Manufacturing process monitoring and predictive analysis method based on data driving

Abstract

The invention relates to the technical field of intelligent operation and maintenance, in particular to a manufacturing process monitoring and predictive analysis method based on data driving. Comprising the following steps: S1, acquiring sensor time sequence data of processing equipment on a production line processing site through embedded sensing equipment; s2, obtaining a quality inspection report of the processed product, wherein the quality inspection report comprises a defective rate and a defective degree of the product; s3, when the defective rate of the product is higher than a preset defective rate threshold value, processing equipment for circulation of the defective product is determined, the processing time of the defective product in each processing equipment is determined according to the production information, sensor time sequence data of a corresponding timestamp range is called, a fault analysis packet is generated, and the fault analysis packet is sent to a server; the fault analysis packet comprises a defective product degree of a defective product, circulation equipment and sensor time sequence data during circulation; and S4, according to the fault analysis packets of the plurality of defective processed products, calculating the correlation between the defective degree and the sensor time sequence data of each processing device to generate an operation and maintenance decision.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology, and specifically to a data-driven method for monitoring and predictive analysis of manufacturing processes. Background Technology

[0002] The Industrial Internet is the result of the integration of global industrial systems with advanced computing, analytics, sensing technologies, and internet connectivity. The essence of the Industrial Internet is to tightly connect and integrate equipment, production lines, factories, suppliers, products, and customers through an open, global industrial-grade network platform, so as to efficiently share various elements and resources in the industrial economy. This will reduce costs and increase efficiency through automated and intelligent production methods, help the manufacturing industry extend its industrial chain, and promote the transformation and development of the manufacturing industry.

[0003] Industrial manufacturing system equipment refers to the mechanical equipment and systems used in the manufacturing process for processing, assembling, testing, and packaging products. In modern industrial production, industrial manufacturing system equipment often achieves efficient, precise, and automated operation through integrated control systems, sensors, and software. The operation and maintenance management of industrial manufacturing system equipment is crucial to ensuring smooth production processes, improving equipment efficiency, and extending equipment lifespan.

[0004] However, in practice, operations management faces various challenges and problems. Currently, most SMEs' operations and maintenance systems primarily rely on stand-alone data collection and display. Some data is collected through manual periodic inspections, manual meter reading, and statistical summarization; other data is generated by individual devices or systems, but data between these devices and systems requires manual import and export. These situations cannot guarantee data accuracy and real-time performance. In traditional operations and maintenance work, personnel generally passively wait before taking action, lacking proactive mechanisms and fault handling management processes, resulting in poor timeliness in fault handling. Summary of the Invention

[0005] The technical problem solved by this invention is to provide a data-driven manufacturing process monitoring and predictive analysis method that can improve the timeliness of equipment fault detection.

[0006] The basic solution provided by this invention is a data-driven method for monitoring and predictive analysis of manufacturing processes, characterized by the following steps: S1. Collect sensor timing data of processing equipment on the production line through embedded sensing devices, and upload the sensor timing data to the cloud platform. S2. Obtain the quality inspection report of the processed products, the quality inspection report including the defect rate and the degree of defect of the products; S3. When the defect rate of a product is higher than the preset defect rate threshold, determine the processing equipment to which the defective product is transferred, and determine the processing time of the defective product in each processing equipment according to the production information. Retrieve the sensor time series data of the corresponding timestamp range and generate a fault analysis package. The fault analysis package includes the defective product's defective degree, the transfer equipment, and the sensor time series data during the transfer. S4. Based on the fault analysis packages of several defective products, calculate the correlation between the degree of defect and the sensor time-series data of each processing equipment, analyze whether the defective product is due to equipment failure or human operation, and generate operation and maintenance decisions when it is due to equipment failure.

[0007] Furthermore, S3 includes the following steps: S31. The cloud platform monitors the defect rate of the production line in real time. When the defect rate exceeds the preset defect rate threshold, fault analysis is triggered. S32. Obtain the unique identification code of the defective product through the MES manufacturing execution system, and trace the processing path of the defective product based on the unique identification code; S33. Obtain the processing timestamp of the defective product on each machine according to the processing path, including the entry time. and departure time ; S34. Retrieve the corresponding devices from the cloud platform. The system collects time-series data from all sensors, including temperature, vibration, current, and pressure, and then links this data with quality inspection reports to generate a fault analysis package.

