electricity consumption in autumn. =4, is the average electricity consumption in winter, is the deviation degree of winter electricity consumption, is the average annual electricity consumption;

[0093] Historical data including the deviation degree of periodic electricity consumption and seasonal electricity consumption are collected. After preprocessing the data, the ARIMAX model is fitted using the historical data to obtain the preliminary form of the electricity consumption deviation model. The model is verified by white noise test, and the fitting and verification process is repeated to obtain a trained model. The trained ARIMAX model is used to predict new input data to obtain the predicted value of the electricity consumption deviation degree.

[0094] Specifically, electricity consumption data of different time periods are obtained from the power grid operation system, a timestamp is assigned to each original data point and the data are sorted to determine the time series of the data, the time domain analysis method is used to extract the periodic components in the time series, and the seasonal fluctuations in the time series are analyzed. The degree of deviation of the periodic electricity consumption is obtained by the ratio of the electricity consumption of one day in the period to the average electricity consumption of the period, and the degree of deviation of the seasonal electricity consumption is obtained by the ratio of the average electricity consumption of each season to the average electricity consumption of the whole year. The ARIMAX model is fitted with historical data to obtain a preliminary form of the electricity consumption deviation model, and the model is verified by a white noise test. The fitting and verification process is repeated to obtain a trained model, and the trained ARIMAX model is used to predict new input data to obtain a predicted value of the electricity consumption deviation.

[0095] In one embodiment of the present invention, the step of assigning different weights to the voltage floating degree, the current floating degree and the frequency floating degree to obtain the comprehensive power floating degree comprises the following steps:

[0096] By collecting grid operation data, including voltage, current and frequency data, through intelligent sensor devices at each node of the grid, statistical features, including extreme values ​​and average values, are extracted from the acquired data through descriptive statistical analysis. The voltage fluctuation degree, current fluctuation degree and frequency fluctuation degree are obtained based on the statistical features and real-time data. An expert team in the power field is organized to evaluate the relative importance of voltage, current and frequency in grid stability based on the experts' experience in power system operation, detection and analysis, and determine the weight coefficients corresponding to the three fluctuation degrees.

[0097] The calculation formula of comprehensive power fluctuation degree is:

[0098] in, The comprehensive power fluctuation degree, is the number of measurements, For the The voltage value obtained by the measurement is is the maximum voltage value, is the minimum voltage value, For the The current value obtained by the measurement is is the maximum current value, is the minimum current value, For the The frequency value obtained by the measurement is is the maximum frequency value, is the minimum frequency value, , and are the weight coefficients corresponding to voltage, current and frequency.

[0099] Specifically, the grid operation data, including voltage, current and frequency data, are collected by intelligent sensor devices at each node of the grid. Statistical features are extracted from the acquired data through descriptive statistical analysis. The maximum and minimum values ​​of voltage, current and frequency are extracted from historical data to understand the extreme conditions of grid operation. The average values ​​of voltage, current and frequency are calculated to reflect the overall level of grid operation. The corresponding floating degree is calculated according to the extreme values ​​and average values ​​of voltage, current and frequency. An expert team in the power field is organized to evaluate the relative importance of voltage, current and frequency in grid stability based on the experts' experience in power system operation, detection and analysis. The weight coefficients corresponding to the three floating degrees are determined, and the three floating degrees are weighted averaged according to their respective weights to obtain the comprehensive power floating degree. , and are the weight coefficients corresponding to voltage, current and frequency, is 0.3, is 0.5, is 0.2.

[0100] In one embodiment of the present invention, the method of extracting the number of charge and discharge times using the sliding window technology and the peak and valley values ​​extracted from the hourly data to obtain the degree of charge and discharge fluctuation of the energy storage device includes the following steps:

[0101] Collect data of energy storage devices on different time scales in real time from the energy storage management system, including key parameters of the energy storage devices, such as the number of charges, discharges, charge capacity, and discharge capacity. After cleaning the collected data, extract time-related features, extract daily peak and valley features from hourly data, and use sliding window technology to extract the average value of the number of charges and discharges in a continuous time period and combine it with the peak and valley features extracted from hourly data to obtain the degree of fluctuation in the charge and discharge of energy storage devices.

