A traffic data self-identification method and system based on big data behavior analysis

Abstract

The present application relates to a method and system for self-identification of flow data based on big data behavior analysis, the method comprising: analyzing the production equipment that exceeds the flow threshold range according to the comparison result of the data flow of each production equipment on the current production line and the flow threshold range corresponding to each production equipment on the current production line; collecting all relevant data of the production equipment that exceeds the flow threshold range, all relevant data include operation history, maintenance records and detailed characteristics of real-time data flow; extracting preset features from all relevant data collected, and inputting the extracted features and real-time data into the fault type prediction model that has been trained, and outputting the predicted fault type. The present application has the following effects: using big data technology for fault prediction can timely discover and deal with potential problems, and reduce the impact of equipment failures on production.

Description

Technical Field

[0001] The present invention relates to the field of big data analysis, and in particular to a flow data self-identification method and system based on big data behavior analysis. Background Art

[0002] Various faults often occur in production equipment during operation, which may lead to reduced production efficiency, damaged product quality and even affect product delivery time.

[0003] Existing technologies usually monitor production equipment by regular inspections or by presetting fixed thresholds. This method cannot dynamically adjust the monitoring strategy according to actual conditions, which can easily lead to false alarms or missed alarms, affecting production efficiency and equipment maintenance costs. Summary of the invention

[0004] In order to use big data technology to predict faults, discover and handle potential problems in a timely manner, and reduce the impact of equipment failures on production, the present application provides a flow data self-identification method and system based on big data behavior analysis.

[0005] In the first aspect, the present application provides a method for self-identification of traffic data based on big data behavior analysis, which adopts the following technical solutions:

[0006] A traffic data self-identification method based on big data behavior analysis, comprising:

[0007] Obtain the data flow of each production equipment on the current production line;

[0008] According to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of production equipment, analyze and determine the flow threshold corresponding to each production equipment on the current production line;

[0009] According to the comparison result of the data flow of each production equipment on the current production line and the flow threshold range corresponding to each production equipment on the current production line, the production equipment exceeding the flow threshold range is analyzed;

[0010] Collect all relevant data of production equipment that exceeds the flow threshold range, including operation history, maintenance records and detailed characteristics of real-time data flow;

[0011] Extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type;

[0012] Analyze whether the predicted fault type is a fault that requires shutdown for processing;

[0013] If not, the data flow of the production equipment with the predicted fault type within the preset time period is obtained and a data flow change table is formed, and the data flow change table and the fault type are sent to the terminal held by the person in charge;

[0014] If yes, then obtain the preset downtime for the production equipment that needs to be shut down to handle the fault, the impact of the production equipment that needs to be shut down to handle the fault on the overall product production efficiency, and the preset product delivery time, and analyze whether it will cause the product delivery to exceed the preset product delivery time;

[0015] If yes, then according to the preset product delivery time and the impact of the production equipment that needs to be shut down for fault processing on the overall product production efficiency, reverse analysis is performed to obtain the shutdown time for fault processing, and according to the corresponding relationship between the time range within which the shutdown time for fault processing falls and the emergency processing plan, the emergency processing plan is analyzed and determined;

[0016] If not, keep the original setting.

[0017] By adopting the above technical solutions, traffic data self-identification can be achieved through big data behavior analysis, which can accurately locate the equipment that exceeds the threshold, predict the type of fault and whether it needs to be shut down for processing. The person in charge can be notified of non-shutdown faults in a timely manner, and the processing time and emergency plan for shutdown faults can be reasonably planned, effectively avoiding delivery delays, improving production efficiency and the stability of product delivery, helping to reduce production losses and ensure the smooth operation of the overall production process.

[0018] Optionally, according to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of the production equipment, the flow threshold corresponding to each production equipment on the current production line is analyzed and determined, including:

[0019] Analyze whether the service life of production equipment exceeds the preset service life;

[0020] If yes, then analyze and determine the flow threshold corresponding to each production equipment on the current production line based on the mapping relationship between the equipment type, the service life range of the production equipment, the working period, the magnitude of the production task and the flow threshold;

[0021] If not, keep the original setting.

[0022] By adopting the above technical solution, the service life of the equipment can be taken into account. For equipment that exceeds the preset service life, the flow threshold can be re-determined based on multiple factors to make the threshold more suitable for the actual condition of the equipment. For equipment that has not exceeded the service life, the original setting can be maintained to ensure efficiency. This effectively improves the accuracy and adaptability of the threshold, reduces flow misjudgment caused by equipment aging, helps optimize the production process, ensures stable operation of equipment, and improves production efficiency and product quality.

[0023] Optionally, according to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of the production equipment, the flow threshold corresponding to each production equipment on the current production line is analyzed and determined, including:

[0024] Analyze whether the production equipment has been modified;

[0025] If yes, then obtain the extent of production efficiency improvement after equipment transformation, and analyze and determine the flow threshold corresponding to each production equipment on the current production line according to the mapping relationship between the extent of production efficiency improvement after equipment transformation, equipment type of production equipment, working hours, magnitude of production tasks and flow threshold;

[0026] If not, keep the original setting.

[0027] By adopting the above technical solution, it is possible to ensure that the flow threshold matches the actual situation of the equipment, improve the efficiency and stability of the production line, avoid equipment overload or inefficiency due to unreasonable flow threshold, thereby optimizing the production process and improving overall production efficiency.

