A visualization-based electric energy equipment and information operation and maintenance system

By introducing intelligent distribution control, data acquisition and big data analysis technologies into power grid equipment, combined with genetic algorithms and BP neural networks, the problem of insufficient performance of low-voltage distribution cabinets is solved, and the safety and data management efficiency of power grid equipment are improved. Accurate prediction and management of power grid operation.

CN113722303BActive Publication Date: 2024-12-03SHENGLONG ELECTRIC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110934508.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-08-16
Publication Date
2024-12-03
Estimated Expiration
2041-08-16

AI Technical Summary

Technical Problem

The existing low-voltage distribution cabinets have insufficient performance in terms of reliability, safety, intelligent control and modularity, resulting in high cost of power grid equipment management and low emergency response capabilities for emergencies, which makes it impossible to adapt to the demand for rapid development of the power grid.

Method used

Adopt visualization-based electrical energy equipment and information operation and maintenance systems, including intelligent distribution control and data acquisition devices, active filtering devices, fuzzy control static dynamic reactive power compensation devices, anti-accident safety closing and closing mechanical device, combined with high precision The electrical energy analysis and decision-making subsystem under intelligent data storage and management subsystem and big data technology use joint genetic algorithms and BP neural networks for data prediction and analysis, improving equipment security, reliability, data acquisition and storage efficiency.

Benefits of technology

Improve the safety, reliability, precision and versatility of power grid equipment, enhance the efficiency of data acquisition, storage and analysis, ensure the safe and reliable operation of the system, reduce false alarms and misoperation phenomena, and realize the operation of the power grid. Accurate prediction and management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113722303B_ABST
    Figure CN113722303B_ABST
Patent Text Reader

Abstract

The present invention discloses a visualization-based electric energy equipment and information operation and maintenance system, which includes an intelligent power distribution control and data acquisition device, a high-precision intelligent data storage and management subsystem, and an electric energy analysis and decision-making subsystem under big data technology. The system combines the characteristics of different types of data such as various analog quantities, component operation and equipment data, video monitoring data, technical parameters, business data, etc. in the power system, a high-precision data acquisition method, a sensor node data prediction technology, and a data access performance optimization technology with integrated caching and pre-fetching technology. The system aims to be the first in China to realize remote seamless real-time interaction of related analog quantities, switch quantities, geographic information, alarm and other information data, and provide users with a multifunctional integrated visualization interaction environment such as real-time monitoring of power data, analysis and alarm of data, prediction of sudden conditions, and remote control of distribution equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of electric power operation, and in particular to an electric energy equipment and information operation and maintenance system based on visualization. Background Art

[0002] With the development of sensor technology, automation control technology, communication network technology, distributed storage technology, big data analysis and artificial intelligence technology, as well as the huge promotion of social, economic and environmental needs, the power grid will truly achieve deep integration with the Internet. The development of "smart grid" is an inevitable development trend of the power industry. The scale of the power system is increasing, the power grid structure is becoming more and more complex, and the consequences of system failures are becoming more and more serious. Multiple major power outages at home and abroad have caused huge economic losses and adverse social impacts. The wide geographical distribution of power grid equipment, many management elements, and large amounts of information data lead to high management costs and low emergency response capabilities for emergencies. Existing equipment technology can no longer adapt to the rapid development of the power grid. It is urgent to develop key technologies for efficient and reliable power distribution and utilization systems to solve power grid accident problems. At this stage, the performance of low-voltage distribution cabinets in terms of reliability, safety, intelligent control, modularization, etc. needs to be improved. The safety and reliability of the operation of the distribution room is still based on uncertain factors such as the skills and sense of responsibility of the on-duty electricians.

[0003] Therefore, the intelligent transformation and upgrading of the power distribution system and the construction of a three-dimensional security system for power plants are in line with national industrial policies and have become the development trend of power plants in recent years. They have broad industrial application needs and important social application value. Summary of the invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides an electric energy equipment and information operation and maintenance system based on visualization.

[0005] The technical solution adopted by the present invention is that the system includes a new generation of intelligent power distribution control and data acquisition device, which improves the safety, reliability, precision, scalability and versatility of the equipment by suppressing three-phase imbalance and compensating for reactive power active filtering device, fuzzy controlled static dynamic reactive power compensation device, and distribution equipment anti-accidental safety opening and closing mechanical device.

