Defective product prediction and early warning method and system for manually assembled motor

Abstract

The invention discloses a defective product prediction and early warning method and system for manually assembled motors, and relates to the field of industrial detection, and the method comprises the steps: obtaining a real-time video stream in a motor assembly process, and determining key motion features of a target person according to the real-time video stream; comparing the key action characteristics with a preset standard action template to obtain an action difference degree; according to the tool type, the performance parameter and the use state information of the assembly tool used by the target person, establishing an applicability model of the assembly tool; inputting the key action features into an applicability model to obtain a tool application score of the target person; acquiring real-time pressure data and real-time temperature data of an assembling environment in a motor assembling process, and determining a motor assembling score of a target person; and when the motor assembly score is lower than a preset qualified threshold value, early warning information is sent to a supervision client. By implementing the method, the application degree of the assembly tool by the assembly personnel can be identified, and the defective motor product can be predicted more comprehensively.

Description

Technical Field

[0001] The present application relates to the field of industrial inspection, and in particular to a method and system for predicting and warning defective products of manually assembled motors. Background Art

[0002] In the manufacturing industry, product quality control is a key link to ensure product reliability, safety and customer satisfaction. With the intensification of market competition and the diversification of consumer demand, companies are paying more and more attention to improving the efficiency of quality management in the production process. Especially in the field of manual assembly of motors, due to the complexity of the assembly process and susceptibility to human factors, the production of defective products often leads to increased costs and reduced production efficiency. Therefore, the development of efficient and accurate defective product prediction and early warning systems has become an important means to improve production quality.

[0003] Among related technologies, defective product prediction technology mainly relies on automated visual inspection systems and machine learning methods. These systems capture images of the assembly process through high-resolution cameras and use image recognition technology to detect defects on the product surface. In addition, some advanced inspection systems also combine sensor data such as temperature and pressure to perform multi-dimensional monitoring through data analysis software.

[0004] However, manual assembly of motors is greatly affected by human factors. Different assemblers have different technical proficiency and use different assembly tools to different degrees. Therefore, the quality of manually assembled motors may vary. The defective product prediction technology in the related art is difficult to effectively identify the degree of assembly personnel's use of assembly tools, and the prediction of defective products is not comprehensive enough. Summary of the invention

[0005] The present application provides a method and system for predicting and warning defective products of manually assembled motors, which are used to identify the degree of assembly personnel's use of assembly tools and to more comprehensively predict defective motors.

[0006] In the first aspect, the present application provides a method for predicting and warning defective products of manually assembled motors, which is applied to a prediction and warning system, and the method includes: obtaining a real-time video stream during the motor assembly process, and determining the key action features of a target person based on the real-time video stream; the key action features include the target person's hand action trajectory, strength, speed and frequency; comparing the key action features with a preset standard action template to obtain the action difference; establishing a suitability model for the assembly tool based on the tool type, performance parameters and usage status information of the assembly tool used by the target person; inputting the key action features into the suitability model to obtain the target person's tool application score; obtaining real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process, and determining the target person's motor assembly score based on the action difference, real-time pressure data, real-time temperature data and tool application score; when the motor assembly score is lower than the preset qualified threshold, sending a warning message to the supervision client.

[0007] In the above embodiment, the prediction and early warning system analyzes the key action features of the target person by acquiring real-time video streams, and compares them with standard action templates to obtain action differences. At the same time, a suitability model is established based on the various attributes of the assembly tools used by the target person, and the key action features are input into the model to obtain a tool application score. Combined with the real-time pressure and temperature data of the assembly environment, the motor assembly score of the target person is determined by comprehensively considering the action difference, environmental data and tool application score. When the score is lower than the preset threshold, an early warning message is issued. Not only the standardization of personnel actions is taken into account, but also the proper use of tools and the influence of the assembly environment are taken into account. The assembly quality is evaluated in multiple dimensions, and the risk of defective motors is predicted more comprehensively, which can provide timely warnings and reduce the production of defective products.

[0008] In combination with some embodiments of the first aspect, in some embodiments, an applicability model of the assembly tools is established based on the tool type, performance parameters and usage status information of the assembly tools used by the target personnel, specifically including: obtaining the tool type of the assembly tools used by the target personnel; the tool types include power tools and hand tools; obtaining the performance parameters of the assembly tools; the performance parameters include rated power, speed, torque, material hardness and size specifications; obtaining the usage status information of the assembly tools; the usage status information includes the actual usage time, maintenance records and failure frequency of the tools; dividing the complexity values ​​of the assembly tools according to the tool types, determining the applicable scope of the assembly tools according to the performance parameters, and determining the reliability values ​​of the assembly tools according to the usage status information; comprehensively considering the complexity values, applicable scope and reliability values ​​of the assembly tools to establish a three-dimensional applicability evaluation model of the assembly tools as a applicability model.

[0009] In the above embodiment, the prediction and early warning system first obtains information such as tool type, performance parameters and usage status, and then quantifies the attributes of the tool from three dimensions: complexity, scope of application and reliability. Finally, a three-dimensional applicability evaluation model is established by integrating the three dimensions. The characteristics of the tool itself and its actual usage are systematically considered to more accurately evaluate its impact on assembly quality. Through more sophisticated tool applicability modeling, the tool application of personnel can be more accurately scored, which improves the comprehensiveness and accuracy of defective product prediction.

[0010] In combination with some embodiments of the first aspect, in some embodiments, real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process are obtained, and the motor assembly score of the target personnel is determined based on the action difference, real-time pressure data, real-time temperature data and tool application score, specifically including: detecting the real-time pressure data of the assembly environment in real time through a pressure sensor, and comparing the real-time pressure data with a preset standard pressure range to obtain the pressure difference; detecting the real-time temperature data of the assembly environment in real time through a temperature sensor, and comparing the real-time temperature data with a preset standard temperature range to obtain the temperature difference; inputting the action difference, pressure difference, temperature difference and tool application score into a preset motor quality assessment model to obtain the motor assembly score of the target personnel.

[0011] In the above embodiment, the prediction and early warning system detects the pressure and temperature data in real time through sensors, compares it with the preset standard range, and obtains the difference of pressure and temperature. Then, the action difference, pressure difference, temperature difference and tool application score are input into the preset motor quality assessment model to obtain the final motor assembly score. The environmental data is fully utilized to make the assessment more comprehensive and objective. At the same time, the quality assessment model is used to quantify the influence weight of each indicator on the assembly quality, making the score more scientific and reasonable. By refining the determination process of the assembly score, the defective product prediction and early warning method is further improved, and the accuracy and reliability of the prediction are improved.

[0012] In combination with some embodiments of the first aspect, in some embodiments, before obtaining a real-time video stream during the motor assembly process and determining the key action characteristics of the target person based on the real-time video stream, the method also includes: obtaining historical assembly data of the target person; the historical assembly data includes assembly time, assembly qualification rate and defective causes of multiple products; obtaining physiological characteristic parameters of the target person; the physiological characteristic parameters include gender, age, vision and reaction speed; inputting the historical assembly data and physiological characteristic parameters into a preset personal skill model to obtain the assembly skill level of the target person; determining the target person whose assembly skill level is lower than the preset level threshold as a key monitoring object, and improving the video acquisition frequency and action recognition accuracy of the key monitoring object.

