An intelligent management system for supercritical foamed protective product production

Abstract

The application discloses an intelligent management system for supercritical foaming protective product production, and relates to the technical field of industrial process data management.The system comprises a demand analysis module, a process constraint generation module, a rolling scheduling module, an execution coordination module, a quality state estimation module and an abnormality closed loop module.The demand analysis module converts orders into batch quality targets.The process constraint generation module determines process relationships and generates feasible process windows.The rolling scheduling module locks scheduling results into executable tasks and scheduling versions under resource constraints.The execution coordination module issues process packages containing batch identifiers, parameter boundaries and interlocking conditions.The quality state estimation module aligns process data with process packages to estimate quality states.The abnormality closed loop module generates correction schemes and verifies them when deviations exceed threshold values.The tracing module stores full-process data.The application aims to solve the problems of scheduling disconnection from processes and difficult tracing of quality abnormalities in supercritical foaming protective product production, and realizes full-process closed loop management from orders to delivery.

Description

Technical Field

[0001] This invention relates to the field of industrial process data management technology, and in particular to an intelligent management system for the production of supercritical foamed protective products. Background Technology

[0002] With the rapid development of sports protection, medical devices, automotive lightweighting and aerospace, the demand for supercritical foam protective products in terms of cushioning, energy absorption, lightweighting and resilience continues to grow, and the market's requirements for product consistency and traceability are also increasing.

[0003] In existing technologies, the production of supercritical foamed protective products involves order management, manufacturing execution, and equipment control systems operating independently, lacking a unified data linkage mechanism. Performance indicators such as protection levels, peak impact limits, and absorbed energy included in orders rely on manual experience to be converted into executable process parameters for the equipment. This fails to adequately consider performance differences between material batches and the capability limits of equipment such as autoclaves and molds, leading to frequent disconnects between production scheduling and process feasibility. When production plans are adjusted, existing systems typically perform a global recalculation of all incomplete tasks, increasing computational overhead and causing frequent changes to already issued tasks, directly impacting on-site production order. Furthermore, the lack of a version control mechanism makes it impossible to effectively identify the corresponding process parameter versions for each task after production plan changes. The risk of incorrect process packages being executed by equipment is difficult to avoid, and the scope of affected tasks cannot be accurately defined in the event of equipment failure, order changes, or quality anomalies. On the other hand, the supercritical foaming production process involves multiple steps such as impregnation, pressurization and holding, depressurization and nucleation, temperature-controlled foaming, and shaping. Real-time operational data from various equipment and finished product testing data are scattered across different systems, lacking a batch-indexed correlation mechanism. When quality anomalies occur, it is impossible to quickly pinpoint the deviation to a specific material batch, process package version, or parameter adjustment record. In the anomaly handling phase, directly adjusting current equipment parameters based on test results will incorrectly subject work-in-process that has already passed through controlled processes to the corrective action, leading to distorted verification conclusions and an inability to accurately evaluate the actual effectiveness of the corrective measures. Furthermore, the entire process data from order confirmation to finished product delivery is scattered across various isolated systems. In the event of product quality disputes or process verification needs, it is difficult to construct a complete chain of evidence within a limited timeframe, hindering quality traceability efficiency and the ability to continuously optimize the production process.

[0004] Therefore, this invention proposes an intelligent management system for the production of supercritical foamed protective products. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent management system for the production of supercritical foamed protective products, in order to solve the problems in the production of supercritical foamed protective products in the prior art, such as fragmented multiple systems, scattered data, disconnect between production planning and process feasibility, inability to quickly locate quality anomalies to specific batches, lack of effective batch identification mechanism for correction and verification, and difficulty in full-process traceability.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] An intelligent management system for the production of supercritical foamed protective products includes:

[0008] Demand Analysis Module: Converts the product structure, material grade, protection level, target energy absorption performance, dimensional tolerance, required quantity, and delivery time limit in protective product orders into batch quality targets with order identifiers;

[0009] Process constraint generation module: Based on batch quality objectives, material batches and product structure, determine the sequence of supercritical fluid impregnation, pressure increase and holding, pressure release and nucleation, temperature-controlled foaming and shaping processes, and generate batch feasible process windows;

[0010] Resource status modeling module: Obtains the available time period, capacity parameters, occupancy status, maintenance status and safety interlock status of manufacturing resources, forming a set of manufacturing resource statuses with timestamps;

[0011] Rolling production scheduling module: Under the constraints of process sequence, batch feasible process window and manufacturing resource status set, rolling production scheduling is performed on the batch to be produced, the equipment allocation, process start and end time and batch switching order of each batch are determined, and the scheduling results located within the preset commitment time domain are locked as executable tasks and scheduling versions;

[0012] Execution coordination module: Generates production process packages based on executable tasks and sends them to the corresponding equipment controllers. The production process package includes at least batch identifier, equipment identifier, process sequence, target parameters, parameter allowable boundaries, and interlocking conditions.

