A power load forecasting method and system based on big data drive

Abstract

The present invention relates to a power load forecasting method and system driven by big data. The method includes the following steps: obtaining regional power station data; dividing the regional power station data by regional electricity consumption types to generate regional power consumption type data; evaluating the regional power complexity of the regional power station data according to the regional power consumption type data to generate regional power complexity data; performing time-series load analysis on the regional power complexity data according to a preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data. The present invention improves the accuracy and reliability of power load forecasting by combining data classification, complexity evaluation, population density analysis and real-time historical data.

Description

Technical Field

[0001] The present invention relates to the technical field of electric load forecasting, and particularly to a big data-driven electric load forecasting method and system. Background Art

[0002] With the improvement of computer technology and data acquisition capabilities, big data technology has gradually been applied to the field of electric load forecasting. The operation data of the power system is huge and diverse, covering multiple aspects such as weather, holidays, and social and economic activities. By deeply analyzing this data, potential factors affecting load changes can be mined. Therefore, the big data-based electric load forecasting method has emerged. In recent years, the rapid development of machine learning and deep learning technologies has brought new opportunities to electric load forecasting. In particular, algorithms such as neural networks, support vector machines, and random forests can establish more accurate forecasting models through training on historical load data and influencing factor data. For example, the long short-term memory network (LSTM) performs well in processing time series data and can effectively capture long-term dependencies. However, currently, traditional load forecasting often does not carefully distinguish different types of electricity consumption, resulting in poor generalization ability of the load forecasting model. At the same time, it lacks targeted processing for different complexity load situations, leading to low accuracy in peak load and base load forecasting, and thus lower overall forecasting accuracy. Summary of the Invention

[0003] Based on this, it is necessary to provide a big data-driven electric load forecasting method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a big data-driven electric load forecasting method includes the following steps:

[0005] Step S1: Obtain regional power station data; divide the regional power station data by regional electricity consumption types to generate regional power consumption type data; evaluate the regional power complexity of the regional power station data according to the regional power consumption type data to generate regional power complexity data;

[0006] Step S2: Perform time-series load analysis on the regional power complexity data according to a preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the high-complexity regional power peak load data and the low-complexity regional power base load data in time to obtain regional power connection period data;

[0007] Step S3: Obtain regional population density data; conduct regional power demand analysis on the regional power connection period data based on the regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate the regional power fluctuation stability of the regional power demand data to generate regional connection period power fluctuation stability data;

[0008] Step S4: Collect historical power load data and real-time power load data of power stations according to the regional power connection period data; conduct power load prediction on the real-time power load data of power stations based on the historical power load data of power stations to obtain a power load prediction curve; adjust the curve curvature of the power load prediction curve through the regional connection period power fluctuation stability data to generate an optimized power load prediction curve.

[0009] The present invention collects basic data of all power stations in a certain region, such as geographical location, capacity, output, etc. Classify these station data according to the specific types of power use (such as residential, industrial, commercial, etc.). Evaluate the power complexity of each region according to factors such as the diversity of power usage types and the distribution of stations to generate complexity data. Analyze the complexity data according to a preset standard complexity threshold. Generate peak power load data for high-complexity regions and base power load data for low-complexity regions according to the high or low of the complexity data. Synchronize the peak load and base load data in time to obtain power connection period data. Collect relevant data on the population density in the region. Based on the population density data, analyze the power connection period data to generate power demand data. Use the fast Fourier transform (FFT) method to evaluate the stability of the power demand data to generate power fluctuation stability data. According to the power connection period data, collect historical power station load data and real-time load data. Use the historical load data to predict the real-time load data to generate a prediction curve. Based on the fluctuation stability data, adjust the curve curvature of the prediction curve to generate an optimized power load prediction curve. By using multi-dimensional data and complex analysis methods, the accuracy of the prediction is significantly improved. By analyzing power complexity and demand fluctuations, rationally allocate power resources to avoid waste and overload. Through in-depth analysis of power demand fluctuations and optimized adjustment of load curves, enhance the stability and reliability of the power grid. Therefore, the present invention improves the accuracy and reliability of power load prediction by combining data classification, complexity evaluation, population density analysis and real-time historical data.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Obtain regional power station data;

[0012] Step S12: Perform data preprocessing on the regional power station data to generate standard regional power station data, where the data preprocessing includes data cleaning, data outlier repair, and data standardization;

[0013] Step S13: Conduct regional electricity consumption type classification on the standard regional power station data to generate regional power electricity consumption type data, where the regional electricity consumption type classification includes regional industrial electricity consumption classification, regional agricultural electricity consumption classification, regional transportation electricity consumption classification, and regional residential electricity consumption classification;

[0014] Step S14: Evaluate the regional power complexity of the standard regional power station data based on the regional power electricity consumption type data to generate regional power complexity data.

[0015] The present invention obtains power-related data from each power station within the region, including but not limited to power load, power generation, power consumption, equipment status, environmental parameters, etc. Clean the noise, missing values, and outliers in the data to ensure data quality. Repair outliers through methods such as interpolation and filling to ensure the continuity and consistency of the data. Standardize the data to unify the data format and dimension and prepare for subsequent analysis. Identify and classify the electricity consumption data of industrial enterprises and analyze the characteristics and patterns of industrial electricity consumption. Identify and classify the electricity consumption data of agricultural production and analyze the characteristics and patterns of agricultural electricity consumption. Identify and classify the electricity consumption data of the transportation system and analyze the characteristics and patterns of transportation electricity consumption. Identify and classify the electricity consumption data of residential life and analyze the characteristics and patterns of residential electricity consumption. Based on the data of different electricity consumption types, evaluate the complexity of the regional power system, including factors such as load volatility, diversity of electricity demand, and seasonal changes, to generate regional power complexity data. Through data cleaning and outlier repair, ensure the accuracy and integrity of the data and provide a reliable data basis for subsequent analysis and prediction. Data standardization makes the data from different sources comparable and consistent, facilitating subsequent comprehensive analysis. Through the detailed classification of electricity consumption types, it is possible to deeply understand the electricity consumption needs and characteristics of various users and provide support for formulating refined power dispatching strategies. Evaluating the complexity of the regional power system helps identify the weak links and potential risks of the system and optimize the operation and management of the power system. The generated standard regional power station data and regional power complexity data provide a scientific basis for the intelligent management and dispatching of the power system, improving the accuracy and effectiveness of decision-making.

[0016] Preferably, step S2 includes the following steps:

[0017] Step S21: Compare the regional power complexity data with a preset standard regional power complexity threshold. When the regional power complexity data is greater than or equal to the preset standard regional power complexity threshold, perform short-term peak load analysis on the corresponding regional power station data to generate high-complexity regional power peak load data;

[0018] Step S22: When the regional power complexity data is less than the preset standard regional power complexity threshold, perform long-term base load analysis on the corresponding regional power station data to generate low-complexity regional power base load data;

[0019] Step S23: Perform power curve conversion on the high-complexity regional power peak load data and the low-complexity regional power base load data to generate a high-complexity regional power time series curve and a low-complexity regional power time series curve;

[0020] Step S24: Overlap the high-complexity regional power time series curve and the low-complexity regional power time series curve to generate a regional power time series overlap curve; synchronize the time of the regional power time series overlap curve to obtain regional power connection period data.

