A method and system for intelligent production line control of brushless motors for automobiles

Abstract

This invention provides an intelligent production line management method and system for brushless motors in automobiles, relating to the field of automotive technology. The method includes: performing comprehensive defect detection to obtain the motor defect distribution; performing process failure mode backtracking to locate P production line equipment at P production process nodes; performing cross-process parameter coupling analysis and reorganizing into M groups of parallel monitoring devices; configuring M edge detection nodes, which communicate with the production line management platform via an IoT gateway protocol and are configured with M multimodal monitoring tasks; after calibration, the M edge detection nodes dynamically allocate computing power to schedule the M multimodal monitoring tasks, executing dynamic calibration triggers driven by deviation diagnosis. This invention solves the technical problem that existing technologies for electric vehicles and other motor vehicles typically rely on manual inspection and experience-based judgment for equipment calibration, lacking automated real-time monitoring and calibration mechanisms, thus affecting product quality.

Description

Technical Field

[0001] This invention relates to the field of motor vehicle technology, specifically to an intelligent production line management method and system for automotive brushless motors. Background Technology

[0002] As a core component of electric vehicles and other motor vehicles, the quality of automotive brushless motors directly affects the performance and safety of the entire vehicle. The production process of automotive brushless motors involves multiple complex steps, each requiring precise operation and strict quality control. However, existing automotive brushless motor production processes often face several challenges. Current technologies typically rely on manual inspection and experience-based judgment for equipment calibration, lacking automated real-time monitoring and calibration mechanisms. Manual operation is often lagging, unable to quickly respond to changes in equipment status. When equipment malfunctions or deviates from its intended path, manual inspection and adjustment take considerable time, failing to promptly identify potential problems. This results in delayed equipment calibration, leading to reduced production efficiency, increased equipment wear and tear, and even product defects, ultimately affecting the quality of the final product. Summary of the Invention

[0003] This application provides an intelligent production line control method and system for automotive brushless motors, aiming to solve the technical problem that the production of electric vehicles and other motor vehicles usually relies on manual inspection and experience judgment for equipment calibration, lacking an automated real-time monitoring and calibration mechanism, thus affecting product quality.

[0004] The first aspect disclosed in this application provides an intelligent production line management method for automotive brushless motors. The method includes: performing comprehensive defect detection on trial-produced automotive brushless motors to obtain motor defect distribution; performing process failure mode backtracking based on the motor defect distribution to locate P production line equipment at P production process nodes; performing cross-process parameter coupling analysis on the P production process nodes to reorganize the P production line equipment into M groups of parallel monitoring devices; configuring M edge detection nodes on the M groups of parallel monitoring devices, wherein the M edge detection nodes are connected to the production line management platform via an IoT gateway protocol, and the M edge detection nodes are configured with M multimodal monitoring tasks; after calibrating the P production line equipment, during the production process of the automotive brushless motor, the M edge detection nodes dynamically allocate and schedule the M multimodal monitoring tasks to perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring devices.

[0005] The second aspect of this application discloses an intelligent production line management system for automotive brushless motors. This system is used in the aforementioned intelligent production line management method for automotive brushless motors. The system includes: a comprehensive defect detection module for performing comprehensive defect detection on trial-produced automotive brushless motors to obtain the motor defect distribution; a failure mode backtracking module for performing process failure mode backtracking based on the motor defect distribution to locate P production line equipment at P production process nodes; and a parameter coupling analysis module for performing cross-process parameter coupling analysis on the P production process nodes to reorganize the P production line equipment. The system comprises M groups of parallel monitoring devices; a detection node configuration module for configuring M edge detection nodes for the M groups of parallel monitoring devices, wherein the M edge detection nodes are connected to the production line management platform via an IoT gateway protocol, and the M edge detection nodes are configured with M multimodal monitoring tasks; and a dynamic calibration trigger module for performing deviation diagnosis-driven dynamic calibration triggering on the M groups of parallel monitoring devices during the production process of automotive brushless motors after the P production line devices have undergone calibration status calibration.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects:

[0007] Comprehensive defect detection identifies various defects in brushless motors during trial production and reveals their distribution, providing a data foundation for subsequent analysis and enabling precise tracking and location of defects during production. By retrospectively analyzing process failure modes based on the defect distribution, the corresponding production process nodes can be traced, and the relevant production line equipment can be accurately located. This provides a clear path for quality control on the production line, helping to identify potential failure causes at the source and implement targeted optimization and adjustments. Cross-process parameter coupling analysis analyzes the correlation between various production process nodes, effectively identifying the mutual influence between different production process nodes. This analytical method can help... This helps to understand the complexity of the production process, identify potential problems between processes, and thus provide a basis for equipment optimization. By configuring M edge detection nodes, it can ensure that the real-time data of each group of parallel monitoring equipment can be fully monitored. The communication between the edge detection nodes and the production line management platform enables the monitoring system to respond to changes in equipment status in a timely manner. The configuration of multimodal monitoring tasks ensures comprehensive monitoring of various key parameters of the equipment. When abnormal equipment status occurs, it can quickly identify and trigger necessary calibration operations, thereby ensuring the accuracy and stability of equipment in the production process, effectively reducing the risk of motor failures in electric vehicles and other motor vehicles during use, and thus improving the overall stability and market competitiveness of electric vehicles and other motor vehicles.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of a smart production line management method for a brushless motor in an automotive application, provided as an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of an intelligent production line control system for an automotive brushless motor, provided as an embodiment of this application.

