Accurate temperature control method and system for regeneration and recovery of injection molding products

Abstract

The invention discloses an accurate temperature control method and system for regeneration and recovery of injection molding products, and relates to the field of temperature control. The method comprises the following steps: analyzing regenerated plastic components in real time through a near-infrared spectrometer, and obtaining a melt index, a thermal stability parameter and a glass transition temperature; establishing a melt apparent viscosity-temperature-shear rate three-dimensional response curved surface model in combination with a rheology theory; the method comprises the following steps: acquiring temperature gradient and pressure data by utilizing a multi-sensor network deployed in each module of a double-screw extruder, calculating an actual shear rate and melt apparent viscosity, and comparing the actual shear rate and melt apparent viscosity with theoretical values to trigger a PID (Proportion Integration Differentiation) neural network to predict temperature fluctuation so as to carry out differential temperature compensation on a feeding section and a melting section; and after each batch of production, model parameters are optimized by adopting a least square method, so that closed-loop intelligent regulation and control in the processing process of the regenerated plastic are realized. The method has the advantages that through the three-dimensional model, the PID neural network differential temperature control is combined, the temperature control problem of the regenerated plastic is solved, the quality stability of products is improved, and the waste rate is reduced.

Description

Technical Field

[0001] This invention relates to the field of carbon accounting, and in particular to a precise temperature control method and system for the recycling of injection molded products. Background Technology

[0002] With the global plastic pollution problem becoming increasingly severe, the efficient recycling of recycled plastics has become an important direction for sustainable development. Injection-molded products, as a significant source of plastic waste, require extremely high temperature control during their recycling process. Traditional recycling processes often employ fixed temperature control modes, which are difficult to adapt to fluctuations in different types of plastics, aging levels, and impurity content. This leads to uneven thermal degradation or plasticization of recycled materials, affecting their mechanical properties and reprocessability. Furthermore, plastics are prone to oxidative cracking at high temperatures, while insufficient temperatures cannot completely remove contaminants, thus hindering the improvement of recycled plastic quality.

[0003] Current temperature control methods for recycled plastics have significant shortcomings. In terms of component analysis, they largely rely on manual sampling or single-spectral analysis, making it difficult to analyze the complex composition of recycled materials and key parameters such as melt flow index and thermal stability in real time. This leads to misjudgments in multi-component systems such as PE / PP blends and ABS composites, creating potential problems for temperature control. Temperature monitoring often uses single-point sensors without constructing a full-process spatial gradient network, only acquiring localized data and failing to capture temperature differences in critical areas such as the mixing and homogenization stages. This can easily lead to localized overheating and degradation or incomplete melting. In terms of modeling, the impact of thermal stability on melt viscosity is often ignored, and a dynamic correction mechanism is lacking. Models struggle to adapt to fluctuations in recycled material performance, resulting in large discrepancies between theoretical and actual values. Temperature control often uses traditional PID algorithms, which suffer from large overshoot and response lag. Furthermore, applying a uniform compensation strategy to the feeding and melting stages fails to consider the differences in characteristics between different modules. When viscosity deviations exceed thresholds, temperature fluctuations cannot be accurately predicted, leading to unstable product quality and high scrap rates. Summary of the Invention

[0004] To improve existing methods and systems, this paper provides a precise temperature control method and system for the recycling of injection molded products. This method uses a spectrometer to detect components, deploys multiple modules of sensors to monitor temperature, builds and optimizes a three-dimensional model, and combines it with PID neural network-based differentiated temperature control. This can solve the temperature control problem of recycled plastics, improve the quality stability of products, and reduce the waste rate.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A precise temperature control method for the recycling of injection-molded products includes:

[0007] Online composition detection of recycled plastic granules was performed using a near-infrared spectrometer. The composition of the recycled plastic granules was analyzed based on the spectral analysis algorithm, and the melt index, thermal stability parameters and glass transition temperature were obtained.

[0008] By deploying fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors in each module of a twin-screw extruder, a spatial temperature gradient monitoring network is constructed to monitor the entire process of injection molded product recycling and acquisition of temperature gradient data. Each module of the twin-screw extruder includes a feeding section, a melting section, a mixing section, a homogenizing section, and a die outlet.

[0009] Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with polymer melt rheology theory, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate was established.

[0010] Based on the pressure data from the melt pressure sensor, the actual shear rate is calculated by combining the screw geometry parameters of the twin-screw extruder and the real-time screw speed. The current temperature feedback value is obtained by the fiber optic temperature sensor and the infrared thermal imager. The actual shear rate and the temperature feedback value are substituted into the three-dimensional response surface model to calculate the real-time melt apparent viscosity.

[0011] The apparent viscosity of the melt is compared with the theoretical value calculated by the three-dimensional response surface model to calculate the viscosity deviation. When the viscosity deviation is greater than the threshold, the PID neural network is activated to predict the temperature fluctuation in the future time period.

[0012] Based on the temperature fluctuation, differentiated temperature compensation is performed on the feeding section and the melting section;

[0013] After each production batch is completed, actual temperature trajectory, shear rate and viscosity deviation data are collected, and the parameters of the three-dimensional response surface model are optimized by fitting using the least squares method.

[0014] Preferably, the step of performing online component detection of recycled plastic particles using a near-infrared spectrometer, and analyzing the composition of the recycled plastic particles according to a spectral analysis algorithm to obtain melt index, thermal stability parameters, and glass transition temperature specifically includes:

[0015] Real-time spectral acquisition of recycled plastic particles was performed using a near-infrared spectrometer, and a Raman spectrometer was simultaneously triggered to obtain molecular vibrational fingerprint spectra.

