Polyurethane foaming mold self-adaptive air extraction control method and system based on multi-section feedback of in-mold pressure

Abstract

The application discloses a polyurethane foaming mold self-adaptive air extraction control method and system based on multi-section feedback of mold internal pressure, the method acquires pressure signals in real time and carries out effectiveness diagnosis through arranging multi-region pressure sensors in a mold cavity; the first-order derivative of pressure is calculated, and dynamic threshold is combined to accurately identify pressure rising section, foaming expansion section and gel curing section; a cascade PID architecture is adopted, a master controller outputs a pressure change rate target based on a multi-region pressure fusion value, and a slave controller adjusts an air extraction rate; when signal failure is detected, a virtual pressure value is reconstructed through a reverse distance weighted interpolation method, and a parameter library is optimized based on product quality feedback self-learning in combination with a genetic algorithm. The application solves the problems that traditional air extraction control relies on fixed timing, has poor anti-interference performance and is difficult to adjust parameters, significantly improves pressure control precision and system robustness, reduces the waste rate, and adapts to flexible foaming production requirements.

Description

Technical Field

[0001] This invention relates to the field of polymer material processing technology, and in particular to an adaptive air extraction control method and system for polyurethane foam molds based on multi-segment feedback of in-mold pressure. Background Technology

[0002] In the polyurethane foaming process, the pressure evolution within the mold cavity directly determines the cell structure and the quality of the final product. Vacuum control, as the core means of regulating internal pressure, directly affects the filling fullness, density uniformity, and surface defect rate. However, existing vacuum control systems generally rely on fixed injection timing or single-point pressure threshold triggering, which cannot adapt to the differences in reaction characteristics of raw materials with different isocyanate indices and combined polyether viscosities. They also struggle to cope with changes in operating conditions such as ambient temperature fluctuations and sensor signal interference, often resulting in over- or under-vacuuming due to stage recognition lag. Furthermore, traditional single-loop PID control has weak anti-interference capabilities when facing the large lag and strong nonlinear characteristics of the foaming process. Sensor failure or control oscillation can easily lead to the scrapping of the entire mold, and parameter tuning is highly dependent on manual experience, making it difficult to meet the efficiency requirements of flexible production with multiple varieties and small batches. Summary of the Invention

[0003] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a complete adaptive air extraction control method and system for polyurethane foam molds based on multi-segment feedback of in-mold pressure, which solves the problems in the prior art where air extraction control relies on fixed injection timing or single-point pressure threshold triggering, and the low accuracy of stage identification leads to defects such as bubble rupture and material shortage shrinkage in the product.

[0004] To achieve the above and other related objectives, the present invention provides the following technical solution:

[0005] An adaptive evacuation control method for polyurethane foam molds based on multi-segment feedback of in-mold pressure is characterized by the following steps: S1, during the foaming molding cycle, pressure signals from multiple monitoring areas within the mold cavity are collected in real time, and the effectiveness of each pressure signal is monitored simultaneously; S2, the first derivative of the pressure signal within a sliding time window is calculated, and based on the first derivative and the pressure value, the current foaming stage is identified online, wherein the foaming stage includes at least a pressure rise stage, a foaming expansion stage, and a gel solidification stage; S3, according to the identified foaming stage, the corresponding target pressure curve and PID control parameter set are retrieved from a preset multi-segment control parameter library, wherein the parameter library is associated with different isocyanate indices and combinations of polymers. A raw material property mapping table for ether viscosity is used, with independent parameter groups pre-stored for raw materials of different properties; S4, the pressure deviation is calculated by comparing the real-time collected pressure value with the target pressure value corresponding to the current stage, and the pumping rate of the pumping system is dynamically adjusted according to the pressure deviation using a closed-loop control algorithm; S5, during the foaming and molding process, the operating status of the closed-loop control system is monitored in real time, and when control instability or signal abnormality is detected, fault-tolerant parameter adjustment is automatically executed to maintain the continuous and stable operation of the pumping control; S6, after each foaming and molding is completed, the target pressure curve and PID control parameters corresponding to the raw material properties in the multi-segment control parameter library are self-learned, optimized, and updated based on the quality inspection results of this molding.

[0006] Furthermore, the validity monitoring of the pressure signal specifically includes at least one of the following methods: when the continuous sampled value remains unchanged or the change is less than the sensor resolution for a period exceeding a preset first timeout threshold, the pressure signal is determined to be invalid; when the sampled value is lower than a preset lower limit of the range or higher than a preset upper limit of the range, the pressure signal is determined to be invalid; when the rate of change of the sampled value exceeds a preset maximum rate of change threshold, the pressure signal is determined to be invalid; when the correlation coefficient between a pressure signal and the pressure signal in an adjacent monitoring area is lower than a preset correlation coefficient threshold, the pressure signal is determined to be invalid; when the pressure signal does not meet any of the above invalidity determination conditions, the pressure signal is determined to be valid.

[0007] Furthermore, the specific method for online identification of the foaming stage includes: when the first derivative of the pressure signal within the sliding time window is continuously greater than a first threshold... When the pressure rises after injection, it is determined that the current stage is in the pressure rise phase; when the first derivative changes from positive to negative and its absolute value is less than the second threshold. When the current stage is determined to be foaming and expanding; when the absolute value of the first derivative is continuously less than the third threshold... When the intramold pressure value is within a preset gel curing pressure range and remains within a preset duration, it is determined that the current stage is the gel curing phase; wherein, the third threshold... The value is less than the second threshold. .

