Intelligent temperature control drying system and method for modified plastic particle production

Abstract

The application discloses an intelligent temperature control drying system and method for modified plastic particle production, and relates to the field of plastic production.The system comprises a sensing module, which is used for continuously acquiring the surface layer temperature and core layer temperature of the modified plastic particles during the drying process, and generating a material temperature change trend signal according to the difference change rate between the two; and an analysis module, which is used for acquiring multi-source time sequence data of the drying inlet air temperature, feeding rate and environmental temperature, and calculating the change rate of each data to extract the working condition disturbance characteristics.The application captures the temperature difference change trend of the surface layer and core layer of the modified plastic particles in real time, and correlates and maps the multi-source working condition disturbance characteristics such as the drying inlet air temperature, feeding rate and environmental temperature, so that the system can predict the deviation time and amplitude before the material temperature actually deviates from the set value, and then output the adjustment instructions of the heating power and hot air circulating air volume in advance, thereby overcoming the temperature overshoot or under-shoot problem caused by the lag response of the traditional feedback control.

Description

Technical Field

[0001] This invention relates to the field of plastic production technology, specifically to an intelligent temperature-controlled drying system and method for producing modified plastic particles. Background Technology

[0002] Modified plastic particles require drying during production to remove internal moisture, preventing defects such as bubbles and silver streaks caused by moisture evaporation during subsequent injection molding or extrusion. Different modified plastics have significantly different drying temperature requirements. For example, non-crystalline plastics like PC and ABS typically require drying temperatures above their glass transition temperature; excessively high temperatures can lead to particle softening, clumping, or even degradation. Therefore, the precision and stability of drying temperature control directly impact product quality. Existing drying equipment often employs hot air circulation drying, using PID temperature controllers or PLC controllers to adjust heating power. Some equipment also incorporates stirring or rotating mechanisms to improve particle heating uniformity. These temperature control strategies and equipment structures are widely used in the industry.

[0003] The invention patent application with application number 201310281260.6 discloses a method for optimizing and saving energy control of temperature in plastic dryers. This application aims to identify and control the dynamic characteristics of air temperature in dryers by adopting a predictive control strategy based on multi-model switching.

[0004] In actual drying systems, material temperature changes have a significant lag characteristic. Existing control strategies mostly adjust only after temperature deviations occur, and their response speed is difficult to match the rate of temperature change, resulting in prominent temperature overshoot and oscillation problems. How to predict the trend of material temperature changes in advance and actively adjust it is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides an intelligent temperature-controlled drying system and method for the production of modified plastic particles, which can effectively solve the problems of the prior art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions; This invention discloses an intelligent temperature-controlled drying system for the production of modified plastic particles, comprising: The system comprises the following modules: a sensing module for continuously acquiring the surface and core temperatures of the modified plastic particles during the drying process, and generating a material temperature change trend signal based on the rate of change of the difference between them; an analysis module for acquiring multi-source time-series data on the drying inlet air temperature, feeding rate, and ambient temperature, and calculating the rate of change of each data point to extract the operating condition disturbance characteristics; a prediction module for mapping the material temperature change trend signal to the operating condition disturbance characteristics, and predicting the time and magnitude of the material temperature deviation from the set value based on the mapping result; a calculation module for calculating the pre-compensation amount of heating power and hot air circulation volume based on the occurrence time and deviation magnitude; an output module for converting the pre-compensation amount into heater control commands and fan control commands, and outputting the control commands in advance of the occurrence time; and an iteration module for acquiring the deviation sequence between the actual material temperature feedback value and the prediction result, and iteratively updating the mapping parameters of the associated mapping based on the deviation sequence. The perception module is interactively connected to the parsing module via a local area network. The parsing module is interactively connected to the prediction module via a local area network. The prediction module is interactively connected to the calculation module via a local area network. The calculation module is interactively connected to the output module via a local area network. The iteration module is interactively connected to the perception module, parsing module, prediction module, and calculation module via a local area network.

[0007] Furthermore, the sensing module includes multiple surface temperature sensing units and multiple core temperature sensing units arranged longitudinally along the drying chamber. Each surface temperature sensing unit is located at the junction of the inner wall of the drying chamber and the modified plastic particle material layer. Each core temperature sensing unit obtains the temperature information inside the modified plastic particle material layer in a non-contact temperature measurement manner. The sensing module synchronously collects the readings of all sensing units according to a preset sampling period, and uses the difference between the surface temperature reading and the core temperature reading at the same longitudinal position at the same sampling time as the material layer temperature difference value at that location. Then, it performs a difference operation on the material layer temperature difference values ​​at two adjacent sampling times, and the resulting difference result is the difference change rate.

[0008] Furthermore, after obtaining the rate of change of the difference, the sensing module compares the rate of change of the difference at the current moment with the rate of change of the difference at the previous moment. When the absolute value of the rate of change of the difference shows a continuous increasing trend, and the surface temperature and core temperature rise synchronously at each moment, it is determined that the modified plastic particles are in a state of surface overheating and a first trend signal is generated accordingly. When the absolute value of the rate of change of the difference shows a continuous decreasing trend, and the rising rate of the surface temperature and the core temperature at each moment tends to zero, it is determined that the modified plastic particles are in a state of uniform internal and external temperature and a second trend signal is generated accordingly. The material temperature change trend signal is composed of the first trend signal and the second trend signal.