[0028] Optionally, according to the magnitude of production efficiency improvement after equipment transformation, the equipment type of production equipment, the working period, the magnitude of production tasks and the mapping relationship between the flow threshold, the flow threshold corresponding to each production equipment on the current production line is analyzed and determined, including:

[0029] Analyze whether the service life of production equipment exceeds the preset service life;

[0030] If yes, then obtain the mapping relationship between the range of service life of the production equipment and the extent of reduction in production efficiency, and analyze and determine the extent of reduction in production efficiency;

[0031] Analyze and determine the actual extent of the improvement in production efficiency after equipment transformation based on the extent of the improvement in production efficiency after equipment transformation and the extent of the reduction in production efficiency caused by exceeding the preset service life;

[0032] According to the actual extent of the production efficiency improvement after the equipment transformation, the equipment type of production equipment, working hours, the magnitude of production tasks and the mapping relationship between the flow threshold, analyze and determine the flow threshold corresponding to each production equipment on the current production line;

[0033] If not, keep the original setting.

[0034] By adopting the above technical solutions, the impact of equipment age on production efficiency is fully considered, and the actual efficiency improvement is accurately determined. Combined with multiple factors such as equipment type and working hours, the flow threshold of each production equipment is accurately analyzed. For equipment that has not exceeded its age, the settings are maintained to ensure efficiency and accuracy. Effectively adapt to equipment conditions, optimize production processes, improve the scientific nature of threshold setting, ensure production stability, and improve product quality and production efficiency.

[0035] Optionally, the method further includes extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type, which are specifically as follows:

[0036] Analyze whether the predicted fault type is output;

[0037] If not, collect detailed data flow records of the production equipment and its preceding equipment that exceed the flow threshold range;

[0038] Calculate the correlation coefficient between the data flow of the device exceeding the threshold and the device preceding it according to a preset correlation calculation method;

[0039] If the correlation coefficient exceeds the preset correlation coefficient value, it is judged that the fault of the current device is greatly correlated with the previous device, and a line graph of the data flow of the device exceeding the threshold and the previous device over time is drawn, and the line graph and the fault situation are sent to the terminal held by the person in charge.

[0040] By adopting the above technical solution, it is possible to confirm in time whether the fault type is output. If not, the fault correlation between devices can be determined by calculating the correlation coefficient. Once the correlation is large, a line graph is drawn and sent to the person in charge to let him or her understand the fault situation intuitively. This helps to quickly locate the source of the problem, improve the efficiency of fault handling, reduce production losses, enhance the person in charge's control and decision-making ability over faults, and ensure smooth production.

[0041] Optionally, according to the correspondence between the time range within which the downtime for handling the fault falls and the emergency treatment plan, the emergency treatment plan is analyzed and determined to include:

[0042] Analyze whether the fault of the current device is closely related to the previous device;

[0043] If not, maintain the original setting;

[0044] If yes, the previous equipment will be repaired first, and the current equipment will continue to be repaired according to the currently determined emergency treatment plan.

[0045] By adopting the above technical solutions, this emergency treatment plan has a good effect. It can determine the correlation of equipment failures. If they are not related, the original settings are maintained, which is efficient and convenient. When there is a correlation, the previous equipment is repaired first to avoid a chain reaction of failures and reduce production delays. It helps to reasonably allocate maintenance resources, improve maintenance efficiency, ensure orderly production, reduce losses caused by failures, and improve overall production stability.

[0046] Optionally, the equipment in front is repaired first, and the current equipment continues to be repaired according to the currently determined emergency treatment plan, including:

[0047] According to the importance of the previous equipment and the current equipment to production, analyze and determine the processing time proportion of different equipment;

[0048] Analyze and determine the downtime for troubleshooting of different devices based on the processing time ratio and downtime for troubleshooting of different devices;

[0049] According to the analysis, the downtime for troubleshooting of different equipment is determined and the equipment is shut down to handle the fault.

[0050] By adopting the above technical solution, the processing time ratio is determined according to the importance of the equipment, and then the accurate downtime processing time is obtained. It can reasonably allocate limited time resources, give priority to important equipment, and minimize production losses. It makes maintenance more targeted and scientific, improves overall maintenance efficiency, stabilizes production order, and enhances the company's ability to deal with equipment failures.

[0051] In the second aspect, the present application provides a traffic data self-identification system based on big data behavior analysis, which adopts the following technical solutions:

[0052] A traffic data self-identification system based on big data behavior analysis includes a memory, a processor, and a program stored in the memory and executable on the processor. The program can be loaded and executed by the processor to implement the traffic data self-identification method based on big data behavior analysis as described in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of a method for self-identification of traffic data based on big data behavior analysis in an embodiment of the present application.

[0054] Figure 2 This is a schematic diagram of a process for analyzing and determining the flow threshold corresponding to each production equipment on the current production line in another embodiment of the present application. Figure 1 .

[0055] Figure 3 This is a schematic diagram of a process for analyzing and determining the flow threshold corresponding to each production equipment on the current production line in another embodiment of the present application. Figure 2 .

[0056] Figure 4 This is a schematic diagram of a process for analyzing and determining the flow threshold corresponding to each production equipment on the current production line in another embodiment of the present application. Figure 3 .

[0057] Figure 5 Another embodiment of the present application is a flowchart of the steps after extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type. DETAILED DESCRIPTION

[0058] The present application is further described in detail below in conjunction with the accompanying drawings.

[0059] Reference Figure 1 , is a traffic data self-identification method based on big data behavior analysis disclosed in this application, comprising:

[0060] Step S100, obtaining the data flow of each production equipment on the current production line.

[0061] Among them, each production equipment is integrated with a data acquisition module, which can accurately obtain the data flow of the equipment in real time. The data flow of each production equipment is mainly obtained in real time through the software for data flow acquisition of production equipment.

[0062] In addition, according to the distribution of production equipment and network topology, a reasonable data acquisition network is designed to ensure the stability and reliability of data acquisition. Industrial Ethernet, wireless sensor network (WSN) and other technologies are used to achieve real-time communication between production equipment and data acquisition system.