[0008] Furthermore, the first threshold Second threshold and the third threshold The setting method includes: pre-storing the free foaming rate curves of polyurethane raw materials under different ambient temperatures; and obtaining the actual rate of rise of the raw materials under the current production environment by linear interpolation based on the measured temperature of the current production environment. ;set up , , ,in , , This is a dimensionless correction factor. The preset reference pressure value, The gel time at the current temperature. This refers to the noise floor amplitude of the pressure sensor. This represents the sampling interval of the pressure sensor.

[0009] Furthermore, in the multi-segment control parameter library: the target pressure curve corresponding to the pressure rise segment is the pressure curve that rises from the initial pressure to the first target pressure at a first rate of increase. The monotonically increasing curve corresponds to PID control parameters including the first proportional coefficient, the first integral coefficient, and the first derivative coefficient; the target pressure curve corresponding to the foaming expansion section is a curve that rises at a second rate from... Rise to the second target pressure The gradual rise curve, where the second rise rate is less than the first rise rate, corresponds to PID control parameters including a second proportional coefficient, a second integral coefficient, and a second derivative coefficient; the target pressure curve corresponding to the gel solidification stage is maintained at a third target pressure. The constant pressure curve corresponds to PID control parameters including the third proportional coefficient, the third integral coefficient, and the third derivative coefficient; among which, .

[0010] Furthermore, the closed-loop control algorithm is a cascaded PID control structure, including a main controller and a secondary controller. The main controller takes the deviation between the weighted fusion value of the pressure values ​​of each monitoring area and the corresponding target pressure value as input, and outputs the target value of the pressure change rate. The secondary controller takes the deviation between the actual pressure change rate and the target change rate output by the main controller as input, and outputs the control quantity of the pumping system. The main controller and the secondary controller each have an independent set of proportional coefficients, integral coefficients, and derivative coefficients. The proportional coefficients, integral coefficients, and derivative coefficients of the main controller and the secondary controller are independently retrieved from the multi-segment control parameter library according to the current foaming stage.

[0011] Furthermore, S5 includes: real-time monitoring of the changing trend of the first derivative; when the first derivative exhibits alternating positive and negative oscillations within a preset time window and the oscillation frequency exceeds a preset frequency threshold and the oscillation amplitude exceeds a preset amplitude threshold, it is determined that the air extraction system is experiencing abnormal oscillation conditions, and the integral coefficient of the PID controller is automatically reduced to a preset proportion while the derivative coefficient is increased to a preset multiple; when the pressure signal of a certain monitoring area is lost or exceeds a reasonable range, the pressure value of that area is estimated by spatial interpolation using the pressure values ​​of adjacent monitoring areas, and a sensor fault alarm signal is issued simultaneously.

[0012] Furthermore, estimating the pressure value of the area using spatial interpolation based on the pressure values ​​of adjacent monitoring areas includes reconstructing the virtual pressure value of the failed monitoring area using inverse distance weighted interpolation. Centered on the failure monitoring area, N effective monitoring areas directly adjacent to it in the spatial topology are selected to construct an interpolation neighborhood; the virtual pressure value The calculation formula is as follows: ;in For the first Pressure values ​​of adjacent effective monitoring areas For the first The weighting coefficients of adjacent effective monitoring areas, and , For the first The Euclidean distance between each adjacent effective monitoring area and the abnormal area; The preset distance attenuation index has a value range of [1,2].

[0013] Furthermore, the specific method for self-learning optimization and updating includes: after each foaming and molding process, obtaining the density uniformity test value and surface quality test value of the product; when the density uniformity test value or surface quality test value is lower than a preset qualified threshold, recording the actual pressure curve and control parameters of each stage in this foaming and molding process as defect samples; using the integral sum of squares of the pressure deviation of the defect samples as the objective function, using a genetic algorithm to optimize and correct the target pressure curve and PID control parameters under the corresponding raw material properties, and updating the optimized parameters to the parameter group of the corresponding raw material properties in the multi-segment control parameter library.

[0014] In another embodiment of the present invention, an adaptive vacuum control system for polyurethane foam molds based on multi-segment feedback of in-mold pressure is provided, comprising: a pressure sensing module, configured with multiple pressure sensors distributed on the feed side, middle and end of the mold cavity, for real-time acquisition of pressure signals in each monitoring area during the foaming cycle, and a built-in signal validity diagnosis unit for synchronously monitoring the validity of each pressure signal; a stage identification module, connected to the pressure sensing module, configured with an edge computing unit for calculating the first derivative of the pressure signal within a sliding time window, and identifying the current foaming stage online based on the first derivative combined with the pressure value; and a parameter scheduling module, connected to the stage identification module, having a built-in multi-segment control parameter library, the parameter library being associated with a raw material property mapping table of different isocyanate indices and combined polyether viscosities, for retrieving the corresponding target pressure curve and PID control parameters according to the identified foaming stage. The system comprises: a closed-loop control module, connected to the parameter scheduling module and the pressure sensing module, configured with a cascaded PID control architecture, including a main controller and a secondary controller; the main controller calculates the deviation between the weighted fusion value of the pressure values ​​in each monitoring area and the target pressure value and outputs the target pressure change rate; the secondary controller calculates the deviation between the actual pressure change rate and the target pressure change rate and outputs the control quantity of the pumping system; a fault-tolerant management module, connected to both the closed-loop control module and the pressure sensing module, performs fault-tolerant parameter adjustments when control instability or signal abnormalities are detected; a self-learning optimization module, connected to both the closed-loop control module and the parameter scheduling module, updates the target pressure curve and PID control parameters in the multi-segment control parameter library based on the molding quality detection results after each foaming and molding process; and an execution module, connected to the closed-loop control module, responds to the control quantity to adjust the pumping rate of the pumping system.

