Data intelligent distribution method based on edge computing industrial computer

Abstract

The present invention belongs to the field of data management technology, and specifically is a data intelligent distribution method based on edge computing industrial computers, including deploying edge computing industrial computers, collecting data in industrial control systems and extracting key information, designing intelligent distribution algorithms and performing data distribution, data backup and intelligent distribution algorithm optimization; the present invention intelligently distributes data to corresponding receiving ends based on intelligent distribution algorithms to achieve fast and accurate data distribution, and provides strong technical support for industrial automation, intelligent manufacturing and other fields, and through monitoring and analyzing corresponding performance indicators in the data distribution process, reasonably analyzes and judges the data distribution performance and continuously optimizes the distribution algorithm, further ensuring the efficiency and accuracy of data distribution, and through evaluating the optimized performance status of the distribution algorithm and timely reminding management personnel to make matching management measures, the present invention has a high degree of intelligence.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a data intelligent distribution method based on an edge computing industrial computer. Background Art

[0002] With the rapid development of industrial informatization, a large amount of data is generated and transmitted in industrial control systems. Traditional data processing methods usually send data to the cloud for processing and distribution. This method has problems such as high data transmission delay and high network bandwidth pressure, which affects the real-time performance and efficiency of industrial control systems. It is difficult to achieve fast and accurate data distribution, and it is impossible to reasonably analyze and judge the data distribution performance and continuously optimize the distribution algorithm. It is also impossible to accurately evaluate the optimization performance of the distribution algorithm, and the degree of intelligence is low;

[0003] In view of the above technical defects, a solution is now proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a data intelligent distribution method based on edge computing industrial computers, which solves the problems that the prior art is difficult to achieve fast and accurate data distribution, and is unable to reasonably analyze and judge the data distribution performance and continuously optimize the distribution algorithm, as well as unable to accurately evaluate the optimization performance status of the distribution algorithm and has a low degree of intelligence.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The data intelligent distribution method based on edge computing industrial computer includes the following steps:

[0007] Step 1: Deploy edge computing industrial computers at each node of the industrial control system;

[0008] Step 2: Use edge computing industrial computers to collect data from industrial control systems in real time and pre-process the data to extract key information;

[0009] Step 3: Design an intelligent distribution algorithm based on the type, importance and target recipient of the data to intelligently distribute the data to the corresponding receiving end;

[0010] Step 4: The edge computing industrial computer works with the cloud server to back up, synchronize and update data;

[0011] Step 5: Continuously optimize the intelligent distribution algorithm by monitoring and analyzing the corresponding performance indicators during the data distribution process.

[0012] Furthermore, in step 2, the specific operation process of preprocessing the data is as follows:

[0013] Data cleaning: During the data cleaning process, the edge computing industrial computer checks the collected data and removes abnormal values ​​or missing values ​​caused by equipment failure and transmission errors;

[0014] Format conversion: Data format conversion is used to convert the collected raw data into a format suitable for subsequent analysis and processing, including converting data from raw format to structured format;

[0015] Standardization: Standardization is to scale or transform data according to certain rules so that data from different sources and dimensions can be compared and analyzed on a unified scale.

[0016] Extraction of key information: The edge computing industrial computer uses corresponding data analysis algorithms and models to conduct in-depth analysis and mining of pre-processed data; the extraction of key information involves equipment failure warning, production efficiency optimization and product quality control.

[0017] Furthermore, the specific process of step three is as follows:

[0018] Gain in-depth understanding of the types and characteristics of data. Data types include text, images, videos, and audio. Each type of data has its own unique processing methods and transmission requirements. Among them, text data requires text analysis and keyword extraction, and video data needs to be compressed and transcoded to adapt to different receiving devices.

[0019] Assess the importance of data, which is measured by factors such as the urgency of the data and the impact on the business. High-importance data should be processed first and ensure that it reaches the target recipient quickly and accurately. For low-importance data, a distribution strategy of delayed transmission or batch processing should be adopted.

[0020] Clarify the characteristics and needs of target recipients, which include different user groups, devices or systems. Different target recipients have different receiving capabilities and processing requirements.

[0021] Design an intelligent distribution algorithm. The algorithm design comprehensively considers the type, importance and needs of the target recipients of the data to formulate an appropriate distribution strategy. For data that is of high importance and requires real-time processing, it is processed using a priority queue to ensure that it can be transmitted to the target recipient first. For data that is of low importance and allows batch processing, it is processed using caching or batch transmission to improve transmission efficiency.