[0013] Quality status estimation module: According to batch identification and material flow delay, the pressure, temperature, supercritical fluid flow, product size, product quality and surface image collected during the production process are aligned with the corresponding production process package to generate batch quality status containing density data from multiple sampling locations, compression displacement data under specified load, unloading recovery data and cycle number data, and estimate foaming ratio, density uniformity, dimensional stability and compression rebound state based on batch quality status;

[0014] Anomaly closed-loop module: When the estimated deviation exceeds the threshold, it generates a parameter correction plan or rearrangement plan based on the batch feasible process window, and verifies the correction result based on the batch quality status after the correction takes effect.

[0015] Preferably, the requirement analysis module establishes a mapping relationship between protective performance indicators and foaming quality indicators, converting at least two of the impact peak, absorbed energy, rebound rate, and durability cycle count into target apparent density, target foaming ratio, allowable density dispersion, allowable permanent compression deformation rate, and dimensional shrinkage rate; when there are multiple candidate quality targets in the mapping results, the manufacturability of the candidate quality targets is screened based on the material's processability range and equipment capabilities.

[0016] Preferably, the process constraint generation module reads the preset process route according to the product structure, determines the sequential relationship of the impregnation, pressure increase and holding, pressure release and nucleation, temperature-controlled foaming and shaping processes; and determines the upper and lower boundaries of pressure, temperature, holding time and pressure release rate based on the melting temperature zone, crystallization characteristics, initial density and moisture content of the material batch; and obtains the batch feasible process window by intersecting the upper and lower boundaries through the rated pressure of the equipment, heating rate, flow rate limit, mold pressure resistance level and safety margin; when the intersection result is an empty set, it outputs an unschedulable production mark and conflict constraints that cause the empty set.

[0017] Preferably, the manufacturing resources in the resource status modeling module include: autoclave, booster pump, temperature control device, mold, transfer device, and post-processing equipment.

[0018] Preferably, the rolling scheduling module takes the weighted sum of delivery delay costs, product changeover costs, equipment idle energy consumption, and quality uncertainty costs as its scheduling objective, and takes the following hard constraints as follows: equipment executes only one task at a time, subsequent processes can only begin after the preceding process is completed, the process parameters of each process selected for scheduling are within the batch feasible process window, and the safety interlock status is effective. When equipment failure, order change, or quality deviation events occur, only tasks outside the preset commitment time domain and tasks affected by the event are re-solved.

[0019] Preferably, the execution coordination module assigns a version number and a verification value to each production process package. The equipment controller allows the task to start after verifying that the batch identifier, equipment identifier, version number, and parameter allowable boundaries are consistent. When the measured pressure, temperature, or pressure relief rate exceeds the parameter allowable boundaries, or the safety interlock status fails, the equipment controller refuses to continue executing the corresponding parameters and sends an abnormal event with a device timestamp to the abnormal closed-loop module.

[0020] Preferably, the quality status estimation module calculates the material flow delay based on the equipment controller timestamp, material entry time, pressure holding duration, transfer duration, and detection time, and divides the continuous sampling data into time windows corresponding to the batch; for missing data, it outputs the data completeness and estimated confidence level, and when the estimated confidence level is lower than a set threshold, it prohibits the automatic generation of parameter correction schemes that expand the parameter change range based on the estimation result.

[0021] Preferably, the abnormal closed-loop module constructs deviation characteristics based on the supercritical fluid dissolution amount, temperature and pressure stability during the pressure holding stage, pressure change rate during the pressure release stage, foaming temperature trajectory, and shaping cooling rate, and determines at least one candidate cause according to the process influence relationship; for each candidate cause, it solves for candidate correction quantities that reduce the predicted quality deviation and limit the parameter change within the batch feasible process window, and selects a parameter correction scheme based on the credibility of the candidate cause, the predicted improvement amount, and the execution risk.

[0022] Preferably, the abnormal closed-loop module takes the moment when the parameter correction instruction is confirmed to be effective by the equipment controller as the starting point, and determines the valid batch based on the product position, remaining holding time and transfer relationship within the equipment; when a preset number of valid batches all meet the batch quality target, the correction is confirmed to be effective; when the deviation of the valid batches is not reduced or a new safety constraint conflict occurs, the parameter correction scheme is revoked and a rearrangement is triggered.

[0023] The present invention also includes a batch traceability module, which is used to associate and store order identifier, batch identifier, material batch, manufacturing resource status set, production scheduling version, production process package, batch quality status, parameter correction scheme and verification results as batch traceability records.

[0024] Compared with the prior art, the beneficial technical effects of the present invention are as follows:

[0025] This invention converts order protection performance into batch quality targets through a demand analysis module, and combines it with a process constraint generation module to form a feasible process window that is jointly constrained by materials, equipment and safety conditions, thereby reducing the disconnect between production scheduling and equipment executable parameters.

[0026] This invention maintains the stability of recent tasks and prevents erroneous process packages from being executed by equipment by committing to rolling production scheduling in the time domain and verifying process package versions.

[0027] This invention aligns process data with inspection data by using batch identification and material flow delay, enabling quality deviations to be located at specific batches, equipment, processes, and process package versions.