[0009] Furthermore, the drying air inlet temperature time series data, feeding rate time series data and ambient temperature time series data acquired by the analysis module are all accompanied by timestamps. The analysis module performs sliding window difference operation on each set of time series data and sets the ratio of the maximum change rate to the minimum change rate of the data set within the current sliding window as the disturbance intensity coefficient of the working condition factor corresponding to the data set. The analysis module assigns preset weights to the rate of change of drying air inlet temperature, the rate of change of feeding rate, and the rate of change of ambient temperature at the same time, and then performs a weighted summation. The weighted summation result is used as the comprehensive disturbance coefficient at that time, and the time when the comprehensive disturbance coefficient exceeds the preset threshold is marked as the time when the disturbance occurs. The operating condition disturbance characteristics include the disturbance intensity coefficients corresponding to each operating condition factor and the time when each disturbance occurs.

[0010] Furthermore, when the prediction module performs the correlation mapping operation, it first uses the time of disturbance occurrence as the time reference point, extracts the material temperature change trend signal segment within a preset time window before and after the reference point, and then uses the disturbance intensity coefficient corresponding to each working condition factor as the weight coefficient to perform weighted correction on the difference change rate of each moment in the signal segment. After correction, a deviation trend sequence is obtained. Based on the deviation trend sequence, the prediction module uses the least squares linear extrapolation method to calculate the predicted deviation time of the material temperature deviation from the set value, and calculates the corresponding deviation amplitude based on the cumulative value of the deviation at each moment in the deviation trend sequence.

[0011] Furthermore, the calculation module is pre-loaded with a heating power compensation coefficient table and a hot air circulation volume compensation coefficient table; The heating power compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the time interval between the predicted deviation time and the current time. The hot air circulation volume compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the current material layer thickness. The calculation module queries the two coefficient tables based on the predicted deviation time and deviation magnitude output by the prediction module. The product of the queried heating power compensation coefficient and the current heating power is used as the pre-compensation amount of heating power, and the product of the queried hot air circulation volume compensation coefficient and the current hot air circulation volume is used as the pre-compensation amount of hot air circulation volume.

[0012] Furthermore, when the output module performs control command conversion, it first converts the pre-compensation amount of the heating power into a duty cycle adjustment amount according to the percentage of the rated power of the heater, and generates a heater duty cycle adjustment command accordingly. At the same time, it converts the pre-compensation amount of the hot air circulation volume into a speed adjustment amount according to the percentage of the rated speed of the fan, and generates a fan speed adjustment command accordingly. The output module outputs the heater duty cycle adjustment command and the fan speed adjustment command before the predicted deviation time. The advance output time of the above commands is equal to the time difference between the predicted deviation time and the current time multiplied by a preset safety margin coefficient. The value range of the safety margin coefficient is initially between 0.05 and 0.35.

[0013] Furthermore, the deviation sequence obtained by the iteration module consists of multiple deviation values ​​arranged in chronological order. Each deviation value is the difference between the measured material temperature feedback value at the same moment and the material temperature deviation value predicted by the prediction module. The iteration module uses the root mean square error of the most recent N deviation values ​​in the deviation sequence as an indicator to evaluate the current mapping accuracy. When the indicator exceeds a preset threshold, the iteration module uses the deviation sequence as a training sample and uses the recursive least squares method to update the mapping parameters of the associated mapping online, and synchronizes the updated mapping parameters to the prediction module. The iterative module also associates and stores the updated mapping parameters with the corresponding deviation sequence identifier.

[0014] On the other hand, a smart temperature-controlled drying method for producing modified plastic particles includes: The system collects real-time temperatures of the particle surface and core layers within the drying chamber, calculates the rate of temperature change, identifies overheating of the surface layer and uniform internal and external temperatures, and outputs a material temperature change trend signal with status codes. It also collects time-series data on inlet air temperature, feeding rate, and ambient temperature with timestamps, calculates the disturbance intensity and comprehensive disturbance coefficient of each factor, and extracts the disturbance characteristics of the operating conditions. The system correlates and maps the material temperature trend signal with the disturbance characteristics of the operating conditions to construct a deviation trend sequence. Through least-squares fitting extrapolation, it predicts the time and magnitude of the material temperature deviation from the set value. Based on the predicted deviation time, deviation magnitude, and material layer thickness, it retrieves a preset compensation coefficient table and calculates the pre-compensation amount for heating power and hot air volume, respectively. These two types of pre-compensation amounts are converted into heater duty cycle and fan speed control commands, matched with a safety margin coefficient, and output control commands in advance before the material temperature deviation occurs. Finally, it collects the deviation sequence between the measured material temperature and the predicted value, calculates the root mean square error, and updates and stores the correlated mapping parameters online using the recursive least squares method after exceeding the limit.

