A carbon fiber product manufacturing and operation management system

Abstract

This invention relates to the field of production operation data flow control technology, and discloses a carbon fiber product manufacturing operation management system, including: an equipment spatiotemporal state matrix generation unit, used to convert equipment volume and energy efficiency data into a capacity state matrix; a queuing delay dynamic correction module, used to fine-tune the priority queue weight when the queuing time reaches a threshold; a discrete state data interpolation module, used to complete the life deduction based on historical temperature when data is interrupted; and a dynamic production scheduling arbitration unit, used to correct the queuing time and output scheduling instructions based on material exposure data, equipment state vectors and heat load matrices through compensation operators. This invention breaks the information gap between the decision-making layer and the underlying thermal boundary, effectively offsets the secondary waiting increment caused by process cycle stretching, reduces the risk of material deterioration, and ensures the deterministic convergence of the production scheduling plan.

Description

Technical Field

[0001] This invention relates to the field of production operation data flow control technology, and more specifically, to a carbon fiber product manufacturing operation management system. Background Technology

[0002] Currently, in workshop operation management, a data processing scheme based on static time windows and fixed capacity matching is usually adopted. The system reads material ledger records and physical capacity parameters of thermoforming equipment to complete batch material allocation and distribution. With the upgrading of manufacturing precision, resin matrix materials such as carbon fiber prepreg exhibit time-varying degradation behavior under ambient temperature storage conditions. The remaining processing life of these materials undergoes physical drift with temperature fluctuations, forming a temporal trajectory constraint. The mainstream data processing architecture separates management business layer data from material property degradation, relying only on static planned time points to perform life control. When the dynamic energy efficiency of downstream molding equipment fluctuates suddenly, and the cumulative exposure time of multiple batches of materials shows random differences, the data matching flow cannot capture the spatiotemporal coupling relationship between material degradation characteristics and equipment transient thermal inertia evolution, resulting in the system data being unable to truly represent the dynamic thermal delay caused by material loading.

[0003] While the load-bearing mechanism can be improved at the physical level by optimizing hardware structures such as roller morphology, simply improving the equipment structure cannot automatically eliminate data lag. Control methods also have shortcomings when facing complex hardware and software interactions. For example, Chinese invention patent CN109598416B discloses a dynamic scheduling system and method for composite material workshops, attempting to utilize IoT-sensing data to process discrete disturbances such as equipment failures. However, the underlying premise of this technology is entirely based on an idealized scenario where process time and the characteristics of the loaded materials are independent. In the scenario of carbon fiber hot-press curing, the stacking and merging of multiple batches of materials drastically changes the flow field inside the tank and introduces physical thermal damping, causing nonlinear stretching of the process cycle due to the lack of material activity. The spatiotemporal coupling perception of trajectory and equipment thermal inertia boundary is a control method that, when faced with large-scale complex tasks, not only fails to offset the secondary waiting increments caused by extended process cycles, but is also prone to control flow divergence due to frequent calculations of high-dimensional tensors, leading to system blockage. To solve the material scrapping risks caused by such state mismatches, the usual improvement approach tends to shorten the global data polling clock cycle or add discrete independent buffer queues to eliminate information feedback lag. However, when faced with large-scale high-frequency order insertion interference or transient thermal distortion of equipment, this linear topology causes an exponential expansion of the high-dimensional tensor solution space of the scheduling state, resulting in data flow blockage within the system and significant algorithm delay. Under the condition of diverging data flow in multi-loop collaborative calculation, the equipment falls into idling deadlock.

[0004] Therefore, the technical problem to be solved by this invention is to combine the time-varying processing life loss trajectory of materials with the thermal damping degradation loss of molding equipment to achieve deterministic convergence scheduling between multiple curing molding equipment and dynamic material sequences while reducing the high-dimensional computational load. Summary of the Invention

[0005] This invention provides a carbon fiber product manufacturing operation management system, the system comprising: The equipment spatiotemporal state matrix generation unit is used to acquire the remaining volume data and energy efficiency data of the curing molding equipment, convert the remaining volume data and energy efficiency data into boundary scalars, and linearly assemble the boundary scalars through the global cluster-level bus to generate a resource capacity occupancy state matrix. The queuing delay dynamic correction module is used to monitor the virtual queuing time of each material batch in the waiting queue, and calculate the priority step coefficient when the virtual queuing time reaches the safety critical threshold, and fine-tune the input priority queue to update the dynamic weight. The discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, update material life deduction data according to the acceleration penalty coefficient, and generate step operator constraints composed of internal state matrix and state feature terms. The dynamic production scheduling arbitration unit is used to construct an input priority queue based on the cumulative heat exposure time series data of each material batch, and combine it with the energy efficiency configuration state vector of the solidification equipment. It uses the heat load feature matrix to be loaded to reorganize the equipment available volume remaining items in the resource capacity occupancy state matrix, calculates the temperature rise rate loss item caused by the candidate material combination through the dynamic compensation operator, corrects the virtual queuing time and outputs the production scheduling instruction.

[0006] Preferably, the device spatiotemporal state matrix generation unit includes local device-level nodes and a global cluster-level bus; the local device-level nodes are used to calculate the remaining available volume and energy efficiency data of the curing and molding device, and convert the remaining available volume and energy efficiency data into boundary scalars and send them to the global cluster-level bus; the global cluster-level bus is used to receive the boundary scalars and generate a resource capacity duty cycle state matrix through linear assembly.

[0007] Preferably, the queuing delay dynamic correction module is used to monitor the virtual queuing time of each batch of materials in the waiting queue in real time, calculate the priority step coefficient when the safety critical threshold is met, and fine-tune the input priority queue through discrete event drive to update the dynamic weight, and convert the delay deviation caused by the temperature rise rate loss term into the step operator constraint between the internal state matrix and the state feature term.

[0008] Preferably, the discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, and update the material life deduction data according to the acceleration penalty coefficient to smooth and compensate the cumulative heat exposure time series data of each material batch, and maintain the continuous evolution expression of dynamic weights in the input priority queue.