[0013] In the above embodiment, the prediction and early warning system obtains the historical assembly data and physiological characteristic parameters of the target personnel, inputs them into the preset personal skill model, and obtains the assembly skill level assessment result. For personnel with lower skill levels, they are identified as key monitoring targets, and their video acquisition frequency and motion recognition accuracy are improved. This method of adjusting the monitoring strategy according to the skill differences of personnel can prevent defective products in a more targeted manner. By identifying high-risk personnel in advance and strengthening the monitoring of their assembly process, the defective product rate can be reduced from the source. At the same time, limited monitoring resources can be more concentrated on key personnel, thereby improving monitoring efficiency and saving costs while ensuring the early warning effect.

[0014] In combination with some embodiments of the first aspect, in some embodiments, obtaining the physiological characteristic parameters of the target person specifically includes: obtaining the facial feature image and motion feature image of the target person through a camera, and evaluating the target person's vision level, concentration and fatigue based on the facial feature image and the motion feature image; determining the target person's hand dexterity based on the motion feature image; integrating the target person's vision level, concentration, fatigue and hand dexterity to obtain the physiological characteristic parameters.

[0015] In the above embodiment, the prediction and warning system obtains facial and motion feature images of personnel through a camera, obtains indicators such as vision, concentration, fatigue, and hand flexibility through analysis and evaluation, and then integrates them into the final physiological characteristic parameters. Image information is fully utilized, and the physiological state of personnel is evaluated from multiple angles through technical means such as face recognition, expression analysis, and motion capture, making the quantification of physiological characteristics more meticulous. At the same time, additional physiological sensors are avoided, reducing the complexity of the system. Incorporating physical fitness factors more comprehensively into personal skill assessment not only takes into account the objective conditions of employees, but also provides richer background information for subsequent assembly process monitoring, making it more humane and personalized, thereby better guiding the improvement of assembly quality.

[0016] In combination with some embodiments of the first aspect, in some embodiments, when the motor assembly score is lower than a preset qualified threshold, after the step of sending a warning message to the supervision client, the method also includes: pushing a video clip of the target personnel's assembly action in the warning message, and determining the feedback returned by the supervision client after receiving the warning message; if the feedback is that the motor needs to be reworked or scrapped, generating an abnormal motor report recording the cause of the abnormality, the person responsible, and the lost working hours; based on the cumulative records of abnormal motor reports, performing statistical analysis on the assembly quality of multiple production links within a preset time period to obtain quality improvement suggestions; generating a work adjustment plan including workstation layout and employee skill requirements.

[0017] In the above embodiment, the prediction and early warning system first pushes the problem video clip in the early warning information to help supervisors quickly locate the problem; then, based on the feedback from the supervisors, if rework or scrapping is required, an abnormal motor report is generated, recording the personnel, time, losses, etc. in detail; then, based on the abnormal reports within a period of time, the entire production process is statistically analyzed to obtain quality improvement suggestions; finally, an optimized work adjustment plan is formed. Extending the passive early warning response to active quality improvement reflects the quality management concept of prevention first and continuous improvement. Through data accumulation and process optimization, a closed loop is formed from the discovery to the solution of quality problems, which not only makes the information flow smoother, but also makes the system more mature and complete.

[0018] In combination with some embodiments of the first aspect, in some embodiments, based on the cumulative records of abnormal motor reports, the assembly quality of multiple production links is statistically analyzed within a preset time period to obtain quality improvement suggestions, specifically including: correlation mining of abnormal motor reports, production task orders, process parameters and equipment status, and construction of a digital twin model of the entire assembly production process; within a preset time period, the assembly process of multiple production links is simulated and deduced through the digital twin model to determine the quality risks and corresponding impact levels of multiple production links; based on the quality risks, the optimized production process parameters and process control plans are determined to obtain quality improvement suggestions.

[0019] In the above embodiment, the prediction and early warning system first summarizes various types of data to build a digital twin model of assembly production; then uses the model to simulate the assembly process over a period of time to identify the quality risks and impacts of each link; finally, optimizes the production process parameters and process control schemes based on this, and forms improvement suggestions. This quality analysis method with the help of digital twin technology can fully mine and correlate various types of data in the production process, simulate the production status through virtual simulation, and predict quality risks. While shortening the analysis cycle and reducing the analysis cost, the model can also be used to evaluate the effects of different optimization schemes to find the best quality improvement path. Fully combining digital means with quality management makes the manufacturing process more transparent and controllable, and makes quality optimization more scientific and efficient.

[0020] In a second aspect, an embodiment of the present application provides a prediction and warning system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the prediction and warning system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when the computer program product is run on a prediction and warning system, enables the prediction and warning system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions. When the instructions are executed on a prediction and warning system, the prediction and warning system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It is understandable that the prediction and early warning system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiment of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By adopting the method of analyzing the key action characteristics of the target personnel based on real-time video streams and comparing them with standard templates, combined with the assembly tool applicability assessment and real-time monitoring data of the assembly environment, the motor assembly quality score is comprehensively determined. Therefore, the quality risks of the motor assembly process can be comprehensively evaluated from multiple dimensions such as personnel operation, tool use and environmental conditions, which effectively solves the limitation of the existing technology that only relies on product surface defect detection and is difficult to identify assembly quality problems caused by human factors, thereby achieving a more comprehensive, accurate and forward-looking prediction and early warning of defective products. This method not only takes into account the standardization of personnel operations, but also takes into account the appropriateness of tool selection and the influence of environmental parameters. Through multi-source data fusion and intelligent analysis, the causes of defective products are more meticulously excavated, quality risks are eliminated in the bud, and the quality management concept of prevention first is fully utilized.

[0025] 2. As the method of evaluating personal skills based on personnel's historical assembly data and physiological characteristic parameters and focusing on monitoring personnel with low skill levels is adopted, high-risk personnel with weak assembly skills can be identified in advance, and supervision and guidance of their operation process can be strengthened in a targeted manner, which effectively solves the problem of ignoring personnel skill differences in existing technologies, and thus realizes refined management that is tailored to individual needs and combines points and surfaces. Through the personal skill model, this method quantifies factors such as employees' historical performance and physical fitness into skill scores, and then monitors personnel by category based on the scores, which not only saves quality management resources but also improves the accuracy of early warning.

[0026] 3. Due to the adoption of a closed-loop processing flow after the early warning information is issued, from feedback from supervisors to the generation of abnormal reports, and then to the overall quality analysis and process optimization methods, it is possible to respond quickly after the problem is discovered, formulate corrective measures, and learn lessons from them, and continuously improve production systems and processes, effectively solving the shortcomings of the existing technology that the early warning processing is superficial, it is difficult to form a closed loop for problem solving, and quality improvement measures cannot be implemented, thereby achieving the goal of comprehensive quality management. This method closely combines early warning information feedback with production practice to make problem response more efficient. At the same time, the horizontal and vertical flow of information between personnel and departments is smoother, which not only improves the efficiency of coordination, but also improves the quality management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a method for predicting and warning defective products of manually assembled motors in an embodiment of the present application; Figure 2 This is another flow chart of the method for predicting and warning defective products of manually assembled motors in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of a physical device of the prediction and early warning system in the embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0030] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0031] The final assembly workshop of a large motor manufacturing company mainly produces multiple series of brushless motors, reduction motors and stepper motors. Due to the changing market demand and frequent product updates, multiple motor models are produced on mixed lines, and manual assembly processes are widely used. However, with the improvement of users' quality requirements and the intensification of competition in the same industry, the problem of defective products in the assembly process has become increasingly prominent. Statistics show that the monthly losses caused by rework and scrapping due to assembly quality problems are as high as millions of yuan, and delivery delays also occur from time to time.

