Shell-based production process management system and method

Abstract

The invention relates to the technical field of shell production management, in particular to a shell-based production process management system and method.The shell-based production process management system is composed of a data layer, a management layer, an execution layer and a control layer, the data layer is used for storing data of the whole shell production process, and the management layer is used for management decision making of shell production; the execution layer comprises a physical equipment unit used for shell processing, the control layer is used for real-time monitoring and stable control of the shell production process, the data layer and the management layer carry out data interaction through a database and middleware, and the management layer and the execution layer transmit a production instruction and a task state through an MES system. According to the production process management system and method based on the shell, design planning, plan scheduling, task execution, equipment management and control, quality detection and data analysis are organically integrated through a four-layer structure design, equipment vacancy and waiting time are reduced, and the stability of a production line is enhanced through edge pre-judgment and localized mistake proofing of a control layer.

Description

Technical Field

[0001] This invention relates to the field of shell production management technology, specifically to a shell-based production process management system and method. Background Technology

[0002] As a critical load-bearing component in mechanical equipment, electronic devices, and industrial products, the production process of housings involves multiple precision machining steps and the coordination of complex process parameters. Traditional housing manufacturing typically relies on manual experience for process design, equipment scheduling, and process control, resulting in pain points such as long design iteration cycles, insufficient production planning flexibility, and low equipment utilization. Especially against the backdrop of increasing demand for multi-variety, small-batch orders, the problems of data fragmentation and delayed decision-making in the traditional model are becoming increasingly prominent, easily leading to drawbacks such as wasted production resources, frequent quality fluctuations, and slow response to anomalies.

[0003] In recent years, some enterprises have introduced digital management systems (such as MES and ERP) to optimize production processes. However, existing solutions mostly focus on improving single aspects and fail to achieve seamless integration across the entire chain of design, planning, execution, and inspection. For example, the disconnect between process design and production planning leads to parameter mismatch; the lack of real-time interaction between equipment status monitoring and task scheduling makes it difficult to dynamically respond to unexpected operating conditions; and quality inspection data is stored in isolation, making it impossible to effectively correlate it with process parameters for defect root cause analysis. Furthermore, traditional systems lack sufficient integration with technologies such as artificial intelligence and edge computing, and process optimization still relies on fixed rule bases, making it difficult to adapt to the intelligent decision-making needs of complex production scenarios.

[0004] Therefore, there is an urgent need to build an integrated and intelligent production process management system, connect the data links of the entire shell manufacturing process, and realize closed-loop optimization of process design, resource scheduling, process control and quality traceability to meet the needs of high-quality, high-efficiency and flexible modern shell manufacturing.

[0005] To this end, we propose a shell-based production process management system and method. Summary of the Invention

[0006] One of the technical problems this application aims to solve is the urgent need to build an integrated and intelligent production process management system, connect the data links of the entire shell manufacturing process, and realize closed-loop optimization of process design, resource scheduling, process control and quality traceability to meet the needs of high-quality, high-efficiency and flexible modern shell manufacturing.

[0007] To address the aforementioned technical problems, this application provides a shell-based production process management system and method, which consists of a four-layer structure: a data layer, a management layer, an execution layer, and a control layer. The data layer is used to store data from the entire shell production process, including design documents, material information, process parameters, equipment status, and quality inspection results, and provides interfaces for data storage, querying, and updating. The management team makes management decisions for shell production, including production planning, task allocation, resource allocation, progress monitoring, and quality control, and interacts with the lower levels through the Manufacturing Execution System (MES). The execution layer includes physical equipment units for housing processing, which include at least CNC machining equipment, laser cutting machines, automated assembly lines, and testing equipment; The control layer is used for real-time monitoring and stable control of the shell production process, including programmable logic controllers (PLCs), sensor networks and automated control systems. The control layer is connected to the execution layer devices via industrial Ethernet or fieldbus. The data layer and the management layer interact through databases and middleware. The management layer and the execution layer communicate through the MES system to transmit production instructions and task status. The execution layer and the control layer communicate through industrial Ethernet or fieldbus to issue instructions and provide status feedback.

[0008] In some embodiments, a design management module is also included. The design management module integrates a CAD / CAM system for rapid design of shell structure and processing technology and generation of process documents, and supports intelligent iterative optimization of design schemes and process parameters based on AI models of historical case sets.

[0009] In some embodiments, a production planning module is also included, which is used to construct a scheduling model containing fuzzy constraints based on the shell production order requirements, the process documents generated by the design management module, and the real-time resource status information in the resource library, and generate a production plan that integrates multiple orders and multiple tasks.

[0010] In some embodiments, the system further includes a task allocation module, which dynamically allocates and optimizes the tasks defined by the production planning module based on the real-time availability of execution layer equipment and the load of operators, and issues task instructions to specific equipment and personnel.