[0016] Near-infrared spectroscopy was used to identify the intensity of characteristic absorption peaks of CH bonds, carbonyl peaks were used to determine the degree of oxidation, Raman spectroscopy was used to analyze the vibrational modes of the CC skeleton, and molecular weight distribution was evaluated based on peak width.

[0017] Based on the comparison of the absorption peak area ratio of CH bond with the standard database, when PE / PP blend is detected, principal component analysis is used to separate overlapping peaks and obtain the melt index.

[0018] Correlate the carbonyl index with thermogravimetric data; for every 10% increase in oxidation degree, the threshold for thermal stability parameters is lowered by 15%.

[0019] By inverting the chain segment compliance through the C-C bond vibration half-width inversion of Raman spectroscopy, a weighted summation model of the glass transition temperatures of each component in the ABS blend was established to obtain the glass transition temperatures.

[0020] Preferably, by deploying fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors in each module of the twin-screw extruder, a spatial temperature gradient monitoring network is constructed to monitor the entire process of injection-molded product recycling and acquisition of temperature gradient data. The twin-screw extruder modules include a feeding section, a melting section, a mixing section, a homogenizing section, and a die exit section, specifically comprising:

[0021] Fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors are deployed at the feeding section, melting section, mixing section, homogenizing section, and die outlet of the twin-screw extruder, respectively.

[0022] Two fiber optic temperature sensors are set at preset distances along the screw axis of each module to obtain temperature data at different locations. An infrared thermal imager captures the temperature distribution image in the area in real time, covering the screw area and the inner wall of the barrel. The melt pressure sensor is installed on the side wall of the barrel of each module and the sensing end is in direct contact with the melt.

[0023] Through the coordinated deployment of the aforementioned sensors, a spatial temperature gradient monitoring network is constructed for the entire process of injection molded product recycling, enabling real-time collection of temperature gradient data and melt pressure data from each module.

[0024] Preferably, the establishment of a three-dimensional response surface model of the melt apparent viscosity-temperature-shear rate based on the acquired melt index, thermal stability parameters, and glass transition temperature data, combined with polymer melt rheology theory, specifically includes:

[0025] Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with the intrinsic relationship between melt viscosity and temperature and shear rate in polymer melt rheology theory, the apparent viscosity, temperature and shear rate of melt are selected as core variables. The intrinsic relationship is that the higher the temperature, the greater the shear rate.

[0026] By collecting apparent viscosity data of recycled plastic melt at different temperatures and shear rates, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate was established by fitting the mathematical correlation between core variables using multiple regression analysis.

[0027] During the model construction process, the influence of thermal stability parameters on melt viscosity was considered, and the calculated melt viscosity under different thermal stability conditions was adjusted by introducing a thermal stability correction coefficient.

[0028] Preferably, the calculation of the actual shear rate based on the pressure data from the melt pressure sensor, combined with the screw geometry parameters and real-time screw speed of the twin-screw extruder, and the acquisition of the current temperature feedback value through a fiber optic temperature sensor and an infrared thermal imager, and the substitution of the actual shear rate and temperature feedback value into the three-dimensional response surface model to calculate the real-time apparent melt viscosity specifically includes:

[0029] Based on real-time pressure data collected by the melt pressure sensor, combined with the screw geometry parameters of the twin-screw extruder and the real-time collected screw speed data, the actual shear rate of the melt under the current screw operating state is calculated by the shear rate calculation formula. The screw geometry parameters include screw groove depth, screw width, screw lead and screw diameter.

[0030] Real-time temperature data of the melt is obtained by deploying fiber optic temperature sensors, and temperature distribution images are obtained by infrared thermal imagers and the average temperature is extracted as an auxiliary value for temperature feedback.

[0031] The real-time temperature data acquired by the fiber optic temperature sensor and the average temperature extracted by the infrared thermal imager are weighted and fused to obtain the current temperature feedback value.

[0032] The calculated actual shear rate and temperature feedback values ​​are input into the three-dimensional response surface model, and the real-time apparent viscosity of the melt is obtained through model calculation.

[0033] Preferably, the step of comparing the apparent viscosity of the melt with the theoretically calculated value of the three-dimensional response surface model to calculate the viscosity deviation, and activating the PID neural network to predict the temperature fluctuation in the future time period when the viscosity deviation is greater than a threshold, specifically includes:

[0034] The obtained real-time apparent viscosity of the melt is compared with the theoretical calculation value of the three-dimensional response surface model under the same actual shear rate and temperature feedback value, and the viscosity deviation between the two is calculated.

[0035] A preset viscosity deviation threshold is set, which is determined based on the quality requirements of the injection molded product and the material characteristics of the recycled plastic. When the calculated viscosity deviation is greater than the preset threshold, the PID neural network model is activated.

[0036] The current viscosity deviation, temperature change rate, and pressure change rate are input into the PID neural network model. The neurons in the input layer convert the parameters into signals that the model can calculate. The hidden layer performs nonlinear integration of the signals, and the proportional, integral, and derivative characteristics of the PID in the output layer are used to calculate the temperature fluctuation in the future time period.

[0037] Preferably, the differential temperature compensation for the feeding section and the melting section based on the temperature fluctuation specifically includes:

[0038] Based on the temperature fluctuation predicted by the PID neural network model, combined with the obtained thermal stability parameters and temperature gradient data of recycled plastic particles, differential temperature compensation is performed on the feeding section and melting section of the twin-screw extruder.