[0063] Assuming that a production line contains three types of production equipment, A, B, and C, we can obtain the data flow of these equipment through the following steps: First, cooperate with the equipment manufacturer to integrate data acquisition modules inside the three types of equipment A, B, and C. These modules can record the data flow generated by the equipment in real time during operation and transmit the data to the data acquisition network through the built-in communication interface.

[0064] Next, we arranged several data acquisition nodes (such as industrial Ethernet switches or wireless sensor network nodes) at the production site according to the distribution of production equipment. These nodes are responsible for receiving data from the equipment data acquisition module and transmitting the data to the data acquisition server in the data center. In the data center, we deployed software specifically for data flow acquisition of production equipment. The software can automatically identify the three types of devices A, B, and C connected to the data acquisition network and configure the corresponding data acquisition parameters according to the device type. The software obtains the data flow of each device in real time and stores it in the database for subsequent analysis.

[0065] Step S200, analyzing and determining the flow threshold corresponding to each production equipment on the current production line according to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of the production equipment.

[0066] The device type refers to the type or model of different production equipment on the production line. Different types of equipment may generate different data traffic patterns when processing tasks. The device type can be obtained by obtaining the device type information through the device label, data records in the management system, or from the device manufacturer.

[0067] Working hours refer to the time period when production equipment is running. The load and data traffic of equipment may vary in different time periods, such as the difference between peak and off-peak periods. Working hours are obtained as follows: the working hours of equipment are determined through the work log of the production line, production plan or real-time monitoring system.

[0068] The magnitude of a production task refers to the amount of tasks or products that the equipment needs to process during production. The magnitude of tasks directly affects the size and pattern of data traffic. The magnitude of a production task can be obtained by reading the current magnitude of the production task from the production plan, work order system, or directly from the production line control system.

[0069] The flow threshold refers to a reasonable data flow range set according to the type of equipment, working hours and production task level. Exceeding this range may mean that the equipment is abnormal.

[0070] The analysis and determination of the flow thresholds corresponding to each production equipment on the current production line are as follows: taking the equipment type, working time period, and magnitude of production tasks of the production equipment as query objects, the flow threshold corresponding to each production equipment is queried from a preset database that stores the mapping relationship between the equipment type, working time period, magnitude of production tasks and flow thresholds.

[0071] For example, for a particular model of assembly machine, a reasonable data traffic threshold might be set between 10MB and 50MB per minute during normal working hours and when processing a moderate amount of tasks.

[0072] Step S300, analyzing the production equipment that exceeds the flow threshold range based on the comparison result of the data flow of each production equipment on the current production line and the flow threshold range corresponding to each production equipment on the current production line.

[0073] Specifically, the system obtains the current data flow of each production device and compares it with the flow threshold determined by the device in step S200. This threshold is set based on multiple factors such as the type of device, working period, and the magnitude of the production task, and represents the data flow range of the device under normal working conditions.

[0074] If a device's data flow exceeds its corresponding threshold, the system will mark it as an abnormal device. This abnormality may mean that there is a problem inside the device, such as worn parts, program errors, or other failures that need attention.

[0075] Suppose we monitor that the data flow of a CNC machine tool suddenly increases significantly during a certain working period, exceeding its normal flow threshold range. This may indicate that a component inside the machine tool may have failed, or there is a problem with the processing program, causing the machine tool to perform repeated or invalid operations, thereby generating abnormally large data flow.

[0076] Step S400, collecting all relevant data of the production equipment that exceeds the flow threshold range, all relevant data include operation history, maintenance records and detailed characteristics of real-time data flow.

[0077] Assume that on a production line, the data flow of a packaging machine in a certain period of time increases abnormally, exceeding its normal flow threshold range. In step S400, the operation history data of the packaging machine will be collected first, such as its last startup time, how many hours it has been running continuously, whether there are frequent shutdown records, etc. Then, its maintenance records will be checked to find out when it was last maintained, whether key components have been replaced, whether there are any known maintenance problems, etc.

[0078] Finally, we will also analyze in detail the characteristics of the real-time data flow of this packaging machine, such as whether the abnormally increased data flow persists, whether there is an obvious fluctuation pattern, and whether the time of the peak is related to a specific production task or operation.

[0079] Step S500, extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type.

[0080] Specifically, the system will extract key features from all relevant data collected in step S400, which may include specific patterns in the equipment operation history, frequent problem points in maintenance records, and specific manifestations of abnormal real-time data flow, etc. These features will be correlated with known failure modes and form the basis for failure prediction.

[0081] Next, the system inputs these extracted features and real-time data into a trained fault type prediction model. This model is trained with historical fault data and corresponding features, and it has the ability to identify different fault types. When new data is input, the model analyzes the similarity between this data and the previously learned patterns to predict the most likely fault type.

[0082] Finally, the system outputs the predicted fault type, possible fault cause and recommended maintenance measures. This information is crucial for quickly locating the problem, developing a maintenance plan and preventing similar faults from happening again.

[0083] Assume that on a packaging production line, the data flow of a carton sealing machine is abnormal and is captured by the system. In step S500, the system first extracts the operation history characteristics of the machine, such as the increase in running time and the increase in the number of downtimes in recent days. At the same time, the system checks the maintenance records and finds that the recent maintenances all involve the replacement of a specific component. In addition, the detailed characteristics of the real-time data flow also show a periodic fluctuation pattern.

[0084] The system inputs these features into the fault type prediction model, which, after analysis and comparison, predicts that the carton sealing machine may have a fault related to the wear of a specific component. The system then outputs this fault type and related maintenance suggestions, such as regularly checking the wear of the component and preparing replacement parts in advance.

[0085] Step S600, analyzing whether the predicted fault type is a fault that requires shutdown processing. If not, executing step S700; if yes, executing step S800.

