Washing machine assembly production line performance monitoring system based on digital twinning
Through the washing machine assembly line performance monitoring system based on digital twins, the multi-type and multi-level models are used to solve the problems of workshop clogging and low production efficiency, and efficient and transparent management and production optimization are achieved, and production efficiency and quality are improved.
Patent Information
- Application Number
- CN202410997701.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-12
AI Technical Summary
The washing machine assembly production line often blocks and shuts down due to man-machine mixing and various small batches of production, resulting in the workshop being frequently blocked and shut down, resulting in reduced production efficiency and increased costs. It is difficult for the existing technology to effectively monitor and manage workshop performance, resulting in poor production efficiency and quality.
Using a washing machine assembly line performance monitoring system based on digital twins, including data acquisition, performance analysis, performance prediction and performance simulation modules, it realizes modularization and segmentation through multi-type and multi-level models (analysis model, simulation model, data-driven model) Layer management, positioning workshop blocked sites, predict performance changes, simulate production plans, and achieve efficient and transparent real-time management.
It realizes efficient and transparent real-time management of the washing machine assembly production line, improves production efficiency and quality, reduces workshop congestion and downtime, optimizes production plans, and reduces costs.
Smart Images

Figure CN118938830A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of real-time monitoring of workshop performance, and more specifically, to a washing machine assembly production line performance monitoring system based on digital twins. Background Art
[0002] The washing machine assembly production line is characterized by mixed flow of people and machines, and small batches of multiple varieties. In the washing machine assembly workshop, the workshop system is often blocked or even shut down due to problems such as frequent personnel transfers, personnel operation errors, missing spare parts, and equipment failures. Frequent product switching may lead to reduced production efficiency, and multi-variety production may increase the complexity of material and resource use, leading to increased costs. Therefore, in the face of complex production fluctuations and changes in product demand, performance monitoring of the workshop production system becomes crucial. Real-time supervision of the production line can effectively improve system performance, productivity and quality.
[0003] A digital twin is a high-fidelity virtual model of a physical system. As one of the important technologies for enterprise digital transformation, the digital twin can effectively analyze, predict and simulate the composition, behavior and performance of the physical workshop through the development of a digital twin performance monitoring system. Summary of the invention
[0004] In order to address the deficiencies in the prior art, the present invention provides a performance monitoring system for a washing machine assembly production line based on digital twins, which realizes modular, hierarchical and standard management, takes into account the performance changes of system components such as machines and workers, and integrates multiple types of models including analytical models, simulation models and data-driven models, as well as multi-level models including unit-level models and system-level models to meet various management needs; it can locate workshop congestion sites, predict workshop performance changes, simulate different production plans, and realize efficient, transparent and real-time management of assembly workshops.
[0005] As a first aspect of the present invention, a washing machine assembly production line performance monitoring system based on digital twin is provided, comprising a data acquisition module, a performance analysis module, a performance prediction module and a performance simulation module, wherein:
[0006] The data acquisition module is used to collect real-time production status data of the washing machine assembly production line;
[0007] The performance analysis module is used to analyze the current production performance of the washing machine assembly production line according to the real-time production status data of the washing machine assembly production line, so as to obtain the current production performance analysis result of the washing machine assembly production line;
[0008] The performance prediction module is used to predict the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the current production performance analysis result of the washing machine assembly line, and adjust the future production plan of the washing machine assembly line according to the future production performance prediction result of the washing machine assembly line;
[0009] The performance simulation module is used to simulate the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the future production plan of the washing machine assembly line to obtain the future production performance simulation result of the washing machine assembly line.
[0010] Furthermore, it also includes a data preprocessing module, a data synchronization module and a data storage module, wherein the data preprocessing module is used to preprocess the real-time production status data of the washing machine assembly production line, the data synchronization module is used to synchronize the real-time production status data of the washing machine assembly production line after preprocessing to the performance monitoring system of the washing machine assembly production line, and the data storage module is used to store the real-time production status data of the washing machine assembly production line after preprocessing;
[0011] The data preprocessing module is used to perform data cleaning and data conversion on the real-time production status data of the washing machine assembly production line, wherein data cleaning includes missing value processing, outlier detection and duplicate data removal, and data conversion includes data type conversion, standardization and normalization, and time series specific conversion.
[0012] Furthermore, the data acquisition module includes a speed sensor, an acceleration sensor, a temperature sensor, an industrial camera and a positioning device. The data acquisition module transmits the collected real-time production status data of the washing machine assembly production line to the production system server through the industrial network, and synchronizes the collected real-time production status data of the washing machine assembly production line to the performance monitoring system of the washing machine assembly production line through the data synchronization module after preprocessing by the data preprocessing module; wherein, the open platform communication unified architecture protocol is used to transmit the collected real-time production status data of the washing machine assembly production line across communication layers, and the unified modeling language UML is used to construct a unified data model to establish a standardized modular database.
[0013] Furthermore, the performance analysis module includes a unit-level performance analysis module and a system-level performance analysis module. The unit-level performance analysis module is used to analyze the production performance of machine objects and individual operators in the washing machine assembly production line, and the system-level performance analysis module is used to analyze the overall production performance of the washing machine assembly production line.
[0014] Furthermore, the current production performance analysis indicators of the washing machine assembly line include throughput, cycle time, capacity utilization, on-time delivery rate, overtime pay rate, employee turnover rate, machine downtime rate, planned maintenance percentage, and overall equipment effectiveness.