[0011] Figure labeling: 10 Comprehensive defect detection module, 20 Failure mode backtracking module, 30 Parameter coupling analysis module, 40 Detection node configuration module, 50 Dynamic calibration triggering module. Detailed Implementation

[0012] This application provides an intelligent production line management method and system for automotive brushless motors, which solves the technical problem that the production of electric vehicles and other motor vehicles usually relies on manual inspection and experience judgment for equipment calibration, lacking an automated real-time monitoring and calibration mechanism, thus affecting product quality.

[0013] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0014] Example 1, as Figure 1 As shown in the figure, this application provides an intelligent production line management method for automotive brushless motors, the method comprising:

[0015] By conducting comprehensive defect testing on the brushless motors of trial-produced automobiles, the distribution of motor defects was obtained.

[0016] On the trial production line for brushless motors, a series of inspection technologies, such as machine vision, vibration analysis, and temperature monitoring, are used to inspect each completed brushless motor for defects. For each detected defect, the type of defect is identified, such as motor winding defects or bearing damage, as well as the spatial location of the defect, such as a specific area of ​​the motor. Each defect is recorded, including its type, spatial location, and frequency of occurrence, ultimately forming a distribution of motor defects.

[0017] Based on the motor defect distribution, process failure modes are traced back to locate P production line equipment at P production process nodes.

[0018] Process failure mode backtracking refers to tracing back to process failure modes in the production process based on the distribution of motor defects. Specifically, based on the distribution of motor defects and the preset process flow, reverse reasoning is performed to analyze which production processes have process problems and establish a correlation with motor defects. Through process failure mode backtracking, it is determined which production processes are the root causes of defects, and each production process node is marked as a potential failure point. Furthermore, the relevant production line equipment is located, which may be the source of the defects. Finally, P production process nodes and P corresponding production line equipment are located.

[0019] Cross-process parameter coupling analysis is performed on the P production process nodes, and the P production line equipment is reorganized into M groups of parallel monitoring equipment.

[0020] Different production process nodes exhibit certain coupling relationships, meaning that changes in the parameters of one production process node can affect the performance of subsequent production process nodes. By analyzing the mutual influences between various production process nodes, such as the correlation between parameters like temperature, pressure, speed, and vibration, we can identify which production process nodes are tightly coupled and which have potential process problems. The cross-process parameter coupling analysis process includes coupling strength calculation and coupling topology graph construction. Based on the coupling analysis results, the P production line devices of the P production process nodes are grouped into M groups of parallel monitoring devices. Each group of parallel monitoring devices has a high degree of correlation and can monitor and analyze in parallel.

[0021] M edge detection nodes are configured for the M groups of parallel monitoring devices. The M edge detection nodes are connected to the production line management platform via an Internet of Things gateway protocol. The M edge detection nodes are configured with M multimodal monitoring tasks.

[0022] To improve monitoring efficiency and reduce data transmission latency, each group of parallel monitoring devices is configured with an edge detection node. These edge detection nodes are located locally on or near the edge of the production line equipment and are responsible for collecting and processing production data in real time. All M edge detection nodes communicate with the production line management platform via an IoT gateway protocol. The IoT gateway is responsible for transmitting the data collected by the edge detection nodes to the production line management platform, ensuring real-time data updates and feedback.

[0023] By connecting with the production line management platform, edge detection nodes can promptly transmit monitored data to the platform. The platform then performs further analysis based on the received data to optimize the production process or provide early warnings of equipment malfunctions. Each edge detection node is configured with multimodal monitoring tasks, which are set according to the characteristics of different production line equipment and production needs. These tasks include, for example, temperature monitoring, current and voltage monitoring, vibration monitoring, and operational status monitoring. The multimodal monitoring task settings ensure that the edge detection nodes can simultaneously monitor different types of data inputs, enhancing the system's ability to monitor equipment status.

[0024] After the P production line equipment is calibrated, during the production of automotive brushless motors, the M edge detection nodes schedule the M multimodal monitoring tasks through dynamic computing power allocation to perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring equipment.

[0025] Before production begins, the P production line devices are calibrated. During the calibration process, the production line devices are adjusted according to preset standards or production requirements to ensure that there are no serious deviations throughout the production process. The calibration includes the calibration of the production line devices' sensors and the adjustment of their operating parameters.

[0026] During the production process, the M edge detection nodes rationally allocate computing resources based on the real-time needs of the production line and the complexity of the monitoring tasks. For example, when a certain production link requires more precise monitoring, the edge detection nodes allocate more computing resources to it to improve monitoring accuracy; while when resources are relatively idle in other production links, the edge detection nodes dynamically adjust computing power to ensure a balance between system efficiency and computing power.

[0027] As the production line equipment operates, the edge detection node continuously monitors the status of the production line equipment. When the monitoring data of a certain production line equipment deviates, for example, when there are abnormal fluctuations in temperature, current or voltage, the edge detection node triggers deviation diagnosis and triggers a calibration command based on the preset deviation standard. This means that when the working status of the production line equipment reaches the preset deviation standard, it will actively perform calibration to ensure that the production line equipment returns to normal working status.

[0028] Furthermore, based on the motor defect distribution, process failure mode backtracking is performed to locate P production line equipment at P production process nodes. The method includes:

[0029] Defect features are extracted from the motor defect distribution to obtain multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the motor production. Based on the multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the motor production, defect patterns are aggregated to obtain a set of defect pattern clusters. Failure process nodes are backtracked and matched according to the set of defect pattern clusters to obtain a set of high-confidence process nodes. The equipment instances corresponding to P production process nodes in the set of high-confidence process nodes are called to obtain the P production line equipment.