[0015] As described above, the adaptive venting control method and system for polyurethane foam molds based on multi-segment feedback of in-mold pressure of the present invention has the following beneficial effects: This solution effectively breaks through the technical bottleneck of traditional venting control through multi-segment feedback and adaptive control mechanisms: First, based on the stage identification logic of pressure change rate and spatial distribution characteristics, combined with a threshold system that dynamically adapts to raw material properties and ambient temperature, it achieves precise segmentation of each foaming stage, and can accurately match the dynamic gas generation and degassing requirements of each stage of pressure rise, foam expansion, and gel solidification, fundamentally reducing defects such as uneven cell structure, material shortage, and overflow; Second, the cascade PID architecture and the inclusion of oscillation suppression and spatial interpolation... The dual-layer fault-tolerance mechanism of the value function significantly improves the control stability of the system, maintaining production continuity even when sensors fail or operating conditions fluctuate drastically, greatly reducing the risk of unplanned downtime. Furthermore, the self-learning optimization module establishes a feedback loop between "in-mold pressure and final quality," enabling iterative optimization of the parameter library through intelligent algorithms, greatly shortening the material changeover and debugging cycle, and continuously driving up the yield rate. In addition, this solution can be implemented using only existing sensors and actuators, without large-scale hardware modifications, and can be flexibly adapted to various polyurethane foaming production lines, combining high control accuracy with low modification costs, providing core technical support for flexible and intelligent foaming production. Attached Figure Description

[0016] Figure 1 The diagram shown is a flowchart of the method of the present invention. Detailed Implementation

[0017] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that, unless otherwise specified, the following embodiments and features described herein can be combined with each other.

[0018] Please see Figure 1This invention provides an adaptive evacuation control method for polyurethane foam molds based on multi-segment feedback of in-mold pressure, comprising the following steps: S1, during the foaming molding cycle, real-time acquisition of pressure signals from multiple monitoring areas within the mold cavity, and simultaneous monitoring of the effectiveness of each pressure signal; S2, calculation of the first derivative of the pressure signal within a sliding time window, and online identification of the current foaming stage based on the first derivative combined with the pressure value, wherein the foaming stage includes at least a pressure rise stage, a foaming expansion stage, and a gel solidification stage; S3, according to the identified foaming stage, retrieval of the corresponding target pressure curve and PID control parameter set from a preset multi-segment control parameter library, wherein the parameter library is associated with different isocyanate indices and combinations. A raw material property mapping table for polyether viscosity is used, with independent parameter groups pre-stored for raw materials with different properties; S4, the pressure deviation is calculated by comparing the real-time collected pressure value with the target pressure value corresponding to the current stage, and the pumping rate of the pumping system is dynamically adjusted according to the pressure deviation using a closed-loop control algorithm; S5, during the foaming and molding process, the operating status of the closed-loop control system is monitored in real time, and when control instability or signal abnormality is detected, fault-tolerant parameter adjustment is automatically executed to maintain the continuous and stable operation of the pumping control; S6, after each foaming and molding is completed, the target pressure curve and PID control parameters corresponding to the raw material properties in the multi-segment control parameter library are self-learned, optimized, and updated based on the quality inspection results of this molding.

[0019] This solution is based on the characteristic that the pressure in the polyurethane foaming process exhibits a regular evolution with the reaction progress. First, during the molding cycle, pressure signals from multiple regions are collected by a sensor array distributed on the feed side, middle, and end of the mold cavity. Simultaneously, signal validity diagnosis is performed to eliminate abnormal data and obtain accurate intramold pressure field information. Then, the first derivative of the pressure signal within the sliding window is calculated. Combined with the absolute pressure value and pre-stored raw material property characteristics, three characteristic stages are accurately divided: the pressure rise stage, the foaming expansion stage, and the gel curing stage. The corresponding target pressure curve and PID parameter group are retrieved from the multi-segment control library of associated raw material parameters. The pumping rate is dynamically adjusted through a cascade PID control architecture to ensure that the intramold pressure always meets the process requirements of each stage.

[0020] During operation, the system monitors the closed-loop control status in real time. If high-frequency oscillations are detected in the first derivative of the pressure, the integral coefficient is adaptively reduced and the differential coefficient is increased to suppress fluctuations. If the pressure signal fails in a certain area, the virtual pressure value is reconstructed based on the spatial topology relationship using the inverse distance weighted interpolation method to ensure the continuous operation of the control logic. After each round of molding, the system collects test data such as product density uniformity and surface quality. If the quality does not meet the standards, the pressure curve and control parameters of this round are recorded as defect samples. Using the integral square sum of pressure deviation as the objective function, the system optimizes the attribute parameter set of the corresponding raw materials through a genetic algorithm to achieve self-iteration of the control strategy and adapt to batch fluctuations of raw materials and changes in the process environment.

[0021] This solution overcomes the limitations of traditional air extraction control that relies on fixed timing or single-point pressure feedback. Through multi-stage adaptive pressure curves and cascade control strategies, it accurately matches the dynamic requirements of gas production and exhaust at each stage of foaming, effectively avoiding defects such as bubble rupture and liquid raw material loss caused by excessive air extraction, as well as defects such as trapped air, material shortage, and uneven density caused by insufficient air extraction. At the same time, relying on signal validity diagnosis and interpolation fault tolerance mechanisms, it significantly reduces the impact of sensor failure and signal interference on production stability, and significantly improves the robustness and environmental adaptability of the control system.

