A method and system for power trading based on multi-dimensional data analysis

Abstract

The present invention relates to the field of data processing technology, and in particular to a method and system for power trading based on multi-dimensional data analysis. The method comprises the following steps: obtaining multi-dimensional power trading data and power grid power flow data; extracting dimensional features from multi-dimensional power trading data using a self-learning neural network to generate dimensional feature power trading data; performing time series analysis on dimensional feature power trading data to generate power trading time series data; extracting demand features from power trading time series data to generate a power demand time series curve; performing dynamic demand peak calculation on the power demand time series curve using a deep learning algorithm to generate a power demand peak; performing demand pattern analysis on the power trading time series data according to the power demand peak to generate power trading demand pattern data; performing implicit association analysis on the power trading demand pattern data to generate demand pattern association data. The present invention realizes efficient and accurate power trading.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an electricity trading method and system based on multi-dimensional data analysis. Background Art

[0002] With the development of the power market and the increasing complexity of power supply and demand, traditional power trading methods face many challenges. Traditional methods usually trade based on the principles of supply and demand matching and price competition, but this method often has problems such as low efficiency and inaccuracy in power trading analysis, resulting in unsatisfactory trading results. In order to meet the needs of the modern power market, a power trading method and system based on multi-dimensional data analysis is proposed. This method comprehensively considers data from multiple dimensions and uses advanced data analysis technology to make power trading decisions in a more accurate and intelligent way. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes an electricity trading method and system based on multi-dimensional data analysis to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a power trading method based on multi-dimensional data analysis, comprising the following steps:

[0005] Step S1: acquiring multi-dimensional power transaction data and power grid power flow data; using a self-learning neural network to extract dimensional features from the multi-dimensional power transaction data to generate dimensional feature power transaction data; performing time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extracting demand features from the power transaction time series data to generate a power demand time series curve;

[0006] Step S2: using a deep learning algorithm to perform dynamic demand peak calculation on the power demand time series curve to generate a power demand peak; performing demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; performing implicit association analysis on the power transaction demand pattern data to generate demand pattern association data;

[0007] Step S3: integrating the power transaction time series data based on the demand pattern association data to construct a power demand pattern library; identifying user transaction nodes for the multi-dimensional power transaction data to generate user transaction node data; constructing a topological structure for the user transaction node data to generate a user transaction node network;

[0008] Step S4: Capture the power demand change rate of the power demand pattern library to generate the power demand evolution law; perform dynamic correlation analysis on the user transaction node network according to the power demand evolution law to construct a dynamic correlation network topology structure;

[0009] Step S5: performing load calculation on the power supply flow data of the power grid to generate power supply load data; performing power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; performing edge load calculation on the dynamic associated network topology structure using the power load distribution map to generate edge load data; performing load defect analysis on the edge load data to generate load defect data;

[0010] Step S6: Optimize the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand transaction strategy of the adaptive load optimization network structure to generate an optimal transaction strategy engine; perform data mining modeling on the optimal transaction strategy engine to construct a power transaction strategy model and execute power transaction operations.

[0011] The present invention can obtain comprehensive power market information by acquiring multi-dimensional power transaction data and power grid power flow data. By using a self-learning neural network to extract dimensional features of power transaction data, key features can be extracted from the original data to reduce redundant information. Through time series analysis and demand feature extraction, power transaction time series data and power demand time series curves can be generated to provide a basis for subsequent analysis. The peak power demand can be calculated by a deep learning algorithm to accurately estimate the peak period of power demand. Demand pattern analysis of power transaction time series data can identify different transaction demand patterns. Implicit association analysis can reveal the association between different demand patterns and provide a basis for subsequent data integration and network construction. Data integration can be performed through demand pattern association data to construct a power demand pattern library containing various different demand patterns. Identifying user transaction nodes and constructing related topological structures can reveal the transaction relationship and dependency relationship between users and form a user transaction node network. By capturing the demand change rate through the power demand pattern library, the evolution law of power demand can be analyzed, including long-term trends and short-term fluctuations. Based on the demand evolution law, the dynamic correlation analysis of the user transaction node network can be carried out to construct a dynamic correlation network topology structure to reflect the dynamic relationship between nodes in the power market. By performing load calculation on the power supply flow data of the power grid, the load situation of the power grid can be obtained. According to the user transaction node network, the power load distribution analysis of the power supply load data can be carried out to understand the distribution of the power load between different nodes. The edge load of the dynamic correlation network topology can be calculated by using the power load distribution diagram to calculate the edge load of each node. Load defect analysis of the edge load data can be carried out to find load anomalies or faults and generate load defect data. By analyzing the load defect data, load anomalies or faults can be identified, and load decisions can be optimized for the dynamic correlation network topology structure, and the load distribution between nodes can be adjusted to improve the reliability and stability of the power system. The optimal supply and demand trading strategy analysis of the adaptive load optimization network structure can formulate the optimal power trading strategy according to the current load situation and demand pattern. The power trading strategy model constructed by data mining modeling can automatically execute power trading operations and improve trading efficiency and profits.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: Acquire multi-dimensional power transaction data and power grid power flow data;

[0014] Step S12: extracting dimensional features from the multi-dimensional power transaction data using a self-learning neural network to generate dimensional feature power transaction data;

[0015] Step S13: performing heterogeneous data fusion on the dimensional feature power transaction data to generate heterogeneous dimensional fusion feature data;

[0016] Step S14: performing time series analysis on the heterogeneous dimension fusion feature data to generate power transaction time series data;

[0017] Step S15: extracting demand characteristics from the power transaction time series data to generate power transaction time series demand data;

[0018] Step S16: Perform time series curve fitting on the power transaction time series demand data to generate a power demand time series curve.

[0019] The present invention can obtain comprehensive power market information by acquiring multi-dimensional power trading data and power grid power flow data. Power trading data includes power supply and power purchase conditions of different trading participants, as well as information such as price and trading volume. Power grid power flow data reflects the supply capacity and load conditions of the power system and is the basic data for power trading. By extracting dimensional features of multi-dimensional power trading data through self-learning neural networks, key features can be extracted from the original data to reduce redundant information. Dimensional feature power trading data contains streamlined and refined features, and has better interpretability and high-dimensional expression capabilities. Through heterogeneous data fusion, power trading data from different data sources or different types are integrated to generate heterogeneous dimensional fusion feature data. Such data fusion can comprehensively consider information from different dimensions, provide more comprehensive and accurate feature descriptions, and enhance the comprehensiveness and interpretability of data. Through time series analysis, the time dimension of heterogeneous dimensional fusion feature data is modeled and analyzed to generate power trading time series data. Power trading time series data shows the trading conditions and trading trends at different time points, and can reveal the dynamic changes and cyclical characteristics of the power market. By extracting demand characteristics from power trading time series data, the characteristics of power demand at different time points can be identified. Power trading time series demand data presents information such as the changing trend of power demand and the difference between peaks and valleys, providing a basis for subsequent demand analysis and decision-making. By fitting the time series curve of power trading time series demand data, a smooth curve representation of power demand can be obtained, namely, the power demand time series curve. The power demand time series curve can more accurately reflect the changing trend and periodicity of power demand, and provide an important reference for demand forecasting and planning in the power market.

[0020] Preferably, step S2 comprises the following steps:

[0021] Step S21: Analyze the demand fluctuation characteristics of the power demand time series curve using a deep learning algorithm to generate power demand fluctuation characteristic data;

[0022] Step S22: using a power dynamic demand peak calculation formula to perform dynamic demand peak calculation on the power demand fluctuation characteristic data to generate a power demand peak;

[0023] Step S23: performing demand trend analysis on the power transaction time series data according to the power demand peak value to generate power demand trend data;

[0024] Step S24: performing demand pattern analysis on the power demand trend data to generate power transaction demand pattern data;

[0025] Step S25: performing a spatiotemporal dependency analysis on the power trading demand pattern data according to the power demand time series curve to generate spatiotemporal demand dependency data;

[0026] Step S26: Perform implicit association analysis on the spatiotemporal demand dependency data to generate demand pattern association data.

[0027] The present invention analyzes the power demand time series curve through a deep learning algorithm, and can extract the fluctuation characteristics of the demand, including information such as amplitude, frequency, and cycle. The power demand fluctuation characteristic data can describe the change pattern of power demand in more detail, and provide a reference for subsequent demand forecasting and regulation. By applying the power dynamic demand peak calculation formula, the power demand fluctuation characteristic data can be processed to obtain the peak value of power demand. The power demand peak refers to the highest point reached by the demand in a certain period of time, which reflects the peak moment of power demand. The calculation of the power demand peak can help the power market and the power system to carry out load scheduling and resource allocation. According to the power demand peak, the power trading time series data is analyzed to obtain the trend information of power demand. The power demand trend data shows the overall change trend of power demand, including year-on-year growth, seasonal changes, cyclical fluctuations, etc. The power demand trend data has important reference value for the planning of the power market and the balance of supply and demand. By performing demand pattern analysis on the power demand trend data, different demand patterns can be identified. The power trading demand pattern data describes the typical patterns of power demand in different time periods, such as steady growth, sudden increase, sudden drop, etc. Power trading demand pattern data can be used for demand forecasting and market regulation, helping decision makers to better respond to different demand situations. By analyzing the power demand time series curve and combining it with the power trading demand pattern data, spatiotemporal demand dependency data can be obtained. Spatiotemporal demand dependency data reflects the correlation of power demand in time and space, such as the degree of demand correlation between different time periods and different regions. Spatiotemporal demand dependency data helps in demand forecasting and resource allocation in the power market and improves the efficiency of supply and demand matching. By performing implicit association analysis on spatiotemporal demand dependency data, the potential association rules and patterns can be discovered. Demand pattern association data reveals the relationship and mutual influence between different demand patterns, which helps to understand the complexity and interdependence of demand. Demand pattern association data provides deeper insights and decision support for demand forecasting, resource allocation and market regulation in the power market.

[0028] Preferably, the calculation formula for the peak value of the dynamic power demand in step S22 is specifically:

[0029]

[0030] P is the peak value of dynamic power demand, A is the load power, B is the power generation, C is the load rate, D is the power demand fluctuation range, n is the system load output frequency, F is the maximum voltage of the power grid system, G is the rated output voltage of the power grid system, l is the total number of power node equipment in the power grid system, b is the power peak control rate, y is the difference between the highest and lowest ambient temperatures, and x is the output efficiency of the power grid unit.

[0031] The present invention calculates the natural logarithm of the product of load power, power generation and load rate through ln(A·B·C). Load power represents the load level of the entire system, power generation represents the power generation capacity of the system, and load rate represents the ratio of load to power generation capacity. Logarithmic operation can smooth data and reduce the impact of extreme values. In the power system, load power, power generation and load rate may experience large fluctuations or extreme values. Taking logarithms can reduce the impact of these fluctuations and make the calculation results more stable. Through logarithmic operations, data of different magnitudes can be compared. Load power, power generation and load rate usually have different magnitudes. After taking the logarithm, the magnitude difference is small, which is convenient for comparing their relative sizes. By multiplying the power demand fluctuation range with the maximum voltage attenuation effect of the power grid system and then dividing it by the rated output voltage of the power grid system, the impact of voltage on the peak power demand can be adjusted. This helps to more accurately estimate the peak demand of the power system and take into account the impact of voltage changes on the peak demand. By calculating the product and taking the logarithm, multiple factors such as the total number of power node devices in the power grid system, the peak load control rate, the maximum and minimum difference in ambient temperature, and the output efficiency of the power grid units are taken into account. This helps to comprehensively evaluate the impact of these factors on the peak value of dynamic power demand. By taking the logarithm, the result of the product can be converted into a smaller numerical range. This helps to reduce the magnitude difference of the values, making the data easier to process and analyze, while also reducing the impact of extreme values. By comprehensively considering these variables, the peak demand level of the power system can be more comprehensively evaluated, and a reference can be provided for system operation and power dispatching.

