Open milling granulation production process control method and system

Abstract

The invention relates to the technical field of open-milling granulation production control, and provides an open-milling granulation production process control method and system, and the method comprises the steps: monitoring and recording the flow state parameters of a melt strand at the outlet of a granulator die head in real time; monitoring and recording the driving motor torque and the driving motor rotating speed of an extrusion screw of the granulator in real time, and judging the fluctuation trend of the melt viscosity according to the driving motor torque and the driving motor rotating speed to obtain melt viscosity fluctuation trend information; based on the melt strand flow state parameters and the melt viscosity fluctuation trend information, executing a judgment process of a starting strategy of a feed-forward compensation mechanism to obtain a feed-forward compensation judgment result; according to the feedforward compensation judgment result, the rotating speed of a rotating cutter head of the granulator or the heating power of a die head is adjusted, so that the open milling granulation production process is controlled. The method has the effect of improving the stability of the open milling granulation production process.

Description

Technical Field

[0001] This invention relates to the technical field of open granulation production control, and specifically to a method and system for controlling the open granulation production process. Background Technology

[0002] In modern industrial production, the granulation process is a crucial step in polymer material processing. Its precise, stable, and efficient control is essential to ensuring final product quality and production efficiency. However, in actual production, some subtle or gradual changes, such as the decrease in heat transfer efficiency caused by scale buildup in the cooling water system, are often difficult to identify and address in a timely manner. Scale formation is a slow process, initially difficult to detect, but its low thermal conductivity significantly increases the thermal resistance between the internal medium and the surface of the rollers, and reduces the cross-sectional area for cooling medium flow, leading to a decrease in roller heat exchange efficiency. At this point, although the surface temperature sensor may show the temperature within the set range, the system's internal energy consumption has quietly increased, and its ability to regulate temperature in response to external disturbances is greatly reduced.

[0003] Faced with this kind of hidden performance degradation, operators, lacking in-depth diagnostic tools, may observe phenomena such as increased roller temperature fluctuations and longer recovery times. To quickly suppress these apparent fluctuations, engineers may manually set the integral time parameter of the roller temperature PID control loop too short. This overly aggressive setting, given that the physical system response is already sluggish due to scale buildup, actually creates a hidden danger for subsequent control instability. When production rates increase and shear heat effects intensify, the control system completely destabilizes when dealing with dynamic heat loads, leading to violent roller temperature fluctuations. These fluctuations severely affect the uniformity of material plasticization, causing drastic and irregular fluctuations in the melt viscosity fed into the granulator.

[0004] The instability of melt viscosity causes the extrusion screw of the pelletizer to be subjected to frequent torque impacts, making it difficult to stabilize the extrusion rate and pressure, resulting in irregular pulsations in the melt flow. Ultimately, the pelletizer is unable to accurately cut the continuously changing melt flow, leading to an abnormally wide range of particle sizes and severely substandard product quality. Simultaneously, the control system repeatedly performs excessive heating and cooling to cope with the instability, significantly increasing energy consumption and drastically reducing production efficiency. Under the interplay of these multiple abnormal factors, traditional open-feed pelletizing process control methods are insufficient to quickly diagnose and effectively resolve this complex problem using conventional means.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] This application discloses a method and system for controlling the open granulation production process, which aims to solve complex problems in the existing open granulation production process, such as instability of the control system due to implicit performance degradation, drastic fluctuations in melt viscosity, substandard product quality, and increased energy consumption.

[0007] The technical solution of this application is as follows: In a first aspect, this application discloses a method for controlling the open-press granulation production process, including: Real-time monitoring and recording of melt flow parameters at the pellet mill die outlet; The torque and speed of the drive motor of the extrusion screw of the granulator are monitored and recorded in real time. Based on the torque and speed of the drive motor, the fluctuation trend of the melt viscosity is determined and the information on the fluctuation trend of the melt viscosity is obtained. Based on the melt flow parameters and melt viscosity fluctuation trend information, the judgment process of the feedforward compensation mechanism activation strategy is executed, and the feedforward compensation judgment result is obtained. Based on the feedforward compensation judgment results, adjust the rotation speed of the pelletizer's rotary cutter head or the heating power of the die head to achieve control of the open granulation production process.

[0008] This technical solution enables real-time monitoring of the melt flow pattern and melt viscosity fluctuation trend, and based on this, a feedforward compensation strategy is executed. This effectively addresses melt viscosity fluctuations caused by hidden factors during the open granulation process, achieving precise control of the pelletizer and die head heating, significantly improving product quality and production stability, and overcoming the limitations of traditional control methods in dealing with complex and hidden problems.

[0009] Furthermore, in some preferred embodiments, the process of determining the initiation strategy of the feedforward compensation mechanism based on the melt flow parameters and melt viscosity fluctuation trend information, and obtaining the feedforward compensation determination result, includes: The flow parameters of the melt stream are corrected in real time to obtain the corrected flow parameters of the melt stream. The performance of the servo motor of the rotary cutter head is evaluated in real time to obtain servo motor performance information; Based on the melt viscosity fluctuation trend information, the corrected melt flow parameters and servo motor performance information are integrated to generate feedforward compensation judgment results. Based on the feedforward compensation judgment result, the pelletizing command is pre-processed in time; Based on the feedforward compensation judgment result, the pelletizing command is subjected to amplitude pre-distortion processing.

[0010] This technical solution enables real-time correction of melt flow parameters and, combined with servo motor performance information, generates more accurate feedforward compensation judgment results. This allows for time pre-processing and amplitude pre-distortion of pelletizing commands, enabling effective intervention before or at the initial stage of melt viscosity fluctuations. This significantly improves pelletizing accuracy and response speed, further optimizing the control effect of the pelletizing process.

[0011] More specifically, in some implementation schemes, the steps of real-time correction of the melt flow parameters to obtain the corrected melt flow parameters include: The first type of non-contact measuring device is pre-deployed to acquire image information of the melt stock, and the image quality characteristics of the first type of non-contact measuring device are acquired based on the image information. Activate the pre-deployed second type of non-contact measuring equipment to obtain the measurement data of the melt strand, and obtain the measurement data of the second type of non-contact measuring equipment; Based on the image quality characteristics of the first type of non-contact measuring equipment, adjust the corresponding image processing parameters; Based on the adjusted image processing parameters, the first type of non-contact measuring device acquires measurement data, and obtains the adjusted measurement data of the first type of non-contact measuring device. Based on the measurement data from the second type of non-contact measuring device, the measurement data from the adjusted first type of non-contact measuring device are corrected to obtain the corrected melt flow parameters.

[0012] This technical solution combines two different types of non-contact measuring devices, dynamically adjusts image processing parameters using the image quality characteristics of the first type of device, and then corrects the measurement data of the first type of device using the measurement data of the second type of device. This effectively improves the accuracy and reliability of melt flow parameter measurement, providing a more solid data foundation for subsequent control decisions.

[0013] Preferably, in some embodiments, the step of real-time correction of the melt flow parameters to obtain the corrected melt flow parameters further includes: When a systematic drift in the measurement characteristics of a Type I or Type II non-contact measuring device is detected, a standard target is introduced for measurement. Based on the deviation between the measured value of the standard target and the actual size of the standard target, a measurement drift compensation function is established; Based on the measurement drift compensation function, the measurement data of the first type of non-contact measuring device are corrected and adjusted to obtain the corrected melt flow parameters. Monitor the long-term trends and cross-correlation of the output data of all measuring devices, and trigger abnormal drift warnings based on the monitoring results.

[0014] This technical solution can effectively correct the systematic drift of the measuring equipment by introducing a standard target and establishing a measurement drift compensation function. At the same time, by monitoring long-term trends and cross-correlation, abnormal drift can be detected and warned in a timely manner, thereby ensuring the long-term accuracy and stability of the measurement data and further improving the robustness of the control system.

[0015] Based on the above, this application further proposes that the steps for monitoring the long-term trend and cross-correlation of the output data of all measuring devices and triggering abnormal drift warnings based on the monitoring results include: During production line downtime or low-load periods, place the standard target in the common measurement area of ​​the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; Within the common measurement area, the measurement values ​​of the standard target by the first type of non-contact measurement device are obtained, and the measurement values ​​of the first type of target are obtained. Within the common measurement area, the measurement values ​​of the standard target by the second type of non-contact measurement device are obtained, and the measurement values ​​of the second type of target are obtained. The measurement deviation of the first type of non-contact measuring device is obtained based on the deviation between the measured value of the first type of target and the actual size of the standard target. The measurement deviation of the second type of non-contact measuring device is obtained based on the deviation between the measured value of the second type of target and the actual size of the standard target. When the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device both exceed the preset threshold, and the deviation direction is consistent and the amplitude deviation is within the preset range, it is determined that common mode drift has occurred. When common-mode drift is detected, an abnormal drift warning is triggered. During normal production operation, continuously monitor the long-term trends of output data from all types of measuring equipment; An abnormal drift warning is triggered when the long-term trend shows that the output data of both the first type of non-contact measuring device and the second type of non-contact measuring device continue to deviate from the historical baseline, and the cross-correlation of the output data of the first type of non-contact measuring device and the second type of non-contact measuring device remains stable.

