A constant temperature control system for polymer granules based on feedforward compensation

Abstract

The present application relates to the technical field of temperature control, in particular to a high polymer pellet constant temperature control system based on feedforward compensation, the present application obtains material conveying section void fraction spatial distribution, generates unevenness index; based on cumulative shear exposure and unevenness index attenuation degree to establish mapping relationship, correct the material state, and reconstruct dynamic lag time and thermal inertia time constant; according to the dynamic parameters, the space-time of feedforward power trajectory is remodeled, the amplitude limiting and collaborative distribution are carried out combined with the power output dynamic response envelope, the temperature regulation instruction and the predicted heat arrival time window are generated; further collect the material near surface heat flow density change rate characteristic inflection point, compare with the predicted time window, correct the mapping relationship according to the phase deviation, realize the adaptive update of heating control parameters. The present application can improve the high polymer pellet constant temperature control precision, reduce the thermal hysteresis effect, and enhance the stability of continuous processing process.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, specifically to a constant temperature control system for polymer granules based on feedforward compensation. Background Technology

[0002] When polymer granules are kept at a constant temperature in the heating zone, existing feedforward compensation control algorithms usually pre-calculate and output a fixed feedforward power trajectory based on external disturbances. The lag time and thermal inertia time constant on which this trajectory is based are mostly fixed parameters pre-calibrated for the material under ideal uniform conditions. They do not take into account the dynamic changes in the spatial distribution unevenness of the material caused by shearing during the conveying process. This makes it difficult for the delayed position of the feedforward power trajectory and the heating development rhythm to match the actual thermal conduction characteristics of the material, resulting in a deviation between the temperature control command and the actual arrival time of heat.

[0003] Meanwhile, existing feedforward compensation algorithms, when limiting the power trajectory, typically only truncate the excess portion according to the upper limit of the actuator's response capability, without processing the resulting heat gap by extending the time axis or transferring it across regions. This leads to a discrepancy between the total compensated heat and the target total before limiting, affecting control accuracy. Furthermore, existing algorithms are mostly open-loop feedforward or rely on fixed parameter models, lacking a closed-loop mechanism to dynamically correct the aforementioned lag time, thermal inertia time constant, and their inherent mapping relationship based on the actual thermal response characteristics of the material. Once the characteristics of the material batch change, the fixed parameters and mapping relationship on which the algorithm is based cannot be adaptively adjusted, causing the control accuracy to gradually deteriorate over time. Moreover, the calculation links in existing algorithms are relatively independent, and parameter updates lack a cross-linking mechanism, making it difficult to promptly transmit measured feedback to the regeneration process of the feedforward power trajectory.

[0004] To address this, a constant temperature control system for polymer granules based on feedforward compensation is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a constant temperature control system for polymer granules based on feedforward compensation. By detecting the porosity distribution of the material to establish a non-uniformity model, and combining shear attenuation to correct dynamic thermal parameters, the feedforward power trajectory is reshaped, and the control model is corrected by feedback of heat flow characteristics, so as to achieve precise constant temperature control of the polymer granule heating process.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A constant temperature control system for polymer granules based on feedforward compensation, comprising: The porosity field sensing unit acquires the spatial distribution of porosity at the material conveying cross section and generates a non-uniformity index. The shear attenuation correction unit establishes a mapping relationship between the cumulative shear exposure and the attenuation degree of the non-uniformity index, performs attenuation correction on the non-uniformity index to obtain the residual non-uniformity, and reconstructs the dynamic hysteresis time and thermal inertia time constant. The waveform reshaping unit reshapes the feedforward power trajectory in time and space according to the dynamic lag time and thermal inertia time constant, and limits and coordinates the reshaped waveform according to the dynamic response envelope of the power output. The reshaped feedforward power trajectory is output as a temperature control command, and the corresponding predicted heat arrival time window is output at the same time. The heat flow anchor closed-loop unit collects the inflection point timestamp of the near-surface heat flow density change rate characteristic of the material at the corresponding location, compares it with the predicted heat arrival time window to obtain the phase deviation, corrects the mapping relationship accordingly, and updates the residual non-uniformity, dynamic lag time, thermal inertia time constant and temperature control command in conjunction with the data.