[0028] This invention improves the reliability of closed-loop regulation by generating a correction scheme within the safe and feasible domain and identifying the first valid batch for verification, thereby avoiding the use of work-in-process that has not yet been corrected to verify the correction effect.

[0029] This invention uses a batch traceability module to associate and store orders, materials, resources, tasks, process packages, quality status, and correction verification results, providing a consistent data foundation for product quality traceability, process review, and production plan optimization. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of an intelligent management system for the production of supercritical foamed protective products provided in an embodiment of the present invention;

[0031] Figure 2 This is a data mapping diagram of order targets, batch process windows, and production process packages in an intelligent management system for the production of supercritical foamed protective products provided in an embodiment of the present invention. Detailed Implementation

[0032] 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.

[0033] The present invention will be further described below with reference to the accompanying drawings and embodiments. The example parameters, thresholds, weights, time ranges, and model types are non-limiting examples and should be configured according to material properties, equipment rated capacity, product standards, and enterprise safety procedures.

[0034] Example 1

[0035] Please refer to Figure 1 In this embodiment, the system is deployed on a foaming production management server and connected to the autoclave controller, booster pump controller, temperature control device, pressure relief valve group, mold identification terminal, transfer terminal, post-processing equipment, and testing terminal via an industrial communication network. The server can consist of one or more processors, memory, database, and communication interfaces. The equipment control loop is still executed by a field controller with safety certification. The management server only issues setpoints or tasks within permissible boundaries and does not replace the overpressure, overtemperature, and emergency stop interlocks of the autoclave itself.

[0036] Please refer to Figure 2The system generates an order identifier O for each order and a batch identifier B for each split production batch. A batch record must contain at least the product model, material batch M, target quantity, and target quality vector. Process route R, feasible process window Production version and process package version Resource records must include at least the resource identifier (E), resource type, capability parameters, calendar, current task, maintenance status, interlock status, and update time. An order identifier is used to uniquely identify each order, and a batch identifier is used to uniquely identify each production batch.

[0037] The target mass vector can be represented as ,in Indicates the apparent density of the target. Indicates the target foaming ratio. Indicates the allowable density dispersion. Indicates the allowable permanent compressive deformation rate. This indicates the allowable dimensional shrinkage rate. Specific indicators and their quantities vary depending on the product type; this formula only illustrates the data relationships.

[0038] The requirements analysis module reads protection level, peak impact limit, absorbed energy, dimensions, and delivery deadline from the order or product data management system. It establishes a mapping relationship between protection performance indicators and foaming quality indicators, converting at least two of the peak impact, absorbed energy, resilience, and durability cycle count into target apparent density, target foaming ratio, allowable density dispersion, allowable permanent compression set, and dimensional shrinkage. This mapping can be established using validated material test data, product design specifications, and historical qualified batches. When multiple candidate target quality vectors exist in the mapping results, the system performs manufacturability screening based on the material's processability range and equipment capabilities: for each candidate vector, the system retrieves the corresponding material's historical qualified range and current equipment capabilities; the candidate vector is marked as manufacturable only if at least one non-empty feasible process window exists; otherwise, the order enters the process review queue.

[0039] The resource status modeling module acquires the available time periods, capacity parameters, occupancy status, maintenance status, and safety interlock status of manufacturing resources, forming a timestamped set of manufacturing resource statuses. These manufacturing resources include autoclaves, booster pumps, temperature control devices, molds, transfer devices, and post-processing equipment. Resource records must include at least the resource identifier E, resource type, capacity parameters, calendar, current task, maintenance status, interlock status, and update time.

[0040] Feasible process window generation. The process constraint generation module reads the preset process route based on the product structure and determines the sequence of processes: impregnation, pressurization and holding, pressure relief and nucleation, temperature-controlled foaming, and shaping. Specifically, material batch attributes include material grade, initial density, moisture content, melting temperature range, crystallization characteristics, and verified supercritical fluid solubility range. Equipment capabilities include the autoclave's rated pressure and effective volume, booster pump flow range, heating and cooling rates, pressure change rate achievable by the pressure relief valve, mold pressure resistance rating, and post-processing temperature range.

[0041] Let the process parameter vector Where P is the holding pressure and T is the impregnation temperature. For the pressure holding time, This is the pressure relief rate. The foaming temperature, Foaming time, The set expression for the batch feasible process window is: (This is the set of cooling rates for the batch design.)

[0042]

[0043] in, and Given from the material and process range, This indicates the material and target quality constraints. Indicates equipment capacity constraints. Indicates safety constraints.

[0044] Based on the melting temperature range, crystallization characteristics, initial density, and moisture content of the material batch, the upper and lower boundaries of pressure, temperature, holding time, and depressurization rate are determined. These boundaries are then intersected by the equipment's rated pressure, heating rate, flow rate limit, mold pressure resistance rating, and safety margin to obtain the batch-feasible process window. The system first obtains the material's process range, then intersects it with the capacity ranges of the autoclave, mold, booster pump, and post-processing equipment, and finally narrows the pressure, temperature, and depressurization rate boundaries according to a preset safety margin. If the set is empty, the system does not generate an executable task, but instead outputs a non-schedulable flag and conflict constraints that cause the empty set, such as: "The lower limit pressure required for the target foaming ratio is higher than the upper limit pressure allowed by the current mold." Process engineers can then change the mold, adjust the product batching method, or re-evaluate the target quality vector.

