Power grid business hall terminal data management method and system based on cloud computing

Abstract

The present invention discloses a data management method and system for power grid business hall terminals based on cloud computing, which relates to the field of data management technology, including: establishing digital communication with power grid business hall terminals, configuring communication link sets, and establishing a link history database; building a link feature analysis model; obtaining real-time usage data of power grid business hall terminals, making usage predictions based on time series networks, establishing prediction data, performing usage analysis of communication link sets, and establishing link feedback; obtaining real-time processor resource usage data, performing dynamic resource reallocation, and establishing an adjustable window for dynamic allocation results of links; performing feature adaptation analysis under link feedback, and establishing fine-tuning coefficients; performing fine-tuning processing of dynamic allocation results of links, and completing data management of power grid business hall terminals based on the fine-tuning processing results. The present invention solves the technical problem that the prior art cannot meet the requirements of efficient data processing and real-time response, and achieves the technical effect of improving data processing speed.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a method and system for managing terminal data in a power grid business hall based on cloud computing. Background Art

[0002] With the rapid development of the power industry and the continuous expansion of the power grid, the data management of power grid business hall terminals has become increasingly important and complex. With the increase in the number of terminals and the diversification of business, the amount of data that power grid business halls need to process has exploded, which has put forward higher requirements for the accuracy, real-time and stability of data management.

[0003] Traditional methods of managing data at power grid business hall terminals usually rely on local storage and processing, which can be used when the data volume is small and the business is simple. However, in today's increasingly complex and large power grid system, traditional methods can no longer meet the needs of efficient data processing and real-time response. Summary of the invention

[0004] The present application provides a cloud computing-based power grid business hall terminal data management method and system, which is used to solve the technical problem that the existing technology cannot meet the requirements of efficient data processing and real-time response.

[0005] In view of the above problems, the present application provides a power grid business hall terminal data management method and system based on cloud computing.

[0006] In a first aspect of the present application, a method for managing terminal data of a power grid business hall based on cloud computing is provided, the method comprising:

[0007] Establish digital communication with the power grid business hall terminal, configure a communication link set based on the digital communication, establish a link history database for each link in the communication link set, and the link history database records the performance data and stability data of the corresponding link; perform data preprocessing on the link history database, and use the data preprocessing result as training data to build a link feature analysis model; obtain real-time usage data of the power grid business hall terminal, and perform usage prediction based on the time series network to establish prediction data, perform usage analysis of the communication link set based on the real-time usage data and the prediction data, and establish link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback; obtain resource usage data of the real-time processor, use the link feedback as a matching feature, perform dynamic reallocation of resources based on the resource usage data, and establish an adjustable window for the dynamic allocation results of each link; perform feature adaptation analysis under link feedback based on the link feature analysis model, and establish a fine-tuning coefficient based on the adaptation analysis result; perform fine-tuning processing of the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and complete data management of the power grid business hall terminal based on the fine-tuning processing result.

[0008] The second aspect of the present application provides a cloud computing-based power grid business hall terminal data management system, the system comprising:

[0009] A digital communication establishment module, which establishes digital communication with the power grid business hall terminal, configures a communication link set based on the digital communication, and establishes a link history database for each link in the communication link set, wherein the link history database records the performance data and stability data of the corresponding link; a link feature analysis model construction module, which performs data preprocessing on the link history database, and uses the data preprocessing result as training data to construct a link feature analysis model; a link feedback establishment module, which obtains real-time usage data of the power grid business hall terminal, and performs usage prediction based on the time series network, establishes prediction data, and performs usage prediction of the communication link set based on the real-time usage data and the prediction data. Use analysis to establish link feedback, wherein the link feedback includes real-time link feedback and predicted link feedback; a resource dynamic allocation module, the resource dynamic allocation module obtains the resource usage data of the real-time processor, uses the link feedback as a matching feature, dynamically reallocates resources based on the resource usage data, and establishes an adjustable window for the dynamic allocation results of each link; a feature adaptation analysis module, the feature adaptation analysis module performs feature adaptation analysis under link feedback based on the link feature analysis model, and establishes a fine-tuning coefficient according to the adaptation analysis result; a data management module, the data management module performs fine-tuning processing on the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and completes the data management of the power grid business hall terminal based on the fine-tuning processing result.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The present application establishes digital communication with the power grid business hall terminal, configures a communication link set based on digital communication, establishes a link history database of each link in the communication link set, and the link history database records the performance data and stability data of the corresponding link; performs data preprocessing on the link history database, and uses the data preprocessing results as training data to build a link feature analysis model; obtains real-time usage data of the power grid business hall terminal, and performs usage prediction based on the time series network to establish prediction data, performs usage analysis of the communication link set based on the real-time usage data and the prediction data, and establishes link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback; obtains resource usage data of the real-time processor, uses link feedback as a matching feature, performs dynamic reallocation of resources based on the resource usage data, and establishes an adjustable window for the dynamic allocation results of each link; performs feature adaptation analysis under link feedback based on the link feature analysis model, and establishes a fine-tuning coefficient based on the adaptation analysis results; performs fine-tuning processing of the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and completes data management of the power grid business hall terminal based on the fine-tuning processing results. The present invention solves the technical problem that the prior art cannot meet the requirements of efficient data processing and real-time response, and achieves the technical effect of improving data processing speed through cloud-based centralized management, real-time monitoring and intelligent analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A schematic flow chart of a method for managing terminal data in a power grid business hall based on cloud computing provided in an embodiment of the present application.