[0032] Preferably, the specific steps of step S21 are:

[0033] Step S211: using a deep learning algorithm to perform demand amplitude analysis on the power demand time series curve to obtain demand fluctuation amplitude data;

[0034] Step S212: performing frequency domain transformation on the demand fluctuation amplitude data to generate demand fluctuation frequency domain data;

[0035] Step S213: performing waveform period detection on the demand fluctuation frequency domain data, thereby generating demand fluctuation period data;

[0036] Step S214: extracting frequency components from the demand fluctuation period data to generate demand fluctuation frequency component data;

[0037] Step S215: performing demand fluctuation characteristic analysis on the power demand time series curve according to the demand fluctuation frequency component data to generate power demand fluctuation characteristic data.

[0038] The present invention performs demand amplitude analysis on the power demand time series curve through a deep learning algorithm, and can obtain the amplitude data of demand fluctuation. The demand fluctuation amplitude data reflects the degree of fluctuation of power demand in different time periods, and can reveal the nature of fluctuation of power demand, such as stable, violent, mild, etc. These data are helpful to understand the stability and variability of power demand, and provide a basis for the regulation and control of the power market and the balance of supply and demand. By performing frequency domain conversion on the demand fluctuation amplitude data, the time domain information of demand fluctuation can be converted into frequency domain information. The demand fluctuation frequency domain data provides the contribution degree of different frequency components to demand fluctuation, and can reveal the frequency characteristics of power demand. These data are helpful to understand the periodicity and frequency distribution of power demand, and are of great significance to the load scheduling and demand forecasting of the power market. By performing waveform period detection on the demand fluctuation frequency domain data, the periodic characteristics of demand fluctuation can be determined. The demand fluctuation period data reflects the periodic changes of power demand, such as the periodicity of daily, weekly, monthly or other longer time periods. These data are helpful to understand the seasonal changes and long-term trends of power demand, and provide guidance for demand forecasting and resource planning in the power market. By extracting the frequency components of the demand fluctuation period data, the contribution degree of different frequency components to demand fluctuation can be obtained. The demand fluctuation frequency component data describes the demand fluctuations within different frequency ranges, such as the amplitude and change trend of low-frequency, medium-frequency, and high-frequency components. These data help to understand the fluctuation characteristics of electricity demand on different time scales and provide a basis for the regulation and analysis of the electricity market. According to the demand fluctuation frequency component data, the electricity demand time series curve is analyzed to extract the fluctuation characteristics of electricity demand, including information such as amplitude and change trend within different frequency ranges. The electricity demand fluctuation characteristic data can more comprehensively describe the change pattern of electricity demand, including the fluctuation nature on different time scales and the contribution of frequency components. These data are helpful for demand forecasting, load scheduling and market analysis in the electricity market, and improve the accuracy and efficiency of supply and demand matching.

[0039] Preferably, the specific steps of step S3 are:

[0040] Step S31: segmenting the power transaction time series data into time series to generate a power time series set;

[0041] Step S32: Sequence-marking the power time series set based on the demand pattern associated data to generate a demand pattern time series sequence;

[0042] Step S33: integrating the demand pattern time series data to construct a power demand pattern library;

[0043] Step S34: performing user transaction node identification on the multi-dimensional power transaction data to generate user transaction node data;

[0044] Step S35: Analyze the transaction behavior of the user transaction node data to generate power transaction behavior data;

[0045] Step S36: construct a topological structure based on the power transaction behavior data and user transaction node data to generate a user transaction node network.

[0046] The present invention organizes and analyzes the power transaction data in chronological order, which facilitates subsequent tasks such as demand pattern identification and transaction behavior analysis. The generated power time series sequence set provides basic data for subsequent steps, making further analysis of the power transaction data more effective and convenient. By marking the power time series sequence set, the demand pattern therein can be identified and divided. The demand pattern refers to a group of transaction sequences with similar characteristics and regularities in the power transaction data. The generated demand pattern time series sequence can be used for subsequent tasks such as demand pattern analysis and demand forecasting. By identifying and analyzing the demand pattern, the potential patterns and regularities in the power transaction data can be revealed, providing a decision-making basis for the power market and supply and demand management. By integrating the data of the demand pattern time series sequence, different demand patterns can be integrated into a power demand pattern library. The power demand pattern library is a collection of multiple demand patterns, each of which represents a class of power transaction sequences with similar demand characteristics. Constructing a power demand pattern library helps to classify and summarize power demand, and provides a more comprehensive and accurate reference for tasks such as demand forecasting and market analysis. By identifying user transaction nodes for multi-dimensional power transaction data, the nodes of different users in power transactions can be determined. User trading nodes refer to specific users in the power trading network, which play an important role in trading. The generated user trading node data can be used for subsequent tasks such as user behavior analysis and user classification. Identifying user trading nodes can better understand the structure and user behavior of the power trading network, and provide insights and decision-making basis for the power market and supply and demand management. By analyzing the trading behavior of user trading node data, the behavioral characteristics and patterns of users in power trading can be obtained. Power trading behavior data describes the trading behavior of users in the trading process, such as buying, selling, adjusting and other operations. The generated power trading behavior data can be used for subsequent tasks such as trading pattern analysis and user behavior prediction. Analyzing user trading behavior helps to understand the trading pattern and user strategy in the power market, and provide reference for market supervision and decision-making. By constructing a topological structure based on power trading behavior data and user trading node data, the connection and relationship between user trading nodes can be established. The user trading node network describes the trading relationship and connection mode between users, and can reveal the trading network structure and interaction mode between users in the power market. The generated user trading node network helps to analyze indicators such as trading flow and market concentration in the power market, and provides reference for market supervision, resource scheduling and other decisions.

[0047] Preferably, the specific steps of step S4 are:

[0048] Step S41: capturing the power demand change rate of the power demand pattern library to generate power demand change rate data;

[0049] Step S42: performing mutation point detection on the power demand time series curve according to the power demand change rate data to generate demand mutation point data;

[0050] Step S43: performing demand evolution analysis on the power demand pattern library based on the demand mutation point data to generate a power demand evolution law;

[0051] Step S44: Performing dynamic association analysis on the user transaction node network according to the law of power demand evolution to generate node dynamic association data;

[0052] Step S45: performing time series change analysis on the node dynamic correlation data based on the power demand time series curve to generate time series structure evolution data;

[0053] Step S46: Perform dynamic correlation analysis on the user transaction node network through the time series structure evolution data to construct a dynamic correlation network topology structure.

[0054] The present invention can calculate and extract the rate of change information of power demand by capturing the rate of change of power demand in the power demand pattern library. The rate of change data of power demand reflects the fluctuation and change trend of power demand. The generated rate of change data of power demand can be used for subsequent demand analysis, fluctuation prediction and other tasks. Analyzing the rate of change of power demand helps to understand the dynamic characteristics and periodicity of demand, and provides a reference for the dispatch and resource optimization of the power market. By performing mutation point detection on the rate of change data of power demand, the position of the mutation point in the power demand time series curve can be determined. The demand mutation point refers to the time point at which a significant change occurs in the demand curve, which may represent a turning point or an important event in power demand. The generated demand mutation point data can be used for subsequent demand analysis, anomaly detection and other tasks. Detecting the mutation point of demand helps to capture the key moments and abnormal situations of demand changes, and provides early warning and decision-making basis for the dispatch and response of the power market. By performing demand evolution analysis on the power demand pattern library based on the demand mutation point data, the evolution law and change trend of power demand can be revealed. The law of evolution of power demand describes the characteristics of the change, cycle and trend of the demand pattern. The generated law of evolution of power demand can be used for subsequent tasks such as demand forecasting and market planning. Analyzing the evolution law of demand helps to understand the long-term trend and periodicity of demand, and provides reference and decision-making basis for the dispatch and resource optimization of the power market. By performing dynamic association analysis on the user trading node network according to the law of power demand evolution, the dynamic association relationship between nodes can be determined. The node dynamic association data describes the connection and interaction mode of user trading nodes at different stages of demand evolution. The generated node dynamic association data can be used for subsequent network analysis, node evolution prediction and other tasks. Analyzing the dynamic association relationship of nodes helps to understand the transaction mode and interaction behavior between users, and provides reference and decision-making basis for the dispatch and market design of the power market. By performing time series change analysis on the node dynamic association data based on the power demand time series curve, the time series evolution characteristics of the node dynamic association relationship can be captured. The time series structure evolution data describes the change and evolution process of the association pattern between nodes over time. The generated time series structure evolution data can be used for subsequent time series analysis, pattern recognition and other tasks. Analyzing the time series evolution characteristics of the node association relationship helps to understand the change of the association pattern between nodes over time, and provides reference and decision-making basis for the dispatch and resource optimization of the power market. The dynamic association network topology reflects the changes and evolution of the association relationship between nodes over time. The generated dynamic association network topology can be used for subsequent network analysis, key node identification and other tasks. Constructing the topology of the dynamic association network helps to understand the changes and evolution of the association pattern between nodes over time, and provides reference and decision-making basis for the dispatch and market design of the power market.

[0055] Preferably, the specific steps of step S5 are:

[0056] Step S51: performing load calculation on the power supply flow data of the power grid to generate power supply load data;

[0057] Step S52: performing power load distribution analysis on the power supply load data according to the user transaction node network to generate power load distribution data;

[0058] Step S53: performing load intensity calculation on the power load distribution data to generate load distribution intensity data;

[0059] Step S54: performing load distribution space mining on the load distribution intensity data to generate a power load distribution map;

[0060] Step S55: Calculate the edge load of the dynamic associated network topology using the power load distribution diagram to generate edge load data;

[0061] Step S56: Perform load defect analysis on the edge load data to generate load defect data.

[0062] The present invention can estimate the current power supply load of the power grid by performing load calculation on the power supply flow data of the power grid. The power supply load data reflects the load size carried by the power grid in a specific time period. The generated power supply load data can be used for subsequent load management, power grid planning and other tasks. Calculating the power supply load helps to understand the load situation and load distribution of the power grid, and provides reference and decision-making basis for the operation and scheduling of the power system. By performing power load distribution analysis on the power supply load data according to the user transaction node network, the distribution of the power load on different nodes can be determined. The power load distribution data describes the load amount and distribution carried by different nodes. The generated power load distribution data can be used for subsequent load balancing, load forecasting and other tasks. Analyzing the distribution of the power load helps to understand the load contribution and load distribution of each node, and provides reference and decision-making basis for load management and resource optimization of the power system. Calculating the load intensity helps to understand the load situation of each node and the degree of imbalance of load distribution, and provides reference and decision-making basis for load management and resource optimization of the power system. By mining the load distribution space of the load distribution intensity data, the distribution pattern and aggregation of the power load in space can be revealed. The power load distribution diagram shows the spatial distribution of the load intensity of each node. By using the power load distribution diagram to calculate the edge load of the dynamically associated network topology, the load of each edge node in the topology can be determined. The edge load data describes the load contribution and load level of the edge node in the power system. By performing load defect analysis on the edge load data, load defects or abnormalities in the edge nodes can be identified. Load defect data describes the abnormal or overloaded load of the edge node.

