Intelligent prediction system and method for user-side power load based on big data

Abstract

The present invention relates to the technical field of power load prediction of AC distribution network, and discloses a user-side power load intelligent prediction system and method based on big data, wherein the system comprises a user-side historical power load information acquisition module, a user-side power load prediction metering module, and a user-side power load prediction result evaluation module; by combining numerical analysis with a user-side target single-month historical power load parameter, accurate statistical processing of the user-side target single-month historical power load growth rate and the user-side power load prediction month power load growth rate is achieved; accurate statistics of the power load of the user-side power load prediction month are performed based on the power load growth rate parameter of the user-side power load prediction month and the user-side target single-month historical power load parameter, thereby improving the precision and accuracy of the user-side power load prediction result.

Description

Technical Field

[0001] The present invention relates to the technical field of power load prediction of AC distribution network, and specifically to a user-side power load intelligent prediction system and method based on big data. Background Art

[0002] The distribution network refers to a power grid that receives electric energy from the transmission network or regional power plants and distributes it locally or step by step according to voltage to various users through distribution facilities. It is composed of overhead lines, cables, towers, distribution transformers, disconnectors, reactive power compensators and some ancillary facilities, and plays an important role in distributing electric energy in the power grid. The traditional power load forecasting at the user end of the distribution network cannot accurately and truly measure the power load parameters at the user end, nor can it scientifically evaluate the rationality of the power load forecasting results at the user end, which reduces the authenticity and reliability of the power load forecasting at the user end of the distribution network.

[0003] A Chinese invention patent with announcement number CN113131476B discloses a method for predicting power load, which collects historical power load data of distribution station equipment, corrects the data once, sorts it in ascending order, and modifies it twice to obtain historical power load data, and combines a machine learning algorithm to predict power load for the predicted time; however, the above technical solution only measures power load prediction results based on data noise reduction processing and machine learning training, and cannot accurately collect historical sample data of power load and scientifically analyze the rationality of power load prediction results, which reduces the accuracy and authenticity of power load prediction results of distribution station equipment. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the problem that the traditional power load forecasting at the user end of the distribution network cannot accurately and truly calculate the power load parameters at the user end, and cannot scientifically evaluate the rationality of the power load forecasting results at the user end, which reduces the authenticity and reliability of the power load forecasting at the user end of the distribution network, the above purposes are achieved: accurately screening the historical power load at the user end, scientifically counting the growth rate of the historical power load at the user end, accurately calculating the power load at the user end on the predicted date, intelligently evaluating the rationality of the power load forecasting results at the user end, and improving the authenticity and reliability of the power load forecasting at the user end of the distribution network.

[0006] (II) Technical solution

[0007] The present invention is implemented by the following technical scheme: a method for intelligent prediction of user-side power load, the method comprising the following steps:

[0008] S1. Collect monthly and time data of power load forecast at the user end;

[0009] S2, constructing and processing the historical month object for sampling the historical power load information of the user end according to the predicted monthly time data of the power load of the user end, and generating the monthly time data of sampling the historical power load of the user end;

[0010] S3, searching and processing the user-side target historical power load information according to the user-side historical power load sampling month time data and the user-side historical single-month power load data, and generating the user-side target single-month historical power load data;

[0011] S4, performing numerical processing on the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data;

[0012] S5. Perform numerical processing on the power load growth rate of the user-side power load forecast month according to the user-side target single-month historical power load growth rate data, generate power load growth rate data of the user-side power load forecast month, and perform power load metering processing on the user-side power load forecast month in combination with the user-side target single-month historical power load data, generate power load data of the user-side power load forecast month;

[0013] S6, performing a rationality analysis and processing of the user-side power load forecast result according to the power load data of the user-side power load forecast month and the user-side historical single-month extreme state power load data, generating rationality analysis data of the user-side power load forecast result, and directly ending the current user-side power load forecast operation when it is abnormal;

[0014] S7. When it is normal, construct the user-side power load forecast feedback data and execute the user-side power load forecast result feedback operation.

[0015] Preferably, the operation steps for collecting the user-side power load forecast monthly time data are as follows:

[0016] S11. Collect the time characteristic text information of the current and next month of the power load forecast of the specific user end online through the data collection dialog box of the distribution network management platform, and generate the monthly time data of the power load forecast of the user end ,in The units include year and month.

[0017] Preferably, the historical month object construction process of sampling the historical power load information of the user end is performed according to the predicted monthly time data of the power load of the user end, and the operation steps of generating the monthly time data of sampling the historical power load of the user end are as follows:

[0018] S21, predicting monthly time data based on the user-side power load Construct the historical monthly time characteristic values ​​of the user-side historical power load information sampling, and generate the user-side historical power load sampling month time data set ,in Indicates the monthly time data related to the user's power load forecast The user-side historical power load sampling month and time data for the corresponding historical month of the previous year, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month; Indicates the monthly time data related to the user's power load forecast The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month; , , , , The units include year and month.