[0007] Preferably, the process of obtaining the non-uniformity index is as follows: In the non-heated section upstream of the heating zone in the material conveying channel, detection signals reflecting the local density differences of the material are simultaneously acquired in multiple directions along the circumference of the conveying cross section; the conveying cross section is divided into multiple sub-regions, and based on the initial assumption that the density of each sub-region is the same, the theoretical signal values ​​that should be obtained in each direction under the initial assumption are calculated according to the propagation path of the detection signals in each direction; the theoretical values ​​are compared with the actual detection signal values ​​acquired in the corresponding directions to obtain the deviation in each direction; based on the range of the sub-regions traversed by the propagation path of the detection signals in each direction, the corresponding deviation is back-allocated to each traversed sub-region according to the path coverage ratio, and the density values ​​of each sub-region are adjusted to obtain the spatial distribution of porosity in each sub-region within the conveying cross section; the difference in porosity between the central region and the edge region of the cross section is extracted as the non-uniformity index.

[0008] Preferably, the process of establishing the mapping relationship is as follows: obtain the non-uniformity index of the material at the conveying section under different working conditions, and obtain a reference non-uniformity index for the same batch of materials when the material reaches the inlet position of the heating zone. The degree of difference between the obtained non-uniformity index and the reference non-uniformity index is determined as the attenuation degree under the corresponding working condition. The cumulative shear exposure amount determined by the material conveying drive speed and the corresponding residence time under each working condition is associated with the attenuation degree one by one to establish a mapping relationship. The current residence time is determined based on the relationship between the current conveying rate and the distance between the conveying section and the entrance of the heating zone, and the current cumulative shear exposure is determined. The attenuation degree corresponding to the current cumulative shear exposure is found in the mapping relationship. If the current cumulative shear exposure does not coincide with the corresponding point established in the mapping relationship, the attenuation degree of the adjacent established corresponding point is used as the transition value according to the degree of proximity. According to the attenuation degree obtained by searching or transitioning the value, the non-uniformity index is reduced proportionally to obtain the residual non-uniformity.

[0009] Preferably, the process of obtaining the dynamic lag time and thermal inertia time constant is as follows: based on the type of material being processed, retrieve the pre-calibrated basic lag time and basic thermal inertia time constant; use the residual non-uniformity, basic lag time, and basic thermal inertia time constant as inputs, and substitute them into a pre-established nonlinear correspondence: the larger the residual non-uniformity, the more significant the thickness and distribution of the insulation layer formed by the difference in density inside the material, and the extension correction applied to the basic lag time and basic thermal inertia time constant increases at an accelerating rate; the smaller the residual non-uniformity, the extension correction decreases at a decelerating rate and approaches zero; superimpose the corresponding extension correction onto the basic lag time and basic thermal inertia time constant respectively to obtain the reconstructed dynamic lag time and thermal inertia time constant.

[0010] Preferably, the spatiotemporal reshaping process of the feedforward power trajectory is as follows: The total heat to be compensated is determined based on the external disturbance; the initial power change rate, duration, and peak shape of the feedforward power trajectory at each time node are determined by combining the dynamic lag time and the thermal inertia time constant; a power output dynamic response envelope is obtained by pre-applying a change test to the temperature control execution link and recording the upper limit of the actual achievable power change rate; the initial power change rate required at each time node is compared point-by-point with the power output dynamic response envelope; if the change rate at a time node does not exceed the envelope, the initial power value of the node is retained; if it exceeds the envelope... The output power value of the node is limited according to the upper limit of the envelope, and the difference between the power values ​​before and after the limit is accumulated at the time node as the gap to be compensated. The time axis is tracked backward to the time node where the required rate of change is lower than the upper limit of the envelope. The gap to be compensated is superimposed on the output power value of the subsequent node so that the cumulative value of the total output heat catches up with the total heat before the limit is set, thus completing the limit. If the gap to be compensated is not caught up before the predicted heat arrival time window approaches, the remaining gap is converted into the increment of the output power sequence to the adjacent heating area, thus completing the coordinated allocation. Spatiotemporal reshaping is completed under the condition that the total compensation heat remains unchanged.

[0011] Preferably, the process of obtaining the temperature control command is as follows: the reshaped feedforward power trajectory is segmented according to the control cycle of each heating zone, and the power values ​​falling within the same zone and the same cycle are converted into the target output power of the zone in the cycle. The corresponding target output power is obtained to form the temperature control command. The process of obtaining the predicted heat arrival time window is as follows: taking the corresponding time node in the temperature control command as the starting time, the conveying time required for the material to be conveyed from the heating zone to the preset observation position is determined based on the real-time conveying rate and the axial distance between the heating zone corresponding to the time node and the preset observation position within the heating zone. The dynamic lag time is added to the conveying time to obtain the total delay required from the application of power to the arrival of heat at the preset observation position. The predicted heat arrival time window is determined by superimposing the total delay on the starting time.