[0032] In related technologies, supervision and management of assembly quality can be achieved by combining post-sampling inspection and manual inspection. Specifically, destructive or non-destructive testing of products is performed to determine whether the assembly quality is qualified; at the same time, special personnel are arranged to conduct regular or irregular inspections of the assembly site to directly observe the operational norms of assembly employees, and to assess and guide them. The following introduces the scenario of using the defective product prediction and early warning method of manually assembled motors in related technologies.

[0033] Xiao Wang, a process engineer at the technical center of the motor company, is responsible for optimizing the assembly process of stepper motors. He learned that the first-time pass rate of the assembly line was only 92% last year, and thousands of defective products were returned due to quality problems such as damage out of the box and functional failure. After investigation, Xiao Wang found that there were common problems in assembly, such as improper operation and wrong use of tools. For example, when connecting the wiring harness, excessive force was often used, resulting in damage to the interface; or the wrong screwdriver was used, causing the screw to slip.

[0034] The defective product prediction and early warning method for manually assembled motors in the embodiment of the present application can achieve a leap from passive detection to active prevention through intelligent perception of the assembly process, data fusion analysis, and a risk warning feedback loop. This can not only minimize the generation of defective products and save quality costs, but also accurately locate the causes of quality problems, continuously optimize and improve the assembly process, and achieve a dual improvement in product quality and production efficiency.

[0035] The following describes a scenario in which the defective product prediction and early warning method for manually assembled motors in this application is used.

[0036] After the motor company introduced this application plan, Xiao Wang was responsible for its specific implementation. They installed a set of visual perception equipment on the stepper motor production line, which can collect high-definition video of the assembly process in real time. Through the hand movement recognition algorithm, the system can accurately analyze the key features of the assembler's hand trajectory, strength, speed and frequency, and compare it with the pre-entered standard movement template to identify the completion of the standard movement. When the deviation is too large, the system will automatically warn. At the same time, the system is also connected to the digital management system of assembly tools, which can accurately judge the applicability of the tools used by employees. In addition, environmental sensors are added to monitor the temperature, humidity, cleanliness, etc. of the assembly site online. The system comprehensively analyzes the impact of assembly actions, tool applicability and environmental factors on assembly quality, derives the quality risk score and issues timely warnings. After half a year of application, the first-time pass rate of the stepper motor assembly line has increased to 98%, quality problems such as product damage out of the box have been greatly reduced, and the cost of rework has been reduced by more than 20%.

[0037] It can be seen that the defective product prediction and early warning method for manually assembled motors in the embodiment of the present application can not only realize the defective product prediction and early warning, but also effectively solve the quality risk problems caused by extensive management of the assembly process, thereby realizing the full process and full factor control of product quality, and promoting the development of assembly production towards intelligence and refinement, which is of great significance for enhancing the market competitiveness of manufacturing enterprises.

[0038] It should be noted that the prediction and early warning system described in this article refers to a computer with data collection and processing functions. Its specific implementation method is similar to that of a server. For the sake of convenience, it will be abbreviated as a system in some cases.

[0039] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , is a flow chart of a method for predicting and warning defective products of manually assembled motors in an embodiment of the present application.

[0040] S101, obtaining a real-time video stream during the motor assembly process, and determining key action features of a target person according to the real-time video stream.

[0041] The prediction and early warning system deploys high-definition cameras at key locations of the manual assembly production line to collect video image information of the assembly process in real time. The system uses computer vision and image processing technology to analyze the video stream frame by frame, and accurately locates and identifies the hand area of ​​each assembly employee through target detection and tracking algorithms. On this basis, the system uses algorithms such as hand bone key point extraction, trajectory tracking, and motion segmentation to dynamically extract the key motion feature parameters of the assembly employee's hand, mainly including hand movement trajectory, speed, acceleration, finger opening and closing angle and frequency, wrist posture and rotation angle, etc. At the same time, it can also identify the category, posture and other information of the assembly object.

[0042] For example, when assembling a brushless motor, the system can analyze the hand movements of the assembly worker to obtain the hand movement trajectory, speed and other parameters when installing the stator, and determine whether the stator is taken in the standard posture. The motion features extracted by the system form time series data, which provides a basic basis for subsequent assembly quality assessment.

[0043] S102: Compare the key action features with the preset standard action template to obtain the action difference.

[0044] After obtaining the key features of the assembly workers' hand movements, the prediction and early warning system compares and analyzes them with the standard action templates pre-entered into the system to obtain the standardization and standardization level of the action execution. Specifically, the standard action template is a standard action video recorded by process experts and senior employees. The system extracts its action feature parameters as the gold standard for judging the assembly actions of employees. The system uses time series similarity algorithms, such as the dynamic time warping (DTW) algorithm, to align and match the actual action sequence of employees with the standard template, calculate the deviation value of the action feature at each time step, and then comprehensively consider the weights of each deviation value to obtain the comprehensive difference of the entire action process. The difference here can be quantified in percentage form. The greater the difference, the greater the deviation between the actual operation of the employee and the standard action template, that is, the more non-standard the assembly action, the higher the quality risk. Through real-time comparison of employee actions with standard templates, the system can accurately judge the execution of each assembly action, promptly discover non-standard actions, and provide data support for deviation correction.

[0045] For example, when the system finds that an employee's hand movement trajectory differs by as much as 30% from the standard template when installing the planetary gear of a reduction motor, it will determine that the employee is likely to have problems such as inaccurate installation position and excessive movement amplitude, and then trigger an early warning to prompt management personnel to intervene and provide guidance.

[0046] S103: Establish an applicability model of the assembly tool according to the tool type, performance parameters and usage status information of the assembly tool used by the target personnel.

[0047] In the process of motor assembly, the selection and use of assembly tools have a direct impact on product quality. To this end, the prediction and early warning system comprehensively considers factors such as the type, performance parameters and usage status of assembly tools to establish a tool suitability evaluation model. First, the system needs to obtain and enter static attribute parameters such as model specifications, accuracy level, range, validity period, etc. of different types of assembly tools (such as screwdrivers, socket wrenches, jacks, etc.). Secondly, the system needs to collect the usage status data of assembly tools in real time, mainly including tool collection records, usage frequency, cumulative usage time, regular inspection and calibration records, etc. After obtaining the above tool attributes and status data, the system uses machine learning algorithms (such as support vector machine SVM, random forest, etc.) to establish a tool suitability evaluation model, that is, a correlation model between various tool parameters and assembly quality. The model can output a tool suitability score to quantitatively evaluate the rationality of tool selection and use.