[0039] For the feeding section, a segmented temperature compensation strategy is adopted. The compensation range is increased in areas where the temperature fluctuation is higher than the threshold, and the compensation range is decreased in areas where the temperature fluctuation is lower than the threshold. The compensation temperature of the feeding section is controlled to not exceed the preheating threshold corresponding to the glass transition temperature of the recycled plastic particles.

[0040] For the molten section, when the thermal stability parameter is below the threshold, a slow heating compensation method is used; when the thermal stability parameter is above the threshold, the temperature compensation rate is increased. During the compensation process, the change in melt pressure is monitored in real time, and the compensation range is dynamically adjusted according to the pressure change.

[0041] Preferably, after each production batch is completed, the actual temperature trajectory, shear rate, and viscosity deviation data are collected, and the parameters of the three-dimensional response surface model are optimized by fitting using the least squares method. This specifically includes:

[0042] After each production batch is completed, collect the actual temperature trajectory data, shear rate data, and viscosity deviation data obtained during the production process of that batch.

[0043] The parameters of the three-dimensional response surface model are fitted and optimized using the least squares method, and the regression coefficients and correction coefficients in the model are updated.

[0044] After optimization, new recycled plastic particle samples are selected for experimental verification. If the error between the melt apparent viscosity calculated by the model and the experimental measurement value is less than the preset verification threshold, the optimized model will be used as the temperature control calculation model for the next batch of production. If the error is greater than the preset verification threshold, the parameters will be optimized again until the model accuracy meets the requirements.

[0045] Furthermore, a precise temperature control system for the recycling of injection-molded products is proposed, including:

[0046] Composition analysis module: The module uses a near-infrared spectrometer and a Raman spectrometer in combination to analyze the melt index, thermal stability parameters and glass transition temperature of recycled plastic particles in real time;

[0047] Temperature gradient monitoring module: The module deploys fiber optic temperature sensors, infrared thermal imagers and melt pressure sensors in each key section of the twin-screw extruder to build a real-time monitoring network for spatial temperature gradient and pressure throughout the entire process;

[0048] Rheological modeling module: Based on composition data and polymer melt rheology theory, the module establishes and dynamically maintains a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate, including thermal stability correction.

[0049] Viscosity monitoring module: The module calculates the actual shear rate based on pressure data, screw parameters and rotation speed, integrates multi-source temperature data, and substitutes it into the rheological model to calculate the real-time apparent viscosity of the melt;

[0050] Temperature control decision module: The module compares the real-time viscosity with the theoretical value calculated by the model, and when the deviation exceeds the threshold, it activates the PID neural network to predict the future temperature fluctuation.

[0051] Temperature compensation module: The module implements differentiated temperature compensation strategies for the feeding section and the melting section based on the predicted fluctuation amount, thermal stability parameters and temperature gradient;

[0052] Processor: The processor is used to handle the calculation process of each formula and the construction calculation process of each model.

[0053] Compared with the prior art, the advantages of the present invention are:

[0054] By combining near-infrared and Raman spectroscopy for online detection of recycled plastic granule composition, key parameters such as melt flow index are accurately obtained, providing a precise initial basis for subsequent temperature control and solving the temperature control difficulties caused by the complex composition of recycled plastics. Secondly, multi-dimensional sensors are deployed in each module of the twin-screw extruder to construct a full-process temperature gradient monitoring network, enabling real-time and comprehensive capture of temperature changes and preventing localized temperature runaway. Furthermore, a three-dimensional response surface model is established based on rheological theory, and a thermal stability correction coefficient is introduced to accurately reflect the relationship between melt viscosity and temperature and shear rate. The model is continuously improved in accuracy through batch data optimization. In addition, when viscosity deviation exceeds a threshold, a PID neural network predicts temperature fluctuations and provides differentiated compensation for the feeding and melting sections, ensuring precise temperature control while adapting to the characteristics of different modules. This effectively improves the quality stability of recycled injection molded products, reduces waste rate, and promotes the efficient recycling of recycled plastics. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the method proposed in this invention;

[0056] Figure 2 This is a schematic diagram illustrating the composition of recycled plastics as proposed in this invention;

[0057] Figure 3 This is a schematic diagram illustrating the acquisition of temperature gradient data proposed in this invention;

[0058] Figure 4 This is a schematic diagram of the three-dimensional response surface model proposed in this invention;

[0059] Figure 5 This is a schematic diagram illustrating the calculation of real-time melt apparent viscosity proposed in this invention;

[0060] Figure 6 This is a schematic diagram of the temperature fluctuation prediction gauge proposed in this invention;

[0061] Figure 7 This is a schematic diagram of the differential temperature compensation proposed in this invention;

[0062] Figure 8 This is a schematic diagram of the optimized three-dimensional response surface model parameters proposed in this invention. Detailed Implementation

[0063] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0064] A precise temperature control system for the recycling of injection-molded products includes:

[0065] Composition analysis module: The module uses a near-infrared spectrometer and a Raman spectrometer in combination to analyze the melt index, thermal stability parameters and glass transition temperature of recycled plastic particles in real time;

[0066] Temperature gradient monitoring module: The module deploys fiber optic temperature sensors, infrared thermal imagers and melt pressure sensors in each key section of the twin-screw extruder to build a real-time monitoring network for spatial temperature gradient and pressure throughout the entire process;

[0067] Rheological modeling module: Based on composition data and polymer melt rheology theory, the module establishes and dynamically maintains a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate, including thermal stability correction.