[0019] Preferably, the steps of searching and processing the target historical power load information of the user end according to the user end historical power load sampling month time data and the user end historical single-month power load data to generate the target historical power load data of the user end for a single month are as follows:

[0020] S31. Establish a user-side historical monthly electricity load data set , ;in Indicates the user's The historical monthly electricity load data of the user end corresponding to the historical month, Indicates the maximum number of historical months, and the user-side historical single-month power load data indicates the single-month power load data of the specific user-side historical records, where The unit is kilowatt;

[0021] S32: Collect the user-side historical power load sampling month and time data The user-side historical power load sampling month time data and the user-side historical single-month power load data set The historical monthly electricity load data of the user side mentioned in According to the historical monthly time feature matching, search for the user-side historical power load data corresponding to the user-side historical power load sampling month time data , and generate a user-side target monthly historical electricity load data set , execute to generate the user-side target single-month historical electricity load data set The specific steps are as follows:

[0022] S321, initialization parameters, update power load, search for the number of crow populations, maximum number of iterations, flight distance ;

[0023] S322, initializing the power load to search for the crow individual in the user's historical monthly power load data set The initial position and memory in the search space, Only the electrical load is used to search for crows that are randomly distributed in a multidimensional search space, i.e. The dimension of the random distribution space of crows searching only by electric load is The user-side historical monthly electricity load data set In the search space; in the first iteration, it is assumed that the power load search crow individual hides the food in the initial position; that is, in the first iteration, it is assumed that the power load search crow individual combines the user-side historical power load sampling month time data with the user-side historical single-month power load data set The historical monthly electricity load data of the user end at the initial position in the search space Make a match;

[0024] S323, calculate the individual fitness value of each power load searching crow, that is, calculate the user-side historical power load sampling month time data and the user-side historical single-month power load data set The user's historical monthly electricity load data in the search space The fitness value of

[0025] S324: Update the power load to search for the crow individual in the user's historical monthly power load data set. The position in the search space, i.e., the historical monthly electricity load data set at the user end Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month time data. The position of the crow is updated by the following formula: ,in Indicates After iterations, use the electric load to search for crow individuals In the user's historical monthly electricity load data set The new position in the search space, Indicates After iterations, use the electric load to search for crow individuals In the user's historical monthly electricity load data set The position in the search space, represents a random number uniformly distributed between

[01] , Indicates After iterations, use the electric load to search for crow individuals In the user's historical monthly electricity load data set The flight distance in the search space, Indicates After iterations, use the electric load to search for crow individuals In the user's historical monthly electricity load data set The food hiding location in the search space, i.e. After iterations, use the electric load to search for crow individuals The user-side historical monthly electricity load data set Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month time data. location;

[0026] S325, determine the feasibility of the new position, determine the feasibility of the new position of each power load search crow individual; if the power load search crow individual new position is feasible, the power load search crow individual will update its position, that is, in the user end historical single month power load data set Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month time data. , the electricity load search crow individual is updated to the historical monthly electricity load data of the user end that successfully matches Otherwise, the power load search crow individual stays at the current user-side historical monthly power load data The location will not be moved to the new user's historical monthly electricity load data. location;

[0027] S326, evaluate the fitness value of the new position, calculate the fitness value of each individual new position of the power load search crow, that is, calculate the historical monthly power load data set of the user end The user-side historical power load sampling month time data in the search space and the user-side historical single-month power load data at the new location The fitness value of

[0028] S327, update memory, if the fitness value of the new position of the power load search crow is greater than the fitness value of the initial position in the memory, the power load search crow updates its memory through the new position, otherwise it does not update its memory; that is, in the user end historical single month power load data set Search for the user-side historical monthly power load data with the largest fitness value for the user-side historical power load sampling month time data ;

[0029] S328: When the maximum number of iterations is met, output the user-side historical monthly power load data that matches the user-side historical power load sampling month and time data. ;

[0030] S329: The user-side historical monthly electricity load data output in step S328 , and generate the user-side target monthly historical electricity load data set through data identification ,in Indicates the user's historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user's historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user's historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user's historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user's historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, including , , , , The unit is kilowatt.

[0031] Preferably, the operation steps of performing numerical processing on the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data are as follows:

[0032] S41, based on the user-side target monthly historical electricity load data set The user-side target single-month historical electricity load data is numerically measured and processed to obtain the user-side target single-month historical electricity load growth rate data set. ,in Indicates the monthly historical power load data of the user end target Relative to the user end target single month historical electricity load data The user-side target monthly historical electricity load growth rate data, including ; Indicates the monthly historical power load data of the user end target Relative to the user end target single month historical electricity load data The user-side target monthly historical electricity load growth rate data, including ; Indicates the monthly historical power load data of the user end target Relative to the user end target single month historical electricity load data The user-side target monthly historical electricity load growth rate data, including , , , All figures are rounded off.