[0015] Furthermore, the performance prediction module includes a unit-level performance prediction module and a system-level performance prediction module, and the production performance prediction of the performance prediction module is divided into three parts: model data processing, prediction model, and performance prediction application;
[0016] Model data processing: In the initial stage of the model data processing, the unit-level performance prediction module emphasizes the collection of specific parameters and detailed data, and the system-level performance prediction module includes more extensive and diverse data processing methods;
[0017] Prediction model: The prediction model includes data processing, feature selection, model training and validation testing. First, in the data processing stage, the collected raw data is cleaned and preprocessed. Then, in the feature selection stage, the key features that affect the production performance of the washing machine assembly line are preliminarily selected, and auxiliary features and secondary features are added on this basis to improve the prediction accuracy of the model. In the model training stage, the processed data is used for training. Finally, in the validation testing stage, a set of test data is used to verify the accuracy of the model, and the parameters of the prediction model are fine-tuned according to the test results to optimize the system performance.
[0018] Performance prediction application: The unit-level performance prediction module is used to predict the production performance of machine objects and individual operators in the washing machine assembly line, and the system-level performance prediction module is used to predict the production performance of the entire washing machine assembly line.
[0019] Furthermore, the key characteristics of the production performance of the washing machine assembly production line include personnel proficiency, typesetting shifts, number of out-of-stock times and out-of-stock time; the auxiliary characteristics of the production performance of the washing machine assembly production line include number of failures and job operation records; and the secondary characteristics of the production performance of the washing machine assembly production line include interval output and average job operation time.
[0020] Furthermore, the performance simulation module first calls the real-time data calibration and adjusts the simulation model based on the initial design, and the relevant parameters are configured according to the process file to obtain the real-time data-driven model; the relevant parameters are calibrated according to the synchronized data through the model interface to realize real-time data-driven simulation, and a unified data model is constructed on the basis of the multi-source heterogeneous real-time production status data of the washing machine assembly line, and a standard library is established to ensure the real-time data transmission for the performance simulation module to call; the performance simulation module maps the actual operating status of the washing machine assembly line to the digital device module to realize the virtual-real synchronization of the simulation model and reality.
[0021] Furthermore, the performance simulation module can simulate the equipment production performance of the washing machine assembly production line, formulate a corresponding equipment maintenance plan, and set the maintenance waiting time, minor maintenance time and major maintenance time according to the factory maintenance records; wherein the corresponding equipment maintenance plan includes post-maintenance, scheduled maintenance, condition maintenance and predictive maintenance.
[0022] Furthermore, the performance prediction module includes a first visualization module, and the performance simulation module includes a second visualization module, the first visualization module is used to display the future production performance prediction results of the washing machine assembly line and the future production plan of the washing machine assembly line, and the second visualization module is used to display the future production performance simulation results of the washing machine assembly line.
[0023] The washing machine assembly production line performance monitoring system based on digital twin provided by the present invention has the following advantages: it realizes modular, hierarchical and standard management, takes into account the performance changes of system components (machines and workers), and integrates multi-type (analysis, simulation and data-driven models) and multi-level (unit level and system level) models to meet various management needs; it can effectively analyze, predict and simulate the composition, behavior and performance of the physical workshop of the washing machine assembly production line, so as to locate the workshop congestion sites, predict the workshop performance changes, simulate different production plans, and realize efficient, transparent and real-time management of the washing machine assembly production line workshop. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation of the present invention.
[0025] Figure 1 This is a structural diagram of the washing machine assembly production line performance monitoring system based on digital twin provided by the present invention.
[0026] Figure 2 A schematic diagram of the composition of the workshop production performance analysis indicators provided by the present invention.
[0027] Figure 3 This is a flow chart for implementing the performance prediction module provided by the present invention.
[0028] Figure 4 This is a specific implementation flow chart of the performance prediction module provided by the present invention.
[0029] Figure 5 This is a flow chart for implementing the performance simulation module provided by the present invention. DETAILED DESCRIPTION
[0030] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the washing machine assembly production line performance monitoring system based on digital twin proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.
[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so as to describe the embodiments of the present invention described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0032] In the interpretation of the present invention, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense unless otherwise specified. For example, the connection can be a fixed connection, or it can be connected through a special interface, or it can be indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0033] Taking a company's washing machine assembly line as a reference, a performance monitoring system based on digital twins is constructed. The washing machine assembly line consists of operators, industrial robots, transportation tools, trucks, and warehouses. There are about 20 operators and 3 industrial robots to assemble washing machines. Based on the Industrial Internet, real-time data of operators, industrial robots, products, materials, and other equipment can be collected.
[0034] The development of a washing machine assembly line performance monitoring system based on digital twins aims to analyze, predict and simulate production performance in order to improve the production process of washing machine assembly lines and quickly identify problems. The key to the digital twin system is to build analysis models, prediction models, simulation models, visualization models, etc., and then integrate these models.
[0035] Specifically, the washing machine assembly production line performance monitoring system based on digital twins was developed by Unity3D, the database was built by MySQL, the relevant data were stored in various lists, and the analysis models were encoded into function classes so that they could be easily rebuilt.