[0022] The validity monitoring of the pressure signal specifically includes at least one of the following methods: when the continuous sampled value remains unchanged or the change is less than the sensor resolution for a period of time exceeding a preset first timeout threshold, the pressure signal is determined to be invalid; when the sampled value is lower than a preset lower limit of the range or higher than a preset upper limit of the range, the pressure signal is determined to be invalid; when the rate of change of the sampled value exceeds a preset maximum rate of change threshold, the pressure signal is determined to be invalid; when the correlation coefficient between a pressure signal and the pressure signal of an adjacent monitoring area is lower than a preset correlation coefficient threshold, the pressure signal is determined to be invalid; when the pressure signal does not meet any of the above invalidation conditions, the pressure signal is determined to be valid.

[0023] This step is the front-end verification stage of pressure signal processing, executed synchronously with pressure signal acquisition. Based on the orderly evolution of intramold pressure during polyurethane foaming, it runs in parallel with multi-dimensional validity judgment logic: For static failures caused by sensor jamming or transmission line interruption, it is identified by determining whether the time for continuous sampling values ​​to remain unchanged or change less than the sensor resolution exceeds a first timeout threshold; for extreme value failures caused by sensor blockage or overload impact, it is identified by determining whether the sampling values ​​exceed the preset upper and lower limits of the range; for signal distortion caused by foam rupture or electromagnetic interference, it is identified by determining whether the rate of change of the sampling values ​​exceeds the maximum rate of change threshold; for logical failures caused by sensor installation misalignment or interference from foreign objects in the local cavity, it is identified by determining whether the correlation coefficient between the current signal and the signals of adjacent monitoring areas is lower than the correlation coefficient threshold. Only when the signal does not trigger any of the invalid judgment conditions is it marked as valid and flows into the subsequent derivative calculation and stage identification stages, ensuring the authenticity of the input data from the source.

[0024] The specific method for online identification of the foaming stage includes: when the first derivative of the pressure signal within the sliding time window is continuously greater than a first threshold. When the pressure rises after injection, it is determined that the current stage is in the pressure rise phase; when the first derivative changes from positive to negative and its absolute value is less than the second threshold. When the current stage is determined to be foaming and expanding; when the absolute value of the first derivative is continuously less than the third threshold... When the intramold pressure value is within a preset gel curing pressure range and remains within a preset duration, it is determined that the current stage is the gel curing phase; wherein, the third threshold... The value is less than the second threshold. .

[0025] This step leverages the strong correlation between the pressure change rate and each reaction stage during the polyurethane foaming reaction. It calculates the first derivative of the pressure signal using a sliding time window to filter out instantaneous noise interference. The dynamic characteristics of the derivative are then matched to the corresponding process characteristics: after injection, the reaction starts rapidly, the gas production rate exceeds the exhaust rate, the in-mold pressure rises sharply, and the first derivative remains above the first threshold of the appropriate raw material rise rate. This indicates the pressure increase phase; as the bubble gradually fills the cavity, the pressure increase slows down, the first derivative changes from positive to negative, and its absolute value is limited to the second threshold that adapts to the gelation process. Within this range, excluding interference from abnormal pressure relief, it is determined to be the foaming expansion stage; once the reaction enters the gelation and solidification period, the pressure tends to stabilize, and the first derivative approaches zero, becoming less than... The third threshold Define steady-state characteristics, and at the same time verify whether the pressure value is within the preset curing range and maintained for a sufficient duration to avoid misjudgment due to instantaneous fluctuations. Based on this, lock the gel curing segment. The entire judgment logic is completely in line with the evolution law of foaming kinetics.

[0026] Compared to traditional methods that rely on fixed injection timing or single-point pressure thresholds to define stages, this solution offers extremely high adaptability and reliability. On one hand, the dynamic determination based on pressure change rate is unaffected by raw material batch fluctuations or changes in ambient temperature, solving the problem that fixed timing cannot adapt to fluctuations in process parameters, and significantly improving the accuracy of stage determination. On the other hand, the hierarchical design of the three-level threshold conforms to the pressure evolution law of each foaming stage. Combined with pressure range verification and duration verification, it can effectively filter out instantaneous interferences such as foam rupture and mechanical vibration, significantly reducing the misjudgment rate.

[0027] The first threshold Second threshold and the third threshold The setting method includes: pre-storing the free foaming rate curves of polyurethane raw materials under different ambient temperatures; and obtaining the actual rate of rise of the raw materials under the current production environment by linear interpolation based on the measured temperature of the current production environment. ;set up , , ,in , , This is a dimensionless correction factor. The preset reference pressure value, The gel time at the current temperature. This refers to the noise floor amplitude of the pressure sensor. This represents the sampling interval of the pressure sensor.

[0028] This step addresses the pain point of traditional fixed thresholds failing to adapt to raw material batch fluctuations and environmental temperature changes, resulting in low stage identification accuracy. It constructs an adaptive threshold calculation logic based on foaming reaction kinetics and the inherent characteristics of the sensor: First, it pre-stores the free foaming rate curves of various polyurethane raw materials under different environmental temperatures as the basic data pool for threshold calculation; then, after real-time acquisition of the measured temperature of the current production environment, it uses linear interpolation to match the actual rate of rise of the raw materials under the current operating conditions. With gel time Subsequently, the reaction characteristic parameters were combined with the sensor characteristic parameters, and a dimensionless correction coefficient was used to adapt to the reactivity of different raw materials, mold cavity resistance, and system safety margin. Three judgment thresholds were calculated respectively: T1 is proportional to the raw material rising rate, matching the rate characteristics of the pressure rise segment; T2 is combined with the reference pressure. Matching the gelation time setting to the rate decay characteristics of the foaming expansion phase; T3 is based on sensor noise floor. The calculation of the sampling interval Δt matches the steady-state noise characteristics of the gel curing section. The hierarchical design of the three components is in complete agreement with the pressure evolution law of the foaming process.