[0033] Preferably, numerical processing of the power load growth rate of the user-side power load forecast month is performed according to the user-side target single-month historical power load growth rate data, the power load growth rate data of the user-side power load forecast month is generated, and the power load metering processing of the user-side power load forecast month is performed in combination with the user-side target single-month historical power load data, and the operation steps of generating the power load data of the user-side power load forecast month are as follows:

[0034] S51, based on the user-side target single-month historical electricity load growth rate data set The user-side target monthly historical electricity load growth rate data is combined with the proportional formula to perform numerical measurement processing on the electricity load growth rate of the user-side electricity load forecast month, and generate the electricity load growth rate data of the user-side electricity load forecast month. ,in , Indicates the monthly time data of the user-side power load forecast The corresponding power load is relative to the user-side target monthly historical power load data The growth rate data of electricity load;

[0035] S52: The user-side power load forecast month power load growth rate data The target monthly historical electricity load data of the user end The power load value of the user-side power load forecast month is measured and processed according to the proportional formula to generate the power load data of the user-side power load forecast month ,in , The unit is kilowatt.

[0036] Preferably, the rationality analysis and processing of the user-side power load forecast result is performed based on the power load data of the user-side power load forecast month and the user-side historical single-month extreme state power load data to generate the rationality analysis data of the user-side power load forecast result. When it is abnormal, the operation steps of directly ending the current user-side power load forecast operation are as follows:

[0037] S61. Establish a historical monthly extreme power load data set at the user end , ;in Indicates The historical monthly extreme power load data of the user end corresponding to the extreme power state type of the user end, Indicates the maximum value of the extreme power consumption state type at the user end; the extreme power consumption state type at the user end includes extreme high temperature weather state, extreme low temperature weather state, extreme flood disaster state, extreme typhoon disaster state and extreme leakage fault state at the user end. The user end historical single-month extreme state power load data indicates abnormal single-month historical power load data at the user end under extreme weather state and extreme power fault state. The unit is kilowatt;

[0038] S62, using the KD tree nearest neighbor search algorithm to calculate the user-side power load forecast monthly power load data The user-side historical monthly extreme power load data set The user-side historical single-month extreme power load data mentioned in Carry out power load value matching, and generate rationality analysis data of user-side power load forecast results based on the power load value matching results ;

[0039] when and The power load value matching is successful, indicating that the user-side power load forecast result meets the user-side power extreme state. At this time, the user-side power load forecast result is unreasonable, and the user-side power load forecast result rationality analysis data is output. If it is abnormal, the user-side power load forecasting operation is terminated directly;

[0040] when and If the power load value is not matched successfully, it means that the power load forecast result at the user end does not meet the extreme power consumption state at the user end. At this time, the power load forecast result at the user end is reasonable, and the rationality analysis data of the power load forecast result at the user end is output. is normal.

[0041] Preferably, when the condition is normal, the operation steps of constructing the user-side power load forecast feedback data and performing the user-side power load forecast result feedback operation are as follows:

[0042] S71, when the user-side power load forecast result rationality analysis data When the user-side power load forecast month time data is normal, , the user-side power load forecast monthly power load data and the rationality analysis data of the user-side power load forecast results Combine data to construct user-side power load forecast feedback data ;

[0043] S72: Feedback the user-side power load forecast data Feedback is pushed to the distribution network management platform through the IoT communication network.

[0044] A user-side power load intelligent prediction system based on big data, used to implement the user-side power load intelligent prediction method, the system includes a user-side historical power load information acquisition module, a user-side power load prediction and metering module, and a user-side power load prediction result evaluation module;

[0045] The user-side historical power load information acquisition module includes a user-side power load forecast month time acquisition unit, a user-side historical power load sampling month time generation unit, a user-side historical single-month power load storage unit, and a user-side historical power load search unit;

[0046] The user-side power load forecast month time collection unit collects the user-side power load forecast month time data through the distribution network management platform; the user-side historical power load sampling month time generation unit constructs and processes the historical month object of the user-side historical power load information sampling according to the user-side power load forecast month time data, and generates the user-side historical power load sampling month time data; the user-side historical single-month power load storage unit stores the user-side historical single-month power load data based on big data; the user-side historical power load search unit searches and processes the user-side target historical power load information according to the user-side historical power load sampling month time data and the user-side historical single-month power load data, and generates the user-side target single-month historical power load data;

[0047] The user-side power load forecasting and metering module includes a user-side historical power load growth rate metering unit, a user-side power load forecasting month growth rate metering unit, and a user-side power load forecasting month power load metering unit;

[0048] The user-side historical electricity load growth rate metering unit performs numerical processing on the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data; the user-side electricity load forecast month growth rate metering unit performs numerical processing on the electricity load growth rate of the user-side electricity load forecast month according to the user-side target single-month historical electricity load growth rate data to generate the user-side electricity load forecast month electricity load growth rate data; the user-side electricity load forecast month electricity load metering unit performs electricity load metering processing on the user-side electricity load forecast month according to the user-side electricity load forecast month electricity load growth rate data and the user-side target single-month historical electricity load data to generate the user-side electricity load forecast month electricity load data;

[0049] The user-side power load forecast result evaluation module includes a user-side historical single-month extreme power load storage unit, a user-side power load forecast result rationality analysis unit, and a user-side power load forecast result output feedback unit;

[0050] The user-side historical single-month extreme state electricity load storage unit stores the user-side historical single-month extreme state electricity load data based on big data; the user-side electricity load forecast result rationality analysis unit performs rationality analysis and processing on the user-side electricity load forecast result based on the user-side electricity load forecast month electricity load data and the user-side historical single-month extreme state electricity load data to generate user-side electricity load forecast result rationality analysis data; the user-side electricity load forecast result output feedback unit constructs the user-side electricity load forecast feedback data and executes the user-side electricity load forecast result feedback operation.