[0036] In this embodiment, a washing machine assembly production line performance monitoring system based on digital twin is provided. Figure 1 As shown, the washing machine assembly production line performance monitoring system based on digital twin includes a data acquisition module, a performance analysis module, a performance prediction module and a performance simulation module, wherein:
[0037] The data acquisition module is used to collect real-time production status data of the washing machine assembly production line;
[0038] The performance analysis module is used to analyze the current production performance of the washing machine assembly production line according to the real-time production status data of the washing machine assembly production line, so as to obtain the current production performance analysis result of the washing machine assembly production line;
[0039] The performance prediction module is used to predict the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the current production performance analysis result of the washing machine assembly line, and adjust the future production plan of the washing machine assembly line according to the future production performance prediction result of the washing machine assembly line;
[0040] The performance simulation module is used to simulate the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the future production plan of the washing machine assembly line to obtain the future production performance simulation result of the washing machine assembly line.
[0041] In the construction of the performance monitoring system for washing machine assembly production line, the first step is performance analysis, the main goal of which is to monitor the performance of the system in real time through mathematical models. Performance analysis is divided into two levels: system-level analysis and unit-level analysis, so as to analyze the unit and global production performance such as time, quality and resource consumption. The washing machine assembly line can be simulated in real time, and the production performance such as processing time, productivity and defective rate can be monitored and optimized through equation-based calculations, so as to effectively improve the delivery time of the washing machine production line, reduce waste, and ensure efficient production maintenance. By monitoring the system performance, managers can deal with sudden problems and optimize the system composition and operation to obtain better production quality and effectiveness. Based on the performance analysis indicators, managers can formulate policies to improve productivity, improve delivery time, reduce waste, and achieve effective economic maintenance.
[0042] In the construction of the performance monitoring system for the washing machine assembly line, the second step is performance prediction. Through real-time data analysis and intelligent algorithms, the production performance at the unit and system levels can be accurately predicted and optimized, effectively improving assembly efficiency. Workers can formulate appropriate maintenance plans based on predictive failures, reduce downtime, and ensure production quality and consistency. In addition, managers can also adjust employee schedules by predicting productivity.
[0043] In the construction of the performance monitoring system for the washing machine assembly line, the third step is performance simulation. By configuring the workshop composition and assisting managers in optimizing production plans, it can effectively optimize the production process, reduce energy consumption, and ensure optimal resource allocation and time management, thereby improving overall production efficiency and reducing production energy consumption. The simulation model configures the current workshop composition status according to a certain time period, and then simulates the new production plan to obtain the optimal production plan arrangement in real time. The simulation model supports managers in making long-term plans and evaluating system-level performance under various factors.
[0044] Preferably, it also includes a data preprocessing module, a data synchronization module and a data storage module, the data preprocessing module is used to preprocess the real-time production status data of the washing machine assembly production line, the data synchronization module is used to synchronize the real-time production status data of the washing machine assembly production line after preprocessing to the washing machine assembly production line performance monitoring system, and the data storage module is used to store the real-time production status data of the washing machine assembly production line after preprocessing.
[0045] Preferably, the data acquisition module includes a speed sensor, an acceleration sensor, a temperature sensor, an industrial camera and a positioning device. The data acquisition module transmits the real-time production status data of the washing machine assembly line collected to the production system server through an industrial network (such as Ethernet TCP / IP), and synchronizes the data to the washing machine assembly line performance monitoring system through the data synchronization module after preprocessing by the data preprocessing module. At the same time, the production status data will be saved in a modular database to build a data model; wherein, the open platform communication unified architecture (OPC UA) protocol is used to transmit multi-heterogeneous industrial big data collected by RFID and other devices across communication layers, and the unified modeling language UML is used to build a unified data model to establish a standardized modular database. In terms of data processing and model optimization, the time-dependent and space-dependent features are constructed in combination with the RFID data and the workshop shift situation to clarify the data structure and relationship, and feature engineering construction innovation is carried out to make the specific model construction method for specific objects more accurate and efficient. The digital twin monitoring system includes performance analysis, performance prediction, performance simulation and visualization based on the architecture designed in the early stage.
[0046] Preferably, the data preprocessing module is used to perform data cleaning and data conversion on the real-time production status data of the washing machine assembly production line, wherein data cleaning includes missing value processing, outlier detection and duplicate data removal, and data conversion includes data type conversion, standardization and time series specific conversion.
[0047] Preferably, the performance analysis module includes a unit-level performance analysis module and a system-level performance analysis module. The unit-level performance analysis module is used to analyze the production performance of machine objects and individual operators in the washing machine assembly production line, and the system-level performance analysis module is used to analyze the overall production performance of the washing machine assembly production line.
[0048] Furthermore, the construction of specific functional modules is divided into unit level and system level, with the purpose of monitoring real-time system performance based on mathematical models, and analyzing and calculating local and global production performance (time, quality and consumption). The performance analysis module serves the shallow analysis of the production system, mainly relying on analysis formulas. Various indicator analysis modules are packaged and sub-packaged in software function functions, which is conducive to rapid analysis and reconstruction. The model is encapsulated into code for reuse. Through model calculation and system visualization, managers can analyze system bottlenecks, formulate optimization plans, and improve output and quality. Quantitative performance indicators reflect the system status and provide a basis for decision-making to improve production efficiency, delivery speed, resource utilization, etc.