[0014] Figure 2 A schematic diagram of the structure of a power grid business hall terminal data management system based on cloud computing provided in an embodiment of the present application.

[0015] Explanation of the accompanying drawings: digital communication establishment module 11, link feature analysis model construction module 12, link feedback establishment module 13, resource dynamic allocation module 14, feature adaptation analysis module 15, data management module 16. DETAILED DESCRIPTION

[0016] This application provides a cloud computing-based power grid business hall terminal data management method and system to solve the technical problem that the existing technology cannot meet the requirements of efficient data processing and real-time response. Through cloud-based centralized management, real-time monitoring and intelligent analysis, the technical effect of improving data processing speed is achieved.

[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0018] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0019] Embodiment 1

[0020] like Figure 1 As shown, the present application provides a power grid business hall terminal data management method based on cloud computing, the method comprising:

[0021] Step S100: Establish digital communication with the power grid business hall terminal, configure a communication link set based on the digital communication, and establish a link history database for each link in the communication link set, wherein the link history database records performance data and stability data of the corresponding link.

[0022] In the embodiment of the present application, when establishing digital communication with the terminal of the power grid business hall, first confirm the network connection between the terminal device and the data management system to ensure the connection at the physical layer. Next, configure the network parameters, such as IP address, subnet mask, gateway, etc., so that the terminal device and the management system can communicate with each other in the same network. Finally, use a security protocol such as TLS / SSL for communication encryption to ensure the security and integrity of data transmission. Through the above process, digital communication with the terminal of the power grid business hall is established.

[0023] Identify and classify the required communication links according to the different functional requirements of the power grid business hall terminals. For example, distinguish between links with large data transmission volume and high bandwidth requirements and links with low bandwidth and delay sensitivity. Allocate appropriate network resources, such as bandwidth, priority, etc. to each link. Use network configuration tools, such as router configuration and switch configuration, to set link parameters. Deploy network monitoring tools, such as SNMP protocol and network monitoring software to monitor the status of each link in real time.

[0024] Use a database management system such as MySQL or PostgreSQL to establish a link history database. The database table structure includes fields such as link ID, performance data, and stability data. Use network performance monitoring tools such as Wireshark and Nagios to collect performance and stability data for each link. Use API or database connection tools to store the collected data in real time in the link history database.

[0025] The link history database records the performance data and stability data of the corresponding link. Performance data includes bandwidth usage, delay time, data transmission rate, etc. Stability data includes connection stability, fault frequency, recovery time, etc.

[0026] Step S200: performing data preprocessing on the link history database, and using the data preprocessing results as training data to construct a link feature analysis model.

[0027] In an embodiment of the present application, when data preprocessing is performed on the link history database, data cleaning is first performed to remove duplicate data to ensure that each record in the link history database is unique. For missing values ​​in link performance or stability data, an appropriate filling strategy is adopted, such as using the average, median or nearest neighbor value of historical data for filling. Based on statistical methods, such as the IQR rule, outliers are identified, and outliers are replaced or deleted as appropriate. Next, data conversion is performed to convert the timestamp into a time format suitable for model processing, such as date, time difference, etc. For link performance data of different dimensions, such as transmission speed, delay, etc., normalization processing is performed, and the data is scaled to a specific range, such as [0,1] or standard normal distribution, through Min-Max normalization, standardization and other methods. Next, statistics such as the average transmission speed, maximum delay, and packet loss rate of the link are calculated as input features of the model.

[0028] Next, the preprocessed data is divided into a training set and a test set. The preprocessed data is divided into 80% training set and 20% validation set by random selection.

[0029] When building a link feature analysis model, select a random forest model for training, use the training set to train the random forest model, use the validation set to validate the model, evaluate the generalization ability and prediction performance of the model, and adjust the model parameters according to the validation results. The link feature analysis model is constructed through the above process. This model is used to analyze and predict the performance and stability of the communication link.

[0030] Step S300: Acquire real-time usage data of the power grid business hall terminal, and make usage predictions based on the time series network to establish prediction data, perform usage analysis of the communication link set based on the real-time usage data and the prediction data, and establish link feedback, wherein the link feedback includes real-time link feedback and predicted link feedback.

[0031] In the embodiment of the present application, a monitoring system such as Zabbix or Nagios is used to deploy agents on the terminals of the power grid business hall to collect real-time usage data. These systems monitor network traffic, bandwidth usage, connection stability and other indicators.

[0032] Next, use prediction is performed based on the time series network. When making predictions, historical use data is first extracted from the database, and a suitable time series model is selected. For example, the LSTM model is used in this application. The model is trained using historical use data to obtain a trained time series model. The real-time use data is input into the trained time series model for prediction to obtain usage prediction data for a period of time in the future.