[0006] The system includes a high-precision intelligent data storage and management subsystem. By combining the characteristics of different types of data in the power system, including various analog quantities, component operation and equipment data, video monitoring data, technical parameters, and business data, it uses high-precision data acquisition methods, sensor node data prediction technology, and data access performance optimization technology with integrated caching and pre-fetching technology to improve data acquisition, storage, and access efficiency.

[0007] The system includes an electric energy analysis and decision-making subsystem based on big data technology. According to the equipment conditions, user electricity consumption and distribution system operation conditions, it considers different types of application scenarios and different types of project conditions, studies the user's electricity load data analysis and correction methods, and designs a distribution business analysis and decision-making module to provide guidance for new projects and post-management of existing projects.

[0008] In the present invention, the high-precision intelligent data storage and management subsystem includes a sensor node data prediction algorithm combining genetic algorithm and BP neural network, a high-precision data measurement method, and a distributed massive heterogeneous data storage technology integrating pre-fetching and caching.

[0009] The research on the electric energy analysis and decision-making subsystem under the big data technology includes a user power load data analysis and correction module and a fast and scalable distribution business data analysis and decision-making method.

[0010] In the present invention, based on the importance of real-time prediction and verification of monitoring data, the real-time data collected by the sensor is tested and analyzed, the prediction technology of power station sensor monitoring data is studied, and the data prediction technology of the joint genetic algorithm and BP neural network is used to realize the prediction of the real-time collected data and verify whether the measured data is a data "bad point".

[0011] In the present invention, an improved genetic algorithm is combined with a BP neural network to form a new load prediction method, which is used to accurately predict load data. First, the improved genetic algorithm is used to optimize the initial connection weights and thresholds of the BP neural network, and then the optimized BP neural network is trained using similar historical data stored in the data center to form a data prediction model. Once the measured data is found to be a data "bad point", the upper computer monitoring system will immediately use the predicted data to replace the measured data, avoiding the sensor from sending a wrong signal and causing the system to have a false alarm and malfunction, thereby ensuring the safe and reliable operation of the system.

[0012] In the present invention, the improved genetic algorithm for optimizing the BP neural network includes:

[0013] Step A. Coding: The improved genetic algorithm is used to optimize the initial weights and thresholds of the BP neural network, and the smoothing factor is also encoded in the form of real number coding;

[0014] Step B. Population initialization;

[0015] Step C. Setting the fitness function;

[0016] Step D. Selection: Use the best retention selection method to sort the individuals in the group according to their fitness function values, and give priority to individuals with larger fitness function values ​​to be inherited to the next generation;

[0017] Step E. Crossover: Since individuals are coded with real numbers, the crossover operation method uses the real number crossover method;

[0018] Step F. Mutation: Improve the local search capability of the genetic algorithm and maintain the diversity of the population to avoid the occurrence of premature maturity;

[0019] Step G. Determine the appropriate value;

[0020] Step H. Decoding: Output the optimal solution.

[0021] In the present invention, a high-precision data measurement method analyzes the influence of the change of the operating state of the power system, especially the frequency change, on the current and voltage measurement accuracy. Considering the problems of cumbersome calculation and poor stability in the existing frequency tracking algorithms such as zero-crossing detection, Fourier transform and wavelet analysis, the Fourier transform method of periodic signals under the condition of sampling frequency mismatch is used to design a frequency tracking algorithm based on FFT transform. Among them, each channel uses up to 256 points of sampling data to form a data warehouse storage connected head to tail, adopts a 64-bit floating point FFT transform, and adds a filtering algorithm to ensure that the deviation between the measured value of current and voltage and the actual value is less than 0.2%, and designs a data measurement method with high calculation accuracy and strong anti-waveform distortion ability.