[0029] The threshold setting logic of this solution breaks through the limitations of traditional empirical setting, and realizes the full-dimensional linkage and adaptation of the threshold with raw material properties, ambient temperature and sensor characteristics: On the one hand, T1 is dynamically adjusted with the raw material rising rate, avoiding misjudgment of the rising stage caused by fluctuations in reactivity; T2 is combined with gel time setting to adapt to the differences in curing rate under different formulations and temperatures; T3 is calculated based on sensor noise floor and sampling interval, so it will not be falsely triggered by high-frequency noise, nor will it miss steady-state characteristics due to excessively high threshold, which greatly improves the accuracy of stage identification and anti-interference ability.

[0030] In the multi-segment control parameter library: the target pressure curve corresponding to the pressure rise segment is the pressure curve that rises from the initial pressure to the first target pressure at a first rise rate. The monotonically increasing curve corresponds to PID control parameters including the first proportional coefficient, the first integral coefficient, and the first derivative coefficient; the target pressure curve corresponding to the foaming expansion section is a curve that rises at a second rate from... Rise to the second target pressure The gradual rise curve, where the second rise rate is less than the first rise rate, corresponds to PID control parameters including a second proportional coefficient, a second integral coefficient, and a second derivative coefficient; the target pressure curve corresponding to the gel solidification stage is maintained at a third target pressure. constant pressure curve, and the corresponding PID control parameters include a third proportional coefficient, a third integral coefficient and a third differential coefficient; wherein .

[0031] In this step, based on the pressure evolution law of the whole polyurethane foaming cycle, differentiated control strategies for each stage are pre-configured in the multi-stage control parameter library: for the pressure rising stage after injection, a monotonically increasing curve rising from the initial pressure to the first target pressure P1 at a first rising rate is designed, which is matched with a PID parameter group focusing on response speed, so as to quickly respond to injection impact and efficiently discharge residual air in the cavity; for the foaming expansion stage, a slow rising curve rising from P1 to the second target pressure P2 at a second rising rate lower than the first rising rate is designed, which is matched with a PID parameter group focusing on anti-interference and overshoot prevention, so as to balance the exhaust demand and the stability of cell structure; for the gel curing stage, a constant pressure curve maintained at the third target pressure P3 is designed, which is matched with a PID parameter group focusing on steady-state accuracy, so as to ensure pressure stability during cell shaping; wherein the gradient of P1<P2≤P3 fully conforms to the physical law of gradual accumulation of gas produced by the foaming reaction, and the parameter library is associated with the raw material property mapping table. Raw materials with different isocyanate indices and combined polyether viscosities correspond to independent curves and parameter groups, which can be directly called after stage identification is completed, without real-time calculation and has low response delay.

[0032] This solution breaks through the limitation of single target pressure and general PID parameters adopted in traditional suction control, and significantly improves the molding quality through the control strategy of matching process requirements by stages: the fast rising curve and large response characteristic parameters in the pressure rising stage can avoid the defects of trapped gas and insufficient filling; the slow rising curve design in the foaming expansion stage can prevent bulging caused by sudden pressure rise and cell breakage caused by too fast suction, so as to ensure uniform and dense cell structure; the constant pressure curve configuration in the gel curing stage can eliminate steady-state pressure errors and avoid defects such as shrinkage cavity and deformation, and the product density uniformity is improved by more than 15% compared with traditional control. Meanwhile, the gradient design of P1<P2≤P3 avoids pressure jump during stage switching, reduces the impact load on the suction system and prolongs the service life of the equipment.

[0033] The closed-loop control algorithm is a cascade PID control structure, including a main controller and a secondary controller; the main controller takes the deviation between the weighted fusion value of the pressure values of each monitoring area and the corresponding target pressure value as input, and outputs the target value of the pressure change rate; the secondary controller takes the deviation between the actual pressure change rate and the target change rate output by the main controller as input, and outputs the control quantity of the suction system; the main controller and the secondary controller each independently have a group of proportional coefficient, integral coefficient and differential coefficient, and each group of proportional coefficient, integral coefficient and differential coefficient of the main controller and the secondary controller are independently called from the multi-section control parameter library according to the current foaming stage.

[0034] This step employs a cascaded PID control architecture to specifically address the shortcomings of traditional single-loop control, such as poor adaptability and weak anti-interference capability. The main controller first performs weighted fusion of pressure values ​​from multiple monitoring areas to eliminate the influence of single-point sensor deviations or local pressure anomalies. Then, it compares the fused pressure value with the target pressure value for the current stage and outputs a target pressure change rate adapted to the current operating conditions, controlling the pumping process from the perspective of pressure steady-state accuracy. The secondary controller takes the actual collected pressure change rate and the target change rate output by the main controller as input, calculates the deviation between the two, and outputs the specific control quantity of the pumping system, implementing the adjustment action from the perspective of dynamic response speed. The main and secondary controllers each have completely independent PID parameter sets, and their parameters are decoupled. In each foaming stage, the corresponding optimal parameter set is matched from a pre-stored multi-segment control parameter library. For example, in the pressure rise stage, a large proportional coefficient is configured for the main controller to improve the response speed, and in the foaming expansion stage, a large derivative coefficient is configured for the secondary controller to suppress pressure oscillations caused by bubble tumbling. The dimensional logic of the parameter sets is completely matched with the signal transmission path to avoid control inaccuracies.

[0035] S5 includes: real-time monitoring of the changing trend of the first derivative; when the first derivative exhibits alternating positive and negative oscillations within a preset time window and the oscillation frequency exceeds a preset frequency threshold and the oscillation amplitude exceeds a preset amplitude threshold, it is determined that the air extraction system is experiencing abnormal oscillation conditions, and the integral coefficient of the PID controller is automatically reduced to a preset ratio while the derivative coefficient is increased to a preset multiple; when the pressure signal of a certain monitoring area is lost or exceeds a reasonable range, the pressure value of that area is estimated by spatial interpolation using the pressure values ​​of adjacent monitoring areas, and a sensor fault alarm signal is issued simultaneously.