[0068] Viscosity monitoring module: The module calculates the actual shear rate based on pressure data, screw parameters and rotation speed, integrates multi-source temperature data, and substitutes it into the rheological model to calculate the real-time apparent viscosity of the melt;

[0069] Temperature control decision module: The module compares the real-time viscosity with the theoretical value calculated by the model, and when the deviation exceeds the threshold, it activates the PID neural network to predict the future temperature fluctuation.

[0070] Temperature compensation module: The module implements differentiated temperature compensation strategies for the feeding section and the melting section based on the predicted fluctuation amount, thermal stability parameters and temperature gradient;

[0071] Processor: The processor is used to handle the calculation process of each formula and the construction calculation process of each model.

[0072] See Figure 1As shown, a precise temperature control method for the recycling of injection molded products includes:

[0073] Step 1: Online composition detection of recycled plastic granules is performed using a near-infrared spectrometer. The composition of the recycled plastic granules is analyzed based on the spectral analysis algorithm to obtain the melt index, thermal stability parameters, and glass transition temperature.

[0074] Step 2: By deploying fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors in each module of the twin-screw extruder, a spatial temperature gradient monitoring network for the entire process of injection molded product recycling is constructed to obtain temperature gradient data. Each module of the twin-screw extruder includes the feeding section, melting section, mixing section, homogenizing section, and die outlet.

[0075] Step 3: Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with polymer melt rheology theory, establish a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate.

[0076] Step 4: Based on the pressure data from the melt pressure sensor, the actual shear rate is calculated by combining the screw geometry parameters of the twin-screw extruder and the real-time screw speed. The current temperature feedback value is obtained through a fiber optic temperature sensor and an infrared thermal imager. The actual shear rate and temperature feedback value are substituted into the three-dimensional response surface model to calculate the real-time apparent viscosity of the melt.

[0077] Step 5: Compare the apparent viscosity of the melt with the theoretical calculation value of the three-dimensional response surface model to calculate the viscosity deviation. When the viscosity deviation is greater than the threshold, start the PID neural network to predict the temperature fluctuation in the future time period.

[0078] Step 6: Perform differentiated temperature compensation for the feeding section and the melting section based on the temperature fluctuation.

[0079] Step 7: After each production batch is completed, collect the actual temperature trajectory, shear rate and viscosity deviation data, and optimize the parameters of the three-dimensional response surface model by fitting the least squares method.

[0080] See Figure 2 As shown, online composition detection of recycled plastic granules is performed using a near-infrared spectrometer. The composition of the recycled plastic granules is analyzed using a spectral analysis algorithm to obtain melt flow index, thermal stability parameters, and glass transition temperature. Specifically, this includes:

[0081] Real-time spectral acquisition of recycled plastic particles was performed using a near-infrared spectrometer, and a Raman spectrometer was simultaneously triggered to obtain molecular vibrational fingerprint spectra.

[0082] Near-infrared spectroscopy was used to identify the intensity of characteristic absorption peaks of CH bonds, carbonyl peaks were used to determine the degree of oxidation, Raman spectroscopy was used to analyze the vibrational modes of the CC skeleton, and molecular weight distribution was evaluated based on peak width.

[0083] Based on the comparison of the absorption peak area ratio of CH bond with the standard database, when PE / PP blend is detected, principal component analysis is used to separate overlapping peaks and obtain the melt index.

[0084] Correlate the carbonyl index with thermogravimetric data; for every 10% increase in oxidation degree, the threshold for thermal stability parameters is lowered by 15%.

[0085] By inverting the chain segment compliance through the C-C bond vibration half-width inversion of Raman spectroscopy, a weighted summation model of the glass transition temperatures of each component in the ABS blend was established to obtain the glass transition temperatures.

[0086] See Figure 3 As shown, by deploying fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors in each module of the twin-screw extruder, a spatial temperature gradient monitoring network is constructed to monitor the entire process of injection-molded product recycling and acquisition of temperature gradient data. The twin-screw extruder modules include a feeding section, a melting section, a mixing section, a homogenizing section, and a die exit section, specifically comprising:

[0087] Fiber optic temperature sensors, infrared thermal imagers, and melt pressure sensors are deployed at the feeding section, melting section, mixing section, homogenizing section, and die outlet of the twin-screw extruder, respectively.

[0088] Two fiber optic temperature sensors are set at preset distances along the screw axis of each module to obtain temperature data at different locations. An infrared thermal imager captures the temperature distribution image in the area in real time, covering the screw area and the inner wall of the barrel. The melt pressure sensor is installed on the side wall of the barrel of each module and the sensing end is in direct contact with the melt.

[0089] Through the coordinated deployment of the aforementioned sensors, a spatial temperature gradient monitoring network is constructed for the entire process of injection molded product recycling, enabling real-time collection of temperature gradient data and melt pressure data from each module.

[0090] Specifically, the outputs of all fiber optic temperature sensors and melt pressure sensors are connected to a multi-channel data acquisition unit, and an infrared thermal imager is connected to the acquisition unit via a network cable or wireless module; the data acquisition unit is connected to a computer terminal via a network cable to build a spatial temperature gradient monitoring network for the entire process of injection molded product recycling.