[0016] This technical solution enables accurate identification and early warning of common-mode drift and abnormal drift by using standard targets for dual-device calibration during downtime or low-load periods, combined with long-term trend and cross-correlation monitoring during normal production. This effectively avoids control errors caused by measurement equipment drift and ensures the continuous stability of the production process.

[0017] In one embodiment, the step of determining common-mode drift as occurring when the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device simultaneously exceed a preset threshold, and the deviation directions are consistent and the amplitude deviations are within a preset range, includes: Obtain production environment parameters; Based on historical calibration data and production environment parameters, adjust the measurement deviation threshold of the first type of non-contact measuring equipment and the measurement deviation threshold of the second type of non-contact measuring equipment. Based on historical calibration data and production environment parameters, adjust the range of similarity between the measurement deviation amplitudes of the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; When the measurement deviation of the first type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the first type of non-contact measuring device, and the measurement deviation of the second type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the second type of non-contact measuring device, and the deviation directions are consistent, and the deviation amplitude is within the range of the adjusted measurement deviation amplitude, it is determined that common mode drift has occurred.

[0018] This technical solution enables the dynamic adjustment of the measurement deviation threshold and amplitude proximity range, making the determination of common mode drift more adaptable to the actual production environment and historical calibration data. This improves the accuracy and sensitivity of common mode drift determination, avoids false alarms or missed alarms, and further enhances the reliability of the system.

[0019] In another implementation, the steps of continuously monitoring the long-term trends of output data from all types of measuring devices during normal production operation include: Obtain the material characteristic parameters of the current material batch to get the feature identifier of the current material batch; Based on the characteristic identifier of the current material batch, a preset material characteristic database is retrieved to obtain the historical measurement data baseline and fluctuation characteristic range that match the current material batch; When there is no matching historical measurement data baseline for the current material batch, after a preset period of stable production, a baseline for the current material batch is established as the historical measurement data baseline. Based on the measurement data of the current batch of materials, the baseline of historical measurement data, and the range of fluctuation characteristics, determine whether there is a trend deviation in the output data of the measuring equipment.

[0020] This technical solution enables personalized monitoring of the long-term trend of measurement equipment output data by combining the characteristics of material batches to establish or retrieve corresponding historical measurement data baselines and fluctuation ranges. It effectively avoids misjudgments caused by differences in material batches, making the judgment of trend deviations more accurate, thereby improving the adaptability and accuracy of the system.

[0021] As an optional solution, when there is no matching historical measurement data baseline for the current material batch, a baseline for the current material batch can be established after a preset period of stable production. The steps for establishing this historical measurement data baseline include: When there is no matching historical measurement data baseline for the current material batch, continuously monitor production process parameters and environmental parameters; When the fluctuations of production process parameters and environmental parameters exceed the preset range, the baseline data acquisition process is paused. Once the fluctuations in production process parameters and environmental parameters have stabilized, the baseline data collection process will continue. Outlier removal is performed on the collected baseline data; Statistical analysis was performed on the baseline data after outlier removal to obtain the statistical analysis results; Based on the statistical analysis results, the average value of the baseline data is calculated as the baseline of the current material batch, and the standard deviation of the baseline data is calculated as the range of fluctuation characteristics.

[0022] This technical solution enables the intelligent control of baseline data collection based on the stability of production process and environmental parameters when there is no matching historical baseline. It also enables the removal of outliers and statistical analysis of the collected data, thereby establishing a more representative and accurate baseline for new material batches. This effectively solves the problem of missing baselines when introducing new materials and ensures the continuous and effective operation of the control system.

[0023] To enhance functionality, the steps for outlier removal from the collected baseline data include: Get the material characteristic parameters of the current material batch; Obtain current production load parameters; Get the current ambient temperature parameters; Adjust the statistical range for outlier judgment based on the material characteristic parameters of the current batch, the current production load parameters, and the current ambient temperature parameters; When the corresponding data point in the collected baseline data exceeds the adjusted statistical range, the corresponding data point is identified as an outlier and removed.

[0024] This technical solution allows for the dynamic adjustment of the statistical range for outlier judgment by combining multiple parameters such as material characteristics, production load, and ambient temperature. This makes the outlier removal process more intelligent and accurate, effectively avoiding false or missed removals caused by a single threshold judgment, and further improving the purity and reliability of baseline data.

[0025] Secondly, this application also discloses an open granulation production process control system for executing open granulation production process control, including: The flow parameter monitoring module is used to monitor and record the flow parameters of the melt stream at the outlet of the pellet mill die in real time. The trend information determination module is used to monitor and record the drive motor torque and drive motor speed of the extrusion screw of the granulator in real time, and determine the fluctuation trend of melt viscosity based on the drive motor torque and drive motor speed to obtain melt viscosity fluctuation trend information. The compensation result judgment module is used to judge the start-up strategy of the feedforward compensation mechanism based on the melt flow parameters and melt viscosity fluctuation trend information, and to obtain the feedforward compensation judgment result. The compensation control execution module is used to adjust the rotation speed of the pelletizer's rotary cutter head or the heating power of the die head based on the feedforward compensation judgment result, so as to realize the control of the open granulation production process.

[0026] This technical solution provides a system that integrates flow parameter monitoring, trend information determination, compensation result judgment, and compensation control execution, enabling comprehensive, real-time, and intelligent control of the open granulation production process. It effectively solves the lag and limitations of traditional control systems in dealing with complex production problems, and significantly improves production efficiency and product quality.

[0027] Beneficial Effects: The open-press granulation production process control method disclosed in this application can accurately determine the fluctuation trend of melt viscosity by real-time monitoring of the melt flow parameters at the granulator die outlet and the torque and speed of the extrusion screw drive motor. Based on this, the method executes a feedforward compensation mechanism to determine the start-up strategy based on the melt flow parameters and melt viscosity fluctuation trend information, and adjusts the rotary cutter speed or die heating power of the pelletizer in a timely manner according to the determination result. This series of measures effectively solves the problems in the prior art caused by hidden factors such as scale accumulation in the cooling water system, resulting in reduced roller heat exchange efficiency, drastic fluctuations in melt viscosity, decreased pelletizing accuracy, and increased energy consumption. Through the feedforward compensation mechanism, this application can intervene before melt viscosity fluctuations significantly affect product quality, avoiding control instability and severe oscillations that may occur when the system response is sluggish in traditional PID control. Therefore, this application can significantly improve the stability of the open-press granulation process, the uniformity of product quality, and reduce production energy consumption, achieving effective control of complex and hidden production problems. Attached Figure Description

[0028] Figure 1 This is a flowchart of a method for controlling the open granulation production process in one embodiment of the present invention; Figure 2 This is one of the process flow diagrams of a method for controlling the open granulation production process in another embodiment of the present invention; Figure 3 This is a system block diagram of an open granulation production process control system according to another embodiment of the present invention; Explanation of reference numerals in the attached figures: 1. Production process control system for granulation; 11. Flow parameter monitoring module; 12. Trend information judgment module; 13. Compensation result judgment module; 14. Compensation control execution module. Detailed Implementation

[0029] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0030] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0031] Traditional open-feed granulation process control methods often struggle to identify and address latent performance degradation, such as scale buildup in the cooling water system. This latent degradation leads to reduced roller heat exchange efficiency, increased system energy consumption, and instability in the control system under dynamic heat loads, resulting in drastic roller temperature fluctuations. Consequently, the melt viscosity exhibits severe and irregular fluctuations, causing frequent torque impacts on the granulator extrusion screw, making it difficult to stabilize extrusion volume and pressure. The melt flow displays irregular pulsations, ultimately preventing the pelletizer from accurately cutting, resulting in severely substandard product quality, significantly increased energy consumption, and a substantial decrease in production efficiency. Given these multiple intertwined anomalies, traditional control methods are insufficient for quickly diagnosing and effectively resolving this complex problem using conventional means.

[0032] In response, this application proposes a method for controlling the open-press granulation production process, combined with... Figure 1 As shown, it includes: S1, monitor and record the flow parameters of the melt stream at the die outlet of the granulator in real time; S2, monitor and record the drive motor torque and drive motor speed of the extrusion screw of the granulator in real time, and determine the fluctuation trend of melt viscosity based on the drive motor torque and drive motor speed to obtain melt viscosity fluctuation trend information; S3, based on the melt flow parameters and melt viscosity fluctuation trend information, executes the judgment process of the feedforward compensation mechanism activation strategy and obtains the feedforward compensation judgment result; S4, based on the feedforward compensation judgment result, adjust the rotation speed of the pelletizer's rotary cutter or the heating power of the die head to achieve control of the open granulation production process.