[0063] Preferably, the specific steps of step S6 are:

[0064] Step S61: optimizing the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimized network structure;

[0065] Step S62: performing supply and demand optimal trading strategy analysis on the adaptive load optimization network structure to generate an optimal trading strategy engine;

[0066] Step S63: Calculate the transaction frequency of the power transaction data to generate power transaction frequency data;

[0067] Step S64: Optimizing the frequency of the optimal trading strategy engine according to the power trading frequency data to generate a frequency optimized trading strategy engine;

[0068] Step S65: Perform data mining modeling on the frequency optimization trading strategy engine to construct a power trading strategy model and execute power trading operations.

[0069] The present invention can improve the network structure to adapt to the load demand by optimizing the load decision of the dynamic associated network topology structure using load defect data. The adaptive load optimization network structure is a network topology structure adjusted and optimized according to the load defect data. The generated adaptive load optimization network structure can improve the load balance, reliability and performance of the system. By optimizing the network structure, the load can be better distributed, the risk of load anomalies and failures can be reduced, and the operating efficiency and stability of the power system can be improved. By analyzing the optimal supply and demand trading strategy of the adaptive load optimization network structure, the optimal supply and demand trading strategy can be determined. The optimal trading strategy engine is a decision engine that provides the optimal supply and demand trading strategy based on the analysis and optimization of the network structure. By calculating the trading frequency of the power trading data, the frequency and trading activity of the transaction can be evaluated. The power trading frequency data reflects the frequency and scale of transactions in the power market. By implementing frequency optimization of the optimal trading strategy engine according to the power trading frequency data, the trading strategy engine can be adjusted to adapt to different trading frequencies. The frequency optimization trading strategy engine is a trading strategy decision engine that is adjusted and optimized according to the trading frequency data. By performing data mining modeling on the frequency optimization trading strategy engine, a power trading strategy model can be constructed for executing power trading operations. The power trading strategy model is based on data mining and modeling of the trading strategy engine, which can predict market trends, analyze trading risks and optimize trading decisions. Through data mining modeling, the laws, trends and correlations in the trading strategy engine can be extracted to provide decision support and guidance for power trading operations. Building a power trading strategy model helps optimize trading decisions, improve trading benefits and reduce risks, thereby promoting the healthy development and sustainable operation of the power market.

[0070] In this specification, a power trading system based on multi-dimensional data analysis is provided, which is used to execute the power trading method based on multi-dimensional data analysis as described above, including:

[0071] The demand feature module is used to obtain multi-dimensional power transaction data and power grid power flow data; use a self-learning neural network to extract dimensional features of the multi-dimensional power transaction data to generate dimensional feature power transaction data; perform time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extract demand features from the power transaction time series data to generate a power demand time series curve;

[0072] The dynamic demand peak module is used to calculate the dynamic demand peak of the power demand time series curve using a deep learning algorithm to generate the power demand peak; perform demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; perform implicit association analysis on the power transaction demand pattern data to generate demand pattern association data;

[0073] The transaction node network module is used to integrate the power transaction time series data based on the demand pattern association data to build a power demand pattern library; identify the user transaction nodes of the multi-dimensional power transaction data to generate user transaction node data; construct the topological structure of the user transaction node data to generate a user transaction node network;

[0074] The dynamic association network module is used to capture the rate of change of power demand in the power demand pattern library and generate the law of power demand evolution; according to the law of power demand evolution, the user transaction node network is dynamically analyzed to construct the dynamic association network topology structure;

[0075] The load distribution module is used to perform load calculation on the power supply flow data of the power grid to generate power supply load data; perform power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; use the power load distribution map to perform edge load calculation on the dynamic associated network topology structure to generate edge load data; perform load defect analysis on the edge load data to generate load defect data;

[0076] The trading strategy model module is used to optimize the load decision of the dynamic associated network topology structure through load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand trading strategy of the adaptive load optimization network structure to generate an optimal trading strategy engine; perform data mining modeling on the optimal trading strategy engine to build a power trading strategy model and execute power trading operations.

[0077] The present invention can generate dimensional feature power trading data by acquiring multidimensional power trading data and power grid power flow data, and extracting dimensional features using a self-learning neural network. This can reduce the dimension of the data and extract key features, providing more accurate and useful data for subsequent analysis and decision-making. By using a deep learning algorithm to perform dynamic demand peak calculation on the power demand time series curve, a power demand peak can be generated. This helps to understand the peak and trough periods of power demand and provides a basis for demand pattern analysis of power trading. According to the power demand peak, the power trading time series data is analyzed for demand patterns, and power trading demand pattern data can be generated to help understand and predict the patterns and trends of trading demand. By analyzing and integrating demand pattern associated data, a power demand pattern library is constructed. This can help identify user trading nodes and construct a user trading node network. The establishment of a user trading node network helps to understand the trading relationship and trading pattern between different users, and provides a basis for the construction of the topological structure of power trading. The power demand pattern library captures the rate of change of power demand and generates the law of evolution of power demand. This helps to understand the changing trend and law of power demand. According to the law of power demand evolution, the user transaction node network is dynamically correlated and analyzed to construct a dynamic correlation network topology. This can better understand and analyze the correlation and influence relationship between different transaction nodes, and provide support for the optimization and decision-making of power transactions. The power supply flow data of the power grid is used for load calculation to generate power supply load data. The power supply load data is analyzed for power load distribution according to the user transaction node network to generate a power load distribution map. This helps to understand the load distribution in different regions and provide a reference for load balancing and resource scheduling. The power load distribution map is used to calculate the edge load of the dynamic correlation network topology to generate edge load data. This can evaluate the load of each node in the network topology and provide a basis for load decision-making and optimization. Load defect analysis of edge load data can identify load anomalies and potential failure risks, and provide guarantees for the stability and reliability of system operation. The dynamic correlation network topology is optimized for load decision-making through load defect data to generate an adaptive load optimization network structure. This helps to improve the load balancing and performance of the network and reduce the risk of load anomalies and failures. By analyzing the optimal supply and demand trading strategy for the adaptive load optimization network structure, the best supply and demand trading strategy can be determined to improve the efficiency and economy of the power market. By conducting data mining modeling on the optimal trading strategy engine, a power trading strategy model can be constructed to perform power trading operations. This can provide data-driven trading decision support and optimize the efficiency and effectiveness of power trading. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 A schematic diagram of the steps of a power trading method based on multi-dimensional data analysis of the present invention;

[0079] Figure 2 Detailed implementation flow chart of step S1;

[0080] Figure 3 Detailed implementation flow chart of step S2;

[0081] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0082] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0083] The present application example provides a method and system for power trading based on multi-dimensional data analysis. The execution subject of the method and system for power trading based on multi-dimensional data analysis includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0084] See also Figures 1 to 4 The present invention provides a power trading method based on multi-dimensional data analysis, and the power trading method based on multi-dimensional data analysis comprises the following steps:

[0085] Step S1: acquiring multi-dimensional power transaction data and power grid power flow data; using a self-learning neural network to extract dimensional features from the multi-dimensional power transaction data to generate dimensional feature power transaction data; performing time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extracting demand features from the power transaction time series data to generate a power demand time series curve;

[0086] Step S2: using a deep learning algorithm to perform dynamic demand peak calculation on the power demand time series curve to generate a power demand peak; performing demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; performing implicit association analysis on the power transaction demand pattern data to generate demand pattern association data;

[0087] Step S3: integrating the power transaction time series data based on the demand pattern association data to construct a power demand pattern library; identifying user transaction nodes for the multi-dimensional power transaction data to generate user transaction node data; constructing a topological structure for the user transaction node data to generate a user transaction node network;

[0088] Step S4: Capture the power demand change rate of the power demand pattern library to generate the power demand evolution law; perform dynamic correlation analysis on the user transaction node network according to the power demand evolution law to construct a dynamic correlation network topology structure;

[0089] Step S5: performing load calculation on the power supply flow data of the power grid to generate power supply load data; performing power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; performing edge load calculation on the dynamic associated network topology structure using the power load distribution map to generate edge load data; performing load defect analysis on the edge load data to generate load defect data;

[0090] Step S6: Optimize the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand transaction strategy of the adaptive load optimization network structure to generate an optimal transaction strategy engine; perform data mining modeling on the optimal transaction strategy engine to construct a power transaction strategy model and execute power transaction operations.

[0091] The present invention can obtain comprehensive power market information by acquiring multi-dimensional power transaction data and power grid power flow data. By using a self-learning neural network to extract dimensional features of power transaction data, key features can be extracted from the original data to reduce redundant information. Through time series analysis and demand feature extraction, power transaction time series data and power demand time series curves can be generated to provide a basis for subsequent analysis. The peak power demand can be calculated by a deep learning algorithm to accurately estimate the peak period of power demand. Demand pattern analysis of power transaction time series data can identify different transaction demand patterns. Implicit association analysis can reveal the association between different demand patterns and provide a basis for subsequent data integration and network construction. Data integration can be performed through demand pattern association data to construct a power demand pattern library containing various different demand patterns. Identifying user transaction nodes and constructing related topological structures can reveal the transaction relationship and dependency relationship between users and form a user transaction node network. By capturing the demand change rate through the power demand pattern library, the evolution law of power demand can be analyzed, including long-term trends and short-term fluctuations. Based on the demand evolution law, the dynamic correlation analysis of the user transaction node network can be carried out to construct a dynamic correlation network topology structure to reflect the dynamic relationship between nodes in the power market. By performing load calculation on the power supply flow data of the power grid, the load situation of the power grid can be obtained. According to the user transaction node network, the power load distribution analysis of the power supply load data can be carried out to understand the distribution of the power load between different nodes. The edge load of the dynamic correlation network topology can be calculated by using the power load distribution diagram to calculate the edge load of each node. Load defect analysis of the edge load data can be carried out to find load anomalies or faults and generate load defect data. By analyzing the load defect data, load anomalies or faults can be identified, and load decisions can be optimized for the dynamic correlation network topology structure, and the load distribution between nodes can be adjusted to improve the reliability and stability of the power system. The optimal supply and demand trading strategy analysis of the adaptive load optimization network structure can formulate the optimal power trading strategy according to the current load situation and demand pattern. The power trading strategy model constructed by data mining modeling can automatically execute power trading operations and improve trading efficiency and profits.