[0045] Rolling production scheduling and process package issuance are implemented, with rolling scheduling executed in discrete time periods or event-driven manner. The system breaks down production batches into impregnation tasks, pressure holding tasks, pressure release tasks, foaming tasks, shaping tasks, and testing tasks, while maintaining the dependencies between these tasks. Continuous occupancy constraints are set for uninterrupted pressure holding and pressure release tasks; shared resource capacity constraints are set for shared booster pumps, transfer devices, and testing equipment.

[0046] The production scheduling objective is to minimize the weighted sum of the costs of delivery delays, product changeovers, idle equipment energy consumption, and quality uncertainties.

[0047]

[0048] Each cost is normalized before summation to a dimensionless normalized cost; the weights are dimensionless coefficients used only to characterize the relative importance of each cost in the overall objective. The total cost of achieving the production target This refers to the amount of delay in batch delivery. Cost of switching materials or molds For equipment no-load and process energy consumption, The cost of quality uncertainty; , , , The weights are configurable and non-negative. Hard constraints include that the equipment executes only one task at a time, the subsequent process can only begin after the preceding process is completed, the process parameters of each process selected for scheduling are within the batch feasible process window, and the safety interlock status is valid. Any objective function optimization must not violate these hard constraints.

[0049] The system is configured with a commitment timeframe, such as 2 hours to a shift. Tasks issued within the commitment timeframe that do not pose a security risk remain locked; tasks outside the commitment timeframe can be adjusted based on new orders, equipment status, or material status. In the event of equipment failure, order changes, or quality deviations, only tasks outside the preset commitment timeframe and tasks affected by the event are re-solved, avoiding frequent global reordering.

[0050] For locked tasks, the execution coordination module generates a production process package based on the executable task and sends it to the corresponding equipment controller. The execution coordination module assigns a version number and checksum to each production process package. A production process package includes at least a batch identifier, equipment identifier, process sequence, target parameters, parameter allowable boundaries, and interlocking conditions. The process package also includes B, E, ... , First, complete the verification process, including verifying the credentials and checksums. The equipment controller allows the task to start after confirming that the batch identifier, equipment identifier, version number, and parameter allowable boundaries are consistent. If the measured pressure, temperature, or depressurization rate exceeds the parameter allowable boundaries, or if the safety interlock fails, the equipment controller refuses to continue executing the corresponding parameters and sends an exception event with the equipment timestamp to the exception closed-loop module. The equipment controller only accepts the task after confirming that the equipment identifier, batch identifier, version, and checksum are consistent and that the current interlock is valid. The equipment returns a status of received, started, effective, completed, or rejected, along with the equipment timestamp.

[0051] Batch data alignment and quality estimation. The quality status estimation module aligns the pressure, temperature, supercritical fluid flow rate, product dimensions, product quality, and surface images collected during the production process with the corresponding production process packages according to the batch identifier and material flow delay. This generates batch quality status data including density data from multiple sampling locations, compression displacement data under specified loads, unloading recovery data, and cycle count data. Specifically, the quality status estimation module receives the pressure and temperature sequence during the pressure holding phase, the cumulative supercritical fluid flow rate, the pressure change rate during the pressure relief phase, the temperature trajectory of the foaming furnace, product quality, dimensional measurement results, and surface images. The system first establishes an initial batch quality status record based on the batch's entry time into the equipment, pressure holding duration, transfer duration, and detection time, and then writes the data for the corresponding time window into this record.

[0052] The quality status estimation module calculates material flow delay based on the equipment controller timestamp, material entry time, pressure holding duration, transfer duration, and detection time, dividing continuous sampling data into time windows corresponding to each batch. For continuous equipment, the start and end boundaries of the batch time window can be determined by the carrier code, mold code, barcode, or equipment encoder position; when only timestamps are available, the system uses the planned start time, equipment confirmation start time, and historical transfer delay estimation boundaries. Completeness is calculated after data alignment. For missing data, the system outputs the data completeness and estimated confidence level. When key pressure sections, pressure relief sections, or detection records are missing, the system lowers the estimated confidence level and does not directly increase the parameter adjustment range after filling with zeros for missing values. When the estimated confidence level is lower than a set threshold, the system prohibits the automatic generation of parameter correction schemes that increase the parameter change range based on the estimation result.

[0053] The foaming ratio, density uniformity, dimensional stability, and compression rebound are estimated based on batch quality status. The foaming ratio can be calculated as the ratio of the original material density to the apparent density after foaming; density uniformity can be estimated using the coefficient of variation of density from multiple sampling locations within the same batch; dimensional stability can be calculated from dimensional changes before and after shaping and after a specified settling time; and compression rebound can be calculated from compression displacement under a specified load, unloading recovery, and the number of cycles. The system can use physical rules, regression models, or independently validated machine learning models for estimation.