[0033] To better understand the open-process granulation production control method proposed in this application, the key terms and implementation environment involved will be explained in detail below.

[0034] "Mel flow parameters" refer to the physical state and flow characteristics of the melt at the die outlet of the granulator, such as the diameter, velocity, temperature, and shape stability of the melt. These parameters directly reflect the uniformity and stability of the melt during the extrusion process and are important indicators for evaluating granulation quality.

[0035] "Mel viscosity fluctuation trend information" refers to the trend of melt viscosity change over time, inferred by monitoring the torque and speed of the extrusion screw in a granulator. Melt viscosity is an important rheological parameter of polymer materials, and its fluctuations directly affect the stability of the extrusion process and product quality. Changes in drive motor torque and speed are closely related to the shear and flow resistance of the melt in the screw, and therefore can serve as an indirect indicator reflecting changes in melt viscosity.

[0036] The "feedforward compensation mechanism" is a control strategy whose core idea is to predict or detect the impact of a disturbance on the system output before or immediately after it occurs, and then take compensatory measures in advance to reduce or eliminate the impact of the disturbance on the system output. In this application, the feedforward compensation mechanism aims to adjust the rotary cutter speed or die heating power of the pelletizer in advance based on the melt flow parameters and melt viscosity fluctuation trend information to cope with the fluctuation of melt viscosity and ensure the stability of the pelletizing process.

[0037] "The rotational speed of the pelletizer's cutter head" refers to the rotational speed of the cutter head used to cut the molten feedstock in the pelletizer. Adjusting this speed directly affects the pelletizing frequency and particle size.

[0038] "Die head heating power" refers to the power of the heater in the die head section of the pellet mill. Adjusting the die head heating power can change the temperature in the die head area, thereby affecting the viscosity and flowability of the melt.

[0039] The implementation environment of this application is typically an open-milling and granulation production line for polymer materials, which includes core equipment such as open mills, granulators, and pelletizers. The entire production process requires precise control of parameters such as temperature, pressure, and speed to ensure product quality.

[0040] The core of the open-press granulation production process control method proposed in this application lies in achieving intelligent feedforward compensation control of the granulation process through real-time monitoring and analysis of key parameters. The implementation methods of each main feature will be described in detail below.

[0041] First, regarding "real-time monitoring and recording of the melt flow parameters at the granulator die outlet".

[0042] This feature can be implemented in the following ways: One method involves visually observing and recording the flow pattern of the molten material. Operators can periodically check the diameter, fracture status, and surface finish of the material at the die exit and manually record these observations. This method is simple and easy to implement, but it is limited by the subjectivity and real-time nature of human judgment.

[0043] Another approach is to install contact sensors, such as thermocouples or pressure sensors, at the pellet mill die outlet to measure the surface temperature or extrusion pressure of the melt stock. These sensors will directly contact the melt stock and convert the measured analog signals into digital signals for recording. This method can provide continuous measurement data, but contact measurements may introduce slight disturbances to the flow pattern of the melt stock, and the sensors are susceptible to wear and contamination.

[0044] Secondly, regarding "real-time monitoring and recording of the drive motor torque and drive motor speed of the granulator extrusion screw, and determining the melt viscosity fluctuation trend based on the drive motor torque and drive motor speed, thus obtaining melt viscosity fluctuation trend information."

[0045] This feature can be implemented in the following ways: One approach is to acquire the torque and speed of the drive motor in real time by installing torque and speed sensors on the motor. These sensors output electrical signals, which are converted into digital signals and recorded by a data acquisition module. Subsequently, the torque and speed data can be mapped to melt viscosity values ​​using empirical formulas or pre-defined lookup tables, and their changes over time can be analyzed to obtain information on the fluctuation trend of melt viscosity. For example, a sustained increase in torque while maintaining a constant speed may indicate an increase in melt viscosity.

[0046] Another approach is to indirectly estimate the drive motor torque and speed using the motor's own current and voltage signals. Modern frequency converters typically provide real-time current and voltage data for the motor, allowing for the estimation of torque and speed using a motor model. These estimates can then be used to determine the fluctuation trend of the melt viscosity. This method eliminates the need for additional sensors, but the estimation accuracy may be slightly lower than direct measurement.

[0047] Secondly, regarding "the judgment process of the feedforward compensation mechanism activation strategy based on melt flow parameters and melt viscosity fluctuation trend information, and the result of the feedforward compensation judgment."

[0048] This feature can be implemented in the following ways: One approach is to set fixed thresholds. For example, when the melt flow parameters (such as the melt diameter) exceed preset upper and lower limits, or when the melt viscosity fluctuation trend information shows that the viscosity change rate exceeds a certain fixed threshold, the system determines that a feedforward compensation mechanism needs to be activated and outputs the corresponding compensation judgment result. This approach is simple, intuitive, and easy to implement, but it may not be able to adapt to complex changes under different materials and production conditions.

[0049] Another approach is to use a rule-based expert system. A set of rules is predefined, such as "if the melt diameter continues to decrease and the melt viscosity shows an upward trend, then initiate feedforward compensation." These rules can be set by experienced engineers based on historical data and production experience. When the monitored parameters meet specific rules, the system generates a feedforward compensation judgment result.

[0050] Finally, regarding "adjusting the rotation speed of the pelletizer's cutter head or the heating power of the die head based on the feedforward compensation judgment results to achieve process control of open granulation production."

[0051] This feature can be implemented in the following ways: One approach is to manually adjust the rotary cutter speed or die heating power of the pelletizer based on the feedforward compensation judgment result. After receiving the compensation judgment result, the operator judges the required adjustment range based on experience and manually adjusts it via the control panel or knob. This method relies on the operator's experience and reaction speed, and may be subject to lag and inaccuracy.

[0052] Another approach is to preset fixed adjustment values. For example, when the feedforward compensation indicates that the pelletizing frequency needs to be increased, the system automatically increases the rotary cutter head speed by a preset fixed value (e.g., 50 RPM). Or, when the judgment indicates that the melt viscosity needs to be reduced, the system automatically increases the die head heating power by a preset fixed value (e.g., 100W). This method can achieve automated adjustment, but the precision and adaptability of the adjustment are limited.

[0053] Optional, combined Figure 2 As shown, S3, based on the melt flow parameters and melt viscosity fluctuation trend information, executes the judgment process of the feedforward compensation mechanism activation strategy, and the steps to obtain the feedforward compensation judgment result include: S31, Real-time correction of melt flow parameters to obtain corrected melt flow parameters; S32 evaluates the performance of the servo motor of the rotary cutter head in real time and obtains servo motor performance information; S33, based on melt viscosity fluctuation trend information, integrates the corrected melt flow parameters and servo motor performance information to generate feedforward compensation judgment results; S34, based on the feedforward compensation judgment result, perform time pre-processing on the pelletizing command; S35 performs amplitude pre-distortion processing on the pelletizing command based on the feedforward compensation judgment result.

[0054] Specifically, real-time correction of melt flow parameters refers to further processing and optimizing the melt flow parameters obtained from the pelletizer die outlet to eliminate the impact of factors such as measurement noise, sensor drift, or environmental interference on data accuracy. The aim is to ensure higher reliability and accuracy of the flow parameters input to the feedforward compensation mechanism. For example, multi-sensor fusion technology can be used to cross-validate and calibrate measurement data from different types of sensors; or adaptive filtering algorithms can be used to smooth and correct the flow parameters based on historical data and real-time fluctuations.

[0055] Real-time evaluation of the servo motor performance of the rotary cutter head refers to continuously monitoring the working status and response characteristics of the servo motor of the pelletizer's rotary cutter head. Servo motor performance information may include, but is not limited to, motor speed stability, torque output accuracy, response time, positioning error, and the presence of wear or aging signs. The purpose is to dynamically understand the actual working capacity and potential limitations of the actuator, so that the dynamic characteristics of the actuator can be fully considered when generating feedforward compensation judgment results, thereby improving the effectiveness of compensation. For example, by analyzing servo motor current, voltage, encoder feedback, and other data, combined with a preset performance model, its various performance indicators can be calculated and evaluated in real time.

[0056] In practical applications, the generation of feedforward compensation judgment results based on melt viscosity fluctuation trend information, combined with corrected melt flow parameters and servo motor performance information, involves using the corrected melt flow parameters, melt viscosity fluctuation trend information, and servo motor performance information as multi-dimensional inputs. These are then comprehensively analyzed and processed by an intelligent decision-making model to arrive at a more comprehensive, accurate, and forward-looking feedforward compensation judgment result. The aim is to leverage the complementarity between multiple information sources to overcome the limitations of a single information source and generate more robust control decisions. For example, machine learning algorithms (such as neural networks and support vector machines) or fuzzy logic systems can be used to perform pattern recognition and correlation analysis on these inputs to predict potential pelletizing problems in the future and provide corresponding compensation suggestions.