[0051] (III) Beneficial effects

[0052] The present invention provides a user-side power load intelligent prediction system and method based on big data. It has the following beneficial effects:

[0053] 1. Accurately collect the monthly time parameters of the user-side power load forecast through the distribution network management platform to provide reliable data support for the scientific forecast of the user-side power load; independently and accurately construct the historical monthly time information of the user-side historical power load information sampling based on the monthly time parameters of the user-side power load forecast; accurately search the user-side historical power load information based on the monthly time parameters of the user-side historical power load sampling combined with the intelligent recognition algorithm and the user-side historical single-month power load parameters based on big data storage, realize intelligent screening of the user-side historical power load information, and improve the efficiency and reliability of intelligent forecasting of the user-side power load.

[0054] 2. By combining numerical analysis with the user-side target single-month historical electricity load parameters, accurate statistical processing of the user-side target single-month historical electricity load growth rate and the user-side electricity load forecast month electricity load growth rate is achieved; accurate statistics of the user-side electricity load forecast month electricity load are performed based on the user-side electricity load forecast month electricity load growth rate parameters and the user-side target single-month historical electricity load parameters, thereby improving the precision and accuracy of the user-side electricity load forecast results.

[0055] 3. By conducting a scientific analysis of the rationality of the user-side power load forecast results based on the power load parameters of the user-side power load forecast month and the user-side historical single-month extreme state power load parameters, intelligent filtering of unreasonable user-side power load forecast results is achieved, thereby improving the quality of user-side power load forecasts; based on the user-side power load forecast month, power load forecast results and power load forecast rationality analysis results, user-side power load forecast feedback information is independently constructed to achieve efficient and accurate collection of user-side power load effective forecast result information, and visual feedback of user-side power load forecast results is performed, thereby improving the applicability and authenticity of the distribution network user-side power load forecast. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A schematic diagram of a module of a user-side power load intelligent prediction system based on big data provided by the present invention;

[0057] Figure 2 This is a flow chart of the user-side power load intelligent prediction method provided by the present invention. DETAILED DESCRIPTION

[0058] 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.

[0059] The embodiments of the user-side power load intelligent prediction system and method based on big data are as follows: Example

[0060] See also Figure 1 - Figure 2 , a method for intelligent prediction of user-side power load, the method comprising the following steps:

[0061] S1. Collect monthly and time data of power load forecast at the user end;

[0062] S2. Construct and process the historical month object for sampling the historical power load information of the user end according to the predicted monthly time data of the power load of the user end, and generate the monthly time data of sampling the historical power load of the user end;

[0063] S3, searching and processing the user-side target historical power load information according to the user-side historical power load sampling month time data and the user-side historical single-month power load data, and generating the user-side target single-month historical power load data;

[0064] S4. Perform numerical processing on the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data;

[0065] S5. Perform numerical processing on the power load growth rate of the user-side power load forecast month according to the user-side target single-month historical power load growth rate data, generate the power load growth rate data of the user-side power load forecast month, and perform power load metering processing on the user-side power load forecast month in combination with the user-side target single-month historical power load data, generate the power load data of the user-side power load forecast month;

[0066] S6. Perform rationality analysis and processing on the user-side power load forecast result based on the power load data of the user-side power load forecast month and the user-side historical single-month extreme power load data, generate rationality analysis data on the user-side power load forecast result, and directly terminate the current user-side power load forecast operation if it is abnormal;

[0067] S7. When it is normal, construct the user-side power load forecast feedback data and execute the user-side power load forecast result feedback operation.

[0068] For further information, see Figure 1 - Figure 2 The steps for collecting the monthly and time data of the user-side power load forecast are as follows:

[0069] S11. Collect the time characteristic text information of the current and next month of the power load forecast of the specific user end online through the data collection dialog box of the distribution network management platform, and generate the monthly time data of the power load forecast of the user end ,in The units include year and month.

[0070] The historical month object for sampling the historical power load information of the user end is constructed and processed according to the predicted monthly time data of the power load of the user end. The operation steps for generating the monthly time data of the historical power load sampling of the user end are as follows:

[0071] S21. Monthly and time data based on user-side power load forecast Construct the historical monthly time characteristic values ​​of the user-side historical power load information sampling, and generate the user-side historical power load sampling month time data set ,in Indicates the monthly time data related to the user-side power load forecast The user-side historical power load sampling month and time data for the corresponding historical month of the previous year, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month; Indicates the monthly time data related to the user-side power load forecast The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates the monthly and time data of the historical power load sampling at the user end The user-side historical power load sampling month and time data corresponding to the previous historical month; , , , , The units include year and month.