[0009] Preferably, the dynamic production scheduling arbitration unit is used to reorganize the available volume remaining items of the equipment in the resource capacity duty cycle matrix at the parameter level using the heat load feature matrix to be loaded; the dynamic production scheduling arbitration unit includes a dynamic compensation operator for the temperature rise of the hot fluid in the curing equipment, the dynamic compensation operator is used to directly read the intrinsic rated heating power of the equipment stored in the resource capacity duty cycle matrix, and calculate the temperature rise rate loss item caused by the candidate material combination entering the equipment.

[0010] Preferably, the calculation rule for the temperature rise rate loss term is reflected as a step-wise correction of energy efficiency degradation caused by the total mass of the material. The dynamic production scheduling arbitration unit linearly assembles and sums the mass of each material batch in the heat load feature matrix to be loaded to obtain the total mass of the material. When the total mass of the material exceeds the loading heat capacity threshold, the dynamic energy efficiency degradation rule is triggered to reduce the virtual temperature rise efficiency benchmark of the curing molding equipment.

[0011] Preferably, the dynamic production scheduling arbitration unit feeds back the curing process cycle extension time caused by the temperature rise rate loss to the waiting queue, automatically rewriting the virtual queuing time of each non-failed material batch in the waiting state, so that it is superimposed with the secondary waiting increment caused by the extension of the thermal inertia of the current batch.

[0012] Preferably, the dynamic scheduling arbitration unit re-introduces the corrected virtual queuing time into the calculation flow of the dynamic urgency index, using the formula: Calculate the dynamic severity index for each material batch; among which, This represents the dynamic urgency index for each batch of materials. This is the cumulative heat exposure time constant extracted from the physical properties of each material batch. This refers to the current cumulative heat exposure time read from the cumulative heat exposure time series data of each material batch. This is the corrected virtual queuing time.

[0013] Preferably, when the modified dynamic urgency index of each material batch crosses the risk threshold... At that time, the dynamic production scheduling arbitration unit intercepts the currently loaded combination and triggers a hard truncation.

[0014] Preferably, the energy efficiency configuration state vector of the curing equipment includes the rated thermal power of the equipment, the thermal flow field uniformity coefficient, and the upper limit of the tank volume; the dynamic production scheduling arbitration unit determines the capacity matching result of the material life degradation state for a single curing processing equipment based on the matching degree between the energy efficiency configuration state vector of the curing equipment and the cumulative heat exposure time series data of each batch of materials.

[0015] The present invention has at least the following beneficial effects: 1. In the manufacturing and operation management of carbon fiber products, the material state time-series trajectory tracking unit collects material life consumption parameters, the equipment spatiotemporal state matrix generation unit obtains the intrinsic rated heating power of the curing and molding equipment, and the dynamic production scheduling arbitration unit extracts the layup monomer mass and flow resistance section feature vector of each batch of materials to construct the heat load feature matrix to be loaded. The cooperating curing equipment hot fluid temperature rise dynamic compensation operator calculates the virtual temperature rise rate loss term caused by the material combination entering the equipment, and converts the extension of the curing process cycle into virtual queuing time compensation for subsequent materials to be processed, so that the urgency calculation adaptively captures the equipment temperature rise lag caused by the loading material, and eliminates the disconnect between the data management flow and the underlying physical and thermal boundary.

[0016] 2. When the dynamic production scheduling arbitration unit incorporates the corrected virtual queuing time into the dynamic urgency index calculation flow, if it identifies that the corrected urgency index of subsequent material batches crosses the risk threshold, it intercepts the current loading combination and activates a hard cutoff rule. The subsequent materials facing the risk of exceeding the time limit are then reverse-woven into the greedy matching scheme of the current tank batch until the remaining available volume of the equipment reaches the saturation boundary. This combination processing method after multi-dimensional thermal inertia correction effectively eliminates the secondary waiting increment caused by the nonlinear stretching of the process cycle, reduces the risk of cumulative heat exposure and deterioration of subsequent materials during temporary storage in the workshop, and ensures the certainty of scheme convergence in continuous production.

[0017] 3. The queuing delay dynamic correction module monitors the virtual queuing time of each material batch in the waiting queue in real time. When the safety critical condition is met, it calculates the priority step coefficient and performs discrete event-driven fine-tuning on the input priority queue to reconstruct the dynamic weight. In conjunction with the discrete state interpolation module, it automatically reads the historical ambient temperature benchmark when the monitoring data stream is interrupted and completes the lifespan deduction update according to the acceleration penalty coefficient. This avoids the calculation chain from diverging due to continuous large-scale optimization. It translates the complex temperature rise distortion into a step-by-step operator constraint between the system's internal state matrix and tensor elements, achieving stable convergence of production scheduling decisions under real physical conditions with lower data processing and transformation overhead. Attached Figure Description

[0018] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings, in which several embodiments of the invention are illustrated by way of example and not limitation, wherein: Figure 1 This is a schematic diagram of the module composition structure of the carbon fiber product manufacturing and operation management system of the present invention; Figure 2 This is an output diagram of the material combination scheme of the carbon fiber product manufacturing and operation management system of the present invention. Detailed Implementation

[0019] The principles and spirit of the present invention will now be described with reference to several exemplary embodiments in conjunction with the accompanying drawings. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0020] A carbon fiber product manufacturing operation management system, the system comprising: The equipment spatiotemporal state matrix generation unit is used to acquire the remaining volume data and energy efficiency data of the curing molding equipment, convert the remaining volume data and energy efficiency data into boundary scalars, and linearly assemble the boundary scalars through the global cluster-level bus to generate a resource capacity occupancy state matrix. The queuing delay dynamic correction module is used to monitor the virtual queuing time of each material batch in the waiting queue, and calculate the priority step coefficient when the virtual queuing time reaches the safety critical threshold, and fine-tune the input priority queue to update the dynamic weight. The discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, update material life deduction data according to the acceleration penalty coefficient, and generate step operator constraints composed of internal state matrix and state feature terms. The dynamic production scheduling arbitration unit is used to construct an input priority queue based on the cumulative heat exposure time series data of each material batch, and combine it with the energy efficiency configuration state vector of the solidification equipment. It uses the heat load feature matrix to be loaded to reorganize the equipment available volume remaining items in the resource capacity occupancy state matrix, calculates the temperature rise rate loss item caused by the candidate material combination through the dynamic compensation operator, corrects the virtual queuing time and outputs the production scheduling instruction.