[0015] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: In this invention, the temperature difference between the surface and core of the modified plastic particles is captured in real time, and the characteristics of multi-source working condition disturbances such as drying air temperature, feeding rate and ambient temperature are correlated and mapped. This enables the system to predict the time and magnitude of deviation before the material temperature actually deviates from the set value, and then output the adjustment command of heating power and hot air circulation volume in advance, effectively overcoming the problem of temperature overshoot or undershoot caused by the lag response of traditional feedback control. Meanwhile, different weighting coefficients are assigned to different disturbance factors, and the prediction time window is dynamically determined based on the thermal inertia time constant and sampling period, so that the calculation of the pre-compensation amount is more in line with the actual thermal response characteristics of the drying chamber, avoiding the insufficient adaptability of a single fixed parameter under varying working conditions. Furthermore, based on the deviation sequence between the actual material temperature feedback and the prediction results, the mapping parameters are updated online, enabling the system to adaptively adjust the fitting coefficients during long-term operation and continuously reduce the prediction error. This ensures that the actual material temperature fluctuation during the drying process is controlled within a narrow range allowed by the process, while maintaining the uniformity of the internal and external temperature of the modified plastic particles and the safety of the material. This reduces energy consumption and equipment wear caused by frequent and large-scale adjustments. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0017] Figure 1 This is a schematic diagram of an intelligent temperature-controlled drying system for the production of modified plastic particles. Figure 2 This is a schematic diagram of a smart temperature-controlled drying method for the production of modified plastic particles. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] The present invention will be further described below with reference to embodiments. Example