[0057] Furthermore, based on the feedforward compensation judgment result, the pelletizing command is pre-processed in time. This means that before the actual pelletizing action needs to be performed, the pelletizing command is issued a certain time in advance, according to the feedforward compensation judgment result. The purpose is to compensate for the inherent response delay of the entire control system (including sensors, controllers, actuators, etc.), ensuring that the pelletizing action can respond more promptly to changes in the melt state, avoiding inaccurate or uneven pelletizing due to delays. For example, if the system detects that the melt viscosity is about to change, and this change requires the pelletizer to respond within 0.2 seconds, but the total system delay is 0.1 seconds, then the pelletizing command will be issued 0.1 seconds in advance.

[0058] Furthermore, based on the feedforward compensation judgment results, amplitude pre-distortion processing is applied to the pelletizing command. This refers to pre-correcting or adjusting the amplitude of the pelletizing command (e.g., the rotational speed of the cutting disc or the adjustment amount of the die heating power) according to the feedforward compensation judgment results. Its purpose is to compensate for the nonlinear response or inherent deviation of the pelletizer actuator under different operating conditions, ensuring that the actual pelletizing action can accurately achieve the expected effect. For example, if the servo motor exhibits a nonlinear response when running at low or high speeds, pre-distortion processing can reversely correct the command amplitude, allowing the actual output speed or power to more accurately match the target value.

[0059] In some preferred embodiments, it is assumed that during the open-feed granulation process, slight drift occurs in the measured data of the melt flow parameters due to batch variations in materials or fluctuations in ambient temperature. Simultaneously, the rotary cutter head servo motor of the pelletizer experiences a slight response delay due to prolonged operation. In this case, the control method of this application first corrects the melt flow parameters in real time, for example, by using multi-sensor fusion or adaptive filtering algorithms to eliminate measurement noise and drift, obtaining more accurate flow parameters. Simultaneously, the system continuously monitors data such as the servo motor's current, voltage, and speed feedback, evaluating its performance indicators such as response time and positioning accuracy, and using this as servo motor performance information. Subsequently, these corrected flow parameters, servo motor performance information, and melt viscosity fluctuation trend information are input into a fusion model, such as a decision model based on neural networks or fuzzy logic, to generate a comprehensive feedforward compensation judgment result. For example, if the judgment result indicates that the melt viscosity is about to increase and the servo motor response is slightly delayed, the system will issue a pelletizing command 0.1 seconds in advance based on the judgment result (time pre-processing) and fine-tune the amplitude of the command to compensate for the nonlinear response of the servo motor (amplitude pre-distortion processing), ensuring that when the melt viscosity actually increases, the pelletizing disc can pelletize at the optimal speed and force, thereby maintaining the uniformity and quality of the pellets.

[0060] Optionally, the steps for real-time correction of the melt flow parameters to obtain the corrected melt flow parameters include: The first type of non-contact measuring device is pre-deployed to acquire image information of the melt stock, and the image quality characteristics of the first type of non-contact measuring device are acquired based on the image information. Activate the pre-deployed second type of non-contact measuring equipment to obtain the measurement data of the melt strand, and obtain the measurement data of the second type of non-contact measuring equipment; Based on the image quality characteristics of the first type of non-contact measuring equipment, adjust the corresponding image processing parameters; Based on the adjusted image processing parameters, the first type of non-contact measuring device acquires measurement data, and obtains the adjusted measurement data of the first type of non-contact measuring device. Based on the measurement data from the second type of non-contact measuring device, the measurement data from the adjusted first type of non-contact measuring device are corrected to obtain the corrected melt flow parameters.

[0061] Specifically, the first type of non-contact measuring equipment can be understood as a vision-based measuring device, such as a high-speed industrial camera or a line scan camera, which is mainly used to capture two-dimensional image information of molten stock. The image information can include visual features such as the outline, surface texture, and fracture characteristics of the molten stock. Image quality features refer to indicators extracted from the acquired image information to evaluate image sharpness, contrast, noise level, etc., such as image sharpness, signal-to-noise ratio, and blurriness, with the aim of assessing the reliability of the image data.

[0062] The second type of non-contact measuring equipment can be understood as a measuring device based on non-visual principles, such as laser diameter gauges, ultrasonic sensors, or infrared sensors. It is mainly used to obtain precise physical dimensions or state data of the molten material, such as diameter, shape, and temperature. The measurement data are quantitative indicators that directly reflect the flow state of the molten material.

[0063] In practical applications, adjusting the corresponding image processing parameters based on the image quality characteristics of the first type of non-contact measuring equipment means that when the image quality characteristics show problems such as blurriness, excessive noise, or insufficient contrast, the system will automatically adjust the parameters of the image acquisition or processing algorithm, such as adjusting the exposure time, gain, filtering intensity, or edge detection threshold. The purpose is to optimize image quality and ensure that the features extracted from the image are more accurate and reliable.

[0064] Furthermore, based on the adjusted image processing parameters, the first type of non-contact measuring device acquires measurement data. The adjusted measurement data of the first type of non-contact measuring device refers to processing or re-acquiring the image after optimizing the image processing parameters to obtain more accurate image-based measurement data of the melt flow state.

[0065] The correction of the adjusted measurement data of the first type of non-contact measuring device based on the measurement data of the second type of non-contact measuring device refers to the fusion and calibration of measurement data from two different principles. For example, the precise physical dimension data provided by the second type of non-contact measuring device can be used as a benchmark to calibrate or compensate the dimension data obtained by the first type of non-contact measuring device based on image analysis, thereby eliminating possible systematic or random errors in the first type of non-contact measuring device. The aim is to improve the overall accuracy and reliability of the final melt flow parameters through multi-source data fusion.

[0066] In some preferred embodiments, a specific example is given below. Assume that on an open-feed pelletizing production line, a first type of non-contact measuring device is configured as a high-speed industrial camera to capture real-time images of the melt stream exiting the pelletizer die. A second type of non-contact measuring device is configured as a laser diameter gauge to accurately measure the real-time diameter of the melt stream.

[0067] When a high-speed industrial camera acquires image information of the molten stock, the image processing module analyzes image quality characteristics such as sharpness and contrast. If blurring or uneven lighting is detected, the system automatically adjusts the camera's exposure time, gain, or image filtering parameters based on the image quality characteristics to optimize image quality. Subsequently, using the adjusted image processing parameters, the contour information of the molten stock is extracted from the optimized image, and the stock diameter and shape parameters based on the image are calculated to obtain the adjusted measurement data for the first type of non-contact measuring device.

[0068] Simultaneously, the laser diameter gauge independently measures the precise diameter of the melt stream in real time, obtaining measurement data from a second type of non-contact measuring device. To improve overall measurement accuracy, the system uses the precise diameter measured by the laser diameter gauge as a reference to correct the diameter calculated from the high-speed industrial camera based on the image. For example, a calibration model can be established to compensate in real time for the deviation between the measurement results of the two devices, or a weighted average or other data fusion algorithm can be used to fuse the two measurement data, where the laser diameter gauge data may be given a higher weight. In this way, the final corrected melt stream flow parameters (such as diameter, shape stability, etc.) will be more accurate and reliable, providing high-quality input for the subsequent feedforward compensation mechanism, thereby ensuring precise control of the pelletizing process.

[0069] Optionally, the steps for real-time correction of the melt flow parameters to obtain the corrected melt flow parameters include: When a systematic drift in the measurement characteristics of a Type I or Type II non-contact measuring device is detected, a standard target is introduced for measurement. Based on the deviation between the measured value of the standard target and the actual size of the standard target, a measurement drift compensation function is established; Based on the measurement drift compensation function, the measurement data of the first type of non-contact measuring device are corrected and adjusted to obtain the corrected melt flow parameters. Monitor the long-term trends and cross-correlation of the output data of all measuring devices, and trigger abnormal drift warnings based on the monitoring results.

[0070] Specifically, "systematic drift" refers to a persistent, directional deviation of the output data of a measuring device from the true value or calibration baseline during long-term operation or under specific environmental conditions. This deviation is not random noise, but is caused by factors such as aging of internal components, sensor performance degradation, and changes in ambient temperature or humidity. When such systematic drift is detected, a "standard target" is introduced for measurement to restore measurement accuracy. A "standard target" is a reference object with known and highly accurate physical dimensions or characteristics, such as a metal rod of known diameter or a plate with specific optical reflectivity. It is placed within the effective measurement area of ​​the measuring device so that the device can measure it.