[0036] This step constructs a two-layer fault-tolerant logic to address two common faults in foaming production: control instability and signal anomalies. On the one hand, it monitors the changing trend of the first derivative of the pressure signal in real time. Only when alternating positive and negative oscillations occur simultaneously within a preset time window, and both the oscillation frequency and amplitude exceed the threshold, is it determined to be an abnormal oscillation of the pumping system. This avoids misjudging normal pressure fluctuations in the foaming expansion section as faults. Subsequently, it automatically retrieves the pre-stored anti-disturbance parameter set, reduces the PID integral coefficient by a preset ratio to weaken the integral accumulation effect and avoid overshoot, and simultaneously amplifies the derivative coefficient by a preset multiple to enhance the system damping effect and suppress high-frequency oscillations. On the other hand, for failure scenarios such as pressure signal loss or exceeding the range, it utilizes the spatial continuity characteristics of the pressure distribution within the mold cavity. Based on the pressure values ​​of the effective monitoring areas adjacent to the failure point, it reconstructs the virtual pressure value of the failure point through spatial interpolation. This simultaneously issues a sensor fault alarm without interrupting the control logic, reserving buffer time for maintenance and troubleshooting.

[0037] The step of estimating the pressure value of the area using spatial interpolation based on the pressure values ​​of adjacent monitoring areas includes: reconstructing the virtual pressure value of the failed monitoring area using inverse distance weighted interpolation. Centered on the failure monitoring area, N effective monitoring areas directly adjacent to it in the spatial topology are selected to construct an interpolation neighborhood; the virtual pressure value The calculation formula is as follows: ;in For the first Pressure values ​​of adjacent effective monitoring areas For the first The weighting coefficients of adjacent effective monitoring areas, and , For the first The Euclidean distance between each adjacent effective monitoring area and the abnormal area; The preset distance attenuation index has a value range of [1,2].

[0038] This step, based on the foaming process characteristics of continuous pressure distribution within the mold cavity and strong correlation between pressures in adjacent areas, designs spatial interpolation fault-tolerant logic for single-point pressure signal failure scenarios: Taking the failure monitoring area as the core, it prioritizes selecting N effective monitoring areas directly connected to the failure point in spatial topology (e.g., within the same flow channel, without physical barriers) to construct interpolation neighborhoods, avoiding the inclusion of interference data from irrelevant areas; when reconstructing virtual pressure values ​​using the inverse distance weighted interpolation method, it assigns a weight coefficient inversely proportional to the distance to each effective neighborhood point. Take distance of The reciprocal of the power (q∈[1,2]) indicates that the closer the point, the greater its contribution to the pressure of the failure point. This highlights the dominant role of the nearest point while also appropriately referencing the pressure information of slightly farther points to offset local noise interference. Finally, a virtual pressure value that conforms to the current foaming condition is obtained through weighted summation, which can be seamlessly integrated into the subsequent cascaded PID control logic.

[0039] The specific method for self-learning optimization and updating includes: after each foaming and molding process, obtaining the density uniformity test value and surface quality test value of the product; when the density uniformity test value or surface quality test value is lower than the preset qualified threshold, recording the actual pressure curve and control parameters of each stage in this foaming and molding process as defect samples; using the integral square sum of the pressure deviations of the defect samples as the objective function, using a genetic algorithm to optimize and correct the target pressure curve and PID control parameters under the corresponding raw material properties, and updating the optimized parameters to the parameter group of the corresponding raw material properties in the multi-segment control parameter library.

[0040] This step constructs a self-evolving logic that links "in-mold pressure control - final product quality" across domains, making up for the shortcomings of traditional vacuum control, which only closes the in-mold pressure and cannot be linked to the quality of the final product. After each round of foaming and molding, the final inspection data such as the density uniformity and surface quality of the product are first acquired. The optimization process is only triggered when the inspection value is lower than the qualified threshold, avoiding meaningless calculations that consume computing power. After triggering, the actual pressure curves of each stage of this round, the PID parameters used, and the corresponding raw material properties are packaged and recorded as defect samples. The integral square of the pressure deviation of the entire round of molding is used as the objective function. This index can comprehensively reflect the tracking accuracy of the pressure on the target curve throughout the entire cycle and is strongly negatively correlated with the product density and surface quality. Then, a genetic algorithm is used to iteratively optimize in the attribute parameter space of the corresponding raw material to obtain a target pressure curve and PID parameters that are more suitable for the raw material. After that, the results are written back to the corresponding parameter group in the multi-segment control parameter library to realize the closed-loop iteration of "production-quality inspection-optimization-reproduction".