[0091] See Figure 4 As shown, based on the obtained melt index, thermal stability parameters, and glass transition temperature data, and combined with polymer melt rheology theory, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate is established, specifically including:

[0092] Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with the intrinsic relationship between melt viscosity and temperature and shear rate in polymer melt rheology theory, the apparent viscosity, temperature and shear rate of melt are selected as core variables. The intrinsic relationship is that the higher the temperature, the greater the shear rate.

[0093] By collecting apparent viscosity data of recycled plastic melt at different temperatures and shear rates, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate was established by fitting the mathematical correlation between core variables using multiple regression analysis.

[0094] During the model construction process, the influence of thermal stability parameters on melt viscosity was considered, and the calculated melt viscosity under different thermal stability conditions was adjusted by introducing a thermal stability correction coefficient.

[0095] Specifically, based on the theory of polymer melt rheology, the function form of the multiple regression analysis is determined. Considering the synergistic effect of temperature and shear rate on viscosity, a quadratic polynomial is selected as the basic fitting function. The function must include a temperature term, a shear rate term, and a temperature-shear rate interaction term.

[0096] After generating the initial regression equation, check the coefficient of determination R² and the significance of the regression coefficients. If a variable term is not significant enough, it can be removed and the model refitted. If R² < 0.9, experimental data needs to be supplemented until the fitting effect meets the standard. Based on the optimized regression equation, plot a three-dimensional surface graph of temperature-shear rate-apparent viscosity to visually observe the trend of viscosity with the two variables. The expression of the three-dimensional response surface model is as follows:

[0097]

[0098] in, This refers to the apparent viscosity of the melt. The absolute temperature of the melt. For shear rate, , The diameter of the screw in a twin-screw extruder. For real-time screw speed, For the depth of the screw groove, This is the consistency coefficient. For activation energy, The gas constant is... The power-law exponent;

[0099] Using thermal weight loss as the core thermal stability index, the relationship between thermal weight loss and viscosity deviation at different temperatures was analyzed. For example, at 180℃, when the thermal weight loss is 0.5%, the viscosity is 5% lower than that of fresh material; at 230℃, when the thermal weight loss is 2%, the viscosity is 15% lower than that of fresh material. Based on this, the calculation method of the thermal stability correction coefficient was determined. The thermal stability correction coefficient is a coefficient less than 1. The higher the thermal weight loss, the smaller the thermal stability correction coefficient. Through linear fitting, the correlation between the thermal stability correction coefficient and the thermal weight loss was obtained.

[0100] Multiply the viscosity calculation value of the preliminary model by the thermal stability correction factor to obtain the corrected viscosity value; select 3-5 sets of experimental data with different thermal weight loss rates, and compare the error between the calculated value of the corrected model and the actual measured value. If the error is too large, adjust the correction factor formula until the prediction accuracy of the corrected model meets the standard.

[0101] By integrating the corrected regression equation and correction coefficients, a complete three-dimensional response surface model of melt apparent viscosity-temperature-shear rate is formed.

[0102] See Figure 5 As shown, based on the pressure data from the melt pressure sensor, combined with the screw geometry parameters and real-time screw speed of the twin-screw extruder, the actual shear rate is calculated. The current temperature feedback value is obtained through a fiber optic temperature sensor and an infrared thermal imager. The actual shear rate and temperature feedback value are then substituted into the three-dimensional response surface model to calculate the real-time apparent melt viscosity. Specifically, this includes:

[0103] Based on real-time pressure data collected by the melt pressure sensor, combined with the screw geometry parameters of the twin-screw extruder and the real-time collected screw speed data, the actual shear rate of the melt under the current screw operating state is calculated by the shear rate calculation formula. The screw geometry parameters include screw groove depth, screw width, screw lead and screw diameter.

[0104] Real-time temperature data of the melt is obtained by deploying fiber optic temperature sensors, and temperature distribution images are obtained by infrared thermal imagers and the average temperature is extracted as an auxiliary value for temperature feedback.

[0105] The real-time temperature data acquired by the fiber optic temperature sensor and the average temperature extracted by the infrared thermal imager are weighted and fused to obtain the current temperature feedback value.

[0106] The calculated actual shear rate and temperature feedback values ​​are input into the three-dimensional response surface model, and the real-time apparent viscosity of the melt is obtained through model calculation.

[0107] Specifically, the module for which the shear rate needs to be calculated is selected, and the screw geometry parameters of that module are retrieved from the parameter library; real-time pressure data from the melt pressure sensor of that module is extracted; the circumferential linear velocity of the screw is calculated, and the linear velocity of the screw surface is converted from the screw rotation speed and screw diameter through mechanical motion; the influence of the linear velocity on shearing is adjusted according to the screw groove depth, and the effective shearing area of ​​the screw groove is corrected with reference to the screw edge width; the initial shear rate is corrected using real-time pressure data. When the pressure is higher than the normal process pressure range of the module, it indicates that the melt flow resistance has increased, and the calculated shear rate value needs to be appropriately increased; the calculation results of the above steps are integrated to obtain the actual shear rate of the melt of that module.

[0108] By fusing point data (optical fiber) and area data (infrared), a more accurate real-time temperature of the module can be obtained as a temperature feedback value.

[0109] Substituting the actual shear rate and temperature feedback value of the same module and the same time stamp into the three-dimensional response surface model, the model outputs the real-time apparent viscosity value of the melt based on the previously fitted viscosity-temperature-shear rate correlation and combined with the thermal stability correction coefficient.