[0022] Furthermore, when designing the intelligent distribution algorithm, the intelligent distribution algorithm must also be adaptive and scalable. The intelligent distribution algorithm needs to be adaptively adjusted according to the actual situation. As the business scale expands and the number of receivers increases, the intelligent distribution algorithm needs to support horizontal expansion to meet higher distribution requirements.

[0023] In addition, the designed intelligent distribution algorithm needs to be fully tested and verified by simulating the data transmission and distribution process in actual scenarios, evaluating the accuracy and efficiency of the intelligent distribution algorithm, and iterating and optimizing the intelligent distribution algorithm based on the test results to improve its performance and stability.

[0024] Furthermore, in step five, by monitoring and analyzing the corresponding performance indicators in the data distribution process, the abnormal detection period is determined based on the corresponding performance indicator data and through performance evaluation analysis; the detection period is set, the number of abnormal detection periods in the detection period is obtained and the ratio is calculated with the total number of detection periods to obtain the distribution abnormality value, and the distribution detection values ​​of all detection periods in the detection period are averaged to obtain the distribution detection condition value, and the distribution detection value with the largest value in the detection period is marked as the distribution detection amplitude value, and the distribution performance value is obtained by numerically calculating the distribution abnormality value, the distribution detection condition value and the distribution detection amplitude value, and the distribution performance value is numerically compared with the preset distribution performance threshold value. If the distribution performance value exceeds the preset distribution performance threshold value, an algorithm optimization signal is generated.

[0025] Furthermore, the specific analysis process of the condition assessment analysis is as follows:

[0026] The distribution speed, success rate, error rate and delay time of the corresponding detection time period are collected, and the distribution detection value is obtained by numerically calculating the distribution speed, success rate, error rate and delay time. The distribution detection value is numerically compared with the preset distribution detection threshold. If the distribution detection value exceeds the preset distribution detection threshold, the corresponding detection time period is marked as an abnormal detection period.

[0027] Furthermore, if the distribution performance value does not exceed the preset distribution performance threshold, the degree of data loss and damage during the transmission process is judged by comparing the original data with the data received after distribution to obtain the accuracy anomaly test value, and the user satisfaction feedback is monitored and analyzed to obtain the satisfaction test value within the test period. The accuracy anomaly test value and the satisfaction test value are numerically calculated to obtain the distribution evaluation value, and the distribution evaluation value is numerically compared with the preset distribution evaluation threshold. If the distribution evaluation value exceeds the preset distribution evaluation threshold, an algorithm optimization signal is generated.

[0028] Furthermore, the specific analysis process of user satisfaction feedback monitoring and analysis is as follows:

[0029] All user feedback information generated within the detection period is obtained, and semantic recognition is performed on each piece of user feedback information one by one. User feedback information containing negative meanings is marked as negative feedback information, the number of negative feedback information within the detection period is marked as a negative feedback detection value, and the ratio of the negative feedback detection value to the total number of user feedback information generated within the detection period is marked as a negative feedback detection ratio value. The satisfaction detection value is obtained by numerically calculating the negative feedback detection value and the negative feedback detection ratio value.

[0030] Further, after the algorithm optimization signal is generated, the generation time of the algorithm optimization signal is collected and marked as the first time, and the time when the algorithm optimization starts is collected and marked as the second time, and the interval between the second time and the first time is marked as the algorithm optimization buffer time;

[0031] An optimization monitoring cycle is set, and a priority buffer set is established for all algorithm optimization buffer times within the optimization monitoring cycle. The priority buffer set is averaged to obtain a priority buffer detection value, and all subsets in the priority buffer set are sorted in descending order according to the values ​​of the algorithm optimization buffer time, and a preset algorithm optimization buffer time threshold is inserted into the priority buffer set, and the number of subsets in the priority buffer set that are located before the preset algorithm optimization buffer time threshold is marked as a priority buffer inefficiency value;

[0032] And through analysis, the algorithm optimization performance value is obtained, the priority detection value, the priority inefficiency value and the algorithm optimization performance value are numerically calculated to obtain the algorithm optimization evaluation value, and the algorithm optimization evaluation value is numerically compared with the preset algorithm optimization evaluation threshold. If the algorithm optimization evaluation value exceeds the preset algorithm optimization evaluation threshold, an algorithm optimization evaluation failure signal is generated; if the algorithm optimization evaluation value does not exceed the preset algorithm optimization evaluation threshold, an algorithm optimization evaluation pass signal is generated.