[0020] This embodiment describes an intelligent temperature-controlled drying system for the production of modified plastic particles, such as... Figure 1 As shown, it includes: The sensing module is used to continuously acquire the surface temperature and core temperature of the modified plastic particles during the drying process, and generate a material temperature change trend signal based on the rate of change of the difference between the two. The sensing module includes multiple surface temperature sensing units and multiple core temperature sensing units arranged longitudinally along the drying chamber. Each surface temperature sensing unit is located at the junction of the inner wall of the drying chamber and the modified plastic particle material layer. Each core temperature sensing unit obtains the temperature information inside the modified plastic particle material layer in a non-contact temperature measurement manner. Preferably, the core layer temperature sensing unit uses an infrared temperature sensor. Each infrared temperature sensor is spaced longitudinally along the drying chamber, positioned outside the observation window or dedicated temperature measurement port. Its sensing head faces inwards from the drying chamber and forms a 30° to 60° angle with the surface of the modified plastic particle layer to avoid interference from hot airflow on the temperature measurement optical path. The diameter of the infrared temperature sensor's measurement spot covering the material layer surface is controlled between 5mm and 15mm to ensure that the measured temperature data represents the temperature level of the core layer region at that longitudinal position. The installation position of each core layer temperature sensing unit should avoid the air inlet and outlet areas of the drying chamber to prevent localized airflow temperature anomalies from causing deviations in the temperature measurement results. The sensing module synchronously collects the readings of all sensing units according to the preset sampling period, and takes the difference between the surface temperature reading and the core temperature reading at the same longitudinal position at the same sampling time as the material layer temperature difference value at that point. Then, it performs a difference operation on the material layer temperature difference values ​​at two adjacent sampling times, and the resulting difference result is the difference change rate. The preset sampling period is user-defined on the system side and changes inversely with the system's temperature control accuracy requirements. That is, the higher the temperature control accuracy requirements, the shorter the preset sampling period, and vice versa. After obtaining the rate of change of the difference, the sensing module compares the rate of change of the difference at the current moment with the rate of change of the difference at the previous moment. When the absolute value of the rate of change of the difference shows a continuous increasing trend, and the surface temperature and core temperature rise synchronously at each moment, it is determined that the modified plastic particles are in a state of surface overheating and a first trend signal is generated accordingly. When the absolute value of the rate of change of the difference shows a continuous decreasing trend, and the rising rate of the surface temperature and the core temperature at each moment tends to zero, it is determined that the modified plastic particles are in a state of uniform internal and external temperature and a second trend signal is generated accordingly. The material temperature change trend signal is composed of a first trend signal and a second trend signal; Specifically, after determining that the surface is overheated, the sensing module concatenates the current difference rate of change value with the preset surface overheating state code word to form a first data frame containing a state identifier field and a rate of change value field, and outputs the first data frame as a first trend signal. After determining that the internal and external temperatures are uniform, the sensing module concatenates the current difference rate of change value with the preset internal and external temperature uniformity state code word to form a second data frame containing a state identifier field and a rate of change value field, and outputs the second data frame as a second trend signal. The binary codes of the status identifier fields in the first data frame and the second data frame are reverse codes of each other; The analysis module is used to acquire multi-source time-series data of drying air inlet temperature, feeding rate and ambient temperature, and to calculate the rate of change of each data to extract the operating condition disturbance characteristics. The time-series data of drying air inlet temperature, feeding rate and ambient temperature obtained by the analysis module are all marked with timestamps. The analysis module performs sliding window difference operation on each set of time-series data and sets the ratio of the maximum change rate to the minimum change rate of the data in the current sliding window as the disturbance intensity coefficient of the working condition factor corresponding to the data in that set. The specific calculation method of sliding window difference operation is as follows: Suppose a set of time series data contains m data points x1, x2, ..., x3 within the sliding window. m The rate of change between two adjacent points Where Δt is the time interval between two adjacent sampling times, The maximum rate of change of this set of data within the current sliding window is r_max = max{r1, r2, ..., The minimum rate of change, r_min, is equal to min{r1, r2, ..., }, then the disturbance intensity coefficient k = r_max / r_min. When r_min = 0, r_min is used as 1 / 10 of the absolute value of the smallest rate of change other than zero in the data set for substitution calculation. The sliding window moves backward along the time axis by one sampling period each time, and the overlap rate between the moved window and the previous window is not less than 50%; The analysis module assigns preset weights to the rate of change of drying air temperature, the rate of change of feeding rate, and the rate of change of ambient temperature at the same time, and then performs a weighted summation. The weighted summation result is used as the comprehensive disturbance coefficient at that time, and the time when the comprehensive disturbance coefficient exceeds the preset threshold is marked as the time when the disturbance occurs. The characteristics of the operating condition disturbance include the disturbance intensity coefficient corresponding to each operating condition factor and the time when each disturbance occurs; The width of the sliding window is the integer value obtained by dividing the ratio of the volume of the drying chamber to the rated air volume of the hot air circulation system by the sampling period and rounding it up, so that the window covers the time span required for the hot air to complete one full cycle in the drying chamber. In the preset weights, the rate of change of drying air temperature corresponds to the first weight coefficient, the rate of change of feeding rate corresponds to the second weight coefficient, and the rate of change of ambient temperature corresponds to the third weight coefficient. The value of the first weight coefficient is greater than that of the second weight coefficient, the value of the second weight coefficient is greater than that of the third weight coefficient, and the sum of the values ​​of the first weight coefficient, the second weight coefficient and the third weight coefficient after normalization is equal to 1. The preset threshold is determined based on the historical value of the comprehensive disturbance coefficient recorded by the system under steady-state drying conditions without external disturbance. Specifically, the sum of the arithmetic mean of the historical values ​​and three times the standard deviation is used as the value of the preset threshold. The prediction module is used to correlate and map the material temperature change trend signal with the operating condition disturbance characteristics, and predict the time and magnitude of the material temperature deviation from the set value based on the mapping result. When the prediction module performs the correlation mapping operation, it first takes the time when the disturbance occurs as the time reference point, extracts the material temperature change trend signal segment within the preset time window before and after the reference point, and then uses the disturbance intensity coefficient corresponding to each working condition factor as the weight coefficient to perform weighted correction on the difference change rate of each time in the signal segment. After correction, the deviation trend sequence is obtained. Based on the deviation trend sequence, the prediction module uses the least squares linear extrapolation method to calculate the predicted deviation time of the material temperature deviation from the set value, and calculates the corresponding deviation amplitude based on the cumulative value of the deviation at each time in the deviation trend sequence. The duration of the preset time window is determined based on the thermal inertia time constant of the drying chamber, specifically taken as any fixed multiple value between 0.5 and 1.5 times the thermal inertia time constant, or as any integer multiple of the sampling period of the sensing module between 3 and 8 times. The prediction module uses the least squares method to perform linear fitting on each data point in the deviation trend sequence that falls within a preset time window, obtaining a fitted straight line representing the change in deviation over time. This fitted straight line is represented as follows: ,in Let t be the deviation of material temperature at time t, a be the intercept parameter of the fitted line, and b be the slope parameter of the fitted line. The prediction module will determine the deviation threshold between the fitted line and the preset deviation. The time corresponding to the intersection point is determined as the predicted deviation time, and is then calculated according to the formula. The predicted deviation time is calculated, where To predict the timing of deviations, the prediction module pre-stores the material safety temperature threshold and process requirement temperature threshold for the modified plastic particles. The deviation threshold is determined by the difference between the material safety temperature threshold and the material temperature deviation from the set value, and the difference between the process requirement temperature threshold and the material temperature deviation from the set value. Specifically, the smaller of the two differences is taken as the deviation threshold. The calculation module is used to calculate the pre-compensation amount of heating power and the pre-compensation amount of hot air circulation volume based on the occurrence time and deviation range. The calculation module has a pre-set heating power compensation coefficient table and a hot air circulation volume compensation coefficient table; The heating power compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the time interval between the predicted deviation time and the current time. The hot air circulation volume compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the current material layer thickness. See the following table for an example of a heating power compensation coefficient table: It should be noted that the boundary value handling rule for each deviation range in the above two compensation coefficient tables is as follows: when the deviation range ΔT is exactly equal to the boundary value of a certain range, it belongs to the smaller interval number (i.e., the one with the lower temperature) of the two adjacent intervals where the boundary value is located. For example, when ΔT=2℃, it belongs to the ΔT<2℃ range; when ΔT=5℃, it belongs to the 2℃≤ΔT<5℃ range; and when ΔT=8℃, it belongs to the 5℃≤ΔT<8℃ range. For cases exceeding the maximum deviation range listed in the table (ΔT≥8℃), if ΔT exceeds 10℃, the compensation coefficient corresponding to the maximum prediction time (10min) or maximum material layer thickness (100mm) in the ΔT≥8℃ column of the table is used. If ΔT is between 8℃ and 10℃, linear interpolation is performed using the compensation coefficient corresponding to the column in the ΔT≥8℃ column that is closest to the current prediction time or material layer thickness. See the following table for an example of a heating power compensation coefficient table: The calculation module queries the two coefficient tables above according to the predicted deviation time and deviation magnitude output by the prediction module. The product of the queried heating power compensation coefficient and the current heating power is used as the pre-compensation amount of heating power, and the product of the queried hot air circulation volume compensation coefficient and the current hot air circulation volume is used as the pre-compensation amount of hot air circulation volume. The output module is used to convert the pre-compensation quantity into heater control commands and fan control commands, and output the control commands in advance of the occurrence time; When the output module performs control command conversion, it first converts the pre-compensation amount of heating power into a duty cycle adjustment amount according to the percentage of the rated power of the heater, and generates a heater duty cycle adjustment command accordingly. At the same time, it converts the pre-compensation amount of hot air circulation volume into a speed adjustment amount according to the percentage of the rated speed of the fan, and generates a fan speed adjustment command accordingly. The output module outputs heater duty cycle adjustment command and fan speed adjustment command before the predicted deviation time. The advance output time of the above commands is equal to the time difference between the predicted deviation time and the current time multiplied by the preset safety margin coefficient. The initial range of the safety margin coefficient is between 0.05 and 0.35. The specific value of the safety margin coefficient is positively correlated with the deviation magnitude output by the prediction module and the current absolute value of the difference change rate. That is, when the deviation magnitude exceeds the preset high threshold or the current absolute value of the difference change rate exceeds the preset high threshold, it indicates that the material temperature deviation is severe and the system's allowable response time is extremely short. The upper limit of the safety margin coefficient range is 0.35. When the deviation magnitude is lower than the preset low threshold and the current absolute value of the difference change rate is lower than the preset low threshold, it indicates that the material temperature deviation is mild and the system has sufficient response margin. The lower limit of the safety margin coefficient range is 0.05. When the deviation magnitude is between the low threshold and the high threshold, or the current absolute value of the difference change rate is between the low threshold and the high threshold, the safety margin coefficient increases linearly with the increase of the deviation magnitude or with the increase of the current absolute value of the difference change rate in the range of 0.05 to 0.35. The iteration module is used to obtain the deviation sequence between the actual material temperature feedback value and the prediction result, and to iteratively update the mapping parameters of the associated mapping based on the deviation sequence. The deviation sequence obtained by the iteration module consists of multiple deviation values ​​arranged in chronological order. Each deviation value is the difference between the measured material temperature feedback value at the same time and the material temperature deviation value predicted by the prediction module. The iteration module uses the root mean square error of the most recent N deviation values ​​in the deviation sequence as an indicator to evaluate the current mapping accuracy. When the indicator exceeds the preset threshold, the iteration module uses the deviation sequence as a training sample and uses the recursive least squares method to update the mapping parameters of the associated mapping online, and synchronizes the updated mapping parameters to the prediction module. The iterative module also associates and stores the updated mapping parameters with the corresponding deviation sequence identifier; The preset threshold is set based on the inherent measurement resolution of the temperature sensing unit in the sensing module and the allowable fluctuation range of the modified plastic particles in the drying process. Specifically, it is the sum of twice the inherent measurement resolution of the temperature sensing unit and 0.5 times the allowable fluctuation range. The allowable fluctuation range is the temperature width within ±3% of the set value. The set value is the target drying temperature set value of the modified plastic particles in the drying process. The mapping parameters of the correlation mapping include the weight allocation coefficient and intercept correction coefficient used by the prediction module when performing linear fitting of the deviation trend sequence using the least squares method. The weight allocation coefficient is used to weight and correct the rate of change of the difference at each time step, and the intercept correction coefficient is used to compensate and adjust the intercept parameter of the fitted line. The iteration module uses the recursive least squares method with the deviation sequence as the training target, iteratively calculates the optimal weight allocation coefficient and optimal intercept correction coefficient that minimize the root mean square error of the deviation sequence, and replaces the weight allocation coefficient and intercept correction coefficient currently used in the prediction module with the optimal weight allocation coefficient and optimal intercept correction coefficient, thereby realizing the online update of the mapping parameters. The specific implementation of the recursive least squares method is as follows: The difference between the measured material temperature deviation at the current moment and the predicted material temperature deviation output by the prediction module is used as the model output, and the weighted correction rate of the difference is used as the model input. The weight allocation coefficient and intercept correction coefficient in the prediction module are updated online. Each recursive update sequentially executes the following four steps: Based on the covariance matrix saved at the previous time step and the model input at the current time step, calculate the gain vector at the current time step; The deviation between the measured material temperature deviation at the current moment and the model prediction value is multiplied by the gain vector to obtain the correction amount of the weight allocation coefficient and the intercept correction coefficient. The correction is added to the weight allocation coefficient and intercept correction coefficient of the previous time step to obtain the updated weight allocation coefficient and intercept correction coefficient of the current time step. The covariance matrix is ​​recursively updated and stored based on the current gain vector and model input, for use in the next time step. A forgetting factor is configured in this recursive process to control the weight of historical data on the current parameter estimation; the forgetting factor ranges from 0.95 to 0.99. The initial weight allocation coefficients and intercept correction coefficients are taken from the calibration values ​​determined during the initial system calibration. The initial covariance matrix is ​​set as a diagonal matrix with all diagonal elements between 1,000 and 10,000. After each recursive update, the updated weight allocation coefficients and intercept correction coefficients are synchronized to the prediction module to replace the corresponding coefficients currently used in the prediction module. The perception module is interconnected with the parsing module via a local area network. The parsing module is interconnected with the prediction module via a local area network. The prediction module is interconnected with the calculation module via a local area network. The calculation module is interconnected with the output module via a local area network. The iteration module is interconnected with the perception module, parsing module, prediction module, and calculation module via a local area network.