[0048] For example, by analyzing the usage history data of a socket wrench, it was found that its cumulative usage time has exceeded 1,000 hours, and it has been used for high-intensity tightening operations for a long time, and its accuracy level is also low. Therefore, it is judged that its applicability score is low, and continued use may damage the thread and cause torque leakage. When the system determines based on the model that the applicability score of the tool used by the assembly employee is lower than the preset threshold, an early warning prompt is issued to remind that the tool needs to be replaced or maintained.

[0049] The input data of the suitability model include the tool type (such as power tools and hand tools), performance parameters (such as rated power, speed, torque, material hardness and size specifications) and usage status information (such as the actual use time, maintenance records and failure frequency of the tools) of the assembly tools used by the target personnel. By analyzing these data, the model can output the complexity value, application scope and reliability value of the assembly tools. Finally, combining the values ​​of these three dimensions, the model can give a three-dimensional suitability evaluation result of the assembly tools. The model is trained using machine learning algorithms such as support vector machines (SVM) and random forests. The training standard is to enable the model to accurately evaluate the rationality of tool selection and use.

[0050] S104. Input key action features into the applicability model to obtain the tool application score of the target person.

[0051] After establishing the assembly tool suitability evaluation model, the prediction and early warning system inputs the key action feature parameters of the assembly workers into the model to automatically evaluate the tool application of the workers. Specifically, the system focuses on the hand action features related to tool application, such as the size and direction of wrist force, the position and angle of finger grip, etc. The system matches and analyzes these action feature data with the tool model specifications, accuracy level and other attribute parameters to determine whether the employees use the tools in accordance with the requirements of the work instructions and use the correct operation methods. At the same time, the system will also evaluate the proficiency of the employees in using tools, judging by the coherence and stability of hand movements and other characteristics. After comprehensively considering the appropriateness of tool selection and the standardization of use methods, the model will output the tool application score of the employee, which quantitatively reflects the rationality and standardization of his tool use. The scoring results can be expressed in percentage or in a graded system (such as A, B, C, D, etc.). When the employee's tool application score is low, the system will warn that there may be problems such as wrong tool selection and improper operation methods, and the management personnel need to provide necessary guidance and training.

[0052] S105, obtaining real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process, and determining the motor assembly score of the target personnel based on the action difference, real-time pressure data, real-time temperature data and tool application score.

[0053] Fluctuations in the environmental conditions of assembly operations can have an impact on product quality. Therefore, the prediction and early warning system also collects real-time pressure and temperature data of the assembly workshop environment as an auxiliary reference for judging assembly quality. By deploying wireless pressure sensors and temperature sensors at the assembly site and establishing a communication connection with the data acquisition terminal, the system can continuously monitor and record changes in pressure and temperature in the assembly environment. After obtaining the environmental parameter data, the system correlates it with indicators such as the difference in the assembly workers' movements and tool application scores to construct a comprehensive quality assessment model for the assembly process. The model uses a weighted scoring method to assign different weight coefficients to each indicator, and ultimately derives a comprehensive quality score for the employee's assembly operation, namely the motor assembly score.

[0054] For example, the system assigns a 40% weight to the action difference, a 30% weight to the tool application score, and a 15% weight to the environmental pressure and temperature data. When an employee assembles the rotor of a stepper motor, the action difference is 10%, the tool application score is B, and the pressure in the assembly room is 101kPa and the temperature is 25°C, both within a reasonable range. After weighted calculation, the employee's motor assembly score is 85 points, indicating that the assembly quality is good and the product is likely to be qualified. By collecting multi-source heterogeneous data of the assembly process and conducting comprehensive evaluation, the system can monitor the assembly quality status more comprehensively and dynamically, reducing evaluation blind spots.

[0055] S106: When the motor assembly score is lower than a preset qualified threshold, a warning message is sent to the supervision client.

[0056] According to the motor assembly scores of assembly employees, the prediction and early warning system can determine the quality risk level of the assembly process in real time and decide whether to issue an early warning message. Specifically, the system pre-sets the qualified threshold of the motor assembly score (such as 80 points). When the employee's score is lower than the threshold, the system will consider that the motor is at risk of failure and requires early warning prompts for timely intervention. The early warning information will be automatically pushed to the mobile APP or PC software of the workshop supervisor, prompting abnormal assembly quality in the form of pop-up reminders, voice broadcasts, flashing signal lights, etc. At the same time, the system will also push the early warning information together with the video clips of the assembly process, so that supervisors can intuitively see the problem and respond quickly. Of course, for general warnings with scores slightly lower than the threshold, the system will only prompt attention, and for serious warnings with scores much lower than the threshold, the system will require the assembly operation to be stopped immediately, the motor to be fully inspected, and the assembly employees to be retrained.

[0057] For example, when an employee assembles the encoder of a servo motor and the motor assembly score is only 65 points, the system will immediately send an early warning message to the workshop director, and attach a video clip of the employee assembling the encoder with problems such as collision, insufficient tightening torque, etc., requiring the motor to be re-inspected and criticizing the employee. Through timely early warning feedback, the system can eliminate quality risks in the bud, minimize the flow of defective products into the next process, and truly realize the forward shift of quality management.

[0058] In the above embodiment, by combining visual perception, sensor detection and algorithm analysis, the human actions, tool applicability and environmental impact of the assembly process are evaluated in real time to fully perceive the assembly quality risks. In practical applications, the application scenarios and effects of this method can be further expanded by introducing functional modules such as human physiological state detection, production line production plan identification, and digital twin simulation.

[0059] For example, the employee fatigue assessment model can be used to determine the impact of employee fatigue status on assembly quality. Through assembly line production planning management, the quality standards and assembly requirements of different products can be distinguished. For example, digital twin technology can be used to predict assembly difficulty and quality risks in advance during the product design and process planning stages.

[0060] The following is a supplement to the scenario of this embodiment.

[0061] In view of the good application effect of the stepper motor assembly line, the motor company decided to promote the system in all motor assembly production lines. However, considering the differences in the characteristics of different production lines, further optimization is needed. For example, in the brushless motor assembly line, it is necessary to focus on the fatigue state of the assembly employees, because this is related to key quality characteristics such as coil winding tension. Therefore, Xiao Wang's team added an employee fatigue assessment model to the system. By combining factors such as the employee's age, length of service, and length of work on duty, the degree of fatigue is judged and associated with the performance of their assembly actions, so as to more comprehensively assess the risks of the assembly process. For example, in the reduction motor assembly line, the production status of different motor models needs to be monitored separately.

[0062] The team further developed a production line production plan management module that can automatically identify the motor model currently produced by the production line and call the corresponding quality standards and process parameters to achieve accurate evaluation of the assembly process of different motor models. Customized intelligent reports have also been developed for different production lines to present the quality trends and risk distribution in the assembly process in a visual way, helping managers to better understand the quality status and optimize production organization.

[0063] In the production preparation stage of the new servo motor project, the system can also simulate the assembly process through process simulation, discover design and process defects in advance, and prevent quality risks from an earlier stage. The continuous optimization and expanded application of the system have laid a solid foundation for the company to achieve quality control of the entire value chain and strongly supported the high-quality development of the enterprise.

[0064] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the defective product prediction and early warning method for manually assembled motors in an embodiment of the present application.