[0054] Let the measured or estimated value of the k-th quality indicator be... The goal is The allowable deviation is The formula for the overall deviation is as follows:

[0055]

[0056] in, Non-negative weights and Any safety-related item exceeding its limit will directly trigger an exception; non-safety indicators can be... An exception is triggered when the threshold is exceeded.

[0057] The abnormal cause and parameter correction: When the estimated deviation exceeds the threshold, the abnormal closed-loop module generates a parameter correction plan or rearrangement plan based on the batch feasible process window, and verifies the correction result based on the batch quality status after the correction takes effect.

[0058] Specifically, the anomaly closed-loop module constructs deviation characteristics based on the supercritical fluid dissolution rate, temperature and pressure stability during the pressure holding stage, pressure change rate during the pressure release stage, foaming temperature trajectory, and shaping cooling rate, and determines at least one candidate cause according to the process influence relationship. When the density dispersion exceeds the limit, the system retrieves pressure holding temperature fluctuations, cumulative supercritical fluid volume, pressure release rate fluctuations, foaming temperature field differences, and mold loading methods related to density uniformity from the process influence relationship. Each candidate cause has corresponding data requirements, adjustable parameters, direction of action, and risk level. When the data is incomplete, the credibility of the candidate cause decreases accordingly.

[0059] For each candidate cause, within the batch feasible process window, the system solves for candidate corrections that reduce predicted quality deviation and limit parameter variation. The parameter correction scheme is then selected based on the candidate cause's credibility, predicted improvement amount, and execution risk. For candidate causes with a credibility threshold, the system... Internal solution candidate correction This causes prediction bias. Reduce, while satisfying safety interlock constraints and:

[0060]

[0061] in, Let i be the change in the i-th parameter. This represents the maximum allowable variation range for this parameter. The selection function can comprehensively consider the credibility of candidate causes, predicted improvement amount, adjustment range, expected energy consumption, and execution risk. If no correction amount satisfies the constraints, the system does not force parameter tuning but instead triggers a rearrangement plan involving equipment replacement, production delay, increased testing, or manual process review.

[0062] The abnormal closed-loop module takes the moment when the parameter correction command is confirmed to be effective by the equipment controller as the starting point, and determines the valid batches based on the product location, remaining holding time, and transfer relationship within the equipment. When a preset number of consecutive valid batches meet the batch quality target, the correction is confirmed to be effective. When the deviation of the valid batches is not reduced or a new safety constraint conflict occurs, the parameter correction scheme is revoked and a rearrangement is triggered.

[0063] The time when the device controller confirms that the correction command has taken effect is recorded as follows: System read The location of each batch is recorded at all times within the autoclave, during transport, and inside the foaming furnace. Batches that have completed the corresponding controlled processes are not considered valid batches; The first batch that subsequently enters the controlled process and fully undergoes the effects of the corrected parameters is determined to be a valid batch. The system is based on And subsequent quality status verification and correction for several batches.

[0064] If N consecutive valid batches meet the target and the overall deviation decreases, the system will mark the correction scheme as valid and update the recommended process center value, but will not automatically relax the safety boundary. If the deviation does not decrease, the confidence level remains insufficient, or a new safety conflict occurs, the system will cancel the automatic correction, restore the most recently verified process package, and trigger a rearrangement or manual handling. N can be configured according to the product risk level.

[0065] For model and traceability management, if a predictive model is used, the training data consists of process characteristics, material properties, equipment status, and audited quality labels from historical batches; the model input does not include the final test values ​​that have not yet been obtained for the same batch to be predicted. Validation data is isolated from training data by material batch or time interval. Candidate models can only be released if the target metric on the independent validation set meets the release threshold and the safety-related false positive rate has not deteriorated.

[0066] Each model release saves the training data range, feature version, parameter summary, validation results, approver, and previous version. In case of model anomalies during production, a rollback to the previous validated version is possible. The model version is written to the traceability record as a field for production process packages and batch quality status.

[0067] The system also includes a batch traceability module. This module associates and stores order identifiers, batch identifiers, material batches, manufacturing resource status sets, production schedule versions, production process packages, batch quality status, parameter correction schemes, and verification results as batch traceability records. These records use non-repeatable event identifiers and timestamps. Events include at least order confirmation, material requisition, resource status update, production schedule release, process package receipt, equipment startup, parameter activation, task completion, data alignment, quality estimation, correction approval, and correction verification. The system can implement integrity verification using database version control, append-only logs, or object storage with verification digests.

[0068] Extended implementation method. The supercritical fluid can be carbon dioxide, nitrogen, or other supercritical media that meet the material foaming requirements. The production form can be batch impregnation in a kettle, continuous extrusion foaming, or a combination of production forms with equivalent process relationships. For continuous production, batch identification can correspond to material segments divided by length, time, or carrier.

[0069] Rolling scheduling can employ mixed-integer programming, constrained programming, heuristic search, reinforcement learning strategies, or combinations of these methods, but its output must satisfy hard constraints. Quality estimation can utilize physical models, statistical models, machine learning models, or model fusion, and must output the estimation results and confidence levels corresponding to each batch.