[0049] Preferably, if Figure 2As shown, the current production performance analysis indicators of the washing machine assembly line include throughput, cycle time, capacity utilization, on-time delivery rate, overtime pay rate, employee turnover rate, machine downtime rate, planned maintenance percentage, and overall equipment effectiveness.
[0050] It should be noted that performance analysis mainly relies on various equations and calculations. The computational cost is low and real-time analysis can be performed. Both unit-level analysis indicators and system-level analysis indicators are constructed based on equations. From the perspective of time, cost and quality, mathematical equations can be used to measure system performance such as actual processing time, productivity and defective rate.
[0051] For unit-level analysis, mean time between failures, mean time to repair, and mean time to failure are key analytical indicators for processing equipment. Mean time between failures refers to the expected time between inherent failures of a mechanical or electronic system during normal system operation. This metric is calculated by taking the average time between system failures. Mean time to failure is a metric for repairable systems, which is calculated as the mean time between failures. Mean time to repair is a basic standard for measuring maintainable repairable items. It represents the average time required to repair a failed component or device. Use the following formula to calculate this indicator: Mean time to repair = Mean time to failure / Total time for fault correction maintenance measures.
[0052] Cost analysis in production includes material costs, unit production costs, and labor costs. Unit production costs are calculated by calculating all possible production costs (materials, labor, variable overhead, machine depreciation, etc.) and dividing this value by the number of units of the product produced. Use the following formula to calculate this indicator: Unit production cost = total production cost / production unit. Material cost is a measure used to compare the amount of material used to produce the expected number of units with the amount of material actually used to produce these units. Use the following formula to calculate this indicator: Material cost usage = materials used / expected materials. Labor costs are a simple indicator: labor as a percentage of cost = labor / total sales. Efficiency is the proportion of output generated to the total possible output, output / optimal output = utilization. Both labor costs and efficiency are related to the unit labor utilization rate. Use the following formula to calculate this indicator: Labor utilization rate = total labor content / (labor content + total idle time). It is necessary to calculate whether a person's working time is spent on work or idle time or other aspects, which will also bring overtime costs. In product production, takt time is the average time between the start of production of one unit and the next unit (takt time = production time / customer demand), and manufacturing conversion time is a measure of the speed or time required for a production line or factory to switch from one product to another different product. Use the following formula to calculate this indicator: Net available time - production time = conversion time. Cycle time is the average time between process completion. Manufacturing cycle time is a related indicator of speed or time. This measure is calculated by subtracting the process start time from the process end time (cycle time = process end time - process start time). At the same time, production cycle efficiency measures the percentage of value-added time to throughput time (manufacturing cycle efficiency = value-added time / throughput time).
[0053] Unplanned expenses such as maintenance will be incurred in product processing and logistics. At this time, analytical indicators such as maintenance cost per unit, unplanned capacity expenditure, and return rate are needed. First-time pass rate (FPY = quality unit / total production) refers to the number of units without rework or scrap defects divided by the number of units entering the same process within a set period of time. Maintenance cost per unit is the total maintenance cost divided by the number of units produced during the measurement period. The total maintenance cost includes maintenance hours, parts and any other costs related to maintenance work (preventive and corrective) calculated in the total cost. Use the following formula to calculate this indicator: Unit maintenance cost per unit maintenance cost = total maintenance cost / number of production units. Unplanned capacity expenditure is an unplanned expense, and the following formula is used to calculate this indicator: Unplanned capacity expenditure = total capacity expenditure-planned capacity expenditure. The return rate measures the frequency of customers rejecting products or requesting returns because they received defective products. On-time delivery is the percentage of time that the manufacturer delivers finished products to customers as planned. On-time delivery is calculated by dividing the number of orders by the number of orders delivered on time. Customer satisfaction can assess customer satisfaction with product quality and delivery and support continuous improvement.
[0054] In the daily production and management of factories, production costs, production performance, production scheduling, and labor management are important issues, and some key system-level performance indicators are needed to analyze them in order to help managers understand whether the factory has the potential to expand production scale and orders or whether production needs to be adjusted. The artificial intelligence and automation adoption rate can measure the extent to which manufacturers incorporate artificial intelligence (AI) and automation technologies into their processes, improve productivity and decision-making, and the digital transformation index can evaluate the application and integration of digital technologies in the manufacturing process to promote innovation and competitiveness.
[0055] Output refers to how much product a machine, line, unit, or plant produces in a given time; capacity utilization measures the total production capacity of a production line. Plant utilization is calculated by dividing actual plant utilization by total production capacity. The following formula is used to calculate this indicator: Actual plant utilization / total production capacity = capacity utilization; Work in progress (WIP) is the number or value of partially finished products. The indicator can be expressed by the following formula: WIP = percentage of completion * sales price - (cost to date / cost estimate). Rolling yield is a metric related to FPY that measures the likelihood that a production line will produce quality units in a multi-step process. Overall equipment effectiveness (OEE = availability * performance * quality) is a way to measure manufacturing productivity.
