A monitoring system, method and device for metal recycling plant manufacturing

Abstract

The application provides a monitoring system, method and device for metal recycling equipment manufacturing, and belongs to the technical field of edge transmission and monitoring. The system comprises: an equipment determination module configured to configure an edge computing device according to a work link baseline consistent with a process feature and in combination with a monitoring feature set; a data analysis module configured to perform data analysis on real-time monitoring information to obtain first progress information of each work link under a corresponding manufacturing process; a data integration module configured to integrate all the first progress information to obtain second progress information; and a project monitoring module configured to compare preset progress information with the second progress information, generate an early warning traceability link, and issue a progress early warning to a management and control end based on a cloud platform. The problems of unreasonable configuration of traditional monitoring equipment, large progress evaluation deviation and untimely early warning are effectively solved, and the management and control precision and manufacturing efficiency of the metal recycling equipment manufacturing process are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of edge transmission and monitoring technology, and in particular to a monitoring system, method and apparatus for manufacturing metal recycling equipment. Background Technology

[0002] The manufacturing process of metal recycling equipment, such as scrap steel shredders, non-ferrous metal sorting machines, and metal balers, involves multiple differentiated manufacturing processes, including cutting, welding, machining, assembly, and debugging. Each process contains several closely related work steps, and the requirements for equipment load, material supply, and process precision vary significantly from step to step. Current technical deficiencies in the monitoring field of metal recycling equipment manufacturing include: edge computing device configuration largely relies on manual experience for allocation, leading to insufficient monitoring resources for some core processes, such as the lack of targeted temperature and weld quality monitoring equipment in the welding process, while resources are redundant for non-core processes. Furthermore, in existing technologies, progress data for each step is stored in different devices, potentially resulting in unused data during data analysis and significant evaluation bias. These factors combined lead to low control precision and efficiency in the metal recycling equipment manufacturing process, which in turn easily results in delays and substandard product quality.

[0003] Therefore, the present invention proposes a monitoring system, method and apparatus for manufacturing metal recycling equipment. Summary of the Invention

[0004] This invention provides a monitoring system, method, and apparatus for manufacturing metal recycling equipment, in order to solve the aforementioned technical problems.

[0005] This invention provides a monitoring system for the manufacturing of metal recycling equipment, comprising:

[0006] The equipment determination module is used to match the corresponding work process baselines based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, and combine the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. It also analyzes the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and constructs a dynamically updated contributing equipment baseline to configure at least one edge computing device for each manufacturing process.

[0007] The data analysis module is used to send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence correlation weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links.

[0008] The data integration module is used to receive the first progress information transmitted by all edge computing devices based on the cloud platform, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project.

[0009] The project monitoring module is used to compare the preset progress information of the target manufacturing project with the second progress information, generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue progress warnings to the management and control terminal based on the cloud platform.

[0010] Preferably, the device determination module includes:

[0011] The feature set configuration unit is used to match the work step baselines corresponding to the process features of each manufacturing process from the feature-step lookup table, and at the same time, combine the deployment features and real-time load status of each edge computing device to configure an adaptive monitoring feature set for the corresponding edge computing device. The adaptive monitoring feature set dynamically increases or decreases the collection dimensions according to the device load.

[0012] The matching unit is used to calculate the matching degree between the baseline of each work process and each monitoring feature set using the cosine similarity algorithm, select monitoring feature sets with matching degrees greater than or equal to the preset degree to form the matching subset of the baseline of each work process, and construct the matching baseline of each manufacturing process corresponding to different monitoring feature sets.

[0013] The extraction unit is used to build an association topology map based on the constructed matching baselines, and extract the link association fluctuation and baseline feature function corresponding to each matching baseline under the preset operating standard. The baseline feature function is obtained by fitting historical operating data through an LSTM neural network.

[0014] The judgment unit is used to synchronously identify the work process baseline and the double-precision attribute and work attribute of each matching baseline, and to determine whether the number of cross-process call levels of the corresponding matching baseline exceeds the dynamic threshold, wherein the dynamic threshold is adjusted in real time according to the complexity of the manufacturing process.

[0015] If the number of cases exceeds the limit, the upstream and downstream derived baselines associated with the corresponding matching baseline will be included in the set of baselines to be coordinated in the global analysis, and key coordination nodes will be marked.

[0016] If the priority is not exceeded, increment the priority of the corresponding work step by 1.

[0017] The coupling analysis unit is used to perform multi-dimensional coupling analysis on all baselines in the associated topology map and the set of baselines to be coordinated, based on the link-related fluctuations and baseline characteristic functions. Combined with the priority increment operation results, it generates a global dynamic fluctuation trend term for each working link baseline and simultaneously obtains the resource redundancy coupling correlation term for the corresponding working link baseline of each edge computing device. The factor determination unit is used to analyze the contribution factor of the resource redundancy coupling correlation term of each edge computing device to the global dynamic fluctuation trend term of each working link baseline using a gradient boosting regression algorithm, and constructs a contributing device baseline with automatically updated contribution factors every hour.

[0018] The equipment deployment unit is used to determine the corresponding final equipment baseline by weighted summation and resource constraints based on the global dynamic fluctuation trend item and contributing equipment baseline of each manufacturing process, and to dynamically match and deploy the edge computing devices involved in the final equipment baseline with the corresponding work link baseline of the manufacturing process.

[0019] Preferably, the data analysis module includes:

[0020] The information acquisition unit is used to collect multi-dimensional real-time work data of each working module in the manufacturing process based on the intelligent monitoring components pre-installed in each manufacturing process. The intelligent monitoring components include temperature sensors, current sensors, laser displacement sensors and ultrasonic weld flaw detectors.

[0021] The information transmission and analysis unit is used to send the real-time monitoring information to the edge computing device corresponding to the manufacturing process and perform data analysis.

[0022] Preferably, the information transmission and analysis unit includes:

[0023] The information output subunit is used to perform edge computing on real-time monitoring information received by the edge computing device and synchronously associate it with the real-time working condition tags of the manufacturing process. The real-time working condition tags are generated by quantification through three dimensions: equipment load, material supply stability and process complexity. The unit outputs initial progress information with working condition attributes and associates it with the baseline of the work process corresponding to the initial progress information.

[0024] The qualified determination subunit is used to extract the total manufacturing volume of each sub-stage based on the initial progress information, integrate historical work information, real-time quality inspection data under real-time working conditions and benchmarking data of similar projects, and determine the dynamic predicted qualified rate of the corresponding sub-stage.

[0025] The weight generation subunit is used to generate the dynamic progress impact weight of the sub-process based on the work characteristics of the sub-process, the dynamic prediction pass rate, the real-time quality inspection anomaly rate and the working condition mutation coefficient. It also determines the sampling intensity of the corresponding sub-process by combining the real-time working condition label, and obtains the historical actual progress of the corresponding sub-process at the historical working moment to determine whether it is qualified or not.

[0026] The information adjustment subunit is used to adjust the initial progress information based on the results of the qualification or failure of each sub-step, historical correction coefficients, and real-time working condition fluctuations to obtain the first progress information of the baseline of the corresponding work step.

[0027] Preferably, the data integration module includes:

[0028] The information sorting unit is used to sort all the first progress information according to the dynamic dependency relationship based on the dependency relationship matrix of the corresponding work process baseline of the cloud platform and each edge computing device to obtain the integrated progress information of the target manufacturing project.

[0029] The information adjustment unit is used to adjust the integrated progress information according to the real-time importance weight of the baseline of each work link to obtain the second progress information of the target manufacturing project.

[0030] Preferably, the project monitoring module includes:

[0031] The deviation determination unit is used to calculate the progress deviation value between the preset progress information and the second progress information of the target manufacturing project, associate the dynamic progress influence weight of each work link with the corresponding sub-link, determine the influence level of the progress deviation, and determine the deviation type of the progress deviation by combining the real-time operating condition fluctuation coefficient, real-time quality inspection anomaly rate and cross-link deviation transmission coefficient of the sub-link. The influence level includes the core influence level, the important influence level and the auxiliary influence level. The early warning generation unit is used to associate and integrate the sub-links corresponding to the progress deviation, the dynamic progress influence weight, the deviation cause information and the intelligent solution to generate an early warning traceability link.

[0032] Preferably, the cloud platform includes:

[0033] The value determination unit is used to determine the warning impact value based on the warning impact range, the criticality of the warning sub-link and the deviation spread speed of the work link baseline associated with the warning tracing link.

[0034] The historical retrieval unit is used to retrieve from the historical database the edge node to which the value consistent with the warning impact value belongs and the historical occupancy status of the edge node at the trigger time to obtain a three-dimensional array, and to place the three-dimensional array in a three-dimensional coordinate system to obtain a three-dimensional discrete map.

[0035] Lock the concentrated intersection points in the three-dimensional discrete graph, wherein there is at least one concentrated intersection point;

[0036] The constraint determination unit is used to obtain the frequency of each surrounding point turning to the central intersection point to obtain the surrounding radiation map, and to perform fitting analysis on the surrounding radiation map, and use the fitting function with the most surrounding points as the initial boundary constraint.