[0011] In some embodiments, the device management module and the quality inspection module are also included; The equipment management module is used to monitor the real-time operating status of the execution layer equipment, and to perform edge state prediction through equipment data collected by the control layer, so as to realize remote equipment control, energy efficiency analysis and fault diagnosis. The quality inspection module uses automated inspection equipment and vision inspection system to perform real-time online inspection of shell products, and automatically associates and binds them with process parameters, equipment status and task batch data in the data layer to form traceable quality records and support defect pattern analysis.

[0012] In some embodiments, a data analysis module is also included. The data analysis module is used to collect the shell design, planning, execution, equipment and quality data stored in the data layer, perform correlation mining and modeling analysis, and provide dynamic decision support for production planning, quality control and process optimization in the management layer.

[0013] In some embodiments, a shell-based production process management method includes the following steps: S1: Utilize the design management module, combined with CAD / CAM systems and AI models, to perform shell design and process planning, and generate optimized process documents; S2: Using the production planning module, based on order requirements, the process documents, and real-time resource constraints, construct and solve the scheduling model to generate a global production plan; S3: Using the task allocation module, combined with the real-time status of equipment and personnel at the execution layer, tasks in the production plan are dynamically allocated to specific workstations; S4: The control layer drives the execution layer devices to perform production tasks, while simultaneously collecting equipment status and production progress data in real time; S5: Use the quality inspection module to perform online inspection of the produced shells and record quality data by associating it with production batch information; S6: Utilize the data analysis module to integrate and analyze design, planning, execution, equipment, and quality data, and generate optimization decisions that are then fed back to management.

[0014] In some embodiments, the steps of using the design management module to perform shell design and process planning, and generate optimized process documents specifically include: calling the AI ​​model based on a historical successful case library to intelligently match and iteratively optimize the current design requirements, and outputting process documents containing key parameter settings.

[0015] In some embodiments, during the step of generating a global production plan, the scheduling model considers the imprecision of equipment capacity, material supply, and personnel skills, and uses a fuzzy constraint modeling method to handle uncertainties.

[0016] In some embodiments, in the step of the control layer driving the execution layer device to perform production tasks and collect data in real time, the control layer performs device edge state prediction and localized error prevention control based on sensor data, and provides real-time feedback of status information to the management layer and the device management module; at the same time, the critical defects detected in the quality inspection step will automatically trigger the task allocation module to pause or adjust the task flow of the relevant workstation.

[0017] This invention has at least the following beneficial effects: 1. Through a four-layer structure—data layer, management layer, execution layer, and control layer—design planning, scheduling, task execution, equipment management, quality inspection, and data analysis are organically integrated. Integrated design management shortens the process preparation cycle; production planning and dynamic task allocation based on real-time status data optimize resource utilization and reduce equipment idle and waiting time; edge prediction and localized error prevention in the control layer enhance production line stability. Multi-layer collaborative operation effectively shortens the overall production cycle and improves overall equipment efficiency (OEE) and shell output per unit time.

[0018] 2. Embedded automated quality inspection equipment and vision systems enable online real-time monitoring and 100% coverage of key dimensions, appearance, and assembly precision of the housing, allowing for timely detection and isolation of defective products. Quality data is automatically and strongly correlated with design parameters, process settings, equipment status, production batches, and other information, constructing a complete product quality traceability chain. Combined with the status prediction and fault diagnosis functions of the equipment management module, the system can quickly locate the root cause of quality fluctuations or anomalies, identify process parameter windows and equipment performance degradation trends, providing clear direction for process optimization and preventative maintenance, significantly reducing batch defect risks and rework costs, and ensuring a high degree of consistency in product quality between batches.

[0019] 3. The data analysis module comprehensively collects and integrates heterogeneous data from multiple sources, including design, planning, execution, equipment, and quality. Through correlation mining, modeling, and in-depth analysis, it reveals bottlenecks, hidden optimization potential, and potential risks in the production process. Based on this, the system provides quantitative and objective support for key management decisions such as optimized scheduling in the production planning module, adjustment of early warning thresholds in the quality inspection module, formulation of predictive maintenance strategies in the equipment management module, and iterative optimization of process parameters. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system composition of the present invention; Figure 2 This is a schematic diagram of the method steps. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1, please refer to Figure 1The present invention provides a technical solution: a shell-based production process management system, which consists of a four-layer structure: data layer, management layer, execution layer and control layer; The data layer is used to store data from the entire shell production process, including design documents, material information, process parameters, equipment status, and quality inspection results, and provides interfaces for data storage, querying, and updating. The management team makes management decisions for shell production, including production planning, task allocation, resource allocation, progress monitoring, and quality control, and interacts with the lower levels through the Manufacturing Execution System (MES). The execution layer includes physical equipment units for housing processing, which include at least CNC machining equipment, laser cutting machines, automated assembly lines, and testing equipment; The control layer is used for real-time monitoring and stable control of the shell production process, including programmable logic controllers (PLCs), sensor networks and automated control systems. The control layer is connected to the execution layer devices via industrial Ethernet or fieldbus. The data layer and the management layer interact through databases and middleware. The management layer and the execution layer communicate through the MES system to transmit production instructions and task status. The execution layer and the control layer communicate through industrial Ethernet or fieldbus to issue instructions and receive status feedback.