[0022] In the present invention, a distributed massive heterogeneous data storage technology integrating pre-fetching and caching is used. By analyzing different types of data structures, sizes and characteristics such as analog quantities in the power system, component operation and equipment data, video surveillance data, technical parameters, business data, etc., different storage methods are adopted for different data. A real-time file system is adopted for real-time data with very high performance requirements, monitoring system operation, and equipment data; a two-level database system is used for structured core business data and power enterprise marketing data; an HDFS distributed file system is adopted for a large amount of historical and unstructured data, video surveillance data, and enterprise management data. At the same time, multi-level caching and pre-fetching technology are introduced to improve the concurrent access efficiency of the storage system. The system can store hundreds of millions of data within 10 years, and the data access response time is accurate to seconds, thereby ensuring the carrying capacity of the network and the processing capacity of the mobile terminal.

[0023] In the present invention, the user power load data analysis and correction module aims at the analysis and correction of user power load data, including cleaning and correction of big data, clustering analysis, association and evolution analysis, using singularity detection technology particle filtering, and combining statistical models to clean, correct and filter data, including using a high-dimensional time series based on the Kalman filtering method of time series analysis to correct power load data, and cleaning and correcting the data by analyzing the correlation of user power data; using hash distributed index and distributed lock index technology to improve the data analysis speed, and using robust principal component analysis methods such as linear discriminant analysis to extract features from the data. On this basis, a clustering method for multi-structure power load data and time-based power load data, and an association rule mining method for power load and user relationships are designed, and then the data correlation relationship and behavioral causal relationship of user feature group features are analyzed, the structure and feedback of spatial economic characteristics and social behavior factors are guided, and abnormal power consumption behavior prediction and early warning are carried out.

[0024] In the present invention, a fast and scalable data analysis and decision-making method for power distribution business adopts cloud computing technology, decomposes data analysis and decision-making tasks into multiple fine-grained subtasks in a distributed parallel environment, and each subtask is scheduled and processed between idle processing nodes, and is merged through specific rules to generate a final decision result for return. It is based on the Hadoop MapReduce distributed parallel data processing cloud computing open source platform, adopts a WEB client log data association rule mining algorithm, an efficient load balancing mechanism and an optimized SPRINT parallel processing method, and uses the parallel computing power of 5000 nodes to accelerate data analysis and processing speed, and increases or decreases hardware nodes according to the actual computing resources required, reducing the cost of equipment maintenance and data backup, so as to realize power saving, power consumption prediction, power distribution network optimization and peak shifting scheduling of typical business big data analysis of intelligent power distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a system block diagram of the present invention;

[0026] Figure 2 It is a processing block diagram of the data analysis and correction technology of the present invention;

[0027] Figure 3 This is a big data analysis diagram of the intelligent power distribution and utilization of the present invention. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application may be combined with each other. The present application is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0029] A visualization-based power equipment and information operation and maintenance system, such as Figure 1As shown, the system includes an intelligent power distribution control and data acquisition device, which improves the safety, reliability, precision, scalability and versatility of the equipment by suppressing three-phase imbalance and compensating for reactive power with an active filter device, a fuzzy-controlled static dynamic reactive power compensation device, and a distribution equipment anti-accidental safety opening and closing mechanical device.

[0030] The system includes a high-precision intelligent data storage and management subsystem. By combining the characteristics of different types of data in the power system, including various analog quantities, component operation and equipment data, video monitoring data, technical parameters, and business data, it uses high-precision data acquisition methods, sensor node data prediction technology, and data access performance optimization technology with integrated caching and pre-fetching technology to improve data acquisition, storage, and access efficiency.

[0031] The system includes an electric energy analysis and decision-making subsystem based on big data technology. According to the equipment conditions, user electricity consumption and distribution system operation conditions, it considers different types of application scenarios and different types of project conditions, studies the user's electricity load data analysis and correction methods, and designs a distribution business analysis and decision-making module to provide guidance for new projects and post-management of existing projects.

[0032] In the present invention, the high-precision intelligent data storage and management subsystem includes a sensor node data prediction algorithm combining genetic algorithm and BP neural network, a high-precision data measurement method, and a distributed massive heterogeneous data storage technology integrating pre-fetching and caching.

[0033] The research on the electric energy analysis and decision-making subsystem under the big data technology includes a user power load data analysis and correction module and a fast and scalable distribution business data analysis and decision-making method.

[0034] In the present invention, based on the importance of real-time prediction and verification of monitoring data, the real-time data collected by the sensor is tested and analyzed, the prediction technology of power station sensor monitoring data is studied, and the data prediction technology of the joint genetic algorithm and BP neural network is used to realize the prediction of the real-time collected data and verify whether the measured data is a data "bad point".