[0071] The "deviation between the measured value of the standard target and the true size of the standard target" refers to the difference between the size data of the standard target obtained by the measuring equipment and the predetermined, precise true size of the standard target. Based on this deviation, a "measurement drift compensation function" can be established. The "measurement drift compensation function" can be a mathematical model (e.g., a linear function, polynomial function, or nonlinear function), a lookup table, or a machine learning-based algorithm. Its function is to map the detected measurement deviation into a correction amount for subsequent measurement data. This function can quantify and predict the drift behavior of the measuring equipment, thus providing a basis for data correction.

[0072] In practical applications, once the measurement drift compensation function is established, the measurement data of the first type of non-contact measuring equipment can be corrected and adjusted according to the measurement drift compensation function. This means that during normal production, the original measurement data acquired by the first type of non-contact measuring equipment, after preliminary adjustment (such as image processing parameter adjustment), will be further corrected using this compensation function to eliminate or reduce errors caused by systematic equipment drift, thereby obtaining the corrected melt flow parameters.

[0073] Furthermore, to achieve continuous monitoring and early warning of the status of measuring equipment, this application also proposes "monitoring the long-term trend and cross-correlation of the output data of all measuring equipment." "Long-term trend" refers to the direction and magnitude of change in the output data of the measuring equipment over a longer time scale, such as whether the data is consistently high or low. "Cross-correlation" refers to whether there is a synchronous change or mutual influence relationship between the output data of different measuring devices (e.g., first-type non-contact measuring devices and second-type non-contact measuring devices). Through continuous monitoring of these indicators, potential equipment anomalies or drift signs can be detected in a timely manner, and "anomaly drift early warning can be triggered based on the monitoring results," so that operators or automated systems can intervene promptly for calibration, maintenance, or troubleshooting.

[0074] In some preferred embodiments, a specific example is given below. Suppose that on an open granulation production line, a first type of non-contact measuring device (e.g., a machine vision-based image measurement system) and a second type of non-contact measuring device (e.g., a laser diameter gauge) are used to measure the diameter of the melt stream in real time. After a period of continuous operation, due to fluctuations in ambient temperature or sensor aging, it is found that the output data of both types of measuring devices begin to show a persistent, unidirectional deviation.

[0075] Specifically, when the system detects this "systematic drift," for example, through historical data analysis revealing that the measured value is consistently higher or lower than the expected baseline, or through periodic comparisons with known reference values, the production line will place a "standard target" with a known diameter of 5.00 mm within the measurement area during a brief downtime or low-load period. A first-type non-contact measuring device and a second-type non-contact measuring device will then measure the standard target. Assume the first-type non-contact measuring device measures a diameter of 5.05 mm, and the second-type non-contact measuring device measures a diameter of 5.04 mm.

[0076] Based on the deviation between these measurements and the true size of the standard target (e.g., a deviation of +0.05 mm for Type I equipment), a measurement drift compensation function can be established. For example, a simple linear compensation function can be established: Correction = Original Measurement - Deviation. Alternatively, if the drift is non-linear, a polynomial function or lookup table can be established. Once the compensation function is established, in subsequent production processes, the melt stream diameter data acquired by the Type I non-contact measuring equipment, after image processing parameter adjustments, will be immediately corrected using this compensation function. For example, if the original measurement is 5.20 mm, after applying a deviation compensation of +0.05 mm, the corrected diameter will be 5.15 mm, resulting in more accurate corrected melt stream flow parameters.

[0077] Simultaneously, the system continuously monitors the long-term trends and cross-correlation of output data from all measuring devices. For example, by plotting the trend of daily average measurements, it can be observed whether the measured values ​​consistently deviate from the historical baseline. If the measured values ​​of both the first and second types of non-contact measuring devices remain consistently high, and the correlation between them (e.g., the difference between their measured values ​​remains within a stable range) remains high, this indicates that common-mode drift may have occurred. When this trend persists for a period of time and exceeds a preset threshold, the system immediately triggers an abnormal drift warning, for example, by displaying a warning message on the control interface and notifying the operator via audible and visual alarms, prompting the need to inspect or recalibrate the measuring devices, thereby ensuring the accuracy and stability of production process control.

[0078] Optionally, the steps for monitoring the long-term trend and cross-correlation of the output data of all measuring devices and triggering abnormal drift warnings based on the monitoring results include: During production line downtime or low-load periods, place the standard target in the common measurement area of ​​the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; Within the common measurement area, the measurement values ​​of the standard target by the first type of non-contact measurement device are obtained, and the measurement values ​​of the first type of target are obtained. Within the common measurement area, the measurement values ​​of the standard target by the second type of non-contact measurement device are obtained, and the measurement values ​​of the second type of target are obtained. The measurement deviation of the first type of non-contact measuring device is obtained based on the deviation between the measured value of the first type of target and the actual size of the standard target. The measurement deviation of the second type of non-contact measuring device is obtained based on the deviation between the measured value of the second type of target and the actual size of the standard target. When the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device both exceed the preset threshold, and the deviation direction is consistent and the amplitude deviation is within the preset range, it is determined that common mode drift has occurred. When common-mode drift is detected, an abnormal drift warning is triggered. During normal production operation, continuously monitor the long-term trends of output data from all types of measuring equipment; An abnormal drift warning is triggered when the long-term trend shows that the output data of both the first type of non-contact measuring device and the second type of non-contact measuring device continue to deviate from the historical baseline, and the cross-correlation of the output data of the first type of non-contact measuring device and the second type of non-contact measuring device remains stable.

[0079] Specifically, in order to more accurately identify abnormal drift of measuring equipment, this application refines the monitoring process into two main scenarios: common mode drift detection during downtime or low-load periods and long-term trend deviation detection during normal production operation.

[0080] The phrase "placing a standard target in the common measurement area of ​​both the first and second types of non-contact measuring equipment during production line downtime or low-load periods" refers to placing a standard target of known size and characteristics in a location simultaneously measured by both non-contact measuring devices (e.g., laser diameter gauges and vision inspection systems) manually or using an automated robotic arm when the production line is not running or under low load, and the material flow is stable or no material is passing through. The purpose of this is to obtain measurement data from both devices on the same standard reference object in a controlled environment for accurate calibration and deviation analysis.

[0081] "Within a common measurement area, the measurement value of the standard target by a first type of non-contact measuring device is obtained, thus obtaining the first type of target measurement value; within the common measurement area, the measurement value of the standard target by a second type of non-contact measuring device is obtained, thus obtaining the second type of target measurement value" means that after the standard target is placed in place, two non-contact measuring devices independently measure the target and record their respective measurement results. For example, the first type of non-contact measuring device might be a high-precision laser diameter gauge, whose measurement value is the diameter of the target; the second type of non-contact measuring device might be an image processing-based vision system, whose measurement value is the actual size of the target's pixel dimensions in the image after calibration.

[0082] "The measurement deviation of the first type of non-contact measuring device is obtained based on the deviation between the measured value of the first type of target and the actual size of the standard target; the measurement deviation of the second type of non-contact measuring device is obtained based on the deviation between the measured value of the second type of target and the actual size of the standard target." This means comparing the target measurement value obtained by each measuring device with the known actual size of the standard target and calculating the respective measurement error. These deviations reflect the accuracy of the device in its current state.

[0083] "When the measurement deviations of the first type of non-contact measuring device and the second type of non-contact measuring device simultaneously exceed a preset threshold, and the deviation directions are consistent while the amplitude deviations are within a preset range, it is determined that common-mode drift has occurred." This means that when the measurement deviations of both devices exceed their respective set allowable ranges, and the directions of these two deviations (e.g., both are too large or both are too small) are the same, and their deviation amplitudes are also close to each other (e.g., the difference does not exceed a certain percentage), then the system determines that common-mode drift has occurred. Common-mode drift usually means that there are common factors affecting all measuring devices, such as changes in ambient temperature, power fluctuations, or drift of a common reference reference.

[0084] "Triggering an abnormal drift warning when common-mode drift is detected" means that once the system identifies common-mode drift, it immediately issues an alarm to the operator or control system, prompting the need to check, calibrate, or take other corrective measures for the measuring equipment.

[0085] "Continuously monitoring the long-term trend of output data from all types of measuring devices during normal production operation" means that when the production line is running normally and materials are passing through, the system will continuously collect and analyze the real-time output data of all measuring devices (including the first type of non-contact measuring devices and the second type of non-contact measuring devices) and track the trend of these data changes over time.