[0033] Then, the real-time usage data and predicted data are combined to analyze the usage of the communication link set. First, the real-time usage data and predicted data are combined to analyze the usage of each communication link. Key performance indicators such as bandwidth utilization, latency, and packet loss rate are calculated. For example, bandwidth utilization is calculated by the ratio of actual bandwidth used to total bandwidth, latency is measured by the round-trip time of data packets, and packet loss rate is calculated by the difference between the number of sent data packets and the number of received data packets. Finally, link feedback is established based on the usage analysis. Link feedback includes real-time link feedback and predicted link feedback. Real-time link feedback is based on real-time data and provides real-time feedback on the current usage and performance of the link. Predicted link feedback is based on predicted data and predicts the usage and performance trend of the link in the future.

[0034] Step S400: Acquire resource usage data of the real-time processor, use the link feedback as a matching feature, dynamically reallocate resources based on the resource usage data, and establish an adjustable window of dynamic allocation results for each link.

[0035] In an embodiment of the present application, a system monitoring tool such as Prometheus, Grafana, etc. is used to deploy an agent on the processor to monitor the usage of resources such as CPU, memory, disk I / O, and network I / O in real time.

[0036] Integrate the acquired real-time resource usage data with the link feedback data. Use the link feedback as feature input and extract key features such as bandwidth utilization, latency, and packet loss rate through feature engineering to analyze the resource requirements of each link. Based on the integrated data, use rule algorithms such as threshold method or priority algorithm to allocate resources. For example, when using the threshold method for resource allocation, define the upper threshold of CPU usage as 80%. When the CPU usage of a link exceeds 80%, increase the CPU resources of the link. Define the lower threshold of memory usage as 20%. When the memory usage of a link is less than 20%, reduce the memory resources of the link.

[0037] Based on link feedback and resource usage data, the initial resource allocation for each link is determined through an optimization algorithm. The resource allocation range for each link is defined, and the minimum and maximum values ​​of resources such as CPU, memory, and bandwidth are set. Sliding window techniques such as Exponential Moving Average are used to adjust the range of the adjustable window so that it can adapt to real-time changes. Through the above process, an adjustable window of dynamic allocation results for each link is established.

[0038] Step S500: performing feature adaptation analysis under link feedback based on the link feature analysis model, and establishing a fine-tuning coefficient according to the adaptation analysis result.

[0039] In the embodiment of the present application, the link feedback data is input into the link feature analysis model to obtain the corresponding performance index. After obtaining the link performance index, feature adaptation analysis is performed. This includes comparing the link performance index with a preset threshold or standard to determine whether the link performance meets the requirements. If the link performance does not meet the requirements, it is necessary to further analyze which features need to be adjusted, as well as the direction and degree of adjustment.

[0040] Based on the results of the feature adaptation analysis, determine the link features that need to be fine-tuned. These features include transmission speed, delay, packet loss rate, etc. For each link feature that needs to be fine-tuned, quantify the degree of fine-tuning. For example, this can be achieved by calculating the difference, ratio, or other mathematical formula between the actual performance and the expected performance. Specifically, if the expected transmission speed is 1Gbps, but the actual speed is 0.8Gbps, then the degree of fine-tuning is (1Gbps-0.8Gbps) / 1Gbps=0.2, which means that the transmission speed needs to be increased by 20%. Convert the quantized degree of fine-tuning into a fine-tuning coefficient. The fine-tuning coefficient is a numerical value used to guide subsequent resource allocation or parameter adjustment. For example, if the quantized degree of fine-tuning is 20%, then the fine-tuning coefficient is 0.2.

[0041] Step S600: fine-tune the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and complete the data management of the power grid business hall terminal based on the fine-tuning processing results.

[0042] In the embodiment of the present application, the direction and degree of adjustment required for each link are determined according to the fine-tuning coefficient. For example, if the fine-tuning coefficient indicates that the bandwidth of a certain link needs to be increased, the bandwidth allocation of the link is adjusted within the adjustable window. Next, check whether the adjusted resource allocation meets the business needs and network performance requirements. If so, keep the current configuration; if not, continue to adjust until the optimal solution is found or the adjustment limit is reached.

[0043] After fine-tuning, the data of the power grid business hall terminal is managed according to the new resource allocation plan. Data management includes optimizing the data transmission path, increasing the data transmission speed, reducing the delay and packet loss rate, etc.

[0044] Furthermore, step S400 in the method provided in the application embodiment also includes:

[0045] Acquire link basic data in the communication link set, wherein the link basic data is the topological structure data of the link; perform network topology dependency analysis of the link based on the link basic data, and construct a first link key value; parse the link feedback, acquire link flow data and link service level agreement data, and establish a second link key value based on the link flow data and the link service level agreement data; normalize the first link key value and the second link key value, and then dynamically reallocate resources based on the normalization result and the resource usage data.

[0046] In the embodiment of the present application, the basic link data in the communication link set is collected through network management protocols such as SNMP, NetFlow, etc. The basic link data includes the topological structure data of the link, such as the connection relationship of the link, routing information, device location, etc.