[0021] By performing real-time evaluation and classification analysis on the regional power complexity, the present invention can dynamically respond to changes in power load, timely identify peak load and base load, and provide accurate data support for power dispatch. Based on the differentiated processing of short-term peak load and long-term base load, the accuracy of power load analysis is improved, ensuring effective power load prediction and management under different complexity conditions. The generated high-complexity regional power time series curve and low-complexity regional power time series curve, as well as the final regional power time series overlap curve, provide detailed time series data for the optimal dispatch of the power system, facilitating scientific decision-making and resource allocation. By analyzing the regional power connection period data, key linkage periods in the power system can be identified, which helps to enhance the stability and reliability of the system and reduce the risks brought by load fluctuations. This refined analysis method and the generated data support the management and optimization of smart grids, can better cope with complex power demand and supply environments, and improve the overall operation efficiency of the power system.

[0022] Preferably, performing short-term peak load analysis on the corresponding regional power station data includes:

[0023] Perform short-term level time series analysis on the corresponding regional power station data to generate first regional power station short-term time series data, where the short-term level time series analysis includes minute level, hour level, and intraday level;

[0024] Perform load fluctuation analysis on the short-term time series data of the first regional power station to generate short-term time series load fluctuation data of the first regional power station; perform load mean calculation on the short-term time series load fluctuation data of the first regional power station to obtain short-term time series intermediate load data of the first regional power station;

[0025] Extract peak loads from the short-term time series load fluctuation data of the first regional power station according to the short-term time series intermediate load data of the first regional power station to generate high-complexity regional power peak load data.

[0026] Through time series analysis at the minute, hour, and intraday levels, the present invention can accurately identify and record peak hours of electricity load, providing refined load management data. Through detailed analysis of load fluctuations, the changing patterns and fluctuation characteristics of electricity load can be identified, providing a scientific basis for the dispatching and optimization of the power system. Through mean calculation, a reference benchmark for electricity load is provided, which helps to identify anomalies and peak loads in the fluctuation data. The generated high-complexity regional power peak load data provides key load information for the power dispatching department, helping it to take corresponding measures during peak load hours to ensure the stable operation of the power system. This meticulous analysis and data processing method enhance the intelligent management ability of the power system, enabling it to better cope with complex and changing electricity demands and improving the reliability and efficiency of power supply.

[0027] Preferably, the long-term base load analysis of the corresponding regional power station data includes:

[0028] Perform long-term time series analysis on the corresponding regional power station data to generate long-term time series data of the second regional power station, where the long-term time series analysis includes weekly, monthly, and yearly levels;

[0029] Perform load fluctuation analysis on the long-term time series data of the second regional power station to generate long-term time series load fluctuation data of the second regional power station; perform load mean calculation on the long-term time series load fluctuation data of the second regional power station to obtain long-term time series intermediate load data of the second regional power station;

[0030] Extract base loads from the long-term time series load fluctuation data of the second regional power station according to the long-term time series intermediate load data of the second regional power station to generate low-complexity regional power base load data.

[0031] The present invention captures the changing trends and patterns of weekly electricity loads through weekly time-series analysis of electricity site data. It conducts monthly time-series analysis of electricity site data to identify the changing patterns and periodicity of monthly electricity loads. It performs annual time-series analysis of electricity site data to analyze the changing patterns of electricity loads at different times throughout the year and identify seasonal and long-term trends. By calculating statistics such as the standard deviation and variance of the load, it conducts load fluctuation analysis on long-term time-series data to identify the amplitude and frequency of electricity load fluctuations. It generates time-series data containing electricity load fluctuation information to provide a basis for subsequent analysis. It calculates the mean of the load fluctuation data to obtain the average load values for different time periods. It generates long-term time-series intermediate load data as a benchmark for measuring the load level. It compares the load fluctuation data with the intermediate load data to extract stable and representative basic load data. It generates data containing basic load information for analyzing the operation of the power system in a low-complexity state. Through weekly, monthly, and annual time-series analysis, it can identify and record the long-term changing trends and periodicity of electricity loads, providing a reliable basis for electricity load forecasting. Through load fluctuation analysis of long-term time-series data, it can identify the changing patterns and fluctuation characteristics of electricity loads, providing a scientific basis for the scheduling and optimization of the power system. Through mean calculation, it provides a reference benchmark for electricity loads, helping to identify the basic load level in the fluctuation data. The generated base load data for the low-complexity area of the power grid provides key load information for the power dispatching department, helping it to take corresponding measures during the base load period to ensure the stable operation of the power system. This detailed analysis and data processing method enhances the intelligent management ability of the power system, enabling it to better respond to complex and changing electricity demands and improving the reliability and efficiency of electricity supply.

[0032] Preferably, step S3 includes the following steps:

[0033] Step S31: Obtain the regional population density data;

[0034] Step S32: Based on the regional population density data, conduct regional electricity demand analysis on the regional electricity connection time period data to generate regional electricity demand data; conduct demand pattern recognition on the regional electricity demand data to generate regional electricity ordinary demand pattern data and regional electricity emergency demand pattern data;

[0035] Step S33: Use the fast Fourier transform method to conduct hierarchical electricity load fluctuation frequency analysis on the regional electricity ordinary demand pattern data and the regional electricity emergency demand pattern data to generate regional electricity ordinary demand fluctuation frequency data and regional electricity emergency demand fluctuation frequency data;

[0036] Step S34: Perform regional power fluctuation stability assessment on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to generate regional power fluctuation stability data during the regional connection period.

[0037] By combining the regional population density data and the power connection period data, the present invention can accurately predict the regional power demand, identify the ordinary and emergency demand patterns, and provide a scientific basis for the rational allocation of power resources. Using the fast Fourier transform method, it is possible to deeply analyze the fluctuation frequency of the power load, reveal the power fluctuation characteristics under different demand patterns, and provide detailed data support for the optimal dispatching of the power system. Through the stability assessment of the fluctuation frequency data, potential unstable factors in the power system can be identified, and measures can be taken in advance to ensure the stable operation of the power system. This meticulous demand analysis and fluctuation frequency analysis method enhance the intelligent management ability of the power system, can better respond to the complex and changeable power demand, and improve the reliability and efficiency of power supply. The generated regional power demand data and fluctuation stability data provide key load information for the power dispatching department, helping it to take corresponding measures under different demand patterns to ensure the stable operation of the power system and the optimal allocation of resources.