[0102] After standardizing the three data of power consumption deviation, comprehensive power fluctuation and energy storage equipment charging and discharging fluctuation, the covariance matrix of the data is calculated to obtain the eigenvalues ​​and eigenvectors of the covariance matrix. After determining the weights of each principal component according to the contribution rate of the eigenvalue, the weight coefficients of each indicator are obtained.

[0103] The calculation formula for the fluctuation degree of charging and discharging of energy storage equipment is:

[0104] in, is the charging and discharging fluctuation degree of the energy storage device. For the The peak value of electricity consumption per day, For the Peak electricity consumption per day, is the total number of days, For the Number of charges per day, For the Number of discharges per day.

[0105] Specifically, data of energy storage devices at different time scales are collected in real time from the energy storage management system, the collected data are cleaned to remove outliers and noise, and the hourly data are processed to obtain the peak and valley characteristics of daily electricity consumption through the hourly data. The hourly data are aggregated into daily data, and the total number of charges and discharges per day are calculated. The sliding window technology is used to move on the time series, and the average number of charges and discharges in each window is calculated. The peak and valley value characteristics and the average value characteristics of the charge and discharge times are combined to obtain a multi-dimensional feature data for describing the degree of charge and discharge fluctuation of the energy storage device.

[0106] In one embodiment of the present invention, by assigning corresponding weights to the predicted value of power consumption deviation, the comprehensive power floating degree and the energy storage device charging and discharging fluctuation degree data, weighted addition is performed, and the power value is combined to obtain the comprehensive evaluation value of the power grid state, including the following steps:

[0107] Calculation formula for comprehensive evaluation value of power grid status:

[0108]

[0109] in, is the comprehensive evaluation value of the power grid status, is the predicted value of the power consumption deviation degree, The comprehensive power fluctuation degree, It is the charging and discharging fluctuation data of energy storage equipment. , and is the corresponding weight coefficient; among them, is the power value at a certain time point, when When the constraints are met, The value of is 0, when When the constraints are not met, is assigned a value of 1.

[0110] Specifically, different weights are assigned to the data of power consumption deviation degree, comprehensive power floating degree and energy storage equipment charging and discharging fluctuation degree, and the power values ​​are added together to obtain the comprehensive evaluation value of the power grid state. According to the power output restriction constraints, the maximum output power and minimum output power of the model are determined, and it is judged whether the power value meets the constraint conditions. Different values ​​are assigned to the power value according to the two situations of satisfaction and non-satisfaction, and the comprehensive evaluation value of the power grid state is obtained. is 0.4, is 0.5, is 0.5.

[0111] In one embodiment of the present invention, the comparing the comprehensive evaluation value of the power grid state with the warning threshold and analyzing the same comprises the following steps:

[0112] Based on historical data and grid operation experience, the 95th percentile of the comprehensive evaluation value of the grid status in historical data is selected as the warning threshold;

[0113] If the comprehensive evaluation value of the power grid status is ≥ the warning threshold, the system will issue a warning signal;

[0114] If the comprehensive evaluation value of the power grid status is less than the warning threshold, the system maintains real-time monitoring.

[0115] Specifically, based on the statistical analysis results of historical data and combined with the power grid operation experience, the 95% quantile of the comprehensive evaluation value of the power grid status in the historical data is selected as the warning threshold, and the currently calculated comprehensive evaluation value of the power grid status is compared with the preset warning threshold. If the comprehensive evaluation value is greater than or equal to the warning threshold, the system will issue a warning signal; if the comprehensive evaluation value is less than the warning threshold, the system will maintain real-time monitoring.

[0116] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A power grid regulation system based on big data, characterized in that: Includes the following modules: Grid information acquisition module: obtains the power supply side, grid side, load side and energy storage measurement data in the grid information, pre-processes the data, and uploads the pre-processed data to the cloud storage service; Grid constraint condition building module: The sampling frequency is set by the Nyquist sampling theorem, and the power output limit constraint condition is built according to the signal and the number of sampling points to obtain the maximum and minimum power values; Power grid big data analysis module: Receives data collected by the power grid information acquisition module, and uses the time domain analysis method to extract features from the data. A model is built based on the periodic power consumption deviation and seasonal power consumption deviation to obtain the power consumption deviation prediction value. Different weights are assigned to the voltage fluctuation, current fluctuation and frequency fluctuation to obtain the comprehensive power fluctuation. The sliding window technology is used to extract the number of charge and discharge times and the peak and valley values ​​extracted from the hourly data to obtain the charging and discharging fluctuation of the energy storage device. Power forecasting model module: By assigning corresponding weights to the power consumption deviation prediction value, the comprehensive power fluctuation degree and the energy storage equipment charging and discharging fluctuation degree data, weighted addition is performed, and the power value is combined to obtain the comprehensive evaluation value of the power grid status; Power grid early warning and adjustment module: compares the comprehensive evaluation value of the power grid status with the early warning threshold and performs analysis.