[0086] "Triggering an anomaly drift warning when long-term trends show that the output data of both Type I and Type II non-contact measuring devices consistently deviate from their historical baselines, and the cross-correlation between their output data remains stable," means that during normal production, if it is found that the output data of two devices consistently deviate from their respective historical normal operating baselines (e.g., average values ​​or setpoints), but the measurement results between the two devices still maintain their original cross-correlation, this indicates that a systematic, non-common-mode drift may have occurred. For example, material properties may change without the device itself drifting, or the device may drift but its relative relationship remains unchanged. In this case, the system will also trigger an anomaly drift warning, suggesting that the material properties may need to be checked or a deeper diagnostic of the device may be required.

[0087] Optionally, when the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device both exceed a preset threshold, and the deviation directions are consistent and the amplitude deviations are within a preset range, the step of determining that common-mode drift has occurred includes: Obtain production environment parameters; Based on historical calibration data and production environment parameters, adjust the measurement deviation threshold of the first type of non-contact measuring equipment and the measurement deviation threshold of the second type of non-contact measuring equipment. Based on historical calibration data and production environment parameters, adjust the range of similarity between the measurement deviation amplitudes of the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; When the measurement deviation of the first type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the first type of non-contact measuring device, and the measurement deviation of the second type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the second type of non-contact measuring device, and the deviation directions are consistent, and the deviation amplitude is within the range of the adjusted measurement deviation amplitude, it is determined that common mode drift has occurred.

[0088] Among these, production environment parameters can be understood as external conditions that affect the performance of measuring equipment or the characteristics of materials, such as ambient temperature, ambient humidity, production line load, material batch characteristics, and equipment uptime. These parameters can be acquired in real time through corresponding sensors or production management systems. Their purpose is to provide dynamic input information for subsequent threshold and range adjustments.

[0089] Furthermore, historical calibration data refers to the measurement equipment deviation data and corresponding environmental parameter records accumulated under different production environment parameters through methods such as standard target measurement or manual calibration. This data can be used to establish a mapping relationship or model between measurement deviations and environmental parameters. Its purpose is to provide empirical evidence and a model basis for adaptive adjustments.

[0090] In practical applications, adjusting the measurement deviation thresholds for both the first and second types of non-contact measuring devices, as well as adjusting the range of similarity in measurement deviation amplitudes between them, can be based on pre-trained machine learning models, lookup tables, or empirical formulas. For example, when the ambient temperature rises, some components of the measuring device may experience thermal expansion, leading to systematic deviations in the measurement results. In this case, the thresholds can be adjusted based on historical data to tolerate a certain degree of normal thermal drift while still effectively identifying abnormal drift. The aim is to enable the common-mode drift judgment criteria to dynamically adapt to changes in actual production conditions, improving the accuracy and robustness of the judgment.

[0091] Therefore, common-mode drift can only be determined to have occurred when the measurement deviation of the first type of non-contact measuring device exceeds the adjusted measurement deviation threshold for the first type of non-contact measuring device, and the measurement deviation of the second type of non-contact measuring device exceeds the adjusted measurement deviation threshold for the second type of non-contact measuring device, and the deviation directions are consistent, while the deviation amplitude is within the range of the adjusted measurement deviation amplitude. This ensures that the determination of common-mode drift is based on the current actual working conditions, rather than a rigid fixed standard.

[0092] In some preferred embodiments, a specific example is given below. Suppose that during the open-feed granulation process, batches of material change, the melt viscosity characteristics of the new batch differ from the previous batch, or the ambient temperature in the production workshop increases due to seasonal changes. Under the traditional fixed threshold determination method, the response characteristics of the measuring equipment may experience slight systematic drift due to the influence of material viscosity or ambient temperature, but this drift may not yet reach a level requiring a warning. However, if the old fixed threshold is used, this normal drift caused by changes in operating conditions may be misjudged as common-mode drift, thereby triggering unnecessary warnings and shutdown checks, resulting in reduced production efficiency.

[0093] Optionally, during normal production operation, the steps for continuously monitoring the long-term trends of output data from all types of measuring devices include: Obtain the material characteristic parameters of the current material batch to get the feature identifier of the current material batch; Based on the characteristic identifier of the current material batch, a preset material characteristic database is retrieved to obtain the historical measurement data baseline and fluctuation characteristic range that match the current material batch; When there is no matching historical measurement data baseline for the current material batch, after a preset period of stable production, a baseline for the current material batch is established as the historical measurement data baseline. Based on the measurement data of the current batch of materials, the baseline of historical measurement data, and the range of fluctuation characteristics, determine whether there is a trend deviation in the output data of the measuring equipment.

[0094] Specifically, obtaining the material characteristic parameters of the current material batch refers to identifying the material batch currently being processed by means of, for example, the material batch number, the material formula code, or the physicochemical properties of the material detected by online sensors (such as melt flow index, density, viscosity, additive content, etc.). This yields a unique identifier for subsequent material batch identification and data matching.

[0095] The system retrieves a pre-defined material property database based on the characteristic identifier of the current material batch. The purpose of this database is to find the most suitable reference baseline for the current material batch. This database can pre-store historical production data for different material types, grades, or formulations, including baseline measurement data (e.g., average value) and allowable fluctuation ranges (e.g., standard deviation or upper and lower limits) under normal and stable production conditions. By matching the characteristic identifier, the system can obtain the historical measurement data baseline and fluctuation range corresponding to the current material batch, thus providing a personalized reference standard for subsequent trend deviation judgment.

[0096] In practical applications, when there is no matching historical measurement data baseline for the current material batch—for example, when a production line processes a new type of material for the first time or when a material batch has not been recorded before—the system will automatically establish a baseline for the current material batch after the production process has been running stably for a preset period of time. This preset period is designed to ensure that the production process has reached a stable state, and that the collected data can accurately reflect the inherent characteristics of the material batch. During this stable production period, the measuring equipment continuously collects data, performs statistical analysis on this data, calculates the average value as the historical measurement data baseline for the material batch, and calculates the standard deviation or determines the fluctuation range as its fluctuation characteristic range.

[0097] Furthermore, based on the measurement data of the current material batch, the historical measurement data baseline, and the fluctuation range, the system determines whether there is a trend deviation in the output data of the measuring device. Specifically, the system compares the real-time collected measurement data of the current material batch with the historical measurement data baseline specific to that material batch, and, in conjunction with its fluctuation range, uses statistical methods (such as control charts, moving averages, exponential smoothing, etc.) to analyze the data trend. If the measurement data continuously and systematically deviates from its specific historical baseline and exceeds the preset fluctuation range, it can be determined that there is a trend deviation in the output data of the measuring device, thereby triggering the corresponding early warning or adjustment mechanism.

[0098] Optionally, when there is no matching historical measurement data baseline for the current material batch, after a preset period of stable production, the steps to establish a baseline for the current material batch as the historical measurement data baseline include: When there is no matching historical measurement data baseline for the current material batch, continuously monitor production process parameters and environmental parameters; When the fluctuations of production process parameters and environmental parameters exceed the preset range, the baseline data acquisition process is paused. Once the fluctuations in production process parameters and environmental parameters have stabilized, the baseline data collection process will continue. Outlier removal is performed on the collected baseline data; Statistical analysis was performed on the baseline data after outlier removal to obtain the statistical analysis results; Based on the statistical analysis results, the average value of the baseline data is calculated as the baseline of the current material batch, and the standard deviation of the baseline data is calculated as the range of fluctuation characteristics.

[0099] Specifically, when there is no matching historical measurement data baseline for the current material batch, to ensure the accuracy and reliability of the established baseline, continuous monitoring of production process parameters and environmental parameters is essential. Production process parameters may include, but are not limited to, extruder screw speed, melt pressure, die temperature, and pelletizing frequency, while environmental parameters may include workshop temperature, humidity, and power supply voltage stability. The purpose of continuous monitoring is to understand the real-time stability of production status and the external environment.

[0100] Furthermore, when fluctuations in production process parameters and environmental parameters are detected to exceed preset ranges—for example, significant fluctuations in screw speed or die temperature, or drastic changes in workshop temperature—the baseline data acquisition process is paused. This measure aims to prevent the acquisition of abnormal data under unstable conditions, thereby contaminating the baseline dataset.

[0101] Once the fluctuations in production process and environmental parameters stabilize—for example, when each parameter returns to its normal fluctuation range and remains so for a period of time—the baseline data acquisition process resumes. This ensures that data acquisition only occurs when production environment and process conditions are stable, thereby guaranteeing the quality of the baseline data.

[0102] In addition, outlier removal is performed on the collected baseline data. Outliers may be caused by transient interference, sensor malfunctions, or brief production fluctuations; failure to remove them will negatively impact the accuracy of the baseline. Baseline data after outlier removal is considered more representative.

[0103] Subsequently, statistical analysis was performed on the baseline data after outlier removal to obtain the statistical analysis results. This statistical analysis may include calculating the mean, median, standard deviation, variance, etc., of the data to comprehensively understand the distribution characteristics of the data.