[0021] Preferably, the device spatiotemporal state matrix generation unit includes local device-level nodes and a global cluster-level bus; the local device-level nodes are used to calculate the remaining available volume and energy efficiency data of the curing and molding device, and convert the remaining available volume and energy efficiency data into boundary scalars and send them to the global cluster-level bus; the global cluster-level bus is used to receive the boundary scalars and generate a resource capacity duty cycle state matrix through linear assembly.

[0022] Preferably, the queuing delay dynamic correction module is used to monitor the virtual queuing time of each batch of materials in the waiting queue in real time, calculate the priority step coefficient when the safety critical threshold is met, and fine-tune the input priority queue through discrete event drive to update the dynamic weight, and convert the delay deviation caused by the temperature rise rate loss term into the step operator constraint between the internal state matrix and the state feature term.

[0023] Preferably, the discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, and update the material life deduction data according to the acceleration penalty coefficient to smooth and compensate the cumulative heat exposure time series data of each material batch, and maintain the continuous evolution expression of dynamic weights in the input priority queue.

[0024] Preferably, the dynamic production scheduling arbitration unit is used to reorganize the available volume remaining items of the equipment in the resource capacity duty cycle matrix at the parameter level using the heat load feature matrix to be loaded; the dynamic production scheduling arbitration unit includes a dynamic compensation operator for the temperature rise of the hot fluid in the curing equipment, the dynamic compensation operator is used to directly read the intrinsic rated heating power of the equipment stored in the resource capacity duty cycle matrix, and calculate the temperature rise rate loss item caused by the candidate material combination entering the equipment.

[0025] Preferably, the calculation rule for the temperature rise rate loss term is reflected as a step-wise correction of energy efficiency degradation caused by the total mass of the material. The dynamic production scheduling arbitration unit linearly assembles and sums the mass of each material batch in the heat load feature matrix to be loaded to obtain the total mass of the material. When the total mass of the material exceeds the loading heat capacity threshold, the dynamic energy efficiency degradation rule is triggered to reduce the virtual temperature rise efficiency benchmark of the curing molding equipment.

[0026] Preferably, the dynamic production scheduling arbitration unit feeds back the curing process cycle extension time caused by the temperature rise rate loss to the waiting queue, automatically rewriting the virtual queuing time of each non-failed material batch in the waiting state, so that it is superimposed with the secondary waiting increment caused by the extension of the thermal inertia of the current batch.

[0027] Preferably, the dynamic scheduling arbitration unit re-introduces the corrected virtual queuing time into the calculation flow of the dynamic urgency index, using the formula: Calculate the dynamic severity index for each material batch; among which, This represents the dynamic urgency index for each batch of materials. This is the cumulative heat exposure time constant extracted from the physical properties of each material batch. This refers to the current cumulative heat exposure time read from the cumulative heat exposure time series data of each material batch. This is the corrected virtual queuing time.

[0028] Preferably, when the modified dynamic urgency index of each material batch crosses the risk threshold... At that time, the dynamic production scheduling arbitration unit intercepts the currently loaded combination and triggers a hard truncation.

[0029] Preferably, the energy efficiency configuration state vector of the curing equipment includes the rated thermal power of the equipment, the thermal flow field uniformity coefficient, and the upper limit of the tank volume; the dynamic production scheduling arbitration unit determines the capacity matching result of the material life degradation state for a single curing processing equipment based on the matching degree between the energy efficiency configuration state vector of the curing equipment and the cumulative heat exposure time series data of each batch of materials.

[0030] Example 1: In a production flow scenario where multiple batches of carbon fiber prepreg are transported concurrently to the subsequent hot pressing and curing equipment, the current carbon fiber product manufacturing operation management system suffers from nonlinear resin curing evolution due to differences in the layup structure and geometry of each batch of carbon fiber prepreg during temporary storage in the workshop. When multiple batches of materials with different geometries and mass resistance are combined into the same autoclave, the overall composite heat capacity and density flow resistance of the materials change the flow field and temperature field response characteristics inside the equipment. The physical differences in the total volume and total mass of the loaded materials directly cause dynamic thermal inertia lag in the curing equipment, resulting in nonlinear stretching of the actual curing process cycle. This leads to unexpected cumulative heat exposure in the workshop for subsequent materials in the waiting queue, causing implicit overdue degradation or production deadlock.

[0031] The system's equipment spatiotemporal state matrix generation unit acquires the remaining volume and energy efficiency data of the curing and molding equipment, converts this data into boundary scalars, and linearly assembles these boundary scalars via a global cluster-level bus to generate a resource capacity duty cycle state matrix. Simultaneously, the system's material state time-series trajectory tracking unit connects to the cryogenic storage and workshop flow monitoring interfaces, collecting the cumulative heat exposure time data stream of specific carbon fiber product batches in the flow environment. It converts physical environment parameters into a structured material lifetime consumption state tensor to track the activity degradation state of each batch of material in the physical environment in real time. Local equipment-level nodes calculate the internal remaining volume and energy efficiency data of a single curing and molding equipment, convert it into boundary scalars, and send it to the global cluster-level bus. The global cluster-level bus then constructs a matrix representing multiple equipment through one-dimensional linear assembly of the received boundary scalars. The resource capacity duty cycle matrix in the storage state reduces communication congestion and high-dimensional tensor computation load during multi-device joint scheduling. Specifically, the global cluster-level bus adopts a time-division reusable control protocol based on round-robin to send data read requests to each local device-level node in sequence. Each node responds to the request and packages its own boundary scalar into a fixed-length 16-bit binary data frame, which is transmitted to the register of the main control processor through the field industrial Ethernet bus. The main control processor fills the received boundary scalars into a pre-allocated one-dimensional continuous memory array in the order of the device physical numbers. Each array element corresponds to the real-time available capacity and energy efficiency benchmark of a single solidified molding device. Thus, the hardware feature matrix reflecting the capacity duty cycle of all equipment in the workshop is directly reconstructed in a one-dimensional linear assembly mapping manner, completing the transformation of the underlying physical state into a discrete data structure.