[0092] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a power trading method and system based on multi-dimensional data analysis of the present invention. In this example, the steps of the power trading method based on multi-dimensional data analysis include:

[0093] Step S1: acquiring multi-dimensional power transaction data and power grid power flow data; using a self-learning neural network to extract dimensional features from the multi-dimensional power transaction data to generate dimensional feature power transaction data; performing time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extracting demand features from the power transaction time series data to generate a power demand time series curve;

[0094] In this embodiment, multidimensional power transaction data is obtained from relevant power markets or power trading platforms. These data may include information such as power market prices, transaction volumes, and transaction types. Power grid power flow data is obtained from power grid operators or relevant institutions. These data usually include information such as total power supply and power supply in each time period. An appropriate neural network model (such as a deep neural network) is selected to construct a self-learning model, which can automatically learn to extract the dimensional features of multidimensional power transaction data. The pre-processed multidimensional power transaction data is input into the self-learning neural network model, and the dimensional features are extracted through the hidden layer of the network to generate dimensional feature power transaction data. An appropriate time series analysis method is selected, such as a time series model (such as ARIMA, LSTM, etc.) or a frequency domain analysis method (such as FFT, Wavelet, etc.) for analyzing power transaction time series data. The selected time series analysis method is applied to model and analyze the power transaction time series data to extract the trend, periodicity, and seasonality characteristics therein. According to the results of the time series analysis, power transaction time series data is generated, which contains the time series change trend and periodicity characteristics of power transactions. Based on the power transaction time series data, appropriate methods and algorithms are used to extract power demand characteristics, such as finding the highest and lowest points, calculating the average value, etc. According to the extraction results of demand characteristics, the power demand characteristics are displayed in a time series manner to generate a power demand time series curve.

[0095] Step S2: using a deep learning algorithm to perform dynamic demand peak calculation on the power demand time series curve to generate a power demand peak; performing demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; performing implicit association analysis on the power transaction demand pattern data to generate demand pattern association data;

[0096] In this embodiment, the power demand time series curve is predicted to obtain the power demand forecast results for a period of time in the future. The peak demand value can be extracted from the prediction results, and can be calculated by finding the highest point or setting a threshold. According to the calculation results of the peak demand value, the peak power demand data is generated, which reflects the peak time and intensity of the power demand. The peak power demand value is matched with the power trading time series data to find the transaction data points related to the peak demand value. The transaction volume at the peak demand moment is analyzed, such as the peak and trough of the transaction volume. The transaction price at the peak demand moment is analyzed, such as the price fluctuation and change trend. The selected implicit association analysis method is applied to mine the power trading demand pattern data to find the association rules or association patterns therein. According to the results of the implicit association analysis, the demand pattern association data is generated, which reflects the association relationship between the power trading demand patterns, such as the simultaneous appearance or mutual exclusion of certain demand patterns with other demand patterns.

[0097] Step S3: integrating the power transaction time series data based on the demand pattern association data to construct a power demand pattern library; identifying user transaction nodes for the multi-dimensional power transaction data to generate user transaction node data; constructing a topological structure for the user transaction node data to generate a user transaction node network;

[0098] In this embodiment, pattern features, such as transaction volume, price, transaction type, etc., are extracted from the matching transaction data. Different users are identified based on the user information in the transaction data. The transaction node of each user is identified, that is, the starting point or end point of the user's transaction behavior. A node set is constructed based on the user information in the user transaction node data, and each node represents a user. According to the relevant information of the transaction node, the edge is assigned a corresponding weight, such as transaction volume, transaction frequency, etc. According to the order of the transaction nodes, the direction of the edge is determined to indicate the order of the transaction. According to the node set and the edge set, a user transaction node network is constructed. The network reflects the transaction relationship between users and the topological structure of transaction behavior.

[0099] Step S4: Capture the power demand change rate of the power demand pattern library to generate the power demand evolution law; perform dynamic correlation analysis on the user transaction node network according to the power demand evolution law to construct a dynamic correlation network topology structure;

[0100] In this embodiment, the difference in power demand in adjacent time periods is calculated and divided by the time interval. The power demand data of each demand pattern is fitted, and then the slope of the fitting curve is calculated. Based on the calculation results of the demand change rate, the law of power demand evolution is generated. This law reflects the changing trend and rate of power demand of different demand patterns over time. The demand evolution law is matched with the user transaction nodes to find the transaction nodes related to the demand evolution. According to the demand evolution law and the attributes of the transaction nodes, the correlation between the nodes is calculated to reflect the strength of the correlation between them. According to the changes in the demand evolution law, the correlation and topology are dynamically updated so that the network can reflect the dynamic changes in power demand. According to the results of the dynamic correlation analysis, a dynamic correlation network topology is constructed. The topology reflects the correlation and changes of user transaction nodes during the evolution of power demand.

[0101] Step S5: performing load calculation on the power supply flow data of the power grid to generate power supply load data; performing power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; performing edge load calculation on the dynamic associated network topology structure using the power load distribution map to generate edge load data; performing load defect analysis on the edge load data to generate load defect data;

[0102] In this embodiment, the power supply flow of adjacent power supply nodes is aggregated to obtain the total power supply load of each node. The total power supply load is proportionally distributed to the connected user nodes to obtain the load value of each user node. According to the load value of the user node, the load distribution of each node is calculated. The load distribution result is visualized as a power load distribution diagram, which can be displayed using a chart or a geographic information system (GIS). According to the results of the load distribution analysis, a power load distribution diagram is generated. The diagram reflects the load conditions of different nodes and regions in the power grid, and can be used to analyze the distribution characteristics and potential bottlenecks of the power load. According to the load conditions of each node in the load distribution diagram, the load is proportionally distributed to the edge nodes. According to the transmission path and load distribution between the edge nodes, the energy loss or delay in the transmission process is calculated. By analyzing the edge load data, abnormal load conditions are detected, such as nodes or edge paths with excessively high or low loads. Potential defects related to load abnormalities are identified, such as insufficient power supply, equipment failure, etc. According to the results of the load defect analysis, load defect data is generated. The data contains load abnormalities and possible causes of defects, which can be used to identify and solve load problems in the power system.

[0103] Step S6: Optimize the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand transaction strategy of the adaptive load optimization network structure to generate an optimal transaction strategy engine; perform data mining modeling on the optimal transaction strategy engine to construct a power transaction strategy model and execute power transaction operations.

[0104] In this embodiment, by adjusting the distribution of loads between nodes, the load is made more balanced to avoid excessive or low loads. The selection of edge paths is optimized to make the path for load transmission in the network more efficient and stable. According to the load defect data, the network structure or node configuration is dynamically adjusted to adapt to load changes and defect repair. According to the network structure and supply and demand data, the supply nodes and demand nodes are matched to determine the supply and demand relationship. Based on the supply and demand relationship and the power market rules, the optimal trading strategy is formulated, including power pricing, transaction volume allocation, etc. Based on the results of supply and demand analysis, the optimal trading strategy engine is generated. The engine provides the ability to make power trading decisions based on real-time supply and demand conditions to achieve the optimal supply and demand matching and trading strategy. Features related to transaction scale, price, time, etc. are extracted from historical transaction data. Select a suitable data mining algorithm or machine learning model, such as a regression model, a decision tree, a neural network, etc. Based on the results of data mining modeling, a power trading strategy model is constructed. The model can predict future trading demand and price trends, and perform power trading operations, including transaction volume allocation, pricing decisions, etc., to achieve the optimization of power trading.

[0105] In this embodiment, reference Figure 2 The above is a schematic flow chart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0106] Step S11: Acquire multi-dimensional power transaction data and power grid power flow data;

[0107] Step S12: extracting dimensional features from the multi-dimensional power transaction data using a self-learning neural network to generate dimensional feature power transaction data;

[0108] Step S13: performing heterogeneous data fusion on the dimensional feature power transaction data to generate heterogeneous dimensional fusion feature data;

[0109] Step S14: performing time series analysis on the heterogeneous dimension fusion feature data to generate power transaction time series data;

[0110] Step S15: extracting demand characteristics from the power transaction time series data to generate power transaction time series demand data;

[0111] Step S16: Perform time series curve fitting on the power transaction time series demand data to generate a power demand time series curve.

[0112] The present invention can obtain comprehensive power market information by acquiring multi-dimensional power trading data and power grid power flow data. Power trading data includes power supply and power purchase conditions of different trading participants, as well as information such as price and trading volume. Power grid power flow data reflects the supply capacity and load conditions of the power system and is the basic data for power trading. By extracting dimensional features of multi-dimensional power trading data through self-learning neural networks, key features can be extracted from the original data to reduce redundant information. Dimensional feature power trading data contains streamlined and refined features, and has better interpretability and high-dimensional expression capabilities. Through heterogeneous data fusion, power trading data from different data sources or different types are integrated to generate heterogeneous dimensional fusion feature data. Such data fusion can comprehensively consider information from different dimensions, provide more comprehensive and accurate feature descriptions, and enhance the comprehensiveness and interpretability of data. Through time series analysis, the time dimension of heterogeneous dimensional fusion feature data is modeled and analyzed to generate power trading time series data. Power trading time series data shows the trading conditions and trading trends at different time points, and can reveal the dynamic changes and cyclical characteristics of the power market. By extracting demand characteristics from power trading time series data, the characteristics of power demand at different time points can be identified. Power trading time series demand data presents information such as the changing trend of power demand and the difference between peaks and valleys, providing a basis for subsequent demand analysis and decision-making. By fitting the time series curve of power trading time series demand data, a smooth curve representation of power demand can be obtained, namely, the power demand time series curve. The power demand time series curve can more accurately reflect the changing trend and periodicity of power demand, and provide an important reference for demand forecasting and planning in the power market.

[0113] In this embodiment, multi-dimensional power trading data is collected, including relevant information such as power trading volume, price, and trading time. At the same time, power grid power flow data is obtained, including the overall situation of power supply, the power supply situation of each node, etc. The multi-dimensional power trading data is used as input to train the self-learning neural network model. The model can learn and extract the key features of the data. The trained self-learning neural network model is used to extract features of the multi-dimensional power trading data. The most representative and distinguishing dimensional features are extracted through the hidden layer of the neural network or other mechanisms. The dimensional feature power trading data and the power grid power flow data are heterogeneously fused. It is possible to consider using a data integration method to integrate the features of the two data sources to form new feature data. Time series analysis is performed on the heterogeneous dimensional fusion feature data. It is possible to consider using time series analysis methods, including stationarity tests, sequence decomposition, autoregressive models, etc., to reveal the time series and trend of the data. According to the results of the time series analysis, the heterogeneous dimensional fusion feature data is converted into power trading time series data. The data contains power trading information that changes over time, such as the trend and periodicity of trading volume and price. Demand features are extracted from power trading time series data. Consider using statistical analysis or machine learning methods to extract important features that affect electricity demand, such as seasonal demand, differences in demand between working days and non-working days, etc. Generate electricity trading time series demand data based on the results of demand feature extraction. This data reflects the changing electricity demand over time and can be used for subsequent demand forecasting and decision analysis. Use a suitable mathematical model or algorithm to fit the time series curve to the electricity trading time series demand data. Consider using methods such as polynomial regression and exponential smoothing to find the best fitting curve. Generate a power demand time series curve based on the results of time series curve fitting. This curve reflects the changing pattern of electricity demand over time and can be used for decision-making tasks such as demand forecasting and load scheduling.

[0114] In this embodiment, reference Figure 3 The above is a schematic flow chart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0115] Step S21: Analyze the demand fluctuation characteristics of the power demand time series curve using a deep learning algorithm to generate power demand fluctuation characteristic data;

[0116] Step S22: using a power dynamic demand peak calculation formula to perform dynamic demand peak calculation on the power demand fluctuation characteristic data to generate a power demand peak;

[0117] Step S23: performing demand trend analysis on the power transaction time series data according to the power demand peak value to generate power demand trend data;

[0118] Step S24: performing demand pattern analysis on the power demand trend data to generate power transaction demand pattern data;

[0119] Step S25: performing a spatiotemporal dependency analysis on the power trading demand pattern data according to the power demand time series curve to generate spatiotemporal demand dependency data;

[0120] Step S26: Perform implicit association analysis on the spatiotemporal demand dependency data to generate demand pattern association data.