[0086] Specifically, the system will refer to the preset fault type classification standards and divide the predicted fault types into two categories: one is the fault that can continue production by adjusting or optimizing operations, that is, non-stop processing faults; the other is the fault that may cause serious damage to the equipment or affect production safety, and requires immediate shutdown for investigation and repair, that is, shutdown processing faults.

[0087] Once the fault type handling method is determined, the system will select the next step to perform according to the classification result. If the fault does not require downtime, the system will go to step S700 to monitor and record data flow changes; if the fault requires downtime, the system will go to step S800 to analyze the impact of downtime on product delivery time.

[0088] Step S700, obtaining the data flow of the production equipment with the predicted fault type within a preset time period and forming a data flow change table, and sending the data flow change table and the fault type to the terminal held by the person in charge.

[0089] In this step, the system will focus on production equipment that is predicted to have failures but does not need to be shut down immediately. By continuously monitoring the data traffic of these devices over a preset time period, the system can generate a detailed data traffic change table to reveal potential fluctuations in equipment performance and problem development trends.

[0090] Importantly, these data flow change tables and fault type information will be sent in real time to the terminal device held by the production line manager, such as a smartphone or tablet. This instant messaging mechanism ensures that the manager can obtain key equipment status updates in the first place, providing strong data support for subsequent maintenance decisions.

[0091] Assume that an injection molding machine on the production line is predicted to have a minor temperature control system failure, but it will not immediately affect the production progress and does not require emergency shutdown. In this case, step S700 will be triggered. The system begins to monitor the data flow of this injection molding machine in the next few hours, paying special attention to data changes related to the temperature control system.

[0092] Through continuous data collection and analysis, the system generates a detailed data flow change table showing the corresponding relationship between temperature fluctuations and time. At the same time, the system also attaches information on the predicted fault type, that is, minor faults predicted by abnormal patterns in the temperature control system data.

[0093] The report was then automatically sent to the production line manager’s mobile app. After reviewing the report, the manager realized that although the fault had a minor impact at the moment, it still required close attention and planned to perform preventive maintenance in the next shift to prevent the fault from getting worse.

[0094] Step S800, obtain the preset downtime for processing the fault of the production equipment that needs to be shut down to process the fault, the impact of the production equipment that needs to be shut down to process the fault on the overall product production efficiency and the preset product delivery time, and analyze whether it will cause the product delivery to exceed the preset product delivery time. If yes, execute step S900; if not, execute step SA00.

[0095] To analyze whether the product delivery will exceed the preset delivery time, you can follow the following steps: First, calculate the time required to complete the production task under normal circumstances. Assume that the original production task requires T0 time to complete. Then, calculate the loss of production efficiency caused by the shutdown of the production equipment that needs to be shut down to deal with the fault. Assume that the overall production efficiency is reduced by R due to the shutdown of the equipment (for example, 100 products were originally produced per hour, and only 80 products can be produced per hour after the shutdown, then R=20%). Next, calculate the downtime T1 of the production equipment that was shut down to deal with the fault. Then, the delay time of the production task caused by the shutdown is T1. Finally, add the time required for normal production T0 and the delay time T1 caused by the shutdown to obtain the actual possible time to complete the production task T=T0+T1×R. If T is greater than the preset product delivery time, it means that the product delivery will exceed the preset delivery time; conversely, if T is less than or equal to the preset product delivery time, it will not cause the product delivery to exceed the preset delivery time. For example, the original production task takes 100 hours to complete, and the production equipment downtime required to handle the fault is 10 hours, which causes a 30% reduction in overall production efficiency. The delay time is 10×30%=3 hours, and the actual completion time is 100+3=103 hours. If the preset product delivery time is 102 hours, then the product delivery will exceed the preset delivery time; if the preset product delivery time is 105 hours, then the product delivery will not exceed the preset delivery time.

[0096] Step S900, based on the preset product delivery time and the impact of the production equipment that needs to be shut down to handle the fault on the overall product production efficiency, reverse analysis is performed to obtain the shutdown time for handling the fault, and based on the correspondence between the time range within which the shutdown time for handling the fault falls and the emergency handling plan, the emergency handling plan is analyzed and determined.

[0097] It should be noted that different production equipment plays different roles in the overall production process. The downtime of some key equipment may have a significant impact on production efficiency, thereby threatening the on-time delivery of products.

[0098] Suppose a core cutting device on a key production line of a manufacturing company fails. The operating efficiency of the device directly affects the output of the entire production line. According to the order requirements, the delivery time of a batch of important products is 15 days later. Under normal circumstances, the device can complete the cutting of 1,000 units of products per day, while the daily output of the entire production line is 3,000 units.

[0099] Impact analysis: The shutdown of the cutting equipment will cause the daily output of the production line to drop to 2,000 units, a decrease of 33%.

[0100] Reverse analysis of downtime: Considering the 15-day delivery deadline and the impact of equipment downtime on production, the system calculates that the maximum allowable downtime is 5 days. This is because even if the equipment is down for 5 days, the remaining 10 days can still produce at a rate of 2,000 units / day, and a total of 20,000 units can be completed, which can meet the minimum delivery requirements.

[0101] Emergency response plan: After determining the maximum downtime, the system will ensure that the fault is resolved and production is resumed within 5 days based on preset emergency response plans, such as quickly purchasing replacement parts, mobilizing professional maintenance personnel, and adjusting other aspects of the production line to mitigate the impact of the downtime.

[0102] Step SA00, maintain the original settings.

[0103] Reference Figure 2 , an implementation scheme for analyzing and determining the flow threshold corresponding to each production equipment on the current production line is as follows:

[0104] Step S210, analyzing whether the service life of the production equipment exceeds the preset service life. If yes, execute step S220; if no, execute step S230.