[0041] In another embodiment of the present invention, an adaptive vacuum control system for a polyurethane foam mold based on multi-segment feedback of in-mold pressure is provided, comprising: a pressure sensing module, configured with multiple pressure sensors distributed on the feed side, middle part, and end of the mold cavity, for real-time acquisition of pressure signals in each monitoring area during the foaming cycle, and having a built-in signal validity diagnosis unit for synchronously monitoring the validity of each pressure signal; a stage identification module, connected to the pressure sensing module, configured with an edge computing unit for calculating the first derivative of the pressure signal within a sliding time window, and identifying the current foaming stage online based on the first derivative combined with the pressure value; and a parameter scheduling module, connected to the stage identification module, having a built-in multi-segment control parameter library, the parameter library being associated with a raw material property mapping table of different isocyanate indices and combined polyether viscosities, for retrieving the corresponding target pressure curve and PID control parameters according to the identified foaming stage. The system comprises: a closed-loop control module, connected to the parameter scheduling module and the pressure sensing module, configured with a cascaded PID control architecture, including a main controller and a secondary controller; the main controller calculates the deviation between the weighted fusion value of the pressure values ​​in each monitoring area and the target pressure value and outputs the target pressure change rate; the secondary controller calculates the deviation between the actual pressure change rate and the target pressure change rate and outputs the control quantity of the pumping system; a fault-tolerant management module, connected to both the closed-loop control module and the pressure sensing module, performs fault-tolerant parameter adjustments when control instability or signal abnormalities are detected; a self-learning optimization module, connected to both the closed-loop control module and the parameter scheduling module, updates the target pressure curve and PID control parameters in the multi-segment control parameter library based on the molding quality detection results after each foaming and molding process; and an execution module, connected to the closed-loop control module, responds to the control quantity to adjust the pumping rate of the pumping system.

[0042] This system, based on a modular architecture, implements a full-process adaptive vacuum control method for polyurethane foam molds based on multi-segment feedback of in-mold pressure. Each module collaborates via signal links to adapt to the nonlinear process characteristics of polyurethane foaming: the pressure sensing module collects multi-region pressure signals through sensor arrays distributed at the inlet, middle, and end of the mold cavity; a built-in signal validity diagnosis unit simultaneously intercepts abnormal data, outputting only valid signals to the stage identification module; the stage identification module, relying on an edge computing unit, performs sliding window first-order derivative calculations locally, combining the absolute pressure value with pre-stored raw material characteristics to accurately identify the current foaming stage and send the results to the parameter scheduling module; the parameter scheduling module then adjusts the parameters based on the current stage and the original material characteristics. The system retrieves the appropriate target pressure curve and PID parameter set from a multi-segment control library containing multiple associated formula parameters, based on the material properties, and sends them to the closed-loop control module. The closed-loop control module implements hierarchical control through a cascaded PID architecture. The main controller integrates the deviations between the pressure values ​​of multiple regions and the target value to output the pressure change rate target, while the secondary controller tracks the deviation between the actual change rate and the target value and outputs the control quantity to drive the execution module to adjust the pumping rate. The fault-tolerant management module simultaneously monitors the control process and signal status, and quickly triggers parameter adjustment or interpolation reconstruction when oscillations or signal failures occur to ensure control continuity. After each molding cycle, the self-learning optimization module iteratively optimizes the parameter library based on the product quality inspection results to achieve self-evolution of the control strategy.

[0043] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. All equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this invention should still be covered by the claims of this invention.

Claims

1. An adaptive air extraction control method for polyurethane foam molds based on multi-segment feedback of in-mold pressure, characterized in that, Includes the following steps: S1, During the foaming molding cycle, pressure signals from multiple monitoring areas within the mold cavity are collected in real time, and the effectiveness of each pressure signal is monitored simultaneously. S2, Calculate the first derivative of the pressure signal within the sliding time window, and based on the first derivative and the pressure value, identify the current foaming stage online. The foaming stage includes at least a pressure rise stage, a foaming expansion stage, and a gel solidification stage. S3. Based on the identified foaming stage, retrieve the corresponding target pressure curve and PID control parameter set from the preset multi-segment control parameter library. The parameter library is associated with a raw material property mapping table with different isocyanate indices and combined polyether viscosities. Independent parameter groups are pre-stored for raw materials with different properties. S4, compare the real-time collected pressure value with the target pressure value corresponding to the current stage to calculate the pressure deviation, and dynamically adjust the pumping rate of the pumping system according to the pressure deviation using a closed-loop control algorithm. S5 monitors the operating status of the closed-loop control system in real time during the foaming process. When control instability or abnormal signal is detected, it automatically performs fault-tolerant parameter adjustment to maintain the continuous and stable operation of the air extraction control. S6. After each foaming and molding process is completed, based on the quality inspection results of this molding, the target pressure curve and PID control parameters corresponding to the raw material properties in the multi-segment control parameter library are self-learned, optimized, and updated.

2. The method according to claim 1, characterized in that, The effectiveness monitoring of the pressure signal specifically includes at least one of the following methods: When the continuous sampled value remains unchanged or the change is less than the sensor resolution for a period of time exceeding a preset first timeout threshold, the pressure signal is determined to be invalid. When the sampled value is lower than the preset lower limit of the range or higher than the preset upper limit of the range, the pressure signal is determined to be invalid. When the rate of change of the sampled value exceeds the preset maximum rate of change threshold, the pressure signal is determined to be invalid. When the correlation coefficient between a pressure signal and the pressure signal in an adjacent monitoring area is lower than the preset correlation coefficient threshold, the pressure signal is determined to be invalid. When the pressure signal does not meet any of the above invalidation conditions, the pressure signal is determined to be valid.

3. The method according to claim 1, characterized in that, The specific method for online identification of the foaming stage includes: When the first derivative of the pressure signal within the sliding time window remains greater than the first threshold At that time, it is determined that the current stage is the pressure rise phase after material injection; When the first derivative changes from positive to negative and its absolute value is less than the second threshold At that time, it is determined that the current stage is the foaming and expansion stage; When the absolute value of the first derivative is continuously less than the third threshold When the pressure value inside the mold is within the preset gel curing pressure range and is maintained for a preset time, it is determined that the current stage is the gel curing stage. Wherein, the third threshold The value is less than the second threshold. .