[0070] The system's modules can be deployed on the same server, edge computing nodes, or multiple collaborative devices. Modules can exchange data via message queues, industrial protocols, or database events. Regardless of the deployment method, the association between order identifiers, batch identifiers, production schedule versions, process package versions, and equipment timestamps is maintained.

[0071] Example 2

[0072] Typical operating scenarios and parameter configurations

[0073] Based on the aforementioned system deployment, this embodiment uses the complete production cycle of a batch of protective gear padding order A as an example to illustrate the typical operation process and preset parameter configuration of each module.

[0074] Order A includes product model P-100, material batch M-202605, target energy absorption performance requirements: impact peak not exceeding 10kN and absorbed energy not less than 25J, dimensional tolerance ±0.5mm, required quantity 500 pieces, and delivery time limit 120 hours. Material batch M-202605 is a TPU base material with a moisture content of 0.12% and a melting temperature range of 140℃ to 180℃. The demand parsing module splits this order into batches B01 and B02, and converts the impact peak and absorbed energy into a target mass vector using mapping rules. The expression is as follows:

[0075]

[0076] The components represent, in order, the target apparent density of 0.55 g / cm³, the target foaming ratio of 15.0, the allowable density dispersion of 5%, the allowable permanent compression set of 8%, and the allowable dimensional shrinkage of 1.5%. Based on a search of the material's workability range and equipment capabilities, this candidate vector was marked as manufacturable.

[0077] The process constraint generation module reads the preset process route based on the product structure and determines the sequence of the impregnation process, the pressure increase and holding process, the pressure release and nucleation process, the temperature-controlled foaming process, and the shaping process. Based on the melting temperature range of 140℃ to 180℃, the crystallization temperature of approximately 120℃, the initial density of 0.08 g / cm³, and the moisture content of material batch M-202605, the upper and lower limits of pressure are determined to be 12MPa to 25MPa, the upper and lower limits of temperature are 150℃ to 170℃, the upper and lower limits of holding time are 30 minutes to 90 minutes, and the upper and lower limits of pressure release rate are 5MPa / s to 30MPa / s. Further, the module intersects with the rated pressure of the autoclave E1 (30MPa) and the mold pressure resistance (20MPa), and shrinks the upper limit by a preset safety margin of 10% to obtain the batch-feasible process window. The expression is:

[0078]

[0079] Within the equipment's capabilities If not empty, proceed with production scheduling.

[0080] The rolling scheduling module addresses the sequence of production processes. Production scheduling is performed under the constraints of resource status sets. The system sets the commitment time domain to 4 hours. Tasks already issued within the commitment time domain remain locked, while tasks outside the commitment time domain are allowed to be adjusted on a rolling basis. The scheduling objective is to minimize the weighted sum of delivery delay costs, product changeover costs, equipment idle energy consumption, and quality uncertainty costs, with weights configured accordingly. 5. 3. For 1, The production schedule is as follows: B01 will be impregnated in autoclave E1 from hour 0 to hour 2, maintained under pressure in the same autoclave from hour 2 to hour 2.75, depressurized from hour 2.75 to hour 2.77, foamed in foaming furnace F1 from hour 2.77 to hour 3.27, shaped and cooled from hour 3.27 to hour 3.77, and inspected from hour 3.77 to hour 4.27. B02 will be postponed until after E1 is released. The production schedule results and version for B01 are as follows. Equal to 1.0 is locked as an executable task.

[0081] The execution collaboration module generates a production process package for B01 and assigns a version number. The values ​​are equal to 1.0 and the checksum CRC32. The production process package includes batch identifier B01, equipment identifier E1, process sequence, and target parameters including holding pressure 15MPa, impregnation temperature 160℃, holding time 45 minutes, pressure relief rate 15MPa / s, foaming temperature 155℃, foaming time 30 minutes, and setting cooling rate 5℃ / min. The allowable parameters are pressure ±2MPa and temperature ±5℃. The interlock conditions are overpressure interlock set at 18.5MPa and overtemperature interlock set at 170℃. The equipment controller allows the task to start after verifying that the batch identifier, equipment identifier, version number, and parameter boundaries are consistent.

[0082] The quality status estimation module collects the pressure sequence from 14.8MPa to 15.3MPa during the pressure holding stage, the temperature trajectory of the foaming furnace, the average deviation of product size measurement results (+0.3mm), and surface images. Based on the equipment controller timestamp, material entry time, and transfer duration, the material flow delay is calculated, and the data is divided into time windows corresponding to B01. Data completeness is assessed. The value is 0.98, with an estimated confidence level of 0.95, which is higher than the set threshold of 0.85. Based on the batch quality status, the foaming ratio is estimated to be 14.8, and the density uniformity is... The measured values ​​were 4.2%, dimensional shrinkage was 1.4%, and compression recovery was 92%. Substituting these measured values ​​into the comprehensive deviation formula in Example 1, the following calculation was obtained: The value is 0.6, which is lower than the deviation threshold of 1.0. The quality is qualified, and the system completes the warehousing and release.