[0037] An optimization unit is used to extract the constraint parameter vector of the initial boundary constraint of each set intersection point and construct a constraint matrix to obtain the constraint feature vector of the constraint matrix, and optimize each constraint parameter vector based on the constraint feature vector to obtain a new boundary constraint.

[0038] The queue determination unit is used to merge the surrounding points of each centralized intersection point that satisfy the new boundary constraints into the corresponding centralized intersection point, count the number of points in the corresponding centralized intersection point after merging, sort all the point counts by size, lock the real-time resource occupancy status of the resource pool of the edge node corresponding to each centralized intersection point, and determine the scheduling queue of the corresponding early warning and tracing link.

[0039] The data packet acquisition unit is used to perform dual verification of the early warning and tracing link by combining the historical associated data cached locally on the edge side of the cloud platform and industry standard data, and to perform compression processing on the data that fails the verification in the early warning and tracing link by combining the data field redundancy of the scheduling queue. The compressed data and the verified data are converted into a structured data packet adapted to the cloud management and control communication protocol.

[0040] The permission reading unit is used to read the preset permission range and real-time work role of each control terminal, match the control permission domain corresponding to the associated work process, filter the control terminals with receiving permissions, embed the control terminal identifier and early warning processing priority tag into the structured data packet, and then push it to the corresponding control terminal based on the cloud platform.

[0041] This invention provides a monitoring method for the manufacturing of metal recycling equipment, comprising:

[0042] Step 1: Based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, match the corresponding work process baselines and combine them with the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. Analyze the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and construct a dynamically updated contributing device baseline to configure at least one edge computing device for each manufacturing process.

[0043] Step 2: Send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence association weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links.

[0044] Step 3: Based on the cloud platform, receive the first progress information transmitted by all edge computing devices, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project;

[0045] Step 4: Compare the preset progress information of the target manufacturing project with the second progress information to generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue a progress warning to the management and control terminal based on the cloud platform.

[0046] The present invention provides a monitoring device for the manufacturing of metal recycling equipment, comprising: a processor and a storage device, wherein the storage device is used to store instructions, and when the processor executes the instructions, the method is implemented.

[0047] Compared with the prior art, the beneficial effects of this application are as follows:

[0048] The device identification module enables precise matching between edge computing devices and the manufacturing process, the data analysis module ensures the real-time and accuracy of progress information, the data integration module enables efficient aggregation of the overall project progress, and the project monitoring module enables rapid early warning and tracing of progress deviations. The entire process of device configuration, data collection, progress analysis, and early warning issuance can be completed without manual intervention, effectively solving the problems of unreasonable configuration of traditional monitoring equipment, large deviations in progress assessment, and untimely early warnings, and significantly improving the control accuracy and manufacturing efficiency of the metal recycling equipment manufacturing process.

[0049] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0052] Figure 1This is a structural diagram of a monitoring system for manufacturing a metal recycling device according to an embodiment of the present invention;

[0053] Figure 2 This is a flowchart illustrating a monitoring method for the manufacturing of a metal recycling device according to an embodiment of the present invention. Detailed Implementation

[0054] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0055] This invention provides a monitoring system for the manufacturing of metal recycling equipment, such as... Figure 1 As shown, it includes:

[0056] The equipment determination module is used to match the corresponding work process baselines based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, and combine the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. It also analyzes the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and constructs a dynamically updated contributing equipment baseline to configure at least one edge computing device for each manufacturing process.

[0057] The data analysis module is used to send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence correlation weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links.

[0058] The data integration module is used to receive the first progress information transmitted by all edge computing devices based on the cloud platform, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project.

[0059] The project monitoring module is used to compare the preset progress information of the target manufacturing project with the second progress information, generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue progress warnings to the management and control terminal based on the cloud platform.

[0060] In this embodiment, the target manufacturing project for metal recycling equipment refers to an engineering project that takes a specific model and specification of metal recycling equipment as the manufacturing object, and includes a complete manufacturing process, clear schedule requirements, and quality standards. For example, a machinery manufacturing company undertakes a project to manufacture 50 scrap steel shredders per year. This project clearly requires that all equipment manufacturing be completed within 3 months, the output particle size of the equipment be ≤10mm, and the service life of the machine body be ≥5 years.

[0061] The process characteristics of a manufacturing process refer to the core attributes of the manufacturing process, including quantifiable or clearly defined characteristic parameters such as processing accuracy requirements, process complexity, material requirements, equipment load range, and process correlation strength. This is achieved by reviewing manufacturing process documents, breaking down process parameters from similar historical projects, and extracting characteristic parameters for each manufacturing process in conjunction with industry standards. For example, the process characteristics of the machining manufacturing process for a scrap steel shredder are: processing accuracy ±0.05mm, including 3 continuous cutting processes, requiring the use of high-strength alloy steel materials, equipment operating load of 5-8kW, and correlation strength with the welding process ≥0.8.

[0062] In this embodiment, the work process baseline is a standardized benchmark for a specific manufacturing process, encompassing core elements such as the composition of sub-processes, their sequential relationships, quality standards, progress thresholds, and resource requirements. For example, in the assembly and manufacturing process of a non-ferrous metal sorting machine, the corresponding work process baseline is: parts cleaning → core component assembly → auxiliary component installation → preliminary debugging. The progress thresholds for each sub-process are 2 hours, 4 hours, 3 hours, and 1.5 hours, respectively. The quality standard is a parts installation deviation of ≤0.1mm, and the resource requirements are 2 skilled workers and 1 set of torque wrench tools.

[0063] In this embodiment, the monitoring feature set of the edge computing device refers to the set of monitoring capability parameters possessed by the edge computing device, including the data types that can be collected, such as temperature, current, displacement, etc., as well as the collection accuracy, collection frequency, data processing capabilities, etc.

[0064] The global dynamic fluctuation trend term refers to the dynamic change trend curve of a work process during the project manufacturing cycle, fitted based on the process-related fluctuations and baseline characteristic functions, combined with cross-process coordination relationships and priority adjustment results. It reflects the fluctuation pattern of a work process affected by internal and external factors. For example, the global dynamic fluctuation trend term of the baseline for the welding process of a metal baler is... Where x is the manufacturing time in hours; y is the fluctuation coefficient, ranging from 0 to 10. The larger the value, the more drastic the fluctuation. This trend indicates that the fluctuation coefficient of the welding process gradually decreases in the first 10 hours and slowly increases after 10 hours. This reflects that the equipment break-in stage in the early stage of welding is more volatile, tends to be stable in the middle stage, and fluctuates slightly in the later stage due to equipment fatigue.

[0065] In this embodiment, the resource redundancy coupling correlation term =

[0066] i1=1 represents CPU, i1=2 represents memory, and i1=3 represents bandwidth, with weights of 0.4, 0.3, and 0.3 respectively.

[0067] In this embodiment, a gradient boosting regression algorithm is used, with resource redundancy coupling correlation terms as input features and the stability index of global dynamic fluctuation trend terms as output labels. The model is trained to obtain the contribution factors of each edge computing device, forming a dynamically updated contribution device baseline. For example, for a certain work process baseline, the contribution factors of the three edge computing devices are 0.85, 0.62, and 0.78, respectively, and the corresponding contribution device baselines are: device A (contribution factor 0.85), device C (contribution factor 0.78), and device B (contribution factor 0.62).

[0068] Real-time monitoring information refers to multi-dimensional real-time data collected through intelligent monitoring components during the manufacturing process, including equipment operating parameters, material supply status, processing quality data, environmental parameters, etc.

[0069] In this embodiment, based on the time sequence relationship diagram of the sub-stages of the work process baseline, the analytic hierarchy process (AHP) is used to determine the time sequence association weight of each sub-stage. For example, a certain work process baseline includes: sub-stage 1 (part processing), sub-stage 2 (part inspection), and sub-stage 3 (part assembly), with time sequence association weights of 0.4, 0.3, and 0.3, respectively. This indicates that sub-stage 1, as a preceding stage, has the greatest impact on the subsequent inspection and assembly stages and has the highest weight.

[0070] In this embodiment, the first progress information refers to information reflecting the real-time progress status of each work step baseline in the manufacturing process, obtained after analysis and processing by an edge computing device. This information includes completed time, remaining time, progress completion rate, and quality compliance status.

[0071] .

[0072] Manufacturing process dependencies refer to the relationships between different manufacturing processes, such as their sequence, resource sharing, and outcome association. These include prerequisite dependencies (where one process must start after another process is completed), parallel dependencies (where two processes run simultaneously and share some resources), and outcome dependencies (where the output of one process is the input of another process).

[0073] Real-time importance weight refers to the weight of the importance of the work process baseline dynamically adjusted based on factors such as the real-time progress of the manufacturing project, changes in market demand, and quality risk level. Specifically, a weight adjustment index system is established, such as project schedule pressure accounting for 0.4, quality risk level accounting for 0.3, and market demand urgency accounting for 0.3. The real-time importance weight of each work process baseline is calculated using the fuzzy comprehensive evaluation method.

[0074] The second progress information refers to the information reflecting the overall progress status of the target manufacturing project, obtained by dynamically sorting all the first progress information according to the manufacturing process dependencies and adjusting it in combination with the real-time importance weight of the work link baseline. It includes the overall project completion rate, the progress ranking of each manufacturing process, key bottleneck links, and the estimated completion time. The overall completion rate is the sum of the products of the baseline progress completion rate of each work link and the real-time importance weight of the corresponding link.