4. The method according to claim 3, characterized in that, The first threshold Second threshold and the third threshold The setup methods include: Pre-store the free foaming rate curves of polyurethane raw materials under different ambient temperatures. Based on the measured temperature of the current production environment, the actual rate of rise of the raw material is obtained by linear interpolation. ; set up , , ,in , , This is a dimensionless correction factor. The preset reference pressure value, The gel time at the current temperature. This refers to the noise floor amplitude of the pressure sensor. This represents the sampling interval of the pressure sensor.

5. The method according to claim 1, characterized in that, In the multi-segment control parameter library: The target pressure curve corresponding to the pressure rise segment is the pressure that rises from the initial pressure to the first target pressure at the first rate of increase. The monotonically increasing curve corresponds to the PID control parameters including the first proportional coefficient, the first integral coefficient, and the first derivative coefficient. The target pressure curve corresponding to the foaming expansion section is from the second rising rate from Rise to the second target pressure The gradual rise curve, the second rise rate is less than the first rise rate, and the corresponding PID control parameters include the second proportional coefficient, the second integral coefficient and the second derivative coefficient. The target pressure curve corresponding to the gel curing section is maintained at the third target pressure. The constant pressure curve corresponds to the PID control parameters including the third proportional coefficient, the third integral coefficient, and the third derivative coefficient. in, .

6. The method according to claim 1, characterized in that, The closed-loop control algorithm is a cascaded PID control structure, including a main controller and a secondary controller; The main controller takes the deviation between the weighted fusion value of the pressure values ​​of each monitoring area and the corresponding target pressure value as input, and outputs the target value of the pressure change rate. The secondary controller takes the deviation between the actual pressure change rate and the target change rate output by the main controller as input and outputs the control quantity of the air extraction system. The main controller and the auxiliary controller each have an independent set of proportional coefficients, integral coefficients and derivative coefficients. The proportional coefficients, integral coefficients and derivative coefficients of the main controller and the auxiliary controller are independently retrieved from the multi-segment control parameter library according to the current foaming stage.

7. The method according to claim 1, characterized in that, S5 includes: The system monitors the changing trend of the first derivative in real time. When the first derivative oscillates alternately between positive and negative within a preset time window and the oscillation frequency exceeds a preset frequency threshold and the oscillation amplitude exceeds a preset amplitude threshold, it determines that the air extraction system is in an abnormal oscillation condition. The system automatically reduces the integral coefficient of the PID controller to a preset ratio and increases the derivative coefficient to a preset multiple. When the pressure signal of a certain monitoring area is lost or exceeds the reasonable range, the pressure value of that area is estimated by spatial interpolation using the pressure values ​​of adjacent monitoring areas, and a sensor fault alarm signal is issued at the same time.

8. The method according to claim 7, characterized in that, The method of estimating the pressure value of a region by spatial interpolation using the pressure values ​​of adjacent monitoring regions includes: The virtual pressure values ​​of the failure monitoring area were reconstructed using the inverse distance weighted interpolation method. Taking the failure monitoring area as the center, N effective monitoring areas that are directly adjacent to it in the spatial topology are selected to construct the interpolation neighborhood; The virtual pressure value The calculation formula is as follows: ;in For the first Pressure values ​​of adjacent effective monitoring areas For the first The weighting coefficients of adjacent effective monitoring areas, and , For the first The Euclidean distance between each adjacent effective monitoring area and the abnormal area; The preset distance attenuation index has a value range of [1,2].

9. The method according to claim 1, characterized in that, The specific methods for self-learning optimization and updating include: After each foaming and molding process, the density uniformity and surface quality of the product are measured. When the density uniformity test value or surface quality test value is lower than the preset qualified threshold, the actual pressure curve and control parameters of each stage in this foaming molding process are recorded as defect samples. Using the integral sum of squares of the pressure deviation of the defective samples as the objective function, a genetic algorithm is used to optimize and correct the target pressure curve and PID control parameters under the corresponding raw material properties, and the optimized parameters are updated to the parameter group of the corresponding raw material properties in the multi-segment control parameter library.

10. A system for implementing the method as claimed in any one of claims 1 to 9 comprises: The pressure sensing module is equipped with multiple pressure sensors distributed on the feed side, middle and end of the mold cavity, for real-time acquisition of pressure signals in each monitoring area during the foaming molding cycle, and has a built-in signal validity diagnosis unit for synchronous monitoring of the validity of each pressure signal. The stage identification module is connected to the pressure sensing module and is equipped with an edge computing unit for calculating the first derivative of the pressure signal within a sliding time window. Based on the first derivative and the pressure value, the current foaming stage is identified online. The parameter scheduling module is connected to the stage identification module and has a built-in multi-segment control parameter library. The parameter library is associated with a raw material property mapping table with different isocyanate indices and combined polyether viscosities. It is used to retrieve the corresponding target pressure curve and PID control parameter set according to the identified foaming stage. The closed-loop control module, connected to the parameter scheduling module and the pressure sensing module, is configured with a cascaded PID control architecture, including a main controller and a secondary controller. The main controller is used to calculate the deviation between the weighted fusion value of the pressure values ​​of each monitoring area and the target pressure value and output the target pressure change rate. The secondary controller is used to calculate the deviation between the actual pressure change rate and the target pressure change rate and output the control quantity of the air extraction system. The fault-tolerant management module is connected to the closed-loop control module and the pressure sensing module respectively, and is used to perform fault-tolerant operations to adjust parameters when control instability or signal abnormality is detected. The self-learning optimization module is connected to the closed-loop control module and the parameter scheduling module. It is used to update the target pressure curve and PID control parameters in the multi-segment control parameter library according to the molding quality detection results after each foaming molding is completed. An execution module, connected to the closed-loop control module, responds to the control quantity to adjust the pumping rate of the pumping system.

Images