Plastic cup extrusion control method and system based on wall thickness dynamic compensation

Abstract

This invention discloses a method and system for controlling the extrusion of plastic cups based on dynamic wall thickness compensation, relating to the field of plastic cup extrusion technology. The method includes: real-time acquisition of plastic cup extrusion data to obtain real-time extrusion state parameters, synchronizing these parameters with real-time wall thickness data to construct a synchronous data sequence for the extrusion process; performing time-series analysis, dynamically predicting wall thickness deviation based on time-series characteristics, and calculating the predicted wall thickness deviation value; constructing wall thickness constraints combined with the predicted wall thickness deviation value for bidirectional constraint, generating a dynamic compensation signal; feeding the dynamic compensation signal back to the synchronous data sequence for compensation verification, and formulating coordinated adjustment commands based on the verification results to dynamically compensate and control the wall thickness of the extruded plastic cup. This solves the technical problem in the prior art where the plastic cup extrusion molding process is affected by multiple factors, leading to uneven wall thickness in the plastic cup. By compensating and controlling the plastic cup extrusion process, the technical effect of improving the uniformity of the plastic cup wall thickness is achieved.

Description

Technical Field

[0001] This invention relates to the field of plastic cup extrusion technology, and more specifically to a plastic cup extrusion control method and system based on dynamic wall thickness compensation. Background Technology

[0002] Plastic cups are widely used in the food, beverage, and disposable product industries. Their production typically employs extrusion blow molding or injection molding processes, with the extrusion process being a critical step affecting the uniformity of wall thickness and dimensional stability. In existing technologies, the molten plastic is affected by temperature, shear force, and pressure changes within the extrusion screw. Even after the cup preform is formed through the die, wall thickness fluctuations still occur due to differences in traction speed, cooling conditions, and material flowability. Traditional extrusion control relies heavily on fixed parameters or empirical adjustments, lacking real-time prediction and dynamic compensation for wall thickness deviations. This results in uneven wall thickness and inconsistent strength distribution in plastic cups, while also increasing scrap rates and raw material waste, impacting production efficiency and the stability of finished product quality. Summary of the Invention

[0003] This application provides a method and system for controlling the extrusion of plastic cups based on dynamic wall thickness compensation, which solves the technical problem in the prior art where the plastic cup extrusion molding process is affected by a variety of factors, resulting in uneven wall thickness of the plastic cup.

[0004] The first aspect of this application provides a method for controlling the extrusion of plastic cups based on dynamic wall thickness compensation, the method comprising: Real-time extrusion data of plastic cups is collected to obtain real-time extrusion state parameters. Real-time wall thickness data of the plastic cups is introduced, and the real-time extrusion state parameters are synchronized with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process. Time series analysis is performed based on the synchronous data sequence, and dynamic prediction of wall thickness deviation is performed according to the time series characteristics to calculate the predicted wall thickness deviation value. Wall thickness constraints are constructed and combined with the predicted wall thickness deviation value for bidirectional constraint to generate a dynamic compensation signal. The dynamic compensation signal is fed back to the synchronous data sequence for compensation verification. Based on the verification results, a coordinated adjustment command is formulated to dynamically compensate and control the wall thickness of the extruded plastic cups.

[0005] A second aspect of this application provides a plastic cup extrusion control system based on dynamic wall thickness compensation, the system comprising: Data Acquisition Module: Acquires real-time extrusion data of plastic cups, obtains real-time extrusion status parameters, incorporates real-time wall thickness data of the plastic cups, and synchronizes the real-time extrusion status parameters with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process; Deviation Prediction Module: Performs time-series analysis based on the synchronous data sequence of the extrusion process, dynamically predicts wall thickness deviation based on time-series characteristics, and calculates the predicted wall thickness deviation value; Compensation Signal Generation Module: Constructs wall thickness constraints and combines them with the predicted wall thickness deviation value to perform bidirectional constraints, generating a dynamic compensation signal; Compensation Control Module: Feeds back the dynamic compensation signal to the synchronous data sequence of the extrusion process for compensation verification, and formulates coordinated adjustment commands based on the verification results to dynamically compensate and control the wall thickness of the extruded plastic cups.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, real-time extrusion data of plastic cups is acquired to obtain real-time extrusion state parameters. Real-time wall thickness data of the plastic cups is then incorporated, and the real-time extrusion state parameters are synchronized with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process. Next, time-series analysis is performed based on the synchronous data sequence, and dynamic prediction of wall thickness deviation is made according to the time-series characteristics, calculating the predicted wall thickness deviation value. Then, wall thickness constraints are constructed and combined with the predicted wall thickness deviation value for bidirectional constraints, generating a dynamic compensation signal. Finally, the dynamic compensation signal is fed back to the synchronous data sequence for compensation verification. Based on the verification results, coordinated adjustment commands are formulated to dynamically compensate and control the wall thickness of the extruded plastic cups. This solves the technical problem in existing technologies where the extrusion molding process of plastic cups is affected by multiple factors, leading to uneven wall thickness. By compensating and controlling the extrusion process, the technical effect of improving the uniformity of the plastic cup wall thickness is achieved. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic flowchart of a plastic cup extrusion control method based on dynamic wall thickness compensation provided in an embodiment of this application; Figure 2 This is a schematic diagram of the simulated wall thickness recovery curve provided in the embodiments of this application; Figure 3 A schematic diagram of a plastic cup extrusion control system based on dynamic wall thickness compensation provided in an embodiment of this application.

[0009] Figure labeling: Data acquisition module 11, deviation prediction module 12, compensation signal generation module 13, compensation control module 14. Detailed Implementation

[0010] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0011] Example 1, as Figure 1 As shown, this application provides a method for controlling the extrusion of plastic cups based on dynamic wall thickness compensation, wherein the method includes: Real-time extrusion data of plastic cups is collected to obtain real-time extrusion status parameters. Real-time wall thickness data of plastic cups is introduced, and the real-time extrusion status parameters are synchronized and aligned with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process.

[0012] Specifically, temperature sensors collect barrel temperature, die temperature, and melt temperature; pressure sensors collect screw front pressure and die outlet pressure; a speed acquisition module collects screw speed; and a traction encoder collects traction speed. This data is then uploaded in real-time to the controller as plastic cup extrusion data. The controller performs filtering, dimension normalization, and outlier removal according to a uniform sampling period to obtain real-time extrusion status parameters. Simultaneously, a laser thickness gauge is installed at the online detection location after the plastic cup is formed to continuously detect the circumferential and axial wall thickness, obtaining real-time wall thickness data with acquisition time stamps.

[0013] Since the wall thickness detection position is usually located after the die, there is a transmission lag between the real-time wall thickness data and the extrusion state parameters. The controller calculates the backtracking time corresponding to the wall thickness data based on the conveying distance, traction speed, and molding cycle of the plastic cup from the die to the wall thickness detection position, and corrects the timestamp of the real-time wall thickness data to its corresponding extrusion action time. Subsequently, the controller uses the extrusion production cycle or a preset time window as a reference to divide the real-time extrusion state parameters such as barrel temperature, die temperature, melt pressure, screw speed, and traction speed into continuous state data segments, and matches the corrected real-time wall thickness data to the corresponding state data segment according to the timestamp. For cases where there are multiple wall thickness detection values ​​within the same time window, the average value is used as the wall thickness parameter for that time window. For time windows with missing wall thickness data, interpolation is performed to complete the data based on the wall thickness change trend of adjacent time windows. Thus, the real-time extrusion state parameters and corresponding wall thickness parameters within each time window are combined to form a synchronous data unit, and arranged continuously in chronological order to construct a synchronous data sequence for the extrusion process, providing a unified timing basis for subsequent wall thickness deviation prediction and dynamic compensation control.