[0023] Specifically, the system's four-layer architecture is data-driven at its core, forming a vertically integrated and horizontally coordinated operating mechanism: The data layer serves as the foundational support, storing shell design documents (such as CAD models and CAM process parameters), order information, equipment operation logs, and quality inspection results through a distributed database. The data layer provides standardized interfaces to support real-time access and dynamic updates of multi-source heterogeneous data.

[0024] The management team makes overall strategic decisions based on data-layer information, and its core modules include: The design management module integrates CAD / CAM tools and AI algorithm libraries. By calling successful process parameter templates from the historical case library and combining them with the current shell structure characteristics, it automatically recommends optimized machining paths, tool parameters, and process sequences, shortening the manual design cycle and avoiding experience errors.

[0025] The production planning module uses fuzzy constraint modeling technology to generate a multi-task scheduling scheme that balances efficiency and robustness by integrating dynamic variables such as order priority, equipment capacity, material inventory and personnel skills, and supports flexible requirements such as order insertion and emergency tasks.

[0026] The task allocation module collects real-time data on the status of execution layer equipment (such as processing progress and fault alarms) and personnel load, and dynamically adjusts the task allocation strategy in the initial plan. For example, it can automatically match high-precision processes to idle high-performance equipment, or intelligently reassign tasks on equipment that suddenly stops, so as to minimize production interruptions.

[0027] The execution layer consists of physical equipment such as CNC machining centers, laser cutting machines, and automated assembly lines. After receiving task instructions from the management layer, it executes specific processing operations.

[0028] The control layer collects equipment operation data (such as spindle vibration, temperature, and current fluctuations) in real time through PLC and industrial IoT terminals. Combined with edge computing capabilities, it predicts the health status of the equipment (such as tool wear warning and transmission component performance degradation analysis). It also fine-tunes equipment parameters through a closed-loop feedback mechanism (such as compensating for machining errors) to ensure machining accuracy and stability.

[0029] Example 2, please refer to Figure 1 The system also includes a design management module, which integrates a CAD / CAM system for rapid design of shell structures and processing technology, as well as generation of process documents. It also supports intelligent iterative optimization of design schemes and process parameters based on AI models using historical case sets.

[0030] Specifically, the core principle of this module is knowledge reuse and data-driven parameter optimization, as follows: The CAD / CAM system is integrated and directly connects to the 3D model data of the design end. It automatically analyzes the geometric features of the shell, such as surface curvature, wall thickness distribution, hole accuracy, etc., and generates basic machining paths and tool trajectories.

[0031] The historical case library is constructed by storing the successfully verified process solutions in a structured manner, including shell features, material properties, equipment models, cutting parameters, etc., and establishing a mapping relationship between feature parameters and process results.

[0032] The AI-powered intelligent optimization engine matches similar historical solutions in the case library based on key features of the housing (such as deep cavity structure and thin-walled areas). It analyzes the correlation between parameter combinations and machining quality (such as surface roughness and deformation) through machine learning algorithms (such as decision trees and neural networks), iteratively optimizes the current process parameters (such as feed rate, spindle speed, and depth of cut), and generates an improvement solution with controllable risks.

[0033] The purpose of this design is to reduce the time required for typical shell process design by more than 50% by automatically matching features and recommending parameters, which relies on repeated trial and error and traditional process design based on human experience. By leveraging data patterns from historical success cases, it avoids the limitations of human experience (such as ignoring cutting vibrations caused by batch variations in materials), ensuring a high first-pass yield for the process design.

[0034] The benefits of this design are: Increased speed of process document generation, with AI automatically outputting 80% of basic processes, requiring engineers only to review key nodes, thus reducing time. Accelerated design-production integration, with CAD models directly supporting CAM programming, eliminating drawing conversion errors and repetitive input. Reduced process defect rate, by using historical parameter boundary constraints (e.g., feed rate ≤0.1mm / r for thin-walled parts) to avoid out-of-tolerance machining. Improved first-piece yield, with AI-optimized parameters stabilizing critical dimension CPK values ​​above 1.33. Reduced material waste in prototype production, decreasing the number of typical shell prototypes from 3-5 to 1-2. Extended tool life: intelligent matching of cutting parameters reduces abnormal tool wear rate by ≥25%.

[0035] Example 3: The system also includes a production planning module. The production planning module is used to construct a scheduling model with fuzzy constraints based on the shell production order requirements, the process documents generated by the design management module, and the real-time resource status information in the resource library, and generate a production plan that integrates multiple orders and multiple tasks.

[0036] Specifically, the core of the module lies in dynamic constraint modeling and multi-objective optimization, as detailed below: The first step is multi-source data fusion, integrating real-time inputs from three sources: order requirements, process documents, and resource status.

[0037] Order requirements include housing type, quantity, delivery date, and priority weight. Process documents are derived from the precise process sequence, unit time, and tooling dependencies in the design management module. Resource status includes real-time machine tool occupancy, tool life count, material inventory level, and personnel skill matching in the equipment library.