[0065] S201, obtaining a real-time video stream during the motor assembly process, and determining key action features of a target person based on the real-time video stream.

[0066] Referring to step S101 , the prediction and warning system determines the key action features of the target person.

[0067] In some embodiments, the prediction and warning system obtains the historical assembly data of the target personnel; the historical assembly data includes the assembly time, assembly qualification rate and defective causes of multiple products; the physiological characteristic parameters of the target personnel are obtained; the physiological characteristic parameters include gender, age, vision and reaction speed; the historical assembly data and physiological characteristic parameters are input into a preset personal skill model to obtain the assembly skill level of the target personnel; the target personnel whose assembly skill level is lower than the preset level threshold are identified as key monitoring objects, and the video acquisition frequency and motion recognition accuracy of the key monitoring objects are improved.

[0068] The input data of the personal skill model includes the historical assembly data of the target personnel (such as the assembly time, assembly qualification rate and reasons for defects of multiple products) and physiological characteristic parameters (such as gender, age, vision and reaction speed). By analyzing these data, the model can output the assembly skill level of the target personnel. The model will mark personnel whose assembly skill level is lower than the preset threshold as key monitoring objects. The model fully considers the historical performance and personal characteristics of employees, and through the training of machine learning algorithms, it can judge the actual operation ability of employees more objectively and dynamically. The training standard is to enable the model to accurately identify high-risk personnel with weak assembly skills.

[0069] Specifically, the prediction and early warning system will further collect and analyze the historical work data and personal physiological characteristics of assembly workers to more comprehensively evaluate their assembly skill level and quality risk level. First, the system will extract the employee's historical assembly data from production management systems such as MES and ERP, mainly including the number of products they participated in assembling, the assembly time of each product, the first-time qualified assembly rate, and the main reasons for defective products. Through statistical analysis of historical data, the system can preliminarily judge the assembly efficiency and quality stability of the employee.

[0070] For example, if the data shows that it takes an employee an average of 30 minutes to assemble a product, and the qualified rate of assembly is 98%, but occasionally the product is scrapped due to operational errors, it can be preliminarily determined that the employee has relatively skilled skills, but the control of details needs to be strengthened. Secondly, the system will also collect key physiological characteristic parameters of employees, such as gender, age, vision, reaction speed, etc., as factors for evaluating their assembly skills. Among them, vision can be obtained through the employee's annual physical examination report, and reaction speed can be tested by the human-computer interaction game designed by the system. By comparing the differences in assembly data of employees under different physiological conditions, we can deeply study the influence of human factors on assembly quality. After obtaining the above two types of data, the system inputs them into the personal skill evaluation model built based on the machine learning algorithm to obtain the comprehensive skill level score of the employee. The model fully considers the employee's historical performance and personal characteristics, and can judge the employee's actual operation ability more objectively and dynamically. When the model determines that an employee's assembly skill level is lower than the preset threshold, the system will mark it as a key monitoring object, and will subsequently increase the video acquisition frequency and action recognition accuracy of the employee's assembly process to more quickly and accurately warn of its assembly quality risks. At the same time, the system will also push the employee's skill assessment report to the workshop supervisor, suggesting that he or she strengthen the employee's skill training and work guidance.

[0071] In some embodiments, the prediction and warning system obtains the facial feature image and motion feature image of the target person through a camera, and evaluates the target person's vision level, concentration and fatigue based on the facial feature image and the motion feature image; determines the target person's hand dexterity based on the motion feature image; integrates the target person's vision level, concentration, fatigue and hand dexterity to obtain physiological feature parameters.

[0072] Specifically, the prediction and early warning system will also use advanced computer vision technology to evaluate the physiological state of employees in real time through image analysis of their facial and motion features, so as to obtain human factor data more dynamically and objectively. Specifically, the system uses high-definition cameras on the assembly line to collect close-up images of employees' faces and full-body motion images in real time, and uses algorithms such as face recognition and human posture estimation to extract and analyze features. First, for facial feature images, the system focuses on analyzing the employee's eye area and evaluates their vision level through indicators such as pupil diameter, blinking frequency, and eye movement trajectory. For example, if it is detected that the employee's pupil diameter is in an enlarged state for a long time and the blinking frequency is less than 15 times per minute, it may indicate that their vision is impaired or fatigue is obvious. At the same time, the system will also analyze the degree of concentration of the employee's facial expression, and judge whether the employee is fully focused through indicators such as the position of the eyebrows, the curvature of the corners of the mouth, and the tension of the facial muscles. For example, when an employee frowns and his eyes are dull, the system can determine that he is not paying attention. In addition, the overall fatigue state of the employee's face can also be estimated through indicators such as the size of dark circles, the degree of relaxation of eye bags, and the dullness of the skin. The hand flexibility of employees is mainly extracted through motion feature images. The system focuses on analyzing indicators such as the flexion and extension range of fingers, the rotation angle of wrists, and the continuity of hand movements. For example, when it is detected that the employee's fingers are stiff and bent, and the movements are slow, it can be determined that the flexibility is poor. By comprehensively analyzing the evaluation results of employees' facial and movement features, the system can dynamically judge the physiological state of employees. When it is found that their eyesight is poor, their attention is not focused, their hand flexibility is poor, or their fatigue level is high, it will generate corresponding early warning prompts and recommend that managers pay attention. At the same time, the system will also associate the evaluation results of the employee's physiological state with their historical work data for dynamic updating of the employee's skill level model, so that it can be closer to the employee's real-time status. Through the multi-dimensional employee status perception of human-machine collaboration, the prediction and early warning system can more finely control the human factor quality risks and achieve dynamic optimization of assembly process management.

[0073] S202: Compare the key action features with the preset standard action template to obtain the action difference.

[0074] Referring to step S102 , the prediction and warning system determines the action difference.

[0075] S203: Establish an applicability model of the assembly tool according to the tool type, performance parameters and usage status information of the assembly tool used by the target personnel.

[0076] Referring to step S103 , the prediction and early warning system will establish a suitability model for the assembly tool.

[0077] In some embodiments, the prediction and warning system will obtain the tool type of assembly tools used by the target personnel; the tool types include power tools and hand tools; obtain the performance parameters of the assembly tools; the performance parameters include rated power, speed, torque, material hardness and size specifications; obtain the usage status information of the assembly tools; the usage status information includes the actual usage time of the tools, maintenance records and failure frequency; divide the complexity value of the assembly tools according to the tool type, determine the applicable scope of the assembly tools according to the performance parameters, and determine the reliability value of the assembly tools according to the usage status information; comprehensively consider the complexity value, applicable scope and reliability value of the assembly tools to establish a three-dimensional applicability evaluation model of the assembly tools as a applicability model.

[0078] Specifically, the prediction and early warning system will also deeply analyze the various attribute parameters of the tools used by assembly employees, and build a tool suitability assessment model to determine whether employees can skillfully and correctly operate the tools in their hands, so as to further predict their assembly quality risks. First, the system will automatically obtain the type information of their handheld tools through the RFID tags worn by employees or visual recognition technology, which are mainly divided into two categories: power tools (such as electric drills, electric wrenches, etc.) and hand tools (such as screwdrivers, socket wrenches, etc.). Generally speaking, power tools are considered complex tools due to their high operating complexity and high risk level; while hand tools are relatively simple. At the same time, the system will also collect key performance parameters of various assembly tools, such as the rated power, no-load speed, maximum torque, etc. of power tools, as well as the material hardness, size specifications, etc. of hand tools. These parameters determine the difficulty and precision requirements of the assembly tasks that the tools can handle, and also put forward corresponding requirements for the skill level of employees.