[0021] In this embodiment, after the system starts the drying operation, the sensing module synchronously collects the surface and core temperatures of particles at multiple points along the longitudinal direction of the drying chamber. It calculates the temperature difference and temperature difference change rate of the material layer according to a custom cycle, quickly distinguishes between two working conditions: surface overheating and uniform internal and external temperature, and outputs a standardized material temperature trend signal. Multi-point synchronous temperature measurement combined with real-time temperature difference calculation can accurately capture the temperature differentiation between the inside and surface of the particles, avoiding local overheating and uneven drying problems from the source. The analysis module synchronously reads time-series data of inlet air temperature, feed rate, and ambient temperature with timestamps. It calculates the disturbance intensity for each operating condition using a sliding window adapted to the full hot air circulation duration, weights these factors according to preset weights to obtain a comprehensive disturbance coefficient, and marks the time when a disturbance exceeds the threshold, forming a complete disturbance characteristic for each operating condition. This multi-source data-linked disturbance identification can proactively detect external disturbances such as feed fluctuations and ambient temperature changes, reducing the predictive lag caused by monitoring only a single parameter. The prediction module uses the disturbance moment as a benchmark to extract the material temperature trend signal for the corresponding time period, and combines it with the disturbance intensity weighted correction to obtain the deviation trend sequence. Through least squares linear extrapolation, it calculates the predicted moment and deviation magnitude of the material temperature exceeding the tolerance. By establishing a correlation mapping between the operating condition disturbance and the particle temperature change, it can realize the advance prediction of temperature deviation and overcome the shortcomings of traditional temperature control with lag adjustment. The calculation module retrieves the built-in dual-dimensional compensation coefficient table, matches the heating power compensation coefficient based on the predicted deviation range and the time interval of the deviation, and matches the air volume compensation coefficient based on the material layer thickness, quickly calculating the pre-compensation amount for heating and hot air volume. The multi-dimensional table lookup compensation can take into account both the heating rate and the uniformity of hot air in the cavity, ensuring that the drying temperature adjustment is accurately matched to the on-site working conditions; The output module converts the pre-compensation amount into heater duty cycle and fan speed adjustment commands. Based on the deviation magnitude and temperature difference change rate, it dynamically adjusts the safety margin coefficient within the 0.05~0.35 range, issuing control commands in advance before the material temperature exceeds the limit. This advance issuance of control commands allows for equipment response buffer time, effectively suppressing large fluctuations in particle temperature and reducing the risk of material overheating and degradation. After on-site control is implemented, the iterative module continuously collects the difference between the measured material temperature and the predicted value, forming a time-series deviation sequence. The accuracy of the mapping model is judged by the root mean square error. When the error exceeds the standard, the associated mapping weights and intercept correction parameters are updated online using the recursive least squares method and synchronized to the prediction module, along with the archived parameters and corresponding deviation data. Continuous online iterative optimization of model parameters improves prediction accuracy as production conditions change. Long-term operation can stably reduce temperature control errors and decrease drying energy consumption and raw material loss.