[0014] Furthermore, real-time acquisition of plastic cup extrusion data to obtain real-time extrusion state parameters, incorporating real-time wall thickness data of the plastic cup, and synchronizing the real-time extrusion state parameters with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process, the method includes: Retrieve extrusion production cycle data, and perform action synchronization analysis on the real-time extrusion status parameters and real-time wall thickness data according to the extrusion production cycle data, and set the reference time of the event trigger frame; segment the real-time extrusion status parameters according to the reference time of the event trigger frame to generate multiple status data segments; extract the time tag of the real-time wall thickness data for correction and construct a retrospective wall thickness time tag; arrange the real-time wall thickness data according to the retrospective wall thickness time tag, and perform time window interpolation alignment based on the real-time wall thickness data and the multiple status data segments to construct the extrusion process synchronization data sequence.

[0015] The system first retrieves the extrusion production cycle data of the plastic cup production line. This data includes the cycle time, conveying distance, and trigger time points for each process, corresponding to the entire process of the plastic cup being extruded from the die, shaped, and traction conveyed to the wall thickness detection station. The controller uses the moment when the plastic cup enters the die area, leaves the die area, or arrives at the detection station as the event trigger event, and defines the corresponding moment as the reference moment of the event trigger frame. By comparing the timing of actions corresponding to the real-time extrusion state parameters and real-time wall thickness data, the system performs action synchronization analysis on the two types of data, establishing the correspondence between extrusion process actions and wall thickness detection actions. Subsequently, the controller segments the real-time extrusion state parameters according to the reference moment of the event trigger frame, grouping parameters such as melt temperature, die temperature, melt pressure, screw speed, and traction speed collected between two adjacent event trigger frames into the same state data segment, thus forming multiple continuous state data segments. Because the wall thickness detection position is located downstream of the die, there is a detection lag in the real-time wall thickness data. Therefore, the controller extracts the acquisition time stamp corresponding to the real-time wall thickness data and calculates the conveying time of the material to reach the detection position by combining the conveying distance between the detection station and the die and the traction speed at the corresponding moment. This conveying time is subtracted from the original acquisition time stamp to obtain a retrospective wall thickness time stamp corresponding to the actual extrusion behavior. Subsequently, the controller re-sorts the real-time wall thickness data according to the retrospective wall thickness time stamp and maps each wall thickness data to the time interval of the corresponding state data segment based on the principle of time proximity. When there are multiple wall thickness data within a certain state data segment, a weighted average method is used to obtain the corresponding wall thickness value; when there is no corresponding wall thickness data within a state data segment, interpolation compensation is performed using wall thickness data within adjacent time windows to match the wall thickness data with the state data segment on a unified time axis. Finally, the time-aligned wall thickness data is merged with the corresponding state data segment and arranged continuously in chronological order to construct a synchronous data sequence for the extrusion process, providing a unified and continuous data foundation for subsequent dynamic prediction of wall thickness deviation.

[0016] Furthermore, the method for constructing the extrusion process synchronization data sequence by combining real-time wall thickness data with the multiple state data segments through time window interpolation and alignment includes: Multiple time windows are set for matching the multiple state data segments, and the multiple time windows correspond to the multiple state data segments; the real-time wall thickness data is queried for missing data according to the multiple time windows, and the target time window with missing data is extracted. The target time window contains a null value identifier; based on the target time window, local missing data calculation is performed according to the null value identifier to generate a wall thickness filling value; the wall thickness filling value is filled into the target time window, the null value identifier is removed, the real-time wall thickness data is updated and aligned with the multiple state data segments to construct the synchronous data sequence of the extrusion process.

[0017] After initial time matching of real-time wall thickness data with multiple status data segments, the controller establishes multiple time windows based on the start and end times of each status data segment. Each time window corresponds to one status data segment, and the length of the time window can be set to one extrusion production cycle or multiple continuous sampling cycles. Subsequently, the controller iterates through the real-time wall thickness data, mapping each wall thickness detection value to the corresponding time window according to the time label, and performs a data integrity check on each time window. When there is no valid wall thickness detection value in a certain time window, the time window is determined as the target time window, and a null value marker is added at the corresponding position; when there is an abnormal wall thickness value in the time window, an anomaly is determined according to the preset wall thickness change threshold, the abnormal data is removed, and a null value marker is also added. For target time windows with null value markers, the controller invokes a local missing value calculation mechanism. Using adjacent valid time windows before and after the target time window as reference intervals, it extracts the wall thickness values ​​and corresponding time positions within these intervals, and calculates the wall thickness change gradient and trend. When valid wall thickness data exists on both sides of the target time window, linear interpolation is used to calculate the wall thickness filling value. When multiple null values ​​appear consecutively in the target time window, trend extrapolation is performed based on the wall thickness change rate of adjacent valid time windows to generate the corresponding wall thickness filling value. When the target time window is located at the sequence boundary, a weighted estimation is performed using the most recent valid wall thickness value combined with the historical average wall thickness value to obtain the wall thickness filling value. After completing the local missing value calculation, the controller writes the generated wall thickness filling value into the corresponding target time window and deletes the null value markers, ensuring that all time windows have corresponding wall thickness data. Subsequently, the updated real-time wall thickness data and the corresponding status data segments are recombined according to a unified time axis. The wall thickness parameters in each time window are associated with and encapsulated with real-time extrusion status parameters such as melt temperature, die temperature, melt pressure, screw speed and traction speed to form a complete data unit. These data units are then arranged continuously in chronological order to finally construct a complete and consistent synchronous data sequence for the extrusion process.

[0018] If effective wall thickness data exists both before and after the target time window, the effective wall thickness values ​​before and after the target time window are extracted. The time interval ratio between the target time window and the effective wall thickness data before the target time window is calculated. Based on the time interval ratio, the effective wall thickness values ​​before and after the target time window are linearly interpolated to obtain the initial wall thickness filling value corresponding to the target time window. Subsequently, the wall thickness change data within the neighborhood of the target time window is extracted, and the initial wall thickness filling value is trend-corrected to generate the target wall thickness filling value.

[0019] When multiple target time windows are consecutively empty, the controller extracts the effective wall thickness values ​​and their time labels from both ends of the consecutive missing intervals, and calculates the wall thickness change rate of the entire missing interval as a trend reference. Based on the relative position ratio of each target time window in the consecutive missing interval, the trend reference is applied to linear extrapolate to calculate the initial wall thickness filling value for each time window. Subsequently, combined with the wall thickness gradient of adjacent effective time windows within the consecutive missing interval, a trend correction is performed on each initial filling value to ensure smooth and continuous wall thickness changes. Finally, the wall thickness filling value corresponding to each target time window is generated and the sequence is updated.