[0075] Preset schedule information refers to the planned schedule information of the target manufacturing project and its various stages based on the project contract and manufacturing plan, including the overall project completion rate, the planned schedule of each manufacturing process, and the baseline planned completion time of each work stage.

[0076] The early warning and tracing link refers to the associated data chain generated when there is a deviation between the second progress information and the preset progress information. This chain includes the identification of the deviation sub-link, the dynamic progress impact weight, the deviation cause analysis, and the intelligent solution. It is used to quickly locate the root cause of the deviation and provide a solution direction. For example, in a target manufacturing project, the overall completion rate of the second progress information is 72%, which is lower than the preset progress completion rate of 75%. The generated early warning and tracing link is as follows: Deviation sub-link: Whole machine debugging → Performance testing → Dynamic progress impact weight: 0.35 → Deviation cause: Calibration deviation of the testing equipment leads to a decrease in testing efficiency, increasing the testing time of each piece of equipment by 30 minutes → Intelligent solution: Immediately activate the backup testing equipment, arrange technicians to perform emergency calibration of the original testing equipment, adjust the testing shifts, and extend the testing working time by 2 hours. Here, the progress deviation value = preset progress completion rate - actual progress completion rate is used to determine the deviation sub-link. Combined with the dynamic progress impact weight algorithm and the cause analysis decision tree model, the input of which is parameters such as the working condition fluctuation coefficient and the quality inspection anomaly rate, and combined with the solution knowledge base, the early warning and tracing link is automatically generated.

[0077] The control terminal refers to the terminal equipment used to receive progress warning information from the cloud platform and remotely control the manufacturing process of metal recycling equipment. This includes industrial control computers, smartphones, tablets, etc. Warning information is received via pop-up alerts, SMS notifications, and sound alarms.

[0078] In this embodiment, the baseline feature function is obtained by fitting historical running data through an LSTM neural network, with an input layer dimension of 10, two hidden layers, 64 neurons per layer, and 100 iterations.

[0079] In this embodiment, the process threshold is adjusted in real time according to the complexity of the manufacturing process. For example, the threshold for a complex process is set to 5, and the threshold for a simple process is set to 3.

[0080] The beneficial effects of the above technical solution are as follows: the equipment determination module enables precise matching between edge computing devices and manufacturing processes, avoiding resource waste and monitoring gaps; the data analysis module ensures the real-time and accuracy of progress information; the data integration module enables efficient aggregation and optimization of the overall project progress; and the project monitoring module enables rapid early warning, accurate tracing, and intelligent solution push for progress deviations, all without the need for manual intervention, significantly improving the control precision and manufacturing efficiency of the metal recycling equipment manufacturing process.

[0081] This invention provides a monitoring system for the manufacturing of metal recycling equipment, wherein the equipment determination module includes:

[0082] The feature set configuration unit is used to match the work step baselines corresponding to the process features of each manufacturing process from the feature-step lookup table, and at the same time, combine the deployment features and real-time load status of each edge computing device to configure an adaptive monitoring feature set for the corresponding edge computing device. The adaptive monitoring feature set dynamically increases or decreases the collection dimensions according to the device load.

[0083] The matching unit is used to calculate the matching degree between the baseline of each work process and each monitoring feature set using the cosine similarity algorithm, select monitoring feature sets with matching degrees greater than or equal to the preset degree to form the matching subset of the baseline of each work process, and construct the matching baseline of each manufacturing process corresponding to different monitoring feature sets.

[0084] The extraction unit is used to build an association topology map based on the constructed matching baselines, and extract the link association fluctuation and baseline feature function corresponding to each matching baseline under the preset operating standard. The baseline feature function is obtained by fitting historical operating data through an LSTM neural network.

[0085] The judgment unit is used to synchronously identify the work process baseline and the double-precision attribute and work attribute of each matching baseline, and to determine whether the number of cross-process call levels of the corresponding matching baseline exceeds the dynamic threshold, wherein the dynamic threshold is adjusted in real time according to the complexity of the manufacturing process.

[0086] If the number of cases exceeds the limit, the upstream and downstream derived baselines associated with the corresponding matching baseline will be included in the set of baselines to be coordinated in the global analysis, and key coordination nodes will be marked.

[0087] If the priority is not exceeded, increment the priority of the corresponding work step by 1.

[0088] The coupling analysis unit is used to perform multi-dimensional coupling analysis on all baselines in the associated topology map and the set of baselines to be coordinated, based on the link-related fluctuations and baseline characteristic functions. Combined with the priority increment operation results, it generates a global dynamic fluctuation trend term for each working link baseline and simultaneously obtains the resource redundancy coupling correlation term for the corresponding working link baseline of each edge computing device. The factor determination unit is used to analyze the contribution factor of the resource redundancy coupling correlation term of each edge computing device to the global dynamic fluctuation trend term of each working link baseline using a gradient boosting regression algorithm, and constructs a contributing device baseline with automatically updated contribution factors every hour.

[0089] The equipment deployment unit is used to determine the corresponding final equipment baseline by weighted summation and resource constraints based on the global dynamic fluctuation trend item and contributing equipment baseline of each manufacturing process, and to dynamically match and deploy the edge computing devices involved in the final equipment baseline with the corresponding work link baseline of the manufacturing process.

[0090] In this embodiment, the formula for calculating the weight of dynamic progress impact is as follows: Where W is the dynamic progress impact weight of the corresponding sub-step; F is the work characteristic weight coefficient of the corresponding sub-step; Q is the dynamic predicted pass rate of the corresponding sub-step; r is the pass rate correction factor of the corresponding sub-step; and A is the real-time quality inspection anomaly rate of the corresponding sub-step. is the anomaly correction factor; T is the operating condition mutation coefficient;

[0091] It should be noted that F is based on the statistical analysis of the process link weight ratio of 100 metal recycling equipment manufacturing projects in the same industry. In the manufacturing of metal recycling equipment, the impact of assembly, testing, and processing links on the overall progress is different. After statistically analyzing the resource input ratio and progress correlation of each link in similar projects, the weight of assembly links is determined to be 0.3, testing links 0.25, processing links 0.35, and other auxiliary links 0.2.

[0092] r is determined based on the historical quality inspection pass rate fluctuation data of the metal recycling equipment manufacturing process over 5 years. When the dynamic predicted pass rate of the sub-process is lower than 90%, the risk of schedule deviation increases, so r is set to 0.8; when the pass rate is between 90% and 98% (normal range), r takes the default value of 1.0; when the pass rate is higher than 98%, the schedule stability is enhanced, so r is set to 1.2.

[0093] Through a gradient experiment on the impact of 20 different quality inspection anomaly rates on schedule, it was found that when the quality inspection anomaly rate > 5%, the probability of rework / adjustment in sub-processes increases significantly, with an impact on schedule exceeding 50%. Take 0.5; when the anomaly rate is ≤5%, the impact is <10%. Take 0.9.

[0094] In this embodiment, Where S is the sampling intensity, μ is the weight correction coefficient, λ is the operating condition correction coefficient, and K is the fluctuation coefficient quantified by real-time operating condition tags, with a value range of 1 to 2. K=1 when the equipment load and material supply are stable, and the larger the value of K is when the equipment load / material supply is more volatile. P1 is the historical qualified stability coefficient, and θ is the stability correction coefficient.

[0095] It should be noted that, in order to compensate for slight errors in the weight calculation, μ adopts the industry-standard correction range of 0.9-1.1, with the default value of the middle value of 1.0.

[0096] Based on the statistical correlation data of the sampling inspection frequency of 50 sets of metal recycling equipment manufacturing operating condition fluctuations, when the operating conditions are stable, the sampling inspection frequency can be reduced, so λ is taken as 0.8; when the operating conditions fluctuate drastically, the sampling inspection frequency needs to be increased, so λ is taken as 1.2.

[0097] In this embodiment, some data from the feature-step lookup table is shown in Table 1:

[0098] Table 1. Partial Data from the Feature-Link Comparison Table

[0099]

[0100] In this embodiment, deployment characteristics refer to the attributes related to the installation and deployment of edge computing devices in the manufacturing workshop, including installation location, communication distance, power supply stability, network connection method, etc.

[0101] In this embodiment, the real-time load status refers to the real-time resource usage of the edge computing device during operation, mainly including indicators such as CPU utilization, memory utilization, and network bandwidth utilization.

[0102] In this embodiment, the adaptive monitoring feature set refers to a set of monitoring parameters that dynamically adjusts the data collection dimensions based on the real-time load of the edge computing device. When the load is high, non-critical collection dimensions are reduced, while when the load is low, all dimensions are collected to balance monitoring accuracy and device operating pressure. For example, edge computing device load thresholds are set as follows: Low load: CPU utilization ≤ 60%, memory utilization ≤ 50%; Medium load: 60% < CPU utilization ≤ 80%, 50% < memory utilization ≤ 70%; High load: CPU utilization > 80%, memory utilization > 70%. When an edge computing device is under low load, the adaptive monitoring feature set includes five collection dimensions: temperature, current, displacement, vibration, and energy consumption; under medium load, three core dimensions (temperature, current, and displacement) are retained; and under high load, only two key dimensions (temperature and current) are retained.