[0022] Example 2: At the implementation level, based on Example 1, this example refers to... Figure 2 A further detailed description is provided of the intelligent temperature-controlled drying system for producing modified plastic particles in Example 1: A smart temperature-controlled drying method for producing modified plastic particles includes: The temperature of the particle surface and core layer in the drying chamber is collected in real time, the temperature difference change rate is calculated, the surface overheating and the internal and external uniform temperature conditions are identified, and the material temperature change trend signal with status code is output. Collect time-series data of inlet air temperature, feed rate, and ambient temperature with timestamps, calculate the disturbance intensity and comprehensive disturbance coefficient of each factor, and extract the disturbance characteristics of the operating conditions; By correlating the material temperature trend signal with the operating condition disturbance characteristics, a deviation trend sequence is constructed, and the time and magnitude of the material temperature deviation from the set value are predicted by least squares fitting extrapolation. Based on the predicted deviation time, deviation range, and material layer thickness, the preset compensation coefficient table is retrieved, and the pre-compensation amount of heating power and the pre-compensation amount of hot air volume are calculated respectively. The two types of pre-compensation quantities are converted into heater duty cycle and fan speed control commands, matched with safety margin coefficients, and control commands are output in advance before the material temperature deviation occurs. Collect the deviation sequence between the measured material temperature and the predicted value, calculate the root mean square error, and update the associated mapping parameters online and store them after exceeding the standard by recursive least squares method.