[0047] Based on the link connection relationship and routing information, identify the key links in the network, such as backbone links, core switch connections, etc. Evaluate the impact of these link failures on the overall performance of the network. Identify redundant links in the network, that is, links that can provide backup in the event of a network failure. Analyze the bandwidth, connection devices and other information of redundant links to evaluate their backup capabilities. Select indicators to evaluate the importance of links, such as connectivity, bandwidth utilization, redundancy, etc. Set weights for each indicator, such as 0.5 for connectivity, 0.3 for bandwidth utilization, and 0.2 for redundancy. Based on the performance of the link on each indicator, such as 1 for connectivity means full connectivity and 0 for disconnection; bandwidth utilization is the ratio of actual bandwidth used to total bandwidth, etc., combine the corresponding weights and calculate the weighted sum as the first link key value.

[0048] Next, analyze the link feedback and extract link traffic data and SLA data from the real-time data. Perform statistics on link traffic data, such as calculating average traffic and peak traffic. Perform statistics on SLA data, such as calculating SLA compliance rate and average response time. Based on the statistical results, set a scoring system for each link, such as 5 points for an SLA compliance rate above 95%, 4 points for 80%-95%, etc. Combine the scores of each item and calculate the weighted sum or weighted product as the second link key value.

[0049] When normalizing the first link key value and the second link key value, a normalized value range is set, such as 0-1 or 1-100. Use linear normalization or maximum and minimum normalization methods to map their values ​​to the set value range.

[0050] Based on the normalized key values ​​and link resource usage data, such as CPU utilization and memory usage, a dynamic resource allocation algorithm, such as a load balancing algorithm, is used to allocate resources.

[0051] Further, the network topology dependency analysis of the link is performed based on the link basic data to construct the first link key value, and the method further includes:

[0052] Perform a connectivity analysis based on the topological structure on the link basic data to establish a structural key value; perform communication association monitoring based on the link basic data, and establish an association key value based on the communication association monitoring result; perform weighted calculation on the structural key value and the association key value to construct a first link key value.

[0053] In the embodiment of the present application, the topological structure data such as the connection relationship, routing information, and device location of the link are extracted from the basic link data. The network topology graph is traversed using the depth-first search or breadth-first search algorithm in graph theory. The reachability between each node in the network is determined, that is, whether there is a path that allows two nodes to communicate with each other.

[0054] For each link, calculate its betweenness as the structural key value. Betweenness refers to the ratio of the number of shortest paths passing through the link to the total number of all shortest paths in the network. The higher the betweenness, the more important the link is in the network.

[0055] By real-time monitoring of network traffic, data packet transmission and other information, the communication data of the link is collected. The data includes the source address, destination address, transmission protocol, communication volume, communication frequency, etc. of the data packet.

[0056] Analyze the collected data packets, identify the source and destination addresses, and determine which links have data communication. Count the communication volume of each link, such as the number of data packets or bytes passing through the link per unit time. Analyze the communication frequency of the link, that is, the number of times the link is used per unit time.

[0057] Next, set up a scoring system for the traffic volume and communication frequency. For example, divide the traffic volume into several levels, each level corresponds to a score, such as traffic volume above 100MB / s is 5 points, 50-100MB / s is 4 points, and so on. According to the scoring system, score the traffic volume and communication frequency of each link, and add the two together, or set different weights according to business needs and perform weighted summation to obtain the associated key value of the link.

[0058] According to the characteristics of the network and business needs, set the weight of the structural key value. For example, if the stability of the network structure is critical to the business, give the structural key value a higher weight, such as 0.6. Similarly, set the weight of the associated key value according to business needs, such as 0.4.

[0059] Next, we perform weighted calculation: the first link key value = structural key value × structural weight + associated key value × associated weight. Assuming that the structural key value of a link is 0.8, with a full score of 1, the associated key value is 4, with a full score of 5, the structural weight is 0.6, and the associated weight is 0.4, then the first link key value of the link = 0.8 × 0.6 + 4 × 0.4 = 0.48 + 1.6 = 2.08.

[0060] The importance of the links is evaluated and ranked according to the calculated first link key values.

[0061] Furthermore, step S500 in the method provided in the application embodiment also includes:

[0062] Based on the link feedback, link flow data and link service level agreement data are obtained; based on the feature analysis model, link features are called, and adaptation analysis of link task execution is performed through the link features and the link flow data and link service level agreement data to generate the adaptation analysis result.

[0063] In the embodiment of the present application, link flow data and link service level agreement data are obtained from link feedback. Link flow data includes real-time flow information of the link, such as the number of packets transmitted per second, the number of bytes transmitted per second, etc. Link service level agreement data includes key indicator data about link performance, such as response time, availability, throughput, etc.

[0064] The acquired link traffic data and link service level agreement data are preprocessed to ensure the accuracy and availability of the data. The preprocessing steps include data cleaning, missing value processing, outlier processing and data conversion.

[0065] Next, the status information of the current link is input into the feature analysis model to call the link features and obtain key indicators such as link bandwidth, delay, jitter, and packet loss rate.