[0020] When the target time window is located at the beginning or end of the time series and there are no valid wall thickness values ​​before or after it, the controller first obtains the most recently available valid wall thickness value in the series and uses the historical average wall thickness value as a reference. The most recently available valid wall thickness value and the historical average wall thickness value are then fused according to a preset weighting ratio to generate the initial wall thickness filling value for the target time window. Subsequently, based on the local variation trend of the historical wall thickness data, the initial filling value is fine-tuned to ensure that the filling value is consistent with the wall thickness variation trend of subsequent sequences. Finally, the corrected wall thickness filling value is filled into the empty positions at the beginning and end of the sequence, achieving continuity and integrity of the sequence boundaries.

[0021] Based on the synchronous data sequence of the extrusion process, time series analysis is performed, and dynamic prediction of wall thickness deviation is made according to the time series characteristics, and the predicted value of wall thickness deviation is calculated.

[0022] Furthermore, based on the synchronous data sequence of the extrusion process, time-series analysis is performed, and dynamic prediction of wall thickness deviation is made according to the time-series characteristics. The predicted wall thickness deviation value is calculated. The method includes: The synchronous data sequence of the extrusion process is decomposed into multiple scale sub-sequence components at various scales. Instantaneous analysis is performed on these multiple scale sub-sequence components to construct morphological feature parameters. Fluctuation analysis is also performed on these multiple scale sub-sequence components to construct statistical feature parameters. The morphological feature parameters and statistical feature parameters are arranged along a time axis to construct multi-scale feature slices. These multi-scale feature slices are stacked in descending order of feature frequency to construct a multi-scale time-series feature map. Operating condition fluctuation analysis is performed based on the multi-scale time-series feature map to identify the target operating condition fluctuation pattern. Dynamic prediction of wall thickness deviation is performed according to the target operating condition fluctuation pattern and the multi-scale time-series feature map, and the predicted wall thickness deviation value is calculated.

[0023] After obtaining the synchronous data sequence of the extrusion process, synchronous data within a continuous time period is first extracted according to a preset analysis cycle. This synchronous data includes melt temperature, die temperature, melt pressure, screw speed, traction speed, and corresponding wall thickness parameters. Subsequently, the synchronous data sequence of the extrusion process is decomposed into multiple scale subsequence components according to different time scales. High-frequency scale subsequences are used to characterize instantaneous disturbances such as sudden changes in die pressure and traction speed fluctuations, while low-frequency scale subsequences are used to characterize slow trends such as melt temperature drift and changes in material flowability. The controller traverses each scale subsequence component, performing instantaneous analysis on each scale subsequence to extract peak values, valley values, peak-valley differences, local slopes, instantaneous rates of change, and the duration of local extrema, constructing morphological characteristic parameters. Simultaneously, fluctuation analysis is performed on each scale subsequence, calculating the mean, variance, standard deviation, fluctuation amplitude, stability of change, and correlation between adjacent time windows, constructing statistical characteristic parameters. Then, the morphological and statistical feature parameters corresponding to the same time position are combined and arranged along the time axis to form multi-scale feature slices. Each multi-scale feature slice is used to characterize the correlation between extrusion state changes and wall thickness changes within the corresponding time window. Further, the controller stacks multiple multi-scale feature slices according to the feature frequencies corresponding to each scale sub-sequence from high to low, placing high-frequency instantaneous disturbance features in the front layer and low-frequency trend change features in the back layer, thereby constructing a multi-scale time-series feature map. Subsequently, the multi-scale time-series feature map is subjected to operating condition fluctuation analysis. The feature distribution in the current map is matched with pre-stored historical operating condition fluctuation patterns to identify the target operating condition fluctuation pattern among pressure fluctuation, temperature drift, traction instability, or combined disturbance types of the current extrusion process. Finally, based on the target operating condition fluctuation pattern, the corresponding wall thickness deviation prediction sub-model is called, and the multi-scale time-series feature map is input into it to calculate the deviation of the actual wall thickness from the target wall thickness within one or more future time windows, obtaining the predicted wall thickness deviation value. This predicted wall thickness deviation value is used for subsequent dynamic compensation signal generation and wall thickness constraint judgment.

[0024] Furthermore, based on the aforementioned multi-scale time-series feature map, operating condition fluctuation analysis is performed to identify target operating condition fluctuation patterns. The method includes: Dimensionality reduction encoding is performed on the multi-scale time-series feature map to generate a multi-scale feature encoding set; feature clustering is performed on the multi-scale feature encoding set to calculate a multi-scale feature representation vector; multiple historical operating condition fluctuation patterns are introduced and similarity calculation is performed with the multi-scale feature representation vector to obtain multiple similarity coefficients; maximum value screening is performed based on the multiple similarity coefficients to determine the target similarity coefficient, and the target similarity coefficient is used as an index to screen the multiple historical operating condition fluctuation patterns to determine the target operating condition fluctuation pattern.

[0025] First, dimensionality reduction encoding is performed on the multi-scale time-series feature map. The morphological and statistical feature parameters corresponding to different frequency layers in the map are uniformly mapped and compressed according to a preset feature dimension to generate a multi-scale feature encoding set. Each encoding unit represents the comprehensive correlation between melt temperature, melt pressure, screw speed, traction speed, and wall thickness changes within a corresponding time period. Next, feature clustering analysis is performed on the multi-scale feature encoding set. Encoding units with similar feature distributions are grouped into the same cluster region. The corresponding multi-scale feature representation vector is calculated based on the center position, distribution density, and dispersion of the encoding units within each cluster region, ensuring that the multi-scale feature representation vector reflects the overall fluctuation characteristics of the current extrusion process. Then, a historical operating condition pattern library is invoked. This library pre-stores multiple historical operating condition fluctuation patterns and their corresponding standard representation vectors. These historical operating condition fluctuation patterns include pressure fluctuation, temperature drift, traction fluctuation, material fluctuation, and composite fluctuation types. The obtained multi-scale feature representation vectors are compared with the standard representation vectors corresponding to each historical operating condition fluctuation pattern to calculate similarity coefficients. These similarity coefficients represent the degree of matching between the current operating condition and the corresponding historical operating condition pattern. After completing all similarity calculations, the controller filters the multiple similarity coefficients for maximum values, determining the target similarity coefficient with the largest value. The historical operating condition fluctuation pattern corresponding to this target similarity coefficient is then used as the target operating condition fluctuation pattern for the current operating condition. When the target similarity coefficient is greater than a preset matching threshold, the corresponding historical operating condition fluctuation pattern is directly determined. When the target similarity coefficient is lower than the preset matching threshold, the current operating condition is marked as a new composite fluctuation pattern and stored in the historical operating condition pattern library for future expansion and updates. The similarity matching threshold can be obtained based on historical production data statistics, i.e., by performing pairwise similarity calculations on the multi-scale feature representation vectors of a large number of historical extrusion operating conditions, and selecting the average similarity or median plus or minus a certain percentage as the threshold range.

[0026] Let the multi-scale feature representation vector of the current extrusion condition be... The standard representation vector of the historical operating condition fluctuation pattern is: Then the similarity between the two vectors Cosine similarity can be used for calculation, and the formula is as follows: ,in: Represents the dot product of vectors; and represents the Euclidean norm of the vectors; the similarity value ranges from [-1, 1], and the closer the value is to 1, the more similar the current working condition is to the historical working condition.