[0030] Defect features are extracted from the distribution of defects in motors. Defect type refers to the type of defect, such as common assembly problems, electrical faults, and mechanical wear that are common in the manufacturing process. Defect spatial location refers to the specific location where the defect occurs, which can be any part of the motor, such as the windings, stator, rotor, or a part of the housing. Defect spatial location helps to determine which production steps or equipment may cause these defects. Defect overall impact refers to the degree of impact of the defect on the whole machine. The impact of each defect will affect the working performance of the whole machine depending on its location and type. For example, winding defects will cause the motor performance to degrade, while housing deformation will affect the appearance or protection performance of the motor. By evaluating the defect overall impact, critical defects can be identified more accurately and prioritized.

[0031] Defect patterns are aggregated based on multiple defect types, spatial locations, and overall impact on the machine. Specifically, if multiple defects have similar types, such as multiple winding defects or rotor malfunctions, they are aggregated into one defect pattern. If multiple defects occur in adjacent areas of the motor or in the same part of the same process, they are also aggregated into one pattern. Defects with similar overall impact on the machine are grouped into the same defect pattern; for example, if multiple defects have similar effects on motor performance, they belong to the same pattern cluster. After aggregation, all defects are divided into a set of defect pattern clusters, with each cluster representing a class of defects with similar characteristics.

[0032] Failure process node backtracking involves analyzing the correlation between defect patterns and various process nodes to identify possible causes of defects. For example, a specific winding defect might be caused by improper machine parameter settings or operation during the winding process. Based on the defect pattern cluster set, combined with the process flow and production parameters, backtracking analysis is performed. By examining the operation records and equipment status of each process step during production, it can be inferred which process nodes have equipment problems that lead to these defects. Through backtracking analysis, the process nodes that cause the defects are identified, and based on the confidence level of the backtracking process, i.e., the correlation between the defect and the process node, a set of high-confidence process nodes is selected. These high-confidence process nodes are the most likely process nodes associated with the defects.

[0033] For each high-confidence process node, its corresponding equipment instance is queried. These equipment instances are the devices on the production line that actually perform the task of that process, such as winding machines, assembly line equipment, and testing equipment. These devices are the key objects that need to be monitored and maintained because they are closely related to the occurrence of defects. Finally, by calling the equipment instances corresponding to the set of high-confidence process nodes, P production line devices are obtained. These production line devices are the devices that need to be focused on in subsequent production processes.

[0034] Furthermore, by performing backtracking matching of failed process nodes based on the defect pattern cluster set to obtain a high-confidence process node set, the method includes:

[0035] A pre-constructed process failure mode library is used, and the potential failure process node set is located by mapping the defect mode cluster set to the process failure mode library. Process parameter records are retrieved based on the production batches of the defect mode cluster set. Multiple process correlation coefficients of multiple failure process nodes in the potential failure process node set are calculated based on the process parameter records to generate an associated process node set. The associated process node set is traversed using a preset correlation threshold to obtain a high-confidence process node set. The equipment instances corresponding to P production process nodes in the high-confidence process node set are called to obtain the P production line equipment.

[0036] A process failure mode library (FLM) is a database containing different failure modes and their corresponding process steps. Each process failure mode describes a specific process problem, such as equipment failure, parameter abnormality, or improper operation. The FLM library typically includes known failure modes, potential defects, and related parameters for different production processes. By mapping a set of defect mode clusters to modes in the FLM library, the failure modes associated with the current defect are identified. In other words, based on the identified defect type, location, and impact, the FLM library is used to find the failure modes associated with these defects, thereby locating the set of potential failure process nodes.

[0037] A production batch refers to motor products manufactured within a certain time period. These products share similar process conditions and equipment status. Based on the production batch of the defect pattern cluster set, the process parameter records of the same production batch are retrieved. The process parameter records include the operating parameters of each process in the production process, such as temperature, pressure, speed, vibration, material ratio, etc. This data helps to analyze and diagnose which processes caused the failure in the production batch.

[0038] The process correlation coefficient is an indicator that measures the degree of mutual influence between processes. For example, a change in the parameter of one process affects the performance of subsequent processes. By analyzing multiple process parameter records, the correlation coefficient between different processes can be calculated. Common methods for calculating correlation coefficients include Pearson correlation coefficient and Spearman's rank correlation coefficient. The process correlation coefficient reflects the degree to which the parameter of one process affects other processes. If the parameter changes of two processes are highly correlated, it means that there is a potential coupling relationship between them, and a failure in one process may affect the performance of the other process. Based on the calculated process correlation coefficients, multiple failed process nodes are grouped according to their correlation to form a set of associated process nodes. The set of associated process nodes represents a group of processes that may influence each other during the production process, and these may be the key nodes leading to defects.

[0039] To identify the most influential associated process nodes from the set of associated process nodes, a correlation threshold is pre-defined. This threshold determines which process nodes are sufficiently correlated; only those with a correlation exceeding this threshold are considered to require close monitoring. The entire set of associated process nodes is iterated through, and based on the comparison between the correlation coefficient and the correlation threshold, those with strong correlations are selected. These are the high-confidence process nodes, representing the processes most likely to cause defects. For each high-confidence process node, its associated equipment instances are invoked, forming P production line devices. These devices are critical and require further calibration or optimization to prevent quality issues during production.