[0072] The steps for searching and processing the target historical power load information of the user end according to the sampling month time data of the user end historical power load and the historical single-month power load data of the user end to generate the target single-month historical power load data of the user end are as follows:

[0073] S31. Establish a user-side historical monthly electricity load data set , ;in Indicates the user's The user-side historical monthly electricity load data corresponding to the historical month, Indicates the maximum number of historical months. The user-side historical monthly electricity load data indicates the monthly electricity load data of the specific user-side historical records. The unit is kilowatt;

[0074] S32, collect the user-side historical power load sampling month and time data The user-side historical power load sampling month time data and the user-side historical single-month power load data collection Historical monthly electricity load data for medium-sized users According to the matching of historical monthly time features, search for the user-side historical electricity load data corresponding to the user-side historical electricity load sampling month time data , and generate a user-side target monthly historical electricity load data set , execute and generate the user-side target single-month historical electricity load data set The specific steps are as follows:

[0075] S321, initialization parameters, update power load, search for the number of crow populations, maximum number of iterations, flight distance ;

[0076] S322, initialize power load to search for the crow individual's historical monthly power load data set at the user end The initial position and memory in the search space, Only the electrical load is used to search for crows that are randomly distributed in a multidimensional search space, i.e. The dimension of the random distribution space of crows searching only by electric load is A collection of historical monthly electricity load data at the user end In the search space; in the first iteration, it is assumed that the power load search crow individual hides the food in the initial position; that is, in the first iteration, it is assumed that the power load search crow individual combines the user-side historical power load sampling month time data with the user-side historical single-month power load data set The historical monthly electricity load data of the user end at the initial position in the search space Make a match;

[0077] S323, calculate the individual fitness value of each power load search crow, that is, calculate the user-side historical power load sampling month time data and the user-side historical single-month power load data set Historical monthly electricity load data of users in the search space The fitness value of

[0078] S324. Update the electricity load to search for the crow individual's historical monthly electricity load data set at the user end The location in the search space, i.e., the historical monthly electricity load data set at the user end Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month and time data. The position of the crow is updated by the following formula: ,in Indicates After iterations, use the electric load to search for crow individuals Historical monthly electricity load data collection at the user end The new position in the search space, Indicates After iterations, use the electric load to search for crow individuals Historical monthly electricity load data collection at the user end The position in the search space, represents a random number uniformly distributed between

[01] , Indicates After iterations, use the electric load to search for crow individuals Historical monthly electricity load data collection at the user end The flight distance in the search space, Indicates After iterations, use the electric load to search for crow individuals Historical monthly electricity load data collection at the user end The food hiding location in the search space, i.e. After iterations, use the electric load to search for crow individuals Historical monthly electricity load data collection at the user end Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month and time data. location;

[0079] S325, determine the feasibility of the new position, determine the feasibility of the new position of each power load searching crow individual; if the power load searching crow individual's new position is feasible, the power load searching crow individual will update its position, that is, the power load data set of the user end for a single month in history Search the search space to find the user-side historical monthly power load data that matches the user-side historical power load sampling month and time data. , the electricity load search crow individual is updated to the historical monthly electricity load data of the matching user end Otherwise, the power load search crow individual stays at the historical monthly power load data of the current user end. Location, will not be moved to the new user's historical monthly electricity load data location;

[0080] S326, evaluate the fitness value of the new position, calculate the fitness value of each individual new position of the power load search crow, that is, calculate the user-side historical single-month power load data set The user-side historical power load sampling month time data in the search space and the user-side historical single-month power load data at the new location The fitness value of

[0081] S327, update memory, if the fitness value of the new position of the power load search crow is greater than the fitness value of the initial position in the memory, the power load search crow updates its memory through the new position, otherwise it does not update its memory; that is, in the user end historical single month power load data set Search for the user-side historical monthly power load data with the largest fitness value for the user-side historical power load sampling month time data ;

[0082] S328. When the maximum number of iterations is met, output the user-side historical monthly power load data that matches the user-side historical power load sampling month and time data. ;

[0083] S329: The user-side historical monthly electricity load data output in step S328 is , and generate the user-side target monthly historical electricity load data set through data identification ,in Indicates the user-side historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user-side historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user-side historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user-side historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, Indicates the user-side historical power load sampling month and time data The corresponding user-side target monthly historical electricity load data, including , , , , The unit is kilowatt.

[0084] Through the user-side electricity load forecast month time collection unit, the distribution network management platform is used to accurately collect the user-side electricity load forecast month time parameters, providing reliable data support for the scientific prediction of the user-side electricity load; the user-side historical electricity load sampling month time generation unit independently and accurately constructs the historical month time information of the user-side historical electricity load information sampling based on the user-side electricity load forecast month time parameters; the user-side historical electricity load search unit accurately searches for the user-side historical electricity load information based on the user-side historical electricity load sampling month time parameters combined with the intelligent recognition algorithm and the user-side historical single-month electricity load parameters based on big data storage, realizes intelligent screening of the user-side historical electricity load information, and improves the efficiency and reliability of intelligent prediction of the user-side electricity load.