[0104] Finally, based on the statistical analysis results, the average value of the baseline data is calculated as the baseline for the current material batch, and the standard deviation of the baseline data is calculated as the range of fluctuation characteristics. The average value represents the typical measurement value of the material batch under stable production conditions, while the standard deviation quantifies the normal fluctuation range around the baseline, providing an important reference for subsequent trend deviation judgment.

[0105] Optionally, the steps for outlier removal from the collected baseline data include: Get the material characteristic parameters of the current material batch; Obtain current production load parameters; Get the current ambient temperature parameters; Adjust the statistical range for outlier judgment based on the material characteristic parameters of the current batch, the current production load parameters, and the current ambient temperature parameters; When the corresponding data point in the collected baseline data exceeds the adjusted statistical range, the corresponding data point is identified as an outlier and removed.

[0106] Among these, obtaining the material characteristic parameters of the current batch refers to acquiring specific properties of the material currently in use, such as melt index, density, molecular weight distribution, and additive content, in real time during the granulation production process through a material management system, online sensors, or manual input. These parameters have a direct and significant impact on the flow pattern and viscosity of the melt. Obtaining the current production load parameters refers to real-time monitoring of the granulator's capacity utilization rate, such as the extruder's screw speed, feed rate, and output. These parameters directly reflect the equipment's operating intensity and production rhythm. Obtaining the current ambient temperature parameters involves measuring and recording the surrounding temperature in real time using environmental sensors deployed in the production workshop or key equipment locations. Changes in ambient temperature affect the physical properties of the material and the operating status of the equipment, thus influencing the distribution of the measurement data.

[0107] Furthermore, adjusting the statistical range for outlier detection based on the material characteristic parameters of the current batch, the current production load parameters, and the current ambient temperature parameters can be understood as establishing a dynamic outlier detection model. This model can be trained based on historical production data to learn the normal fluctuation range of baseline data under different combinations of material characteristics, production loads, and ambient temperatures. For example, machine learning algorithms (such as regression models or neural networks) or pre-defined lookup tables can be used to establish the mapping relationship between these dynamic parameters and the statistical range. When the above parameters change, the corresponding statistical range will also be dynamically adjusted to more accurately reflect the data distribution characteristics under the current production conditions.

[0108] In practical applications, when data points in the collected baseline data exceed the adjusted statistical range, these data points are identified as outliers and removed. This means marking data points that deviate from the current dynamic statistical range as anomalies. For example, the three-standard-deviation principle (3σ principle) in statistics can be used, but here the standard deviation and mean are calculated based on the dynamically adjusted statistical range. Data points exceeding this range will be automatically identified and removed from the baseline dataset to ensure the accuracy and reliability of subsequent statistical analysis.

[0109] This application also discloses a control system for an open granulation production process, used to execute open granulation production process control, combined with... Figure 3 As shown, the open granulation production process control system 1 includes: The flow parameter monitoring module 11 is used to monitor and record the flow parameters of the melt stream at the outlet of the pellet mill die head in real time. The trend information determination module 12 is used to monitor and record the drive motor torque and drive motor speed of the extrusion screw of the granulator in real time, and determine the fluctuation trend of melt viscosity based on the drive motor torque and drive motor speed to obtain melt viscosity fluctuation trend information. The compensation result judgment module 13 is used to judge the start-up strategy of the feedforward compensation mechanism based on the melt flow parameters and melt viscosity fluctuation trend information, and to obtain the feedforward compensation judgment result. The compensation control execution module 14 is used to adjust the rotation speed of the pelletizer's rotary cutter head or the heating power of the die head based on the feedforward compensation judgment result, so as to realize the control of the open granulation production process.

[0110] To better understand the open-process granulation production control system proposed in this application, the key modules and implementation environment involved will be described in detail below.

[0111] "Mel flow parameters" refer to the physical state and flow characteristics of the melt at the die outlet of the granulator, such as the diameter, velocity, temperature, and shape stability of the melt. These parameters directly reflect the uniformity and stability of the melt during the extrusion process and are important indicators for evaluating granulation quality.

[0112] "Mel viscosity fluctuation trend information" refers to the trend of melt viscosity change over time, inferred by monitoring the torque and speed of the extrusion screw in a granulator. Melt viscosity is an important rheological parameter of polymer materials, and its fluctuations directly affect the stability of the extrusion process and product quality. Changes in drive motor torque and speed are closely related to the shear and flow resistance of the melt in the screw, and therefore can serve as an indirect indicator reflecting changes in melt viscosity.

[0113] The "feedforward compensation mechanism" is a control strategy whose core idea is to predict or detect the impact of a disturbance on the system output before or immediately after it occurs, and then take compensatory measures in advance to reduce or eliminate the impact of the disturbance on the system output. In this application, the feedforward compensation mechanism aims to adjust the rotary cutter speed or die heating power of the pelletizer in advance based on the melt flow parameters and melt viscosity fluctuation trend information to cope with the fluctuation of melt viscosity and ensure the stability of the pelletizing process.

[0114] "The rotational speed of the pelletizer's cutter head" refers to the rotational speed of the cutter head used to cut the molten feedstock in the pelletizer. Adjusting this speed directly affects the pelletizing frequency and particle size.

[0115] "Die head heating power" refers to the power of the heater in the die head section of the pellet mill. Adjusting the die head heating power can change the temperature in the die head area, thereby affecting the viscosity and flowability of the melt.

[0116] The implementation environment of this application is typically an open-milling and granulation production line for polymer materials, which includes core equipment such as open mills, granulators (including extrusion screws and dies), and pelletizers. The entire production process requires precise control of parameters such as temperature, pressure, and speed to ensure product quality.

[0117] The core of the open granulation production process control system proposed in this application lies in achieving intelligent feedforward compensation control of the granulation process through the coordinated work of various functional modules. The implementation methods of each main module will be described in detail below.

[0118] First, regarding the "flow parameter monitoring module".

[0119] The above embodiments have already described the real-time monitoring and recording of the melt flow parameters at the pelletizer die outlet, and will not be repeated here. It should be emphasized that the flow parameter monitoring module, as a hardware or software entity, can be configured in various forms.

[0120] One implementation approach is that the flow parameter monitoring module can be a standalone hardware module integrating an image acquisition device (such as an industrial camera) and an image processing unit. The image acquisition device is used to acquire image information of the molten material, while the image processing unit analyzes the images to extract flow parameters such as the diameter and shape stability of the material.

[0121] Another implementation approach is that the flow parameter monitoring module can be a hardware module containing contact sensors (such as temperature sensors or pressure sensors) and a data acquisition unit. The contact sensors directly measure the temperature or pressure of the molten material, while the data acquisition unit is responsible for converting analog signals into digital signals and recording them.

[0122] Secondly, regarding the "Trend Information Judgment Module".

[0123] The above embodiments have already described the real-time monitoring and recording of the drive motor torque and drive motor speed of the granulator extrusion screw, and the determination of the melt viscosity fluctuation trend based on the drive motor torque and drive motor speed to obtain melt viscosity fluctuation trend information, which will not be repeated here. It should be emphasized that the trend information determination module, as a hardware or software entity, can be configured in various forms.

[0124] One implementation involves a trend information determination module that includes a torque sensor and a speed sensor, and integrates a microcontroller or digital signal processor. The sensors acquire real-time torque and speed data of the drive motor, and the microcontroller converts this data into melt viscosity fluctuation trend information according to a preset algorithm or lookup table.

[0125] Another approach is that the trend information determination module can be a software module deployed on an industrial control computer. By reading the drive motor current and voltage signals provided by the frequency converter, the torque and speed are estimated using a motor model, thereby determining the fluctuation trend of the melt viscosity.

[0126] Secondly, regarding the "compensation result judgment module".

[0127] The above embodiments have already described the judgment process for executing the feedforward compensation mechanism activation strategy based on melt flow parameters and melt viscosity fluctuation trend information, and the content of obtaining the feedforward compensation judgment result, which will not be repeated here. It should be emphasized that the compensation result judgment module, as a hardware or software entity, can be configured in various forms.

[0128] One implementation approach is to have the compensation result judgment module be a software module based on a rule engine, with a set of preset judgment rules. When the received melt flow parameters and melt viscosity fluctuation trend information meet the specific rules, the module outputs the feedforward compensation judgment result.

[0129] Another approach is to have the compensation result judgment module be a software module integrating a machine learning model. This model, trained on historical data, can predict whether a feedforward compensation mechanism needs to be activated based on real-time flow parameters and viscosity fluctuation trends, and output the judgment result.

[0130] Finally, regarding the "compensation control execution module".

[0131] The above embodiments have already described how the rotational speed of the pelletizer's cutter head or the heating power of the die head is adjusted based on the feedforward compensation judgment result to achieve control of the open-milling pelletizing production process, and will not be repeated here. It should be emphasized that the compensation control execution module, as a hardware or software entity, can be configured in various forms.