[0040] Furthermore, cross-process parameter coupling analysis is performed on the P production process nodes, and the P production line equipment is reorganized into M groups of parallel monitoring equipment. The method includes:

[0041] Extract multiple historical time-series defect chain records covering the P production process nodes; combine and enumerate the P production process nodes to obtain... Group production process nodes; calculate the above based on the multiple historical time-series defect chain records. Group production process nodes The coupling strength of each process; based on the aforementioned The process coupling strength topology connects the P production process nodes to obtain a process coupling topology graph; based on a preset coupling strength threshold, the process coupling topology is traversed and low-weight edges are pruned to obtain M independent connected subgraphs; based on the M independent connected subgraphs, the monitoring groups of the P production line equipment are mapped, and the M groups of parallel monitoring equipment are output.

[0042] Historical time-series defect chain records refer to time-series data on the occurrence and evolution of defects during the production process. These records include the occurrence time, type, scope of impact, and evolution trajectory of defects over time. For P production process nodes, multiple corresponding historical time-series defect chain records are extracted to analyze which process nodes are often the source of defects in the production process, and then make improvements.

[0043] Combinatorial enumeration refers to generating all possible pairings between production process nodes. For P production process nodes, enumerating all possible node pairs results in a number of combinations. That is, there is a combination between every two production process nodes.

[0044] Process coupling strength refers to the degree of influence of changes in the operation or parameters of one production process node on another. Higher coupling strength indicates a stronger correlation between the two production process nodes, suggesting a potential mutual influence. By analyzing multiple historical time-series defect chain records, the degree of influence between these production process nodes can be quantified. Specifically, process coupling strength is calculated by analyzing process relationships within historical time-series defect chain records. Specific calculation methods can be based on statistical analysis of historical data, such as calculating Pearson correlation coefficients, covariance, or other correlation measures. These statistical methods can quantify the process coupling strength between two production process nodes, representing the degree of mutual influence between the nodes.

[0045] A process coupling topology graph is a graphical structure used to represent the coupling relationships between different production process nodes. In this process coupling topology graph, each production process node is treated as a node, and the process coupling strength between production process nodes is used as the weight of the edge. If the coupling strength between two production process nodes is high, the edge connecting them will have a larger weight; conversely, if the coupling strength is low, the edge weight will be smaller.

[0046] To reduce the impact of loosely coupled production process nodes, a coupling strength threshold is pre-defined. If the coupling strength between two production process nodes is lower than this threshold, their connection is considered weak and can be removed from the topology graph. Low-weight edge pruning reduces redundant connections in the topology graph, preserving the relationships between production process nodes that have a significant impact on the production process. The process coupling topology graph is traversed, and pruning is determined based on the edge weights. If the weight of an edge is less than the coupling strength threshold, that edge is deleted. The pruned graph is more concise and intuitive. After pruning, M independent connected subgraphs are obtained, where M is a positive integer. The nodes in each independent connected subgraph represent a group of highly related processes with high coupling strength, typically representing a group of interdependent process nodes in the production process.

[0047] Each independent connected subgraph represents a set of production process nodes with strong coupling relationships. Based on these production process nodes, the corresponding production equipment is mapped to the corresponding monitoring group. Each production process node is executed by one or more production line devices. Therefore, the process nodes in each independent connected subgraph will be connected with the corresponding equipment to form a set of parallel monitoring devices. The parallel monitoring device group means that these devices can be monitored simultaneously, ensuring that the correlation between processes in each production link is handled in a timely manner. If a device malfunctions, the node affecting the entire production process can be quickly located and corrected.

[0048] Furthermore, based on the multiple historical time-series defect chain records, the calculation of the Group production process nodes The method for determining the coupling strength of individual processes includes:

[0049] Based on the first set of production process nodes, extract the first set of time-series defect chains covering the first defect type from the multiple historical time-series defect chain records; calculate the Pearson correlation coefficient of the first set of time-series defect chains to obtain the coupling strength of the first type of process; similarly calculate the coupling strength of the first set of production process nodes in K types of defects in K process chains; weight the K process coupling strengths according to the K performance impact correlations of the K types of defects, and output the coupling strength of the first process; 335 similarly calculate the... The production process nodes of the group The coupling strength of each process.

[0050] The first set of production process nodes is Any group of production process nodes is taken as the current analysis object. From multiple historical time-series defect chain records, the first group of production process nodes covering the first defect type is selected as the first time-series defect chain. This means that if the first group of production process nodes is related to a certain type of defect, such as mechanical defects or electrical faults, then the time-series data related to these defect types are extracted from the historical defect chain records and used as a time-series defect chain for subsequent analysis.

[0051] The Pearson correlation coefficient is a statistical method for measuring the strength and direction of the linear relationship between two variables. In this step, the Pearson correlation coefficient is used to measure the correlation of the first set of time-series defect chains. The Pearson correlation coefficient ranges from -1 to 1. A value close to 1 indicates a strong positive correlation between the two production process nodes; a value close to -1 indicates a strong negative correlation; and a value close to 0 indicates no linear relationship between the two process nodes. By calculating the Pearson correlation coefficient, the coupling strength between the first set of production process nodes is determined, thus obtaining the coupling strength of the first type of process.

[0052] In actual production, different types of defects have different effects on production process nodes. For K types of defects, repeat the above steps to calculate the coupling strength of the first group of production process nodes under different defect types. The analogy here means applying a similar calculation method. Finally, we obtain the coupling strength of K processes under K types of defects, which reflects the influence of different defect types on the coupling relationship between the first group of production process nodes.