[0105] The equipment life of production equipment can be extracted from the equipment management system or equipment archives. These data usually include the equipment's purchase date, commissioning date, and scheduled service life.

[0106] Take a production line with multiple CNC machine tools as an example. First, we extract the purchase date and expected service life of each CNC machine tool from the equipment management system. Assume that the purchase date of one of the machine tools is January 1, 2015, and the expected service life is 10 years. Then, by January 1, 2023, this machine tool will have been in use for 8 years.

[0107] Next, we compare the 8 years that the machine tool has been in use with the preset service life of 10 years. Since 8 years is less than 10 years, it means that the machine tool has not exceeded the preset service life. According to the logic of step S210, we will execute step S230, that is, maintain the flow threshold for this machine tool in the original setting.

[0108] Step S220, analyze and determine the flow threshold corresponding to each production equipment on the current production line based on the mapping relationship between the equipment type of the production equipment, the service life range of the production equipment, the working period, the magnitude of the production task and the flow threshold.

[0109] The specific age range can be extracted from the device management system or calculated based on the purchase date and expected service life of the device. The age of the device will affect its performance and stability, which in turn affects the setting of the flow threshold.

[0110] Take a production line that includes packaging machines and quality inspection equipment as an example. Assume that a packaging machine has been used for 8 years, which falls within the service life range of 6-10 years. This packaging machine works 8 hours a day and mainly undertakes medium-volume packaging tasks. Based on this information, we can find the corresponding traffic threshold from the preset mapping relationship.

[0111] Specifically, for this packaging machine, since its service life is in the medium range, its working hours are fixed every day, and the production task volume is moderate, we can set a flow threshold that can both ensure stable operation of the equipment and meet production needs. For example, the maximum packaging volume per hour can be set to 1,000 pieces as the flow threshold of this packaging machine.

[0112] Step S230, maintaining the original settings.

[0113] Reference Figure 3 , another implementation scheme for analyzing and determining the flow threshold corresponding to each production equipment on the current production line is as follows:

[0114] Step S2A0, analyzing whether the production equipment has been modified. If yes, execute step S2B0; if no, execute step S2C0.

[0115] Whether the production equipment has been modified can be extracted from the equipment management system. The equipment management system usually records the historical information of equipment purchase, installation, maintenance, modification, etc. By analyzing these records, we can determine which equipment has been modified.

[0116] Assume that there is a production line with multiple packaging machines. In step S2A0, we first extract the historical records of each device from the equipment management system. By analyzing these records, we find that one of the packaging machines underwent an equipment modification a year ago, mainly upgrading the conveyor belt and motor system to improve its production efficiency.

[0117] Step S2B0, obtain the extent of the improvement in production efficiency after equipment transformation, and analyze and determine the flow threshold corresponding to each production equipment on the current production line based on the mapping relationship between the extent of the improvement in production efficiency after equipment transformation, the equipment type of production equipment, the working period, the magnitude of production tasks and the flow threshold.

[0118] The extent of production efficiency improvement after equipment transformation can be obtained by comparing the production efficiency data before and after the equipment transformation. Specifically, the production data before and after the equipment transformation can be collected, such as the output per unit time, production cycle, etc., and the improvement in production efficiency can be quantified by calculating the efficiency improvement percentage before and after the transformation.

[0119] Suppose a CNC machine tool on a production line has undergone equipment modification to improve its cutting speed and precision. In order to determine the flow threshold of this equipment after the modification, we first collect production efficiency data before and after the modification. By comparison, we find that the production efficiency of the machine tool has increased by 20% after the modification.

[0120] Next, we confirmed that the machine tool is a high-precision CNC machine tool, which mainly processes precision parts. At the same time, we learned from the production line scheduling system that the machine tool's working hours are 8 hours a day, and the production task volume is moderate.

[0121] Based on the above information, we can combine factors such as equipment type, working hours, production task volume, and production efficiency improvement to determine the flow threshold of this CNC machine tool after transformation according to the preset mapping relationship. For example, considering the improvement of production efficiency and the demand of production tasks after transformation, we can appropriately increase the flow threshold of the machine tool to ensure that it can give full play to the performance advantages after transformation and meet the actual needs of the production line.

[0122] Step S2C0, maintaining the original settings.

[0123] Reference Figure 4 , another implementation scheme for analyzing and determining the flow threshold corresponding to each production equipment on the current production line is as follows:

[0124] Step S2a0, analyzing whether the service life of the production equipment exceeds the preset service life. If yes, execute step S2b0; if no, execute step S2e0.

[0125] Step S2b0, obtaining a mapping relationship between the range of service life of the production equipment and the extent of reduction in production efficiency, and analyzing and determining the extent of reduction in production efficiency.

[0126] In step S2b0, the key feature is the mapping relationship between the service life range of the production equipment and the magnitude of the reduction in production efficiency. The acquisition of this feature mainly depends on the analysis of the performance data of the production equipment over the years and the performance decay curve or data provided by the equipment manufacturer.

[0127] First, it is necessary to extract the purchase date, commissioning date, and preset service life of the production equipment from the equipment management system. Then, based on the actual service life of the equipment, determine the range of service life it falls into.

[0128] Next, by combining the performance test data, maintenance records, and failure rate information of the production equipment over the years, the relationship between the age of the equipment and the degree of reduction in production efficiency can be analyzed. This relationship may exist in the form of a performance decay curve, which shows the trend of equipment performance gradually decreasing with the increase of service life.

[0129] Assume that a cutting machine on a production line has been in use for 10 years, and the preset service life is 15 years. According to the records and performance data in the equipment management system, we found that the performance of this cutting machine was stable in the first 5 years of use, and the production efficiency did not show a significant decline. However, after 5 years of use, its production efficiency began to gradually decrease at a rate of about 2% per year.