[0103] In this invention, the cosine similarity algorithm is used to calculate the matching degree between the baseline of the work process and the monitoring feature set. For example, the feature parameters of the work process baseline are quantified into a vector X = [0.8 (processing accuracy requirement), 0.7 (process complexity), 0.6 (material compatibility), 0.9 (monitoring real-time requirement)], and the capability parameters of the monitoring feature set are quantified into a vector Y = [0.75 (acquisition accuracy), 0.68 (multi-process compatibility), 0.59 (material monitoring range), 0.88 (data transmission delay)], and then calculated using the cosine similarity formula. The calculated matching degree is 0.97.

[0104] In this embodiment, by analyzing the matching data of 100 sets of historical work process baselines and monitoring feature sets, it was found that when the matching degree is ≥0.8, the monitoring coverage of the work process by the monitoring feature set is ≥90%. Therefore, the preset degree is set to 0.8. For example, after the cosine similarity calculation of a certain work process baseline, the matching degrees of 3 monitoring feature sets are 0.92, 0.85 and 0.78 respectively. Among them, the first two have a matching degree ≥0.8. Therefore, the matching subset of the work process baseline is {monitoring feature set A (0.92) and monitoring feature set B (0.85)}.

[0105] In this embodiment, the matching baseline refers to the benchmark data constructed based on the correspondence between the manufacturing process and the matching subset, which includes the manufacturing process, the matching subset, and the monitoring feature set parameters. For example, if the matching subset corresponding to the welding process is {monitoring feature set A, monitoring feature set B}, then the corresponding matching baseline is: welding process - monitoring feature set A (temperature, weld flaw detection, current acquisition), welding process - monitoring feature set B (temperature, vibration, energy consumption acquisition).

[0106] In this embodiment, the Neo4j graph database is used to construct an association topology graph. Nodes store the ID, name, and core parameters of the matching baselines, while edges store information such as association type and association strength. For example, in the association topology graph of a target manufacturing project, the nodes include: machining matching baseline, welding matching baseline, and assembly matching baseline. The relationship between the edges is: machining matching baseline → welding matching baseline (pre-process association), welding matching baseline - assembly matching baseline (data interaction association).

[0107] Preset operating standards refer to the standard parameter ranges pre-defined during the operation of the matching baseline, including data acquisition accuracy standards, operational stability standards, response time standards, etc., used to extract the correlation fluctuations and baseline characteristic functions. For example, the preset operating standards for the matching baseline are: data acquisition accuracy ±0.1mm, operational stability (i.e., continuous fault-free operation time ≥24h), and response time ≤100ms.

[0108] Link-related fluctuation refers to the degree of mutual influence between the operational status fluctuations of the matching baseline and other related operational links under the preset operating standard. The value range is 0-1. By collecting the fluctuation parameters of each operational link in historical operating data, such as dimensional deviation fluctuations and temperature fluctuations, the correlation coefficient is calculated and used as the link-related fluctuation value.

[0109] The baseline feature function refers to a function obtained by fitting historical running data through an LSTM neural network, reflecting the change of the running state of the matching baseline over time, and is used to quantify the dynamic characteristics of the matching baseline. For example, the baseline feature function of a certain matching baseline is f(t) = 0.5sin(0.2t) + 0.8, where t is the running time in hours; f(t) is the running state value, ranging from 0 to 2. This function indicates that the running state of the matching baseline fluctuates periodically over time, with a fluctuation amplitude of 0.5 and a baseline value of 0.8.

[0110] The stage double precision attribute refers to the precision requirement attribute of a work stage, including precision parameters in two dimensions: machining precision and inspection precision. It is used to determine the rationality of the cross-stage calling hierarchy of the matching baseline. For example, the stage double precision attribute of a certain work stage is: machining precision ±0.03mm, inspection precision ±0.01mm.

[0111] Work attributes refer to the core functions, job types, and resource requirements of a work process, and are used to help determine the rationality of cross-process calling hierarchy. For example, the work attributes of a certain work process are: core function: surface treatment of parts, job type: chemical treatment, resource requirements: degreasing agent, drying equipment.

[0112] In this embodiment, based on the association topology map, the upstream and downstream related links of the work link corresponding to the matching baseline are traced, and the number of levels of the related links is counted. Direct association is level one, and indirect association is level two and above. For example, if the work link corresponding to a certain matching baseline is the assembly of core components, it is necessary to call the parts processing (level one), parts inspection (level one), and parts transportation (level two, which is the downstream related link of parts processing). Then the number of cross-link call levels is 2.

[0113] In this embodiment, a manufacturing process complexity assessment index system is established, with the number of work steps accounting for 0.4, process complexity accounting for 0.3, and resource requirement types accounting for 0.3. The complexity level is calculated using the fuzzy comprehensive evaluation method, and the dynamic threshold is adjusted according to the level. For example, the manufacturing process complexity levels are set as follows: Simple: 1-3 work steps, Medium: 4-6 work steps, Complex: ≥7 work steps, with corresponding dynamic thresholds of 2, 3, and 4, respectively. If a manufacturing process contains 8 work steps, it is classified as complex, and the dynamic threshold is 4. For example, if the number of cross-step call levels of matching baseline A is 5, which exceeds the dynamic threshold of 4, then the set of baselines to be coordinated is: {matching baseline A, upstream derived baseline B, downstream derived baseline C, upstream derived baseline D (upstream related baseline of baseline B)}. It should be noted that based on the association relationship of the set of baselines to be coordinated, the number of related baselines of each baseline is counted, and the top 20% of baselines with the most related baselines are marked as key coordination nodes.

[0114] Multi-dimensional coupling analysis refers to the correlation analysis of all baselines in the correlation topology map and the set of baselines to be coordinated from multiple dimensions such as data interaction, resource sharing, progress correlation, and quality impact. It is used to generate global dynamic fluctuation trend items and resource redundancy coupling correlation items.

[0115] The final equipment baseline refers to the benchmark used to determine the final suitable edge computing devices, calculated by weighted summation and resource constraints based on the global dynamic fluctuation trend of the manufacturing process and the contributing equipment baseline. For example, if the stability index of the global dynamic fluctuation trend of a manufacturing process is 0.85, the contribution factors of three edge computing devices are 0.82, 0.65, and 0.73, respectively, and the resource constraints are: equipment operating cost ≤ 500 yuan / day and communication distance ≤ 150m, the equipment comprehensive score is calculated by weighted summation. The device with the highest score that meets the resource constraints is selected to form the final equipment baseline. Comprehensive score = Global dynamic fluctuation trend stability index × 0.6 + Contribution factor × 0.4.

[0116] The beneficial effects of the above technical solution are as follows: the feature set configuration unit enables adaptive configuration of the monitoring feature set; the matching unit ensures a high degree of matching between the baseline of the work process and the monitoring feature set; the extraction unit accurately obtains the core features of the baseline; the judgment unit reasonably divides the scope of collaborative analysis; the coupling analysis unit realizes multi-dimensional correlation analysis; the factor determination unit quantifies the contribution of the device; and the device deployment unit completes accurate deployment. The synergistic effect of each unit significantly improves the scientificity and rationality of the edge computing device configuration.

[0117] This invention provides a monitoring system for the manufacturing of metal recycling equipment, wherein the data analysis module includes:

[0118] The information acquisition unit is used to collect multi-dimensional real-time work data of each working module in the manufacturing process according to the intelligent monitoring components pre-installed in each manufacturing process. The intelligent monitoring components include temperature sensors, current sensors, laser displacement sensors and ultrasonic weld flaw detectors.

[0119] The information transmission and analysis unit is used to send the real-time monitoring information to the edge computing device corresponding to the manufacturing process and perform data analysis.

[0120] In this embodiment, the temperature sensor is PT100; the current sensor is ACS712; the laser displacement sensor is KEYENCELK-G5000; and the ultrasonic weld flaw detector is HS610e.

[0121] Multi-dimensional real-time work data refers to various data reflecting the operational status of the manufacturing process collected through intelligent monitoring components, including equipment operating parameters such as temperature, current, and speed; processing quality data such as dimensional deviations and weld defects; material supply data such as supply speed and material balance; and environmental data such as workshop temperature and humidity.

[0122] A work module refers to an independent work unit with a specific function in the manufacturing process. Each manufacturing process contains several work modules. For example, the welding manufacturing process of a metal baler includes four work modules: weld pretreatment module, welding operation module, weld inspection module, and weld cooling module. The welding operation module corresponds to the spot welding fixation, continuous welding, and weld shaping sub-steps.

[0123] The beneficial effects of the above technical solution are: the information acquisition unit enables multi-dimensional and high-precision real-time data acquisition, ensuring the comprehensiveness and accuracy of monitoring data; the information transmission and analysis unit enables rapid data transmission and efficient analysis, providing a reliable guarantee for the accurate generation of subsequent first-stage progress information.

[0124] This invention provides a monitoring system for the manufacturing of metal recycling equipment, wherein the information transmission and analysis unit includes:

[0125] The information output subunit is used to perform edge computing on real-time monitoring information received by the edge computing device and synchronously associate it with the real-time working condition tags of the manufacturing process. The real-time working condition tags are generated by quantification through three dimensions: equipment load, material supply stability and process complexity. The unit outputs initial progress information with working condition attributes and associates it with the baseline of the work process corresponding to the initial progress information.