[0066] When performing adaptation analysis for link task execution, first determine the specific requirements of the task for link performance, such as the data transmission speed to be achieved, the maximum allowable delay, etc. These requirements are pre-set. Next, compare the actual performance of the link, i.e., based on link characteristics, traffic data, and SLA data, with the task requirements. For example, if the data transmission speed required by the task is 1Gbps, and the actual bandwidth of the link is only 500Mbps, the link cannot meet the task requirements. If the performance of the link fully meets the task requirements, the task is directly executed on the link. If the performance of the link partially meets the requirements, such as the bandwidth is slightly lower than the requirement but the delay is within an acceptable range, you can consider optimizing the execution method of the task, such as adjusting the packet size, reducing concurrent requests, etc., to reduce the link load. If the current link cannot meet the task requirements and there are other available links, the task can be transferred to other links for execution. For example, task A requires a bandwidth of 1Gbps and a delay of 10ms. The bandwidth of link 1 is 1.2Gbps, but the delay is 20ms. In this case, choose to optimize the execution method of task A, such as reducing the packet size to reduce the dependence on delay, and execute the task on link 1. Task B requires 99.99% availability. Link 2 has an availability of 99.9%, but has a short history of failures. In this case, Task B is transferred to Link 3, which has higher availability.

[0067] Through the above adaptation analysis, the adaptation analysis results are generated. The adaptation analysis results include link status, task requirements, adaptation decisions and recommended measures. The link status describes the current performance status of the link in detail, including the specific values ​​of key indicators such as bandwidth, delay, jitter, and packet loss rate. The task requirements clearly list the specific requirements of the task for link performance. The adaptation decision describes the process and results of the adaptation decision in detail, including the selected adaptation strategy, such as direct execution, optimization adjustment, task transfer and its reasons. The recommended measures provide specific recommended measures based on the results of the adaptation analysis, such as optimizing the task execution method, upgrading network equipment, etc.

[0068] Furthermore, step S300 in the method provided in the application embodiment also includes:

[0069] A usage database of the power grid business hall terminal is established, and the usage database is divided into time series periods to establish a time series data set; a first prediction subnetwork is constructed based on the time series data set, wherein the first prediction subnetwork is a processing subnetwork of the time series network; a second prediction subnetwork is established, wherein the second prediction subnetwork is a processing subnetwork of the time series network, and the second prediction subnetwork is communicatively connected with the first prediction subnetwork to receive real-time data to perform prediction correction; usage prediction is completed based on the first prediction subnetwork and the second prediction subnetwork in the time series network, and prediction data is established.

[0070] In the embodiment of the present application, user usage data is continuously collected from the log system and application of the power grid business hall terminal. These data include the number of user visits, visit duration, operation type, operation time, etc.

[0071] These data are stored in the database according to a unified data format and timestamp, and a usage database for power grid business hall terminals is established.

[0072] According to the operational characteristics of the power grid business hall, select appropriate time intervals, such as hours, days, weeks, months, etc. to divide the time series period. Arrange the data in chronological order to form a time series data set.

[0073] The first prediction subnetwork is responsible for learning the inherent laws and patterns of data from the time series data set. You can choose to use time series models such as recurrent neural networks, long short-term memory networks, gated recurrent units, etc. to construct the first prediction subnetwork. In this application, a long short-term memory network is used to construct the first prediction subnetwork. During the construction process, the first prediction subnetwork is trained using historical time series data. A supervised learning method is used to take past time series data as input and actual usage data as labels to train model parameters. Use loss functions (such as mean square error) to evaluate model performance, and optimize model parameters through back propagation algorithms. Cross-validation is performed during training to ensure the generalization ability of the model. After training, the first prediction subnetwork can generate preliminary prediction results, such as the number of visits and average visit duration in the next hour.

[0074] Next, the second prediction subnetwork is established. The second prediction subnetwork is also a processing subnetwork of the time series network, which is used to receive real-time data and perform prediction correction. The second prediction subnetwork is connected to the first prediction subnetwork and receives real-time data, such as user access data at the current time node, as input. Based on the real-time data and the preliminary prediction results of the first prediction subnetwork, the second prediction subnetwork performs prediction correction. Similar to the first prediction subnetwork, the second prediction subnetwork uses real-time data and preliminary prediction results as input, and actual usage data as labels for training. The establishment of the second prediction subnetwork is completed through training.

[0075] Finally, the first prediction subnetwork is used to predict the time series data set to generate preliminary prediction results. These prediction results usually cover a longer period of time, such as one day, one week, or one month of usage. Based on the preliminary prediction, the second prediction subnetwork begins to receive real-time data. The preliminary prediction results are revised based on the real-time data. The revised prediction results are closer to the actual usage. The revised prediction results are output as the final prediction data.

[0076] Furthermore, the method further includes completing the usage prediction based on the first prediction subnetwork and the second prediction subnetwork in the time series network and establishing prediction data:

[0077] Obtain the current time node and use the current time node as the first input data; obtain the current crowd flow data and use the current crowd flow data as the second input data; input the first input data and the second input data into the time series network, receive the first input data through the first prediction sub-network, perform prediction analysis, and generate an initial prediction result; synchronously input the initial prediction result and the second input data into the second prediction sub-network, perform revised prediction, and establish prediction data.

[0078] In the embodiment of the present application, the current time node is obtained using the system clock or timestamp, and the current time node is used as the first input data for predictive analysis of the time series network.