[0012] Preferably, the process of correcting the mapping relationship is as follows: based on whether the currently processed material has phase change endothermic characteristics, a characteristic inflection point interval is selected as the characteristic inflection point interval, and the timestamp corresponding to the characteristic inflection point interval is extracted; the characteristic inflection point timestamp is compared with the predicted heat arrival time window to obtain the phase deviation; if the characteristic inflection point timestamp is different from the predicted heat arrival time window, the attenuation degree of the mapping relationship in the region corresponding to the current cumulative shear exposure is determined, and the attenuation degree in the region is adjusted according to the phase deviation amplitude; after the adjustment is completed, the attenuation degree between the region and the adjacent cumulative shear exposure region is processed to obtain the corrected mapping relationship.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention obtains the spatial distribution of the void ratio of the conveying section, constructs a non-uniformity index, and establishes a non-uniformity attenuation mapping relationship by combining the cumulative shear exposure amount. It then quantifies and corrects the state changes of the material before it enters the heating zone, further reconstructs the dynamic lag time and thermal inertia time constant, enabling the control process to perceive the differences in heat transfer inside the material in advance, achieve more accurate thermal response prediction, reduce overshoot, undertemperature and temperature fluctuation problems caused by thermal response delay, and improve the stability of constant temperature control.

[0014] 2. This invention reshapes the feedforward power trajectory in time and space based on dynamic lag time and thermal inertia time constant, and performs amplitude limiting processing in combination with the dynamic response envelope of actual power output. The power change exceeding the response capability is redistributed along the time axis, and at the same time, it provides collaborative compensation to adjacent heating areas when necessary, so that the total compensation heat remains consistent while satisfying the dynamic constraints of the actuator, thereby improving the heat transfer efficiency and heating process control accuracy.

[0015] 3. This invention collects the inflection point timestamps of the near-surface heat flux density change rate characteristics of materials and compares them with the predicted heat arrival time window. Based on the obtained phase deviation, it dynamically corrects the shear attenuation mapping relationship and simultaneously updates the residual non-uniformity, dynamic lag time, thermal inertia time constant, and temperature control commands. This enables the control model to adaptively adjust with changes in the actual heat transfer state, avoiding model mismatch problems in long-term operation and improving the system's continuous control capability and reliability in industrial applications. Attached Figure Description

[0016] Figure 1 A schematic diagram of a constant temperature control system for polymer granules based on feedforward compensation is provided for this invention. Figure 2 This is a schematic diagram of the constant temperature control process for polymer granules provided by the present invention. Figure 3 A schematic diagram of the spatiotemporal reshaping process of the feedforward power trajectory provided by the present invention; Figure 4 This is a schematic diagram of the process for correcting the mapping relationship provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0018] Example 1: The data processing of this system is divided into four levels that are transmitted sequentially and fed back from the end to the middle, corresponding to the porosity field sensing unit, shear attenuation correction unit, waveform reshaping unit and heat flow anchor point closed loop unit, respectively. The following descriptions are based on their technical features.

[0019] Please see Figure 1 and Figure 2This invention provides a constant temperature control system for polymer granules based on feedforward compensation. The technical solution is as follows: a void ratio field sensing unit acquires the spatial distribution of void ratio in the material conveying section and generates a non-uniformity index; a shear attenuation correction unit establishes a mapping relationship between the cumulative shear exposure and the attenuation degree of the non-uniformity index, performs attenuation correction on the non-uniformity index to obtain residual non-uniformity, and reconstructs the dynamic lag time and thermal inertia time constant; a waveform reshaping unit reshapes the feedforward power trajectory spatiotemporally according to the dynamic lag time and thermal inertia time constant, and limits and coordinates the reshaped waveform according to the dynamic response envelope of the power output, outputs the reshaped feedforward power trajectory as a temperature control command, and outputs the corresponding predicted heat arrival time window; a heat flux anchor point closed-loop unit collects the inflection point timestamp of the near-surface heat flux density change rate characteristic of the material at the corresponding location, compares it with the predicted heat arrival time window to obtain the phase deviation, corrects the mapping relationship accordingly, and updates the residual non-uniformity, dynamic lag time, thermal inertia time constant, and temperature control command in linkage.