[0038] Next, fuzzy constraint modeling is used to transform traditional rigid constraints, such as the daily production capacity of equipment A being 100 units, into elastic ranges. For example, equipment A can complete 100 units under normal operating conditions with a membership degree of 1.0, but the membership degree drops to 0.7 when overloaded to 120 units. The impact of uncertainties (such as the probability of material delivery delays and the risk of sudden equipment failures) on the plan is quantified through membership functions.

[0039] Finally, the scheduling engine performs calculations with multiple objectives, including minimizing on-time order delivery rate, overall equipment utilization rate, and changeover frequency. Heuristic algorithms (such as improved genetic algorithms) are used to solve for the Pareto optimal solution set within seconds, generating executable daily / weekly plans.

[0040] The purpose of this design is that traditional fixed production schedules cannot respond to disturbances such as equipment downtime and order changes. This module achieves dynamic reconfiguration of the plan through flexible modeling. It establishes a three-dimensional mapping relationship between orders, processes, and resources to avoid the risk of line stoppages caused by equipment contention and material supply disruptions. While maximizing capacity utilization, it reserves reasonable buffers to cope with uncertainties (such as controlling the load of key equipment within the 85% range).

[0041] The advantages of this design are that it improves the on-time order fulfillment rate, reduces the response time for urgent orders from hours to minutes, reduces equipment idle rate, increases the material availability rate, and reduces production line changeover waste caused by plan adjustments.

[0042] Example 4: The system also includes a task allocation module. The task allocation module is used to dynamically allocate and optimize the tasks set by the production planning module based on the real-time availability of execution layer equipment and the load of operators, and to issue task instructions to specific equipment and personnel.

[0043] Specifically, the task allocation module operates based on a real-time perception-dynamic decision-making-closed-loop execution mechanism, and its core principle is as follows: First, there is multi-dimensional status perception, which obtains key parameters such as the operating status (idle, processing, fault code), remaining processing time, and tool life count of execution layer equipment such as machining centers and laser cutting machines in real time through the equipment PLC interface. The integrated personnel management system monitors operators' skills and qualifications (such as five-axis machine tool operation licenses), current task progress, and physiological load (such as continuous working time alarms).

[0044] Then comes constraint-driven dynamic optimization, which breaks down tasks in the production plan into process units and establishes a mapping relationship between task attributes (process requirements, priority, delivery date) and resource capabilities (equipment accuracy level, personnel skills, tooling and fixture matching). Using rule engines (such as finite capacity scheduling rules) and operations research algorithms (such as mixed integer programming), the task sequence is rearranged in real time to minimize the total delay time and maximize equipment utilization when disturbances such as equipment failure and order insertion occur.

[0045] Finally, tiered instructions are issued. For high-priority urgent tasks, expedited instructions are sent directly to the equipment through the MES system to seize the resources of current low-priority tasks. For routine tasks, work orders are generated according to the optimization results and pushed to the equipment HMI interface and personnel mobile terminals simultaneously to ensure unambiguous reception at the execution end.

[0046] The purpose of this design is to address the inefficiencies inherent in traditional manual task assignment, such as assigning high-precision tasks to low-performance equipment and consuming skilled workers with trivial tasks. This module ensures that high-value resources are dedicated to high-value processes through precise task-resource matching rules. In the event of sudden equipment failures, material shortages, or other abnormal operating conditions, the task flow can be quickly reconstructed without manual intervention, maintaining the continuous operation of the production line. Based on equipment OEE (Overall Equipment Effectiveness) thresholds and personnel fatigue models, the workload of each node is dynamically balanced to avoid quality risks or safety hazards caused by localized overload.

[0047] The advantages of this design are: by clustering intelligent tasks and processing similar processes together, the equipment switchover preparation time is reduced; relying on the automatic task migration mechanism, the recovery time from abnormal downtime is reduced from an average of 2 hours to less than 20 minutes; key processes (such as the finishing of sealing surfaces) are 100% allocated to high-stability equipment with a CPK (process capability index) ≥1.67; and processing errors caused by mismatch of personnel skills are reduced.

[0048] Example 5: The system also includes an equipment management module and a quality inspection module. The equipment management module is used to monitor the real-time operating status of the execution layer equipment, and to perform edge state prediction through equipment data collected by the control layer, so as to realize remote equipment control, energy efficiency analysis and fault diagnosis. The quality inspection module uses automated inspection equipment and vision inspection system to perform real-time online inspection of the shell products, and automatically associates and binds them with process parameters, equipment status and task batch data in the data layer to form traceable quality records and support defect pattern analysis.