[0121] The present invention analyzes the power demand time series curve through a deep learning algorithm, and can extract the fluctuation characteristics of the demand, including information such as amplitude, frequency, and cycle. The power demand fluctuation characteristic data can describe the change pattern of power demand in more detail, and provide a reference for subsequent demand forecasting and regulation. By applying the power dynamic demand peak calculation formula, the power demand fluctuation characteristic data can be processed to obtain the peak value of power demand. The power demand peak refers to the highest point reached by the demand in a certain period of time, which reflects the peak moment of power demand. The calculation of the power demand peak can help the power market and the power system to carry out load scheduling and resource allocation. According to the power demand peak, the power trading time series data is analyzed to obtain the trend information of power demand. The power demand trend data shows the overall change trend of power demand, including year-on-year growth, seasonal changes, cyclical fluctuations, etc. The power demand trend data has important reference value for the planning of the power market and the balance of supply and demand. By performing demand pattern analysis on the power demand trend data, different demand patterns can be identified. The power trading demand pattern data describes the typical patterns of power demand in different time periods, such as steady growth, sudden increase, sudden drop, etc. Power trading demand pattern data can be used for demand forecasting and market regulation, helping decision makers to better respond to different demand situations. By analyzing the power demand time series curve and combining it with the power trading demand pattern data, spatiotemporal demand dependency data can be obtained. Spatiotemporal demand dependency data reflects the correlation of power demand in time and space, such as the degree of demand correlation between different time periods and different regions. Spatiotemporal demand dependency data helps in demand forecasting and resource allocation in the power market and improves the efficiency of supply and demand matching. By performing implicit association analysis on spatiotemporal demand dependency data, the potential association rules and patterns can be discovered. Demand pattern association data reveals the relationship and mutual influence between different demand patterns, which helps to understand the complexity and interdependence of demand. Demand pattern association data provides deeper insights and decision support for demand forecasting, resource allocation and market regulation in the power market.

[0122] In this embodiment, a deep learning algorithm suitable for fluctuation feature analysis is selected, such as a recurrent neural network (RNN) or a long short-term memory network (LSTM). Constructing a deep learning model includes setting a network structure, selecting an activation function, and the like. Using the trained deep learning model, the power demand time series curve is predicted. The fluctuation feature data output by the model is extracted, which may include fluctuation amplitude, fluctuation frequency, and the like. According to the power demand fluctuation feature data, a calculation formula is designed to determine the dynamic demand peak. The selection of the formula should take into account the characteristics of power demand and the definition of the demand peak. Using the power demand fluctuation feature data and the calculation formula, the dynamic demand peak at each moment is calculated. The peak reflects the highest point of power demand and can be used for subsequent demand trend and pattern analysis. Using the power demand peak data, the demand trend analysis of the power trading time series data is performed. It can be considered to use methods such as moving average and exponential smoothing to find the overall change trend of demand. According to the results of the demand trend analysis, the power demand trend data is generated. This data reflects the overall change trend of power demand and can be used for demand forecasting and decision making. Using the power demand trend data, the demand patterns such as periodicity and seasonality are analyzed. Methods such as periodic decomposition and Fourier analysis can be applied to identify important pattern components of demand. Based on the results of the demand pattern analysis, the power trading demand pattern data is generated. This data contains different patterns of power demand and can be used to identify the laws and characteristics of demand. The time-space dependency of demand is analyzed using the power demand time series curve data and demand pattern data. Time series analysis methods, spatial statistical methods, etc. can be considered to explore the correlation and dependency of demand at different times and spatial locations. Based on the results of the time-space dependency analysis, the time-space demand dependency data is generated. This data reflects the correlation of demand in time and space and can be used to optimize power trading and supply-demand balance. Using the time-space demand dependency data, association rule mining or other machine learning methods are applied to discover the implicit correlation between demand patterns. This can help understand the inherent laws and mutual influences between demand patterns. Based on the results of the implicit association analysis, demand pattern association data is generated. This data contains the correlation between demand patterns and can be used to predict and optimize power trading strategies.

[0123] In this embodiment, the calculation formula for the peak value of the dynamic power demand in step S22 is specifically:

[0124]

[0125] P is the peak value of dynamic power demand, A is the load power, B is the power generation, C is the load rate, D is the power demand fluctuation range, n is the system load output frequency, F is the maximum voltage of the power grid system, G is the rated output voltage of the power grid system, l is the total number of power node equipment in the power grid system, b is the power peak control rate, y is the difference between the highest and lowest ambient temperatures, and x is the output efficiency of the power grid unit.

[0126] The present invention calculates the natural logarithm of the product of load power, power generation and load rate through ln(A·B·C). Load power represents the load level of the entire system, power generation represents the power generation capacity of the system, and load rate represents the ratio of load to power generation capacity. Logarithmic operation can smooth data and reduce the impact of extreme values. In the power system, load power, power generation and load rate may experience large fluctuations or extreme values. Taking logarithms can reduce the impact of these fluctuations and make the calculation results more stable. Through logarithmic operations, data of different magnitudes can be compared. Load power, power generation and load rate usually have different magnitudes. After taking the logarithm, the magnitude difference is small, which is convenient for comparing their relative sizes. By multiplying the power demand fluctuation range with the maximum voltage attenuation effect of the power grid system and then dividing it by the rated output voltage of the power grid system, the impact of voltage on the peak power demand can be adjusted. This helps to more accurately estimate the peak demand of the power system and take into account the impact of voltage changes on the peak demand. By calculating the product and taking the logarithm, multiple factors such as the total number of power node devices in the power grid system, the peak load control rate, the maximum and minimum difference in ambient temperature, and the output efficiency of the power grid units are taken into account. This helps to comprehensively evaluate the impact of these factors on the peak value of dynamic power demand. By taking the logarithm, the result of the product can be converted into a smaller numerical range. This helps to reduce the magnitude difference of the values, making the data easier to process and analyze, while also reducing the impact of extreme values. By comprehensively considering these variables, the peak demand level of the power system can be more comprehensively evaluated, and a reference can be provided for system operation and power dispatching.

[0127] In this embodiment, the specific steps of step S21 are:

[0128] Step S211: using a deep learning algorithm to perform demand amplitude analysis on the power demand time series curve to obtain demand fluctuation amplitude data;

[0129] Step S212: performing frequency domain transformation on the demand fluctuation amplitude data to generate demand fluctuation frequency domain data;

[0130] Step S213: performing waveform period detection on the demand fluctuation frequency domain data, thereby generating demand fluctuation period data;

[0131] Step S214: extracting frequency components from the demand fluctuation period data to generate demand fluctuation frequency component data;

[0132] Step S215: performing demand fluctuation characteristic analysis on the power demand time series curve according to the demand fluctuation frequency component data to generate power demand fluctuation characteristic data.

[0133] The present invention performs demand amplitude analysis on the power demand time series curve through a deep learning algorithm, and can obtain the amplitude data of demand fluctuation. The demand fluctuation amplitude data reflects the degree of fluctuation of power demand in different time periods, and can reveal the nature of fluctuation of power demand, such as stable, violent, mild, etc. These data are helpful to understand the stability and variability of power demand, and provide a basis for the regulation and control of the power market and the balance of supply and demand. By performing frequency domain conversion on the demand fluctuation amplitude data, the time domain information of demand fluctuation can be converted into frequency domain information. The demand fluctuation frequency domain data provides the contribution degree of different frequency components to demand fluctuation, and can reveal the frequency characteristics of power demand. These data are helpful to understand the periodicity and frequency distribution of power demand, and are of great significance to the load scheduling and demand forecasting of the power market. By performing waveform period detection on the demand fluctuation frequency domain data, the periodic characteristics of demand fluctuation can be determined. The demand fluctuation period data reflects the periodic changes of power demand, such as the periodicity of daily, weekly, monthly or other longer time periods. These data are helpful to understand the seasonal changes and long-term trends of power demand, and provide guidance for demand forecasting and resource planning in the power market. By extracting the frequency components of the demand fluctuation period data, the contribution degree of different frequency components to demand fluctuation can be obtained. The demand fluctuation frequency component data describes the demand fluctuations within different frequency ranges, such as the amplitude and change trend of low-frequency, medium-frequency, and high-frequency components. These data help to understand the fluctuation characteristics of electricity demand on different time scales and provide a basis for the regulation and analysis of the electricity market. According to the demand fluctuation frequency component data, the electricity demand time series curve is analyzed to extract the fluctuation characteristics of electricity demand, including information such as amplitude and change trend within different frequency ranges. The electricity demand fluctuation characteristic data can more comprehensively describe the change pattern of electricity demand, including the fluctuation nature on different time scales and the contribution of frequency components. These data are helpful for demand forecasting, load scheduling and market analysis in the electricity market, and improve the accuracy and efficiency of supply and demand matching.

[0134] In this embodiment, a deep learning algorithm suitable for demand amplitude analysis is selected, such as a convolutional neural network (CNN) or an autoencoder. A deep learning model is constructed, including defining a network structure and selecting an appropriate activation function. The deep learning model is trained using a training set. The model parameters are adjusted through a back propagation algorithm and an optimizer to minimize the amplitude prediction error. The power demand time series curve is predicted using the trained deep learning model. The amplitude data output by the model is extracted to represent the fluctuation amplitude of the power demand. The demand fluctuation amplitude data is applied to a frequency domain conversion algorithm, such as Fourier transform, to convert the time domain data into frequency domain data. Spectral data is extracted from the result of the frequency domain conversion to represent the distribution of power demand fluctuations at different frequencies. The waveform period detection of the demand fluctuation frequency domain data can use autocorrelation function, peak detection or other related methods. This will help determine the main period of demand fluctuations. According to the results of the waveform period detection, demand fluctuation period data is generated. The data represents the main period of power demand fluctuations and can be used for subsequent analysis and prediction. The frequency analysis of the demand fluctuation period data can use fast Fourier transform (FFT) or other spectrum analysis methods. The frequency components in the fluctuation period data are extracted. Based on the results of frequency component extraction, demand fluctuation frequency component data is generated. This data represents the main frequency components of power demand fluctuations and can be used for further analysis and feature extraction. Using demand fluctuation frequency component data, feature extraction is performed on the power demand time series curve. It is possible to consider extracting features such as fluctuation amplitude, energy distribution of frequency components, and phase information. Based on the results of feature extraction, power demand fluctuation feature data is generated. This data contains the fluctuation characteristics of power demand and can be used for demand analysis and forecasting.

[0135] In this embodiment, reference Figure 4 The above is a schematic flow chart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0136] Step S31: segmenting the power transaction time series data into time series to generate a power time series set;

[0137] Step S32: Sequence-marking the power time series set based on the demand pattern associated data to generate a demand pattern time series sequence;

[0138] Step S33: integrating the demand pattern time series data to construct a power demand pattern library;

[0139] Step S34: performing user transaction node identification on the multi-dimensional power transaction data to generate user transaction node data;

[0140] Step S35: Analyze the transaction behavior of the user transaction node data to generate power transaction behavior data;

[0141] Step S36: construct a topological structure based on the power transaction behavior data and user transaction node data to generate a user transaction node network.