[0053] Different defect types have varying degrees of impact on the production process and the performance of the final product. Some defect types have a significant impact on the overall product quality or the stability of the production process, while others have a smaller impact. Performance impact correlation is an indicator used to measure the degree of impact of each defect type on the production process. Based on the performance impact correlation of each defect type, the process coupling strength is weighted. Specifically, the process coupling strength of each defect type is multiplied by its corresponding performance impact correlation. Thus, defect types with a greater impact will have a larger influence on the final coupling strength calculation result, while defect types with a smaller impact will have a smaller impact. Finally, the first process coupling strength is output.

[0054] for For each group of production process nodes, the same analysis process as the first group of production process nodes is used to obtain the corresponding... The coupling strength of each process.

[0055] Furthermore, the method involves configuring M edge detection nodes for the M groups of parallel monitoring devices, and includes:

[0056] The P equipment operating baselines of the P production line equipment are obtained interactively; the P equipment operating baselines are reorganized according to the M sets of parallel monitoring equipment to obtain M sets of equipment operating baselines; based on the M sets of equipment operating baselines, M multimodal monitoring tasks are set on the M edge detection nodes, wherein the multimodal monitoring tasks are used to perform multi-threaded operation error monitoring of parallel equipment.

[0057] Equipment operating baselines refer to the standard or reference values ​​of various parameters of equipment under normal operating conditions. These are typically measured when the equipment is stable and fault-free, reflecting its performance under normal operating conditions. These operating baselines are obtained from various production line devices through interactive communication with the equipment control system or data acquisition system. This process retrieves historical operating data or standard setting values ​​from the equipment. Obtaining these operating baselines provides a reference for subsequent monitoring tasks and helps determine if any equipment anomalies have occurred.

[0058] Based on the equipment members of the M groups of parallel monitoring equipment, the operating baselines of P equipment are reorganized and redistributed to the M groups. This ensures that the operating baselines of each group of equipment can adapt to the common characteristics or operating environment of the equipment within the group. The reorganized operating baselines of the M groups of equipment are used for the monitoring tasks of their respective parallel monitoring equipment groups, helping to analyze the performance of each equipment group more accurately.

[0059] Based on the operating baseline of M groups of equipment, M multimodal monitoring tasks are set for each group of equipment on M edge detection nodes. These multimodal monitoring tasks are executed in a multi-threaded manner, meaning each edge detection node can handle multiple monitoring tasks simultaneously, thereby improving monitoring efficiency and real-time performance. Through these multimodal monitoring tasks, the edge detection nodes monitor the operating status of the equipment in real time. When certain parameters of the equipment deviate from the operating baseline, potential errors are immediately identified and diagnosed. These operational error monitoring tasks help to detect potential equipment problems in advance, such as overheating or abnormal current, ensuring the smooth operation of the production process.

[0060] Furthermore, the method also includes:

[0061] Based on the defect pattern cluster set, L associated process parameters of the P production line equipment are extracted; using the L associated process parameters as search conditions, L benchmark datasets are extracted from the historical good product production operation logs; the benchmark center value and dynamic deviation threshold of the L benchmark datasets are calculated to generate calibration benchmark rules; and the calibration benchmark rules are used to perform calibration status calibration processing on the P production line equipment.

[0062] Based on the defect pattern cluster set, L associated process parameters are extracted from P production line equipment, where L is a positive integer. The associated process parameters are variables that are closely related to the production process and equipment performance. They are closely related to equipment performance, defect patterns and production quality. They can be equipment operating status parameters or factors affecting the production environment.

[0063] Historical good-product production operation logs record normal operating data of equipment during historical production, including detailed parameter records of each piece of equipment under normal operating conditions, such as temperature, current, and vibration. These logs provide data references for the standard operation of the equipment. Based on L related process parameters, L relevant benchmark datasets are selected from the historical good-product production operation logs. Each benchmark dataset corresponds to a set of standard operating data for the equipment. By retrieving these data, the performance of the equipment under optimal operating conditions is determined.

[0064] For L benchmark datasets, the benchmark center value is calculated. The benchmark center value is representative of a set of normally operating data, usually the mean or median of the dataset. This value represents the operating level of the equipment under ideal conditions. The dynamic deviation threshold is the standard deviation of the benchmark dataset, representing the range of parameter fluctuations under normal operating conditions, and is used to determine whether abnormal fluctuations have occurred in the equipment's operating parameters. By calculating the benchmark center value and the dynamic deviation threshold, calibration benchmark rules are generated. The calibration benchmark rules define the standard parameter range that the equipment should follow under normal operating conditions.

[0065] By using calibration benchmark rules, the operating status of P production line equipment is calibrated. This process adjusts the operating status of the equipment to conform to the standard range defined in the calibration benchmark rules. The purpose is to ensure that the equipment can operate within the predetermined working range and avoid production defects or efficiency problems caused by the equipment deviating from the standard operating status. Through calibration status calibration, it can be ensured that the equipment always maintains the best working status during the production process, thereby improving production quality and efficiency.

[0066] Furthermore, in the production process of automotive brushless motors, the M edge detection nodes dynamically allocate and schedule the M multimodal monitoring tasks to perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring devices. The method includes:

[0067] Based on the first set of historical deviation correction frequencies of the first set of parallel monitoring devices, the monitoring computing power of the first set of parallel monitoring devices is pre-allocated; the first edge detection node schedules the first set of multimodal monitoring subtasks in the first multimodal monitoring task, and performs real-time parameter monitoring of the first set of parallel monitoring devices based on the first set of device operating baseline mapping; when any task in the first set of multimodal monitoring subtasks detects that the real-time parameters deviate from the dynamic deviation threshold of the first set of device operating baseline, a device calibration command is triggered.