[0049] Specifically, the equipment management module is designed based on edge intelligence and a state prediction mechanism. The core technical principles are as follows: First, multi-source data collaborative sensing: real-time collection of equipment operating data (spindle current, guide rail vibration frequency, lubricating oil temperature, etc., over 200 parameters) via industrial IoT terminals; deployment of edge computing nodes to compress high-dimensional data, reducing transmission latency to milliseconds. Second, state prediction model construction: establishing degradation prediction models for equipment health indicators (such as spindle bearing wear coefficient and servo motor life index); training decision tree algorithms using historical fault data to identify abnormal patterns in advance; and a threshold mechanism based on working condition self-learning (e.g., triggering an early warning if spindle vibration intensity exceeds the baseline by 30% for 5 minutes). Finally, a closed-loop control strategy: remote intervention for controllable anomalies (e.g., compensating feed rate to suppress tool chatter); and triggering maintenance work orders for predictive faults (e.g., ball screw fatigue risk) and synchronizing with the production scheduling module.

[0050] The purpose of this design is to establish a baseline model of equipment energy efficiency (such as no-load power curve), automatically shut down inefficiently operating equipment, reduce unplanned downtime by more than 85% through early fault intervention mechanisms, and increase the service life of core components such as spindles and guide rails by 20% through predictive maintenance strategies.

[0051] The benefits of this design include a 75% reduction in emergency repair costs and spare parts procurement expenses for sudden equipment failures. The overall equipment effectiveness (OEE) of the core machining center is increased to over 90%. Precision drift caused by equipment deterioration is reduced by 95%.

[0052] The design principle of the quality inspection module is to adopt a full-element traceability and defect pattern closed-loop mechanism, the specific details of which are as follows: First, online inspection technologies are integrated. The vision system performs 100% online scanning of 3,000 inspection items, including housing assembly gaps and geometric tolerances; the laser measuring instrument performs high-precision point cloud comparison on key mounting surfaces (resolution 0.001mm). Next, multi-source data is intelligently correlated, establishing a four-dimensional binding relationship between quality data, process parameters, equipment fingerprints, and operators (e.g., automatically associating out-of-tolerance workpieces with the machine tool number and spindle temperature rise data during machining); and using a time-series database to achieve millisecond-level backtracking of batch data. Finally, defect root cause diagnosis is implemented, using a defect pattern classifier based on the random forest algorithm (e.g., achieving a 99% recognition rate for typical defects such as cutting vibration marks, tool chip adhesion, and clamping deformation); automatically generating SPC (Statistical Process Control) reports and pushing them to the responsible process unit.

[0053] The purpose of this design is to solve the problem of hidden defects, such as microcracks, flowing into the next process under the traditional sampling inspection model; and to drive dynamic optimization of process parameters with quality data, such as automatically triggering the revision of cutting parameters due to abnormal surface roughness of the sealing surface.

[0054] The advantages of this design are that the scrap rate is reduced to below 0.15%, the critical dimension CPK (process capability index) is stabilized above 1.67, and the batch quality analysis time is reduced from 8 hours of manual analysis to 10 minutes of automatic report generation by the system.

[0055] Example 6: The system also includes a data analysis module. The data analysis module is used to collect the shell design, planning, execution, equipment and quality data stored in the data layer, perform correlation mining and modeling analysis, and provide dynamic decision support for production planning, quality control and process optimization in the management layer.

[0056] Specifically, the core of the data analysis module lies in end-to-end data fusion and deep decision modeling. Its operating principle is based on a three-layer architecture: full-domain data integration, correlation modeling and analysis, and decision generation. Full-domain data integration involves real-time extraction of heterogeneous data sources from the data layer, such as design documents (CAD parameters), scheduling plans (work order priorities), execution logs (equipment OEE), and quality reports (defect codes). Cross-process data associations are established through timestamps and batch numbers, such as design change version number - process parameter adjustment record - corresponding batch quality inspection results. Correlation modeling and analysis involves constructing feature engineering pipelines to extract key indicators (such as equipment vibration spectrum characteristic values ​​and thermal maps of spatial distribution of quality defects). Time-series analysis algorithms (such as LSTM prediction models) are applied to locate the correlation between equipment performance degradation and batch quality fluctuations; for example, when the slope of the spindle temperature rise trend is >0.5℃ / hour, the probability of borehole diameter deviation increases by 80% after 3 days. Decision generation, based on clustering and attribution analysis, generates actionable suggestions, such as reducing the cutting speed in thin-walled areas by 15% to reduce deformation risk; a decision tree rule base is established to dynamically optimize management strategies, such as triggering the capacity assessment process if the plan achievement rate is less than 90% for three consecutive days.

[0057] The purpose of this design is to connect data from discrete stages such as design, planning, production, and quality, eliminating the break in the causal chain caused by traditional segmented management, such as the inability to pinpoint the root cause of process delays. Through multi-dimensional data cross-analysis, process bottlenecks can be identified early, such as a 12% decrease in overall line efficiency due to a material change; the analysis conclusions are fed back to upstream modules in real time, such as adjusting the process parameter library based on the distribution of quality defects.

[0058] The advantages of this design are: compared to traditional manual traceability meetings, the root cause analysis of production anomalies is shortened from an average of 4 hours to within 10 minutes; the process optimization solution generation cycle is compressed from weekly to real-time dynamic iteration. Bottleneck equipment utilization is improved by 28% through a task-equipment matching model; and material waste in high-defect processes is reduced by 42% based on a quality-cost correlation model. A parameter self-optimization closed loop is established, such as automatically updating the cutting parameter recommendation table for every 10,000 products.