[0079] Get the current online access data from the log system or API of the power grid business hall terminal. This data includes the current number of visits, the number of visiting users, etc. The current online access data is used as the second input data and provided to the time series network for analysis.

[0080] The first input data and the second input data are preprocessed to adapt to the input format of the time series network. The preprocessed data is input into the first prediction sub-network to generate an initial prediction result. Next, the initial prediction result and the second input data are combined into a modified input vector, which is input into the second prediction sub-network for modified prediction analysis to generate the final prediction result.

[0081] Furthermore, step S600 in the method provided in the application embodiment also includes:

[0082] Perform real-time link sorting on the communication link set to generate a sequential sorting result; based on the sequential sorting result, call the corresponding fine-tuning coefficient to perform optimization adjustment based on the adjustable window, and establish adjustment feedback; update the dynamic allocation result through the adjustment feedback to generate an iterative window constraint; perform subsequent fine-tuning processing constraint management based on the iterative window constraint.

[0083] In the embodiment of the present application, monitoring tools such as SNMP protocol, NetFlow, sFlow, etc. are deployed to collect real-time performance data from network devices. These data reflect the bandwidth usage, delay, packet loss rate, etc. of the link in real time.

[0084] Perform weighted calculation on the performance data of each link. For example, the weight of bandwidth utilization is 0.4, the weight of delay is 0.3, and the weight of packet loss rate is 0.3. Then the priority score = 0.4 × bandwidth utilization + 0.3 × delay + 0.3 × packet loss rate. Sort the links according to the priority scores to generate a sequential sorting result. Next, call the corresponding fine-tuning coefficient according to the link sorting result to perform optimization adjustment based on the adjustable window. For example, the fine-tuning coefficient of the high priority link is 1.0, the medium priority link is 0.5, and the low priority link is 0.2. Determine the initial resource allocation for each link based on the sorting result and the fine-tuning coefficient. For example, assuming the initial bandwidth allocation is 100Mbps, for the high priority link, the bandwidth allocation is 100Mbps × 1.0 = 100Mbps; for the medium priority link, it is 100Mbps × 0.5 = 50Mbps.

[0085] Next, determine the adjustable window range for resources such as bandwidth and delay. For example, the bandwidth adjustment range is [50Mbps, 200Mbps], and the delay adjustment range is [10ms, 50ms]. Adjust link resource allocation based on the fine-tuning coefficient and real-time monitoring data. For example, if the current bandwidth utilization of a high-priority link is 90%, the fine-tuning coefficient is 1.0, and the bandwidth adjustment range is [50Mbps, 200Mbps], the bandwidth allocation can be increased to 200Mbps, which is the maximum value. Use network configuration tools, such as router and switch configuration, to make actual resource adjustments.

[0086] Use monitoring tools such as Prometheus and Grafana to collect performance data of the adjusted link, including bandwidth utilization, latency, packet loss rate, etc. Compare the data before and after the adjustment, evaluate the adjustment effect, and generate a feedback report. For example, check whether the bandwidth utilization is reduced and whether the latency is improved.

[0087] Conduct in-depth analysis of adjustment feedback to identify deficiencies and areas for improvement in resource allocation. Update the resource allocation strategy for the link based on the analysis results. For example, if the bandwidth utilization of a link is still higher than 90% after multiple adjustments, further increase its bandwidth limit.

[0088] Based on the adjustment feedback and dynamic allocation results, set the iteration window parameters for the next round of adjustments. For example, the adjustment frequency is once an hour, and the bandwidth adjustment range is [60Mbps, 220Mbps]. Develop specific iteration window constraints. For example, if the bandwidth utilization rate exceeds 85%, increase the bandwidth allocation by 10%; if it is less than 20%, reduce the bandwidth allocation by 10%.

[0089] When fine-tuning constraint management, use network monitoring tools to continuously monitor the real-time performance data of each link. Automatically adjust resources based on real-time monitoring data and iteration window constraints. For example, use automation tools such as Ansible and Puppet to automatically increase bandwidth allocation by 10% when bandwidth utilization exceeds 85%.

[0090] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0091] The present application establishes digital communication with the power grid business hall terminal, configures a communication link set based on digital communication, establishes a link history database of each link in the communication link set, and the link history database records the performance data and stability data of the corresponding link; performs data preprocessing on the link history database, and uses the data preprocessing results as training data to build a link feature analysis model; obtains real-time usage data of the power grid business hall terminal, and performs usage prediction based on the time series network to establish prediction data, performs usage analysis of the communication link set based on the real-time usage data and the prediction data, and establishes link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback; obtains resource usage data of the real-time processor, uses link feedback as a matching feature, performs dynamic reallocation of resources based on the resource usage data, and establishes an adjustable window for the dynamic allocation results of each link; performs feature adaptation analysis under link feedback based on the link feature analysis model, and establishes a fine-tuning coefficient based on the adaptation analysis results; performs fine-tuning processing of the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and completes data management of the power grid business hall terminal based on the fine-tuning processing results. The present invention solves the technical problem that the prior art cannot meet the requirements of efficient data processing and real-time response, and achieves the technical effect of improving data processing speed through cloud-based centralized management, real-time monitoring and intelligent analysis.