[0033] Furthermore, the analysis and acquisition method of the algorithm optimization performance value is as follows:

[0034] The interval duration between the first moment and the previous adjacent algorithm optimization completion moment is marked as the algorithm optimal interval value, and the algorithm optimal interval value is numerically compared with the preset algorithm optimal interval threshold. If the algorithm optimal interval value does not exceed the preset algorithm optimal interval threshold, the corresponding algorithm optimal interval value is marked as the optimized low interval value, and the number of optimized low interval values ​​within the optimization monitoring period is obtained and marked as the algorithm optimization performance value.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. In the present invention, by deploying edge computing industrial computers at various nodes of the industrial control system, the edge computing industrial computers collect data in the industrial control system in real time and extract key information, design intelligent distribution algorithms according to the type, importance and target recipients of the data, and intelligently distribute the data to the corresponding receiving end based on the intelligent distribution algorithm, so as to achieve fast and accurate data distribution, and provide strong technical support for the fields of industrial automation, intelligent manufacturing, etc.;

[0037] 2. In the present invention, by monitoring and analyzing the corresponding performance indicators in the data distribution process, the data distribution performance is reasonably analyzed and judged, and the distribution algorithm is continuously optimized to further ensure the efficiency and accuracy of data distribution. By evaluating the optimized performance status of the distribution algorithm, the management personnel are promptly reminded to conduct cause investigation and analysis and take matching management measures, which has a high degree of intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0039] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Embodiment 1: Figure 1 As shown, the data intelligent distribution method based on edge computing industrial computer proposed in the present invention includes the following steps:

[0042] Step 1: Deploy edge computing industrial computers at various nodes of the industrial control system. These nodes can be sensors, actuators and other equipment on the production line. It should be noted that in industrial control systems, a large number of sensors and equipment are deployed on production lines, machines and other key links. They can capture real-time data on equipment status, production efficiency, product quality and other aspects.

[0043] Step 2: Collect data from industrial control systems in real time through edge computing industrial computers. As the core device for data collection, edge computing industrial computers communicate with the sensors and devices connected to them to obtain these data in real time. The collection frequency and accuracy of these data can be set according to actual needs to ensure the integrity and accuracy of the data; and pre-process the data to extract key information. Since the data in industrial control systems are often complex and diverse, a series of pre-processing operations are required before analyzing and extracting key information; this step not only improves the accuracy and reliability of the data, but also provides strong support for the company's production management and decision-making; the specific operation process of pre-processing the data is as follows:

[0044] Data cleaning: During the data cleaning process, the edge computing industrial computer checks the collected data and removes abnormal values ​​or missing values ​​caused by equipment failure, transmission errors, etc. At the same time, factors that may affect the accuracy of the data, such as outliers, will also be processed accordingly;

[0045] Format conversion: Data format conversion is used to convert the collected raw data into a format suitable for subsequent analysis and processing, such as converting data from a raw format to a structured format for more efficient storage and query;

[0046] Standardization: Standardization is to scale or transform data according to certain rules, so that data from different sources and dimensions can be compared and analyzed on a unified scale. After preprocessing, the quality of the data is significantly improved, providing a solid foundation for subsequent key information extraction.

[0047] Extraction of key information: The edge computing industrial computer uses the corresponding data analysis algorithms and models to conduct in-depth analysis and mining of pre-processed data. These algorithms and models can be selected and adjusted according to actual needs to achieve effective extraction of specific key information; the extraction of key information involves multiple aspects, such as equipment failure warning, production efficiency optimization, product quality control, etc. Through real-time monitoring and analysis of these key information, enterprises can promptly discover problems in the production process and take corresponding measures to adjust and optimize, thereby improving production efficiency and product quality and reducing operating costs.

[0048] Step 3: Design an intelligent distribution algorithm based on the data type, importance, and target recipients to intelligently distribute the data to the corresponding recipients. By deeply understanding the characteristics of the data and the needs of the recipients, formulate a suitable distribution strategy, and continuously optimize the performance and stability of the algorithm, you can ensure that the data can be intelligently and efficiently distributed to the corresponding recipients. The specific process is as follows:

[0049] Have a deep understanding of the types and characteristics of data. Data types include text, images, videos, and audio. Each type of data has its own unique processing methods and transmission requirements. For example, text data requires text analysis and keyword extraction, and video data requires compression and transcoding to adapt to different receiving devices. When designing intelligent distribution algorithms, fully consider the type of data and formulate corresponding processing strategies based on its characteristics.