[0085] For further information, see Figure 1 - Figure 2 , based on the user-side target single-month historical electricity load data, the user-side target historical electricity load growth rate is numerically processed, and the operation steps for generating the user-side target single-month historical electricity load growth rate data are as follows:

[0086] S41. Based on the user-side target monthly historical electricity load data collection The user-side target single-month historical electricity load data is used to perform numerical measurement processing on the growth rate of the user-side target historical electricity load, and a user-side target single-month historical electricity load growth rate data set is generated. ,in Indicates the user-side target monthly historical electricity load data Relative to the user-side target monthly historical electricity load data The user-side target monthly historical electricity load growth rate data, including ; Indicates the user-side target monthly historical electricity load data Relative to the user-side target monthly historical electricity load data The user-side target monthly historical electricity load growth rate data, including ; Indicates the user-side target monthly historical electricity load data Relative to the user-side target monthly historical electricity load data The user-side target monthly historical electricity load growth rate data, including , , , All figures are rounded off.

[0087] According to the user-side target single-month historical power load growth rate data, the power load growth rate of the user-side power load forecast month is numerically processed to generate the power load growth rate data of the user-side power load forecast month and combined with the user-side target single-month historical power load data to perform power load metering processing for the user-side power load forecast month. The operation steps for generating the power load data of the user-side power load forecast month are as follows:

[0088] S51. Data collection based on the monthly historical electricity load growth rate of the user-side target The user-side target monthly historical electricity load growth rate data is combined with the proportional formula to perform numerical measurement processing on the electricity load growth rate of the user-side electricity load forecast month, generating the electricity load growth rate data of the user-side electricity load forecast month. ,in , Indicates the monthly time data of the user-side power load forecast The corresponding power load is relative to the user-side target monthly historical power load data The growth rate data of electricity load;

[0089] S52, forecast the monthly electricity load growth rate data of the user end electricity load The monthly historical power load data of the user end target The power load value of the user-side power load forecast month is measured and processed according to the proportional formula to generate the power load data of the user-side power load forecast month ,in , The unit is kilowatt.

[0090] Through the cooperation of the user-side historical electricity load growth rate metering unit and the user-side electricity load forecast month growth rate metering unit, based on the user-side target single-month historical electricity load parameters combined with numerical analysis, accurate statistical processing of the user-side target single-month historical electricity load growth rate and the user-side electricity load forecast month electricity load growth rate is achieved; the user-side electricity load forecast month electricity load metering unit performs accurate statistics of the user-side electricity load forecast month based on the user-side electricity load growth rate parameters of the user-side electricity load forecast month and the user-side target single-month historical electricity load parameters, thereby improving the precision and accuracy of the user-side electricity load forecast results.

[0091] For further information, see Figure 1 - Figure 2, perform rationality analysis and processing on the user-side power load forecast results based on the monthly power load data of the user-side power load forecast and the user-side historical single-month extreme power load data, generate rationality analysis data on the user-side power load forecast results, and directly end the current user-side power load forecast operation if it is abnormal. The following are the steps:

[0092] S61. Establish a historical monthly extreme power load data set at the user end , ;in Indicates The historical monthly extreme power load data of the user end corresponding to the extreme power state type of the user end, Indicates the maximum value of the extreme power consumption state type at the user end; the extreme power consumption state types at the user end include extreme high temperature weather state, extreme low temperature weather state, extreme flood disaster state, extreme typhoon disaster state and extreme leakage fault state at the user end. The historical single-month extreme power load data at the user end indicates abnormal single-month historical power load data at the user end under extreme weather conditions and extreme power fault conditions. The unit is kilowatt;

[0093] S62, using the KD tree nearest neighbor search algorithm to predict the monthly power load data of the user end power load The data set of extreme power load in a single month at the user end Historical single-month extreme power load data for medium-sized users Carry out power load value matching, and generate rationality analysis data of user-side power load forecast results based on the power load value matching results ;

[0094] when and The power load value matching is successful, indicating that the user-side power load forecast result meets the user-side power extreme state. At this time, the user-side power load forecast result is unreasonable, and the user-side power load forecast result rationality analysis data is output. If it is abnormal, the user-side power load forecasting operation is terminated directly;

[0095] when and If the power load value is not matched successfully, it means that the power load forecast result at the user end does not meet the extreme power consumption state at the user end. At this time, the power load forecast result at the user end is reasonable, and the rationality analysis data of the power load forecast result at the user end is output. is normal.

[0096] When it is normal, the operation steps of constructing the user-side power load forecast feedback data and executing the user-side power load forecast result feedback operation are as follows:

[0097] S71. When the user-side power load forecast result is reasonable, the analysis data When it is normal, the user-side power load forecast month and time data , User-side power load forecast monthly power load data And the rationality analysis data of the user-side power load forecast results Combine data to construct user-side power load forecast feedback data ;

[0098] S72, user-side power load forecast feedback data Feedback is pushed to the distribution network management platform through the IoT communication network.