[0038] Preferably, step S34 includes the following steps:

[0039] Step S341: Calculate the mean value of the fluctuation amplitude of the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to obtain the regional power demand fluctuation amplitude reference data;

[0040] Step S342: Based on the regional power demand fluctuation amplitude reference data, perform regional power fluctuation amplitude overlimit analysis on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to generate regional power demand fluctuation amplitude overlimit data and regional power emergency demand fluctuation overlimit data;

[0041] Step S343: According to the regional power demand fluctuation amplitude overlimit data and the regional power emergency demand fluctuation overlimit data, perform regional power fluctuation stability assessment to generate regional power fluctuation stability data during the regional connection period.

[0042] By calculating the mean value of the fluctuation amplitude, the present invention can accurately evaluate the fluctuation amplitude of power demand, providing a scientific basis for identifying abnormal fluctuations. Through the over-limit analysis of the fluctuation frequency data of ordinary demand and emergency demand, the abnormal fluctuations of power demand can be identified, and potential problems can be discovered and handled in a timely manner. Through the stability assessment, the unstable factors in the power system can be identified, and measures can be taken in advance to ensure the stable operation of the power system. This meticulous fluctuation analysis and stability assessment method enhance the intelligent management ability of the power system, enabling it to better cope with complex and changing power demands, and improving the reliability and efficiency of power supply. The generated stability data provides key load information for the power dispatching department, helping it to take corresponding measures under different demand patterns to ensure the stable operation of the power system and the optimal allocation of resources.

[0043] Preferably, step S4 includes the following steps:

[0044] Step S41: Collect time-series power loads according to the regional power connection period data to obtain historical power station power load data and real-time power station power load data;

[0045] Step S42: Use machine learning methods to predict the real-time power load data of the power station based on the historical power station power load data to obtain power load prediction data;

[0046] Step S43: Visualize the power load prediction data to generate a power load prediction curve; adjust the curve curvature of the power load prediction curve through the regional connection period power fluctuation stability data to generate an optimized power load prediction curve.

[0047] Through machine learning methods, the present invention can accurately predict power loads based on historical data and real-time data, improving the accuracy and reliability of the prediction. Through time-series power load collection, the changes in power loads are monitored in real time, providing data support for timely adjustment and response. The generated power load prediction curve makes the prediction results more intuitive, facilitating analysis and decision-making, and improving the efficiency of power management. Through curve curvature adjustment, combined with power fluctuation stability data, the prediction results can be optimized, enhancing the stability and response ability of the power system. This data-driven prediction and optimization method enhances the intelligent management ability of the power system, enabling it to better cope with complex and changing power demands, and improving the reliability and efficiency of power supply.

[0048] Preferably, step S42 includes the following steps:

[0049] Step S421: Divide the historical power station power load data into a model training set and a model test set;

[0050] Step S422: Use a linear programming algorithm to train the model training set to generate a power load training model; perform model testing and iteration on the power load training model through the model test set to generate a power load prediction model.

[0051] Step S423: Import the real-time power load data of the power station into the power load prediction model for power load prediction to generate power load prediction data.

[0052] In the present invention, the historical power load data of the power station is divided into a model training set and a model test set, usually divided according to a ratio (such as 70:30 or 80:20), to ensure that the data of the training set and the test set are representative and independent. Use a linear programming algorithm to train the model training set to generate a preliminary power load training model. The linear programming algorithm finds the optimal parameters for power load prediction by optimizing the objective function. Use the model test set to test the power load training model and evaluate the prediction performance of the model. By iteratively adjusting the model parameters, the prediction accuracy of the model is improved, and finally an optimized power load prediction model is generated. Import the real-time load data of the power station into the power load prediction model that has been trained and tested and optimized. Use the prediction model to predict the real-time data to generate power load prediction data, providing prediction information on future power demand. By reasonably dividing the historical data, the independence of model training and testing is ensured, and the generalization ability and prediction accuracy of the model are improved. Use a linear programming algorithm for model training, and by iteratively optimizing the model parameters, the prediction performance of the model is improved to generate an accurate power load prediction model. By importing real-time data into the prediction model, power load prediction data can be generated in real time, reflecting the change of power demand in a timely manner, and supporting dynamic power dispatching and management. Through iterative optimization and testing, the high accuracy and reliability of the prediction model are ensured, providing a scientific basis for the stable operation of the power system. This prediction method based on historical data and real-time data enhances the intelligent management ability of the power system, can better cope with the complex and changeable power demand, and improve the reliability and efficiency of power supply.

[0053] In this specification, a power load prediction system based on big data drive is provided, including:

[0054] A power complexity analysis module, configured to obtain regional power station data; perform regional power consumption type division on the regional power station data to generate regional power consumption type data; evaluate the regional power complexity of the regional power station data according to the regional power consumption type data to generate regional power complexity data.

[0055] The peak load analysis module is used to perform time-series load analysis on regional power complexity data according to a preset standard regional power complexity threshold, generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the time of the high-complexity regional power peak load data and the low-complexity regional power base load data to obtain regional power connection period data;

[0056] The power demand analysis module is used to obtain regional population density data; perform regional power demand analysis on regional power connection period data based on the regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate the regional power fluctuation stability of the regional power demand data to generate regional connection period power fluctuation stability data;

[0057] The load forecasting module is used to collect time-series power loads according to the regional power connection period data to obtain historical power station power load data and real-time power station power load data; perform power load forecasting on the real-time power station power load data based on the historical power station power load data to obtain a power load forecasting curve; adjust the curve curvature of the power load forecasting curve through the regional connection period power fluctuation stability data to generate an optimized power load forecasting curve.

[0058] The beneficial effects of the present invention are as follows: By collecting relevant data of all power stations in a specific area, classifying these data to determine different types of electricity consumption demands, such as industrial electricity consumption, commercial electricity consumption, and residential electricity consumption, etc., generating a dataset of electricity consumption types according to the classification results. Using the electricity consumption type data to evaluate the complexity of the regional power station data, and the evaluation criteria can include electricity consumption, diversity of electricity consumption patterns, etc., generating regional power complexity data. Analyzing the complexity data to identify the peak load in high-complexity regions and the base load in low-complexity regions. Synchronizing the high-complexity and low-complexity data in time to obtain regional power connection period data. Collecting regional population density data related to electricity demand. Based on the population density data, analyzing the connection period data to generate electricity demand data. Using FFT to evaluate the fluctuation stability of the electricity demand data, generating regional connection period power fluctuation stability data. Based on the regional power connection period data, collecting historical power station load data and real-time load data. Using the historical load data to predict the real-time load data to obtain a power load prediction curve. Adjusting the prediction curve through the fluctuation stability data to generate an optimized power load prediction curve. Through multi-dimensional data analysis and complexity evaluation, providing a more accurate power load prediction. It helps to efficiently allocate power resources, ensure stable and efficient power supply. Through the fluctuation stability evaluation, reducing the impact of power fluctuations on the system and enhancing the stability and reliability of the overall system. Therefore, the present invention combines data classification, complexity evaluation, population density analysis, and real-time and historical data to conduct power load prediction, improving the accuracy and reliability of power load prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the step flow of a power load prediction method driven by big data;

[0060] Figure 2 is Figure 1 a schematic diagram of the detailed implementation steps of step S2 in

[0061] Figure 3 is Figure 1 a schematic diagram of the detailed implementation steps of step S3 in

[0062] Figure 4 is Figure 1 a schematic diagram of the detailed implementation steps of step S4 in

[0063] The realization, functional features, and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] The technical method of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work belong to the scope of protection of the present invention.