[0008] Furthermore, S3 also includes the following steps: S351. Quantify the measurement values ​​of defective products in various types of defects according to preset standards. Defect types include functional defects, appearance defects and safety defects. S352. Based on the measured values ​​of each type of defect and the preset tolerance threshold, generate the weighted defect index of the defective product. :

[0009] Where n is the number of defect types contained in the defective product. This represents the actual measured value of the i-th defect type. This represents the allowable tolerance threshold for the i-th defect type. This represents the sensitivity coefficient for the i-th defect type. represents the weighting coefficient for the i-th defect type, and S represents the defect diffusion coefficient; S353, Weighted Defect Index It correlates with the time-series data of the sensor.

[0010] Furthermore, S4 includes the following steps: S41. Based on the sensor time-series data of the defective products on each processing equipment, obtain the statistical characteristics x of the defective products during each processing period on each equipment; S42. Calculate the correlation coefficient based on the weighted defect index (WDI) and statistical characteristic x of the defective products. :

[0011] Where n is the number of defective product samples. This represents the value of feature k for the i-th defective sample in device j. This represents the mean of feature k across all samples. This represents the mean of the Defect Index (WDI) for all samples. S43. When at least one device j exists... If the equipment is faulty, the defective product is determined to be mainly due to equipment malfunction; otherwise, the defective product is determined to be mainly due to human error.

[0012] Furthermore, it also includes the following steps: S5. When the main cause of the fault is equipment malfunction, generate dynamic maintenance work orders and preventive maintenance strategies based on the output list of faulty equipment and the characteristics of highly relevant sensors. S51. When there are faulty devices, locate the number of faulty devices and their node positions on the production line. S52. Based on the correlation between each faulty device and the Weighted Defect Index (WDI) Determine the repair priority index and correlation of each faulty device. The higher the value, the higher the maintenance priority index. S53. Based on the maintenance priority index, sort the faulty equipment, and starting from the equipment with the highest priority, evaluate the expected reduction of the weighted defect index (WDI) after repairing each piece of equipment in turn, accumulate the expected reduction, and when the accumulated reduction is sufficient to reduce the weighted defect index below the preset threshold, stop the selection and generate a set of priority maintenance equipment. S55. Generate dynamic maintenance work orders, and only perform immediate maintenance on equipment in the priority maintenance equipment set, while the remaining equipment is included in the planned maintenance list.

[0013] Furthermore, it also includes the following steps: S6. When the main cause of the fault is human error, generate a precise training plan and optimize the operating procedures.

[0014] The principles and advantages of this invention are as follows: This invention uses embedded sensing devices to collect multi-dimensional sensor time-series data from production line processing equipment in real time and uploads it to a cloud platform. Simultaneously, it obtains a quality inspection report containing the defect rate and defect severity. When the defect rate exceeds a preset threshold, a fault analysis mechanism is automatically triggered. The MES (Manufacturing Execution System) is used to trace the unique identifier of defective products to accurately locate their processing path and the processing timestamp of each piece of equipment. Full-dimensional sensor data within the corresponding time interval is retrieved to generate a fault analysis package deeply bound to the defect severity. Furthermore, by quantifying the measured values ​​of functional defects, appearance defects, and safety defects of defective products, and combining tolerance thresholds, sensitivity coefficients, and weights... A weighted defect index model is constructed using coefficients and defect diffusion coefficients. This index is then correlated with time-series data from equipment sensors using Pearson correlation analysis at the statistical feature level. A correlation coefficient threshold is used to distinguish between inherent equipment faults and human error as the root cause. Based on the diagnostic results, targeted maintenance strategies are dynamically generated. If the fault is equipment-related, a work order containing priority, maintenance plan, and preventative maintenance measures is automatically generated by matching the maintenance knowledge base. If the fault is human error, the problematic process and operator are identified. Training and assessment programs, including standard operation video tutorials, error replay analysis, and a dual-confirmation mechanism, are developed for the problematic process and operator. Finally, a closed-loop optimization engine is formed by continuously collecting historical maintenance performance data, adaptively adjusting the correlation coefficient threshold and maintenance cycle parameters. This achieves full-process automation of fault warning, root cause location, proactive maintenance, and strategy iteration, effectively improving equipment reliability, product quality stability, and production process continuity. It completely changes the passive situation in traditional maintenance, characterized by data silos, low efficiency of manual traceability, strong subjectivity in fault attribution, and delayed maintenance response lacking continuous optimization. This provides industrial manufacturing systems with a highly integrated, self-evolving, and precisely decision-making intelligent maintenance paradigm. The core principle of this method lies in establishing a spatiotemporal correlation channel between real-time equipment monitoring data and product quality data. It establishes a quantifiable causal relationship between fluctuations in equipment status in the physical world and quantitative defect indicators in the virtual world through mathematical modeling. It uses statistical analysis methods to replace manual experience-based judgment and enables continuous evolution of operation and maintenance strategies through a closed-loop feedback mechanism. Essentially, it constructs an autonomous industrial intelligent system driven by data, with algorithms as the core of decision-making and execution as the final step. This system achieves predictive, proactive, and adaptive equipment health management without human intervention, thereby fundamentally reconstructing the operational mode of industrial operation and maintenance and providing solid technical support for the digital transformation and upgrading of the manufacturing industry.