[0050] Assess the importance of data, which is measured based on factors such as the urgency of the data and the impact on the business. For example, high-importance data, such as emergency notifications and critical business data, needs to be processed first and ensure that it reaches the target recipient quickly and accurately. For low-importance data, more flexible distribution strategies can be adopted, such as delayed transmission or batch processing.

[0051] Clarify the characteristics and needs of the target recipients. The target recipients include different user groups, devices or systems. Different target recipients have different receiving capabilities and processing requirements. For example, some users may need to receive data in real time and process it immediately, while others may only need to receive summary reports of the data on a regular basis. Therefore, when designing intelligent distribution algorithms, it is necessary to fully consider the characteristics and needs of the target recipients to ensure that the data can be distributed to them in the most appropriate way.

[0052] Design an intelligent distribution algorithm. The algorithm design comprehensively considers the type, importance and needs of the target recipients of the data to formulate an appropriate distribution strategy. For example, for data that is of high importance and requires real-time processing, a priority queue is used to ensure that it can be transmitted to the target recipient first. For data that is of low importance and allows batch processing, a cache or batch transmission method is used to improve transmission efficiency.

[0053] Furthermore, when designing an intelligent distribution algorithm, the intelligent distribution algorithm should also be adaptive and scalable. Since the type and importance of data may change over time, and the needs of the target recipients may also be adjusted at any time, the algorithm needs to be able to adapt according to the actual situation. At the same time, as the scale of business expands and the number of recipients increases, the algorithm should also be able to support horizontal expansion to meet higher distribution needs; and the designed intelligent distribution algorithm needs to be fully tested and verified, by simulating the data transmission and distribution process in actual scenarios, evaluating the accuracy and efficiency of the intelligent distribution algorithm, and iterating and optimizing the intelligent distribution algorithm based on the test results to improve its performance and stability.

[0054] Step 4: The edge computing industrial computer works with the cloud server to back up, synchronize and update data to ensure data integrity and consistency;

[0055] Step 5: Continuously optimize the intelligent distribution algorithm by monitoring and analyzing the corresponding performance indicators during the data distribution process. The specific process of analyzing the corresponding performance indicators is as follows:

[0056] Based on the corresponding performance indicator data and through performance evaluation and analysis, determine the abnormal detection period; specifically: collect the distribution speed, success rate, error rate and delay time of the corresponding detection period, among which, the slower the distribution speed, the lower the success rate, the higher the error rate and the longer the delay time, the worse the data distribution status of the corresponding detection period;

[0057] The distribution speed FS, success rate FG, error rate FY and delay time FL are numerically calculated by the formula FX=(a3*FY+a4*FL) / (a1*FS+a2*FG)to obtain the distribution detection value FX, wherein a1, a2, a3, a4 are preset proportional coefficients, and the values ​​of a1, a2, a3, a4 are all positive numbers; and the larger the value of the distribution detection value FX, the worse the data distribution status of the corresponding detection period; the distribution detection value FX is numerically compared with the preset distribution detection threshold value, if the distribution detection value FX exceeds the preset distribution detection threshold value, it indicates that the data distribution status of the corresponding detection period is poor, and the corresponding detection period is marked as an abnormal detection period;

[0058] Set a detection cycle, preferably, the detection cycle is 36 hours; obtain the number of abnormal detection time periods within the detection cycle and calculate the ratio of it to the total number of detection time periods to obtain the distribution abnormality value, and calculate the average of the distribution detection values ​​of all detection time periods within the detection cycle to obtain the distribution detection value, and mark the distribution detection value with the largest value within the detection cycle as the distribution detection amplitude value;

[0059] The distribution performance value GY is obtained by numerically calculating the distribution abnormality value GK, the distribution detection value GP and the distribution detection amplitude value GM through the formula GY=kp1*GK+(kp2*GP+kp3*GM) / 2, wherein kp1, kp2 and kp3 are preset proportional coefficients, kp1>kp2>kp3>0; and the larger the value of the distribution performance value GY is, the worse the data distribution performance condition in the detection period is in general; the distribution performance value GY is numerically compared with the preset distribution performance threshold value, and if the distribution performance value GY exceeds the preset distribution performance threshold value, it indicates that the data distribution performance condition in the detection period is in general poor, and an algorithm optimization signal is generated;

[0060] If the distribution performance value GY does not exceed the preset distribution performance threshold, the degree of data loss and damage during the transmission process is determined by comparing the original data with the data received after distribution to obtain the accuracy error detection value, and the user satisfaction feedback is monitored and analyzed to obtain the satisfaction detection value within the detection period. The larger the accuracy error detection value and the larger the satisfaction detection value, the worse the distribution status within the detection period.