[0083] In the abnormal triggering scenario, during the pressure holding process of batch B10, the system monitored a measured foaming temperature of 148℃, exceeding the allowable boundary of 150℃ to 160℃, triggering an abnormal closed-loop. Based on the foaming temperature trajectory deviation characteristics and the process influence relationship, the abnormal closed-loop module determined the candidate cause as insufficient heating power of the temperature control device, with a confidence level of 0.90. This is based on the process parameter vector in Example 1. exist The internal algorithm solves for candidate correction values, which must satisfy the variation range constraint given in Example 1. In this scenario, the foaming temperature is increased by 7°C, which meets the safety constraint of the maximum allowable variation range of 10°C for this parameter, and the prediction deviation can be reduced from 1.8 to 0.5. The selection function comprehensively considers the credibility of candidate causes, the predicted improvement amount, and the execution risk, and confirms that the correction scheme is to adjust the foaming temperature setpoint from 155°C to 162°C. This value is still within the range of... The upper limit is 165℃, and a correction order is issued.

[0084] The equipment controller confirms the correction command takes effect at 10.5 hours. The system reads the batch position within each piece of equipment at that moment. Batches that have completed the foaming process and entered the shaping stage are no longer affected by the correction and are not considered for verification. Batches currently in the autoclave impregnation stage but have not yet entered the foaming process will be affected by the corrected foaming temperature parameter, and the system identifies these batches as the first valid batch. The system confirms the correction as valid when three consecutive valid batches meet the quality target. In actual operation, if the overall deviation of three consecutive valid batches drops below 0.8, meeting the deviation threshold, the system marks the correction scheme as valid and updates the recommended process center value to a foaming temperature of 162℃. If the overall deviation of subsequent batches rises back to 1.2 or new safety constraint conflicts occur, the system cancels the automatic correction, restores the most recently verified process package (foaming temperature 155℃), and triggers a rearrangement.

[0085] In low-confidence scenarios, if pressure sensor data is continuously missing during the depressurization phase, the data integrity is compromised. The value is 0.62, and the estimated confidence level is 0.42, which is lower than the set threshold of 0.85. Therefore, the system prohibits the automatic generation of correction schemes that increase the range of pressure relief rate changes. Instead, it arranges offline density and bubble size detection. The detection results are written back before deciding whether to form a candidate correction scheme.

[0086] In equipment failure scenarios, if the high-pressure reactor E1 reports an invalid interlock before batch B15 starts, the execution coordination module refuses to issue a pressure increase target, the resource status modeling module sets E1 to unavailable, and the rolling scheduling module selects E2 with compatible capabilities for B15 or adjusts the delivery date. If the mold pressure resistance limit of E2 is 10MPa, which is lower than the original process window... The lower limit is 12MPa. The two do not intersect. The system outputs a conflict constraint and notifies the process personnel to replace the mold or re-evaluate the target quality vector.

[0087] The batch traceability module will include order A, batches B01 to B13, material batch M-202605, manufacturing resource status set, and production scheduling version. The data is stored together with the following information: versions of production process packages (1.0 to 1.2), batch quality status, parameter correction plan (i.e., adjusting the foaming temperature from 155℃ to 162℃), and verification results (i.e., three consecutive qualified batches from B11 to B13), forming a complete batch traceability record.

[0088] In this embodiment, the preset parameters, namely the commitment time domain of 4 hours, weights α of 5, β of 3, γ of 1, δ of 2, deviation threshold of 1.0, confidence threshold of 0.85, number of consecutive valid batches N of 3, safety margin of 10%, maximum correction range of 10℃, and interlocking boundaries of 18.5MPa and 170℃, are all exemplary configurations. In actual applications, they should be adjusted according to product standards, equipment capabilities, and enterprise safety procedures.

[0089] In summary, this invention converts order protection performance into batch quality targets through a demand parsing module, and combines this with a process constraint generation module to form a feasible process window constrained by materials, equipment, and safety conditions, reducing the disconnect between production scheduling and equipment executable parameters. Through committed time-domain rolling scheduling and process package version verification, it maintains the stability of recent tasks and prevents erroneous process packages from being executed by equipment during order changes, equipment failures, or material batch fluctuations. By aligning process data with inspection data through batch identification and material flow delays, it can pinpoint quality deviations to specific batches, equipment, processes, and process package versions. By generating correction schemes within the safe feasible domain and identifying the first valid batch for verification, it avoids using work-in-process that has not yet been corrected to verify the correction effect, improving the reliability of closed-loop adjustments. The batch traceability module associates and stores orders, materials, resources, tasks, process packages, quality status, and correction verification results, providing a consistent data foundation for product quality traceability, process review, and production plan optimization.