[0065] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0066] It should be understood that although terms such as "first" and "second" may be used here to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit can be called the second unit, and similarly the second unit can be called the first unit. The term "and / or" used here includes any and all combinations of one or more of the listed associated items.

[0067] To achieve the above object, please refer to Figures 1 to 4 , a power load forecasting method based on big data drive, the method includes the following steps:

[0068] Step S1: Obtain regional power station data; divide the regional power station data by regional electricity consumption types to generate regional power consumption type data; evaluate the regional power complexity of the regional power station data according to the regional power consumption type data to generate regional power complexity data;

[0069] Step S2: Perform time-series load analysis on the regional power complexity data according to a preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the high-complexity regional power peak load data and the low-complexity regional power base load data in time to obtain regional power connection period data;

[0070] Step S3: Obtain regional population density data; perform regional power demand analysis on the regional power connection period data based on the regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate the regional power fluctuation stability of the regional power demand data to generate regional connection period power fluctuation stability data;

[0071] Step S4: Collect time-series power load according to the regional power connection period data to obtain historical power load data of power stations and real-time power load data of power stations; perform power load forecasting on the real-time power load data of power stations based on the historical power load data of power stations to obtain a power load forecasting curve; adjust the curve curvature of the power load forecasting curve through the regional connection period power fluctuation stability data to generate an optimized power load forecasting curve.

[0072] The present invention collects basic data of all power stations in a certain area, such as geographical location, capacity, output, etc. Classify these station data according to the specific types of power use (such as residential, industrial, commercial, etc.). Evaluate the power complexity of each area according to factors such as the diversity of electricity consumption types and the distribution of stations to generate complexity data. Analyze the complexity data according to a preset standard complexity threshold. Generate peak power load data for high-complexity areas and base power load data for low-complexity areas according to the level of the complexity data. Synchronize the peak load and base load data in terms of time to obtain power connection period data. Collect relevant data on population density in the area. Analyze the power connection period data based on the population density data to generate power demand data. Use the fast Fourier transform (FFT) method to evaluate the stability of the power demand data and generate power fluctuation stability data. Collect historical power station load data and real-time load data according to the power connection period data. Use the historical load data to forecast the real-time load data and generate a forecasting curve. Based on the fluctuation stability data, adjust the curve curvature of the forecasting curve to generate an optimized power load forecasting curve. By using multi-dimensional data and complex analysis methods, the accuracy of forecasting is significantly improved. By analyzing power complexity and demand fluctuations, power resources are rationally allocated to avoid waste and overload. Through in-depth analysis of power demand fluctuations and optimized adjustment of load curves, the stability and reliability of the power grid are enhanced. Therefore, the present invention improves the accuracy and reliability of power load forecasting by combining data classification, complexity evaluation, population density analysis and real-time historical data.

[0073] In the embodiment of the present invention, refer to Figure 1 As described, it is a schematic diagram of the step flow of a power load forecasting method based on big data drive of the present invention. In this example, the power load forecasting method based on big data drive includes the following steps:

[0074] Step S1: Obtain regional power station data; divide the regional power station data by regional power use types to generate regional power use type data; evaluate the regional power complexity of the regional power station data according to the regional power use type data to generate regional power complexity data;

[0075] In the embodiment of the present invention, by extracting power station data of a specific area from the power company database, including information such as station location, capacity, type, equipment configuration, etc., to ensure the integrity and accuracy of the data, and performing data cleaning and correction when necessary. Classify the stations according to the type of electricity consumption (such as residential electricity consumption, industrial electricity consumption, commercial electricity consumption, agricultural electricity consumption, etc.). Use historical electricity consumption data and user type information, and through clustering analysis or other classification algorithms, generate regional power consumption type data. Integrate the electricity consumption type information of each station to form a complete regional power consumption type data set. The data set should contain detailed information such as the electricity consumption type, electricity consumption, and electricity consumption time period of each station. Define complexity evaluation indicators, such as electricity consumption type diversity, electricity consumption volatility, load stability, equipment complexity, etc. Adopt the weighted scoring method or other evaluation methods to evaluate the complexity of each power station in the region according to each indicator. Summarize the complexity evaluation results to form regional power complexity data. The data should contain the complexity score of each power station and the corresponding evaluation indicator weights.

[0076] Step S2: Perform a time-series load analysis on the regional power complexity data according to a preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the time of the high-complexity regional power peak load data and the low-complexity regional power base load data to obtain regional power connection period data;

[0077] In the embodiment of the present invention, by setting one or more standard complexity thresholds according to historical data and expert suggestions to distinguish high-complexity and low-complexity regions. For example, set the threshold to 0.6, those with a complexity score higher than 0.6 are high-complexity regions, and those lower than 0.6 are low-complexity regions. Perform a time-series load analysis on the regional power complexity data to extract load data at different time periods. Use historical load data, combined with the complexity score, to analyze the load change trend at different time periods. According to the complexity threshold, divide the regional power load data into high-complexity regional power peak load data and low-complexity regional power base load data. The high-complexity regional power peak load data includes the load data of stations with a complexity score higher than the threshold during high-load periods. The low-complexity regional power base load data includes the load data of stations with a complexity score lower than the threshold during low-load periods. Perform a time synchronization analysis on the high-complexity regional power peak load data and the low-complexity regional power base load data. Find the load data of the high-complexity region and the low-complexity region within the same time period to generate regional power connection period data.

[0078] Step S3: Obtain regional population density data; conduct regional power demand analysis on the regional power connection period data based on the regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate the regional power fluctuation stability of the regional power demand data to generate regional connection period power fluctuation stability data;

[0079] In the embodiment of the present invention, by obtaining the latest regional population density data, including information such as population distribution and residential density. Ensure the accuracy and timeliness of the data, and perform necessary data cleaning and preprocessing. Combine the population density data with the regional power connection period data to analyze the power demand in different regions. Assume that the higher the population density in a region, the greater the power demand. Establish a regional power demand model according to the relationship between population density and power load. Calculate the power demand in different time periods and different regions according to the regional power demand model. Generate a regional power demand data set containing information such as timestamp, region, and demand. Use the fast Fourier transform (FFT) method to perform frequency domain analysis on the regional power demand data. Analyze the frequency components of the power demand and evaluate the stability of the power fluctuation. Generate the power fluctuation stability data for the regional connection period according to the FFT analysis result. This data set should contain information such as timestamp, region, and fluctuation stability index.