[0068] The first group of parallel monitoring devices is any one of the M groups of parallel monitoring devices, serving as the current analysis object. The historical deviation correction frequency of the first group refers to the number of times parameter deviations occurred and corrections were performed during the historical production process of the first group of parallel monitoring devices. This reflects the frequency of anomalies occurring during the operation of the first group of parallel monitoring devices. Devices with higher frequencies require more monitoring resources to ensure timely handling of anomalies. Based on the historical deviation correction frequency of the first group, monitoring computing power is pre-allocated to the first group of parallel monitoring devices. If the historical deviation correction frequency of the first group is high, more computing power resources are allocated to the first group of parallel monitoring devices to ensure they have sufficient processing power in real-time monitoring for rapid detection and response to potential anomalies.

[0069] The first multimodal monitoring task refers to the task of simultaneously monitoring multiple device parameters. The first set of multimodal monitoring subtasks within this task is scheduled to be executed at the first edge detection node. Each subtask is responsible for monitoring a specific parameter, ensuring the comprehensiveness and efficiency of the monitoring task. After task scheduling, the first set of device operating baselines is used as a reference to ensure that the range of monitored parameters matches the device operating baseline. This helps to accurately identify whether any abnormalities have occurred in the device.

[0070] When the monitored real-time parameters deviate from the dynamic deviation threshold of the first set of equipment operating baselines, an automatic equipment calibration command is triggered. This command includes adjusting equipment parameters or resetting certain equipment settings to ensure the equipment returns to normal operating conditions. This means the equipment needs adjustment to return to normal operation, which not only reduces the need for manual intervention but also improves the stability and efficiency of the production line.

[0071] Furthermore, after performing global calibration of the first group of parallel monitoring devices in response to the device calibration command, the monitoring computing power allocation of the first group of parallel monitoring devices is updated.

[0072] Global calibration refers to the unified adjustment of the first group of parallel monitoring devices. Upon receiving a calibration command, the status of all devices in the first group is comprehensively checked and adjusted centrally. This means it's not just the calibration of individual devices, but the overall calibration of the entire device group, ensuring that the devices within the group maintain a consistent operating state when working collaboratively. After calibration, the status of the devices in the current production process is assessed to determine if the allocation of monitoring computing power needs adjustment. For example, if a device becomes more stable after calibration, the computing power allocation for its monitoring is reduced, freeing up resources for other devices that require more monitoring. If a device still exhibits some instability after calibration, the computing power allocation for that device is increased to ensure real-time monitoring and data accuracy.

[0073] In summary, the intelligent production line control method for automotive brushless motors provided in this application has the following technical effects:

[0074] Comprehensive defect detection identifies various defects in brushless motors during trial production and reveals their distribution, providing a data foundation for subsequent analysis and enabling precise tracking and location of defects during production. By retrospectively analyzing process failure modes based on the defect distribution, the corresponding production process nodes can be traced, and the relevant production line equipment can be accurately located. This provides a clear path for quality control on the production line, helping to identify potential failure causes at the source and implement targeted optimization and adjustments. Cross-process parameter coupling analysis analyzes the correlation between various production process nodes, effectively identifying the mutual influence between different production process nodes. This analytical method can help... This helps to understand the complexity of the production process, identify potential problems between processes, and thus provide a basis for equipment optimization. By configuring M edge detection nodes, it can ensure that the real-time data of each group of parallel monitoring equipment can be fully monitored. The communication between the edge detection nodes and the production line management platform enables the monitoring system to respond to changes in equipment status in a timely manner. The configuration of multimodal monitoring tasks ensures comprehensive monitoring of various key parameters of the equipment. When abnormal equipment status occurs, it can quickly identify and trigger necessary calibration operations, thereby ensuring the accuracy and stability of equipment in the production process, effectively reducing the risk of motor failures in electric vehicles and other motor vehicles during use, and thus improving the overall stability and market competitiveness of electric vehicles and other motor vehicles.

[0075] Example 2, based on the same inventive concept as the intelligent production line control method for a brushless motor in an automobile described in the previous examples, such as... Figure 2 As shown in the figure, this application embodiment provides an intelligent production line management system for automotive brushless motors, the system comprising:

[0076] The comprehensive defect detection module 10 is used to perform comprehensive defect detection on the trial-produced automotive brushless motor to obtain the motor defect distribution; the failure mode backtracking module 20 is used to perform process failure mode backtracking based on the motor defect distribution to locate P production line equipment at P production process nodes; the parameter coupling analysis module 30 is used to perform cross-process parameter coupling analysis on the P production process nodes and reorganize the P production line equipment into M groups of parallel monitoring equipment; the detection node configuration module 40 is used to configure M edge detection nodes for the M groups of parallel monitoring equipment, wherein the M edge detection nodes are connected to the production line management platform through an IoT gateway protocol, and the M edge detection nodes are configured with M multimodal monitoring tasks; the dynamic calibration triggering module 50 is used to perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring equipment during the automotive brushless motor production process after the P production line equipment has been calibrated.

[0077] Furthermore, the failure mode backtracking module 20 is used to perform the following operation steps:

[0078] Defect features are extracted from the motor defect distribution to obtain multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the motor production. Based on the multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the motor production, defect patterns are aggregated to obtain a set of defect pattern clusters. Failure process nodes are backtracked and matched according to the set of defect pattern clusters to obtain a set of high-confidence process nodes. The equipment instances corresponding to P production process nodes in the set of high-confidence process nodes are called to obtain the P production line equipment.