[0015] Simultaneously, the correlation analysis results between the weighted defect index and equipment sensor data are directly applied to the maintenance decision-making process. First, based on locating the faulty equipment, a quantitative evaluation model for the maintenance priority index is introduced. Its core basis is the correlation coefficient between the equipment and the weighted defect index; a higher correlation indicates a greater impact of the equipment on product quality defects, thus increasing its urgency for maintenance and consequently raising its priority index. Subsequently, a dynamic determination algorithm based on the expected reduction in maintenance scope is used. The system predicts the reduction effect of each piece of equipment on the overall weighted defect index after maintenance, in descending order of maintenance priority, and accumulates these expected reductions in real time. When the accumulated value reaches a level sufficient to restore the system's defect level below a safe threshold, the selection of subsequent equipment is automatically stopped, forming the final set of equipment requiring immediate maintenance. The essence of this principle is to transform maintenance decision-making from a traditional passive model based on fixed cycles or manual experience into a dynamic optimization process that uses real-time data correlation as input and aims to eliminate defect impact. Its underlying logic is to accurately quantify the contribution of each faulty piece of equipment to quality problems and predict the chain reaction of improvement effects from maintenance actions, thereby achieving optimal allocation of maintenance resources while ensuring production continuity.

[0016] Compared to existing technologies, this solution significantly improves the intelligence and precision of the operation and maintenance process. Traditional operation and maintenance often involves indiscriminate shutdowns for repair of all identified faulty equipment, easily leading to excessive production downtime and wasted resources. This solution, through prioritization and expected effect assessment, accurately identifies the core equipment most critical to current product quality issues and performs immediate repairs only on these devices, while the remaining equipment is included in planned maintenance schedules. This fundamentally reduces unnecessary emergency downtime and optimizes the efficiency of maintenance manpower and material resources. Furthermore, this solution possesses strong adaptability, dynamically adjusting maintenance strategies based on real-time production quality data. This ensures that operation and maintenance decisions are always aligned with the most pressing production quality goals, effectively accelerating the closed-loop resolution of quality issues and reducing the overall costs associated with over-maintenance or delays. Ultimately, this improves the overall reliability and operational efficiency of the manufacturing system. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of an embodiment of a data-driven manufacturing process monitoring and predictive analysis method according to the present invention. Detailed Implementation

[0018] The following detailed description illustrates the specific implementation method: The basic implementation examples are as follows: Figure 1 As shown: A data-driven manufacturing process monitoring and predictive analysis method, characterized by the following steps: S1. Collect sensor timing data of processing equipment on the production line through embedded sensing devices, and upload the sensor timing data to the cloud platform. S2. Obtain the quality inspection report of the processed products, the quality inspection report including the defect rate and the degree of defect of the products; S3. When the defect rate of a product is higher than the preset defect rate threshold, determine the processing equipment to which the defective product is transferred, and determine the processing time of the defective product in each processing equipment according to the production information. Retrieve the sensor time series data of the corresponding timestamp range and generate a fault analysis package. The fault analysis package includes the defective product's defective degree, the transfer equipment, and the sensor time series data during the transfer. S4. Based on the fault analysis packages of several defective products, calculate the correlation between the degree of defect and the sensor time-series data of each processing equipment, analyze whether the defective product is due to equipment failure or human operation, and generate operation and maintenance decisions when it is due to equipment failure.