[0035] Genetic algorithm is a random global search and optimization method developed by imitating the biological evolution mechanism in nature. Its essence is an efficient, parallel, global search method that mainly solves the optimization problem of the model. Improving the fitness value is to filter out abnormally large fitness values ​​in the population, because the algorithm always fluctuates around the optimal solution when it is close to it.

[0036] In order to improve the diversity of the population, the isofitness value is introduced into the genetic algorithm. The isofitness value means that there are individuals with approximately the same fitness in the population. (1) The fitness values ​​of the individuals in the population are sorted and the average fitness value is solved. (2) Individuals with fitness values ​​greater than the average fitness value are screened out as samples and selected in turn to form a new population. (3) The population is stopped when the population size is reached. The population contains only excellent individuals and individuals with small fitness are considered defective. The purpose of this is to increase the rate at which the genetic algorithm transmits information.

[0037] In the present invention, an improved genetic algorithm is combined with a BP neural network to form a new load prediction method, which is used to accurately predict load data. First, the improved genetic algorithm is used to optimize the initial connection weights and thresholds of the BP neural network, and then the optimized BP neural network is trained using similar historical data stored in the data center to form a data prediction model. Once the measured data is found to be a data "bad point", the upper computer monitoring system will immediately use the predicted data to replace the measured data, avoiding the sensor from sending a wrong signal and causing the system to have a false alarm and malfunction, thereby ensuring the safe and reliable operation of the system.

[0038] In the present invention, the steps of optimizing the BP neural network using the improved genetic algorithm are as follows:

[0039] Step 1. Coding: The improved genetic algorithm is used to optimize the initial weights and thresholds of the BP neural network, and the smoothing factor is also encoded in the form of real number coding.

[0040] Step 2. Population initialization: The initial population selected in this paper is 100.

[0041] Step 3. Fitness function: Based on the error function E of the output node of the neural network, the individual fitness function f is:

[0042] f = 1 / (E+1).

[0043] Step 4. Selection: Use the best retention selection method. Sort the individuals in the group according to their fitness function values, and give priority to selecting individuals with larger fitness function values ​​to be inherited to the next generation.

[0044] Step 5. Crossover: Since individuals are coded with real numbers, the crossover operation method uses the real number crossover method. Assuming that after k iterations, the crossover function is described as follows:

[0045]

[0046] Step 6. Mutation: The purpose of mutation is to improve the local search capability of the genetic algorithm and maintain the diversity of the population, and also to avoid the occurrence of premature maturity.

[0047]

[0048] Where: a max Gene a ij The upper bound of a min Gene a ij The lower bound of 2 (1-g / Gmax) 2 , r 2 is a random number; g is the current iteration number, G max is the maximum number of evolutions, and r is a random number between [0,1].

[0049] Step 7. Determine the appropriate value.

[0050] Step 8. Decoding: Output the optimal solution.

[0051] In the present invention, a high-precision data measurement method analyzes the influence of the change of the operating state of the power system, especially the frequency change, on the current and voltage measurement accuracy. Considering the problems of cumbersome calculation and poor stability in the existing frequency tracking algorithms such as zero-crossing detection, Fourier transform and wavelet analysis, the Fourier transform method of periodic signals under the condition of sampling frequency mismatch is used to design a frequency tracking algorithm based on FFT transform. Among them, each channel uses up to 256 points of sampling data to form a data warehouse storage connected head to tail, adopts a 64-bit floating point FFT transform, and adds a filtering algorithm to ensure that the deviation between the measured value of current and voltage and the actual value is less than 0.2%, and designs a data measurement method with high calculation accuracy and strong anti-waveform distortion ability.

[0052] In the present invention, a distributed massive heterogeneous data storage technology integrating pre-fetching and caching is used. By analyzing different types of data structures, sizes and characteristics such as analog quantities in the power system, component operation and equipment data, video surveillance data, technical parameters, business data, etc., different storage methods are adopted for different data. A real-time file system is adopted for real-time data with very high performance requirements, monitoring system operation, and equipment data; a two-level database system is used for structured core business data and power enterprise marketing data; an HDFS distributed file system is adopted for a large amount of historical and unstructured data, video surveillance data, and enterprise management data. At the same time, multi-level caching and pre-fetching technology are introduced to improve the concurrent access efficiency of the storage system. The system can store hundreds of millions of data within 10 years, and the data access response time is accurate to seconds, thereby ensuring the carrying capacity of the network and the processing capacity of the mobile terminal.