[0032] The system's dynamic production scheduling and arbitration unit is connected to both the material state time-series trajectory tracking unit and the equipment spatiotemporal state matrix generation unit. It receives the material lifetime consumption state tensor and the resource capacity duty cycle state matrix, extracts the lifetime consumption parameters of the current candidate material batches, and simultaneously extracts the intrinsic structure parameters corresponding to each batch of materials from the transfer interface. Finally, it constructs a feature matrix of the heat load to be loaded based on the candidate combination sequence arranged in descending order of preliminary urgency. Then utilize the feature matrix of the heat load to be loaded. Reorganize the device available volume surplus items in the resource capacity duty cycle matrix ,in, The characteristic matrix of the heat load to be loaded. For the remaining available volume of the equipment, intrinsic structural parameters include the mass of the ply unit of the material. With the characteristic vector of the flow resistance section ,in This is an index for candidate material batches. For the first The quality of the layup monomers in each candidate material batch. For the first The flow resistance cross-sectional characteristic vector of each candidate material batch, and the lifetime consumption parameter is represented by the current cumulative heat exposure time. ,in For the first The current cumulative heat exposure time for each candidate material batch.

[0033] The dynamic scheduling and arbitration unit is equipped with a dynamic compensation operator for the temperature rise of the hot fluid in the curing equipment. This operator directly reads the intrinsic rated heating power of the equipment stored in the resource capacity duty cycle matrix and the characteristic matrix of the heat load to be loaded. The mass of each batch of material layup monomers One-dimensional linear assembly summation is performed to obtain the total mass of the materials. When the total mass of the materials exceeds a preset loading heat capacity threshold, a dynamic energy efficiency degradation rule is triggered, directly reducing the virtual temperature rise efficiency benchmark of the curing equipment. The dynamic temperature rise rate loss term caused by the current candidate material combination entering the equipment is calculated. And will be affected by the dynamic temperature rise rate loss term. The resulting curing process cycle stretching time is fed forward to the pre-waiting queue, automatically rewriting the virtual queuing time for each batch of non-failed materials in the waiting state. This causes the secondary waiting increment caused by the extension of the current tank's thermal inertia to be superimposed, where This is the loss term based on the dynamic temperature rise rate. This is the virtual queuing time. The characteristic matrix of the heat load to be loaded. For the first The mass of the layup monomers in each material batch, subscript Index for candidate material batches.

[0034] The dynamic scheduling arbitration unit will adjust the virtual queuing time. Substitute it back into the calculation flow of the dynamic criticality index, and use the formula... Calculate the dynamic urgency index for each material batch. ,in, For the first Dynamic urgency index for each batch of materials. This refers to the critical time for physical failure of the material. This represents the current cumulative heat exposure time. This is the virtual queuing time, indicated by the subscript. For material batch indexing, when the dynamic urgency index of each material batch... Crossing the preset risk threshold At this time, the dynamic production scheduling arbitration unit intercepts the current loading combination and triggers a hard cutoff rule, forcibly weaving subsequent materials facing the risk of overdue into the greedy matching scheme of the current tank batch, until the remaining available volume of the equipment is reached. Upon reaching the saturation boundary, a unique tank material combination scheme, after multi-dimensional thermal inertia correction, is output as the operation control command, where the risk threshold is... It is a dimensionless pure number, with its value range set between 1.5 and 2.0, representing the remaining available volume of the equipment. The unit is cubic meters. To establish the thermodynamic correspondence between macroscopic time control parameters and microscopic resin matrix property degradation, this invention uses preliminary thermodynamic experiments to measure the average crosslinking and curing activation energy of the current system resin in the room temperature range. The material physical failure critical time introduced in the dynamic urgency index is not a static fixed constant, but a dynamic variable obtained by combining the temperature fluctuation of the circulation environment collected in real time by the distributed sensor array in the workshop with the equivalent rate integral correction of the Arrhenius empirical formula. When the circulation environment temperature increases, the system automatically decreases the value of the physical failure critical time, and vice versa. This reduces the nonlinear cumulative loss of the resin crosslinking reaction rate at the microscopic level to a time scale constraint at the macroscopic scheduling layer, ensuring that the urgency judgment in the time dimension is highly consistent with the actual material property degradation trajectory in physical essence. The queuing delay dynamic correction module is based on discrete event dynamics. The system theory controls the time delay deviation and adjusts the state transition trajectory of the queuing sequence when nonlinear stretching occurs in the downstream node process cycle. The controller input receives the current cumulative heat exposure time and initial priority weight of each material batch. The system clock synchronization accuracy is maintained within 1 millisecond, providing a deterministic enabling environment. The queuing delay dynamic correction module monitors the virtual queuing time in real time by reading the timer register values ​​corresponding to each material batch in the workshop fieldbus. When the virtual queuing time crosses the safety critical threshold, the controller calls the step correction program, using the ratio of the virtual queuing time to the material physical failure critical time as the input variable, and calculates the step-by-step priority step coefficient. The controller directly accumulates this priority step coefficient to the dynamic weight item of the corresponding material batch in the input priority queue, changing the discrete event triggering order of the input priority queue and reducing the response delay of production scheduling instructions when there are fluctuations in physical and thermal boundaries.