[0110] See Figure 6 As shown, the apparent viscosity of the melt is compared with the theoretically calculated value from the three-dimensional response surface model to calculate the viscosity deviation. When the viscosity deviation exceeds a threshold, the PID neural network is activated to predict the temperature fluctuation over the future time period. Specifically, this includes:

[0111] The obtained real-time apparent viscosity of the melt is compared with the theoretical calculation value of the three-dimensional response surface model under the same actual shear rate and temperature feedback value, and the viscosity deviation between the two is calculated.

[0112] A preset viscosity deviation threshold is set, which is determined based on the quality requirements of the injection molded product and the material characteristics of the recycled plastic. When the calculated viscosity deviation is greater than the preset threshold, the PID neural network model is activated.

[0113] The current viscosity deviation, temperature change rate, and pressure change rate are input into the PID neural network model. The neurons in the input layer convert the parameters into signals that the model can calculate. The hidden layer performs nonlinear integration of the signals, and the proportional, integral, and derivative characteristics of the PID in the output layer are used to calculate the temperature fluctuation in the future time period.

[0114] Specifically, the three preprocessed parameters are input into the model's input layer, with each parameter corresponding to one input neuron. The input layer neurons convert the parameters into data signals that the model can recognize, and at the same time label the parameter type.

[0115] The hidden layer performs nonlinear operations on the input signal through multiple neurons to capture the correlation between parameters. For example, an increase in viscosity deviation and a negative temperature change rate mean that the temperature is too low, resulting in high viscosity, and the temperature needs to be adjusted positively. The hidden layer weakens secondary signals by weighting, such as pressure change rate fluctuations, which have lower weights, and strengthens primary signals, such as viscosity deviations, which have higher weights, to ensure that the integrated signal can accurately reflect the core influencing factors of temperature fluctuations.

[0116] The output layer neurons combine the three characteristics of PID: proportional, integral, and derivative characteristics. Based on the response speed of the temperature regulation of the twin-screw extruder, the future prediction time period is set to the next 10 seconds, and the temperature fluctuation during this time period is calculated.

[0117] Compare the predicted fluctuation with the maximum adjustment capacity of the heating system. If the fluctuation is within the adjustment range, output the final fluctuation. If it exceeds the range, indicate that the fluctuation exceeds the adjustment capacity and correct the fluctuation to the maximum adjustment value.

[0118] See Figure 7 As shown, based on the temperature fluctuation, the differentiated temperature compensation for the feeding section and the melting section specifically includes:

[0119] Based on the temperature fluctuation predicted by the PID neural network model, combined with the obtained thermal stability parameters and temperature gradient data of recycled plastic particles, differential temperature compensation is performed on the feeding section and melting section of the twin-screw extruder.

[0120] For the feeding section, a segmented temperature compensation strategy is adopted. The compensation range is increased in areas where the temperature fluctuation is higher than the threshold, and the compensation range is decreased in areas where the temperature fluctuation is lower than the threshold. The compensation temperature of the feeding section is controlled to not exceed the preheating threshold corresponding to the glass transition temperature of the recycled plastic particles.

[0121] For the molten section, when the thermal stability parameter is below the threshold, a slow heating compensation method is used; when the thermal stability parameter is above the threshold, the temperature compensation rate is increased. During the compensation process, the change in melt pressure is monitored in real time, and the compensation range is dynamically adjusted according to the pressure change.

[0122] Specifically, the formula for the compensation amount in the feeding section is:

[0123]

[0124] in, This is the compensation amount for the feeding section. The first compensation coefficient related to materials, For the predicted temperature fluctuation, To compensate for the time, This refers to the heat conduction delay time in the feeding section.

[0125] The formula for the compensation amount in the molten section is:

[0126]

[0127] in, This is the compensation amount for the molten section. This is the material-related second compensation coefficient. For the predicted temperature fluctuation, This is the baseline value for shear rate.

[0128] See Figure 8 As shown, after each production batch, actual temperature trajectory, shear rate, and viscosity deviation data are collected. The parameters of the three-dimensional response surface model are then optimized using the least squares method, specifically including:

[0129] After each production batch is completed, collect the actual temperature trajectory data, shear rate data, and viscosity deviation data obtained during the production process of that batch.

[0130] The parameters of the three-dimensional response surface model are fitted and optimized using the least squares method, and the regression coefficients and correction coefficients in the model are updated.

[0131] After optimization, new recycled plastic particle samples are selected for experimental verification. If the error between the melt apparent viscosity calculated by the model and the experimental measurement value is less than the preset verification threshold, the optimized model will be used as the temperature control calculation model for the next batch of production. If the error is greater than the preset verification threshold, the parameters will be optimized again until the model accuracy meets the requirements.

[0132] Specifically, based on the characteristics of this batch of recycled plastics, the range of values ​​for thermal stability parameters was fixed, and the regression relationship between temperature, shear rate, and viscosity was optimized to avoid fitting confusion caused by adjusting multiple variables simultaneously. The temperature and shear rate data of the fitted group were substituted into the model, and the regression coefficients were adjusted by the least squares method to gradually reduce the deviation between the model-calculated base viscosity and the actual viscosity. The thermal stability parameters of the fitted group were substituted into the model, and the correction coefficients were adjusted to further reduce the deviation between the corrected viscosity and the actual viscosity.

[0133] Each time the parameters are adjusted, the average deviation of the fitted group is calculated. If the deviation does not decrease significantly after three consecutive adjustments, the fitting is paused, and the data distribution blind spots are checked. Data is added before continuing.