[0099] Through the user-side power load forecast result rationality analysis unit, the rationality of the user-side power load forecast result is scientifically analyzed according to the power load parameters of the user-side power load forecast month and the user-side historical single-month extreme state power load parameters, so as to realize intelligent filtering of unreasonable user-side power load forecast results and improve the quality of user-side power load forecast; the user-side power load forecast result output feedback unit independently constructs the user-side power load forecast feedback information based on the user-side power load forecast month, power load forecast results and power load forecast rationality analysis results, realizes efficient and accurate collection of user-side power load effective forecast result information, and conducts visual feedback of user-side power load forecast results, so as to improve the applicability and authenticity of the power load forecast of the distribution network user. Example

[0100] See also Figure 1 - Figure 2 , a user-side power load intelligent prediction system based on big data, which is used to realize the user-side power load intelligent prediction method. The system includes a user-side historical power load information acquisition module, a user-side power load prediction and metering module, and a user-side power load prediction result evaluation module;

[0101] The user-side historical power load information acquisition module includes a user-side power load forecast month time acquisition unit, a user-side historical power load sampling month time generation unit, a user-side historical single-month power load storage unit, and a user-side historical power load search unit;

[0102] The user-side electricity load forecast month time collection unit collects the user-side electricity load forecast month time data through the distribution network management platform; the user-side historical electricity load sampling month time generation unit constructs and processes the historical month object sampling of the user-side historical electricity load information based on the user-side electricity load forecast month time data, and generates the user-side historical electricity load sampling month time data; the user-side historical single-month electricity load storage unit stores the user-side historical single-month electricity load data based on big data storage; the user-side historical electricity load search unit searches and processes the user-side target historical electricity load information based on the user-side historical electricity load sampling month time data and the user-side historical single-month electricity load data, and generates the user-side target single-month historical electricity load data;

[0103] The user-side electricity load forecasting metering module includes a user-side historical electricity load growth rate metering unit, a user-side electricity load forecasting month growth rate metering unit, and a user-side electricity load forecasting month electricity load metering unit;

[0104] The user-side historical electricity load growth rate metering unit performs numerical processing of the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data; the user-side electricity load forecast month growth rate metering unit performs numerical processing of the electricity load growth rate of the user-side electricity load forecast month based on the user-side target single-month historical electricity load growth rate data to generate the user-side electricity load forecast month electricity load growth rate data; the user-side electricity load forecast month electricity load metering unit performs electricity load metering processing of the user-side electricity load forecast month based on the electricity load growth rate data of the user-side electricity load forecast month and the user-side target single-month historical electricity load data to generate the user-side electricity load forecast month electricity load data;

[0105] The user-side power load forecast result evaluation module includes a user-side historical single-month extreme power load storage unit, a user-side power load forecast result rationality analysis unit, and a user-side power load forecast result output feedback unit;

[0106] The user-side historical single-month extreme power load storage unit stores the user-side historical single-month extreme power load data based on big data; the user-side power load forecast result rationality analysis unit performs rationality analysis on the user-side power load forecast result based on the user-side power load forecast month power load data and the user-side historical single-month extreme power load data to generate user-side power load forecast result rationality analysis data; the user-side power load forecast result output feedback unit constructs the user-side power load forecast feedback data and executes the user-side power load forecast result feedback operation.

[0107] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The intelligent prediction method of user-side power load is characterized by: The method comprises the following steps: S1. Collect monthly and time data of power load forecast at the user end; The S1 comprises the following steps: S11. Collect the time characteristic text information of the current and next month of the power load forecast of the specific user end online through the data collection dialog box of the distribution network management platform, and generate the monthly time data of the power load forecast of the user end ,in The units include year and month; S2, constructing and processing the historical month object for sampling the historical power load information of the user end according to the predicted monthly time data of the power load of the user end, and generating the monthly time data of sampling the historical power load of the user end; The S2 comprises the following steps: S21, predicting monthly time data based on the user-side power load Construct the historical monthly time characteristic values ​​of the user-side historical power load information sampling, and generate the user-side historical power load sampling month time data set ,in Indicates and The user-side historical power load sampling month and time data for the corresponding historical month of the previous year, Indicates and The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates and The user-side historical power load sampling month and time data corresponding to the previous historical month; Indicates and The user-side historical power load sampling month and time data corresponding to the previous historical month, Indicates and The user-side historical power load sampling month and time data corresponding to the previous historical month; , , , , The units include year and month; S3, searching and processing the user-side target historical power load information according to the user-side historical power load sampling month time data and the user-side historical single-month power load data, and generating the user-side target single-month historical power load data; S4, performing numerical processing on the user-side target historical electricity load growth rate based on the user-side target single-month historical electricity load data to generate the user-side target single-month historical electricity load growth rate data; S5. Perform numerical processing on the power load growth rate of the user-side power load forecast month according to the user-side target single-month historical power load growth rate data, generate power load growth rate data of the user-side power load forecast month, and perform power load metering processing on the user-side power load forecast month in combination with the user-side target single-month historical power load data, generate power load data of the user-side power load forecast month; S6, performing a rationality analysis and processing of the user-side power load forecast result according to the power load data of the user-side power load forecast month and the user-side historical single-month extreme state power load data, generating rationality analysis data of the user-side power load forecast result, and directly ending the current user-side power load forecast operation when it is abnormal; S7. When it is normal, construct the user-side power load forecast feedback data and execute the user-side power load forecast result feedback operation.