[0061] The accuracy anomaly detection value WS and the satisfaction detection value WS are numerically calculated using the formula WX=ty1*WS+ty2*WS to obtain the distribution evaluation value WX, where ty1 and ty2 are preset proportional coefficients, and the values ​​of ty1 and ty2 are both positive numbers; and the larger the distribution evaluation value WX is, the worse the data distribution condition within the detection period is; the distribution evaluation value WX is numerically compared with the preset distribution evaluation threshold value. If the distribution evaluation value WX exceeds the preset distribution evaluation threshold value, it indicates that the data distribution condition within the detection period is poor and the intelligent distribution algorithm needs to be optimized in time, then an algorithm optimization signal is generated.

[0062] Furthermore, the specific analysis process of user satisfaction feedback monitoring and analysis is as follows: all user feedback information generated within the detection period is obtained, each piece of user feedback information is semantically identified one by one, user feedback information containing negative meanings is marked as negative feedback information, the number of negative feedback information within the detection period is marked as a negative feedback detection value, and the ratio of the negative feedback detection value to the total number of user feedback information generated within the detection period is marked as a negative feedback detection value, and the negative feedback detection value WF and the negative feedback detection value WK are numerically calculated by the formula WS=b1*WF / b2+b2*WK to obtain the satisfaction detection value WS; wherein b1 and b2 are preset proportional coefficients, and b2>b1>0.

[0063] Embodiment 2: The difference between this embodiment and embodiment 1 is that after the algorithm optimization signal is generated, the generation time of the algorithm optimization signal is collected and marked as the first time, and the time when the algorithm optimization starts is collected and marked as the second time, and the interval between the second time and the first time is marked as the algorithm optimization buffer time; wherein, the larger the value of the algorithm optimization buffer time, the less timely the corresponding optimization is and the slower the response efficiency is;

[0064] An optimization monitoring period is set, preferably, the optimization monitoring period is thirty days; a priority buffer set is established for all algorithm optimization buffer times within the optimization monitoring period, the priority buffer set is averaged to obtain a priority buffer detection value, and all subsets in the priority buffer set are sorted in descending order according to the values ​​of the algorithm optimization buffer time, and a preset algorithm optimization buffer time threshold is inserted into the priority buffer set, and the number of subsets in the priority buffer set that are located before the preset algorithm optimization buffer time threshold is marked as a priority buffer inefficiency value;

[0065] And the algorithm optimization performance value is obtained through analysis, specifically: the interval duration between the first moment and the adjacent last set of algorithm optimization completion moments is marked as the algorithm optimization interval value, the algorithm optimization interval value is numerically compared with the preset algorithm optimization interval threshold, if the algorithm optimization interval value does not exceed the preset algorithm optimization interval threshold, then the corresponding algorithm optimization interval value is marked as the optimized low interval value, the number of optimized low interval values ​​in the optimization monitoring period is obtained and marked as the algorithm optimization performance value;

[0066] By formula The optimization detection value YK, the optimization inefficiency value YG and the algorithm optimization performance value YM are numerically calculated to obtain the algorithm optimization evaluation value YP, wherein wq1, wq2 and wq3 are preset proportional coefficients, and the values ​​of wq1, wq2 and wq3 are all positive numbers; and the larger the value of the algorithm optimization evaluation value YP is, the worse the optimization performance of the intelligent distribution algorithm is during the optimization monitoring period;

[0067] The algorithm optimization evaluation value YP is numerically compared with the preset algorithm optimization evaluation threshold. If the algorithm optimization evaluation value YP exceeds the preset algorithm optimization evaluation threshold, indicating that the optimization performance of the intelligent distribution algorithm during the optimization monitoring period is poor, then an algorithm optimization evaluation failure signal is generated; if the algorithm optimization evaluation value YP does not exceed the preset algorithm optimization evaluation threshold, indicating that the optimization performance of the intelligent distribution algorithm during the optimization monitoring period is good, then an algorithm optimization evaluation qualification signal is generated. When the algorithm optimization evaluation failure signal is generated, the management personnel are reminded to conduct timely cause investigation and analysis and make matching management measures, with a high degree of intelligence.