[0053] In the present invention, the user power load data analysis and correction module aims at the analysis and correction of user power load data, including cleaning and correction of big data, clustering analysis, association and evolution analysis, using singular point detection technology particle filtering, and combining statistical models to clean, correct and filter data, including using high-dimensional time series based on time series analysis Kalman filtering method to correct power load data, by analyzing the correlation of user power data, to clean and correct the data; using hash distributed index and distributed lock index technology to improve data analysis speed, using linear discriminant analysis and other robust principal component analysis methods to extract data features, on this basis, design multi-structure power load data and time-based power load data clustering methods, power load and user relationship-oriented association rule mining methods, and then analyze the data association relationship and behavioral causal relationship of user feature group features, guide the structure and feedback of spatial economic characteristics and social behavior factors, and predict and warn abnormal power consumption behavior. The main methods and technologies used are as follows Figure 2 shown.

[0054] In the present invention, Figure 3 As shown in the figure, a fast and scalable data analysis and decision-making method for power distribution business adopts cloud computing technology to decompose data analysis and decision-making tasks into multiple fine-grained subtasks in a distributed parallel environment. Each subtask is scheduled and processed between idle processing nodes, and is merged through specific rules to generate the final decision result and return it. It is based on the Hadoop Map Reduce distributed parallel data processing cloud computing open source platform, adopts the WEB client log data association rule mining algorithm, efficient load balancing mechanism and optimized SPRINT parallel processing method, and uses the parallel computing power of 5000 nodes to accelerate the data analysis and processing speed. It also increases or decreases hardware nodes according to the actual computing resources needed, reduces the cost of equipment maintenance and data backup, and realizes power saving, power consumption prediction, distribution network optimization and peak-shifting scheduling of typical business big data analysis of intelligent power distribution.

[0055] Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A visualization-based electric energy equipment and information operation and maintenance system, characterized in that ,The system includes intelligent power distribution control and data acquisition device, ,high-precision intelligent data storage and management subsystem, and power ,analysis and decision making subsystem based on big data technology; The intelligent power distribution control and data acquisition device includes an active filter device for reactive power compensation, a fuzzy-controlled static dynamic reactive power compensation device, and a mechanical device for preventing accidental safety opening and closing of power distribution equipment; The high-precision intelligent data storage and management subsystem, which processes the data characteristics of various analog quantities and component operations in the power system, as well as equipment data, video monitoring data, technical parameters, and business data, using the sensor node data prediction algorithm of the joint genetic algorithm and BP neural network, the high-precision data measurement method, and the distributed massive heterogeneous data storage technology with integrated pre-fetching and caching; Among them, the improved genetic algorithm is combined with the BP neural network to form a new load prediction method. This method is used to accurately predict the load data. First, the improved genetic algorithm is used to optimize the initial connection weights and thresholds of the BP neural network. Then, the optimized BP neural network is trained using similar historical data stored in the data center to form a data prediction model. Once the measured data is found to be a "bad point" in the data, the upper computer monitoring system will immediately use the predicted data to replace the measured data to avoid the sensor sending a wrong signal and causing the system to have a false alarm and malfunction. The high-precision data measurement method includes: analyzing the influence of the change of the operation state of the power system on the measurement accuracy of the current and voltage, using the Fourier transform method of the periodic signal under the condition of sampling frequency mismatch, and the frequency tracking algorithm based on the FFT transform to form the data warehouse storage connected head to tail, using the 64-bit floating point FFT transform, and adding the filtering algorithm to ensure that the deviation between the measured value and the actual value of the current and voltage is less than 0.2%; the change of the operation state of the power system includes the frequency change; The distributed massive heterogeneous data storage technology with integrated pre-fetching and caching includes: by analyzing the data structure, size and characteristics of analog quantities, component operation and equipment data, video surveillance data, technical parameters and business data of the power system, different storage methods are used for different data. For real-time data with very high performance requirements, real-time file system is used for monitoring system operation and equipment data; a two-level database system is used for structured core business data and power enterprise marketing data; HDFS distributed file system is used for a large amount of historical and unstructured data, video surveillance data and enterprise management data. At the same time, multi-level cache and pre-fetching technology are introduced to improve the concurrent access efficiency of the storage system. The system stores hundreds of millions of data in 10 years, and the data access response time is accurate to seconds. Among them, the steps of using the improved genetic algorithm to optimize the BP neural network are: Step 1. Coding: using an improved genetic algorithm to optimize the initial weights and thresholds of the BP neural network, and encoding the smoothing factor of the BP neural network in the form of real number coding; Step 2. Population initialization: the initial population is 100; Step 3. Fitness function: The error function E of the output node of the neural network, the individual fitness function f is: f = 1 / (E+1); Step 4. Selection: Use the best retention selection method; sort the individuals in the group according to their fitness function values, and give priority to selecting the individuals with the largest fitness function value to be inherited to the next generation; Step 5. Crossover: Individuals are coded with real numbers, and the crossover operation method uses the real number crossover method. Assuming that after k iterations, the crossover function is described as follows: Step 6. Mutation: The mutation is described as follows: Where: a max Gene a ij The upper bound of a min Gene a ij The lower bound of f(g) = r2(1-g / G max ) 2 , r2 is a random number; g is the current iteration number, G max is the maximum number of evolutions, r is a random number between [0,1]; Step 7. Determine the appropriate value; Step 8. Decoding: output the optimal solution; The electric energy analysis and decision-making subsystem under the big data technology, according to the equipment conditions, user power consumption and distribution system operation conditions, takes into account different types of application scenarios and different types of project conditions, studies the user power load data analysis and correction methods, and designs the distribution business analysis and decision module. The distribution business analysis and decision module is used for the guidance of new projects and the post-management of existing projects.