[0142] The present invention organizes and analyzes the power transaction data in chronological order, which facilitates subsequent tasks such as demand pattern identification and transaction behavior analysis. The generated power time series sequence set provides basic data for subsequent steps, making further analysis of the power transaction data more effective and convenient. By marking the power time series sequence set, the demand pattern therein can be identified and divided. The demand pattern refers to a group of transaction sequences with similar characteristics and regularities in the power transaction data. The generated demand pattern time series sequence can be used for subsequent tasks such as demand pattern analysis and demand forecasting. By identifying and analyzing the demand pattern, the potential patterns and regularities in the power transaction data can be revealed, providing a decision-making basis for the power market and supply and demand management. By integrating the data of the demand pattern time series sequence, different demand patterns can be integrated into a power demand pattern library. The power demand pattern library is a collection of multiple demand patterns, each of which represents a class of power transaction sequences with similar demand characteristics. Constructing a power demand pattern library helps to classify and summarize power demand, and provides a more comprehensive and accurate reference for tasks such as demand forecasting and market analysis. By identifying user transaction nodes for multi-dimensional power transaction data, the nodes of different users in power transactions can be determined. User trading nodes refer to specific users in the power trading network, which play an important role in trading. The generated user trading node data can be used for subsequent tasks such as user behavior analysis and user classification. Identifying user trading nodes can better understand the structure and user behavior of the power trading network, and provide insights and decision-making basis for the power market and supply and demand management. By analyzing the trading behavior of user trading node data, the behavioral characteristics and patterns of users in power trading can be obtained. Power trading behavior data describes the trading behavior of users in the trading process, such as buying, selling, adjusting and other operations. The generated power trading behavior data can be used for subsequent tasks such as trading pattern analysis and user behavior prediction. Analyzing user trading behavior helps to understand the trading pattern and user strategy in the power market, and provide reference for market supervision and decision-making. By constructing a topological structure based on power trading behavior data and user trading node data, the connection and relationship between user trading nodes can be established. The user trading node network describes the trading relationship and connection mode between users, and can reveal the trading network structure and interaction mode between users in the power market. The generated user trading node network helps to analyze indicators such as trading flow and market concentration in the power market, and provides reference for market supervision, resource scheduling and other decisions.

[0143] In this embodiment, the power transaction time series data is segmented according to a certain time window size to form multiple time series. A sliding window method can be used, each sliding a certain time step. The segmented time series sequences are combined to form a power time series set. Each sequence represents power transaction information within a time window. For the power time series set, the demand pattern associated data is used for sequence marking. The sequence matching the known demand pattern is marked as the corresponding demand pattern. According to the result of the sequence marking, the sequence matching the demand pattern in the power time series set is extracted to form a demand pattern time series sequence. These sequences represent the occurrence of specific demand patterns in power transaction data. All demand pattern time series sequences are merged together to form a demand pattern library. Ensure that each demand pattern time series sequence has a corresponding tag for subsequent demand pattern analysis and prediction. The demand pattern library is stored as a data structure, such as a table or a database, for fast retrieval and query. Each demand pattern is associated with a corresponding tag, which can be used for the identification and analysis of the demand pattern. According to specific rules or algorithms, the user transaction node is identified for the power transaction data. Whether the transaction belongs to the same user can be judged based on dimensions such as transaction time and transaction volume. The transaction data identified as the same user are classified into the corresponding user transaction nodes. Each user transaction node contains the transaction information of the user. The user transaction node data is analyzed to study the behavior pattern of power transactions. The statistical characteristics of transaction frequency and transaction volume can be calculated, and transaction rules and abnormal transaction behaviors can be identified. Based on the results of the transaction behavior analysis, a power transaction behavior data set is generated. This data set contains the behavioral characteristics of each user transaction node and can be used for subsequent transaction behavior prediction and analysis. Based on the user transaction node data and power transaction behavior data, the topological structure of the user transaction node network is constructed. The network can be constructed based on indicators such as transaction relationships and transaction frequencies between transaction nodes. The user transaction nodes are used as nodes of the network, and the edges of the network are constructed based on the transaction relationships in the transaction behavior data. The weight of the edge can represent the transaction frequency or other related indicators. By analyzing the generated user transaction node network, network indicators such as degree centrality and betweenness centrality of the nodes can be calculated, and important nodes and community structures can be identified.

[0144] In this embodiment, the specific steps of step S4 are:

[0145] Step S41: capturing the power demand change rate of the power demand pattern library to generate power demand change rate data;

[0146] Step S42: performing mutation point detection on the power demand time series curve according to the power demand change rate data to generate demand mutation point data;

[0147] Step S43: performing demand evolution analysis on the power demand pattern library based on the demand mutation point data to generate a power demand evolution law;

[0148] Step S44: Performing dynamic association analysis on the user transaction node network according to the law of power demand evolution to generate node dynamic association data;

[0149] Step S45: performing time series change analysis on the node dynamic correlation data based on the power demand time series curve to generate time series structure evolution data;

[0150] Step S46: Perform dynamic correlation analysis on the user transaction node network through the time series structure evolution data to construct a dynamic correlation network topology structure.

[0151] The present invention can calculate and extract the rate of change information of power demand by capturing the rate of change of power demand in the power demand pattern library. The rate of change data of power demand reflects the fluctuation and change trend of power demand. The generated rate of change data of power demand can be used for subsequent demand analysis, fluctuation prediction and other tasks. Analyzing the rate of change of power demand helps to understand the dynamic characteristics and periodicity of demand, and provides a reference for the dispatch and resource optimization of the power market. By performing mutation point detection on the rate of change data of power demand, the position of the mutation point in the power demand time series curve can be determined. The demand mutation point refers to the time point at which a significant change occurs in the demand curve, which may represent a turning point or an important event in power demand. The generated demand mutation point data can be used for subsequent demand analysis, anomaly detection and other tasks. Detecting the mutation point of demand helps to capture the key moments and abnormal situations of demand changes, and provides early warning and decision-making basis for the dispatch and response of the power market. By performing demand evolution analysis on the power demand pattern library based on the demand mutation point data, the evolution law and change trend of power demand can be revealed. The law of evolution of power demand describes the characteristics of the change, cycle and trend of the demand pattern. The generated law of evolution of power demand can be used for subsequent tasks such as demand forecasting and market planning. Analyzing the evolution law of demand helps to understand the long-term trend and periodicity of demand, and provides reference and decision-making basis for the dispatch and resource optimization of the power market. By performing dynamic association analysis on the user trading node network according to the law of power demand evolution, the dynamic association relationship between nodes can be determined. The node dynamic association data describes the connection and interaction mode of user trading nodes at different stages of demand evolution. The generated node dynamic association data can be used for subsequent network analysis, node evolution prediction and other tasks. Analyzing the dynamic association relationship of nodes helps to understand the transaction mode and interaction behavior between users, and provides reference and decision-making basis for the dispatch and market design of the power market. By performing time series change analysis on the node dynamic association data based on the power demand time series curve, the time series evolution characteristics of the node dynamic association relationship can be captured. The time series structure evolution data describes the change and evolution process of the association pattern between nodes over time. The generated time series structure evolution data can be used for subsequent time series analysis, pattern recognition and other tasks. Analyzing the time series evolution characteristics of the node association relationship helps to understand the change of the association pattern between nodes over time, and provides reference and decision-making basis for the dispatch and resource optimization of the power market. The dynamic association network topology reflects the changes and evolution of the association relationship between nodes over time. The generated dynamic association network topology can be used for subsequent network analysis, key node identification and other tasks. Constructing the topology of the dynamic association network helps to understand the changes and evolution of the association pattern between nodes over time, and provides reference and decision-making basis for the dispatch and market design of the power market.

[0152] In this embodiment, for the time series data in each power demand pattern, the power demand change rate between adjacent time points is calculated. The calculated power demand change rate is associated with the corresponding time point to form power demand change rate data. Each data point contains time and corresponding change rate value. The power demand time series curve is matched with the corresponding power demand change rate data. Ensure that the time point and change rate data are aligned. The power demand change rate data is analyzed using the selected mutation point detection algorithm to identify the mutation point. The mutation point indicates the location where a significant change has occurred in the time series curve. The detected demand mutation point is associated with the corresponding time point to form demand mutation point data. Each data point contains the mark of the time and mutation point. The demand mutation point data is matched with the power demand pattern library to ensure that the time point and mutation point data are aligned. Based on the demand mutation point data, the evolution law of different demand patterns in the power demand pattern library is analyzed. Indicators such as the frequency of occurrence, duration, and evolution trend of the demand pattern can be calculated. The results of the demand evolution analysis are sorted to form the law of power demand evolution. Each law describes the evolution characteristics and regular changes of different demand patterns. The law of power demand evolution is matched with the user transaction node network to ensure that the time points and node data are aligned. Based on the law of power demand evolution, the dynamic association of the user transaction node network is analyzed. Indicators such as the degree of association, the strength of association, and the change of association between nodes can be calculated. The results of the dynamic association analysis are sorted to form node dynamic association data. Each data point contains time, node pairs, and the corresponding degree of association or strength of association. Prepare node dynamic association data, which contains time, node pairs, and the corresponding degree of association or strength of association. According to the current time point, filter the data corresponding to the time point in the node dynamic association data. Analyze the node pairs and the corresponding degree of association or strength of association in the node dynamic association data to understand the time series changes between nodes. Record the node pairs, degree of association or strength of association, and the corresponding power demand values ​​as time series structure evolution data. Traverse each time point in the time series structure evolution data. According to the node pairs and degree of association or strength of association at the current time point, the association relationship between nodes is constructed. According to the association relationship, the dynamic association network topology is updated or expanded. The node association relationship at the current time point is recorded as part of the dynamic association network topology.

[0153] In this embodiment, the specific steps of step S5 are:

[0154] Step S51: performing load calculation on the power supply flow data of the power grid to generate power supply load data;

[0155] Step S52: performing power load distribution analysis on the power supply load data according to the user transaction node network to generate power load distribution data;

[0156] Step S53: performing load intensity calculation on the power load distribution data to generate load distribution intensity data;

[0157] Step S54: performing load distribution space mining on the load distribution intensity data to generate a power load distribution map;

[0158] Step S55: Calculate the edge load of the dynamic associated network topology using the power load distribution diagram to generate edge load data;

[0159] Step S56: Perform load defect analysis on the edge load data to generate load defect data.

[0160] The present invention can estimate the current power supply load of the power grid by performing load calculation on the power supply flow data of the power grid. The power supply load data reflects the load size carried by the power grid in a specific time period. The generated power supply load data can be used for subsequent load management, power grid planning and other tasks. Calculating the power supply load helps to understand the load situation and load distribution of the power grid, and provides reference and decision-making basis for the operation and scheduling of the power system. By performing power load distribution analysis on the power supply load data according to the user transaction node network, the distribution of the power load on different nodes can be determined. The power load distribution data describes the load amount and distribution carried by different nodes. The generated power load distribution data can be used for subsequent load balancing, load forecasting and other tasks. Analyzing the distribution of the power load helps to understand the load contribution and load distribution of each node, and provides reference and decision-making basis for load management and resource optimization of the power system. Calculating the load intensity helps to understand the load situation of each node and the degree of imbalance of load distribution, and provides reference and decision-making basis for load management and resource optimization of the power system. By mining the load distribution space of the load distribution intensity data, the distribution pattern and aggregation of the power load in space can be revealed. The power load distribution diagram shows the spatial distribution of the load intensity of each node. By using the power load distribution diagram to calculate the edge load of the dynamically associated network topology, the load of each edge node in the topology can be determined. The edge load data describes the load contribution and load level of the edge node in the power system. By performing load defect analysis on the edge load data, load defects or abnormalities in the edge nodes can be identified. Load defect data describes the abnormal or overloaded load of the edge node.