[0027] Furthermore, the method for dynamically predicting wall thickness deviation based on the target operating condition fluctuation pattern and the multi-scale time series feature map, and calculating the predicted wall thickness deviation value, includes: Multiple wall thickness deviation prediction sub-models are constructed, and these sub-models are associated with multiple historical operating condition fluctuation patterns to construct a pattern-model mapping table. Based on the target operating condition fluctuation pattern, the target wall thickness deviation prediction sub-model is determined by searching the pattern-model mapping table. High-frequency instantaneous calculations are performed on the multi-scale time-series feature map according to the multi-scale feature slices to obtain the instantaneous energy fluctuation index. Low-frequency change calculations are performed on the multi-scale time-series feature map according to the multi-scale feature slices to obtain the wall thickness trend change rate. The gain of the target wall thickness deviation prediction sub-model is adjusted based on the instantaneous energy fluctuation index and the wall thickness trend change rate to construct an optimized target wall thickness deviation prediction sub-model. The multi-scale time-series feature map is synchronized to the optimized target wall thickness deviation prediction sub-model for prediction, and the predicted wall thickness deviation value is calculated.

[0028] Preferably, multiple wall thickness deviation prediction sub-models are constructed, each corresponding to a historical operating condition fluctuation pattern. For example, pressure fluctuation type operating conditions correspond to a pressure-sensitive prediction sub-model, temperature drift type operating conditions correspond to a temperature trend prediction sub-model, traction fluctuation type operating conditions correspond to a traction disturbance prediction sub-model, and composite fluctuation type operating conditions correspond to a multi-factor coupled prediction sub-model. The controller associates and stores the model number, applicable operating condition type, input feature range, and output wall thickness deviation type of each wall thickness deviation prediction sub-model with multiple historical operating condition fluctuation patterns, constructing a pattern-model mapping relationship table.

[0029] Specifically, based on the identified target operating condition fluctuation patterns, a search is performed in the pattern-model mapping table to determine the target wall thickness deviation prediction sub-model that matches the current operating condition. Subsequently, the controller performs high-frequency instantaneous calculations based on multi-scale feature slices in the multi-scale time-series feature map, extracting the amplitude changes, rates of change, and durations corresponding to pressure abrupt changes, traction speed fluctuations, and instantaneous wall thickness offsets in the high-frequency layer, and weighting these values ​​to obtain the instantaneous energy fluctuation index, which characterizes the intensity of the impact of short-term disturbances in the current extrusion process on the wall thickness deviation. Simultaneously, low-frequency change calculations are performed based on the low-frequency layer features in the multi-scale time-series feature map, extracting the direction of change, slope of change, and cumulative offset of the wall thickness parameters relative to the target wall thickness within a continuous time window, and calculating the wall thickness trend change rate, which characterizes the trend of wall thickness deviation continuously increasing or decreasing over time. Subsequently, the instantaneous energy fluctuation index and the wall thickness trend change rate are used as gain adjustment factors to dynamically correct the disturbance response weights, trend prediction weights, and output compensation coefficients in the target wall thickness deviation prediction sub-model. When the instantaneous energy fluctuation index is high, the model's response weights to high-frequency disturbance features are increased; when the wall thickness trend change rate is high, the model's prediction weights to low-frequency trend features are increased, thereby constructing the target wall thickness deviation prediction optimization sub-model. Finally, the multi-scale time-series feature map is input into the target wall thickness deviation prediction optimization sub-model to predict the deviation direction and magnitude of the wall thickness relative to the target wall thickness within subsequent time windows, outputting the predicted wall thickness deviation value, providing data basis for the subsequent generation of dynamic compensation signals.

[0030] The construction of the wall thickness deviation prediction sub-model includes: acquiring historical plastic cup extrusion production data, including historical melt temperature, historical die temperature, historical melt pressure, historical screw speed, historical traction speed, historical wall thickness data, and corresponding production cycle data; performing time tag correction and time window alignment on the historical plastic cup extrusion production data according to the aforementioned synchronization alignment method to generate a historical extrusion process synchronization data sequence; extracting historical multi-scale time series feature maps based on the historical extrusion process synchronization data sequence, and dividing the historical multi-scale time series feature maps into pressure fluctuation type sample sets, temperature drift type sample sets, traction fluctuation type sample sets, material fluctuation type sample sets, and composite fluctuation type sample sets according to the historical operating condition fluctuation characteristics; using the multi-scale time series feature maps in each sample set as model input, and using the actual wall thickness after the corresponding time window as the model input. The difference between the current thickness and the target wall thickness is used as the model output label to establish training sample pairs. The corresponding wall thickness deviation prediction sub-models are trained using the training sample pairs for each working condition, enabling each sub-model to learn the mapping relationship between extrusion state changes and wall thickness deviation changes under the corresponding working condition. After training, validation samples are used to verify the prediction error of each sub-model. When the prediction error is less than a preset error threshold, the sub-model is designated as a callable model, and its applicable working condition type, input feature range, prediction time window length, and output wall thickness deviation type are recorded. Finally, each sub-model is bound and stored with its corresponding historical working condition fluctuation pattern to form a pattern-model mapping table, enabling the rapid retrieval of the corresponding target wall thickness deviation prediction sub-model based on the target working condition fluctuation pattern during real-time control.

[0031] A bidirectional constraint is constructed by combining the predicted wall thickness deviation with the wall thickness constraint to generate a dynamic compensation signal.

[0032] After obtaining the predicted wall thickness deviation, the controller first retrieves the design requirements and production parameters of the plastic cup, including the minimum allowable wall thickness, maximum allowable wall thickness, and historical average wall thickness parameters, and sets the target wall thickness accordingly. Then, the difference between the minimum allowable wall thickness and the target wall thickness is used to obtain the lower limit relaxation amount, and the difference between the maximum allowable wall thickness and the target wall thickness is used to obtain the upper limit relaxation amount, constructing an asymmetric interval. The lower limit relaxation amount is used to limit the compensation constraint for thinner wall thickness, and the upper limit relaxation amount is used to limit the compensation constraint for thicker wall thickness. The controller maps the predicted wall thickness deviation to this asymmetric interval, determines whether the wall thickness is too thin or too thick based on the deviation direction, and calculates the corresponding compensation signal amplitude according to the deviation magnitude, achieving bidirectional constraint adjustment. For different deviation magnitudes, the upper and lower limit relaxation amounts are dynamically adjusted to form a real-time updated asymmetric wall thickness constraint boundary, enabling the compensation signal to simultaneously meet the correction requirements for both thinning and thickening. Finally, the controller combines the asymmetric wall thickness constraint boundary with the predicted wall thickness deviation to generate a dynamic compensation signal. The dynamic compensation signal includes an adjustment amount for controlling the die opening and an adjustment amount for adjusting the traction speed, which is used to correct the wall thickness deviation in real time during the extrusion process and realize closed-loop control.