[0080] Step S4: Collect time-series power loads according to the regional power connection period data to obtain historical power station power load data and real-time power station power load data; conduct power load prediction on the real-time power station power load data based on the historical power station power load data to obtain a power load prediction curve; adjust the curvature of the power load prediction curve through the regional connection period power fluctuation stability data to generate an optimized power load prediction curve.

[0081] In the embodiment of the present invention, obtain historical power load data and real-time power load data from the monitoring system of the power station. Ensure that the data collection frequencies are consistent, and process any missing values or outliers. Use the historical power load data to train a prediction model (such as ARIMA, LSTM, etc.). Take the real-time power load data as input and use the prediction model to generate a power load prediction curve for a future period of time. Adjust the power load prediction curve according to the regional connection period power fluctuation stability data. Consider the stability index in the fluctuation stability data to optimize the curvature of the prediction curve to reduce the prediction error.

[0082] Preferably, step S1 includes the following steps:

[0083] Step S11: Obtain regional power station data;

[0084] Step S12: performing data preprocessing on the regional power site data to generate standard regional power site data, wherein the data preprocessing includes data cleaning, data outlier repair and data standardization;

[0085] Step S13: dividing the standard regional power site data into regional power consumption types to generate regional power consumption type data, wherein the regional power consumption type division includes regional industrial power consumption division, regional agricultural power consumption division, regional transportation power consumption division and regional living power consumption division;

[0086] Step S14: Perform regional power complexity evaluation on standard regional power site data according to regional power consumption type data to generate regional power complexity data.

[0087] In an embodiment of the present invention, the power site data of a specific area is extracted from the power company database. The data includes but is not limited to site location, capacity, type, equipment configuration, historical power consumption data, etc. Ensure the integrity and accuracy of the data. Check the missing values, duplicate values ​​and outliers in the data. Delete or fill in missing values ​​and remove duplicate records. Use statistical methods (such as box plot analysis) to identify and process outliers. Use interpolation, mean substitution or regression models to repair outliers. Ensure that the repaired data conforms to the actual situation and business logic. Standardize the numerical data (such as normalization or Z-score standardization) to eliminate the dimensional differences between different features. Use appropriate standardization methods to make the data evenly distributed and improve the model training effect. According to the distribution and power consumption records of industrial enterprises, classify the records related to industrial power consumption in the power site data. Use clustering algorithms (such as K-means) to segment industrial electricity consumption and identify different types of industrial electricity consumption characteristics. Specifically, use clustering algorithms to divide industrial electricity consumption data into different clusters. The characteristics of different types of industrial electricity consumption can be described in the following ways: Perform statistical analysis on each cluster, such as calculating the center of each cluster (centroid) to describe its average power, peak power, power consumption time distribution and other characteristics. Use mathematical formulas to describe the characteristics of different clusters, such as mean, variance, distribution function, etc. For example, suppose there are two clusters and , its characteristics can be described as: and Respectively represent clusters and The average electricity consumption of and Respectively represent clusters and The variance of electricity consumption. According to agricultural facilities and electricity consumption records, classify the records related to agricultural electricity consumption in the power station data. Analyze the time series characteristics of agricultural electricity consumption to identify the electricity consumption patterns in different seasons and production stages. According to transportation infrastructure (such as electric vehicle charging stations, traffic lights, etc.) and electricity consumption records, classify the records related to transportation electricity consumption in the power station data. Identify the electricity consumption characteristics during peak and off-peak traffic periods. According to the distribution of residential areas and electricity consumption records, classify the records related to domestic electricity consumption in the power station data. Analyze the daily patterns of domestic electricity consumption to identify the electricity consumption characteristics during morning and evening peak hours and on weekends. Determine the key indicators for evaluating the power complexity of the assessment area, such as the diversity of electricity consumption types, electricity consumption volatility, load stability, equipment complexity, etc. Each indicator should have a clear calculation method and weight. For each power station, calculate the complexity score based on its electricity consumption type data. Use the weighted scoring method or other evaluation methods to aggregate the scores of each indicator into a total complexity score. Aggregate the complexity scores of each power station to form a regional power complexity dataset. This dataset should include the complexity score of each station and the relevant indicator information.

[0088] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0089] Step S21: Compare the regional power complexity data with a preset standard regional power complexity threshold. When the regional power complexity data is greater than or equal to the preset standard regional power complexity threshold, perform a short-term peak load analysis on the corresponding regional power station data to generate high-complexity regional power peak load data;

[0090] Step S22: When the regional power complexity data is less than the preset standard regional power complexity threshold, perform a long-term base load analysis on the corresponding regional power station data to generate low-complexity regional power base load data;

[0091] Step S23: Perform a power curve conversion on the high-complexity regional power peak load data and the low-complexity regional power base load data to generate a high-complexity regional power time series curve and a low-complexity regional power time series curve;

[0092] Step S24: Overlap the high-complexity regional power time series curve and the low-complexity regional power time series curve to generate a regional power time series overlap curve; synchronize the time of the regional power time series overlap curve to obtain the regional power connection period data.

[0093] In the embodiments of the present invention, one or more standard thresholds are set according to historical data and expert opinions. For example, the threshold is set to 0.6. The regional power complexity data is compared with the preset standard threshold to determine whether it belongs to a high-complexity region. When the regional power complexity data is greater than or equal to the preset threshold, short-term peak load analysis is performed. Short-term load analysis is carried out on the power station data in the high-complexity region to identify the peak load periods. High-complexity region power peak load data is generated, including the peak load values and corresponding stations for each time period. When the regional power complexity data is less than the preset standard threshold, long-term base load analysis is performed. Long-term load analysis is carried out on the power station data in the low-complexity region to identify the base load periods. Low-complexity region power base load data is generated, including the base load values and corresponding stations for each time period. Time series curve conversion is performed on the high-complexity region power peak load data and the low-complexity region power base load data. High-complexity region power time series curves and low-complexity region power time series curves are generated to show the power load changes in different time periods. The high-complexity region power time series curves and the low-complexity region power time series curves are overlapped to generate a regional power time series overlap curve. Through superposition analysis, the high- and low-complexity power load characteristics in the same time period are identified. Time synchronization processing is performed on the regional power time series overlap curve to identify the connected time periods of the power load. Regional power connected time period data is generated, including the power load characteristics during the connected time periods.