[0056] Plant uptime and downtime measure production performance at a system level, using the following formulas: Plant uptime = Plant run time (production) / Total available time for operation or production; Plant run time = Total available operation time - Planned and unplanned downtime / downtime; Equipment downtime = 1 - (Plant run time / Total time available for operation or production). Production achievement measures the percentage of time that the target production level is achieved within the specified schedule. This metric is calculated using the following formula: Production completion = Time to achieve production target / Total time in the schedule. Overtime as a percentage of total work hours calculates an organization's overtime hours compared to standard work hours, highlighting production scheduling issues (Overtime hours / Normal work hours * 100% = Overtime rate). Labor utilization measures the percentage of time employees spend on production activities, highlighting areas for better labor management and training. In the full-scale new product development stage, the new product introduction rate measures the time it takes to design, develop and produce a new product, New product introduction rate = Full-scale production start date - Product development start date, Cash cycle time is the time between the purchase of a manufacturing plant or the purchase of inventory and the receipt of payment for the sale of the produced or purchased inventory.
[0057] Supply Chain Efficiency Supply Chain Efficiency tracks the performance of the entire supply chain, identifying bottlenecks and optimization opportunities. Unused capacity expenditure measures how much available capacity provided by machines, personnel, etc. is supported but not used. This metric is calculated using the following formula: Unused capacity expenditure = Fixed cost x (1-Company output / Company maximum possible output). Inventory turnover analyzes the speed of inventory sales, revealing insights into demand and production efficiency. Demand forecasting is a measure that helps manufacturers understand the amount of raw materials required to meet expected customer demand. Expected customer demand = Raw materials * Production rate. Asset utilization is a metric used to analyze efficiency. This metric is calculated using the following formula: Asset utilization = (actual output / maximum capacity) * 100. Opportunity gap is a metric related to asset utilization. The opportunity gap is the difference between what an asset can produce and what it actually produces. This metric is calculated using the following formula: Maximum capacity - actual output = opportunity gap.
[0058] Additionally, health and safety incidents is the number of health and safety incidents recorded over a period of time, and measures the number of violations that occur in a year. Avoided costs are the estimated savings based on preventive actions. Using machine breakdown as an example, the avoidable cost KPI can be calculated by subtracting the preventive maintenance costs from the assumed repair costs plus the productivity losses associated with downtime. The relevant calculation is the ratio of unplanned repairs to scheduled repairs. (Repair costs + production losses - preventive maintenance costs = avoided costs; emergency repair incidents / planned repair incidents = unplanned repair ratio).
[0059] The depth of system performance analysis depends on the status data and activity trajectories of the manufacturing unit collected by sensors and smart sensing devices. The digital twin analysis models of the workshop are designed to monitor real-time system performance in an intuitive way. They are all built based on equations. Constant values such as standard time and single task consumption are regarded as static properties of unit-level objects, but system performance will fluctuate due to random interference and unexpected events. Mathematical equations can be used to measure system performance (actual processing time, productivity and defective rate) from the perspective of time, cost and quality. Figure 2 The system performance is described by equations of key performance indicators shown. Evaluation models with certain interfaces can be constructed based on the equations. The modular model reads variables from the data model and outputs the results to the visualization model. The performance analysis model is encapsulated in the code. Quantifiable and measurable performance indicators reflect the operating status of the physical system. By monitoring system performance indicators, managers can handle unexpected problems, optimize system composition and operation, increase productivity, improve delivery time, reduce waste, and achieve economically effective maintenance.
[0060] Preferably, if Figure 3 As shown, the performance prediction module includes a unit-level performance prediction module and a system-level performance prediction module. The production performance prediction of the performance prediction module is divided into three parts: model data processing, prediction model, and performance prediction application;
[0061] Model data processing: In the initial stage of the model data processing, the unit-level performance prediction module emphasizes the collection of specific parameters and detailed data, and the system-level performance prediction module includes a wider range of and more diverse data processing methods; for the workshop object of the present invention, there are many manual workers involved and production is easily affected by multiple factors, so it is necessary to refine data collection, and it is necessary to consider the work shifts, positions, and skill levels of personnel; the operating status of the equipment, the number of failures, and maintenance records; the spare parts inventory level, the number of out-of-stock times, the duration of out-of-stock times, and the impact on production; multi-dimensional data such as job operation records, operation error types, occurrence times, and durations. At the same time, the workshop situation is complex, and data missing, abnormal, and duplicate problems may occur. The data is also multi-source heterogeneous data, so data preprocessing requires data cleaning (missing value processing, abnormal value detection, duplicate data removal), data conversion (data type conversion, standardization, time series specific conversion), etc.
[0062] Prediction model: The prediction model includes data processing, feature selection, model training and verification test; first, in the data processing stage, the collected raw data is cleaned and preprocessed; then, in the feature selection stage, the key features that affect the production performance of the washing machine assembly production line are preliminarily selected, and auxiliary features and secondary features are added on this basis to improve the prediction accuracy of the model; in the model training stage, the processed data is used for training; finally, in the verification test stage, a set of test data is used to verify the accuracy of the model, and the parameters of the prediction model are fine-tuned according to the test results to optimize the system performance; among them, for the workshop object of the present invention, considering the production mode and production status of the washing machine assembly production line, it is necessary to preliminarily select the key features that affect the production performance of the washing machine assembly production line, and add auxiliary features and secondary features on the basis of the key features to improve the prediction accuracy; the key features will have a serious impact on the production performance and need to be considered first, the auxiliary features will affect the performance prediction level to a certain extent, and the secondary features will be less affected by sudden fluctuations and will also be an important reference. After completing the feature engineering, data segmentation and integration are carried out and training is carried out, and the optimal training parameters and models are selected by continuously optimizing the training parameters.