[0092] Embodiment 2

[0093] Based on the same inventive concept as the power grid business hall terminal data management method based on cloud computing in the aforementioned embodiment, Figure 2 As shown, the present application provides a cloud computing-based power grid business hall terminal data management system, and the system and method embodiments in the present application embodiments are based on the same inventive concept. The system includes:

[0094] A digital communication establishment module 11, wherein the digital communication establishment module 11 establishes digital communication with a power grid business hall terminal, configures a communication link set based on the digital communication, and establishes a link history database for each link in the communication link set, wherein the link history database records performance data and stability data of the corresponding link;

[0095] A link feature analysis model building module 12, wherein the link feature analysis model building module 12 performs data preprocessing on the link history database and uses the data preprocessing result as training data to build a link feature analysis model;

[0096] A link feedback establishment module 13, wherein the link feedback establishment module 13 obtains real-time usage data of the power grid business hall terminal, and performs usage prediction according to the time series network, establishes prediction data, performs usage analysis of the communication link set according to the real-time usage data and the prediction data, and establishes link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback;

[0097] A resource dynamic allocation module 14, the resource dynamic allocation module 14 obtains resource usage data of the real-time processor, uses the link feedback as a matching feature, dynamically reallocates resources based on the resource usage data, and establishes an adjustable window of dynamic allocation results for each link;

[0098] A feature adaptation analysis module 15, wherein the feature adaptation analysis module 15 performs feature adaptation analysis under link feedback based on the link feature analysis model, and establishes a fine-tuning coefficient according to the adaptation analysis result;

[0099] The data management module 16 performs fine-tuning processing on the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and completes the data management of the power grid business hall terminal based on the fine-tuning processing result.

[0100] Furthermore, the system is also used to implement the following functions:

[0101] Acquire basic link data in the communication link set, wherein the basic link data is topological structure data of the link;

[0102] Performing a network topology dependency analysis of the link based on the link basic data to construct a first link key value;

[0103] Parse the link feedback, obtain link flow data and link service level agreement data, and establish a second link key value according to the link flow data and the link service level agreement data;

[0104] After normalizing the first link key value and the second link key value, dynamically reallocate resources according to the normalization result and the resource usage data.

[0105] Furthermore, the system is also used to implement the following functions:

[0106] Performing a connectivity analysis based on a topological structure on the basic link data to establish a structural key value;

[0107] Performing communication association monitoring based on the link basic data, and establishing association key values ​​based on the communication association monitoring results;

[0108] A weighted calculation is performed on the structure key value and the association key value to construct a first link key value.

[0109] Furthermore, the system is also used to implement the following functions:

[0110] Acquire link traffic data and link service level agreement data based on the link feedback;

[0111] Based on the feature analysis model, the link feature is called, and the adaptation analysis of the link task execution is performed through the link feature and the link traffic data and link service level agreement data to generate the adaptation analysis result.

[0112] Furthermore, the system is also used to implement the following functions:

[0113] Establishing a usage database of power grid business hall terminals, and dividing the usage database into time series periods to establish a time series data set;

[0114] Constructing a first prediction subnetwork based on the time series data set, wherein the first prediction subnetwork is a processing subnetwork of the time series network;

[0115] Establishing a second prediction subnetwork, wherein the second prediction subnetwork is a processing subnetwork of the time series network, and the second prediction subnetwork is communicatively connected with the first prediction subnetwork, and is used to receive real-time data to perform prediction correction;

[0116] Based on the first prediction subnetwork and the second prediction subnetwork in the time series network, usage prediction is completed to establish prediction data.

[0117] Furthermore, the system is also used to implement the following functions:

[0118] Obtaining a current time node, and using the current time node as first input data;

[0119] Acquire current crowd flow data, and use the current crowd flow data as second input data;

[0120] Inputting the first input data and the second input data into the time series network, receiving the first input data through the first prediction subnetwork, performing prediction analysis, and generating an initial prediction result;

[0121] The initial prediction result and the second input data are synchronously input into the second prediction subnetwork, and a revised prediction is performed to establish prediction data.

[0122] Furthermore, the system is also used to implement the following functions:

[0123] Perform real-time link sorting on the communication link set and generate a sequential sorting result;

[0124] Based on the sequence sorting result, the corresponding fine-tuning coefficient is called to perform optimization adjustment based on the adjustable window, and adjustment feedback is established;

[0125] Update the dynamic allocation result through the adjustment feedback to generate an iteration window constraint;

[0126] Subsequent fine-tuning processing constraint management is performed based on the iteration window constraint.

[0127] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0128] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0129] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for managing terminal data in a power grid business hall based on cloud computing, characterized in that: The method comprises: Establish digital communication with the power grid business hall terminal, configure a communication link set based on the digital communication, and establish a link history database for each link in the communication link set, wherein the link history database records performance data and stability data of the corresponding link; Performing data preprocessing on the link history database, and using the data preprocessing results as training data to build a link feature analysis model; Acquire real-time usage data of the power grid business hall terminal, and make usage predictions based on the time series network to establish prediction data, perform usage analysis of the communication link set based on the real-time usage data and the prediction data, and establish link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback; Acquire resource usage data of the real-time processor, use the link feedback as a matching feature, dynamically reallocate resources based on the resource usage data, and establish an adjustable window of dynamic allocation results for each link; Performing feature adaptation analysis under link feedback based on the link feature analysis model, and establishing a fine-tuning coefficient according to the adaptation analysis result; The dynamic allocation results of each link are fine-tuned by using the fine-tuning coefficient and the adjustable window, and the data management of the power grid business hall terminal is completed based on the fine-tuning processing results.