[0020] Further, the process of obtaining the non-uniformity index is as follows: In the non-heated section upstream of the heating area in the material conveying channel, detection signals reflecting the local density differences of the material are simultaneously acquired in multiple directions along the circumference of the conveying cross section. The channel wall of the non-heated section is made of a material structure that allows the detection signals to penetrate effectively. The conveying cross section is divided into multiple sub-regions. Taking the initial assumption that the density of each sub-region is the same, the theoretical signal value that should be obtained in each direction is calculated according to the propagation path of the detection signal in each direction. The theoretical value is compared with the actual detection signal value acquired in the corresponding direction to obtain the deviation in each direction. According to the range of the sub-regions traversed by the propagation path of the detection signal in each direction, the corresponding deviation is back-allocated to each sub-region according to the path coverage ratio, and the density value of each sub-region is adjusted accordingly. The theoretical value calculation, comparison and back-allocation adjustment are repeated until the deviation in each direction tends to stabilize, and the porosity spatial distribution of each sub-region in the conveying cross section is obtained. The difference in porosity between the central region and the edge region of the cross section is extracted as the non-uniformity index.

[0021] Specifically, in the non-heated section where the porosity field sensing unit is located, the wall of the conveying channel is composed of an inner wear-resistant composite lining and an outer high-strength insulating fiber wound pressure-bearing layer, replacing the original integral metal structure of this section. This section does not involve material heating or melting; it only bears conveying pressure. The aforementioned composite material structure meets the strength requirements while allowing the detection signal applied by the external electrodes to effectively penetrate the material, overcoming the shielding effect of the metal wall. The detection electrodes are arranged in multiple circumferential directions along the conveying cross-section, simultaneously acquiring detection signals from the material passing through the cross-section from each direction. These detection signals are preferably capacitance or dielectric response signals acquired through a multi-electrode array.

[0022] When processing the detection signal, the transmission cross section is first divided into multiple sub-regions along the radial and circumferential directions, with the initial assumption that each sub-region has the same density. Based on the geometric orientation of the propagation path corresponding to the detection signal in each direction within the cross section, the sub-regions through which the path passes in sequence and the relative length proportion within each sub-region are determined. Based on the density value of the initial assumption and the geometric relationship of the path, the theoretical signal value that should be obtained in each direction under this assumption is calculated and compared with the actual measured value in that direction to obtain the deviation in each direction. The deviation in each direction is distributed according to the length proportion of the path through each sub-region in that direction, and then substituted back to the corresponding sub-region and its density value is adjusted. The adjusted value is used as a new round of assumptions, and the above theoretical calculation, comparison and substitution process is repeated.

[0023] The stopping criterion for the above iterations is predetermined during the equipment debugging phase: the above iterations are repeated for calibration samples with known spatial distribution of porosity, and the percentage decrease in the deviation of each iteration relative to the previous iteration is recorded. As the number of iterations increases, this percentage decreases gradually and eventually remains at a consistently low level. The percentage decrease corresponding to this level is set as the preset convergence threshold. In actual operation, once the percentage decrease in deviation between two adjacent iterations is lower than this threshold, convergence is determined, the iteration stops, and the density value of each sub-region at this time is taken as the spatial distribution result of porosity of that cross section. The degree of difference between the density values ​​of the central region and the edge region (the absolute value of the difference between the two) is extracted as the non-uniformity index and transmitted to the shear attenuation correction unit along with the timestamp.

[0024] Further, the process of establishing the mapping relationship is as follows: under various combinations of material conveying drive speed and real-time material conveying rate in the non-heated section, the non-uniformity index of the material at the conveying cross section under the corresponding working condition is obtained respectively. When the material reaches the inlet position of the heating zone, a reference non-uniformity index is obtained for the same batch of material. The degree of difference between the index at the conveying cross section and the reference index is determined as the attenuation degree under the corresponding working condition. The cumulative shear exposure amount determined by the material conveying drive speed and the corresponding residence time under each working condition is associated with its corresponding attenuation degree one by one to establish a mapping relationship between the cumulative shear exposure amount and the attenuation degree; the residual non-uniformity index is determined as follows: The process of obtaining uniformity is as follows: The current material conveying drive speed and the current material conveying rate in the non-heated section are acquired in real time. The current residence time is determined based on the relationship between the conveying rate and the distance between the conveying section and the entrance of the heating area. Both determine the current cumulative shear exposure. The attenuation degree corresponding to the current cumulative shear exposure is searched in the mapping relationship. If the current cumulative shear exposure does not coincide with the corresponding point already established in the mapping relationship, a transitional value is selected based on the proximity of the attenuation degree of adjacent established corresponding points. The non-uniformity index is proportionally reduced according to the attenuation degree obtained from the search or transitional value to obtain the residual non-uniformity.