[0094] Preferably, performing short-term peak load analysis on the corresponding regional power station data includes:

[0095] Performing short-term level time series analysis on the corresponding regional power station data to generate first regional power station short-term time series data, where the short-term level time series analysis includes minute level, hour level, and intraday level;

[0096] Performing load fluctuation analysis on the first regional power station short-term time series data to generate first regional power station short-term time series load fluctuation data; calculating the load mean value of the first regional power station short-term time series load fluctuation data to obtain first regional power station short-term time series intermediate load data;

[0097] Extracting the peak load from the first regional power station short-term time series load fluctuation data according to the first regional power station short-term time series intermediate load data to generate high-complexity region power peak load data.

[0098] In the embodiments of the present invention, by performing time segmentation on the data of regional power stations, time series data at the minute level, hour level, and intraday level are generated. The segmented data is stored as different time series data sets. Load fluctuation calculation is performed on each time series data set to generate short-term time series load fluctuation data of the first regional power station. The load mean value is calculated for each load fluctuation data set to generate short-term time series intermediate load data of the first regional power station. Peak load is extracted according to the intermediate load data to generate high-complexity regional power peak load data. Specifically, the resample method of Pandas is used to resample the data, and the average load data at the minute level, hour level, and intraday level are calculated respectively. The diff method is used to calculate the load change amount between adjacent time points to generate load fluctuation data. The mean value of the load fluctuation data is calculated to obtain the intermediate load data. The load fluctuation data and the intermediate load data are compared, and the data with load fluctuation greater than the intermediate load is extracted as the peak load data.

[0099] Preferably, the long-term base load analysis of the corresponding regional power station data includes:

[0100] Perform long-term time series analysis on the corresponding regional power station data to generate long-term time series data of the second regional power station, where the long-term time series analysis includes weekly level, monthly level, and yearly level;

[0101] Perform load fluctuation analysis on the long-term time series data of the second regional power station to generate long-term time series load fluctuation data of the second regional power station; calculate the load mean value of the long-term time series load fluctuation data of the second regional power station to obtain long-term time series intermediate load data of the second regional power station;

[0102] Extract the basic load from the long-term time series load fluctuation data of the second regional power station according to the long-term time series intermediate load data of the second regional power station to generate low-complexity regional power base load data.

[0103] In the embodiments of the present invention, time resampling of data is performed by using the resample method of the Pandas library to generate time series data at the weekly, monthly, and annual levels. The diff method is used to calculate the load difference between adjacent time points to obtain load fluctuation data. The mean method is used to calculate the mean value of the load fluctuation data. The load fluctuation data is compared with the intermediate load data, and the load values less than or equal to the intermediate load are extracted as the basic load. Specifically, the resample method of Pandas is used to resample the data, and the average load data at the weekly, monthly, and annual levels are calculated respectively. The diff method is used to calculate the load change amount between adjacent time points to generate load fluctuation data. The mean value of the load fluctuation data is calculated to obtain the intermediate load data of the long-term time series. The load fluctuation data is compared with the intermediate load data, and the data with load fluctuations less than or equal to the intermediate load is extracted as the basic load data.

[0104] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0105] Step S31: Obtain regional population density data;

[0106] Step S32: Based on the regional population density data, perform regional power demand analysis on the regional power connection time period data to generate regional power demand data; perform demand pattern recognition on the regional power demand data to generate regional power ordinary demand pattern data and regional power emergency demand pattern data;

[0107] Step S33: Use the fast Fourier transform method to perform hierarchical power load fluctuation frequency analysis on the regional power ordinary demand pattern data and the regional power emergency demand pattern data to generate regional power ordinary demand fluctuation frequency data and regional power emergency demand fluctuation frequency data;

[0108] Step S34: Perform regional power fluctuation stability assessment on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to generate regional connection time period power fluctuation stability data.

[0109] In an embodiment of the present invention, by obtaining population density information at different geographical locations in a region, as an important input variable for power demand analysis, the information is obtained specifically through census data, demographic data or satellite image analysis, and is collected at the granularity of regional units (such as cities, counties, townships, etc.). Population density data is correlated with historical power usage data to understand the power demand trends in different regions. Based on historical data and prediction models, regional power demand data is generated, including predictions of power demand for each time period. The generated regional power demand data is analyzed to identify normal demand patterns and emergency demand patterns. The normal demand pattern is specifically a typical demand pattern on daily working days or non-working days, and the emergency demand pattern is specifically a demand pattern under emergencies or extreme weather conditions. The fast Fourier transform (FFT) method is used for frequency analysis to evaluate the fluctuation characteristics of power demand data. The normal demand pattern data and the emergency demand pattern data are respectively input into the FFT algorithm to obtain their spectrum information, where the mathematical formula of the FFT algorithm is as follows: ,in is the input signal of discrete time series, is the corresponding frequency domain representation, is the length of the sequence (assuming ), It is an imaginary unit. Through spectrum analysis, the main fluctuation frequency and amplitude of power load are determined. The stability of regional power fluctuation is evaluated, that is, the fluctuation degree and frequency stability of power load under different demand modes. The results of ordinary demand fluctuation frequency data and emergency demand fluctuation frequency data are compared and analyzed. According to the analysis results, the power fluctuation stability data of regional interconnection period is generated to formulate corresponding power regulation and stability measures.

[0110] Preferably, step S34 includes the following steps:

[0111] Step S341: Calculate the fluctuation amplitude mean of the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to obtain the regional power demand fluctuation amplitude benchmark data;

[0112] Step S342: performing regional power fluctuation range over-limit analysis on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data based on the regional power demand fluctuation range benchmark data, and generating regional power demand fluctuation range over-limit data and regional power emergency demand fluctuation over-limit data;

[0113] Step S343: perform regional power fluctuation stability assessment based on regional power demand fluctuation amplitude exceeding limit data and regional power emergency demand fluctuation exceeding limit data, and generate regional interconnection period power fluctuation stability data.

[0114] In the embodiments of the present invention, the mean value of the fluctuation amplitudes of the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data is calculated to determine the benchmark of the fluctuation amplitude. Statistical analysis is performed on the fluctuation frequency data of each demand mode (ordinary demand and emergency demand), and the mean value of its fluctuation amplitude is calculated. Statistical software such as NumPy and Pandas in Python is used for data processing and calculation. Based on the fluctuation amplitude benchmark data obtained in step S341, it is analyzed whether the fluctuation amplitudes of ordinary demand and emergency demand exceed the preset threshold. The mean value of the fluctuation amplitude calculated in step S341 is used as the benchmark. Overrun analysis is performed on the ordinary demand fluctuation frequency data and the emergency demand fluctuation frequency data to determine whether the fluctuation amplitude exceeds the set threshold. If the fluctuation amplitude exceeds the threshold, the corresponding regional power demand fluctuation amplitude overrun data and regional power emergency demand overrun data are generated. Power fluctuation stability assessment is performed according to the overrun data to determine the power fluctuation stability of the region under different demand modes. Analyze the regional power demand fluctuation amplitude overrun data and regional power emergency demand overrun data generated in step S342. Combining the operating requirements and safety standards of the power system, the stability of power fluctuations under each demand mode is evaluated. Generate regional connection period power fluctuation stability data, which can be used to optimize the dispatching and management strategies of the power system.