[0130] By comparing with the performance degradation curve provided by the equipment manufacturer, we confirmed the rationality of this reduction. Therefore, in step S2b0, we can calculate that the reduction in production efficiency is 10% (i.e. 5 years × 2% / year) based on the service life of the cutting machine (10 years) and the known performance degradation rate (2% per year).

[0131] This reduction in production efficiency will serve as a key input for analysis in subsequent steps to determine the true extent of the increase in production efficiency after the equipment modification, and ultimately determine the flow threshold corresponding to the cutting machine.

[0132] Step S2c0, analyzing and determining the actual extent of the improvement in production efficiency after the equipment transformation based on the extent of the improvement in production efficiency after the equipment transformation and the extent of the reduction in production efficiency caused by exceeding the preset service life.

[0133] For example, we have an injection molding machine on a production line with a preset service life of 10 years. It is currently 12 years old and recently underwent an equipment modification.

[0134] First, we collected production data before and after the transformation and found that the production efficiency after the transformation increased by 15% compared with before the transformation.

[0135] Then, based on the performance degradation curve provided by the manufacturer, we found that the production efficiency of the equipment will decrease by 2% every year after it has been used for more than 10 years. Therefore, the production efficiency of this injection molding machine has decreased by 4% due to its service life (2 years × 2% / year).

[0136] Next, in step S2c0, we need to analyze and determine the true extent of the production efficiency improvement after the equipment transformation. This true extent is the efficiency improvement after the transformation minus the efficiency reduction due to overuse, that is: 15%-4%=11%. Finally, this true improvement will serve as an important input parameter for analyzing and determining the flow threshold in the subsequent steps.

[0137] Step S2d0, analyze and determine the flow threshold corresponding to each production equipment on the current production line based on the mapping relationship between the actual extent of production efficiency improvement after equipment transformation, equipment type, working hours, magnitude of production tasks and flow threshold.

[0138] Assume that there is a production line with multiple injection molding machines, one of which has recently undergone equipment modification to improve its injection speed and molding quality. At the same time, the service life of this injection molding machine has exceeded the preset lifespan, resulting in a decrease in production efficiency.

[0139] In step S2d0, we first collected the production data of this injection molding machine before and after the transformation, and analyzed that the production efficiency increased by 15% after the transformation, but due to exceeding the service life, the production efficiency decreased by 5%. Therefore, the actual increase in production efficiency after the equipment transformation is 10% (i.e. 15%-5%).

[0140] Next, we confirmed that the type of equipment for this injection molding machine is a high-precision injection molding machine, which is mainly responsible for the injection molding of precision parts. From the production plan, we learned that the machine mainly operates during the daytime and the current production task volume is moderate.

[0141] Based on the above information, we can combine the actual extent of the improvement in production efficiency after equipment transformation (10%), equipment type (high-precision injection molding machine), working hours (daytime) and production task volume (moderate) and other characteristics to determine the current flow threshold corresponding to this injection molding machine according to the preset mapping relationship.

[0142] Step S2e0, maintain the original settings.

[0143] Reference Figure 5A method for self-identification of traffic data based on big data behavior analysis also includes steps after extracting preset features from all relevant data collected, inputting the extracted features and real-time data into a trained fault type prediction model, and outputting the predicted fault type, which are as follows:

[0144] Step SA00, analyzing whether to output the predicted fault type.

[0145] Specifically, if the model successfully predicts one or more fault types based on the input data, the feature value of this step should be "yes"; if the model fails to make a clear prediction of the fault type, the feature value is "no".

[0146] Assume that our fault type prediction model has been trained with a large amount of historical data and can effectively analyze traffic data and predict possible fault types. After real-time data is input into the model for analysis, the model successfully predicts a fault type called "abnormal traffic fluctuation" based on the data characteristics.

[0147] At this point, in step SA00, we have obtained the key feature - "the predicted fault type has been output", and the feature value is "yes". Next, the system will carry out subsequent fault handling processes based on this feature value, such as alarming, recording fault information, etc.

[0148] On the contrary, if the model fails to predict any fault type based on the input data, then in step SA00, the feature value we obtain is "No". At this point, the system may enter step SB00 to further collect relevant data for in-depth analysis of the cause.

[0149] Step SB00, if no, collect detailed data flow records of the production equipment and its preceding equipment that exceed the flow threshold range.

[0150] During the production process, sensors or software monitoring tools installed on the equipment can collect data flow information of each device in real time. This information not only includes the amount of data transmission, but also may include multiple dimensions such as data transmission speed, frequency, and stability.

[0151] Suppose there is an automated production line that includes multiple links such as assembly, testing, and packaging, and each link is completed by different equipment. At a certain moment, the detection system finds that the data flow in the assembly link on the production line is abnormal, that is, the data flow exceeds the preset threshold.

[0152] At this point, the system will enter step SB00 and begin collecting detailed data flow records of the assembly equipment and its preceding equipment (possibly the feeding equipment). These records may include information such as the data transmission volume, transmission peak, and average speed of each device in the past few minutes or hours. By deeply analyzing this data, engineers can more accurately determine the cause of the failure, such as whether the unstable data transmission of the feeding equipment has caused the abnormal data flow of the assembly equipment.

[0153] Step SC00, calculating the correlation coefficient between the data flow of the device exceeding the threshold and the data flow of the preceding device according to a preset correlation calculation method.

[0154] A commonly used correlation calculation method is the Pearson Correlation Coefficient. This method measures the strength and direction of the linear relationship between two variables. Its value range is [-1,1], where 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no correlation.