[0132] One implementation is that the compensation control execution module can be a hardware module containing a digital output interface, through which it sends speed adjustment commands to the frequency converter of the pelletizer or power adjustment commands to the power controller of the die heater.

[0133] Another implementation method is that the compensation control execution module can be a software module deployed in a host computer or PLC, which communicates with the pelletizer controller and die head heating controller through industrial communication protocols (such as Modbus, Profinet) to send adjustment commands.

[0134] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the open granulation production process, characterized in that, include: Real-time monitoring and recording of melt flow parameters at the pellet mill die outlet; The torque and speed of the drive motor of the extrusion screw of the granulator are monitored and recorded in real time. Based on the torque and speed of the drive motor, the fluctuation trend of the melt viscosity is determined and the information on the fluctuation trend of the melt viscosity is obtained. Based on the melt flow parameters and the melt viscosity fluctuation trend information, the judgment process of the feedforward compensation mechanism activation strategy is executed to obtain the feedforward compensation judgment result; Based on the feedforward compensation judgment result, adjust the rotation speed of the pelletizer's rotary cutter or the heating power of the die head to achieve control of the open granulation production process.

2. The method for controlling the open granulation production process according to claim 1, characterized in that, The process of determining the start-up strategy of the feedforward compensation mechanism based on the melt flow parameters and the melt viscosity fluctuation trend information, and obtaining the feedforward compensation determination result, includes the following steps: The flow parameters of the melt stream are corrected in real time to obtain the corrected flow parameters of the melt stream. The performance of the servo motor of the rotary cutter head is evaluated in real time to obtain servo motor performance information; Based on the melt viscosity fluctuation trend information, the corrected melt flow parameters and the servo motor performance information are integrated to generate a feedforward compensation judgment result. Based on the feedforward compensation judgment result, the pelletizing command is pre-processed in time. Based on the feedforward compensation judgment result, the pelletizing command is subjected to amplitude pre-distortion processing.

3. The method for controlling the open granulation production process according to claim 2, characterized in that, The step of real-time correction of the melt flow parameters to obtain the corrected melt flow parameters includes: The first type of non-contact measuring device is pre-deployed to acquire image information of the melt stock, and the image quality characteristics of the first type of non-contact measuring device are acquired based on the image information. Activate the pre-deployed second type of non-contact measuring equipment to obtain the measurement data of the melt strand, and obtain the measurement data of the second type of non-contact measuring equipment; Based on the image quality characteristics of the first type of non-contact measuring equipment, adjust the corresponding image processing parameters; Based on the adjusted image processing parameters, the first type of non-contact measuring device acquires measurement data, and obtains the adjusted measurement data of the first type of non-contact measuring device. Based on the measurement data from the second type of non-contact measuring device, the measurement data from the adjusted first type of non-contact measuring device are corrected to obtain the corrected melt flow parameters.

4. The method for controlling the open granulation production process according to claim 3, characterized in that, The step of real-time correction of the melt flow parameters to obtain the corrected melt flow parameters includes: When a systematic drift in the measurement characteristics of a Type I or Type II non-contact measuring device is detected, a standard target is introduced for measurement. A measurement drift compensation function is established based on the deviation between the measured value of the standard target and the actual size of the standard target; Based on the measurement drift compensation function, the measurement data of the first type of non-contact measuring device are corrected and adjusted to obtain the corrected melt flow parameters. Monitor the long-term trends and cross-correlation of the output data of all measuring devices, and trigger abnormal drift warnings based on the monitoring results.

5. The method for controlling the open granulation production process according to claim 4, characterized in that, The steps for monitoring the long-term trend and cross-correlation of the output data of all measuring devices, and triggering an abnormal drift warning based on the monitoring results, include: During production line downtime or low-load periods, place the standard target in the common measurement area of ​​the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; Within the common measurement area, the measurement value of the standard target by the first type of non-contact measurement device is obtained, and the measurement value of the first type of target is obtained; Within the common measurement area, the measurement value of the standard target by the second type of non-contact measurement device is obtained, and the measurement value of the second type of target is obtained; The measurement deviation of the first type of non-contact measuring device is obtained based on the deviation between the measured value of the first type of target and the actual size of the standard target. The measurement deviation of the second type of non-contact measuring device is obtained based on the deviation between the measured value of the second type of target and the actual size of the standard target. When the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device both exceed the preset threshold, and the deviation direction is consistent and the amplitude deviation is within the preset range, it is determined that common mode drift has occurred. When common-mode drift is detected, an abnormal drift warning is triggered. During normal production operation, continuously monitor the long-term trends of output data from all types of measuring equipment; An abnormal drift warning is triggered when the long-term trend shows that the output data of both the first type of non-contact measuring device and the second type of non-contact measuring device continuously deviate from the historical baseline, and the cross-correlation of the output data of the first type of non-contact measuring device and the second type of non-contact measuring device remains stable.

6. The method for controlling the open granulation production process according to claim 5, characterized in that, The step of determining common-mode drift as occurring when the measurement deviation of the first type of non-contact measuring device and the measurement deviation of the second type of non-contact measuring device simultaneously exceed a preset threshold, and the deviation directions are consistent and the amplitude deviations are within a preset range, includes: Obtain production environment parameters; Based on historical calibration data and the production environment parameters, adjust the measurement deviation threshold of the first type of non-contact measuring device and the measurement deviation threshold of the second type of non-contact measuring device; Based on historical calibration data and the production environment parameters, adjust the range of similarity between the measurement deviation amplitudes of the first type of non-contact measuring equipment and the second type of non-contact measuring equipment; When the measurement deviation of the first type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the first type of non-contact measuring device, and the measurement deviation of the second type of non-contact measuring device exceeds the adjusted measurement deviation threshold of the second type of non-contact measuring device, and the deviation directions are consistent, and the deviation amplitude is within the range of the adjusted measurement deviation amplitude, it is determined that common mode drift has occurred.

7. The method for controlling the open granulation production process according to claim 5, characterized in that, The steps for continuously monitoring the long-term trends of output data from all types of measuring devices during normal production operation include: Obtain the material characteristic parameters of the current material batch to get the feature identifier of the current material batch; Based on the characteristic identifier of the current material batch, a preset material characteristic database is retrieved to obtain the historical measurement data baseline and fluctuation characteristic range that match the current material batch; When there is no matching historical measurement data baseline for the current material batch, after a preset period of stable production, a baseline for the current material batch is established as the historical measurement data baseline. Based on the measurement data of the current material batch, the baseline of the historical measurement data, and the fluctuation characteristic range, determine whether there is a trend deviation in the output data of the measuring device.

8. The method for controlling the open granulation production process according to claim 7, characterized in that, When there is no matching historical measurement data baseline for the current material batch, the step of establishing a baseline for the current material batch as the historical measurement data baseline after a preset period of stable production includes: When there is no matching historical measurement data baseline for the current material batch, continuously monitor production process parameters and environmental parameters; When the fluctuations of the production process parameters and the environmental parameters exceed the preset range, the baseline data acquisition process is paused. Once the fluctuations in the production process parameters and the environmental parameters have stabilized, the baseline data acquisition process continues. Outlier removal is performed on the collected baseline data; Statistical analysis was performed on the baseline data after outlier removal to obtain the statistical analysis results; Based on the statistical analysis results, the average value of the baseline data is calculated as the baseline of the current material batch, and the standard deviation of the baseline data is calculated as the fluctuation characteristic range.

9. The method for controlling the open granulation production process according to claim 8, characterized in that, The step of outlier removal from the collected baseline data includes: Get the material characteristic parameters of the current material batch; Obtain current production load parameters; Get the current ambient temperature parameters; Adjust the statistical range for outlier judgment based on the material characteristic parameters of the current batch, the current production load parameters, and the current ambient temperature parameters; When the corresponding data point in the collected baseline data exceeds the adjusted statistical range, the corresponding data point is identified as an outlier and removed.

10. A control system for an open granulation production process, used to execute control of the open granulation production process, characterized in that, include: The flow parameter monitoring module is used to monitor and record the flow parameters of the melt stream at the outlet of the pellet mill die in real time. The trend information determination module is used to monitor and record the drive motor torque and drive motor speed of the extrusion screw of the granulator in real time, and determine the fluctuation trend of melt viscosity based on the drive motor torque and drive motor speed to obtain melt viscosity fluctuation trend information. The compensation result judgment module is used to determine the start-up strategy of the feedforward compensation mechanism based on the melt flow parameters and the melt viscosity fluctuation trend information, and to obtain the feedforward compensation judgment result. The compensation control execution module is used to adjust the rotation speed of the pelletizer's rotary cutter head or the heating power of the die head based on the feedforward compensation judgment result, so as to realize the control of the open granulation production process.

Images