[0068] The working principle of the present invention is as follows: when in use, edge computing industrial computers are deployed at various nodes of the industrial control system. The edge computing industrial computers collect data in the industrial control system in real time and extract key information. Intelligent distribution algorithms are designed according to the type, importance and target recipients of the data. Based on the intelligent distribution algorithm, the data is intelligently distributed to the corresponding receiving end to achieve fast and accurate data distribution, providing strong technical support for industrial automation, intelligent manufacturing and other fields; and by monitoring and analyzing the corresponding performance indicators in the data distribution process, the data distribution performance is reasonably analyzed and judged, and the distribution algorithm is continuously optimized to further ensure the efficiency and accuracy of data distribution, as well as by accurately evaluating the optimization performance status of the distribution algorithm to generate an algorithm optimization evaluation qualified signal or an algorithm optimization evaluation unqualified signal, and when the algorithm optimization evaluation unqualified signal is generated, the management personnel are reminded to conduct timely cause investigation and analysis and make matching management measures, with a high degree of intelligence.

[0069] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that technicians in the relevant technical field can understand and use the present invention well. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A data intelligent distribution method based on edge computing industrial computers, characterized in that: The following steps are involved: Step 1: Deploy edge computing industrial computers at each node of the industrial control system; Step 2: Use edge computing industrial computers to collect data from industrial control systems in real time and pre-process the data to extract key information; Step 3: Design an intelligent distribution algorithm based on the type, importance and target recipient of the data to intelligently distribute the data to the corresponding receiving end; Step 4: The edge computing industrial computer works with the cloud server to back up, synchronize and update data; Step 5: Continuously optimize the intelligent distribution algorithm by monitoring and analyzing the corresponding performance indicators during the data distribution process; In step 5, by monitoring and analyzing the corresponding performance indicators in the data distribution process, the abnormal inspection period is determined based on the corresponding performance indicator data and through performance evaluation analysis; A detection period is set, the number of abnormal detection periods within the detection period is obtained and the ratio thereof is calculated to the total number of detection periods to obtain a distribution abnormality value, and the distribution detection values ​​of all detection periods within the detection period are averaged to obtain a distribution detection condition value, and the distribution detection value with the largest value within the detection period is marked as a distribution detection amplitude value, and the distribution abnormality value, the distribution detection condition value and the distribution detection amplitude value are numerically calculated to obtain a distribution performance value, and if the distribution performance value exceeds a preset distribution performance threshold, an algorithm optimization signal is generated; The specific analysis process of the performance evaluation is as follows: The distribution speed, success rate, error rate and delay time of the corresponding detection period are collected, and the distribution detection value is obtained by numerically calculating the distribution speed, success rate, error rate and delay time. The distribution detection value is numerically compared with the preset distribution detection threshold. If the distribution detection value exceeds the preset distribution detection threshold, the corresponding detection period is marked as an abnormal detection period; If the distribution performance value does not exceed the preset distribution performance threshold, the degree of data loss and damage during the transmission process is determined by comparing the original data with the data received after distribution to obtain the accuracy anomaly detection value, and the satisfaction detection value within the detection period is obtained by monitoring and analyzing the user satisfaction feedback. The accuracy anomaly detection value and the satisfaction detection value are numerically calculated to obtain the distribution evaluation value. If the distribution evaluation value exceeds the preset distribution evaluation threshold, an algorithm optimization signal is generated; The specific analysis process of user satisfaction feedback monitoring and analysis is as follows: All user feedback information generated within the detection period is obtained, and semantic recognition is performed on each piece of user feedback information one by one, and user feedback information containing negative meanings is marked as negative feedback information, and the number of negative feedback information within the detection period is marked as a negative feedback detection value, and the ratio of the negative feedback detection value to the total number of user feedback information generated within the detection period is marked as a negative feedback detection ratio value, and the satisfaction detection value is obtained by numerically calculating the negative feedback detection value and the negative feedback detection ratio value; After the algorithm optimization signal is generated, the generation time of the algorithm optimization signal is collected and marked as the first time, and the time when the algorithm optimization starts is collected and marked as the second time, and the interval between the second time and the first time is marked as the algorithm optimization buffer time; An optimization monitoring cycle is set, and a priority buffer set is established for all algorithm optimization buffer times within the optimization monitoring cycle. The priority buffer set is averaged to obtain a priority buffer detection value, and all subsets in the priority buffer set are sorted in descending order according to the values ​​of the algorithm optimization buffer time, and a preset algorithm optimization buffer time threshold is inserted into the priority buffer set, and the number of subsets in the priority buffer set that are located before the preset algorithm optimization buffer time threshold is marked as a priority buffer inefficiency value; And the algorithm optimization performance value is obtained through analysis, and the priority detection value, priority inefficiency value and algorithm optimization performance value are numerically calculated to obtain the algorithm optimization evaluation value. If the algorithm optimization evaluation value exceeds the preset algorithm optimization evaluation threshold, an algorithm optimization evaluation unqualified signal is generated; if the algorithm optimization evaluation value does not exceed the preset algorithm optimization evaluation threshold, an algorithm optimization evaluation qualified signal is generated; The analysis and acquisition method of the algorithm optimization performance value is as follows: The interval duration between the first moment and the previous adjacent algorithm optimization completion moment is marked as the algorithm optimal interval value, and the algorithm optimal interval value is numerically compared with the preset algorithm optimal interval threshold. If the algorithm optimal interval value does not exceed the preset algorithm optimal interval threshold, the corresponding algorithm optimal interval value is marked as the optimized low interval value, and the number of optimized low interval values ​​within the optimization monitoring period is obtained and marked as the algorithm optimization performance value.