[0019] S3 includes the following steps: S31. The cloud platform monitors the defect rate of the production line in real time. When the defect rate exceeds the preset defect rate threshold, fault analysis is triggered. S32. Obtain the unique identification code of the defective product through the MES manufacturing execution system, and trace the processing path of the defective product based on the unique identification code; S33. Obtain the processing timestamp of the defective product on each machine according to the processing path, including the entry time. and departure time ; S34. Retrieve the corresponding devices from the cloud platform. The system collects time-series data from all sensors, including temperature, vibration, current, and pressure, and then links this data with quality inspection reports to generate a fault analysis package.

[0020] S3 further includes the following steps: S351. Quantify the measurement values ​​of defective products in various types of defects according to preset standards. Defect types include functional defects, appearance defects and safety defects. S352. Based on the measured values ​​of each type of defect and the preset tolerance threshold, generate the weighted defect index of the defective product. :

[0021] Where n is the number of defect types contained in the defective product. This represents the actual measured value of the i-th defect type. This represents the allowable tolerance threshold for the i-th defect type. This represents the sensitivity coefficient for the i-th defect type. represents the weighting coefficient for the i-th defect type, and S represents the defect diffusion coefficient; S353, Weighted Defect Index It correlates with the time-series data of the sensor.

[0022] S4 includes the following steps: S41. Based on the sensor time-series data of the defective products on each processing equipment, obtain the statistical characteristics x of the defective products during each processing period on each equipment; S42. Calculate the correlation coefficient based on the weighted defect index (WDI) and statistical characteristic x of the defective products. :

[0023] Where n is the number of defective product samples. This represents the value of feature k for the i-th defective sample in device j. This represents the mean of feature k across all samples. This represents the mean of the Defect Index (WDI) for all samples. S43. When at least one device j exists... If the equipment is faulty, the defective product is determined to be mainly due to equipment malfunction; otherwise, the defective product is determined to be mainly due to human error.

[0024] It also includes the following steps: S5. When the main cause of the fault is equipment malfunction, generate dynamic maintenance work orders and preventive maintenance strategies based on the output list of faulty equipment and the characteristics of highly relevant sensors. S51. When there are faulty devices, locate the number of faulty devices and their node positions on the production line. S52. Based on the correlation between each faulty device and the Weighted Defect Index (WDI) Determine the repair priority index and correlation of each faulty device. The higher the value, the higher the maintenance priority index. S53. Based on the maintenance priority index, sort the faulty equipment, and starting from the equipment with the highest priority, evaluate the expected reduction of the weighted defect index (WDI) after repairing each piece of equipment in turn, accumulate the expected reduction, and when the accumulated reduction is sufficient to reduce the weighted defect index below the preset threshold, stop the selection and generate a set of priority maintenance equipment. S55. Generate dynamic maintenance work orders, and only perform immediate maintenance on equipment in the priority maintenance equipment set, while the remaining equipment is included in the planned maintenance list.

[0025] It also includes the following steps: S6. When the main cause of the fault is human error, generate a precise training plan and optimize the operating procedures.