[0035] When the aforementioned dynamic production scheduling arbitration unit outputs operation control commands, the system's internal dynamic queuing delay correction module monitors in real time the virtual queuing time of material batches in the waiting queue. When virtual queuing time is detected The judgment conditions are met. At that time, the queuing delay dynamic correction module calculates the priority step coefficient. The specific calculation formula is as follows: ,in, This is the priority step coefficient. This is the virtual queuing time. This refers to the critical time for physical failure of the material. The current cumulative heat exposure time is indicated by the subscript. Index the material batches and use the calculated priority step factor. The system overlays and rewrites the input priority queue of the dynamic production scheduling arbitration unit, fine-tuning the core scheduling sequence through discrete event-driven methods to avoid computational chain divergence. Simultaneously, the system's internal discrete state data interpolation module automatically reads the historical ambient temperature baseline of the corresponding material batch when the monitoring data stream is interrupted, and updates the lifetime deduction data according to a preset acceleration penalty coefficient. This outputs a step operator constraint composed of an internal state matrix and state feature terms, maintaining the continuous evolution of the dynamic weights in the input priority queue until the cumulative heat exposure time data stream is restored. The discrete state data interpolation module is based on thermodynamic steady-state heat transfer. The model updates the historical ambient temperature baseline by using the slow drift thermal equilibrium characteristics of ambient temperature under long-term operating conditions to correct the baseline data. The module includes an internal time-series cyclic buffer with a capacity of 30 days to store the historical time-series discrete sequences of ambient temperature collected by the temperature sensor array. The sampling frequency of the data acquisition interface is set to 0.1 Hz. The controller accumulates the continuous running time of the system in real time. When the continuous running time reaches an integer multiple of 24 hours, the reconstruction program is triggered. The reconstruction program retrieves all ambient temperature data within the current sliding time window, calculates the variance and mean drift of all ambient temperature data, and updates the historical ambient temperature baseline.

[0036] The weights of the long-term historical data are reduced using a time decay factor of 0.15, so that the updated historical ambient temperature benchmark smoothly tracks the physical drift of the mean of the actual ambient temperature in the workshop. In the event of a sudden disconnection of the workshop network bus or a failure of the telemetry signal, a replacement reference background value is directly provided. In this process, the internal state matrix is ​​defined as a two-dimensional discrete state data set composed of the remaining processing life of each candidate material batch, the instantaneous heat exposure accumulation time, and the workshop benchmark temperature at the last moment before the interruption. Its dimension is the total number of material batches multiplied by 3. The state feature term is a single-valued scalar obtained by multiplying the standard deviation of the historical ambient temperature fluctuation by the acceleration penalty coefficient. The step operator constraint is specifically manifested as a boundary difference inequality judgment condition, which limits the step increment of the dynamic weight in the priority queue in the next scheduling cycle to not exceed the product of the state feature term and the corresponding feature value in the internal state matrix. Through the mandatory constraint of this mathematical boundary, the temperature uncertainty distortion during the data interruption is transformed into a deterministic step-by-step constraint, controlling the gradual evolution of the scheduling sequence.

[0037] Because the dynamic scheduling arbitration unit integrates the processing mechanism based on material temperature and heat exposure status with the energy efficiency configuration state vector of the curing equipment, it eliminates the overdue prediction residuals of materials in the workshop flow. This allows the processing scheduling instructions for multiple batches of materials to be fully adapted to the actual process cycle stretching characteristics of the curing equipment under specific loading combinations. Even when faced with large-scale order insertion disturbances, it can still maintain millisecond-level decision response latency. The dynamic scheduling arbitration unit adaptively intercepts and prioritizes the digestion of subsequent material batches that are on the edge of physical mutation before active consumption, keeping the overdue scrap rate of secondary materials in the continuous production environment below 0.5%, and keeping the system in a stable and convergent operating state.

[0038] Example 2: In an experiment verifying the long-term operational stability and production scheduling convergence speed of a carbon fiber product manufacturing operation management system, the experimental platform was constructed in a digital hot-pressing molding and curing workshop data network environment simulating multi-configuration, multi-batch concurrent flow. Environmental parameters during the actual flow process were acquired through a temperature distributed sensor array with 0.1℃ resolution and ±0.5℃ static measurement accuracy, and a material flow data acquisition bus with a 10ms response delay. The determination of the material life consumption status tracking cycle was controlled by two technical factors: the crosslinking and curing kinetic rate of the polymer resin and the bandwidth load of the industrial control network in the workshop. This tracking cycle was set to balance the real-time capture of material activity degradation characteristics by high-frequency data updates with the communication blockage caused by high-density concurrent telemetry data streams on the industrial-grade fieldbus. Its operation mode was characterized by dynamically adjusting the tracking cycle based on the wavelet transform energy spectral density of the workshop ambient temperature envelope bandwidth: when the ambient temperature envelope bandwidth was at a certain value, the tracking cycle was adjusted accordingly. When the ambient temperature is between 22℃ and 28℃ and the variance of fluctuation is less than 0.5, the material activity evolution rate is relatively in a steady-state period. The tracking period is set to the upper limit of the range, i.e., 30 seconds. The specific operation steps are as follows: the system collects a discrete sequence of workshop ambient temperature for 2 consecutive hours, and performs a 3-level multi-scale wavelet decomposition on the temperature data stream using the fourth-order Daubey wavelet basis function to separate the high-frequency detail coefficients and low-frequency approximation coefficients. The sum of squares of the third-level high-frequency detail coefficients is calculated and defined as the energy spectral density index characterizing the transient and violent fluctuations in ambient temperature. When the energy spectral density index crosses the preset fluctuation intensity threshold of 0.85, it indicates that there is a sudden thermal disturbance in the external environment. The system automatically reduces the material life consumption status tracking period from 30 seconds to 5 seconds through a preset inverse proportional mapping logic, thereby improving the sensitivity to capture temperature disturbances and mitigating the residual influence of high-frequency noise on activity prediction. In this experimental environment, an amplitude of [missing value] is actively superimposed. Random Gaussian thermocouple thermal noise perturbation at ℃, and introduced Random network data packet loss failures are used to simulate data flow transmission conditions in complex industrial settings.

[0039] In a comparative experiment on high-intensity order insertion scheduling across multiple tanks, the flow data streams of 10 concurrent material batches were simultaneously input to both the control group using a traditional static time window transaction processing architecture and the test group using the method of this invention. The initial cumulative heat exposure time of each candidate material batch was also included. The scores were randomly distributed between 35 and 85. For the first Initial cumulative heat exposure time for each candidate material batch, in minutes, subscript The candidate material batch index; due to the aforementioned 5.2% data loss fault, the control group lost the integral trajectory of the temperature rise segment of the 4th and 7th batches of materials in the temporary storage area for 18.5 minutes due to the lack of a missing state completion mechanism, resulting in the omission of residual values ​​in the lifetime consumption parameters stored in its system. The test group used the discrete state data interpolation module to retrieve the historical environmental temperature benchmark of the corresponding material batch, and reconstructed a continuous lifetime deduction sequence through the acceleration penalty coefficient operator based on least squares fitting. This smoothly corrected the original lifetime data with power frequency interference and random jumps into a measurement fingerprint data stream with physical continuity.