[0079] Furthermore, the failure mode backtracking module 20 is used to perform the following operation steps:

[0080] A pre-constructed process failure mode library is used, and the potential failure process node set is located by mapping the defect mode cluster set to the process failure mode library. Based on the production batches of the defect mode cluster set, process parameter records are retrieved. Based on the process parameter records, multiple process correlation coefficients of multiple failure process nodes in the potential failure process node set are calculated to generate an associated process node set. The associated process node set is traversed using a preset correlation threshold to obtain a high-confidence process node set. The equipment instances corresponding to P production process nodes in the high-confidence process node set are called to obtain the P production line equipment.

[0081] Furthermore, the parameter coupling analysis module 30 is used to perform the following operation steps:

[0082] Extract multiple historical time-series defect chain records covering the P production process nodes; combine and enumerate the P production process nodes to obtain... Group production process nodes; calculate the above based on the multiple historical time-series defect chain records. Group production process nodes The coupling strength of each process; based on the aforementioned The process coupling strength topology connects the P production process nodes to obtain a process coupling topology graph; based on a preset coupling strength threshold, the process coupling topology is traversed and low-weight edges are pruned to obtain M independent connected subgraphs; based on the M independent connected subgraphs, the monitoring groups of the P production line equipment are mapped, and the M groups of parallel monitoring equipment are output.

[0083] Furthermore, the parameter coupling analysis module 30 is used to perform the following operation steps:

[0084] Based on the first set of production process nodes, extract the first set of time-series defect chains covering the first defect type from the multiple historical time-series defect chain records; calculate the Pearson correlation coefficient of the first set of time-series defect chains to obtain the coupling strength of the first type of process; similarly calculate the coupling strength of the first set of production process nodes in K types of defects in K process chains; weight the K process coupling strengths according to the K performance impact correlations of the K types of defects, and output the coupling strength of the first process; similarly calculate the coupling strength of the first set of production process nodes in K process chains covering K types of defects in K process chains. The production process nodes of the group The coupling strength of each process.

[0085] Furthermore, the detection node configuration module 40 is used to perform the following operation steps:

[0086] The P equipment operating baselines of the P production line equipment are obtained interactively; the P equipment operating baselines are reorganized according to the M sets of parallel monitoring equipment to obtain M sets of equipment operating baselines; based on the M sets of equipment operating baselines, M multimodal monitoring tasks are set on the M edge detection nodes, wherein the multimodal monitoring tasks are used to perform multi-threaded operation error monitoring of parallel equipment.

[0087] Furthermore, the failure mode backtracking module 20 is used to perform the following operation steps:

[0088] Based on the defect pattern cluster set, L associated process parameters of the P production line equipment are extracted; using the L associated process parameters as search conditions, L benchmark datasets are extracted from the historical good product production operation logs; the benchmark center value and dynamic deviation threshold of the L benchmark datasets are calculated to generate calibration benchmark rules; and the calibration benchmark rules are used to perform calibration status calibration processing on the P production line equipment.

[0089] Furthermore, the dynamic calibration trigger module 50 is used to perform the following operation steps:

[0090] Based on the first set of historical deviation correction frequencies of the first set of parallel monitoring devices, the monitoring computing power of the first set of parallel monitoring devices is pre-allocated; the first edge detection node schedules the first set of multimodal monitoring subtasks in the first multimodal monitoring task, and performs real-time parameter monitoring of the first set of parallel monitoring devices based on the first set of device operating baseline mapping; when any task in the first set of multimodal monitoring subtasks detects that the real-time parameters deviate from the dynamic deviation threshold of the first set of device operating baseline, a device calibration command is triggered.

[0091] Furthermore, after performing global calibration of the first group of parallel monitoring devices in response to the device calibration command, the monitoring computing power allocation of the first group of parallel monitoring devices is updated.

[0092] Through the foregoing detailed description of an intelligent production line management method for an automotive brushless motor, those skilled in the art can clearly understand the intelligent production line management system for an automotive brushless motor in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to in the method section.

[0093] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent production line control of brushless motors for automobiles, characterized in that, The method includes: By conducting comprehensive defect testing on the brushless motors of trial-produced automobiles, the distribution of motor defects was obtained. Based on the motor defect distribution, process failure modes are traced back to locate P production line equipment at P production process nodes; Perform cross-process parameter coupling analysis on the P production process nodes, and reorganize the P production line equipment into M groups of parallel monitoring equipment; M edge detection nodes are configured for the M groups of parallel monitoring devices, wherein the M edge detection nodes are connected to the production line management and control platform through an Internet of Things gateway protocol, and the M edge detection nodes are configured with M multimodal monitoring tasks; After the P production line equipment is calibrated, during the production of automotive brushless motors, the M edge detection nodes schedule the M multimodal monitoring tasks through dynamic computing power allocation to perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring equipment. The method for backtracking process failure modes based on the motor defect distribution to locate P production line equipment at P production process nodes includes: Defect features are extracted from the motor defect distribution to obtain multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the motor manufacturing defects. Based on the multiple defect types, multiple defect spatial locations, and multiple defects' overall impact on the machine, the defect patterns of the multiple motor production defects are aggregated to obtain a set of defect pattern clusters; Based on the defect pattern cluster set, backtracking and matching of failed process nodes are performed to obtain a high-confidence process node set; Call the equipment instances corresponding to the P production process nodes in the high-confidence process node set to obtain the P production line equipment; The method of performing backtracking matching of failed process nodes based on the defect pattern cluster set to obtain a high-confidence process node set includes: A pre-built process failure mode library is constructed, and the potential failure process node set is located by mapping the defect mode cluster set to the process failure mode library; Based on the production batch of the defect pattern cluster set, retrieve the process parameter records; Based on the process parameter records, calculate the process correlation coefficients of multiple failed process nodes in the potential failure process node set, and generate an associated process node set; The set of associated process nodes is traversed using a preset correlation threshold to obtain a set of high-confidence process nodes; The equipment instances corresponding to the P production process nodes in the high-confidence process node set are called to obtain the P production line equipment.