[0033] Furthermore, the process of constructing wall thickness constraints includes the following methods: The minimum allowable wall thickness parameter is determined by retrieving the cup body's compressive strength, and the maximum allowable wall thickness parameter is determined by retrieving the cup body's molding cycle. A target wall thickness parameter is set by introducing historical average wall thickness parameters. The difference between the minimum allowable wall thickness parameter and the target wall thickness is calculated to generate a first difference value, and the difference between the maximum allowable wall thickness parameter and the target wall thickness is calculated to generate a second difference value. The first difference value is used as a lower limit relaxation value, and the second difference value is used as an upper limit relaxation value. An asymmetric interval is constructed based on the lower limit relaxation value and the upper limit relaxation value. The predicted wall thickness deviation value is mapped to the asymmetric interval for dynamic adjustment, and the parameter fluctuation amplitude value is calculated. Based on the parameter fluctuation amplitude value, the lower limit relaxation value and the upper limit relaxation value are dynamically adjusted to update the asymmetric interval and construct an asymmetric wall thickness constraint boundary. The asymmetric wall thickness constraint boundary is added to the wall thickness constraint conditions.

[0034] First, the product design parameters and process constraint parameters of the plastic cup are retrieved. Based on the compressive strength requirements that the cup body must meet during stacking, handling, and transportation, the minimum allowable wall thickness parameter to ensure structural strength is determined. Simultaneously, based on the molding cycle, cooling time, and demolding stability requirements of the plastic cup, the maximum allowable wall thickness parameter is determined without affecting molding efficiency and cooling quality. Then, the historical average wall thickness parameter of plastic cups of the same specification under stable production conditions is introduced and set as the target wall thickness parameter. The difference between the minimum allowable wall thickness parameter and the target wall thickness parameter is calculated to obtain a first difference value, which is used as the lower limit relaxation amount for the downward allowable offset of the target wall thickness. The difference between the maximum allowable wall thickness parameter and the target wall thickness parameter is calculated to obtain a second difference value, which is used as the upper limit relaxation amount for the upward allowable offset of the target wall thickness. Since a thinner plastic cup wall directly affects compressive strength, while a thicker wall affects cooling cycle and material consumption, the lower and upper limit relaxation amounts can be set separately according to different quality constraints, forming an asymmetric interval centered on the target wall thickness parameter but with different allowable offset ranges. Furthermore, the predicted wall thickness deviation is mapped to the asymmetric interval, and the degree of proximity of the predicted wall thickness deviation to the lower and upper limits of relaxation is determined. The parameter fluctuation amplitude is calculated, which characterizes the impact of the current predicted wall thickness deviation on the boundary of the asymmetric interval. When the parameter fluctuation amplitude is large and the wall thickness deviation approaches the lower limit boundary, the lower limit relaxation is tightened and the thinning compensation constraint strength is increased. When the parameter fluctuation amplitude is large and the wall thickness deviation approaches the upper limit boundary, the upper limit relaxation is tightened and the thickening suppression constraint strength is increased. When the parameter fluctuation amplitude is within a stable range, the original relaxation is maintained or slowly restored. Through the above dynamic adjustments, the controller updates the asymmetric interval, constructs the asymmetric wall thickness constraint boundary, and adds this asymmetric wall thickness constraint boundary to the wall thickness constraint conditions, so that the subsequent dynamic compensation signal is simultaneously subject to both minimum wall thickness strength constraints and maximum wall thickness forming efficiency constraints during generation.

[0035] The dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process for compensation verification. Based on the verification results, a coordinated adjustment command is formulated to dynamically compensate and control the wall thickness of the extruded plastic cup.

[0036] The controller feeds back the dynamic compensation signal to the synchronous data sequence of the extrusion process. Based on the melt temperature, die temperature, melt pressure, screw speed, traction speed, and predicted wall thickness deviation within the current time window, it simulates the wall thickness change response after the dynamic compensation signal is applied, obtaining the corresponding simulated wall thickness recovery result. Subsequently, the controller compares the simulated wall thickness recovery result with the pre-constructed wall thickness constraint conditions to determine whether the simulated compensated wall thickness falls within the asymmetric wall thickness constraint boundary. It also calculates the reduction in wall thickness deviation, recovery speed, and overshoot before and after compensation, forming a verification result. When the verification result meets the preset compensation pass condition, the controller parses the dynamic compensation signal into adjustment components corresponding to different execution objects, including a die opening adjustment component for adjusting the melt flow rate and a traction speed adjustment component for adjusting the cup blank stretching degree. The controller generates a coordinated adjustment command based on each adjustment component and sends it to the die opening adjustment actuator and the traction drive mechanism according to a preset control cycle, so that the die opening and traction speed act in concert according to the same compensation target, thereby dynamically compensating and controlling the wall thickness of the extruded plastic cup. When the verification result does not meet the preset compensation pass condition, the controller corrects the amplitude or direction of the dynamic compensation signal according to the deviation between the simulated wall thickness recovery result and the wall thickness constraint condition, and performs the compensation verification again until the verification is passed before generating the coordinated adjustment command. This avoids the unverified compensation signal from directly acting on the production process, which could lead to over-compensation of the wall thickness or control oscillation.

[0037] Furthermore, the dynamic compensation signal is fed back to the extrusion process synchronization data sequence for compensation verification, and a coordinated adjustment command is formulated based on the verification results. The method includes: The dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process to simulate the wall thickness compensation response and construct a simulated wall thickness recovery curve. The simulated wall thickness recovery curve is compared with the asymmetric interval to calculate the compensation effect evaluation index. The dynamic compensation signal is judged to meet the verification pass condition according to the compensation effect evaluation index. When the dynamic compensation signal meets the verification pass condition, the dynamic compensation signal is decoupled and calculated to generate a die opening execution command and a traction speed execution command. The die opening execution command and the traction speed execution command are simulated and executed for coordinated correction to generate a coordinated adjustment command. When the dynamic compensation signal does not meet the verification pass condition, the dynamic compensation signal is iteratively corrected to generate a dynamic compensation update signal for secondary verification until the dynamic compensation signal meets the verification pass condition.

[0038] After generating the dynamic compensation signal, it is not immediately sent to the extrusion actuator. Instead, the dynamic compensation signal is first fed back to the synchronous data sequence of the extrusion process. Using the melt temperature, die temperature, melt pressure, screw speed, traction speed, and predicted wall thickness deviation within the current time window and adjacent historical time windows as simulation inputs, the simulation process of the plastic cup wall thickness recovering over time after the dynamic compensation signal is applied is simulated, constructing a simulated wall thickness recovery curve. Subsequently, the simulated wall thickness recovery curve is compared with the previously constructed asymmetric interval to determine whether the simulated recovered wall thickness falls within the allowable constraint range. The compensation effect evaluation index is calculated, including the wall thickness deviation reduction rate, recovery time, overshoot amplitude, and stabilization time. When the wall thickness deviation reduction rate is greater than a preset reduction rate threshold, the recovery time is less than a preset recovery time threshold, the overshoot amplitude is less than a preset overshoot threshold, and the stabilized wall thickness falls within the asymmetric interval, the dynamic compensation signal is deemed to have met the verification pass conditions. At this point, the dynamic compensation signal is decoupled and calculated. The compensation component used to change the melt output is converted into a die opening execution command, and the compensation component used to change the cup blank stretching degree is converted into a traction speed execution command. Further, the die opening execution command and the traction speed execution command are simulated in a virtual execution environment to determine whether the two types of execution commands satisfy a coordinated relationship in terms of action direction, action amplitude, and response time. If there is an action conflict or asynchronous response, the die opening execution amount and the traction speed execution amount are proportionally corrected so that the wall thickness change direction after their combined action is consistent with the dynamic compensation target. A final coordinated adjustment command is generated and sent to the corresponding actuator for dynamic wall thickness compensation control. When the compensation effect evaluation index does not meet the verification pass conditions, the controller iteratively corrects the compensation direction, compensation amplitude, or compensation duration of the dynamic compensation signal based on the deviation between the simulated wall thickness recovery curve and the asymmetric interval, generating a dynamic compensation update signal. The wall thickness compensation simulation response and secondary verification are performed again until the dynamic compensation signal meets the verification pass conditions, at which point a coordinated adjustment command is generated.