[0040] When the autoclave equipment is used to package the first batch of materials, the dynamic scheduling and arbitration unit of the test group extracts the intrinsic structure parameters of each batch of materials to construct the characteristic matrix of the heat load to be loaded. Intrinsic structural parameters include the mass of the ply unit. With the characteristic vector of the flow resistance section ,in For the first The quality of the layup monomers in each candidate material batch. For the first The flow resistance cross-sectional characteristic vector of each candidate material batch, with subscript... The candidate material batch index is used, and the actual available volume remaining item of the autoclave is extracted by the equipment spatiotemporal state matrix generation unit. It is 0.42 cubic meters, of which For the remaining available capacity of the equipment, since the total mass of the material in the first batch reached 820.6 kg, exceeding the preset loading heat capacity threshold of 500.0 kg, the dynamic compensation operator inside the dynamic production scheduling arbitration unit was activated, and the loss item caused by the increase in composite heat capacity was calculated. The temperature increase was 0.38°C per minute, which extended the curing process cycle of this batch from the standard 120.0 minutes to 148.3 minutes. The system automatically fed forward this 28.3-minute process delay to the pre-waiting queue, automatically rewriting the virtual queuing time of each batch of non-expired materials in the waiting state. This allows the system to superimpose the secondary waiting increment caused by the extended thermal inertia of the current tank, and automatically obtain the critical point of the material's overdue degradation and deterioration.

[0041] To verify the risk threshold To ensure the reasonableness of the numerical range, the experimental design included an out-of-range control group stress test that deviated from the protection range. This was done when the risk threshold was considered. When adjusted to below the protection lower limit of 1.30, the hard cutoff rule is triggered too sensitively, causing subsequent materials in a safe state to be frequently reverse-woven into the current loading combination. This specific numerical range is determined based on stability boundary pressure test results under high-intensity continuous production conditions in the workshop. Specifically, when the value is below 1.5, the boundary for triggering hard cutoff is too lenient, causing a large number of materials that are not yet close to the danger zone to be forcibly merged prematurely, disrupting the original optimal volume allocation scheme. Conversely, when the value is above 2.0, the system's perception of the time delay caused by process stretching is too sluggish, resulting in a delayed response to interception intervention. This causes the accumulated exposure time of materials at the bottom of the waiting queue to directly cross the critical collapse point. Only by constricting this control boundary within the working window of 1.5 to 2.0 can the secondary waiting risk be accurately eliminated while ensuring the saturated utilization rate of the equipment volume, within the remaining available equipment volume. While overload occurs, the system frequently increases the conflict correction count of the input priority queue in each event cycle, causing the computation chain to oscillate and diverge. This causes the system's scheduling instruction output decision response latency to surge from the normal millisecond level to 1420.5 milliseconds, reducing the convergence reliability of multi-tank joint scheduling.

[0042] When the risk threshold is When adjusted to 2.20, which is higher than the protection limit, the internal data flow analysis of the system shows that due to the delayed timing of the interception intervention, the secondary waiting delay caused by the process cycle stretching could not be eliminated by feedforward. The 9th and 10th batches of carbon fiber prepreg in the subsequent temporary storage queue failed to receive priority scheduling arbitration before reaching the active physical mutation point. Their actual cumulative heat exposure time exceeded the failure threshold, leading to irreversible localized premature gelation and deterioration of the matrix resin. The secondary material overdue scrap rate rose to 5.4%. Maintaining a stable operating window of 1.5 to 2.0 as defined in the invention, and setting it to 1.65 in this typical operating condition, the queuing delay dynamic correction module and the dynamic scheduling arbitration unit respond collaboratively: through the formula Calculate the dynamic urgency index for each material batch. ,in, As a risk threshold, For the first Dynamic urgency index for each batch of materials. This refers to the critical time for physical failure of the material. This represents the current cumulative heat exposure time. This is the virtual queuing time, indicated by the subscript. For material batch indexing, the priority step coefficient is calculated using a formula. It is 0.0024 per minute, of which Using priority step coefficients, accurate discrete event-driven fine-tuning is implemented on the input priority queue. With a decision latency maintained at a high-speed convergence state of 4.2 milliseconds, the problem of cumulative heat exposure exceeding the expiration period of subsequent materials is solved. The over-expiration scrap rate of secondary materials produced in multiple batches of continuous curing is stably controlled at 0.15%, objectively demonstrating the engineering practical value of this system architecture in resisting interference from underlying thermal physical residuals and ensuring deterministic convergence of high-dimensional data management decisions.

[0043] Example 3: This example combines Figures 1 to 2 A description of a carbon fiber product manufacturing and operation management system, such as... Figure 1 As shown, the equipment spatiotemporal state matrix generation unit acquires the remaining volume data and energy efficiency data and generates a resource capacity duty cycle state matrix. The generated resource capacity duty cycle state matrix is ​​provided as input to the dynamic production scheduling arbitration unit. The discrete state data interpolation module reads the historical ambient temperature benchmark and generates step operator constraints. These step operator constraints are input as constraint conditions to the dynamic production scheduling arbitration unit. The dynamic production scheduling arbitration unit constructs a priority queue based on the cumulative heat exposure time series data, reorganizes the equipment volume using the heat load feature matrix to be loaded, and calculates the temperature rise rate loss term to correct the queuing time. At the same time, the dynamic production scheduling arbitration unit transmits relevant information to the queuing delay dynamic correction module by feedback correction of the virtual queuing time. The queuing delay dynamic correction module monitors the virtual queuing time and fine-tunes the input priority queue. The fine-tuned input priority queue is fed back to the dynamic production scheduling arbitration unit as the input priority queue. Finally, the dynamic production scheduling arbitration unit outputs a production scheduling instruction to offset secondary waiting increments and ensure deterministic convergence.