[0079] For example, an electric wrench with a higher torque is more suitable for assembling large and heavy parts, which requires employees to have high strength control and stability; while a small flat-blade screwdriver is more suitable for assembling precision electronic components, which requires employees to have higher hand-eye coordination. In addition, the system will dynamically track the usage status data of each tool, including the cumulative usage time, maintenance records, number of failures, etc., which to a certain extent reflects the reliability of the tool in the assembly task.

[0080] When a tool has a high failure frequency, it may cause assembly task interruption or quality abnormality. Based on the above tool data, the system constructs a three-dimensional tool suitability evaluation model. The model starts from the three dimensions of tool complexity, tool application scope and tool reliability, uses the hierarchical analysis method to determine the weight of each indicator, and finally calculates the comprehensive score of the suitability of each tool.

[0081] When an employee holds a tool for assembly work, the system will compare its motion characteristics with the tool's suitability score in real time. If the two do not match, a quality warning will be generated and the risk will be prompted. For example, when an employee is accustomed to using a high-torque electric wrench to assemble small screws, the system will immediately warn the employee that there may be problems with improper force control. For another example, when an employee is assigned to a brand new assembly station and faces some tools that have never been used, the system will prompt him to operate with caution to avoid quality accidents caused by tool incompatibility. Through the construction of a tool suitability evaluation model, the prediction and warning system can consider the degree of match between employees and tools as an important influencing factor of assembly quality, and carry out targeted employee skills training and tool management optimization based on this, ultimately achieving a deep integration of employees, tools, and processes, and improving the overall quality level of assembly operations.

[0082] S204: Input key action features into the applicability model to obtain the tool application score of the target person.

[0083] Referring to step S104, the prediction and early warning system determines the tool application score.

[0084] S205, obtaining real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process, and determining the motor assembly score of the target personnel based on the action difference, real-time pressure data, real-time temperature data and tool application score.

[0085] Referring to step S105 , the prediction and warning system determines the motor assembly score.

[0086] In some embodiments, the prediction and warning system will detect the real-time pressure data of the assembly environment through a pressure sensor, and compare the real-time pressure data with a preset standard pressure range to obtain the pressure difference; detect the real-time temperature data of the assembly environment through a temperature sensor, and compare the real-time temperature data with a preset standard temperature range to obtain the temperature difference; input the motion difference, pressure difference, temperature difference and tool application score into a preset motor quality assessment model to obtain the motor assembly score of the target personnel.

[0087] The input data of the motor quality assessment model include the difference in the assembly workers' movements, tool application scores, and real-time pressure and temperature data of the assembly environment. The model uses a weighted scoring method to assign different weight coefficients to each indicator, and finally outputs the comprehensive quality score of the employee's motor assembly. The higher the score, the greater the probability that the employee will assemble a qualified product under the current circumstances. The model uses a decision tree algorithm and is trained through a large amount of historical assembly data. The training standard is to enable it to accurately predict the impact of various combinations of factors on motor quality.

[0088] Specifically, the prediction and warning system will also sense the pressure and temperature parameters of the assembly environment in real time, and take them into consideration as key environmental factors affecting the quality of motor assembly. First, the system collects the air pressure data of the local environment in real time through high-precision pressure sensors installed around the assembly station, and compares it with the preset standard pressure range (such as 1 atmosphere ± 0.1 atmosphere) to calculate the pressure difference. When the ambient pressure is too high or too low, it may affect the normal progress of certain assembly processes, such as glue curing and vacuum adsorption. For example, when assembling highly sealed electronic equipment, if the ambient pressure fluctuates too much, it may cause the internal components to become damp and fail. At the same time, the system will also monitor the temperature distribution of the assembly room in real time through a distributed temperature sensor network, especially some local areas that are more sensitive to temperature. When the actual temperature exceeds the preset standard range (such as 25 ± 5 ° C), the system will calculate the temperature difference and warn of possible quality risks.

[0089] For example, during the assembly of a servo motor, if the ambient temperature in the stator winding area is too high, the winding insulation material will age or break down, causing the motor performance to deteriorate. Therefore, real-time monitoring of the assembly environment temperature is crucial for quality control of certain precision assembly processes. After obtaining the pressure difference and temperature difference, the system will input them into the pre-trained motor quality assessment model together with the employee's action difference, tool application score and other factors. The model uses a decision tree algorithm, and through training with a large amount of historical assembly data, it can accurately predict the impact of various factor combinations on motor quality. The output of the model is the employee's motor assembly score. The higher the score, the greater the probability that the employee will assemble a qualified product under the current environmental conditions. Once the score is lower than the preset threshold, the system will issue a quality warning to relevant employees and managers, indicating that there is a risk of abnormal pressure or temperature in the assembly environment, and recommend taking measures to adjust it, such as turning on the dehumidification device, adjusting the air conditioning temperature, etc. At the same time, the system will dynamically adjust the quality inspection frequency and inspection items of the affected process to timely discover and isolate problem products and prevent unqualified products from flowing into the next process. By incorporating the assembly environment into the scope of motor quality assessment, the predictive warning system can more comprehensively and dynamically monitor the various factors that affect the assembly quality, so that high-quality products can be stably output even in a complex and changeable workshop environment.

[0090] S206: When the motor assembly score is lower than a preset qualified threshold, a warning message is sent to the supervision client.

[0091] Referring to step S106 , the prediction and warning system will send a warning message to the supervision client when the motor assembly score is lower than the preset qualified threshold.

[0092] S207: Push the assembly action video clip of the target person in the warning information, and determine the feedback returned by the supervision client after receiving the warning information.

[0093] When the prediction and early warning system determines that the motor assembly score of an assembly employee is lower than the preset threshold, it will automatically trigger the push of the early warning message. At the same time, the system will also intelligently extract the key video clips of the employee's assembly of the current motor and send them together with the early warning message to the client device of the workshop supervisor. These video clips are usually 10-30 seconds long, focusing on the details of the employee's assembly operation, such as irregular actions and improper use of tools, which are prone to quality problems. Through the video screen, the supervisor can intuitively see the triggering reason of the early warning, quickly judge the motor quality risk level, and whether the employee's operation needs to be corrected. At the same time, the system also supports supervisors to view the real-time video monitoring of the assembly station at any time through the client, and can fast forward, replay, pause and other controls on the video. After receiving the early warning information, the supervisor needs to go to the assembly site for verification and feedback the disposal opinions on the client, mainly including: the motor needs to be reworked, the motor needs to be scrapped, the motor is temporarily shelved, the motor quality is qualified and the assembly continues, etc. The system will automatically collect and record the feedback of the supervisor for subsequent responsibility tracing and performance appraisal.