2. The method according to claim 1, characterized in that The method of using the link feedback as a matching feature, dynamically reallocating resources based on the resource usage data, and establishing an adjustable window of dynamic allocation results for each link also includes: Acquire basic link data in the communication link set, wherein the basic link data is topological structure data of the link; Performing a network topology dependency analysis of the link based on the link basic data to construct a first link key value; Parse the link feedback, obtain link flow data and link service level agreement data, and establish a second link key value according to the link flow data and the link service level agreement data; After normalizing the first link key value and the second link key value, dynamically reallocate resources according to the normalization result and the resource usage data.

3. The method according to claim 2, characterized in that The performing of network topology dependency analysis of the link based on the link basic data to construct the first link key value further includes: Performing a connectivity analysis based on a topological structure on the basic link data to establish a structural key value; Performing communication association monitoring based on the link basic data, and establishing association key values ​​based on the communication association monitoring results; A weighted calculation is performed on the structure key value and the association key value to construct a first link key value.

4. The method according to claim 1, characterized in that The performing feature adaptation analysis under link feedback based on the link feature analysis model and establishing a fine-tuning coefficient according to the adaptation analysis result also includes: Acquire link traffic data and link service level agreement data based on the link feedback; Based on the feature analysis model, the link features are called, and the adaptation analysis of the link task execution is performed through the link features and the link traffic data and link service level agreement data to generate the adaptation analysis results.

5. The method according to claim 1, characterized in that The method of performing usage prediction based on the time series network and establishing prediction data also includes: Establishing a usage database of power grid business hall terminals, and dividing the usage database into time series periods to establish a time series data set; Constructing a first prediction subnetwork based on the time series data set, wherein the first prediction subnetwork is a processing subnetwork of the time series network; Establishing a second prediction subnetwork, wherein the second prediction subnetwork is a processing subnetwork of the time series network, and the second prediction subnetwork is communicatively connected with the first prediction subnetwork, and is used to receive real-time data to perform prediction correction; Based on the first prediction subnetwork and the second prediction subnetwork in the time series network, usage prediction is completed to establish prediction data.

6. The method according to claim 5, characterized in that The method of completing the usage prediction based on the first prediction subnetwork and the second prediction subnetwork in the time series network and establishing the prediction data further includes: Obtaining a current time node, and using the current time node as first input data; Acquire current crowd flow data, and use the current crowd flow data as second input data; Inputting the first input data and the second input data into the time series network, receiving the first input data through the first prediction subnetwork, performing prediction analysis, and generating an initial prediction result; The initial prediction result and the second input data are synchronously input into the second prediction subnetwork, and a revised prediction is performed to establish prediction data.

7. The method according to claim 1, characterized in that The fine-tuning processing of the dynamic allocation result of each link by the fine-tuning coefficient and the adjustable window also includes: Perform real-time link sorting on the communication link set and generate a sequential sorting result; Based on the sequence sorting result, the corresponding fine-tuning coefficient is called to perform optimization adjustment based on the adjustable window, and adjustment feedback is established; Update the dynamic allocation result through the adjustment feedback to generate an iteration window constraint; Subsequent fine-tuning processing constraint management is performed based on the iteration window constraint.

8. The power grid business hall terminal data management system based on cloud computing is characterized by: The system comprises: A digital communication establishment module, wherein the digital communication establishment module establishes digital communication with a power grid business hall terminal, configures a communication link set based on the digital communication, and establishes a link history database for each link in the communication link set, wherein the link history database records performance data and stability data of the corresponding link; A link feature analysis model construction module, wherein the link feature analysis model construction module performs data preprocessing on the link history database and uses the data preprocessing result as training data to construct a link feature analysis model; A link feedback establishment module, wherein the link feedback establishment module obtains real-time usage data of the power grid business hall terminal, and performs usage prediction according to the time series network, establishes prediction data, performs usage analysis of the communication link set according to the real-time usage data and the prediction data, and establishes link feedback, wherein the link feedback includes real-time link feedback and prediction link feedback; A resource dynamic allocation module, which obtains resource usage data of the real-time processor, uses the link feedback as a matching feature, dynamically reallocates resources based on the resource usage data, and establishes an adjustable window of dynamic allocation results for each link; A feature adaptation analysis module, wherein the feature adaptation analysis module performs feature adaptation analysis under link feedback based on the link feature analysis model, and establishes a fine-tuning coefficient according to the adaptation analysis result; A data management module, wherein the data management module performs fine-tuning processing on the dynamic allocation results of each link through the fine-tuning coefficient and the adjustable window, and completes data management of the power grid business hall terminal based on the fine-tuning processing results.

Images