[0155] During the calculation process, the data flow records of the device that exceeds the flow threshold (recorded as device A) and the device in front of it (recorded as device B) over a period of time are first collected. Then, according to the formula of the Pearson correlation coefficient, the mean, standard deviation, and covariance of the data flow of device A and device B are calculated respectively. Finally, these statistics are used to calculate the correlation coefficient between the data flow of the two devices.

[0156] Suppose that while monitoring network traffic, we find that the data traffic of device A (a server) is abnormal and exceeds the set threshold. After step SB00, we collect detailed data traffic records of device A and its preceding device B (possibly a router) in the past hour. Next, in step SC00, we use the Pearson correlation coefficient to analyze the relationship between the data traffic of the two devices. The calculation results show that there is a significant positive correlation between the data traffic of device A and device B, with a correlation coefficient of 0.85. This indicates that when the data traffic of device B increases, the data traffic of device A also tends to increase, and the changes between the two are highly synchronized.

[0157] Step SD00, if the correlation coefficient exceeds the preset correlation coefficient value, it is judged that the failure of the current device is greatly correlated with the previous device, and a line graph of the data flow of the device exceeding the threshold and the previous device over time is drawn, and the line graph and the fault situation are sent to the terminal held by the person in charge.

[0158] Specifically, data acquisition is achieved through data monitoring tools deployed on production equipment, which can capture the data flow of equipment in real time and record it in the database. The database continuously accumulates this data, providing a basis for subsequent line chart drawing.

[0159] When drawing a line graph, the horizontal axis usually represents time, which can clearly reflect the trend of traffic data over time, while the vertical axis represents the data traffic size of each device. By comparing the traffic data of the device that exceeds the threshold and the device in front of it in the same time period, the relationship between the two can be intuitively displayed.

[0160] According to the corresponding relationship between the time range of the downtime for fault handling and the emergency treatment plan, the emergency treatment plan is analyzed and determined to include:

[0161] Step S910, analyzing whether the failure of the current device is significantly related to the previous device. If not, executing step S920; if yes, executing step S930.

[0162] Step S920, maintaining the original settings.

[0163] Step S930: The equipment in front of the equipment is repaired first, and the current equipment continues to be repaired according to the currently determined emergency treatment plan.

[0164] Assume that in an automated production line, device A is the preceding device of device B. When device A fails, it may cause insufficient input materials for device B, which in turn causes the failure of device B. At this point, we can assume that the failure of device A is highly correlated with the failure of device B.

[0165] In order to obtain this feature, we collected the failure data of device A and device B in the past period of time. By analyzing this data, we found that whenever device A fails, device B will also fail soon, and the time interval between the two shows a certain regularity. This shows that the failure of device A is closely related to the failure of device B.

[0166] Based on this feature, when a similar situation occurs again, we can give priority to repairing and processing equipment A, while allowing equipment B to continue to operate according to the currently determined emergency treatment plan.

[0167] Priority will be given to repairing the equipment in front, and the current equipment will continue to be repaired according to the currently determined emergency treatment plan, including:

[0168] Step S931, based on the preset importance of the previous equipment and the current equipment to production, analyze and determine the processing time proportion of different equipment.

[0169] In step S931, the key feature is "the importance of the previous equipment and the current equipment to production". The acquisition of this feature requires comprehensive consideration of multiple dimensions, including but not limited to the production capacity of the equipment, the frequency of failures, the difficulty of maintenance, the ease of obtaining spare parts, and the criticality of the equipment in the production process.

[0170] In order to quantify this feature, we can design an evaluation system to score each dimension and finally summarize it into a comprehensive score to indicate the importance of the equipment to production. For example, we can set the production capacity score to 30%, the failure frequency score to 20%, the maintenance difficulty score to 15%, the ease of obtaining spare parts score to 15%, and the criticality score in the production process to 20%. Then, score according to the actual situation of each dimension, and finally take the weighted average to get the comprehensive importance score of the equipment.

[0171] Assume that there are two key devices in a production line: device X (previous device) and device Y (current device). We evaluate the importance of these two devices to determine their processing time share.

[0172] For example, we can set an importance score from 1 to 10 for each device, where 10 represents the highest importance. The score can be based on multiple dimensions, such as contribution to production efficiency, production loss after failure, etc.

[0173] Equipment X: Since it is the bottleneck of the production line, its failure will directly lead to the stagnation of the entire production line. We assign it a higher importance score, for example 8.

[0174] Equipment Y: Although equipment Y is also important, its failure will not cause the entire production line to come to a standstill. It will only slow down production. Therefore, we assign it a slightly lower importance score, such as 6.

[0175] Based on the importance scores of the two devices, we can directly calculate their processing time proportions. For example, if the total processing time is 100 time units, then the processing time of device X can be 8 / (8+6)×100=57 time units, while the processing time of device Y is 43 time units.

[0176] Step S932: Analyze and determine the downtime for processing faults of different devices based on the processing time proportions of different devices and the downtime for processing faults.

[0177] Step S933, performing equipment shutdown processing fault processing according to the shutdown processing fault time of different equipment determined through analysis.

[0178] For example, suppose we have two devices, device E (previous device) and device F (current device). Through the analysis of steps S931 and S932, we get the following data: the processing time of device E accounts for 60%, and the processing time of device F accounts for 40%. The total downtime of the device to handle the fault is 10 hours.

[0179] Based on these data, we can calculate: The downtime for equipment E to handle the failure = 10 hours × 60% = 6 hours. The downtime for equipment F to handle the failure = 10 hours × 40% = 4 hours.

[0180] Therefore, in step S933, we will shut down device E to handle the fault for 6 hours, and device F will continue to operate according to the current emergency handling plan during the 6 hours when device E handles the fault. After device E completes the handling, device F will be shut down for another 4 hours to handle the fault.