[0115] As an example of the present invention, refer to Figure 4 As shown, step S4 in this example includes:

[0116] Step S41: Collect time-series power loads according to the regional power connection period data to obtain historical power station power load data and real-time power station power load data;

[0117] Step S42: Use machine learning methods to predict the real-time power station power load data based on the historical power station power load data to obtain power load prediction data;

[0118] Step S43: Visualize the power load prediction data to generate a power load prediction curve; adjust the curve curvature of the power load prediction curve through the regional connection period power fluctuation stability data to generate an optimized power load prediction curve.

[0119] In the embodiments of the present invention, the historical power load data and real-time power load data of power stations within the collection area are collected as the basis for power load prediction and optimization. Through historical records or databases, detailed load data of power stations over a past period of time are obtained, usually in units of hours or smaller time granularity. The current power load data of power stations are monitored and collected in real time to ensure the timeliness and accuracy of the data. Suitable machine learning models are selected, such as regression models (linear regression, polynomial regression), time series models (ARIMA, Prophet), neural network models (LSTM, GRU), etc. Feature extraction and preprocessing are performed on the historical load data, including seasonal adjustment, trend analysis, etc. The selected model is trained and optimized using historical data to ensure that the model can accurately predict the power load in different time periods. The real-time collected power load data are input into the trained model to generate real-time prediction results of the power load. Data visualization tools (such as Matplotlib, Seaborn) are used to convert the power load prediction data into an intuitive curve graph to display the prediction results and historical data. According to the regional power fluctuation stability data generated in step S34, the curvature and change rate of the power load prediction curve are adjusted. This includes increasing the conservativeness of the prediction during high fluctuations to ensure the stability of the power system under large fluctuations.

[0120] Preferably, step S42 includes the following steps:

[0121] Step S421: Divide the historical power load data of power stations into a model training set and a model test set;

[0122] Step S422: Use the linear programming algorithm to train the model on the model training set to generate a power load training model; perform model test iterations on the power load training model through the model test set to generate a power load prediction model;

[0123] Step S423: Import the real-time power load data of power stations into the power load prediction model for power load prediction to generate power load prediction data.

[0124] In the embodiments of the present invention, by sorting and cleaning the historical power load data of power stations, the integrity and consistency of the data are ensured. Usually, the random division or time series division method is adopted to ensure that the training set and the test set are representative and have a sufficient amount of data. A common division ratio is to use most of the data (for example, 80%) as the training set and the remaining part (for example, 20%) as the test set. Select a suitable linear programming algorithm, such as linear regression or other suitable optimization algorithms, where the mathematical formula of the linear programming algorithm is ; where represents the observed load, represents the predicted load, is the input feature, which are the coefficients of the linear regression model optimized by the linear programming algorithm. Using the historical data in the training set, the selected model is trained to learn the patterns and trends of the power load. The trained model is evaluated using the test set, comparing the predicted results with the actual data to evaluate the accuracy and generalization ability of the model. According to the evaluation results, the model is adjusted and optimized to improve the accuracy and stability of the prediction. The real-time collected power load data is used as the input of the model. The model makes predictions based on the real-time data and generates the predicted power load data. The prediction results are output as real-time power load prediction data for the power system operators to perform real-time scheduling and management.

[0125] In this specification, a big data-driven power load prediction system is provided, including:

[0126] A power complexity analysis module, which is used to obtain the regional power station data; divide the regional power station data by regional electricity consumption types to generate regional power consumption type data; evaluate the regional power complexity of the regional power station data according to the regional power consumption type data to generate regional power complexity data;

[0127] A load peak analysis module, which is used to perform time-series load analysis on the regional power complexity data according to a preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the high-complexity regional power peak load data and the low-complexity regional power base load data in time to obtain regional power connection period data;

[0128] A power demand analysis module, which is used to obtain the regional population density data; perform regional power demand analysis on the regional power connection period data based on the regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate the regional power fluctuation stability of the regional power demand data to generate regional connection period power fluctuation stability data;

[0129] A load prediction module, which is used to collect time-series power load of the power station according to the regional power connection period data to obtain historical power station power load data and real-time power load data of the power station; perform power load prediction on the real-time power load data of the power station based on the historical power station power load data to obtain a power load prediction curve; adjust the curve curvature of the power load prediction curve through the regional connection period power fluctuation stability data to generate an optimized power load prediction curve.

[0130] The beneficial effects of the present invention are as follows: By collecting relevant data of all power stations in a specific area, classifying these data to determine different types of electricity consumption demands, such as industrial electricity consumption, commercial electricity consumption, and residential electricity consumption, etc., generating an electricity consumption type dataset according to the classification results. Using the electricity consumption type data, evaluating the complexity of the regional power station data, and the evaluation criteria can include electricity consumption, diversity of electricity consumption patterns, etc., to generate regional power complexity data. Analyzing the complexity data to identify the peak load in high-complexity areas and the base load in low-complexity areas. Synchronizing the high-complexity and low-complexity data in time to obtain regional power connection period data. Collecting regional population density data related to electricity demand. Based on the population density data, analyzing the connection period data to generate electricity demand data. Using FFT to evaluate the fluctuation stability of the electricity demand data to generate regional connection period power fluctuation stability data. Based on the regional power connection period data, collecting historical power station load data and real-time load data. Using the historical load data to predict the real-time load data to obtain a power load prediction curve. Adjusting the prediction curve through the fluctuation stability data to generate an optimized power load prediction curve. Through multi-dimensional data analysis and complexity evaluation, providing a more accurate power load prediction, which helps to efficiently allocate power resources, ensure stable and efficient power supply. Through the fluctuation stability evaluation, reducing the impact of power fluctuations on the system and improving the overall stability and reliability of the system. Therefore, the present invention conducts power load prediction through data classification, complexity evaluation, population density analysis, and the combination of real-time and historical data, improving the accuracy and reliability of power load prediction.

[0131] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to cover all changes falling within the meaning and scope of the equivalent elements of the application documents within the present invention.

[0132] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power load forecasting method based on big data drive, characterized in that: The following steps are involved: Step S1: Acquire regional power site data; Divide regional power site data into regional power consumption types and generate regional power consumption type data; Conduct regional power complexity assessment on regional power site data based on regional power consumption type data to generate regional power complexity data; Step S2: Compare the regional power complexity data with the preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; synchronize the high-complexity regional power peak load data and the low-complexity regional power base load data to obtain regional power interconnection period data; Step S3: Acquire regional population density data; perform regional power demand analysis on regional power interconnection period data based on regional population density data to generate regional power demand data; use fast Fourier transform method to evaluate regional power fluctuation stability on regional power demand data to generate regional interconnection period power fluctuation stability data; Step S4: collecting time-series power load according to the regional power interconnection period data to obtain historical power site power load data and real-time power site power load data; Based on the historical power site power load data, the real-time power load data of the power site is used to predict the power load, and a power load prediction curve is obtained; The curvature of the power load forecast curve is adjusted by using the power fluctuation stability data during the regional interconnection period to generate an optimized power load forecast curve.