2. The visualization-based electric energy equipment and information operation and maintenance system according to claim 1, characterized in that: The electric energy analysis and decision-making subsystem under the big data technology uses a user power load data analysis and correction module and a fast and scalable distribution business data analysis and decision-making method to complete electric energy analysis and decision-making.

3. The visualization-based electric energy equipment and information operation and maintenance system according to claim 1, characterized in that: Based on the importance of real-time prediction and verification of monitoring data, the real-time data collected by sensors is tested and analyzed to study the prediction technology of power station sensor monitoring data. The data prediction technology of joint genetic algorithm and BP neural network is used to predict the real-time collected data and verify whether the measured data is a "bad point" of data.

4. A visualization-based electric energy equipment and information operation and maintenance system as described in claim 2, characterized in that the user power load data analysis and correction module includes cleaning and correction of big data, cluster analysis, and correlation and evolution analysis; using singular point detection technology particle filtering, combined with statistical models to clean, correct and filter data; using high-dimensional time series based on time series analysis Kalman filtering method to correct power load data, by analyzing the correlation of user power data, the data is cleaned and corrected; using hash distributed index and distributed lock index technology to improve data analysis speed; Linear discriminant analysis and robust principal component analysis were used to extract features from the data; Based on the clustering method of multi-structure power load data and time-based power load data, and the association rule mining method for power load and user relationship, the data association relationship and behavioral causal relationship of user feature group characteristics are analyzed, the structure and feedback of spatial economic characteristics and social behavior factors are guided, and abnormal power consumption behavior is predicted and warned.

5. A visualization-based electric energy equipment and information operation and maintenance system as described in claim 1, characterized in that the fast and scalable distribution business data analysis and decision-making method adopts cloud computing technology to decompose the data analysis and decision-making task into multiple fine-grained subtasks in a distributed parallel environment. Each subtask is scheduled and processed between idle processing nodes, and merged to generate the final decision result and return it. It is based on the Hadoop Map Reduce distributed parallel data processing cloud computing open source platform, adopts the WEB client log data association rule mining algorithm, efficient load balancing mechanism and optimized SPRINT parallel processing method, and uses the parallel computing power of 5000 hardware nodes to accelerate the data analysis and processing speed, and increases or decreases the 5000 hardware nodes according to the actual computing resources needed.