[0026] This invention collects multi-dimensional sensor time-series data from production line processing equipment in real time using embedded sensing devices and uploads it to a cloud platform. Simultaneously, it obtains a quality inspection report containing the defect rate and degree of defect. When the defect rate exceeds a preset threshold, a fault analysis mechanism is automatically triggered. The MES (Manufacturing Execution System) is used to trace the unique identifier of defective products to accurately locate their processing path and the processing timestamp of each piece of equipment. Full-dimensional sensor data within the corresponding time interval is retrieved to generate a fault analysis package deeply bound to the degree of defect. Furthermore, by quantifying the measured values ​​of functional defects, appearance defects, and safety defects of defective products, a weighted defect index model is constructed by combining tolerance thresholds, sensitivity coefficients, weighting coefficients, and defect diffusion coefficients. This index is then used to perform Pearson correlation calculations with the equipment sensor time-series data at the statistical characteristic level. The correlation coefficient threshold is used as the criterion to distinguish between inherent equipment faults and the root cause of human operational errors. Based on diagnostic results, targeted operation and maintenance strategies are dynamically generated. If the problem is equipment failure, a work order containing priority, repair plan, and preventive maintenance measures is automatically generated by matching the maintenance knowledge base. If the problem is human error, the problematic process and operator are accurately located, and training and assessment programs such as standard operation video tutorials, error operation playback analysis, and dual-person confirmation mechanisms are developed. Finally, by continuously collecting historical operation and maintenance effect data, a closed-loop optimization engine is formed, which adaptively adjusts the correlation coefficient threshold and maintenance cycle parameters to achieve full-process automation of fault warning, root cause location, proactive maintenance, and strategy iteration. This effectively improves equipment reliability, product quality stability, and production process continuity, and completely changes the passive situation of traditional operation and maintenance, such as data silos, low efficiency of manual traceability, strong subjectivity of fault attribution, and delayed maintenance response without continuous optimization. It provides industrial manufacturing systems with a highly integrated, self-evolving, and precise decision-making intelligent operation and maintenance paradigm. The core principle of this method lies in establishing a spatiotemporal correlation channel between real-time equipment monitoring data and product quality data. It establishes a quantifiable causal relationship between fluctuations in equipment status in the physical world and quantitative defect indicators in the virtual world through mathematical modeling. It uses statistical analysis methods to replace manual experience-based judgment and enables continuous evolution of operation and maintenance strategies through a closed-loop feedback mechanism. Essentially, it constructs an autonomous industrial intelligent system driven by data, with algorithms as the core of decision-making and execution as the final step. This system achieves predictive, proactive, and adaptive equipment health management without human intervention, thereby fundamentally reconstructing the operational mode of industrial operation and maintenance and providing solid technical support for the digital transformation and upgrading of the manufacturing industry.

[0027] Simultaneously, the correlation analysis results between the weighted defect index and equipment sensor data are directly applied to the maintenance decision-making process. First, based on locating the faulty equipment, a quantitative evaluation model for the maintenance priority index is introduced. Its core basis is the correlation coefficient between the equipment and the weighted defect index; a higher correlation indicates a greater impact of the equipment on product quality defects, thus increasing its urgency for maintenance and consequently raising its priority index. Subsequently, a dynamic determination algorithm based on the expected reduction in maintenance scope is used. The system predicts the reduction effect of each piece of equipment on the overall weighted defect index after maintenance, in descending order of maintenance priority, and accumulates these expected reductions in real time. When the accumulated value reaches a level sufficient to restore the system's defect level below a safe threshold, the selection of subsequent equipment is automatically stopped, forming the final set of equipment requiring immediate maintenance. The essence of this principle is to transform maintenance decision-making from a traditional passive model based on fixed cycles or manual experience into a dynamic optimization process that uses real-time data correlation as input and aims to eliminate defect impact. Its underlying logic is to accurately quantify the contribution of each faulty piece of equipment to quality problems and predict the chain reaction of improvement effects from maintenance actions, thereby achieving optimal allocation of maintenance resources while ensuring production continuity.

[0028] Compared to existing technologies, this solution significantly improves the intelligence and precision of the operation and maintenance process. Traditional operation and maintenance often involves indiscriminate shutdowns for repair of all identified faulty equipment, easily leading to excessive production downtime and wasted resources. This solution, through prioritization and expected effect assessment, accurately identifies the core equipment most critical to current product quality issues and performs immediate repairs only on these devices, while the remaining equipment is included in planned maintenance schedules. This fundamentally reduces unnecessary emergency downtime and optimizes the efficiency of maintenance manpower and material resources. Furthermore, this solution possesses strong adaptability, dynamically adjusting maintenance strategies based on real-time production quality data. This ensures that operation and maintenance decisions are always aligned with the most pressing production quality goals, effectively accelerating the closed-loop resolution of quality issues and reducing the overall costs associated with over-maintenance or delays. Ultimately, this improves the overall reliability and operational efficiency of the manufacturing system.