[0063] Performance prediction application: The unit-level performance prediction module is used to predict the production performance of machine objects and individual operators in the washing machine assembly line, and the system-level performance prediction module is used to predict the production performance of the entire washing machine assembly line.
[0064] Preferably, the key characteristics of the production performance of the washing machine assembly production line include personnel proficiency, typesetting shifts, number of out-of-stock times and out-of-stock time; the auxiliary characteristics of the production performance of the washing machine assembly production line include number of failures and job operation records; and the secondary characteristics of the production performance of the washing machine assembly production line include interval output and average job operation time.
[0065] Furthermore, performance prediction focuses on using intelligent algorithms and statistical models for short-term performance prediction to achieve production progress prediction, production efficiency prediction, bottleneck prediction, etc. based on real-time data, which is conducive to managers to optimize production processes and adjust production plans. In terms of technical methods, model data processing is first carried out to enter the field of prediction models and algorithms, and the statistical model algorithm is innovatively combined with a specially constructed intelligent algorithm to traverse the various stages of model selection, training and verification testing. Starting with algorithm selection, it involves statistical methods, machine learning or deep learning technology, and obtains feature engineering and model training, and then continues to model verification and testing to evaluate the proficiency and accuracy of the model. Finally, the model is optimized to achieve peak system performance. From the computational level, the new application model that combines statistical models and intelligent algorithms achieves a good balance between real-time and accuracy in the actual application effect of the model, and uses intelligent algorithms to autonomously adjust the time window to balance latency and accuracy.
[0066] It should be noted that in the application field of performance prediction, unit-level prediction is mainly used to predict and improve the efficiency and performance of a single device, while system-level prediction is aimed at the performance of the entire system or production line, such as output prediction and system optimization. Workers can formulate appropriate maintenance plans based on predictive faults to avoid unexpected downtime and achieve unit-level performance prediction; managers can adjust employee schedules by predicting work efficiency based on personal factors, implement unit-level performance prediction and system-level performance prediction to formulate short-term policies and monitor the changing state of interests. Unit-level system performance prediction mainly focuses on predicting the performance of a single unit or equipment in the workshop, which can adjust the processing and assembly personnel of the position or optimize the process steps of a specific position; system-level performance prediction emphasizes the monitoring and prediction of the overall production performance of the workshop, and promotes more refined resource allocation and scheduling based on the output prediction of the entire system. Performance prediction focuses more on short-term performance prediction using intelligent algorithms and statistical models. The workshop digital twin prediction model aims to predict system performance in the presence of random interference, and mainly relies on historical data and complex real-time data to predict performance. Static components consist of three-dimensional models and three-dimensional environments, while dynamic elements are driven by various data such as analysis and real-time data. Some simulation models only describe the static mode of system performance, but since there are many random events in the actual workshop and system parameters change frequently, data-driven methods can predict the performance status more accurately.
[0067] Specifically, Figure 4 As shown, the system performance prediction method of the present invention is introduced by taking the workshop output as an example. First, data collection is performed, and data information related to the output such as personnel proficiency, work shift, job processing time, product RFID quantity and other data information are preliminarily screened; then data processing is performed, the main purpose of which is to improve data quality and convert it into a suitable format. In particular, it is necessary to perform preliminary statistical processing on data in different time intervals for subsequent prediction of the output; further feature engineering is constructed, because the time series algorithm is selected for prediction, the constructed features need to have obvious time dependence, and at the same time, in order to accurately predict the output, the spatial feature sequence of the job, the location data of the product and the output of different jobs in a period of time are processed, especially in order to represent the workshop production period, the 24-hour time sine and cosine processing is performed to achieve continuous operation of feature quantity; finally, through model optimization, through the network layer parameters, activation function, number of iterations, learning rate, and multiple different training parameters of the number of neurons are continuously adaptively adjusted, and finally a stable and fast-converging loss function image is obtained, and finally the model is applied to prediction, and the output of the entire workshop is output. By predicting this parameter, the evaluation and prediction of the workshop production performance can be improved, the workshop production plan can be adjusted, and the production efficiency can be further improved.
[0068] Preferably, if Figure 5 As shown, the performance simulation module first calls the real-time data calibration and adjusts the simulation model based on the initial design, and the relevant parameters are configured according to the process file to obtain the real-time data-driven model; the relevant parameters are calibrated according to the synchronized data through the model interface to realize the real-time data-driven simulation, and a unified data model is constructed on the basis of the multi-source heterogeneous real-time production status data of the washing machine assembly line, and a standard library is established to ensure the real-time data transmission for the performance simulation module to call; the performance simulation module maps the real operating status of the washing machine assembly line to the digital device module to realize the virtual-real synchronization of the simulation model and reality.
[0069] Specifically, Plant Simulation is used to simulate the discrete events of the workshop. The workstations of assembly operators, industrial robots, etc. are configured according to the workshop layout, and initial parameters are set based on the standard processing time. In order to characterize the changes in the processing performance of the workshop components, the performance model of the workshop unit components is integrated to improve the realism of the workshop simulation model. Figure 5 For the construction process of the performance simulation module, although the simulation model does not update the workshop model in real time, it will evaluate the performance status of the workshop components based on real-time data. Once the status changes, the performance parameters in the workshop simulation model will also be adjusted, which can ensure the reliable simulation of the new production plan. When managers need to adjust the production plan, such as the production of new batches of products, personnel transfers, and equipment, they can use the simulation function module to formulate the optimal plan. The output plan can be generated as a report, and the simulation results can be visualized, which is conducive to the efficient management of managers.