[0044] like Figure 2 As shown, the feature vectors of the layup unit mass and flow resistance section of each batch of materials are extracted and a feature matrix of the heat load to be loaded is constructed. Then, the remaining available volume of the equipment is reorganized to determine whether the total mass of the material exceeds the loading heat capacity threshold. If it does, the dynamic energy efficiency degradation rule is triggered and the dynamic temperature rise rate loss term is calculated to enter the step of rewriting the virtual queuing time superimposed with the secondary waiting increment. If it does not, the step of rewriting the virtual queuing time superimposed with the secondary waiting increment is directly entered. Then, the dynamic urgency index of each batch of materials is calculated, and it is further determined whether the dynamic urgency index exceeds the risk threshold. If it does, the hard truncation rule is triggered to force the greedy matching scheme of the current tank and output the unique tank material combination scheme. If it does not, the unique tank material combination scheme is directly output.

[0045] Example 4: When five curing and molding machines coexist in a clustered deployment architecture and face the situation of material influx into the temporary storage area, the system's equipment spatiotemporal state matrix generation unit distributes task signals according to the available volume of each curing and molding machine from largest to smallest. The dynamic production scheduling arbitration unit reorganizes the resource capacity duty cycle state matrix to initiate the discrete step screening rules and retrieves the layup monomer quality of each candidate material batch in the pre-waiting queue. Compared with the current cumulative heat exposure time Calculate the initial severity level of each material batch, then arrange the initial severity levels in descending order of value and construct a dimension-wise array. The priority sequence to be loaded, where For the first The quality of the layup monomers in each candidate material batch. For the first Current cumulative heat exposure time for each candidate material batch The total number of candidate material batches, indicated by subscript. As the index for candidate material batches, the first material batch in the priority sequence to be loaded is selected again as the seed node, and the subsequent candidate sequences are traversed to calculate the composite heat capacity and total flow resistance characteristics after merging the seed node with each subsequent material batch. Then, the loading heat capacity threshold of the calculated total material mass is less than or equal to 500.0 kg, and the flow resistance cross-sectional feature vector of each batch is... When the algebraic sum is less than the preset flow resistance upper limit, the corresponding subsequent material batches will be temporarily incorporated into the current tank batch material combination scheme, where For the first The flow resistance cross-sectional feature vector of each candidate material batch is finally added, along with the remaining available equipment volume due to the current subsequent material batch. Less than zero or the resulting dynamic temperature rise rate loss term This leads to a dynamic criticality index for subsequent unprocessed materials. Crossing the risk threshold When this happens, the current subsequent material batch is removed and the matching is carried over to the next index node until the remaining available volume of the current curing equipment is reached. When the capacity utilization rate reaches its upper limit saturation boundary, a unique tank material combination scheme is output, in which... This refers to the remaining available volume of the equipment. This is the loss term based on the dynamic temperature rise rate. For the first Dynamic urgency index for each batch of materials. To mitigate the risk threshold and prevent system oscillations caused by repeated jumps in the temperature rise rate loss term due to frequent material rejection and recalculation, this invention introduces a unidirectional, non-decreasing dynamic temperature rise rate loss term caching and truncation mechanism in the matching logic. In the current single-iteration search process, after each rejection of unqualified subsequent material batches, the dynamic temperature rise rate loss term under the new combination does not immediately complete a complete reverse backtracking calculation. Instead, it maintains the currently accumulated maximum temperature rise rate loss value as the control benchmark for this round of matching until all nodes in the current index sequence have been traversed, or when the remaining available volume of the current autoclave reaches the saturation boundary. Only then does the control loop uniformly implement a single closed-loop convergence judgment, thereby forcibly breaking the infinite nested state of calculation and ensuring deterministic convergence of the production scheduling decision process within a finite number of discrete steps.

[0046] The system's discrete-state data interpolation module stores historical discrete time-series sequences of ambient temperatures collected by a distributed temperature sensor array over the past 30 days in its internal time-series cyclic buffer. When the system reaches a trigger point after 24 hours of continuous operation, the discrete-state data interpolation module retrieves all ambient temperature data within the sliding time window stored in the time-series cyclic buffer, calculates the variance and mean shift of the ambient temperature during that period to update the historical ambient temperature baseline, and superimposes a dimensionless time decay factor. The weighting of long-term historical data is reduced, and the calculation formula for the historical ambient temperature baseline is updated as follows: ,in, For the updated historical ambient temperature benchmark, The historical ambient temperature baseline before the update This represents the average temperature within the current sliding time window. The time decay factor is set to 0.15. The discrete state data interpolation module periodically reconstructs the historical ambient temperature benchmark. When the workshop bus is disconnected or the telemetry signal data stream is interrupted, the discrete state data interpolation module calls the current temperature and humidity field historical ambient temperature benchmark and updates the material life deduction data in combination with the acceleration penalty coefficient, so as to maintain the causal continuous expression of the dynamic urgency index in the input priority queue.

[0047] Example 5: When the system faces the access of a new curing molding device or the reconstruction of the workshop deployment environment, the energy efficiency data of the device's spatiotemporal state matrix generation unit will experience a baseline drift due to the difference in the thermal capacity of the physical hardware. To establish the control starting point, the system initiates a pre-state calibration procedure, controls the target curing molding device to heat up under no-load conditions, and uses a distributed temperature sensor array to capture the transient temperature rise curve. The processor calculates the intrinsic thermal inertia dissipation coefficient of the device based on the slope of the temperature rise curve in the linear region. ,in Let be the intrinsic thermal inertia dissipation coefficient, and use the intrinsic thermal inertia dissipation coefficient Update the energy efficiency configuration parameters in the resource capacity duty cycle state matrix to complete the initial state definition of the system under specific physical hardware boundaries.