[0126] The qualified determination subunit is used to extract the total manufacturing volume of each sub-stage based on the initial progress information, integrate historical work information, real-time quality inspection data under real-time working conditions and benchmarking data of similar projects, and determine the dynamic predicted qualified rate of the corresponding sub-stage.

[0127] The weight generation subunit is used to generate the dynamic progress impact weight of the sub-process based on the work characteristics of the sub-process, the dynamic prediction pass rate, the real-time quality inspection anomaly rate and the working condition mutation coefficient. It also determines the sampling intensity of the corresponding sub-process by combining the real-time working condition label, and obtains the historical actual progress of the corresponding sub-process at the historical working moment to determine whether it is qualified or not.

[0128] The information adjustment subunit is used to adjust the initial progress information based on the results of the qualification or failure of each sub-step, historical correction coefficients, and real-time working condition fluctuations to obtain the first progress information of the baseline of the corresponding work step.

[0129] In this embodiment, the real-time operating condition label refers to a label generated by quantifying three dimensions: equipment load, material supply stability, and process complexity. It is used to characterize the real-time operating condition of the manufacturing process. Each dimension takes a value of 0-1, and the weighted sum is used to obtain the operating condition label value. Among them, equipment load is 0.4, material supply stability is 0.3, and process complexity is 0.3.

[0130] In this embodiment, the total manufacturing volume refers to the total workload indicators such as the number of products, processing length, and assembly times that need to be completed in each sub-step of the work process baseline.

[0131] Historical work information refers to the operational data of similar sub-processes in the previous manufacturing project, including historical pass rates, historical progress completion status, and historical anomaly handling records.

[0132] Real-time quality inspection data refers to the sub-process operation quality inspection data obtained through intelligent monitoring components or manual sampling under the current real-time operating conditions, including dimensional deviations, number of defects, performance parameters, etc. For example, the real-time quality inspection data of a certain sub-process (machining) under the current operating conditions is: 10 workpieces were sampled, and the dimensional deviations were all ≤0.03mm (qualified), there were no surface defects, the performance parameters met the standards, and the real-time quality inspection pass rate was 100%.

[0133] Benchmarking data for similar projects refers to the quality and progress data of corresponding sub-stages in metal recycling equipment manufacturing projects previously completed by other companies or by this company that are similar to the current target manufacturing project. For example, if the current target project is the manufacturing of the XGP-100 scrap steel shredder, the benchmarking data for the XGP-80 scrap steel shredder manufacturing project shows that the average pass rate for the corresponding machining sub-stage is 97%, and the completion time is 3.8 hours.

[0134] The dynamic predicted pass rate refers to the pass rate predicted after the completion of the current sub-stage by integrating historical work information, real-time quality inspection data under real-time working conditions, and benchmarking data of similar projects. For example, if the historical average pass rate of a certain sub-stage is 97.9%, the real-time quality inspection pass rate is 100%, and the benchmarking pass rate of similar projects is 97%, then the pass rate is calculated by weighting the historical data (0.4), the real-time data (0.5), and the benchmarking data (0.1).

[0135] In this embodiment, the work characteristics refer to the core operation type, technical difficulty, quality requirement level, and other characteristics of the sub-process, which are used to generate dynamic progress impact weights. For example, the work characteristics of a certain sub-process are: operation type: precision welding, technical difficulty: high, quality requirement level: level one, which is the highest level.

[0136] The real-time quality inspection anomaly rate refers to the ratio of the number of quality anomalies that occur during the sub-process operation to the total number of operations under the current real-time operating conditions.

[0137] In this embodiment, the deviation of each working condition dimension from the baseline working condition is calculated in real time, and the weighted sum is used to obtain the working condition mutation coefficient. The calculation is performed every 15 minutes. The working condition dimensions include, for example, equipment load dimension, material supply stability dimension, and process complexity dimension.

[0138] In this embodiment, a table is established to correspond to the sampling intensity and the real-time working condition label. For example, label values ​​0-3: sampling ratio 5%, 3 pieces per hour; 3-7: sampling ratio 10%, 5 pieces per hour; 7-10: sampling ratio 15%, 8 pieces per hour.

[0139] In this embodiment, the progress data of sub-processes at different historical working moments are extracted from the historical project progress database, sorted in chronological order, and marked as qualified. Those that meet the planned progress are qualified, and those that do not are unqualified.

[0140] Calculate the deviation rate between the actual historical progress and the planned progress of each sub-stage. The deviation rate is calculated as (actual completion time - planned completion time) / planned completion time. The average deviation rate plus 1.0 is taken as the historical correction coefficient.

[0141] In this embodiment, the absolute value of the difference between two consecutive updated real-time operating condition label values ​​is calculated, and the fluctuation level is divided according to the value: 0-0.5: slight fluctuation; 0.5-1.0: moderate fluctuation; >1.0: severe fluctuation.

[0142] In this embodiment, if the initial progress completion rate is 60%, the historical correction coefficient is 1.02, and the real-time working condition medium fluctuation adjustment coefficient is 1.05, then the adjusted progress completion rate = 60% × 1.02 × 1.05 ≈ 64.26%.

[0143] The beneficial effects of the above technical solution are as follows: the information output sub-unit quickly generates initial progress information with working condition attributes; the qualification determination sub-unit accurately predicts the qualification rate of sub-stages; the weight generation sub-unit quantifies the impact of sub-stages on the progress and reasonably sets the sampling intensity; and the information adjustment sub-unit optimizes the progress information by combining multi-dimensional parameters, which significantly improves the accuracy of the generated first progress information and provides high-quality data support for the subsequent integration of the overall project progress.

[0144] This invention provides a monitoring system for the manufacturing of metal recycling equipment, wherein the data integration module includes:

[0145] The information sorting unit is used to sort all the first progress information according to the dynamic dependency relationship based on the dependency relationship matrix of the corresponding work process baseline of the cloud platform and each edge computing device to obtain the integrated progress information of the target manufacturing project.

[0146] The information adjustment unit is used to adjust the integrated progress information according to the real-time importance weight of the baseline of each work link to obtain the second progress information of the target manufacturing project.

[0147] In this embodiment, the dependency matrix refers to the dependency relationships between the baselines of each work stage, represented in matrix form. The rows and columns of the matrix are the IDs of the work stage baselines, and the matrix elements are 0 (no dependency) or 1 (dependency). For example, a target manufacturing project contains 3 work stage baselines: ID: HJ-001, HJ-002, HJ-003. The dependency matrix is ​​shown in Table 2.

[0148] Table 2 Dependency Matrix

[0149]

[0150] In this embodiment, dynamic dependency sorting refers to sorting based on the dependency matrix according to the order of dependencies between the baselines of the work processes, ensuring that the sorted progress information conforms to the actual operating logic of the manufacturing process. For example, according to the above dependency matrix, the dynamic dependency sorting result is: HJ-001→HJ-002→HJ-003. At this time, the integrated progress information is: Sorting 1: HJ-001: First progress information: Completion rate 85%, remaining time 1h; Sorting 2: HJ-002: First progress information: Completion rate 70%, remaining time 2.5h; Sorting 3: HJ-003: First progress information: Completion rate 60%, remaining time 3h.

[0151] The beneficial effects of the above technical solution are: the information arrangement unit realizes the orderly integration of the first progress information, ensuring that the progress information conforms to the manufacturing process dependency logic; the information adjustment unit optimizes the overall progress calculation by combining real-time importance weights, so that the generated second progress information can accurately reflect the overall progress status of the target manufacturing project, providing reliable overall progress data for project monitoring.

[0152] This invention provides a monitoring system for the manufacturing of metal recycling equipment, wherein the project monitoring module includes:

[0153] The deviation determination unit is used to calculate the progress deviation value between the preset progress information and the second progress information of the target manufacturing project, associate the dynamic progress influence weight of each work link with the corresponding sub-link, determine the influence level corresponding to the progress deviation, and determine the deviation type of the progress deviation by combining the real-time operating condition fluctuation coefficient, real-time quality inspection anomaly rate and cross-link deviation transmission coefficient of the sub-link. The influence level includes the core influence level, the important influence level and the auxiliary influence level.

[0154] The early warning generation unit is used to associate and integrate the sub-links corresponding to the progress deviation, the dynamic progress impact weight, the deviation cause information and the intelligent solution to generate an early warning tracing link.

[0155] In this embodiment, the schedule deviation value refers to the difference between the preset schedule information and the second schedule information of the target manufacturing project. It is used to quantify the degree of schedule deviation. A positive schedule deviation value indicates that the schedule is lagging behind, and a negative schedule indicates that the schedule is ahead of schedule.

[0156] Impact level refers to the hierarchy based on the dynamic schedule impact weight of the sub-stages corresponding to the schedule deviation. It is used to assess the degree of impact of the deviation on the overall project, including core impact level, significant impact level, and secondary impact level. For example, a dynamic schedule impact weight ≥ 0.5 is the core impact level; 0.3 ≤ weight < 0.5 is the significant impact level; and weight < 0.3 is the secondary impact level.