[0029] The above are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A data-driven manufacturing process monitoring and predictive analysis method, characterized in that: Includes the following steps: S1. Collect sensor timing data of processing equipment on the production line through embedded sensing devices, and upload the sensor timing data to the cloud platform. S2. Obtain the quality inspection report of the processed products, the quality inspection report including the defect rate and the degree of defect of the products; S3. When the defect rate of a product is higher than the preset defect rate threshold, determine the processing equipment to which the defective product is transferred, and determine the processing time of the defective product in each processing equipment according to the production information. Retrieve the sensor time series data of the corresponding timestamp range and generate a fault analysis package. The fault analysis package includes the defective product's defective degree, the transfer equipment, and the sensor time series data during the transfer. S4. Based on the fault analysis packages of several defective products, calculate the correlation between the degree of defect and the sensor time series data of each processing equipment, analyze whether the defective product is due to equipment failure or human operation, and generate operation and maintenance decisions when it is due to equipment failure. S3 further includes the following steps: S351. Quantify the measurement values ​​of defective products in various types of defects according to preset standards. Defect types include functional defects, appearance defects and safety defects. S352. Based on the measured values ​​of each type of defect and the preset tolerance threshold, generate the weighted defect index of the defective product. : Where n is the number of defect types contained in the defective product. This represents the actual measured value of the i-th defect type. This represents the allowable tolerance threshold for the i-th defect type. This represents the sensitivity coefficient for the i-th defect type. represents the weighting coefficient for the i-th defect type, and S represents the defect diffusion coefficient; S353, Weighted Defect Index It correlates with the time-series data of the sensor. S4 includes the following steps: S41. Based on the sensor time-series data of the defective products on each processing equipment, obtain the statistical characteristics x of the defective products during each processing period on each equipment; S42. Calculate the correlation coefficient based on the weighted defect index (WDI) and statistical characteristic x of the defective products. : Where n is the number of defective product samples. This represents the value of feature k for the i-th defective sample in device j. This represents the mean of feature k across all samples. This represents the mean of the Defect Index (WDI) for all samples. S43. When at least one device j exists... hour, If the preset correlation threshold is used, the device is determined to be faulty, and the main cause of the defective product is determined to be the device malfunction; otherwise, the main cause of the defective product is determined to be human error. It also includes the following steps: S5. When the main cause of the fault is equipment malfunction, a dynamic maintenance work order and a preventive maintenance strategy are generated based on the output list of faulty equipment and the characteristics of highly relevant sensors. S5 includes the following steps: S51. When there are faulty devices, locate the number of faulty devices and their node positions on the production line. S52. Based on the correlation between each faulty device and the Weighted Defect Index (WDI) Determine the repair priority index and correlation of each faulty device. The higher the value, the higher the maintenance priority index. S53. Based on the maintenance priority index, sort the faulty equipment, and starting from the equipment with the highest priority, evaluate the expected reduction of the weighted defect index (WDI) after repairing each piece of equipment in turn, accumulate the expected reduction, and when the accumulated reduction is sufficient to reduce the weighted defect index below the preset threshold, stop the selection and generate a set of priority maintenance equipment. S55. Generate dynamic maintenance work orders, and only perform immediate maintenance on equipment in the priority maintenance equipment set, while the remaining equipment is included in the planned maintenance list.

2. The data-driven manufacturing process monitoring and predictive analysis method according to claim 1, characterized in that: S3 includes the following steps: S31. The cloud platform monitors the defect rate of the production line in real time. When the defect rate exceeds the preset defect rate threshold, fault analysis is triggered. S32. Obtain the unique identification code of the defective product through the MES manufacturing execution system, and trace the processing path of the defective product based on the unique identification code; S33. Obtain the processing timestamp of the defective product on each machine according to the processing path, including the entry time. and departure time ; S34. Retrieve the corresponding devices from the cloud platform. The system collects time-series data from all sensors, including temperature, vibration, current, and pressure, and then links this data with quality inspection reports to generate a fault analysis package.

3. The data-driven manufacturing process monitoring and predictive analysis method according to claim 1, characterized in that: It also includes the following steps: S6. When the main cause of the fault is human error, generate a precise training plan and optimize the operating procedures.

Images