[0039] Furthermore, such as Figure 2 As shown, the dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process to simulate the wall thickness compensation response and construct a simulated wall thickness recovery curve. The method includes: Based on the dynamic compensation signal, the die opening adjustment component and the traction speed adjustment component are extracted; a virtual simulation network is constructed, and the die opening adjustment component is applied to the virtual simulation network as the die opening boundary condition, and the traction speed adjustment component is applied to the virtual simulation network as the traction boundary condition; according to the die opening boundary condition and the traction boundary condition, the virtual simulation network is driven to perform forward extrapolation on the synchronous data sequence of the extrusion process, generating a multi-directional simulated wall thickness parameter set for the plastic cup, which includes a circumferential simulated wall thickness distribution sequence and an axial simulated wall thickness distribution sequence; based on the circumferential simulated wall thickness distribution sequence and the axial simulated wall thickness distribution sequence, wall thickness recovery analysis is performed, and the simulated wall thickness recovery curve is plotted.

[0040] First, the dynamic compensation signal is analyzed, extracting the component used to change the melt flow rate as the die opening adjustment component and the component used to adjust the cup preform stretching speed as the traction speed adjustment component. Then, a virtual simulation network is constructed, based on the extruder fluid dynamics model, die flow field model, and traction mechanism dynamics model, capable of simulating the flow and forming behavior of the plastic melt under different operating conditions. The die opening adjustment component is applied to the die boundary conditions of the virtual simulation network, and the traction speed adjustment component is applied to the traction boundary conditions of the network. Using the current extrusion process synchronous data sequence as the initial state input, the virtual simulation network is driven to perform forward simulation, calculating the wall thickness change of the plastic cup within each time window under dynamic compensation. Through simulation, a multi-directional simulated wall thickness parameter set for the plastic cup is generated, including a circumferential simulated wall thickness distribution sequence to represent the wall thickness uniformity of the plastic cup in the circumferential direction, and an axial simulated wall thickness distribution sequence to represent the wall thickness change of the plastic cup along the height direction. Wall thickness recovery analysis was performed based on circumferential and axial simulated wall thickness distribution sequences. Parameters such as maximum wall thickness deviation, average wall thickness recovery rate, and fluctuation range were extracted, and the analysis results were plotted as simulated wall thickness recovery curves. These simulated wall thickness recovery curves were used to evaluate the effect of dynamic compensation signals on wall thickness uniformity and deviation correction, providing a basis for subsequent compensation verification and coordinated adjustment command generation.

[0041] In summary, the embodiments of this application have at least the following technical effects: First, real-time extrusion data of plastic cups is acquired to obtain real-time extrusion state parameters. Real-time wall thickness data of the plastic cups is then incorporated, and the real-time extrusion state parameters are synchronized with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process. Next, time-series analysis is performed based on the synchronous data sequence, and dynamic prediction of wall thickness deviation is made according to the time-series characteristics, calculating the predicted wall thickness deviation value. Then, wall thickness constraints are constructed and combined with the predicted wall thickness deviation value for bidirectional constraints, generating a dynamic compensation signal. Finally, the dynamic compensation signal is fed back to the synchronous data sequence for compensation verification. Based on the verification results, coordinated adjustment commands are formulated to dynamically compensate and control the wall thickness of the extruded plastic cups. This solves the technical problem in existing technologies where the extrusion molding process of plastic cups is affected by multiple factors, leading to uneven wall thickness. By compensating and controlling the extrusion process, the technical effect of improving the uniformity of the plastic cup wall thickness is achieved.

[0042] Example 2 is based on the same inventive concept as the plastic cup extrusion control method based on dynamic wall thickness compensation in the previous examples, such as... Figure 3 As shown, this application provides a plastic cup extrusion control system based on dynamic wall thickness compensation, wherein the system includes: Data acquisition module 11: Acquires real-time extrusion data of plastic cups, obtains real-time extrusion state parameters, introduces real-time wall thickness data of plastic cups, and synchronizes the real-time extrusion state parameters with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process; Deviation prediction module 12: Performs time-series analysis based on the synchronous data sequence for the extrusion process, dynamically predicts wall thickness deviation based on time-series characteristics, and calculates the predicted wall thickness deviation value; Compensation signal generation module 13: Constructs wall thickness constraint conditions and combines them with the predicted wall thickness deviation value to perform bidirectional constraints, generating a dynamic compensation signal; Compensation control module 14: Feeds back the dynamic compensation signal to the synchronous data sequence for the extrusion process for compensation verification, and formulates coordinated adjustment instructions based on the verification results to dynamically compensate and control the wall thickness of the extruded plastic cup.

[0043] Furthermore, the data acquisition module 11 is used to perform the following methods: Retrieve extrusion production cycle data, and perform action synchronization analysis on the real-time extrusion status parameters and real-time wall thickness data according to the extrusion production cycle data, and set the reference time of the event trigger frame; segment the real-time extrusion status parameters according to the reference time of the event trigger frame to generate multiple status data segments; extract the time tag of the real-time wall thickness data for correction and construct a retrospective wall thickness time tag; arrange the real-time wall thickness data according to the retrospective wall thickness time tag, and perform time window interpolation alignment based on the real-time wall thickness data and the multiple status data segments to construct the extrusion process synchronization data sequence.

[0044] Furthermore, the data acquisition module 11 is used to perform the following methods: Multiple time windows are set for matching the multiple state data segments, and the multiple time windows correspond to the multiple state data segments; the real-time wall thickness data is queried for missing data according to the multiple time windows, and the target time window with missing data is extracted. The target time window contains a null value identifier; based on the target time window, local missing data calculation is performed according to the null value identifier to generate a wall thickness filling value; the wall thickness filling value is filled into the target time window, the null value identifier is removed, the real-time wall thickness data is updated and aligned with the multiple state data segments to construct the synchronous data sequence of the extrusion process.

[0045] Furthermore, the deviation prediction module 12 is used to perform the following method: The synchronous data sequence of the extrusion process is decomposed into multiple scale sub-sequence components at various scales. Instantaneous analysis is performed on these multiple scale sub-sequence components to construct morphological feature parameters. Fluctuation analysis is also performed on these multiple scale sub-sequence components to construct statistical feature parameters. The morphological feature parameters and statistical feature parameters are arranged along a time axis to construct multi-scale feature slices. These multi-scale feature slices are stacked in descending order of feature frequency to construct a multi-scale time-series feature map. Operating condition fluctuation analysis is performed based on the multi-scale time-series feature map to identify the target operating condition fluctuation pattern. Dynamic prediction of wall thickness deviation is performed according to the target operating condition fluctuation pattern and the multi-scale time-series feature map, and the predicted wall thickness deviation value is calculated.