[0157] In this embodiment, a correspondence table between real-time operating condition fluctuation amplitude and fluctuation coefficient is established: 0-0.5: 0.2; 0.5-1.0: 0.5; >1.0: 0.8, and the corresponding real-time operating condition fluctuation coefficient is matched according to the real-time operating condition fluctuation amplitude.

[0158] The cross-stage deviation transmission coefficient refers to the degree of transmission of schedule deviations between different related sub-stages. It is used to assess whether the deviation will spread to other sub-stages. The value ranges from 0 to 1. For example, if the correlation between the welding sub-stage and the assembly sub-stage is 0.8 (high correlation), then the cross-stage deviation transmission coefficient is 0.7; if the correlation between the welding sub-stage and the cleaning sub-stage is 0.3 (low correlation), then the transmission coefficient is 0.2.

[0159] In this embodiment, parameters such as the impact level, real-time operating condition fluctuation coefficient, cross-stage deviation transmission coefficient, and deviation cause are input into the decision tree algorithm to automatically determine the deviation type. For example, the schedule deviation caused by insufficient material supply is a resource shortage type, and the deviation type includes resource shortage type, process abnormal type, operating condition fluctuation type, and related impact type.

[0160] Deviation cause information refers to a description of the specific reasons for schedule deviations, including the type of cause, influencing factors, time of occurrence, and scope of impact. For example, deviation cause information might include: Cause type: resource shortage; Influencing factor: delay in supply of core components; Time of occurrence: Day 20 of the project; Scope of impact: assembly sub-stage of core components, causing a 5% delay in the schedule of this stage; and data such as material supply records and process parameter logs. By using fishbone diagram analysis to pinpoint the root cause of the deviation, the deviation cause information is compiled.

[0161] Intelligent solutions refer to targeted solutions automatically generated based on deviation type, deviation cause, and historical solution database. These solutions include specific operational steps, required resources, estimated resolution time, and responsible department.

[0162] The beneficial effects of the above technical solution are: the deviation determination unit accurately quantifies the schedule deviation, divides the impact level and locates the cause of the deviation, and the early warning generation unit integrates multi-dimensional information to generate a complete early warning and traceability link, which provides precise guidance for management personnel to quickly handle deviations and recover schedule losses, and effectively reduces the risk of project delays.

[0163] This invention provides a monitoring system for the manufacturing of metal recycling equipment, the cloud platform comprising:

[0164] The value determination unit is used to determine the warning impact value based on the warning impact range, the criticality of the warning sub-link and the deviation spread speed of the work link baseline associated with the warning tracing link.

[0165] The historical retrieval unit is used to retrieve from the historical database the edge node to which the value consistent with the warning impact value belongs and the historical occupancy status of the edge node at the trigger time to obtain a three-dimensional array, and to place the three-dimensional array in a three-dimensional coordinate system to obtain a three-dimensional discrete map.

[0166] Lock the concentrated intersection points in the three-dimensional discrete graph, wherein there is at least one concentrated intersection point;

[0167] The constraint determination unit is used to obtain the frequency of each surrounding point turning to the central intersection point to obtain the surrounding radiation map, and to perform fitting analysis on the surrounding radiation map, and use the fitting function with the most surrounding points as the initial boundary constraint.

[0168] An optimization unit is used to extract the constraint parameter vector of the initial boundary constraint of each set intersection point and construct a constraint matrix to obtain the constraint feature vector of the constraint matrix, and optimize each constraint parameter vector based on the constraint feature vector to obtain a new boundary constraint.

[0169] The queue determination unit is used to merge the surrounding points of each centralized intersection point that satisfy the new boundary constraints into the corresponding centralized intersection point, count the number of points in the corresponding centralized intersection point after merging, sort all the point counts by size, lock the real-time resource occupancy status of the resource pool of the edge node corresponding to each centralized intersection point, and determine the scheduling queue of the corresponding early warning and tracing link.

[0170] The data packet acquisition unit is used to perform dual verification of the early warning and tracing link by combining the historical associated data cached locally on the edge side of the cloud platform and industry standard data, and to perform compression processing on the data that fails the verification in the early warning and tracing link by combining the data field redundancy of the scheduling queue. The compressed data and the verified data are converted into a structured data packet adapted to the cloud management and control communication protocol.

[0171] The permission reading unit is used to read the preset permission range and real-time work role of each control terminal, match the control permission domain corresponding to the associated work process, filter the control terminals with receiving permissions, embed the control terminal identifier and early warning processing priority tag into the structured data packet, and then push it to the corresponding control terminal based on the cloud platform.

[0172] In this embodiment, ;

[0173] Wherein, YJ is the warning impact value; Fw is the range value of the warning impact area; Fz is the range value of the maximum impact area; Ns is the warning sub-link involved under the baseline of the corresponding work link; Nz is the total number of sub-links under the baseline of the corresponding work link; gc is the normalized value of the sum of the criticality of all warning sub-links; vc is the mean of the deviation diffusion rate of all warning sub-links; and vmax is the maximum diffusion rate among all warning sub-links.

[0174] In this embodiment, based on the associated topology map, the upstream and downstream associated manufacturing processes and work links of the early warning associated work links are traced, and the number of affected nodes is counted as the range value of the early warning impact range.

[0175] The criticality of a warning sub-link refers to the importance level of the sub-link involved in the warning within the target manufacturing project. For example, if a certain warning sub-link has a dynamic progress impact weight of 0.35, a quality risk level of 0.8, and a resource input ratio of 0.25, the criticality can be calculated by weighted summation as 0.35×0.4+0.8×0.3+0.25×0.3=0.445, where 0.4, 0.3, and 0.3 are the corresponding weights.

[0176] In this embodiment, the deviation status of associated nodes is monitored in real time after the warning is issued, the number of newly affected nodes per unit time is counted, and the deviation spread rate is calculated. For example, if the deviation of a certain warning sub-link affects 3 associated sub-links within 2 hours, the deviation spread rate is 3 / 2 = 1.5 nodes / hour.

[0177] The warning impact value is an indicator that quantifies the severity of a warning, calculated based on the scope of the warning's impact, the criticality of the warning sub-links, and the rate of deviation spread. Its value ranges from 0 to 5.

[0178] The historical resource utilization status of the corresponding edge node at the trigger time refers to the resource utilization status of the edge node when a warning with the same impact value as the current warning was triggered in the past, namely CPU utilization, memory utilization, and bandwidth utilization. For example, if the current warning impact value is 0.8, and the warning record with a warning impact value of 0.8 is retrieved from the historical database, the historical resource utilization status of the edge node BJ-001 corresponding to its trigger time t01 is: CPU utilization 70%, memory utilization 65%, and bandwidth utilization 50%.

[0179] A three-dimensional array refers to an array structure that uses edge node number, trigger time, and historical occupancy status parameters as three-dimensional data to store edge node occupancy information corresponding to historical warnings. For example, the three-dimensional array corresponding to a certain historical warning is ["BJ-001", "2024-03-15 10:00", [70%, 65%, 50%]], where the first dimension is the node number, the second dimension is the trigger time, and the third dimension is the three occupancy status parameters.

[0180] A three-dimensional discrete graph refers to a discrete point graph formed by mapping data from a three-dimensional array onto a three-dimensional coordinate system. The three axes of the three-dimensional coordinate system are the edge node number (quantized to a numerical value), the trigger time, and the historical occupancy status parameter (normalized numerical value). Each three-dimensional array corresponds to one discrete point. For example, mapping the three-dimensional array ["BJ-001" (quantized to 1), "t01" (quantized to t01), [70% (0.7), 65% (0.65), 50% (0.5)]] onto a three-dimensional coordinate system forms three discrete points with coordinates (1, t01, 0.7), (1, t01, 0.65), and (1, t01, 0.5). Multiple similar data points constitute a three-dimensional discrete graph.

[0181] A clustered intersection point (COP) is a coordinate point in a 3D discrete graph where discrete points are densely clustered, with significantly more COPs surrounding it than in other areas. It is used to determine the central tendency of resource occupancy. For example, in a 3D discrete graph, the coordinate (1, t0x, 0.68) is surrounded by 12 COPs, while other areas have a maximum of 3 COPs; therefore, this coordinate point is a COP. By using the DBSCAN algorithm to perform cluster analysis on the discrete points in the 3D discrete graph, the coordinate points corresponding to the cluster centers are the COPs, ensuring that there is at least one COP.

[0182] Surrounding points refer to discrete points within a certain distance range around the intersection of clustered points. The distance range is determined by the radius parameter of the clustering algorithm, such as a radius of 0.05. Based on the normalized coordinate values, the Euclidean distance between each discrete point and the intersection of clustered points is calculated. Discrete points with a distance ≤ the preset radius are determined as surrounding points.

[0183] Point frequency refers to the frequency with which each surrounding point moves towards the convergent intersection point, that is, the number of times the historical warnings corresponding to the surrounding point and the historical warnings corresponding to the convergent intersection point match in terms of resource usage change trends. For example, if the historical warnings corresponding to surrounding point A have 8 instances where the resource usage change trend matches the historical warnings corresponding to the convergent intersection point, then the point frequency is 8.