[0094] For example, when the system finds that an employee has multiple problems with incorrect magnetic steel assembly angles when assembling the rotor of a stepper motor, it will send an early warning message to the production supervisor and quality supervisor, and attach a video clip of the employee's hand movements when assembling the magnetic steel. After receiving the early warning, the two supervisors immediately went to the assembly line to check after viewing the video, and finally determined that the rotor needed to be reworked and criticized and educated the employee.

[0095] S208. If the feedback is that the motor needs to be reworked or scrapped, an abnormal motor report is generated that records the cause of the abnormality, the person responsible, and the lost working time.

[0096] When the supervisor selects "motor needs to be reworked" or "motor needs to be scrapped" in the feedback of the early warning information, the prediction and early warning system automatically generates an abnormal motor report. The report records in detail the reasons for the abnormal motor quality, the information of the responsible persons such as the assembly staff, the lost working hours caused by rework or scrapping, etc. in the form of a document. The abnormal reasons mainly refer to the problems prompted by the system in the early warning information, such as irregular assembly movements, improper use of tools, interference in the assembly environment, etc., and other problems found by the supervisor during the on-site inspection will also be supplemented. The responsible person information includes the name, work number, and team of the relevant assembly employee, as well as the name and position of the supervisor such as the workshop supervisor and quality supervisor. The lost working hours mainly refer to the time cost of rework or scrapping, which requires comprehensive consideration of factors such as the number of motors to be reworked or scrapped, the assembly time of a single motor, and the time for rework or scrapping. The system automatically estimates the lost working hours based on the preset empirical formula and converts them into economic costs.

[0097] For example, according to the abnormal motor report generated by the system, it can be seen that the photoelectric encoder of a servo motor needs to be reworked due to improper operation of the assembly staff. After preliminary calculation, the rework takes about 50 minutes, which is equivalent to a loss of 0.83 working hours, which is equivalent to an economic loss of about 150 yuan. By generating abnormal motor reports, the processing process and results of each quality warning event can be recorded in detail, which is convenient for subsequent responsibility tracing and can also provide data reference for management work such as performance appraisal and cost accounting.

[0098] S209. Based on the accumulated records of abnormal motor reports, statistical analysis is performed on the assembly quality of multiple production links within a preset time period to obtain quality improvement suggestions.

[0099] The prediction and early warning system will continuously track and record each abnormal motor event, regularly conduct statistical analysis on these data, and form an assembly quality improvement report. Usually a report is formed once a week or a month, focusing on analyzing the occurrence of quality warnings in each production link during the time period, including the number and frequency of warning events, the assembly processes and assembly employees involved, the main causes of warnings, the extent of losses, etc. After comparing and analyzing the data of different production links, the system uses the expert experience knowledge base and the weight distribution of the warning reasons to output targeted suggestions for assembly quality improvement. These suggestions may involve optimizing assembly process flow, improving assembly tooling and fixtures, strengthening employee skills training, and improving workshop management systems.

[0100] For example, the monthly quality improvement report shows that the stator winding process of the brushless motor assembly line has a high number of quality warnings. The main reasons are that the assembly workers lack skills, the operation time is long, and the body fatigue leads to deformation of the movement. In response to this situation, the system recommends optimizing the assembly procedures of the winding process to reduce workers' repetitive labor; improving winding tooling to increase the degree of automation; increasing personnel skills training and rotation frequency; improving working hours quota and reward and punishment system, etc. Through systematic mining of the causes of quality problems in each link, the prediction and warning system can output highly operational quality improvement measures to help assembly workshops achieve refined and streamlined quality management.

[0101] In some embodiments, the prediction and warning system will conduct correlation mining on abnormal motor reports, production task orders, process parameters and equipment status to build a digital twin model of the entire assembly production process; within a preset time period, the assembly process of multiple production links will be simulated and deduced through the digital twin model to determine the quality risks and corresponding impact levels of multiple production links; based on the quality risks, the optimized production process parameters and process control plans will be determined to obtain quality improvement suggestions.

[0102] The digital twin model uses virtual simulation, 3D modeling and other technologies to construct a digital mapping of the assembly workshop, including personnel, equipment, materials, environment and other elements. The input data of the model includes abnormal motor reports, production task orders, process parameter settings and equipment status data. Through the simulation and deduction of the model, the quality risk level of each production link and the degree of impact on the whole can be output, and various optimization schemes can be simulated and evaluated, and finally actionable quality improvement suggestions can be output. The training of the model follows the rules in the expert experience knowledge base, and the standard is to enable it to accurately predict quality risks through simulation and provide effective response plans.

[0103] Specifically, the prediction and early warning system will also achieve panoramic quality management and predictive optimization of the production process by building a digital twin model of the entire assembly production process. First, the system will use big data mining technology to conduct correlation analysis on abnormal motor reports, production task orders, process parameter settings, and equipment status data in various production management systems to mine the key factors affecting production quality and their interaction rules. On this basis, the system uses virtual simulation, 3D modeling and other technologies to build a digital twin model of the assembly workshop, including digital mapping of various elements such as personnel, equipment, materials, and environment. Through this model, all aspects of assembly production can be synchronized and highly restored in real time in the virtual space. Based on the digital twin model, the system can simulate the workshop production process within a preset time period (such as the next week). By setting different production tasks and process parameter combination schemes, the system can fully simulate the production process under each scheme and evaluate the quality risk level of each production link and the degree of impact on the overall production. For example, when the production task volume of a batch of products is large and the process difficulty is high, the system simulation finds that the assembly, debugging, and testing links are facing greater quality risks. Among them, the risk of increased defective rate due to insufficient employee skills in the assembly link is the most prominent, which may lead to delayed delivery of production tasks. By summarizing the quality risk assessment results of each link, the system can identify the bottlenecks that restrict the improvement of production quality. On this basis, the system will automatically search for solutions to quality risks, such as optimizing the production task sequence, adjusting process parameters, adding quality inspection points, etc., and verify the improvement effect of each solution through the simulation model, and finally select the optimal process control strategy combination to form an actionable quality improvement suggestion for managers to refer to and make decisions.

[0104] For example, in response to the skill risks in the above-mentioned assembly process, the system recommends taking measures such as conducting employee skill training in advance, reasonably allocating task saturation, and increasing the frequency of first-piece inspection. After simulation verification, the employee defect rate can be reduced by 2 percentage points, and the probability of on-time delivery of production tasks can be increased to more than 95%. Through simulation deduction and optimized decision-making driven by digital twins, the prediction and early warning system can transform passive post-event quality management into active pre-event risk prediction and process optimization, while ensuring product quality, achieving continuous improvement in production organization methods, and ultimately bringing about a dual improvement in quality and efficiency. In the future, digital twin technology is expected to become an important means of assembly production quality management in the era of intelligent manufacturing.

[0105] S210. Generate a work adjustment plan including workstation layout and employee skill requirements.