2. The intelligent production line control method for a brushless automotive motor as described in claim 1, characterized in that, The method involves performing cross-process parameter coupling analysis on the P production process nodes and reorganizing the P production line equipment into M groups of parallel monitoring devices. Extract multiple historical time-series defect chain records covering the P production process nodes; By combining and enumerating the P production process nodes, we can obtain Group production process nodes; Based on the aforementioned multiple historical time-series defect chain records, calculate the... Group production process nodes The coupling strength of each process step; According to the above The process coupling strength topology connects the P production process nodes to obtain the process coupling topology diagram; Based on a preset coupling strength threshold, the process coupling topology is traversed and low-weight edges are pruned to obtain M independent connected subgraphs. Based on the M independent connected subgraphs, perform monitoring group mapping for the P production line equipment, and output the M groups of parallel monitoring equipment.

3. The intelligent production line control method for a brushless automotive motor as described in claim 2, characterized in that, Based on the aforementioned multiple historical time-series defect chain records, calculate the... Group production process nodes The method for determining the coupling strength of individual processes includes: Based on the first set of production process nodes, extract the first set of time-series defect chains covering the first defect type from the multiple historical time-series defect chain records; Calculate the Pearson correlation coefficient of the first set of time-series defect chains to obtain the coupling strength of the first type of process. By analogy, calculate the coupling strength of the first group of production process nodes in the K processes of K defect types; Based on the correlation of the K performance effects of the K types of defects, the coupling strength of the K processes is weighted and the coupling strength of the first process is output. By analogy, the calculation of the above The production process nodes of the group The coupling strength of each process.

4. The intelligent production line control method for a brushless automotive motor as described in claim 1, characterized in that, The method involves configuring M edge detection nodes for the M groups of parallel monitoring devices, wherein the method includes: Interactively obtain the operating baselines of the P production line devices for each of the P devices; Based on the M groups of parallel monitoring devices, the operating baselines of the P devices are reorganized to obtain the operating baselines of the M groups of devices; Based on the operating baseline of the M groups of devices, M multimodal monitoring tasks are set on the M edge detection nodes, wherein the multimodal monitoring tasks are used to perform multi-threaded operation error monitoring of parallel devices.

5. The intelligent production line control method for a brushless automotive motor as described in claim 1, characterized in that, The method further includes: Based on the defect pattern cluster set, extract L associated process parameters of the P production line equipment; Using the L associated process parameters as search conditions, L benchmark datasets are extracted from the historical good product production operation logs; Calculate the benchmark center value and dynamic deviation threshold of the L benchmark datasets to generate calibration benchmark rules; The calibration benchmark rules are used to perform calibration status calibration on the P production line equipment.

6. The intelligent production line control method for a brushless automotive motor as described in claim 4, characterized in that, In the production process of brushless motors for automobiles, the M edge detection nodes schedule the M multimodal monitoring tasks through dynamic computing power allocation, and perform dynamic calibration triggering driven by deviation diagnosis on the M groups of parallel monitoring devices. The method includes: Based on the first set of historical deviation correction frequencies of the first set of parallel monitoring devices, the monitoring computing power of the first set of parallel monitoring devices is pre-allocated. The first edge detection node schedules the first group of multimodal monitoring subtasks in the first multimodal monitoring task, and performs real-time parameter monitoring of the first group of parallel monitoring devices based on the first group of device operation baseline mapping; When any task in the first group of multimodal monitoring subtasks detects a dynamic deviation threshold in real-time parameters that deviates from the operating baseline of the first group of devices, a device calibration command is triggered.

7. The intelligent production line control method for a brushless automotive motor as described in claim 6, characterized in that, After performing global calibration of the first group of parallel monitoring devices in response to the device calibration command, the monitoring computing power allocation of the first group of parallel monitoring devices is updated.

8. An intelligent production line control system for automotive brushless motors, characterized in that, A smart production line control method for implementing the brushless motor for automobiles according to any one of claims 1-7, the system comprising: The comprehensive defect detection module is used to perform comprehensive defect detection on the brushless motor of a trial-produced automobile to obtain the defect distribution of the motor. The failure mode backtracking module is used to backtrack the process failure modes based on the motor defect distribution and locate P production line equipment at P production process nodes. The parameter coupling analysis module is used to perform cross-process parameter coupling analysis on the P production process nodes and reorganize the P production line equipment into M groups of parallel monitoring equipment. The detection node configuration module is used to configure M edge detection nodes for the M groups of parallel monitoring devices. The M edge detection nodes are connected to the production line management and control platform through an Internet of Things gateway protocol, and the M edge detection nodes are configured with M multimodal monitoring tasks. The dynamic calibration trigger module is used to perform dynamic calibration triggering driven by deviation diagnosis on the M sets of parallel monitoring equipment during the production process of automotive brushless motors after the P production line equipment have been calibrated.

Images