[0184] A peripheral radiation map is a radial graph drawn with a central intersection point as the center and the frequency of surrounding points as the radiation intensity. It is used to visually demonstrate the degree of correlation between surrounding points and the central intersection point. For example, if the central intersection point is located at the center, the surrounding points are distributed according to their distance from the center. The higher the frequency of the surrounding points, the thicker the radiation lines or the darker the color, thus forming a peripheral radiation map.

[0185] Initial boundary constraints refer to selecting the fitting function with the most surrounding points as the boundary constraint condition after fitting the surrounding radiation map. This is used to define the range of surrounding points closely related to the concentrated intersection points. For example, performing a polynomial fitting on the surrounding radiation map yields the fitting function. , where x is the distance from the intersection of the points, y is the point frequency threshold, the curve corresponding to this function covers the most surrounding points (8), so as the initial boundary constraint, at this time, the constraint parameter vector is [0.03,0.2,0.5], at this time, it is a matrix composed of the constraint parameter vectors of multiple intersection points.

[0186] Constraint eigenvectors are eigenvectors obtained by performing eigenvalue decomposition on the constraint matrix. They are used to characterize the core features of the constraint matrix. The linalg.eig function of the NumPy library is used to perform eigenvalue decomposition on the constraint matrix, and the eigenvector corresponding to the largest eigenvalue is selected as the constraint eigenvector.

[0187] The new boundary constraint refers to the boundary constraint condition obtained by optimizing the constraint parameter vector of the initial boundary constraint based on the constraint feature vector. Compared with the initial boundary constraint, it has a higher accuracy in filtering surrounding points. For example, given the initial constraint parameter vector [0.03, 0.2, 0.5] and the constraint feature vector [0.35, 0.52, 0.78], the optimized new constraint parameter vector is obtained by weighted summation of the vectors: [0.03×0.7+0.35×0.3, 0.2×0.7+0.52×0.3, 0.5×0.7+0.78×0.3] = [0.126, 0.296, 0.584]. The corresponding fitting function is... As a new boundary constraint.

[0188] The resource pool refers to the complete set of available resources of the edge computing device corresponding to the edge node, including CPU resources, memory resources, bandwidth resources, etc. The real-time resource occupancy status refers to the current resource usage of the resource pool, that is, the ratio of occupied resources to total resources, including CPU utilization, memory utilization, bandwidth utilization, etc.

[0189] In this embodiment, the scheduling queue refers to the transmission scheduling order of the early warning and tracing links determined based on the sorting result of the number of points at the centralized intersection and the real-time resource occupancy status of the edge nodes. The early warning and tracing links with higher priority are transmitted first. For example, if the number of points after merging the centralized intersections of the three early warning and tracing links are 15, 10, and 8 respectively, the sorting order is 15 > 10 > 8. Combined with the real-time resource occupancy status of the corresponding edge nodes (all under low load), the scheduling queue is determined as follows: Early Warning Link 1 (number of points 15) → Early Warning Link 2 (number of points 10) → Early Warning Link 3 (number of points 8).

[0190] In this embodiment, the target compression ratio C needs to be determined before the data compression process is performed:

[0191] ;

[0192] Where C is the target compression ratio; R1 is the data field redundancy; L is the warning impact domain level; Lmax is the maximum impact domain level; α1, β1, α2, and α3 are dynamic weight coefficients, and α1+β1+α2+α3=1. Real-time resource utilization rate of edge nodes; Prioritize the scheduling queue; To verify the correlation of key domains in data that failed the verification; K1 is the cloud management communication protocol adaptation coefficient; This is a compression efficiency correction factor; The lowest effective compression ratio; This is the highest safe compression ratio.

[0193] It should be noted that α1, β1, α2, and α3 were obtained by scoring the importance of the monitoring data dimensions of metal recycling equipment using the Analytic Hierarchy Process (AHP). For the four dimensions of "dynamic progress impact weight, resource redundancy coupling correlation item, early warning impact domain level, and historical correlation data", five industry technicians were invited to score their proportion in the compression priority. The final weights are: dynamic progress impact weight (α1=0.4), resource redundancy coupling correlation item (β1=0.2), early warning impact domain level (α2=0.3), and historical correlation data (α3=0.1), which meet the requirement that the weight sum is 1.

[0194] Historical related data cached locally on the edge side refers to historical early warning data, solution execution effect data, and similar project processing records stored on the edge side of the cloud platform that are related to the current early warning and tracing link.

[0195] Dual verification refers to combining historical correlation data cached locally on the edge side with industry standard data to perform two verifications on the accuracy, rationality, and compliance of the early warning and tracing link, ensuring that the early warning information is accurate. For example, the first verification (historical data verification): compares the consistency of the current deviation causes in the early warning and tracing link with the historical deviation causes of similar deviations; a deviation rate of ≤5% is considered a pass. The second verification (industry standard verification): checks whether the solution for the early warning and tracing link meets the "deviation handling response time" requirement in the industry standard; if it does, the verification is considered a pass.

[0196] In this embodiment, the number of duplicate and invalid fields in the early warning and tracing data is counted, and the ratio of these counts to the total number of fields is calculated to obtain the data field redundancy.

[0197] The cloud management communication protocol refers to the standardized communication protocol for data transmission between the cloud platform and the management terminal, using the MQTT protocol. Structured data packets refer to standardized data packets formed by integrating verified data with compressed redundant data, organized according to the format required by the cloud management communication protocol. For example, the structured data packet format is: Header: {'Device Identifier':'YJ-001','Timestamp':'2024-05-2014:30:00','Data Length':1024}; Body: {'Deviation Sub-link':'System Debugging','Dynamic Progress Impact Weight':0.35,…}; Tail: {'Checksum':'8F3D2A'}.

[0198] The preset permission range refers to the range of early warning information that can be received and processed by each control terminal in advance. It is determined based on the department to which the control terminal belongs and the user role. That is, when a user logs into the control terminal, the system verifies the user's identity and reads their real-time work role, and associates the corresponding permission configuration.

[0199] Control permission domains refer to the work processes they correspond to. For example, the control permission domains corresponding to the whole machine debugging work process are the production department permission domain and the technical department permission domain.

[0200] The control terminal identifier refers to the unique identification code assigned to each control terminal, which is used to distinguish different control terminals and ensure accurate data packet delivery.

[0201] The warning processing priority label refers to the warning processing priority determined based on the warning impact value. It is divided into three levels: Emergency (red label), Important (yellow label), and General (blue label), used to remind users at the control end to prioritize the processing of high-priority warnings. For example: a warning impact value ≥ 3.0 is Emergency, 1.5 ≤ warning impact value < 3.0 is Important, and a warning impact value < 1.5 is General. A warning impact value of 0.8 corresponds to the warning processing priority label: General (blue).

[0202] The beneficial effects of the above technical solution are as follows: the queue determination unit quantifies the severity of the early warning, the historical retrieval unit and the constraint determination unit accurately locate the concentrated trend of resource occupation, the optimization unit improves the accuracy of boundary constraints, the scheduling queue determination unit ensures transmission efficiency, the data packet acquisition unit ensures data accuracy and compatibility, and the permission reading unit enables accurate push of early warnings, thereby improving the early warning transmission and scheduling efficiency of the cloud platform and significantly enhancing the timeliness and pertinence of early warning processing.

[0203] This invention provides a monitoring method for the manufacturing of metal recycling equipment, such as... Figure 2 As shown, it includes:

[0204] Step 1: Based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, match the corresponding work process baselines and combine them with the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. Analyze the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and construct a dynamically updated contributing device baseline to configure at least one edge computing device for each manufacturing process.

[0205] Step 2: Send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence association weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links.

[0206] Step 3: Based on the cloud platform, receive the first progress information transmitted by all edge computing devices, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project;

[0207] Step 4: Compare the preset progress information of the target manufacturing project with the second progress information to generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue a progress warning to the management and control terminal based on the cloud platform.

[0208] The present invention provides a monitoring device for the manufacturing of metal recycling equipment, comprising: a processor and a storage device, wherein the storage device is used to store instructions, and when the processor executes the instructions, the method is implemented.

[0209] The beneficial effects of the above technical solution are as follows: the equipment determination module enables precise matching between edge computing devices and manufacturing processes; the data analysis module ensures the real-time and accuracy of progress information; the data integration module enables efficient summarization of the overall project progress; and the project monitoring module enables rapid early warning and tracing of progress deviations. The entire process can be completed without manual intervention, including equipment configuration, data collection, progress analysis, and early warning issuance. This effectively solves the problems of unreasonable configuration of traditional monitoring equipment, large deviations in progress assessment, and untimely early warnings, and significantly improves the control accuracy and manufacturing efficiency of the metal recycling equipment manufacturing process.

[0210] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A monitoring system for the manufacture of metal recycling equipment, characterized in that, include: The equipment determination module is used to match the corresponding work process baselines based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, and combine the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. It also analyzes the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and constructs a dynamically updated contributing equipment baseline to configure at least one edge computing device for each manufacturing process. The data analysis module is used to send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence correlation weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links. The data integration module is used to receive the first progress information transmitted by all edge computing devices based on the cloud platform, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project. The project monitoring module is used to compare the preset progress information of the target manufacturing project with the second progress information, generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue progress warnings to the management and control terminal based on the cloud platform.