[0046] Furthermore, the deviation prediction module 12 is used to perform the following method: Dimensionality reduction encoding is performed on the multi-scale time-series feature map to generate a multi-scale feature encoding set; feature clustering is performed on the multi-scale feature encoding set to calculate a multi-scale feature representation vector; multiple historical operating condition fluctuation patterns are introduced and similarity calculation is performed with the multi-scale feature representation vector to obtain multiple similarity coefficients; maximum value screening is performed based on the multiple similarity coefficients to determine the target similarity coefficient, and the target similarity coefficient is used as an index to screen the multiple historical operating condition fluctuation patterns to determine the target operating condition fluctuation pattern.

[0047] Furthermore, the deviation prediction module 12 is used to perform the following method: Multiple wall thickness deviation prediction sub-models are constructed, and these sub-models are associated with multiple historical operating condition fluctuation patterns to construct a pattern-model mapping table. Based on the target operating condition fluctuation pattern, the target wall thickness deviation prediction sub-model is determined by searching the pattern-model mapping table. High-frequency instantaneous calculations are performed on the multi-scale time-series feature map according to the multi-scale feature slices to obtain the instantaneous energy fluctuation index. Low-frequency change calculations are performed on the multi-scale time-series feature map according to the multi-scale feature slices to obtain the wall thickness trend change rate. The gain of the target wall thickness deviation prediction sub-model is adjusted based on the instantaneous energy fluctuation index and the wall thickness trend change rate to construct an optimized target wall thickness deviation prediction sub-model. The multi-scale time-series feature map is synchronized to the optimized target wall thickness deviation prediction sub-model for prediction, and the predicted wall thickness deviation value is calculated.

[0048] Furthermore, the compensation signal generation module 13 is used to perform the following method: The minimum allowable wall thickness parameter is determined by retrieving the cup body's compressive strength, and the maximum allowable wall thickness parameter is determined by retrieving the cup body's molding cycle. A target wall thickness parameter is set by introducing historical average wall thickness parameters. The difference between the minimum allowable wall thickness parameter and the target wall thickness is calculated to generate a first difference value, and the difference between the maximum allowable wall thickness parameter and the target wall thickness is calculated to generate a second difference value. The first difference value is used as a lower limit relaxation value, and the second difference value is used as an upper limit relaxation value. An asymmetric interval is constructed based on the lower limit relaxation value and the upper limit relaxation value. The predicted wall thickness deviation value is mapped to the asymmetric interval for dynamic adjustment, and the parameter fluctuation amplitude value is calculated. Based on the parameter fluctuation amplitude value, the lower limit relaxation value and the upper limit relaxation value are dynamically adjusted to update the asymmetric interval and construct an asymmetric wall thickness constraint boundary. The asymmetric wall thickness constraint boundary is added to the wall thickness constraint conditions.

[0049] Furthermore, the compensation control module 14 is used to perform the following method: The dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process to simulate the wall thickness compensation response and construct a simulated wall thickness recovery curve. The simulated wall thickness recovery curve is compared with the asymmetric interval to calculate the compensation effect evaluation index. The dynamic compensation signal is judged to meet the verification pass condition according to the compensation effect evaluation index. When the dynamic compensation signal meets the verification pass condition, the dynamic compensation signal is decoupled and calculated to generate a die opening execution command and a traction speed execution command. The die opening execution command and the traction speed execution command are simulated and executed for coordinated correction to generate a coordinated adjustment command. When the dynamic compensation signal does not meet the verification pass condition, the dynamic compensation signal is iteratively corrected to generate a dynamic compensation update signal for secondary verification until the dynamic compensation signal meets the verification pass condition.

[0050] Furthermore, the compensation control module 14 is used to perform the following method: Based on the dynamic compensation signal, the die opening adjustment component and the traction speed adjustment component are extracted; a virtual simulation network is constructed, and the die opening adjustment component is applied to the virtual simulation network as the die opening boundary condition, and the traction speed adjustment component is applied to the virtual simulation network as the traction boundary condition; according to the die opening boundary condition and the traction boundary condition, the virtual simulation network is driven to perform forward extrapolation on the synchronous data sequence of the extrusion process, generating a multi-directional simulated wall thickness parameter set for the plastic cup, which includes a circumferential simulated wall thickness distribution sequence and an axial simulated wall thickness distribution sequence; based on the circumferential simulated wall thickness distribution sequence and the axial simulated wall thickness distribution sequence, wall thickness recovery analysis is performed, and the simulated wall thickness recovery curve is plotted.

[0051] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for controlling the extrusion of plastic cups based on dynamic wall thickness compensation, characterized in that, The method includes: Real-time extrusion data of plastic cups is collected to obtain real-time extrusion status parameters. Real-time wall thickness data of plastic cups is introduced, and the real-time extrusion status parameters are synchronized and aligned with the real-time wall thickness data to construct a synchronous data sequence of the extrusion process. Based on the synchronous data sequence of the extrusion process, a time series analysis is performed, and the wall thickness deviation is dynamically predicted according to the time series characteristics, and the predicted value of the wall thickness deviation is calculated. A bidirectional constraint is constructed by combining the predicted wall thickness deviation with the wall thickness constraint, and a dynamic compensation signal is generated. The dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process for compensation verification. Based on the verification results, a coordinated adjustment command is formulated to dynamically compensate and control the wall thickness of the extruded plastic cup.

2. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 1, characterized in that, Real-time acquisition of plastic cup extrusion data to obtain real-time extrusion state parameters, incorporating real-time wall thickness data of the plastic cup, and synchronizing the real-time extrusion state parameters with the real-time wall thickness data to construct a synchronous data sequence for the extrusion process, the method includes: Retrieve extrusion production cycle data, perform action synchronization analysis on the real-time extrusion state parameters and the real-time wall thickness data according to the extrusion production cycle data, and set the reference time for the event trigger frame; The real-time extrusion status parameters are segmented according to the reference time of the event trigger frame to generate multiple status data segments; Extract the time stamps from the real-time wall thickness data, correct them, and construct a retrospective wall thickness time stamp; The real-time wall thickness data is arranged according to the backtracking wall thickness time label, and time window interpolation is performed based on the real-time wall thickness data and the multiple state data segments to construct the synchronous data sequence of the extrusion process.

3. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 2, characterized in that, The method involves constructing a synchronous data sequence for the extrusion process by interpolating and aligning time windows based on real-time wall thickness data and multiple state data segments. Multiple time windows are set for matching the multiple state data segments, and the multiple time windows have a corresponding relationship with the multiple state data segments; The real-time wall thickness data is queried for missing data according to the multiple time windows, and the target time window with missing data is extracted. The target time window contains a null value identifier. Based on the target time window, local missing value calculations are performed according to the null value identifier to generate wall thickness filling values; The wall thickness filling value is filled into the target time window, the null value identifier is removed, the real-time wall thickness data is updated and aligned with the multiple state data segments to construct the extrusion process synchronization data sequence.

4. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 1, characterized in that, Based on the synchronous data sequence of the extrusion process, time-series analysis is performed, and dynamic prediction of wall thickness deviation is made according to the time-series characteristics. The predicted wall thickness deviation value is calculated. The method includes: The synchronous data sequence of the extrusion process is decomposed into multiple scales and time series components. Instantaneous analysis is performed on the multiple scale subsequence components to construct morphological feature parameters; Fluctuation analysis is performed on the multiple scale subsequence components to construct statistical characteristic parameters; The morphological feature parameters and the statistical feature parameters are arranged along the time axis to construct multi-scale feature slices; The multi-scale feature slices are stacked in descending order of feature frequency to construct a multi-scale temporal feature map. Based on the multi-scale time-series feature map, operating condition fluctuation analysis is performed to identify target operating condition fluctuation patterns. Based on the target operating condition fluctuation pattern and the multi-scale time series feature map, dynamic prediction of wall thickness deviation is performed, and the predicted value of wall thickness deviation is calculated.

5. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 4, characterized in that, Based on the multi-scale time-series feature map, operating condition fluctuation analysis is performed to identify target operating condition fluctuation patterns. The method includes: Dimensionality reduction encoding is performed based on the multi-scale temporal feature map to generate a multi-scale feature encoding set; Perform feature clustering on the multi-scale feature encoding set and calculate the multi-scale feature representation vector; Multiple historical operating condition fluctuation patterns are introduced and similarity calculations are performed with the multi-scale feature representation vector to obtain multiple similarity coefficients; Based on the multiple similarity coefficients, a maximum value is selected to determine the target similarity coefficient. The target similarity coefficient is then used as an index to select the multiple historical operating condition fluctuation patterns to determine the target operating condition fluctuation pattern.

6. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 5, characterized in that, Dynamic prediction of wall thickness deviation is performed based on the target operating condition fluctuation pattern and the multi-scale time series feature map, and the predicted wall thickness deviation value is calculated. The method includes: Construct multiple sub-models for predicting wall thickness deviation, and associate and map these sub-models with the multiple historical operating condition fluctuation patterns to construct a pattern-model mapping relationship table. Based on the target operating condition fluctuation pattern, the target wall thickness deviation prediction sub-model is determined by searching the pattern-model mapping relationship table. Based on the multi-scale temporal feature map, high-frequency instantaneous calculations are performed according to the multi-scale feature slices to obtain the instantaneous energy fluctuation index; Based on the multi-scale temporal feature map, low-frequency variation calculations are performed according to the multi-scale feature slices to obtain the wall thickness trend change rate. The gain of the target wall thickness deviation prediction sub-model is adjusted based on the instantaneous energy fluctuation index and the wall thickness trend change rate to construct the target wall thickness deviation prediction optimization sub-model; The multi-scale temporal feature map is synchronized to the target wall thickness deviation prediction optimization sub-model for prediction, and the predicted wall thickness deviation value is calculated.

7. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 1, characterized in that, The process of constructing wall thickness constraints includes the following methods: The minimum allowable wall thickness parameter is determined by retrieving the cup body's compressive strength, and the maximum allowable wall thickness parameter is determined by retrieving the cup body's molding cycle. A target wall thickness parameter is set by introducing a historical average wall thickness parameter. The difference between the minimum allowable wall thickness parameter and the target wall thickness is calculated to generate a first difference value. The difference between the maximum allowable wall thickness parameter and the target wall thickness is calculated to generate a second difference value. The first difference is used as the lower limit relaxation amount, the second difference is used as the upper limit relaxation amount, and an asymmetric interval is constructed based on the lower limit relaxation amount and the upper limit relaxation amount; The predicted wall thickness deviation is mapped to the asymmetric interval for dynamic adjustment, and the parameter fluctuation amplitude is calculated. Based on the fluctuation amplitude of the parameters, the lower limit relaxation amount and the upper limit relaxation amount are dynamically adjusted to update the asymmetric interval and construct the asymmetric wall thickness constraint boundary. Add the asymmetric wall thickness constraint boundary to the wall thickness constraint condition.

8. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 7, characterized in that, The dynamic compensation signal is fed back to the extrusion process synchronization data sequence for compensation verification, and a coordinated adjustment command is formulated based on the verification result. The method includes: The dynamic compensation signal is fed back to the synchronous data sequence of the extrusion process to simulate the wall thickness compensation response and construct a simulated wall thickness recovery curve. The simulated wall thickness recovery curve is compared with the asymmetric interval to calculate the compensation effect evaluation index; Determine whether the dynamic compensation signal meets the verification pass conditions according to the compensation effect evaluation index; When the dynamic compensation signal meets the verification pass condition, the dynamic compensation signal is decoupled and calculated to generate the die opening execution command and the traction speed execution command. Simulate the execution of the die opening degree execution command and the traction speed execution command, perform coordinated correction, and generate a coordinated adjustment command; If the dynamic compensation signal does not meet the verification pass condition, the dynamic compensation signal is iteratively corrected to generate a dynamic compensation update signal for secondary verification until the dynamic compensation signal meets the verification pass condition.

9. The plastic cup extrusion control method based on dynamic wall thickness compensation as described in claim 8, characterized in that, The method involves feeding the dynamic compensation signal back to the synchronous data sequence of the extrusion process to simulate the wall thickness compensation response and constructing a simulated wall thickness recovery curve. Based on the dynamic compensation signal, extract the die opening adjustment component and the traction speed adjustment component; A virtual simulation network is constructed, and the die opening adjustment component is applied to the virtual simulation network as a die opening boundary condition, and the traction speed adjustment component is applied to the virtual simulation network as a traction boundary condition. According to the die boundary conditions and the traction boundary conditions, the virtual simulation network is driven to extrapolate the synchronous data sequence of the extrusion process to generate a multi-directional simulated wall thickness parameter set for the plastic cup. The multi-directional simulated wall thickness parameter set for the plastic cup includes a circumferential simulated wall thickness distribution sequence and an axial simulated wall thickness distribution sequence. Based on the circumferential simulated wall thickness distribution sequence and the axial simulated wall thickness distribution sequence, wall thickness recovery analysis is performed, and the simulated wall thickness recovery curve is plotted.

10. A plastic cup extrusion control system based on dynamic wall thickness compensation, characterized in that, For implementing the plastic cup extrusion control method based on dynamic wall thickness compensation according to any one of claims 1-9, the system comprises: Data acquisition module: Real-time acquisition of plastic cup extrusion data, obtaining real-time extrusion status parameters, introducing real-time wall thickness data of the plastic cup, synchronizing and aligning the real-time extrusion status parameters with the real-time wall thickness data, and constructing a synchronous data sequence for the extrusion process; Deviation prediction module: Based on the synchronous data sequence of the extrusion process, it performs time series analysis, dynamically predicts the wall thickness deviation according to the time series characteristics, and calculates the predicted wall thickness deviation value; Compensation signal generation module: Constructs wall thickness constraint conditions and combines them with the predicted wall thickness deviation value to perform bidirectional constraints and generate dynamic compensation signals; The compensation control module feeds back the dynamic compensation signal to the synchronous data sequence of the extrusion process for compensation verification, and formulates coordinated adjustment instructions based on the verification results to dynamically compensate and control the wall thickness of the extruded plastic cup.

Images