2. The method for intelligent prediction of user-side power load according to claim 1, characterized in that: The S3 comprises the following steps: S31. Establish a user-side historical monthly electricity load data set , ;in Indicates the user's The historical monthly electricity load data of the user end corresponding to the historical month, Represents the maximum number of historical months, where The unit is kilowatt; S32, the The user-side historical power load sampling month and time data mentioned in As stated in According to the historical monthly time feature matching, search for the corresponding monthly time data of the historical power load sampling at the user end , and generate a user-side target monthly historical electricity load data set , execute to generate the user-side target single-month historical electricity load data set The specific steps are as follows: S321, initialization parameters, update power load, search for the number of crow populations, maximum number of iterations, flight distance ; S322, initializing the power load to search for crow individuals in the The initial position and memory in the search space, Only the electrical load is used to search for crows that are randomly distributed in a multidimensional search space, i.e. The dimension of the random distribution space of crows searching only by electric load is of In the search space; in the first iteration, it is assumed that the power load search crow individual hides the food in the initial position; that is, in the first iteration, it is assumed that the power load search crow individual compares the user-side historical power load sampling month time data with the The initial position in the search space Make a match; S323, calculate the individual fitness value of each power load search crow, that is, calculate the monthly time data of the user-side historical power load sampling and the The search space The fitness value of S324, update the power load to search for crow individuals in the The position in the search space, i.e., Search the search space to find the one that matches the user-side historical power load sampling month and time data. location; S325, determine the feasibility of the new position, determine the feasibility of the new position of each individual crow using the electric load to search for the crow; if the new position of the individual crow using the electric load to search for the crow is feasible, the individual crow using the electric load to search for the crow will update its position, that is, Search the search space to find the one that matches the user-side historical power load sampling month and time data. , the power load search crow individual is updated to the matching successful Otherwise, the electric load search crow individual stays at the current position and will not move to the new location; S326, evaluate the fitness value of the new position, calculate the fitness value of each individual crow searching for electricity load at the new position, that is, calculate the fitness value of the new position of each crow searching for electricity load. The user-side historical power load sampling month and time data in the search space and the new location The fitness value of S327, updating memory, if the fitness value of the new position of the electric load search crow is greater than the fitness value of the initial position in the memory, the electric load search crow updates its memory through the new position, otherwise it does not update its memory; that is, in the Search for the one with the largest fitness value for the user-side historical power load sampling month and time data ; S328. When the maximum number of iterations is met, output the data that matches the user-side historical power load sampling month and time data. ; S329, the output in step S328 , and generate the user-side target monthly historical electricity load data set through data identification ,in Indicates the The corresponding user-side target monthly historical electricity load data, Indicates the The corresponding user-side target monthly historical electricity load data, Indicates the The corresponding user-side target monthly historical electricity load data, Indicates the The corresponding user-side target monthly historical electricity load data, Indicates the The corresponding user-side target monthly historical electricity load data, including , , , , The unit is kilowatt.

3. The method for intelligent prediction of user-side power load according to claim 2 is characterized in that: The S4 comprises the following steps: S41. According to the The user-side target single-month historical electricity load data is numerically measured and processed to obtain the user-side target single-month historical electricity load growth rate data set. ,in Indicates the Relative to the The user-side target monthly historical electricity load growth rate data, including ; Indicates the Relative to the The user-side target monthly historical electricity load growth rate data, including ; Indicates the Relative to the The user-side target monthly historical electricity load growth rate data, including , , , All figures are rounded off.

4. The method for intelligent prediction of user-side power load according to claim 3 is characterized in that: The S5 comprises the following steps: S51, according to The user-side target monthly historical electricity load growth rate data is combined with the proportional formula to perform numerical measurement processing on the electricity load growth rate of the user-side electricity load forecast month, and generate the electricity load growth rate data of the user-side electricity load forecast month. ,in , Indicates the The corresponding power load is relative to the The growth rate data of electricity load; S52, the With the The power load value of the user-side power load forecast month is measured and processed according to the proportional formula to generate the power load data of the user-side power load forecast month ,in , The unit is kilowatt.

5. The method for intelligent prediction of user-side power load according to claim 4 is characterized in that: The S6 comprises the following steps: S61. Establish a historical monthly extreme power load data set at the user end , ;in Indicates The historical monthly extreme power load data of the user end corresponding to the extreme power state type of the user end, Indicates the maximum value of the extreme state type of the user's power consumption. The unit is kilowatt; S62, using the KD tree nearest neighbor search algorithm to With the As stated in Carry out power load value matching, and generate rationality analysis data of user-side power load forecast results based on the power load value matching results ; when and If the power load value matching is successful, the rationality analysis data of the power load forecast result at the user end will be output. If it is abnormal, the user-side power load forecasting operation is terminated directly; when and If the power load value is not matched successfully, the rationality analysis data of the power load forecast result at the user end will be output. is normal.

6. The method for intelligent prediction of user-side power load according to claim 5 is characterized in that: The S7 comprises the following steps: S71, when the When it is normal, , and stated Combine data to construct user-side power load forecast feedback data ; S72, the Feedback is pushed to the distribution network management platform through the IoT communication network.

7. A user-side power load intelligent prediction system based on big data, used to implement the user-side power load intelligent prediction method according to any one of claims 1 to 6, characterized in that: The system comprises a user-side historical power load information acquisition module, a user-side power load prediction and metering module, and a user-side power load prediction result evaluation module.

Images