2. The monitoring system for manufacturing metal recycling equipment according to claim 1, characterized in that, The device determination module includes: The feature set configuration unit is used to match the work step baselines corresponding to the process features of each manufacturing process from the feature-step lookup table, and at the same time, combine the deployment features and real-time load status of each edge computing device to configure an adaptive monitoring feature set for the corresponding edge computing device. The adaptive monitoring feature set dynamically increases or decreases the collection dimensions according to the device load. The matching unit is used to calculate the matching degree between the baseline of each work process and each monitoring feature set using the cosine similarity algorithm, select monitoring feature sets with matching degrees greater than or equal to the preset degree to form the matching subset of the baseline of each work process, and construct the matching baseline of each manufacturing process corresponding to different monitoring feature sets. The extraction unit is used to build an association topology map based on the constructed matching baselines, and extract the link association fluctuation and baseline feature function corresponding to each matching baseline under the preset operating standard. The baseline feature function is obtained by fitting historical operating data through an LSTM neural network. The judgment unit is used to synchronously identify the work process baseline and the double-precision attribute and work attribute of each matching baseline, and to determine whether the number of cross-process call levels of the corresponding matching baseline exceeds the dynamic threshold, wherein the dynamic threshold is adjusted in real time according to the complexity of the manufacturing process. If the number of cases exceeds the limit, the upstream and downstream derived baselines associated with the corresponding matching baseline will be included in the set of baselines to be coordinated in the global analysis, and key coordination nodes will be marked. If the priority is not exceeded, increment the priority of the corresponding work step by 1. The coupling analysis unit is used to perform multi-dimensional coupling analysis on all baselines in the associated topology map and the set of baselines to be coordinated, based on the link-related fluctuations and baseline characteristic functions. Combined with the priority increment operation results, it generates a global dynamic fluctuation trend term for each working link baseline and simultaneously obtains the resource redundancy coupling correlation term for the corresponding working link baseline of each edge computing device. The factor determination unit is used to analyze the contribution factor of the resource redundancy coupling correlation term of each edge computing device to the global dynamic fluctuation trend term of each working link baseline using a gradient boosting regression algorithm, and constructs a contributing device baseline with automatically updated contribution factors every hour. The equipment deployment unit is used to determine the corresponding final equipment baseline by weighted summation and resource constraints based on the global dynamic fluctuation trend item and contributing equipment baseline of each manufacturing process, and to dynamically match and deploy the edge computing devices involved in the final equipment baseline with the corresponding work link baseline of the manufacturing process.

3. The monitoring system for manufacturing metal recycling equipment according to claim 1, characterized in that, The data analysis module includes: The information acquisition unit is used to collect multi-dimensional real-time work data of each working module in the manufacturing process according to the intelligent monitoring components pre-installed in each manufacturing process. The intelligent monitoring components include temperature sensors, current sensors, laser displacement sensors and ultrasonic weld flaw detectors. The information transmission and analysis unit is used to send the real-time monitoring information to the edge computing device corresponding to the manufacturing process and perform data analysis.

4. The monitoring system for manufacturing metal recycling equipment according to claim 3, characterized in that, The information transmission and analysis unit includes: The information output subunit is used to perform edge computing on real-time monitoring information received by the edge computing device and synchronously associate it with the real-time working condition tags of the manufacturing process. The real-time working condition tags are generated by quantification through three dimensions: equipment load, material supply stability and process complexity. The unit outputs initial progress information with working condition attributes and associates it with the baseline of the work process corresponding to the initial progress information. The qualified determination subunit is used to extract the total manufacturing volume of each sub-stage based on the initial progress information, integrate historical work information, real-time quality inspection data under real-time working conditions and benchmarking data of similar projects, and determine the dynamic predicted qualified rate of the corresponding sub-stage. The weight generation subunit is used to generate the dynamic progress impact weight of the sub-process based on the work characteristics of the sub-process, the dynamic prediction pass rate, the real-time quality inspection anomaly rate and the working condition mutation coefficient. It also determines the sampling intensity of the corresponding sub-process by combining the real-time working condition label, and obtains the historical actual progress of the corresponding sub-process at the historical working moment to determine whether it is qualified or not. The information adjustment subunit is used to adjust the initial progress information based on the results of the qualification or failure of each sub-step, historical correction coefficients, and real-time working condition fluctuations to obtain the first progress information of the baseline of the corresponding work step.

5. The monitoring system for manufacturing metal recycling equipment according to claim 1, characterized in that, The data integration module includes: The information sorting unit is used to sort all the first progress information according to the dynamic dependency relationship based on the dependency relationship matrix of the corresponding work process baseline of the cloud platform and each edge computing device to obtain the integrated progress information of the target manufacturing project. The information adjustment unit is used to adjust the integrated progress information according to the real-time importance weight of the baseline of each work link to obtain the second progress information of the target manufacturing project.

6. The monitoring system for manufacturing metal recycling equipment according to claim 4, characterized in that, The project monitoring module includes: The deviation determination unit is used to calculate the progress deviation value between the preset progress information and the second progress information of the target manufacturing project, associate the dynamic progress influence weight of each work link with the corresponding sub-link, determine the influence level of the progress deviation, and determine the deviation type of the progress deviation by combining the real-time operating condition fluctuation coefficient, real-time quality inspection anomaly rate and cross-link deviation transmission coefficient of the sub-link. The influence level includes the core influence level, the important influence level and the auxiliary influence level. The early warning generation unit is used to associate and integrate the sub-links corresponding to the progress deviation, the dynamic progress influence weight, the deviation cause information and the intelligent solution to generate an early warning traceability link.

7. The monitoring system for manufacturing metal recycling equipment according to claim 1, characterized in that, The cloud platform includes: The value determination unit is used to determine the warning impact value based on the warning impact range, the criticality of the warning sub-link and the deviation spread speed of the work link baseline associated with the warning tracing link. The historical retrieval unit is used to retrieve from the historical database the edge node to which the value consistent with the warning impact value belongs and the historical occupancy status of the edge node at the trigger time to obtain a three-dimensional array, and to place the three-dimensional array in a three-dimensional coordinate system to obtain a three-dimensional discrete map. Lock the concentrated intersection points in the three-dimensional discrete graph, wherein there is at least one concentrated intersection point; The constraint determination unit is used to obtain the frequency of each surrounding point turning to the central intersection point to obtain the surrounding radiation map, and to perform fitting analysis on the surrounding radiation map, and use the fitting function with the most surrounding points as the initial boundary constraint. An optimization unit is used to extract the constraint parameter vector of the initial boundary constraint of each set intersection point and construct a constraint matrix to obtain the constraint feature vector of the constraint matrix, and optimize each constraint parameter vector based on the constraint feature vector to obtain a new boundary constraint. The queue determination unit is used to merge the surrounding points of each centralized intersection point that satisfy the new boundary constraints into the corresponding centralized intersection point, count the number of points in the corresponding centralized intersection point after merging, sort all the point counts by size, lock the real-time resource occupancy status of the resource pool of the edge node corresponding to each centralized intersection point, and determine the scheduling queue of the corresponding early warning and tracing link. The data packet acquisition unit is used to perform dual verification of the early warning and tracing link by combining the historical associated data cached locally on the edge side of the cloud platform and industry standard data, and to perform compression processing on the data that fails the verification in the early warning and tracing link by combining the data field redundancy of the scheduling queue. The compressed data and the verified data are converted into a structured data packet adapted to the cloud management and control communication protocol. The permission reading unit is used to read the preset permission range and real-time work role of each control terminal, match the control permission domain corresponding to the associated work process, filter the control terminals with receiving permissions, embed the control terminal identifier and early warning processing priority tag into the structured data packet, and then push it to the corresponding control terminal based on the cloud platform.

8. A monitoring method for the manufacturing of metal recycling equipment, characterized in that, include: Step 1: Based on the process characteristics of multiple manufacturing processes involved in the target manufacturing project of metal recycling equipment, match the corresponding work process baselines and combine them with the monitoring feature set of each edge computing device to generate a global dynamic fluctuation trend item for each work process baseline. Analyze the contribution of the resource redundancy coupling correlation item of the edge computing device to the global dynamic fluctuation trend item, and construct a dynamically updated contributing device baseline to configure at least one edge computing device for each manufacturing process. Step 2: Send the real-time monitoring information of each manufacturing process to the corresponding edge computing device, and perform data analysis on the real-time monitoring information based on the time sequence association weight of the sub-links to obtain the first progress information of the baseline of each work link under the corresponding manufacturing process. Each manufacturing process contains several work link baselines, and each work link baseline contains several sub-links. Step 3: Based on the cloud platform, receive the first progress information transmitted by all edge computing devices, dynamically sort them according to the manufacturing process dependencies, and adjust them in combination with the real-time importance weight of the work process baseline to obtain the second progress information of the target manufacturing project; Step 4: Compare the preset progress information of the target manufacturing project with the second progress information to generate an early warning and traceability link that includes deviation sub-links, dynamic progress impact weights, deviation causes and intelligent solutions, and issue a progress warning to the management and control terminal based on the cloud platform.

9. A monitoring device for a metal recycling equipment manufacturing process, comprising: A processor and a storage device, the storage device being used to store instructions that, when executed by the processor, implement the method according to claim 8.

Images