[0070] Specifically, for industrial robots in washing machine assembly lines, the main factor affecting life performance is the performance change of rotating parts, including transmission gears, bearings, etc. Therefore, in the industrial robot unit model, the Weibull model is used to describe the degradation process of these key components. By integrating the degradation model, the performance simulation module can simulate the equipment production performance of the washing machine assembly line and formulate a corresponding equipment maintenance plan.
[0071]
[0072] As shown in the formula, α is the scale parameter and β is the shape parameter.
[0073] Therefore, the failure rate formula is:
[0074]
[0075] The reliability formula is:
[0076]
[0077] After setting the performance degradation model, different maintenance strategies can be set accordingly, such as post-maintenance, scheduled maintenance, condition maintenance, predictive maintenance, etc. According to the plant maintenance records, set parameters such as maintenance waiting time, minor maintenance time, and major maintenance time.
[0078] For operators in the washing machine assembly line, statistical models are used to describe the changes in their production efficiency, which is mainly reflected in the changes in unit operation time. For example, probability models such as normal distribution are used to characterize their change patterns. In general, the performance change model is integrated into the performance simulation module to optimize the personnel task scheduling plan in the production system.
[0079] Preferably, the performance prediction module includes a first visualization module, and the performance simulation module includes a second visualization module, the first visualization module is used to display the future production performance prediction results of the washing machine assembly line and the future production plan of the washing machine assembly line, and the second visualization module is used to display the future production performance simulation results of the washing machine assembly line.
[0080] It should be noted that unit-level performance analysis is performed based on the unit components of the system (operators, processing equipment, transmission equipment, etc.), while system-level performance analysis is performed based on the overall system. Unit-level performance prediction is performed based on the unit components of the system (operators, processing equipment, transmission equipment, etc.), while system-level performance prediction is performed based on the overall system. Performance simulation is performed based on real-time production status and production plans to determine key indicators of the production system, such as system bottlenecks and expected completion time.
[0081] In an embodiment of the present invention, data is acquired from the factory and stored in a database / data model. When a new solution needs to be evaluated, the simulation model will be rebuilt; when the simulation encounters bottlenecks and problems, highlights and red alarms will be displayed in the three-dimensional scene. The results are displayed on a digital panel to assist managers in making decisions, and to achieve result visualization and decision support. Through the real-time transmission of production status data and production plans for special objects, long-term and short-term predictions of multi-level production system performance at the unit level and system level are obtained. Combined with statistical model algorithms and intelligent algorithms, the various stages of the model are selected, trained and verified, and tested to complete the analysis, prediction and simulation of the production system performance. By constructing a multi-level real-time performance monitoring system for production systems at the unit level and system level, intelligent functions such as production optimization, production early warning, and fault tracking can be effectively carried out.
[0082] The main innovation of the present invention is to predict the performance of the simulation production line unit to evaluate the bottleneck, adjust the production plan, and predict the production progress. The present invention highlights the hierarchical (system level to unit level) digital twin modeling; highlights the combination of multiple models, such as analytical models that focus on analytical calculations, prediction models that are essentially algorithms, and simulation models that are essentially discrete event simulations.
[0083] The washing machine assembly production line performance monitoring system based on digital twin provided in this embodiment can realize multi-level, modular and standardized system development, and help improve the robustness and reusability of the system. The washing machine assembly production line mainly sets sensors to track the production of work-in-progress, and uses RFID to identify the production batch number of each work-in-progress, the processing time at each workstation, the production date, etc., and the operator can upload the quality of the work-in-progress. These sampled data are transmitted through an industrial network (such as Ethernet TCP / IP). After the collected data is transmitted to the production system server, it needs to undergo certain data preprocessing, including data cleaning, NULL value filling, etc., and then synchronized to the digital twin monitoring system. At the same time, the production data will be saved in a modular database to build a data model. The constructed data model is a real-time status of a production system. After being injected into the system, the system updates the system performance analysis, prediction and visualization modules in real time.
[0084] First, a washing machine assembly production line performance monitoring system based on digital twins is constructed according to the performance monitoring requirements of the workshop. The washing machine assembly production line performance monitoring system based on digital twins supports rapid reconstruction of the software architecture under demand changes. The washing machine assembly production line performance monitoring system based on digital twins provided by the embodiment of the present invention takes into account the performance changes of the components (machines and workers) of the washing machine assembly production line, and integrates multi-type (analysis, simulation and data-driven models) and multi-level (unit level and system level) models to meet various management needs. In addition, the hybrid mechanism also realizes transient and long-term performance analysis, including evaluation, prediction and simulation.
[0085] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.