2. The power grid regulation system based on big data according to claim 1, characterized in that: The method of obtaining the power supply side, grid side, load side and energy storage measurement data in the grid information, preprocessing the data, and uploading the preprocessed data to the cloud storage service includes the following steps: On the power supply side, power generation data, including power generation and power generation time, are collected in real time by connecting to the power plant's automation system interface. On the grid side, grid operation data, including voltage, current and frequency, are collected by intelligent sensor devices at each node of the grid. On the load side, user electricity consumption data, including power consumption during peak, normal and off-peak hours and total daily, weekly and monthly electricity consumption, are collected by deploying smart meters. For energy storage measurement, the operation data of energy storage equipment is obtained through the energy storage management system, including the number of battery charges, discharges, charge and discharge amounts. The collected data is pre-processed, including removing outliers and filling in missing data, and converting data from different sources into a unified format. The processed data is uploaded to the cloud storage service, and the grid big data analysis module downloads the data from the cloud for analysis.

3. The power grid regulation system based on big data according to claim 1, characterized in that: The method of setting the sampling frequency by the Nyquist sampling theorem, constructing the power output limit constraint condition according to the signal and the number of sampling points, and obtaining the maximum and minimum power values ​​includes the following steps: By sampling the signal, setting the sampling frequency according to the Nyquist sampling theorem, using sampling software to discretize the analog signal to obtain digital sample points, and calculating the square sum of the signal divided by the number of sampling points to obtain the power between each data point; Power calculation formula: ； in, is the power corresponding to the sample point, For each sample point, is the index of the sample point, is the total number of sampling points; Traverse all the calculated power values ​​and find the maximum power value As the maximum power in signal processing, the minimum power value found As the minimum power in signal processing; The power output limit constraint is constructed by obtaining the maximum power value and the minimum power value: ； in, is the minimum power value, is the power value at a certain time point, is the maximum power value.

4. The power grid regulation system based on big data according to claim 1, characterized in that: The receiving power grid information acquisition module collects data, extracts features from the data using a time domain analysis method, and constructs a model to obtain a power consumption deviation prediction value by using the periodic power consumption deviation degree and the seasonal power consumption deviation degree, including the following steps: Obtain the raw data of electricity consumption during peak hours, normal hours, and off-peak hours, as well as the total electricity consumption per day, week, and month, assign a timestamp to each raw data, sort the data points by timestamp, convert the raw data into a time series format, extract the key features of the time series through time domain analysis methods, including periodicity and seasonal fluctuations, identify the periodic components in the data by drawing a time series graph, and for each identified period, calculate the average electricity consumption within each period; The calculation formula of average power consumption in a cycle is: ； in, is the average power consumption during the period, For the Daily electricity consumption, is the total number of days; The degree of deviation of cycle power consumption is obtained by the ratio of power consumption on one day in the cycle to the average power consumption of the cycle; The calculation formula of the deviation degree of cycle power consumption is: ； in, is the degree of deviation of cycle power consumption, For the Daily electricity consumption, is the average power consumption during the period; The seasonal electricity consumption deviation is obtained by the ratio of the average electricity consumption in each season to the average electricity consumption throughout the year; Calculation formula for seasonal electricity consumption deviation: ； in, is the seasonal electricity consumption deviation degree, when =1, is the average electricity consumption in spring, is the deviation degree of electricity consumption in spring. =2, is the average electricity consumption in summer, is the degree of deviation of electricity consumption in summer. =3, is the average electricity consumption in autumn, is the deviation degree of electricity consumption in autumn. =4, is the average electricity consumption in winter, is the deviation degree of winter electricity consumption, is the average annual electricity consumption; Historical data including the deviation degree of periodic electricity consumption and seasonal electricity consumption are collected. After preprocessing the data, the ARIMAX model is fitted using the historical data to obtain the preliminary form of the electricity consumption deviation model. The model is verified by white noise test, and the fitting and verification process is repeated to obtain a trained model. The trained ARIMAX model is used to predict new input data to obtain the predicted value of the electricity consumption deviation degree.