[0181] This approach ensures that equipment that is more important to production is handled first within a limited time, thereby minimizing production losses. At the same time, it also ensures that all equipment can receive the necessary maintenance treatment.

[0182] Based on the same inventive concept, an embodiment of the present invention provides a traffic data self-identification system based on big data behavior analysis, including a memory and a processor, wherein the memory stores a data stream that can be executed on the processor to implement the following Figures 1 to 5 Procedure for either method.

[0183] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A traffic data self-identification method based on big data behavior analysis, characterized in that: include: Obtain the data flow of each production equipment on the current production line; According to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of production equipment, analyze and determine the flow threshold corresponding to each production equipment on the current production line; According to the comparison result of the data flow of each production equipment on the current production line and the flow threshold range corresponding to each production equipment on the current production line, the production equipment exceeding the flow threshold range is analyzed; Collect all relevant data of production equipment that exceeds the flow threshold range, including operation history, maintenance records and detailed characteristics of real-time data flow; Extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type; Analyze whether the predicted fault type is a fault that requires shutdown for processing; If not, the data flow of the production equipment with the predicted fault type within the preset time period is obtained and a data flow change table is formed, and the data flow change table and the fault type are sent to the terminal held by the person in charge; If yes, then obtain the preset downtime for the production equipment that needs to be shut down to handle the fault, the impact of the production equipment that needs to be shut down to handle the fault on the overall product production efficiency, and the preset product delivery time, and analyze whether it will cause the product delivery to exceed the preset product delivery time; If yes, then according to the preset product delivery time and the impact of the production equipment that needs to be shut down for fault processing on the overall product production efficiency, reverse analysis is performed to obtain the shutdown time for fault processing, and according to the corresponding relationship between the time range within which the shutdown time for fault processing falls and the emergency processing plan, the emergency processing plan is analyzed and determined; If not, maintain the original setting; The method further includes extracting preset features from all relevant data collected, inputting the extracted features and real-time data into the trained fault type prediction model, and outputting the predicted fault type, which are as follows: Analyze whether the predicted fault type is output; If not, collect detailed data flow records of the production equipment and its preceding equipment that exceed the flow threshold range; According to a preset correlation calculation method, the correlation coefficient between the data flow of the production equipment exceeding the flow threshold range and the equipment preceding it is calculated; If the correlation coefficient exceeds the preset correlation coefficient value, it is determined that the fault of the current device is highly correlated with the previous device, and a line graph of the data flow of the production device that exceeds the flow threshold range and the previous device over time is drawn, and the line graph and the fault situation are sent to the terminal held by the person in charge; According to the corresponding relationship between the time range of the downtime for fault handling and the emergency treatment plan, the emergency treatment plan is analyzed and determined to include: Analyze whether the fault of the current device is closely related to the previous device; If not, maintain the original setting; If yes, the previous equipment will be repaired first, and the current equipment will continue to be repaired according to the currently determined emergency treatment plan; Priority will be given to repairing the equipment in front, and the current equipment will continue to be repaired according to the currently determined emergency treatment plan, including: According to the importance of the previous equipment and the current equipment to production, analyze and determine the processing time proportion of different equipment; Analyze and determine the downtime for troubleshooting of different devices based on the processing time ratio and downtime for troubleshooting of different devices; According to the analysis, the downtime for troubleshooting of different equipment is determined and the equipment is shut down to handle the fault.

2. According to claim 1, a method for self-identification of traffic data based on big data behavior analysis is characterized in that: According to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of production equipment, the flow threshold corresponding to each production equipment on the current production line is analyzed and determined, including: Analyze whether the service life of production equipment exceeds the preset service life; If yes, then analyze and determine the flow threshold corresponding to each production equipment on the current production line based on the mapping relationship between the equipment type, the service life range of the production equipment, the working period, the magnitude of the production task and the flow threshold; If not, keep the original setting.

3. According to the method of self-identification of traffic data based on big data behavior analysis in claim 1, it is characterized in that: According to the mapping relationship between the equipment type, working period, magnitude of production tasks and flow threshold of production equipment, the flow threshold corresponding to each production equipment on the current production line is analyzed and determined, including: Analyze whether the production equipment has been modified; If yes, then obtain the extent of production efficiency improvement after equipment transformation, and analyze and determine the flow threshold corresponding to each production equipment on the current production line according to the mapping relationship between the extent of production efficiency improvement after equipment transformation, equipment type of production equipment, working hours, magnitude of production tasks and flow threshold; If not, keep the original setting.

4. According to claim 3, a method for self-identification of traffic data based on big data behavior analysis is characterized in that: According to the mapping relationship between the degree of production efficiency improvement after equipment transformation, the equipment type, working hours, magnitude of production tasks and flow thresholds, the flow thresholds corresponding to each production equipment on the current production line are analyzed and determined, including: Analyze whether the service life of production equipment exceeds the preset service life; If yes, then obtain the mapping relationship between the range of service life of the production equipment and the extent of reduction in production efficiency, and analyze and determine the extent of reduction in production efficiency; Analyze and determine the actual extent of the improvement in production efficiency after equipment transformation based on the extent of the improvement in production efficiency after equipment transformation and the extent of the reduction in production efficiency caused by exceeding the preset service life; According to the actual extent of the production efficiency improvement after the equipment transformation, the equipment type of production equipment, working hours, the magnitude of production tasks and the mapping relationship between the flow threshold, analyze and determine the flow threshold corresponding to each production equipment on the current production line; If not, keep the original setting.

5. A traffic data self-identification system based on big data behavior analysis, characterized in that: It includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program can be loaded and executed by the processor to implement a traffic data self-identification method based on big data behavior analysis as described in any one of claims 1 to 4.

Images