[0134] The optimized model is tested using data from the "reserved verification group". The temperature, shear rate, and thermal stability parameters of the reserved group are substituted into the new model, the viscosity value is calculated and compared with the actual value. If the average deviation of the reserved group is ≤2%, the parameter optimization is deemed effective and proceeds to the next experimental verification step. If the deviation is >2%, the model returns to the step-by-step fitting optimization stage, the weights of the regression coefficients are adjusted, and the model is re-optimized.

[0135] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0136] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0137] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for precise temperature control of injection molded article recycling recovery, characterized by, The method comprises the following steps: The composition of the recycled plastic particles is detected online by a near-infrared spectrometer, the composition of the recycled plastic particles is analyzed according to a spectral analysis algorithm, the melt index, the thermal stability parameter and the glass transition temperature are obtained; A temperature gradient monitoring network for the whole process of recycling of injection molded products is constructed by deploying fiber temperature sensors, infrared thermal imagers and melt pressure sensors in each module of a double-screw extruder, including the feeding section, the melting section, the mixing section, the homogenizing section and the die outlet, so as to obtain temperature gradient data; Based on the obtained melt index, thermal stability parameter and glass transition temperature data, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate is established according to the polymer melt rheology theory; Based on the pressure data of the melt pressure sensor, the actual shear rate is calculated by combining the screw geometry parameters and the real-time screw speed of the double-screw extruder, the current temperature feedback value is obtained by the fiber temperature sensor and the infrared thermal imager, the actual shear rate and the temperature feedback value are substituted into the three-dimensional response surface model, and the real-time melt apparent viscosity is calculated; The melt apparent viscosity is compared with the theoretical calculation value of the three-dimensional response surface model, the viscosity deviation is calculated, and when the viscosity deviation is greater than a threshold value, a PID neural network is started to predict the temperature fluctuation in a future time period; Based on the temperature fluctuation, differential temperature compensation is performed on the feeding section and the melting section; After each production batch is completed, the actual temperature trajectory, the shear rate and the viscosity deviation data are collected, and the three-dimensional response surface model parameters are fitted and optimized by the least square method.

2. The method of claim 1, wherein the method is characterized by, The method for detecting the composition of the recycled plastic particles online by the near-infrared spectrometer, analyzing the composition of the recycled plastic particles according to the spectral analysis algorithm and obtaining the melt index, the thermal stability parameter and the glass transition temperature specifically comprises the following steps: Real-time spectrum acquisition of the recycled plastic particles is performed by the near-infrared spectrometer, and a molecular vibration fingerprint spectrum is obtained by synchronously triggering a Raman spectrometer; The characteristic absorption peak intensity of C-H bonding is identified by near-infrared spectrum analysis, the oxidation degree is judged by the carbonyl peak, the C-C skeleton vibration mode is analyzed by Raman spectrum analysis, and the molecular weight distribution is evaluated according to the peak width; When PE / PP blending is detected, the melt index is obtained by separating the overlapping peaks by using the principal component analysis method based on the comparison of the C-H bond absorption peak area ratio with a standard database; The carbonyl index is related to the thermal weight loss data, and the thermal stability parameter threshold value is lowered by 15% for each 10% increase in the oxidation degree; The chain segment flexibility is inversely calculated according to the C-C bond vibration half-width of the Raman spectrum, a weighted summation model of the glass transition temperature of each component is established for ABS blends, and the glass transition temperature is obtained.

3. The method of claim 1, wherein the method is used for precise temperature control of injection molded articles during recycling. The method for constructing the temperature gradient monitoring network for the whole process of recycling of injection molded products by deploying the fiber temperature sensors, the infrared thermal imagers and the melt pressure sensors in each module of the double-screw extruder, and obtaining the temperature gradient data, wherein the modules of the double-screw extruder include the feeding section, the melting section, the mixing section, the homogenizing section and the die outlet specifically comprises the following steps: The fiber temperature sensors, the infrared thermal imagers and the melt pressure sensors are respectively deployed in the feeding section, the melting section, the mixing section, the homogenizing section and the die outlet of the double-screw extruder; Two optical fiber temperature sensors are arranged at a preset distance along the screw axis direction of each module to obtain temperature data at different positions, and an infrared thermal imager captures temperature distribution images in real time, covering the screw area and the inner wall of the cylinder. Through the cooperative deployment of the above sensors, a spatial temperature gradient monitoring network for the whole process of recycling and recycling of injection molded products is constructed, and temperature gradient data and melt pressure data of each module are collected in real time.

4. The method of claim 1, wherein the method is used for precise temperature control of injection molded articles during recycling. Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with the polymer melt rheology theory, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate is established, which specifically includes: Based on the obtained melt index, thermal stability parameters and glass transition temperature data, and combined with the inherent correlation between melt viscosity, temperature and shear rate in the polymer melt rheology theory, the melt apparent viscosity, temperature and shear rate are selected as the core variables, and the inherent correlation is that the higher the temperature and the greater the shear rate; By collecting the apparent viscosity data of the recycled plastic melt at different temperatures and different shear rates, a three-dimensional response surface model of melt apparent viscosity-temperature-shear rate is established by using multiple regression analysis method to fit the mathematical correlation between the core variables. In the model construction process, combined with the influence of thermal stability parameters on melt viscosity, by introducing a thermal stability correction coefficient, the melt viscosity calculation results under different thermal stability conditions are adjusted.