[0059] Example 7, a production process management method based on a shell, includes the following steps: S1: Utilize the design management module, combined with CAD / CAM systems and AI models, to perform shell design and process planning, and generate optimized process documents; S2: Using the production planning module, based on order requirements, the process documents, and real-time resource constraints, construct and solve the scheduling model to generate a global production plan; S3: Using the task allocation module, combined with the real-time status of equipment and personnel at the execution layer, tasks in the production plan are dynamically allocated to specific workstations; S4: The control layer drives the execution layer devices to perform production tasks, while simultaneously collecting equipment status and production progress data in real time; S5: Use the quality inspection module to perform online inspection of the produced shells and record quality data by associating it with production batch information; S6: Utilize the data analysis module to integrate and analyze design, planning, execution, equipment, and quality data, and generate optimization decisions that are then fed back to management.

[0060] Specifically, the implementation process of this method includes the following steps: In the process design and optimization module, the design management module analyzes the 3D model of the shell through AI model, automatically matches similar features (such as thin-walled structure and curved surface curvature) in historical cases, and generates an initial process plan; further, combined with constraints such as material properties and equipment performance, it performs multi-objective iterative optimization of cutting parameters (such as feed rate and speed), outputs executable process documents and synchronizes them to the production planning module.

[0061] The production planning module generates and dynamically schedules production plans based on order delivery dates, process durations in process documents, and current resource occupancy. It constructs a fuzzy constraint model (e.g., using membership functions to describe equipment capacity elasticity) and calculates the optimal production sequence. The task allocation module monitors equipment status changes (such as sudden failures or efficiency deviations) in real time during plan execution and achieves dynamic correction by rearranging processes, adjusting task priorities, or switching to backup equipment.

[0062] The production process is controlled in a closed loop. The control layer uses a sensor network to monitor the machining status in real time. For example, it can analyze vibration signals to determine whether the tool has reached the replacement threshold. If an abnormality is detected, an alarm is triggered immediately and the task is suspended. At the same time, the edge controller adaptively adjusts the equipment parameters (such as reducing the feed rate) based on real-time machining data (such as cutting force fluctuations) to avoid dimensional deviations caused by tool wear.

[0063] Quality traceability and feedback optimization: The quality inspection module uses a machine vision system to measure key dimensions of the housing (such as hole diameter tolerance and assembly surface flatness) online. The inspection data is automatically associated and stored with information such as process parameters, equipment number, and operators. The data analysis module traces the root causes of abnormal batches (e.g., identifying temperature anomalies in a certain piece of equipment within a specific time period) and feeds the conclusions back to management, triggering process parameter revisions or equipment maintenance plan updates, thus forming a continuous improvement closed loop.

[0064] Example 8, the steps of using the design management module to design the shell and plan the process, and generate optimized process documents specifically include: calling the AI ​​model based on the historical successful case library to intelligently match and iteratively optimize the parameters for the current design requirements, and outputting process documents containing key parameter settings.

[0065] Specifically, the core of this step lies in knowledge transfer and constraint-driven parameter optimization, which includes: feature matching engine, parameter iteration mechanism, and process document generation.

[0066] The feature matching engine extracts the geometric features of the current shell 3D model (such as surface curvature, thin-wall region thickness, and aperture tolerance requirements), quantifies them into feature vectors, and retrieves successful cases with similarity > 85% from the historical case library (based on the Euclidean distance algorithm) to match the corresponding process templates and parameter combinations.

[0067] The parameter iteration mechanism uses the matched case as the initial value and optimizes the parameters according to the current working condition constraints (material hardness variation, equipment model difference): for example, in the historical case, the milling feed rate of aluminum alloy shell is 0.15mm / r - the current material is detected as high silicon aluminum alloy - the cutting force model calculation is triggered - optimized to 0.12mm / r to avoid chipping; through multi-objective optimization algorithms (such as NSGA-II), the machining efficiency (upper limit of spindle speed), surface quality (roughness Ra≤0.8μm), tool life (cutting force threshold) and other indicators are balanced.

[0068] Process document generation outputs an executable process card containing process sequence, recommended values ​​for cutting parameters (speed / feed / depth of cut), and error prevention points (such as clamping pressure ≤0.5MPa for thin-walled areas); it also marks parameters that exceed historical safety boundaries and mandates manual review.

[0069] The purpose of this design is to address the shortcomings of traditional process design, which relies on individual engineer experience and is prone to design flaws due to cognitive blind spots (such as ignoring resonance caused by batch-to-batch material variations). By systematically utilizing historical best practices, it ensures that design benchmarks align with mass production verification conclusions. This avoids repeated modifications to process parameters on prototypes (a process that traditionally requires 3-5 rounds of debugging) and directly guides first-piece production through pre-verified parameter combinations. It transforms discrete expert experience into a quantifiable and iterative digital process library.