[0161] In this embodiment, the load calculation is performed on the power flow data at each time point to obtain the power load data. The calculated load value is associated with the corresponding time point to form the power load data. Each data point contains the time and the corresponding load value. The power load data is matched with the user transaction node network to ensure that the time point and the node data are aligned. Traverse each time point in the power load data. According to the load value at the current time point, the load is distributed to the corresponding node. According to the connection relationship of the nodes, the power load distribution between the nodes is calculated. For each node in the power load distribution data, its load intensity is calculated. The calculated load intensity value is associated with the corresponding node to form the load distribution intensity data. Each data point contains a node and a corresponding load intensity value. According to the nodes and the corresponding load intensity values ​​in the load distribution intensity data, the spatial position of the node is determined. According to the connection relationship and load intensity value between the nodes, the connection lines between the nodes are drawn to form a power load distribution diagram. According to the results of the load distribution space mining, a power load distribution diagram is generated, in which the nodes represent power supply points and the connection lines represent the load relationship between the nodes. Prepare dynamic association network topology data, including the connection relationship between the nodes and their related attributes. Calculate the edge load based on the nodes in the edge area and their connection relationships. Factors such as the load strength of the node and the load strength of the connection line can be considered. Traverse each node in the edge load data. Compare the edge load of the node with the preset load threshold to determine whether there is a load defect. If the load exceeds the threshold, record the node and its related information as load defect data.

[0162] In this embodiment, the specific steps of step S6 are:

[0163] Step S61: optimizing the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimized network structure;

[0164] Step S62: performing supply and demand optimal trading strategy analysis on the adaptive load optimization network structure to generate an optimal trading strategy engine;

[0165] Step S63: Calculate the transaction frequency of the power transaction data to generate power transaction frequency data;

[0166] Step S64: Optimizing the frequency of the optimal trading strategy engine according to the power trading frequency data to generate a frequency optimized trading strategy engine;

[0167] Step S65: Perform data mining modeling on the frequency optimization trading strategy engine to construct a power trading strategy model and execute power trading operations.

[0168] The present invention can improve the network structure to adapt to the load demand by optimizing the load decision of the dynamic associated network topology structure using load defect data. The adaptive load optimization network structure is a network topology structure adjusted and optimized according to the load defect data. The generated adaptive load optimization network structure can improve the load balance, reliability and performance of the system. By optimizing the network structure, the load can be better distributed, the risk of load anomalies and failures can be reduced, and the operating efficiency and stability of the power system can be improved. By analyzing the optimal supply and demand trading strategy of the adaptive load optimization network structure, the optimal supply and demand trading strategy can be determined. The optimal trading strategy engine is a decision engine that provides the optimal supply and demand trading strategy based on the analysis and optimization of the network structure. By calculating the trading frequency of the power trading data, the frequency and trading activity of the transaction can be evaluated. The power trading frequency data reflects the frequency and scale of transactions in the power market. By implementing frequency optimization of the optimal trading strategy engine according to the power trading frequency data, the trading strategy engine can be adjusted to adapt to different trading frequencies. The frequency optimization trading strategy engine is a trading strategy decision engine that is adjusted and optimized according to the trading frequency data. By performing data mining modeling on the frequency optimization trading strategy engine, a power trading strategy model can be constructed for executing power trading operations. The power trading strategy model is based on data mining and modeling of the trading strategy engine, which can predict market trends, analyze trading risks and optimize trading decisions. Through data mining modeling, the laws, trends and correlations in the trading strategy engine can be extracted to provide decision support and guidance for power trading operations. Building a power trading strategy model helps optimize trading decisions, improve trading benefits and reduce risks, thereby promoting the healthy development and sustainable operation of the power market.

[0169] In this embodiment, the load defect data is analyzed to identify the location and cause of the load defect. According to the location and cause of the load defect, the dynamically associated network topology is adjusted and optimized to solve the load defect problem. The optimized network structure should take into account factors such as load balancing, capacity adjustment and reliability of power supply, and generate an adaptive load optimization network structure, including the connection relationship between nodes and the update of related attributes. The nodes in the network structure and their related attributes, including load, supply capacity, demand, etc., are analyzed. Based on the connection relationship and related attributes between the nodes, the optimal supply and demand trading strategy is determined to achieve load balancing and optimal resource utilization. Considering the timing and dynamic changes of supply and demand, the optimal trading strategy that adapts to different time periods and load conditions is formulated. According to the results of the analysis of the optimal supply and demand trading strategy, an optimal trading strategy engine is generated. The engine can automatically adjust the supply and demand trading strategy according to the real-time load conditions and demand changes. The calculated trading frequency value is associated with the corresponding time period to form power trading frequency data. Each data point contains a time period and a corresponding trading frequency value. The power trading frequency data is analyzed to understand the trading The time distribution and frequency changes of power transactions can be analyzed. According to the changes in transaction frequency, the execution frequency of the optimal transaction strategy engine is adjusted. For high-frequency transaction periods, the frequency of engine execution is increased to meet real-time needs. For low-frequency transaction periods, the frequency of engine execution is reduced to reduce resource consumption. The optimized engine should be able to automatically adjust strategies under different transaction frequencies to achieve more efficient transactions. According to the results of frequency optimization, a frequency-optimized transaction strategy engine is generated. The engine can automatically adjust the execution frequency of transaction strategies according to changes in power transaction frequency to adapt to different transaction needs. The transaction data executed by the engine is analyzed, including transaction nodes, transaction time, transaction volume and other information. According to the transaction data, features are extracted and data preprocessing is performed, including data cleaning, feature selection, etc. A power transaction strategy model is constructed using machine learning algorithms or statistical analysis methods. The model can predict transaction needs, optimize transaction strategies, make decisions on transaction behaviors, etc. According to the constructed power transaction strategy model, power transaction operations are executed. The model can make transaction decisions and optimizations based on real-time transaction data and environmental changes to achieve efficient power transaction operations.

[0170] In this embodiment, a power trading system based on multi-dimensional data analysis is provided, which is used to execute the power trading method based on multi-dimensional data analysis as described above, including:

[0171] The demand feature module is used to obtain multi-dimensional power transaction data and power grid power flow data; use a self-learning neural network to extract dimensional features of the multi-dimensional power transaction data to generate dimensional feature power transaction data; perform time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extract demand features from the power transaction time series data to generate a power demand time series curve;

[0172] The dynamic demand peak module is used to calculate the dynamic demand peak of the power demand time series curve using a deep learning algorithm to generate the power demand peak; perform demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; perform implicit association analysis on the power transaction demand pattern data to generate demand pattern association data;

[0173] The transaction node network module is used to integrate the power transaction time series data based on the demand pattern association data to build a power demand pattern library; identify the user transaction nodes of the multi-dimensional power transaction data to generate user transaction node data; construct the topological structure of the user transaction node data to generate a user transaction node network;

[0174] The dynamic association network module is used to capture the rate of change of power demand in the power demand pattern library and generate the law of power demand evolution; according to the law of power demand evolution, the user transaction node network is dynamically analyzed to construct the dynamic association network topology structure;

[0175] The load distribution module is used to perform load calculation on the power supply flow data of the power grid to generate power supply load data; perform power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; use the power load distribution map to perform edge load calculation on the dynamic associated network topology structure to generate edge load data; perform load defect analysis on the edge load data to generate load defect data;

[0176] The trading strategy model module is used to optimize the load decision of the dynamic associated network topology structure through load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand trading strategy of the adaptive load optimization network structure to generate an optimal trading strategy engine; perform data mining modeling on the optimal trading strategy engine to build a power trading strategy model and execute power trading operations.

[0177] The present invention can generate dimensional feature power trading data by acquiring multidimensional power trading data and power grid power flow data, and extracting dimensional features using a self-learning neural network. This can reduce the dimension of the data and extract key features, providing more accurate and useful data for subsequent analysis and decision-making. By using a deep learning algorithm to perform dynamic demand peak calculation on the power demand time series curve, a power demand peak can be generated. This helps to understand the peak and trough periods of power demand and provides a basis for demand pattern analysis of power trading. According to the power demand peak, the power trading time series data is analyzed for demand patterns, and power trading demand pattern data can be generated to help understand and predict the patterns and trends of trading demand. By analyzing and integrating demand pattern associated data, a power demand pattern library is constructed. This can help identify user trading nodes and construct a user trading node network. The establishment of a user trading node network helps to understand the trading relationship and trading pattern between different users, and provides a basis for the construction of the topological structure of power trading. The power demand pattern library captures the rate of change of power demand and generates the law of evolution of power demand. This helps to understand the changing trend and law of power demand. According to the law of power demand evolution, the user transaction node network is dynamically correlated and analyzed to construct a dynamic correlation network topology. This can better understand and analyze the correlation and influence relationship between different transaction nodes, and provide support for the optimization and decision-making of power transactions. The power supply flow data of the power grid is used for load calculation to generate power supply load data. The power supply load data is analyzed for power load distribution according to the user transaction node network to generate a power load distribution map. This helps to understand the load distribution in different regions and provide a reference for load balancing and resource scheduling. The power load distribution map is used to calculate the edge load of the dynamic correlation network topology to generate edge load data. This can evaluate the load of each node in the network topology and provide a basis for load decision-making and optimization. Load defect analysis of edge load data can identify load anomalies and potential failure risks, and provide guarantees for the stability and reliability of system operation. The dynamic correlation network topology is optimized for load decision-making through load defect data to generate an adaptive load optimization network structure. This helps to improve the load balancing and performance of the network and reduce the risk of load anomalies and failures. By analyzing the optimal supply and demand trading strategy for the adaptive load optimization network structure, the best supply and demand trading strategy can be determined to improve the efficiency and economy of the power market. By conducting data mining modeling on the optimal trading strategy engine, a power trading strategy model can be constructed to perform power trading operations. This can provide data-driven trading decision support and optimize the efficiency and effectiveness of power trading.