[0106] On the basis of generating suggestions for improving assembly quality, the prediction and early warning system can further design and form an employee work adjustment plan. The plan mainly re-evaluates and adjusts the employee's current workstation based on the employee's motor assembly score and abnormal record to match his actual work skill level. At the same time, the plan will also put forward targeted training requirements for the employee's skill shortcomings to promote the improvement of his comprehensive ability. When formulating the adjustment plan, the system mainly considers the following factors: the employee's learning ability and skill improvement potential, the production capacity balance and beat balance requirements of the assembly line, the employee's rationalization suggestions and willingness to transfer positions, etc. Based on the transfer rules in the expert knowledge base, the system uses combined optimization methods such as the Hungarian algorithm and genetic algorithm to optimally match employees and workstations to form a set of feasible work adjustment plans. For example, according to the adjustment plan generated by the system, employee A, who originally manually assembled parts, was not flexible enough and frequently had quality abnormalities. It was recommended that he be transferred to the product packaging station; while employee B had significantly improved the assembly quality and efficiency after several rounds of training. It was recommended that he be promoted to the assembly team leader to be responsible for teaching new employees. At the same time, the adjustment plan generated by the system also provides decision-making references for managers, enabling them to break away from empirical and arbitrary personnel management and truly achieve teaching according to students' aptitude and matching people with positions.

[0107] In the embodiment of the present application, due to the introduction of a multi-dimensional quality prediction model that integrates historical assembly data, employee physiological characteristics, assembly tool properties, and environmental parameters, and the use of digital twin technology to simulate and evaluate the entire production process, a set of assembly production intelligent quality management system integrating data collection, risk warning, and optimized decision-making is constructed, which effectively solves the problems of lagging preventive measures, incomplete analysis of influencing factors, and lack of quantitative verification of improvement plans in the traditional assembly quality management model, thereby realizing the transformation from passive detection to active prevention, and moving assembly quality management forward to all links of the production process, so as to prevent quality problems before they occur, significantly improving the quality level and operational efficiency of assembly production, and providing strong support for lean production and agile management in the era of intelligent manufacturing.

[0108] The following describes the prediction and early warning system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the structure of a physical device of the prediction and early warning system in an embodiment of the present application.

[0109] It should be noted that Figure 3 The structure of the prediction and warning system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0110] like Figure 3As shown, the prediction and warning system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0111] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0112] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0113] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0114] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0115] Specifically, the prediction and warning system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the defective product prediction and warning method for manually assembled motors provided in the above embodiment is implemented.

[0116] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the prediction and warning system described in the above embodiment; or may exist independently without being assembled into the prediction and warning system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the prediction and warning system, the prediction and warning system implements the defective prediction and warning method for manually assembled motors provided in the above embodiment.

[0117] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0118] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0119] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A method for predicting and warning defective products of manually assembled motors, characterized in that: Applied to a prediction and early warning system, the method comprises: Acquire a real-time video stream during the motor assembly process, and determine key motion features of the target person based on the real-time video stream; the key motion features include the hand motion trajectory, strength, speed and frequency of the target person; Comparing the key action features with a preset standard action template to obtain the action difference; Establishing a suitability model of the assembly tool according to the tool type, performance parameters and usage status information of the assembly tool used by the target personnel; Inputting the key action features into the applicability model to obtain the tool application score of the target person; Acquire real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process, and determine the motor assembly score of the target person according to the action difference, the real-time pressure data, the real-time temperature data and the tool application score; When the motor assembly score is lower than a preset qualified threshold, a warning message is sent to the supervision client.

2. The method according to claim 1, characterized in that The establishing of the applicability model of the assembly tool according to the tool type, performance parameters and usage status information of the assembly tool used by the target personnel specifically includes: Obtaining the tool type of the assembly tool used by the target person; the tool type includes power tools and hand tools; Obtaining performance parameters of the assembly tool; the performance parameters include rated power, rotation speed, torque, material hardness and size specifications; Acquiring usage status information of the assembly tool; the usage status information includes actual usage time, maintenance records and failure frequency of the tool; Classifying the complexity values ​​of the assembly tools according to the tool types, determining the applicable scope of the assembly tools according to the performance parameters, and determining the reliability values ​​of the assembly tools according to the usage status information; A three-dimensional applicability evaluation model of the assembly tool is established as a applicability model by combining the complexity value, application scope and reliability value of the assembly tool.

3. The method according to claim 1, characterized in that The acquiring of real-time pressure data and real-time temperature data of the assembly environment during the motor assembly process, and determining the motor assembly score of the target person according to the action difference, the real-time pressure data, the real-time temperature data and the tool application score, specifically includes: The real-time pressure data of the assembly environment is detected by a pressure sensor in real time, and the real-time pressure data is compared with a preset standard pressure range to obtain a pressure difference; The real-time temperature data of the assembly environment is detected by a temperature sensor in real time, and the real-time temperature data is compared with a preset standard temperature range to obtain a temperature difference; The action difference, the pressure difference, the temperature difference and the tool application score are input into a preset motor quality assessment model to obtain the motor assembly score of the target person.

4. The method according to claim 1, characterized in that: Before the step of acquiring the real-time video stream during the motor assembly process and determining the key action features of the target person according to the real-time video stream, the method further includes: Obtaining historical assembly data of a target person; the historical assembly data includes assembly time, assembly qualification rate and defect reasons of multiple products; Acquiring physiological characteristic parameters of the target person; the physiological characteristic parameters include gender, age, vision and reaction speed; Inputting the historical assembly data and the physiological characteristic parameters into a preset personal skill model to obtain the assembly skill level of the target person; The target personnel whose assembly skill level is lower than the preset level threshold are determined as key monitoring objects, and the video acquisition frequency and action recognition accuracy of the key monitoring objects are improved.

5. The method according to claim 4, characterized in that The step of obtaining the physiological characteristic parameters of the target person specifically includes: Acquire a facial feature image and a motion feature image of a target person through a camera, and evaluate the target person's vision level, concentration, and fatigue level based on the facial feature image and the motion feature image; Determining the hand dexterity of the target person according to the action feature image; The vision level, concentration, fatigue level and hand dexterity of the target person are integrated to obtain physiological characteristic parameters.

6. The method according to claim 1, characterized in that After the step of sending a warning message to a supervision client when the motor assembly score is lower than a preset qualified threshold, the method further includes: Pushing the assembly action video clip of the target person in the warning information, and determining the feedback returned by the supervision client after receiving the warning information; If the feedback is that the motor needs to be reworked or scrapped, an abnormal motor report is generated that records the cause of the abnormality, the person responsible, and the lost working time; Based on the accumulated records of the abnormal motor reports, statistical analysis is performed on the assembly quality of multiple production links within a preset time period to obtain quality improvement suggestions; Generate work adjustment plans including workstation layout and employee skill requirements.

7. The method according to claim 6, characterized in that According to the accumulated records of the abnormal motor reports, statistical analysis is performed on the assembly quality of multiple production links within a preset time period to obtain quality improvement suggestions, specifically including: Conduct correlation mining on abnormal motor reports, production task orders, process parameters and equipment status to build a digital twin model of the entire assembly production process; Within a preset time period, the assembly process of multiple production links is simulated and deduced through the digital twin model to determine the quality risks of the multiple production links and the corresponding impact levels; Based on the quality risks, the optimized production process parameters and process control solutions are determined to obtain quality improvement suggestions.

8. A prediction and early warning system, characterized in that: The prediction and warning system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the prediction and warning system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a prediction and warning system, the prediction and warning system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a prediction and early warning system, the prediction and early warning system is enabled to execute the method according to any one of claims 1 to 7.

Images