[0070] The advantages of this design are: the typical shell process design time is reduced from 8 hours to within 1.5 hours, with AI automatically generating 80% of the basic parameters; the error rate of process document versions is reduced by 90%, and the latest CAD model and equipment parameters are automatically associated; the first-piece pass rate is increased to over 98%, and the historical case library avoids known failure modes; the critical dimension process capability CPK≥1.67, and the risk of out-of-tolerance is suppressed through parameter boundary constraints; material waste in trial production is reduced by 60%, and the number of trial cuts is reduced to less than 1; tool life is extended by 30%, and intelligent matching of cutting parameters avoids abnormal wear.

[0071] In Example 9, during the step of generating a global production plan, the scheduling model considers the imprecision of equipment capacity, material supply, and personnel skills, and uses a fuzzy constraint modeling method to handle uncertainties.

[0072] Specifically, this step is based on the core logic of elastic modeling and uncertainty quantification, which includes the transformation of imprecise factors, multi-constraint collaborative optimization, and perturbation adaptive mechanism.

[0073] The transformation of imprecise factors involves converting traditional rigid parameters such as equipment capacity, material supply cycle, and personnel skill level into quantifiable flexible ranges. For example, when the daily capacity of equipment is under normal operating conditions, the membership degree is 1.0 corresponding to 100 units, and when the load reaches 120 units, the membership degree drops to 0.8. Fuzzy membership functions are established to quantify uncertainty. The probability of equipment failure is fitted with historical downtime data as a triangular fuzzy number with a minimum of 0.5% and a maximum of 2%. The material delay risk is divided into confidence intervals according to supplier rating.

[0074] Multi-constraint collaborative optimization involves constructing a three-dimensional elastic constraint network of equipment, materials, and personnel, setting dynamic weights such as a weight of 0.9 for the capacity constraint of critical equipment and 0.6 for the constraint of ordinary materials; and using a fuzzy programming algorithm to solve for the feasible solution with the highest confidence, taking into account both efficiency and robustness.

[0075] The disturbance adaptive mechanism monitors changes in constraints in real time, such as sudden equipment failures causing the capacity membership to drop sharply from 0.9 to 0.3, triggering online rescheduling of plans; and absorbs fluctuations through buffer pool design, such as reserving a 15% flexible capacity pool to cope with emergency order insertions.

[0076] The purpose of this design is to ensure that rigid scheduling models completely fail in the event of equipment failure or material delays, while fuzzy constraint modeling, through flexible intervals, guarantees that over 80% of core tasks can still be executed. This avoids pursuing 100% equipment utilization at the expense of zero planning tolerance (e.g., controlling the load on critical equipment within a safe range of membership ≥ 0.8). It effectively addresses the constant uncertainties such as supplier fluctuations and personnel turnover in multi-variety, small-batch production.

[0077] The benefits of this design are: 70% fewer planning interruptions due to sudden disturbances, equipment failures, and material shortages; on-time delivery rate of core orders >95%; response time for emergency orders reduced from hours to within 10 minutes; decreased equipment idle rate; increased material availability rate; reduced manual adjustment time for planners; and reduced capacity waste due to planning errors.

[0078] In Example 10, in the step of the control layer driving the execution layer equipment to perform production tasks and collect data in real time, the control layer performs equipment edge state prediction and localized error prevention control based on sensor data, and provides real-time feedback of status information to the management layer and equipment management module; at the same time, the key defects detected in the quality inspection step will automatically trigger the task allocation module to pause or adjust the task flow of the relevant workstation.

[0079] Specifically, the core of this mechanism lies in constructing a real-time closed-loop control chain and a defect propagation blocking network, namely edge intelligent prediction error prevention and quality defect linkage control. Edge-based intelligent prediction and error prevention includes both the device layer and the data channel. The device layer involves embedding lightweight analysis models into execution units such as CNC systems to analyze sensor data (such as spindle vibration spectra and temperature rise curves) in real time. When an abnormal pattern is detected (such as a surge in energy at the characteristic frequency of tool wear), localized control is immediately triggered (automatically reducing the feed rate by 30% to prevent tool breakage), without waiting for instructions from the upper layer. The data channel transmits key status parameters (current fluctuation trends, positioning errors) back to the device management module via a 5G private network in milliseconds, forming a device health profile.

[0080] Quality defect linkage control includes a defect binding mechanism and task freezing logic. The defect binding mechanism refers to the vision system automatically associating the workpiece's processing equipment ID, operator number, and raw material batch number with the detected porosity defect on the sealing surface, locking the data chain. The task freezing logic means that when a defect exceeds a threshold, such as a critical dimension exceeding the tolerance by 0.1mm, the entire batch of tasks is immediately frozen. A blocking command, such as "pause task at station 3 of equipment A," is sent to the task allocation module, and a root cause analysis process, such as tracing the melting temperature, is triggered.

[0081] The purpose of this design is to proactively intercept 75% of potential quality issues caused by parameter drift, such as cutting vibration marks, through edge control of the equipment; and to immediately freeze the task flow for any escaped defects, preventing defective parts from flowing into the assembly process and causing hundreds of times the loss. Directly connecting equipment-level sensor data and quality inspection results with the production scheduling system avoids the lag caused by traditional manual fault reporting.