5. The power grid regulation system based on big data according to claim 1, characterized in that: The step of assigning different weights to the voltage floating degree, the current floating degree and the frequency floating degree to obtain the comprehensive power floating degree comprises the following steps: By collecting grid operation data, including voltage, current and frequency data, through intelligent sensor devices at each node of the grid, statistical features, including extreme values ​​and average values, are extracted from the acquired data through descriptive statistical analysis. The voltage fluctuation degree, current fluctuation degree and frequency fluctuation degree are obtained based on the statistical features and real-time data. An expert team in the power field is organized to evaluate the relative importance of voltage, current and frequency in grid stability based on the experts' experience in power system operation, detection and analysis, and determine the weight coefficients corresponding to the three fluctuation degrees. The calculation formula of comprehensive power fluctuation degree is: ； in, The comprehensive power fluctuation degree, is the number of measurements, For the The voltage value obtained by the measurement is is the maximum voltage value, is the minimum voltage value, For the The current value obtained by the measurement is is the maximum current value, is the minimum current value, For the The frequency value obtained by the measurement is is the maximum frequency value, is the minimum frequency value, , and are the weight coefficients corresponding to voltage, current and frequency.

6. The power grid regulation system based on big data according to claim 1, characterized in that: The method of extracting the number of charge and discharge times using the sliding window technology and the peak and valley values ​​extracted from the hourly data to obtain the degree of charge and discharge fluctuation of the energy storage device includes the following steps: Collect data of energy storage devices on different time scales in real time from the energy storage management system, including key parameters of the energy storage devices, such as the number of charges, discharges, charge capacity, and discharge capacity. After cleaning the collected data, extract time-related features, extract daily peak and valley features from hourly data, and use sliding window technology to extract the average value of the number of charges and discharges in a continuous time period and combine it with the peak and valley features extracted from hourly data to obtain the degree of fluctuation in the charge and discharge of energy storage devices. The calculation formula for the fluctuation degree of charging and discharging of energy storage equipment is: ； in, is the charging and discharging fluctuation degree of the energy storage device. For the The peak value of electricity consumption per day, For the Peak electricity consumption per day, is the total number of days, For the Number of charges per day, For the The number of discharges per day.

7. The power grid regulation system based on big data according to claim 1, characterized in that: The method allocates corresponding weights to the power consumption deviation degree prediction value, the comprehensive power floating degree and the energy storage device charging and discharging fluctuation degree data, performs weighted addition, and obtains the comprehensive evaluation value of the power grid state by combining the power values, including the following steps: After standardizing the three data of power consumption deviation, comprehensive power fluctuation and energy storage equipment charging and discharging fluctuation, the covariance matrix of the data is calculated to obtain the eigenvalues ​​and eigenvectors of the covariance matrix. After determining the weights of each principal component according to the contribution rate of the eigenvalue, the weight coefficients of each indicator are obtained. Calculation formula for comprehensive evaluation value of power grid status: ； in, is the comprehensive evaluation value of the power grid status, is the predicted value of the power consumption deviation degree, The comprehensive power fluctuation degree, It is the charging and discharging fluctuation data of energy storage equipment. , and is the corresponding weight coefficient; among them, is the power value at a certain time point, when When the constraints are met, The value of is 0, when When the constraints are not met, is assigned a value of 1.

8. The power grid regulation system based on big data according to claim 1, characterized in that: The method of comparing the comprehensive evaluation value of the power grid state with the early warning threshold and analyzing the same comprises the following steps: Based on historical data and grid operation experience, the 95th percentile of the comprehensive evaluation value of the grid status in historical data is selected as the warning threshold; If the comprehensive evaluation value of the power grid status is ≥ the warning threshold, the system will issue a warning signal; If the comprehensive evaluation value of the power grid status is less than the warning threshold, the system maintains real-time monitoring.

Images