5. The method of claim 1, wherein the method is used for precise temperature control of injection molded articles during recycling. Based on the pressure data of the melt pressure sensor, combined with the screw geometry parameters of the twin-screw extruder and the real-time screw speed, the actual shear rate is calculated, the current temperature feedback value is obtained through the optical fiber temperature sensor and the infrared thermal imager, and the actual shear rate and the temperature feedback value are substituted into the three-dimensional response surface model to calculate the real-time melt apparent viscosity, which specifically includes: Based on the real-time pressure data collected by the melt pressure sensor, combined with the screw geometry parameters of the twin-screw extruder and the real-time screw speed data, the actual shear rate of the melt under the current screw operating state is calculated through the shear rate calculation formula, and the screw geometry parameters include screw channel depth, screw rib width, screw lead and screw diameter. Real-time temperature data of the melt is obtained through the deployed optical fiber temperature sensor, temperature distribution images are obtained through the infrared thermal imager, and the average temperature is extracted as the temperature feedback auxiliary value; The real-time temperature data obtained by the optical fiber temperature sensor and the average temperature extracted by the infrared thermal imager are weighted and fused to obtain the current temperature feedback value; The calculated actual shear rate and temperature feedback value are input into the three-dimensional response surface model to calculate the real-time melt apparent viscosity.

6. The method of claim 1, wherein the method is used for precision temperature control of injection molded articles during recycling. The melt apparent viscosity is compared with the theoretical calculation value of the three-dimensional response surface model, the viscosity deviation is calculated, and when the viscosity deviation is greater than the threshold, the PID neural network is started to predict the temperature fluctuation in the future time period, which specifically includes: The real-time melt apparent viscosity obtained is compared with the theoretical calculation value of the three-dimensional response surface model under the same actual shear rate and temperature feedback value, and the viscosity deviation between the two is calculated. A preset viscosity deviation threshold is determined according to the quality requirements of the injection molded product and the material characteristics of the recycled plastic, and when the calculated viscosity deviation is greater than the preset threshold, the PID neural network model is started; The current viscosity deviation, temperature change rate and pressure change rate are input into the PID neural network model as input parameters, the parameters are converted into signals that can be calculated by the model through the neurons of the input layer, the signals are nonlinearly integrated through the hidden layer, and the temperature fluctuation in the future time period is calculated through the proportional, integral and differential characteristics of the output layer PID.

7. The method of claim 1, wherein the method is used for precision temperature control of injection molded articles during recycling. The differential temperature compensation for the feeding section and the melting section based on the temperature fluctuation specifically includes: Based on the temperature fluctuation predicted by the PID neural network model, combined with the obtained recycled plastic particle thermal stability parameters and temperature gradient data, differential temperature compensation is performed on the feeding section and the melting section of the twin-screw extruder; For the feeding section, a segmented temperature compensation strategy is adopted, the compensation amplitude is increased in the region where the temperature fluctuation is higher than the threshold, the compensation amplitude is reduced in the region where the temperature fluctuation is lower than the threshold, and the compensation temperature of the feeding section is controlled not to exceed the preheating threshold corresponding to the glass transition temperature of the recycled plastic particles; For the melting section, when the thermal stability parameter is lower than the threshold, a slow warming compensation method is used, and when the thermal stability parameter is higher than the threshold, the temperature compensation rate is increased. The melt pressure change is monitored in real time during the compensation process, and the compensation amplitude is dynamically adjusted according to the pressure change.

8. The method of claim 1, wherein the method is used for precision temperature control of injection molded articles during recycling. The actual temperature trajectory, shear rate and viscosity deviation data are collected after each production batch, and the three-dimensional response surface model parameters are fitted and optimized by the least squares method, specifically including: After completing one production batch, the actual temperature trajectory data, shear rate data and viscosity deviation data obtained during the production of the batch are collected; The parameters of the three-dimensional response surface model are fitted and optimized by the least squares method, and the regression coefficients and correction coefficients in the model are updated; After optimization, new recycled plastic particle samples are selected for experimental verification. If the error between the calculated melt apparent viscosity and the experimental measured value is less than the preset verification threshold, the optimized model is used as the temperature control calculation model for the next batch of production. If the error is greater than the preset verification threshold, the parameter optimization is performed again until the model accuracy meets the requirements.

9. A precision temperature control system for injection molded article recycling, for implementing a precision temperature control method for injection molded article recycling according to any one of claims 1-8, characterized in that, It includes: The composition analysis module: the module uses near-infrared spectrometer and Raman spectrometer to analyze the melt index, thermal stability parameter and glass transition temperature of the recycled plastic particles in real time; The temperature gradient monitoring module: the module deploys optical fiber temperature sensors, infrared thermal imagers and melt pressure sensors at key sections of the twin-screw extruder to build a real-time monitoring network for the whole process space temperature gradient and pressure; The rheology modeling module: the module establishes and dynamically maintains a melt apparent viscosity-temperature-shear rate three-dimensional response surface model containing thermal stability correction based on the composition data and polymer melt rheology theory; The viscosity monitoring module: the module calculates the actual shear rate according to the pressure data, screw parameters and rotational speed, and calculates the real-time melt apparent viscosity by combining the rheological model with the multi-source temperature data. Temperature control decision module: the module compares the real-time viscosity with the model theoretical value to calculate the deviation, and starts the PID neural network to predict the future temperature fluctuation when the threshold is exceeded; Temperature compensation module: the module performs differentiated temperature compensation strategies for the feeding section and the melting section according to the predicted fluctuation, the thermal stability parameter and the temperature gradient; Processor: the processor is used for processing the calculation process of each formula and the construction calculation process of each model.

Images