2. The power load forecasting method based on big data drive according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire regional power site data; Step S12: performing data preprocessing on the regional power site data to generate standard regional power site data, wherein the data preprocessing includes data cleaning, data outlier repair and data standardization; Step S13: dividing the standard regional power site data into regional power consumption types to generate regional power consumption type data, wherein the regional power consumption type division includes regional industrial power consumption division, regional agricultural power consumption division, regional transportation power consumption division and regional living power consumption division; Step S14: Perform regional power complexity evaluation on standard regional power site data according to regional power consumption type data to generate regional power complexity data.

3. The power load forecasting method based on big data drive according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: Compare the regional power complexity data with a preset standard regional power complexity threshold. When the regional power complexity data is greater than or equal to the preset standard regional power complexity threshold, perform short-term peak load analysis on the corresponding regional power site data to generate high-complexity regional power peak load data. Step S22: when the regional power complexity data is less than a preset standard regional power complexity threshold, a long-term base load analysis is performed on the corresponding regional power site data to generate low-complexity regional power base load data; Step S23: performing power curve conversion on the power peak load data of the high-complexity region and the power base load data of the low-complexity region to generate a power time series curve of the high-complexity region and a power time series curve of the low-complexity region; Step S24: Overlapping the high-complexity regional power timing curve and the low-complexity regional power timing curve to generate a regional power timing overlap curve; performing time synchronization on the regional power timing overlap curve to obtain regional power interconnection period data.

4. The method for predicting power load based on big data drive according to claim 3 is characterized in that: The short-term peak load analysis of the corresponding regional power station data includes: Performing short-term time series analysis on the corresponding regional power site data to generate short-term time series data of the first regional power site, wherein the short-term time series analysis includes minute level, hour level and intraday level; Perform load fluctuation analysis on the short-term time series data of the first regional power site to generate the short-term time series load fluctuation data of the first regional power site; perform load mean value calculation on the short-term time series load fluctuation data of the first regional power site to obtain the short-term time series intermediate load data of the first regional power site; The peak load is extracted from the short-term time series load fluctuation data of the first regional power site according to the short-term time series intermediate load data of the first regional power site to generate high-complexity regional power peak load data.

5. The power load forecasting method based on big data drive according to claim 3 is characterized in that: Long-term baseload analysis of corresponding regional power station data includes: Performing a long-term time series analysis on the corresponding regional power site data to generate long-term time series data of the second regional power site, wherein the long-term time series analysis includes a weekly level, a monthly level, and a yearly level; Perform load fluctuation analysis on the long-term time series data of the second regional power site to generate the long-term time series load fluctuation data of the second regional power site; perform load mean calculation on the long-term time series load fluctuation data of the second regional power site to obtain the long-term time series intermediate load data of the second regional power site; The long-term time-series load fluctuation data of the second regional power site is subjected to basic load extraction according to the long-term time-series intermediate load data of the second regional power site to generate low-complexity regional power base load data.

6. The method for predicting power load based on big data drive according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: Obtain regional population density data; Step S32: performing regional power demand analysis on the regional power interconnection period data based on the regional population density data to generate regional power demand data; performing demand pattern recognition on the regional power demand data to generate regional power ordinary demand pattern data and regional power emergency demand pattern data; Step S33: using a fast Fourier transform method to perform hierarchical power load fluctuation frequency analysis on the regional power normal demand pattern data and the regional power emergency demand pattern data, to generate regional power normal demand fluctuation frequency data and regional power emergency demand fluctuation frequency data; Step S34: Perform regional power fluctuation stability assessment on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to generate regional interconnection period power fluctuation stability data.

7. The method for predicting power load based on big data drive according to claim 6 is characterized in that: Step S34 includes the following steps: Step S341: Calculate the fluctuation amplitude mean of the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data to obtain the regional power demand fluctuation amplitude benchmark data; Step S342: performing regional power fluctuation range over-limit analysis on the regional power ordinary demand fluctuation frequency data and the regional power emergency demand fluctuation frequency data based on the regional power demand fluctuation range benchmark data, and generating regional power demand fluctuation range over-limit data and regional power emergency demand fluctuation over-limit data; Step S343: perform regional power fluctuation stability assessment based on regional power demand fluctuation amplitude exceeding limit data and regional power emergency demand fluctuation exceeding limit data, and generate regional interconnection period power fluctuation stability data.

8. The method for predicting power load based on big data drive according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: collecting time-series power load according to the regional power interconnection period data to obtain historical power site power load data and real-time power site power load data; Step S42: using a machine learning method to perform power load forecasting on the real-time power load data of the power site based on the historical power load data of the power site, to obtain power load forecasting data; Step S43: Visualize the power load forecast data to generate a power load forecast curve; adjust the curvature of the power load forecast curve using the power fluctuation stability data during the regional interconnection period to generate a power load forecast optimization curve.

9. The method for predicting power load based on big data drive according to claim 8, characterized in that: Step S42 includes the following steps: Step S421: dividing the historical power site power load data into a data set to obtain a model training set and a model test set; Step S422: Performing model training on the model training set using a linear programming algorithm to generate a power load training model; performing model testing iterations on the power load training model using a model testing set to generate a power load prediction model; Step S423: Import the real-time power load data of the power site into the power load prediction model to perform power load prediction and generate power load prediction data.

10. A power load forecasting system based on big data, characterized in that: Used to execute the power load forecasting method based on big data drive as claimed in claim 1, the power load forecasting system based on big data drive comprises: The power complexity analysis module is used to obtain regional power site data; classify the regional power site data into regional power consumption types to generate regional power consumption type data; perform regional power complexity evaluation on the regional power site data based on the regional power consumption type data to generate regional power complexity data; The load peak analysis module is used to compare the regional power complexity data with the preset standard regional power complexity threshold to generate high-complexity regional power peak load data and low-complexity regional power base load data; time synchronization is performed on the high-complexity regional power peak load data and the low-complexity regional power base load data to obtain regional power interconnection period data; The power demand analysis module is used to obtain regional population density data; perform regional power demand analysis on regional power interconnection period data based on regional population density data to generate regional power demand data; use the fast Fourier transform method to evaluate regional power fluctuation stability on regional power demand data to generate regional interconnection period power fluctuation stability data; The load forecasting module is used to collect time-series power load according to the regional power interconnection period data, obtain historical power site power load data and power site power load real-time data; forecast the power load of the power site real-time data based on the historical power site power load data to obtain the power load forecasting curve; adjust the curvature of the power load forecasting curve through the power fluctuation stability data of the regional interconnection period to generate the power load forecasting optimization curve.

Images