2. The data intelligent distribution method based on edge computing industrial computer according to claim 1 is characterized in that: In step 2, the specific operation process of preprocessing the data is as follows: Data cleaning: During the data cleaning process, the edge computing industrial computer checks the collected data and removes abnormal values ​​or missing values ​​caused by equipment failure and transmission errors; Format conversion: Data format conversion is used to convert the collected raw data into a format suitable for subsequent analysis and processing, including converting data from raw format to structured format; Standardization: Standardization is to scale or transform data according to certain rules so that data from different sources and dimensions can be compared and analyzed on a unified scale. Extraction of key information: The edge computing industrial computer uses corresponding data analysis algorithms and models to conduct in-depth analysis and mining of pre-processed data; the extraction of key information involves equipment failure warning, production efficiency optimization and product quality control.

3. The data intelligent distribution method based on edge computing industrial computer according to claim 2 is characterized in that: The specific process of step three is as follows: Gain in-depth understanding of the types and characteristics of data. Data types include text, images, videos, and audio. Each type of data has its own unique processing methods and transmission requirements. Among them, text data requires text analysis and keyword extraction, and video data needs to be compressed and transcoded to adapt to different receiving devices. Assess the importance of data, which is measured by factors such as the urgency of the data and the impact on the business. High-importance data should be processed first and ensure that it reaches the target recipient quickly and accurately. For low-importance data, a distribution strategy of delayed transmission or batch processing should be adopted. Clarify the characteristics and needs of target recipients, which include different user groups, devices or systems. Different target recipients have different receiving capabilities and processing requirements. Design an intelligent distribution algorithm. The algorithm design comprehensively considers the type, importance and needs of the target recipients of the data to formulate an appropriate distribution strategy. For data that is of high importance and requires real-time processing, a priority queue is used to ensure that it can be transmitted to the target recipient first. For data that is of low importance and allows batch processing, a cache or batch transmission method is used to improve transmission efficiency.

4. The data intelligent distribution method based on edge computing industrial computer according to claim 3 is characterized in that: When designing the intelligent distribution algorithm, the intelligent distribution algorithm must also be adaptive and scalable. The intelligent distribution algorithm needs to be adaptively adjusted according to the actual situation. As the business scale expands and the number of receivers increases, the intelligent distribution algorithm needs to support horizontal expansion to meet higher distribution requirements. In addition, the designed intelligent distribution algorithm needs to be fully tested and verified by simulating the data transmission and distribution process in actual scenarios, evaluating the accuracy and efficiency of the intelligent distribution algorithm, and iterating and optimizing the intelligent distribution algorithm based on the test results to improve its performance and stability.

Images