Claims
1. A washing machine assembly production line performance monitoring system based on digital twin, characterized in that: It includes data acquisition module, performance analysis module, performance prediction module and performance simulation module, among which: The data acquisition module is used to collect real-time production status data of the washing machine assembly production line; The performance analysis module is used to analyze the current production performance of the washing machine assembly production line according to the real-time production status data of the washing machine assembly production line, so as to obtain the current production performance analysis result of the washing machine assembly production line; The performance prediction module is used to predict the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the current production performance analysis result of the washing machine assembly line, and adjust the future production plan of the washing machine assembly line according to the future production performance prediction result of the washing machine assembly line; The performance simulation module is used to simulate the future production performance of the washing machine assembly line according to the real-time production status data of the washing machine assembly line and the future production plan of the washing machine assembly line to obtain the future production performance simulation result of the washing machine assembly line.
2. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: It also includes a data preprocessing module, a data synchronization module and a data storage module, wherein the data preprocessing module is used to preprocess the real-time production status data of the washing machine assembly production line, the data synchronization module is used to synchronize the real-time production status data of the washing machine assembly production line after preprocessing to the performance monitoring system of the washing machine assembly production line, and the data storage module is used to store the real-time production status data of the washing machine assembly production line after preprocessing; Among them, the data preprocessing module is used to perform data cleaning and data conversion on the real-time production status data of the washing machine assembly production line, wherein data cleaning includes missing value processing, outlier detection and duplicate data removal, and data conversion includes data type conversion, standardization and normalization and time series specific conversion.
3. A washing machine assembly production line performance monitoring system based on digital twin according to claim 2, characterized in that: The data acquisition module includes a speed sensor, an acceleration sensor, a temperature sensor, an industrial camera and a positioning device. The data acquisition module transmits the collected real-time production status data of the washing machine assembly production line to the production system server through the industrial network, and synchronizes the collected real-time production status data of the washing machine assembly production line to the performance monitoring system of the washing machine assembly production line through the data synchronization module after being preprocessed by the data preprocessing module; wherein, the open platform communication unified architecture protocol is used to transmit the collected real-time production status data of the washing machine assembly production line across communication layers, and the unified modeling language UML is used to build a unified data model to establish a standardized modular database.
4. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: The performance analysis module includes a unit-level performance analysis module and a system-level performance analysis module. The unit-level performance analysis module is used to analyze the production performance of machine objects and individual operators in the washing machine assembly production line, and the system-level performance analysis module is used to analyze the overall production performance of the washing machine assembly production line.
5. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: Current production performance analysis metrics for the washing machine assembly line include throughput, cycle time, capacity utilization, on-time delivery, overtime pay rate, employee turnover, machine downtime rate, planned maintenance percentage, and overall equipment effectiveness.
6. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: The performance prediction module includes a unit-level performance prediction module and a system-level performance prediction module. The production performance prediction of the performance prediction module is divided into three parts: model data processing, prediction model, and performance prediction application; Model data processing: In the initial stage of the model data processing, the unit-level performance prediction module emphasizes the collection of specific parameters and detailed data, and the system-level performance prediction module includes more extensive and diverse data processing methods; Prediction model: The prediction model includes data processing, feature selection, model training and validation testing. First, in the data processing stage, the collected raw data is cleaned and preprocessed. Then, in the feature selection stage, the key features that affect the production performance of the washing machine assembly line are preliminarily selected, and auxiliary features and secondary features are added on this basis to improve the prediction accuracy of the model. In the model training stage, the processed data is used for training. Finally, in the validation testing stage, a set of test data is used to verify the accuracy of the model, and the parameters of the prediction model are fine-tuned according to the test results to optimize the system performance. Performance prediction application: The unit-level performance prediction module is used to predict the production performance of machine objects and individual operators in the washing machine assembly line, and the system-level performance prediction module is used to predict the production performance of the entire washing machine assembly line.
7. A washing machine assembly production line performance monitoring system based on digital twin according to claim 6, characterized in that: The key characteristics of the production performance of the washing machine assembly production line include personnel proficiency, typesetting shifts, number of out-of-stock times and out-of-stock time; the auxiliary characteristics of the production performance of the washing machine assembly production line include the number of failures and job operation records; and the secondary characteristics of the production performance of the washing machine assembly production line include interval output and average job operation time.
8. The washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: The performance simulation module first calls the real-time data calibration and adjusts the simulation model based on the initial design, and the relevant parameters are configured according to the process file to obtain the real-time data driven model; Relevant parameters are calibrated according to synchronized data through the model interface to realize real-time data-driven simulation. Based on the multi-source heterogeneous real-time production status data of the washing machine assembly line, a unified data model is constructed and a standard library is established to ensure the real-time data transmission for the performance simulation module to call. The performance simulation module maps the actual operating status of the washing machine assembly line to the digital device module to realize the virtual-real synchronization of the simulation model and reality.
9. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: The performance simulation module can simulate the equipment production performance of the washing machine assembly production line, formulate a corresponding equipment maintenance plan, and set the maintenance waiting time, minor maintenance time and major maintenance time according to the factory maintenance record; wherein the corresponding equipment maintenance plan includes post-maintenance, scheduled maintenance, condition maintenance and predictive maintenance.
10. A washing machine assembly production line performance monitoring system based on digital twin according to claim 1, characterized in that: The performance prediction module includes a first visualization module, and the performance simulation module includes a second visualization module. The first visualization module is used to display the future production performance prediction results of the washing machine assembly production line and the future production plan of the washing machine assembly production line, and the second visualization module is used to display the future production performance simulation results of the washing machine assembly production line.
Citation Information
Cited By
Discrete manufacturing equipment efficiency maximization strategy determination system and method
CN120045008A