[0048] When the system faces the situation of batch replacement of new carbon fiber prepreg materials, the material state time trajectory tracking unit faces a matching fault due to the difference in crosslinking activation energy of different resin systems. The system calls the material feature mapping procedure, reads the in vitro differential scanning calorimetry fingerprint data, and calculates the material physical failure critical time of the current batch of materials through a two-layer fully connected network architecture. And the critical time for physical failure of materials Write the dynamic urgency index into the address register of the dynamic production scheduling arbitration unit. Maintaining a stable working window around the risk threshold for convergence, subsequent batches of materials to be processed are restored to a closed-loop flow state with zero degradation overflow. This two-layer fully connected network architecture consists of an input layer, a hidden layer with 64 nodes, and a single-output output layer. The network's input vector contains three discrete features extracted from the in vitro differential scanning calorimetry curve: the initial exothermic temperature, the peak exothermic temperature, and the total exothermic enthalpy. The activation function of the hidden layer is a linear rectified function, while the output layer uses an identity activation function to directly output the smoothed time scalar. The dataset used for network training comes from the correlation records of standard deviation scanning calorimetry test data and actual physical failure time of 120 sets of known carbon fiber prepreg samples with different formulation ratios and storage histories. The weight matrix is ​​calibrated and converged through the error backpropagation algorithm, achieving an accurate mapping from the intrinsic reaction characteristics of the material to the macroscopic failure time parameters.

[0049] The above description is only a few preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, technical solutions formed by replacing the above-mentioned features with the technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A carbon fiber product manufacturing and operation management system, characterized in that, The system includes: The equipment spatiotemporal state matrix generation unit is used to acquire the remaining volume data and energy efficiency data of the curing molding equipment, convert the remaining volume data and energy efficiency data into boundary scalars, and linearly assemble the boundary scalars through the global cluster-level bus to generate a resource capacity occupancy state matrix. The queuing delay dynamic correction module is used to monitor the virtual queuing time of each material batch in the waiting queue, and calculate the priority step coefficient when the virtual queuing time reaches the safety critical threshold, and fine-tune the input priority queue to update the dynamic weight. The discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, update material life deduction data according to the acceleration penalty coefficient, and generate step operator constraints composed of internal state matrix and state feature terms. The dynamic production scheduling arbitration unit is used to construct an input priority queue based on the cumulative heat exposure time series data of each material batch, and combine it with the energy efficiency configuration state vector of the solidification equipment. It uses the heat load feature matrix to be loaded to reorganize the equipment available volume remaining items in the resource capacity occupancy state matrix, calculates the temperature rise rate loss item caused by the candidate material combination through the dynamic compensation operator, corrects the virtual queuing time and outputs the production scheduling instruction.

2. The carbon fiber product manufacturing and operation management system according to claim 1, characterized in that, The device spatiotemporal state matrix generation unit includes local device-level nodes and a global cluster-level bus. The local device-level nodes are used to calculate the available volume remaining items and energy efficiency data of the curing and molding equipment, and convert the available volume remaining items and energy efficiency data into boundary scalars and send them to the global cluster-level bus. The global cluster-level bus is used to receive the boundary scalars and generate the resource capacity duty cycle state matrix through linear assembly.

3. The carbon fiber product manufacturing operation management system according to claim 1, characterized in that, The queuing delay dynamic correction module is used to monitor the virtual queuing time of each batch of materials in the waiting queue in real time. When the safety critical threshold is met, the priority step coefficient is calculated, and the input priority queue is fine-tuned through discrete events to update the dynamic weight. The delay deviation caused by the temperature rise rate loss term is converted into a step operator constraint between the internal state matrix and the state feature term.

4. The carbon fiber product manufacturing operation management system according to claim 1, characterized in that, The discrete state data interpolation module is used to read historical ambient temperature references when the monitoring data stream is interrupted, and update the material life deduction data according to the acceleration penalty coefficient to smooth and compensate the cumulative heat exposure time series data of each material batch, and maintain the continuous evolution expression of the dynamic weights in the input priority queue.

5. The carbon fiber product manufacturing operation management system according to claim 1, characterized in that, The dynamic production scheduling arbitration unit is used to reorganize the available volume remaining items of the equipment in the resource capacity duty cycle matrix at the parameter level using the heat load feature matrix to be loaded; the dynamic production scheduling arbitration unit includes a dynamic compensation operator for the temperature rise of the hot fluid in the curing equipment. The dynamic compensation operator is used to directly read the intrinsic rated heating power of the equipment stored in the resource capacity duty cycle matrix and calculate the temperature rise rate loss term caused by the combination of candidate materials entering the equipment.

6. The carbon fiber product manufacturing operation management system according to claim 5, characterized in that, The calculation rule for the temperature rise rate loss term is reflected as a step-wise correction of energy efficiency degradation caused by the total mass of the material. The dynamic production scheduling arbitration unit linearly sums the mass of each material batch in the heat load feature matrix to be loaded to obtain the total mass of the material. When the total mass of the material exceeds the loading heat capacity threshold, the dynamic energy efficiency degradation rule is triggered to reduce the virtual temperature rise efficiency benchmark of the curing molding equipment.

7. The carbon fiber product manufacturing operation management system according to claim 6, characterized in that, The dynamic production scheduling arbitration unit feeds back the curing process cycle extension time caused by the temperature rise rate loss to the waiting queue, automatically rewriting the virtual queuing time of each batch of non-failed materials in the waiting state, and adding the secondary waiting increment caused by the extension of the thermal inertia of the current batch.

8. The carbon fiber product manufacturing operation management system according to claim 7, characterized in that, The dynamic scheduling arbitration unit re-introduces the corrected virtual queuing time into the calculation flow of the dynamic urgency index, using the formula: Calculate the dynamic severity index for each material batch; among which, This represents the dynamic urgency index for each batch of materials. This is the cumulative heat exposure time constant extracted from the physical properties of each material batch. This refers to the current cumulative heat exposure time read from the cumulative heat exposure time series data of each material batch. This is the corrected virtual queuing time.

9. A carbon fiber product manufacturing operation management system according to claim 8, characterized in that, When the revised dynamic severity index of each material batch crosses the risk threshold At that time, the dynamic production scheduling arbitration unit intercepts the currently loaded combination and triggers a hard truncation.

10. A carbon fiber product manufacturing operation management system according to claim 1, characterized in that, The energy efficiency configuration state vector of the curing equipment includes the equipment's rated thermal power, heat flow field uniformity coefficient, and upper limit value of the tank volume; The dynamic production scheduling arbitration unit determines the capacity matching result of a single curing processing equipment based on the matching degree between the energy efficiency configuration state vector of the curing equipment and the cumulative heat exposure time series data of each batch of materials.

Images