[0090] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent management system for the production of supercritical foamed protective products, characterized in that, include: Demand Analysis Module: Converts the product structure, material grade, protection level, target energy absorption performance, dimensional tolerance, required quantity, and delivery time limit in protective product orders into batch quality targets with order identifiers; Process constraint generation module: Based on batch quality objectives, material batches and product structure, determine the sequence of supercritical fluid impregnation, pressure increase and holding, pressure release and nucleation, temperature-controlled foaming and shaping processes, and generate batch feasible process windows; Resource status modeling module: Obtains the available time period, capacity parameters, occupancy status, maintenance status and safety interlock status of manufacturing resources, forming a set of manufacturing resource statuses with timestamps; Rolling production scheduling module: Under the constraints of process sequence, batch feasible process window and manufacturing resource status set, rolling production scheduling is performed on the batch to be produced, the equipment allocation, process start and end time and batch switching order of each batch are determined, and the scheduling results located within the preset commitment time domain are locked as executable tasks and scheduling versions; Execution coordination module: Generates production process packages based on executable tasks and sends them to the corresponding equipment controllers. The production process package includes at least batch identifier, equipment identifier, process sequence, target parameters, parameter allowable boundaries, and interlocking conditions. Quality status estimation module: According to batch identification and material flow delay, the pressure, temperature, supercritical fluid flow, product size, product quality and surface image collected during the production process are aligned with the corresponding production process package to generate batch quality status containing density data from multiple sampling locations, compression displacement data under specified load, unloading recovery data and cycle number data, and estimate foaming ratio, density uniformity, dimensional stability and compression rebound state based on batch quality status; Anomaly closed-loop module: When the estimated deviation exceeds the threshold, it generates a parameter correction plan or rearrangement plan based on the batch feasible process window, and verifies the correction result based on the batch quality status after the correction takes effect.

2. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The requirement analysis module establishes a mapping relationship between protective performance indicators and foaming quality indicators, converting at least two of the following into target apparent density, target foaming ratio, allowable density dispersion, allowable permanent compression set, and dimensional shrinkage rate: when there are multiple candidate quality targets in the mapping results, the manufacturability of the candidate quality targets is screened based on the material's processability range and equipment capabilities.

3. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The process constraint generation module reads the preset process route according to the product structure, determines the sequence of impregnation, pressure increase and holding, pressure release and nucleation, temperature-controlled foaming, and shaping processes; and determines the upper and lower boundaries of pressure, temperature, holding time, and pressure release rate based on the melting temperature zone, crystallization characteristics, initial density, and moisture content of the material batch. It then intersects the upper and lower boundaries using the rated pressure of the equipment, heating rate, flow rate limit, mold pressure resistance level, and safety margin to obtain the batch's feasible process window. When the intersection result is an empty set, it outputs an unschedulable production flag and the conflict constraints that caused the empty set.

4. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The manufacturing resources in the resource status modeling module include: autoclave, booster pump, temperature control device, mold, transfer device, and post-processing equipment.

5. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The rolling scheduling module aims to minimize the weighted sum of delivery delay costs, product changeover costs, equipment idle energy consumption, and quality uncertainty costs. It also imposes hard constraints such as equipment executing only one task at a time, subsequent processes only starting after the preceding process is completed, process parameters for each selected process being within the batch feasible process window, and safety interlocks being in effect. In the event of equipment failure, order changes, or quality deviations, the module only re-solves the problems for tasks outside the preset commitment time domain and tasks affected by the event.

6. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The execution coordination module assigns a version number and a verification value to each production process package. The equipment controller allows the task to start after verifying that the batch identifier, equipment identifier, version number, and parameter allowable boundaries are consistent. When the measured pressure, temperature, or pressure relief rate exceeds the parameter allowable boundaries, or the safety interlock status fails, the equipment controller refuses to continue executing the corresponding parameters and sends an abnormal event with the equipment timestamp to the abnormal closed-loop module.

7. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The quality status estimation module calculates the material flow delay based on the equipment controller timestamp, material entry time, pressure holding duration, transfer duration, and detection time, and divides the continuous sampling data into time windows corresponding to the batch. For missing data, output the data completeness and estimated confidence level. When the estimated confidence level is lower than a set threshold, automatically generate parameter correction schemes that expand the range of parameter changes based on the estimation result is prohibited.

8. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The abnormal closed-loop module constructs deviation characteristics based on the supercritical fluid dissolution amount, temperature and pressure stability during the pressure holding stage, pressure change rate during the pressure release stage, foaming temperature trajectory, and shaping cooling rate. It determines at least one candidate cause according to the process influence relationship. For each candidate cause, it solves for the candidate correction amount that reduces the predicted quality deviation and limits the parameter change within the batch feasible process window. It then selects the parameter correction scheme based on the credibility of the candidate cause, the predicted improvement amount, and the execution risk.

9. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: The abnormal closed-loop module takes the moment when the parameter correction command is confirmed to be effective by the equipment controller as the starting point, and determines the valid batch based on the product position, remaining holding time and transfer relationship in the equipment. When a preset number of valid batches meet the batch quality target, the correction is confirmed to be effective. When the deviation of the valid batches is not reduced or a new safety constraint conflict occurs, the parameter correction scheme is revoked and a rearrangement is triggered.

10. The intelligent management system for the production of supercritical foamed protective products according to claim 1, characterized in that: It also includes a batch traceability module, which is used to associate and store order identifiers, batch identifiers, material batches, manufacturing resource status sets, production scheduling versions, production process packages, batch quality status, parameter correction schemes and verification results as batch traceability records.

Images