[0178] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0179] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0180] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A power trading method based on multi-dimensional data analysis, characterized in that: The following steps are involved: Step S1: acquiring multi-dimensional power transaction data and power grid power flow data; using a self-learning neural network to extract dimensional features from the multi-dimensional power transaction data to generate dimensional feature power transaction data; Perform time series analysis on the dimensional characteristic power transaction data to generate power transaction time series data; extract demand characteristics from the power transaction time series data to generate a power demand time series curve; Step S2: Calculate the dynamic demand peak value of the power demand time series curve using a deep learning algorithm to generate a power demand peak value; Performing demand pattern analysis on power transaction time series data according to power demand peaks to generate power transaction demand pattern data; Conduct implicit correlation analysis on power transaction demand pattern data to generate demand pattern correlation data; Step S3: integrating the power transaction time series data based on the demand pattern associated data to build a power demand pattern library; Identify user transaction nodes for multi-dimensional power transaction data to generate user transaction node data; construct a topological structure for the user transaction node data to generate a user transaction node network; Step S4: Capture the power demand change rate of the power demand pattern library to generate the power demand evolution law; perform dynamic correlation analysis on the user transaction node network according to the power demand evolution law, and construct a dynamic correlation network topology structure; wherein the specific steps of step S4 are: Step S41: capturing the power demand change rate of the power demand pattern library to generate power demand change rate data; Step S42: performing mutation point detection on the power demand time series curve according to the power demand change rate data to generate demand mutation point data; Step S43: performing demand evolution analysis on the power demand pattern library based on the demand mutation point data to generate a power demand evolution law; Step S44: performing dynamic association analysis on the user transaction node network according to the law of power demand evolution to generate node dynamic association data; Step S45: performing time series change analysis on the node dynamic association data based on the power demand time series curve to generate time series structure evolution data; wherein step S45 includes: Based on the law of electricity demand evolution, the dynamic association of the user transaction node network is analyzed, the association degree, association strength and association change index between nodes are calculated, the results of dynamic association analysis are sorted out to form node dynamic association data, each data point contains time, node pair and corresponding association degree or association strength, node dynamic association data is prepared, which contains time, node pair and corresponding association degree and association strength, according to the current time point, the data corresponding to the time point in the node dynamic association data is filtered, the node pairs and corresponding association degree or association strength in the node dynamic association data are analyzed to understand the time series changes between nodes, and the node pairs, association degree and association strength and the corresponding electricity demand value are recorded as time series structure evolution data; Step S46: Perform dynamic association analysis on the user transaction node network through the time series structure evolution data to construct a dynamic association network topology structure, wherein step S46 includes: Traverse each time point in the time series structure evolution data, build the association relationship between nodes according to the node pairs, association degree and association strength at the current time point, update the dynamic association network topology structure according to the association relationship, and record the node association relationship at the current time point as part of the dynamic association network topology structure; Step S5: performing load calculation on the power supply flow data of the power grid to generate power supply load data; performing power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; performing edge load calculation on the dynamic associated network topology structure using the power load distribution map to generate edge load data; performing load defect analysis on the edge load data to generate load defect data; wherein the specific steps of step S5 are: Step S51: performing load calculation on the power supply flow data of the power grid to generate power supply load data; Step S52: Performing power load distribution analysis on the power supply load data according to the user transaction node network to generate power load distribution data; wherein step S52 includes: Match the power supply load data with the user transaction node network to ensure that the time points and node data are aligned, traverse each time point in the power supply load data, distribute the load to the corresponding nodes according to the load value at the current time point, and calculate the power load distribution between the nodes according to the connection relationship of the nodes; Step S53: performing load intensity calculation on the power load distribution data to generate load distribution intensity data; wherein step S53 includes: For each node in the power load distribution data, the load intensity is calculated, and the calculated load intensity value is associated with the corresponding node to form load distribution intensity data, where each data point contains a node and a corresponding load intensity value; Step S54: performing load distribution space mining on the load distribution intensity data to generate a power load distribution map; wherein step S54 includes: According to the nodes in the load distribution intensity data and the corresponding load intensity values, the spatial positions of the nodes are determined, and according to the connection relationship between the nodes and the load intensity values, the connection lines between the nodes are drawn to form a power load distribution diagram, in which the nodes represent power supply points and the connection lines represent the load relationship between the nodes; Step S55: Calculate the edge load of the dynamic associated network topology using the power load distribution diagram to generate edge load data; wherein step S55 includes: Prepare dynamic association network topology data, including the connection relationship between nodes, and calculate edge load based on the nodes in the edge area and their connection relationship; Step S56: performing load defect analysis on the edge load data to generate load defect data, wherein step S56 includes: Traverse each node in the edge load data, compare the edge load of the node with the preset load threshold, and when the load exceeds the threshold, judge the node as load defect data; Step S6: optimize the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimization network structure; analyze the supply and demand optimal trading strategy of the adaptive load optimization network structure to generate an optimal trading strategy engine; perform data mining modeling on the optimal trading strategy engine to construct an electric power trading strategy model and execute electric power trading operations, wherein the specific steps of step S6 are: Step S61: optimizing the load decision of the dynamic associated network topology structure through the load defect data to generate an adaptive load optimized network structure; wherein step S61 includes: Analyze the load defect data, identify the location and cause of the load defect, and adjust and optimize the dynamic associated network topology according to the location and cause of the load defect to solve the load defect problem. The optimized network structure includes load balancing, capacity adjustment and reliability factors of power supply, and generates an adaptive load optimization network structure, including the connection relationship between nodes; Step S62: Analyze the optimal trading strategy of supply and demand for the adaptive load optimization network structure to generate an optimal trading strategy engine; wherein step S62 includes: Analyze the nodes and related properties in the network structure, including load, supply capacity and demand. Based on the connection relationship and related properties between nodes, determine the optimal supply and demand trading strategy to achieve load balancing and optimal resource utilization. According to the timing and dynamic changes of supply and demand, formulate the optimal trading strategy that adapts to different time periods and load conditions. According to the results of the analysis of the optimal trading strategy for supply and demand, generate the optimal trading strategy engine; Step S63: Calculating the transaction frequency of the multi-dimensional power transaction data to generate power transaction frequency data; Step S64: Optimizing the frequency of the optimal trading strategy engine according to the power trading frequency data to generate a frequency optimized trading strategy engine; Step S65: Perform data mining modeling on the frequency optimization trading strategy engine to construct a power trading strategy model and execute power trading operations, wherein step S65 includes: According to the changes in the frequency of electricity trading, the execution frequency of trading strategies is automatically adjusted to adapt to different trading needs. The trading data executed by the analysis engine, including trading nodes, trading time and trading volume information, is extracted based on the trading data and data preprocessing is performed, including data cleaning and feature selection. Using machine learning algorithms, an electricity trading strategy model is constructed. The model can be used to predict trading needs, optimize trading strategies and make decisions on trading behaviors.

2. The power trading method based on multi-dimensional data analysis according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: Acquire multi-dimensional power transaction data and power grid power flow data; Step S12: extracting dimensional features from the multi-dimensional power transaction data using a self-learning neural network to generate dimensional feature power transaction data; Step S13: performing heterogeneous data fusion on the dimensional feature power transaction data to generate heterogeneous dimensional fusion feature data; Step S14: performing time series analysis on the heterogeneous dimension fusion feature data to generate power transaction time series data; Step S15: extracting demand characteristics from the power transaction time series data to generate power transaction time series demand data; Step S16: Perform time series curve fitting on the power transaction time series demand data to generate a power demand time series curve.

3. The power trading method based on multi-dimensional data analysis according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: Analyze the demand fluctuation characteristics of the power demand time series curve using a deep learning algorithm to generate power demand fluctuation characteristic data; Step S22: using a power dynamic demand peak calculation formula to perform dynamic demand peak calculation on the power demand fluctuation characteristic data to generate a power demand peak; Step S23: performing demand trend analysis on the power transaction time series data according to the power demand peak value to generate power demand trend data; Step S24: performing demand pattern analysis on the power demand trend data to generate power transaction demand pattern data; Step S25: performing a spatiotemporal dependency analysis on the power trading demand pattern data according to the power demand time series curve to generate spatiotemporal demand dependency data; Step S26: Perform implicit association analysis on the spatiotemporal demand dependency data to generate demand pattern association data.

4. The power trading method based on multi-dimensional data analysis according to claim 3 is characterized in that: The specific formula for calculating the peak value of the dynamic power demand in step S22 is: ； is the peak value of dynamic power demand, is the load power, Power generation, is the load rate, is the fluctuation range of electricity demand, is the system load output frequency, is the maximum voltage of the power grid system, Output rated voltage for the power grid system. is the total number of power node devices in the power grid system, is the peak load control rate of electricity, is the difference between the highest and lowest ambient temperatures, is the output efficiency of the power grid unit.

5. The power trading method based on multi-dimensional data analysis according to claim 3 is characterized in that: The specific steps of step S21 are: Step S211: using a deep learning algorithm to perform demand amplitude analysis on the power demand time series curve to obtain demand fluctuation amplitude data; Step S212: performing frequency domain conversion on the demand fluctuation amplitude data to generate demand fluctuation frequency domain data; Step S213: performing waveform period detection on the demand fluctuation frequency domain data, thereby generating demand fluctuation period data; Step S214: extracting frequency components from the demand fluctuation period data to generate demand fluctuation frequency component data; Step S215: performing demand fluctuation characteristic analysis on the power demand time series curve according to the demand fluctuation frequency component data to generate power demand fluctuation characteristic data.

6. The power trading method based on multi-dimensional data analysis according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: segmenting the power transaction time series data into time series to generate a power time series set; Step S32: Sequence-marking the power time series set based on the demand pattern associated data to generate a demand pattern time series sequence; Step S33: integrating the demand pattern time series data to construct a power demand pattern library; Step S34: performing user transaction node identification on the multi-dimensional power transaction data to generate user transaction node data; Step S35: Analyze the transaction behavior of the user transaction node data to generate power transaction behavior data; Step S36: construct a topological structure based on the power transaction behavior data and the user transaction node data to generate a user transaction node network.

7. An electricity trading system based on multi-dimensional data analysis, characterized in that: The method for executing the power trading method based on multi-dimensional data analysis as claimed in claim 1 comprises: The demand feature module is used to obtain multi-dimensional power transaction data and power grid power flow data; use a self-learning neural network to extract dimensional features of the multi-dimensional power transaction data to generate dimensional feature power transaction data; perform time series analysis on the dimensional feature power transaction data to generate power transaction time series data; extract demand features from the power transaction time series data to generate a power demand time series curve; The dynamic demand peak module is used to calculate the dynamic demand peak of the power demand time series curve using a deep learning algorithm to generate the power demand peak; perform demand pattern analysis on the power transaction time series data according to the power demand peak to generate power transaction demand pattern data; perform implicit association analysis on the power transaction demand pattern data to generate demand pattern association data; The transaction node network module is used to integrate the power transaction time series data based on the demand pattern association data to build a power demand pattern library; identify the user transaction nodes of the multi-dimensional power transaction data to generate user transaction node data; construct the topological structure of the user transaction node data to generate a user transaction node network; The dynamic association network module is used to capture the rate of change of power demand in the power demand pattern library and generate the law of power demand evolution; according to the law of power demand evolution, the user transaction node network is dynamically analyzed to construct the dynamic association network topology structure; The load distribution module is used to perform load calculation on the power supply flow data of the power grid to generate power supply load data; perform power load distribution analysis on the power supply load data according to the user transaction node network to generate a power load distribution map; use the power load distribution map to perform edge load calculation on the dynamic associated network topology structure to generate edge load data; perform load defect analysis on the edge load data to generate load defect data; The trading strategy model module is used to optimize the load decision of the dynamic associated network topology structure through load defect data to generate an adaptive load optimization network structure; analyze the optimal supply and demand trading strategy of the adaptive load optimization network structure to generate an optimal trading strategy engine; perform data mining modeling on the optimal trading strategy engine to build a power trading strategy model and execute power trading operations.

Images