[0023] Referring to the above embodiments, the system and method in these embodiments capture the temperature difference between the surface and core layers of modified plastic particles in real time, and perform correlation mapping based on the multi-source operating condition disturbance characteristics such as drying air inlet temperature, feeding rate, and ambient temperature. This allows the system to predict the time and magnitude of deviation before the material temperature actually deviates from the set value, and thus output adjustment commands for heating power and hot air circulation volume in advance. This effectively overcomes the temperature overshoot or undershoot problem caused by the delayed response of traditional feedback control. At the same time, different weighting coefficients are assigned to different disturbance factors, and the system is based on the thermal inertia time constant and sampling period. Dynamically determining the prediction time window makes the calculation of the pre-compensation amount more consistent with the actual thermodynamic response characteristics of the drying chamber, avoiding the insufficient adaptability of a single fixed parameter under varying operating conditions. At the same time, based on the deviation sequence between the actual material temperature feedback and the prediction result, the mapping parameters are updated online, enabling the system to adaptively adjust the fitting coefficient during long-term operation and continuously reduce the prediction error. Thus, while ensuring the uniformity of temperature inside and outside the modified plastic particles and the safety of the material, the actual material temperature fluctuation range during the drying process is controlled within a narrow range allowed by the process, thereby reducing energy consumption and equipment wear caused by frequent and large-scale adjustments.

[0024] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent temperature-controlled drying system for the production of modified plastic particles, characterized in that, include: The sensing module is used to continuously acquire the surface temperature and core temperature of the modified plastic particles during the drying process, and generate a material temperature change trend signal based on the rate of change of the difference between the two. The analysis module is used to acquire multi-source time-series data of drying air inlet temperature, feeding rate and ambient temperature, and to calculate the rate of change of each data to extract the operating condition disturbance characteristics. The prediction module is used to correlate and map the material temperature change trend signal with the operating condition disturbance characteristics, and predict the time and magnitude of the material temperature deviation from the set value based on the mapping result. The calculation module is used to calculate the pre-compensation amount of heating power and the pre-compensation amount of hot air circulation volume based on the occurrence time and deviation range. The output module is used to convert the pre-compensation amount into heater control commands and fan control commands, and output the control commands in advance of the occurrence time; The iteration module is used to obtain the deviation sequence between the actual material temperature feedback value and the prediction result, and to iteratively update the mapping parameters of the associated mapping based on the deviation sequence.

2. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 1, characterized in that, The sensing module includes multiple surface temperature sensing units and multiple core temperature sensing units arranged longitudinally along the drying chamber. Each surface temperature sensing unit is located at the junction of the inner wall of the drying chamber and the modified plastic particle material layer. Each core temperature sensing unit obtains the temperature information inside the modified plastic particle material layer in a non-contact temperature measurement manner. The sensing module synchronously collects the readings of all sensing units according to a preset sampling period, and uses the difference between the surface temperature reading and the core temperature reading at the same longitudinal position at the same sampling time as the material layer temperature difference value at that location. Then, it performs a difference operation on the material layer temperature difference values ​​at two adjacent sampling times, and the resulting difference result is the difference change rate.

3. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 2, characterized in that, After obtaining the rate of change of the difference, the sensing module compares the rate of change of the difference at the current moment with the rate of change of the difference at the previous moment. When the absolute value of the rate of change of the difference shows a continuous increasing trend, and the surface temperature and core temperature rise synchronously at each moment, it is determined that the modified plastic particles are in a state of surface overheating and a first trend signal is generated accordingly. When the absolute value of the rate of change of the difference shows a continuous decreasing trend, and the rising rate of the surface temperature and the core temperature at each moment tends to zero, it is determined that the modified plastic particles are in a state of uniform internal and external temperature and a second trend signal is generated accordingly. The material temperature change trend signal is composed of the first trend signal and the second trend signal.

4. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 1, characterized in that, The time-series data of drying air inlet temperature, feeding rate and ambient temperature obtained by the analysis module are all marked with timestamps. The analysis module performs sliding window difference operation on each set of time-series data and sets the ratio of the maximum change rate to the minimum change rate of the data in the current sliding window as the disturbance intensity coefficient of the working condition factor corresponding to the data in that set. The analysis module assigns preset weights to the rate of change of drying air inlet temperature, the rate of change of feeding rate, and the rate of change of ambient temperature at the same time, and then performs a weighted summation. The weighted summation result is used as the comprehensive disturbance coefficient at that time, and the time when the comprehensive disturbance coefficient exceeds the preset threshold is marked as the time when the disturbance occurs. The operating condition disturbance characteristics include the disturbance intensity coefficients corresponding to each operating condition factor and the time when each disturbance occurs.

5. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 4, characterized in that, When the prediction module performs the correlation mapping operation, it first uses the time when the disturbance occurs as the time reference point, extracts the material temperature change trend signal segment within a preset time window before and after the reference point, and then uses the disturbance intensity coefficient corresponding to each working condition factor as the weight coefficient to perform weighted correction on the difference change rate of each moment in the signal segment. After correction, a deviation trend sequence is obtained. Based on the deviation trend sequence, the prediction module uses the least squares linear extrapolation method to calculate the predicted deviation time of the material temperature deviation from the set value, and calculates the corresponding deviation amplitude based on the cumulative deviation value of each moment in the deviation trend sequence.

6. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 5, characterized in that, The calculation module has a pre-set heating power compensation coefficient table and a hot air circulation volume compensation coefficient table. The heating power compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the time interval between the predicted deviation time and the current time. The hot air circulation volume compensation coefficient table is divided into multiple compensation coefficient values ​​according to two dimensions: the deviation range and the current material layer thickness. The calculation module queries the two coefficient tables based on the predicted deviation time and deviation magnitude output by the prediction module. The product of the queried heating power compensation coefficient and the current heating power is used as the pre-compensation amount of heating power, and the product of the queried hot air circulation volume compensation coefficient and the current hot air circulation volume is used as the pre-compensation amount of hot air circulation volume.

7. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 1, characterized in that, When the output module performs control command conversion, it first converts the pre-compensation amount of the heating power into a duty cycle adjustment amount according to the percentage of the rated power of the heater, and generates a heater duty cycle adjustment command accordingly. At the same time, it converts the pre-compensation amount of the hot air circulation volume into a speed adjustment amount according to the percentage of the rated speed of the fan, and generates a fan speed adjustment command accordingly. The output module outputs the heater duty cycle adjustment command and the fan speed adjustment command before the predicted deviation time. The advance output time of the above commands is equal to the time difference between the predicted deviation time and the current time multiplied by a preset safety margin coefficient. The value range of the safety margin coefficient is initially between 0.05 and 0.

35.

8. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 1, characterized in that, The deviation sequence obtained by the iterative module consists of multiple deviation values ​​arranged in chronological order. Each deviation value is the difference between the measured material temperature feedback value at the same moment and the material temperature deviation value predicted by the prediction module. The iterative module uses the root mean square error of the most recent N deviation values ​​in the deviation sequence as an indicator to evaluate the current mapping accuracy. When the indicator exceeds a preset threshold, the iterative module uses the deviation sequence as a training sample and uses the recursive least squares method to update the mapping parameters of the associated mapping online, and synchronizes the updated mapping parameters to the prediction module. The iterative module also associates and stores the updated mapping parameters with the corresponding deviation sequence identifier.

9. The intelligent temperature-controlled drying system for producing modified plastic particles according to claim 1, characterized in that, The perception module is interactively connected to the parsing module via a local area network. The parsing module is interactively connected to the prediction module via a local area network. The prediction module is interactively connected to the calculation module via a local area network. The calculation module is interactively connected to the output module via a local area network. The iteration module is interactively connected to the perception module, parsing module, prediction module, and calculation module via a local area network.

10. A smart temperature-controlled drying method for producing modified plastic particles, wherein the method is an implementation method of the smart temperature-controlled drying system for producing modified plastic particles as described in any one of claims 1-9, characterized in that, include: The temperature of the particle surface and core layer in the drying chamber is collected in real time, the temperature difference change rate is calculated, the surface overheating and the internal and external uniform temperature conditions are identified, and the material temperature change trend signal with status code is output. Collect time-series data of inlet air temperature, feed rate, and ambient temperature with timestamps, calculate the disturbance intensity and comprehensive disturbance coefficient of each factor, and extract the disturbance characteristics of the operating conditions; By correlating the material temperature trend signal with the operating condition disturbance characteristics, a deviation trend sequence is constructed, and the time and magnitude of the material temperature deviation from the set value are predicted by least squares fitting extrapolation. Based on the predicted deviation time, deviation range, and material layer thickness, the preset compensation coefficient table is retrieved, and the pre-compensation amount of heating power and the pre-compensation amount of hot air volume are calculated respectively. The two types of pre-compensation quantities are converted into heater duty cycle and fan speed control commands, matched with safety margin coefficients, and control commands are output in advance before the material temperature deviation occurs. Collect the deviation sequence between the measured material temperature and the predicted value, calculate the root mean square error, and update the associated mapping parameters online and store them after exceeding the standard by recursive least squares method.

Images