[0082] The advantages of this design are that the incidence of batch quality accidents is reduced to zero, defective batches are automatically frozen within 30 seconds; rework costs are reduced, serious failures such as abnormal spindle breakage are reduced, the duration of unexpected downtime is reduced, the root cause of defects is traced from cross-departmental meetings to automatic system association, and process parameter optimization instructions are pushed to the equipment for execution within 1 minute.

[0083] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0084] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A shell-based production process management system, characterized in that: It consists of a four-layer structure: data layer, management layer, execution layer, and control layer. The data layer is used to store data from the entire shell production process, including design documents, material information, process parameters, equipment status, and quality inspection results, and provides interfaces for data storage, querying, and updating. The management layer is used for management decisions in shell production, including production planning, task allocation, resource allocation, progress monitoring and quality control, and interacts with the lower layer through the Manufacturing Execution System (MES). The execution layer includes physical equipment units for housing processing, and the physical equipment units include at least CNC machining equipment, laser cutting machines, automated assembly lines, and testing equipment; The control layer is used for real-time monitoring and stable control of the shell production process, including a programmable logic controller (PLC), a sensor network, and an automated control system. The control layer is connected to the execution layer devices via industrial Ethernet or fieldbus. The data layer and the management layer interact through a database and middleware. The management layer and the execution layer communicate through the MES system to transmit production instructions and task status. The execution layer and the control layer communicate through the industrial Ethernet or fieldbus to issue instructions and provide status feedback.

2. The shell-based production process management system according to claim 1, characterized in that: It also includes a design management module, which integrates a CAD / CAM system for rapid design of shell structures and processing technology and generation of process documents, and supports intelligent iterative optimization of design schemes and process parameters based on AI models of historical case sets.

3. The shell-based production process management system according to claim 1, characterized in that: It also includes a production planning module, which is used to construct a scheduling model with fuzzy constraints based on the shell production order requirements, the process documents generated by the design management module, and the real-time resource status information in the resource library, and generate a production plan that integrates multiple orders and multiple tasks.

4. The shell-based production process management system according to claim 1, characterized in that: It also includes a task allocation module, which is used to dynamically allocate and optimize the tasks set by the production planning module based on the real-time availability of execution layer equipment and the load of operators, and to issue task instructions to specific equipment and personnel.

5. A shell-based production process management system according to claim 1, characterized in that: It also includes an equipment management module and a quality inspection module; The device management module is used to monitor the real-time operating status of the execution layer devices, and to perform edge state prediction through device data collected by the control layer, so as to realize remote device control, energy efficiency analysis and fault diagnosis. The quality inspection module uses automated inspection equipment and a vision inspection system to perform real-time online inspection of the shell products, and automatically associates and binds them with process parameters, equipment status, and task batch data in the data layer to form traceable quality records and support defect pattern analysis.

6. A shell-based production process management system according to claim 1, characterized in that: It also includes a data analysis module, which is used to collect data on the entire process of shell design, planning, execution, equipment and quality stored in the data layer, and to perform correlation mining and modeling analysis to provide dynamic decision support for production planning, quality control and process optimization in the management level.

7. A shell-based production process management method, comprising a shell-based production process management system according to any one of claims 1-6, characterized in that, Includes the following steps: S1: Utilize the design management module, combined with CAD / CAM systems and AI models, to perform shell design and process planning, and generate optimized process documents; S2: Using the production planning module, based on order requirements, the process documents, and real-time resource constraints, construct and solve the scheduling model to generate a global production plan; S3: Using the task allocation module, combined with the real-time status of equipment and personnel at the execution layer, tasks in the production plan are dynamically allocated to specific workstations; S4: The control layer drives the execution layer devices to perform production tasks, while simultaneously collecting equipment status and production progress data in real time; S5: Use the quality inspection module to perform online inspection of the produced shells and record quality data by associating it with production batch information; S6: Utilize the data analysis module to integrate and analyze design, planning, execution, equipment, and quality data, and generate optimization decisions that are then fed back to management.

8. The production process management method based on a shell according to claim 7, characterized in that: The steps of using the design management module for shell design and process planning, and generating optimized process documents, specifically include: calling the AI ​​model based on the historical successful case library to intelligently match and iteratively optimize the parameters for the current design requirements, and outputting process documents containing key parameter settings.

9. A production process management method based on a shell according to claim 7, characterized in that: In the step of generating a global production plan, the scheduling model considers the imprecision of equipment capacity, material supply, and personnel skills, and uses a fuzzy constraint modeling method to handle uncertainties.

10. A production process management method based on a shell according to claim 7, characterized in that: In the step of the control layer driving the execution layer equipment to perform production tasks and collect data in real time, the control layer performs equipment edge state prediction and localized error prevention control based on sensor data, and provides real-time feedback of status information to the management layer and equipment management module. At the same time, the key defects detected in the quality inspection step will automatically trigger the task allocation module to pause or adjust the task flow of the relevant workstation.

Images