[0025] Specifically, the mapping relationship is established during the equipment commissioning phase. For multiple combinations of material conveying drive speeds and residence times within the actual operating range of the equipment, a known degree of spatial non-uniformity is artificially created in the calibration materials beforehand, making the original non-uniformity of the materials under this combination controllable and known. The equipment is run under this combination, and the non-uniformity index of the batch is measured at the sampling section. At the same time, an auxiliary sampling device with the same detection principle is temporarily deployed at the entrance of the heating zone to measure the reference non-uniformity index of the distribution of the same batch of materials when it arrives at the entrance using the same reconstruction process. The degree of difference between the two represents the degree to which the original non-uniform distribution of the material under this combination is weakened by the conveying mixing effect during the process of conveying from the sampling section to the entrance, and is recorded as the attenuation degree of this combination.

[0026] The cumulative shear exposure of the material is determined by combining the material conveying drive speed and the residence time of the material between the collection section and the inlet of the heating zone: the higher the speed, the stronger the relative shear fault effect per unit conveying time; the residence time is divided into several continuous time periods, and the instantaneous shear intensity of each time period is represented by the corresponding drive speed; the total amount obtained by accumulating the intensities of each time period in chronological order is the cumulative shear exposure; the longer the residence time, the longer the effect is applied; the cumulative shear exposure is a combination of the intensity and duration of the effect. The cumulative shear exposure of each calibration combination and its corresponding attenuation degree are recorded one by one to form a mapping relationship covering the actual operating range of the equipment; after calibration, the temporary device at the inlet is removed.

[0027] During real-time operation, the current residence time is determined based on the current drive speed and the real-time conveying rate and axial distance of the material between the collection section and the heating zone inlet. These two factors together determine the current cumulative shear exposure. This exposure is then compared to each calibration point in the mapping relationship: if the difference is less than a preset proximity threshold, it is considered an overlap, and the attenuation level at that point is directly used; if it falls between two adjacent calibration points, weights are assigned according to the numerical proximity of the current value to the points on either side, with greater weight for closer values, resulting in a weighted transition value; if it exceeds both ends of the calibrated range, the attenuation level at the boundary point is used. The preset proximity threshold is determined based on the calibration point interval density, taking a smaller proportion of this interval to ensure that most current values ​​are obtained through weighted transition values, avoiding the situation where a threshold that is too wide would incorrectly identify different current values ​​as corresponding to the same attenuation level. When the current cumulative shear exposure falls between two adjacent calibration points, the numerical difference between the current value and the two calibration points is calculated, and the sum of the differences is used as the allocation base. The proportion of the difference between the current value and one calibration point to the allocation base is used as the weight of the attenuation degree at the other calibration point. That is, the closer the current value is to a calibration point, the smaller the proportion of the difference to the allocation base, and the greater the weight of the other calibration point. After obtaining the weights corresponding to the attenuation degree at the two calibration points in this way, the attenuation degrees of the two calibration points are weighted and summed according to their respective weights. The result is the transition value. This allocation method ensures that the weights transition continuously as the current value changes position between the two calibration points. The closer the current value is to a calibration point, the greater the proportion of the attenuation degree of that calibration point in the transition value, ensuring a smooth transition between the transition value and the attenuation degrees already calibrated on both sides without abrupt changes. After obtaining the attenuation degree, the corresponding part is subtracted from the non-uniformity index according to its represented proportion, and the remaining part is the residual non-uniformity.

[0028] Furthermore, the process of obtaining the dynamic lag time and thermal inertia time constant is as follows: Based on the type of material being processed, the pre-calibrated basic lag time and basic thermal inertia time constant are retrieved. These basic lag time and basic thermal inertia time constant characterize the inherent delay characteristics of heat conduction from the inner wall of the heating zone to the interior of the material under conditions of uniform porosity spatial distribution. These inherent delay characteristics are identified through experiments applying a step heating response to the material under standard uniform conditions. The residual non-uniformity, along with the basic lag time and the basic thermal inertia time constant, are used as inputs, and... The pre-established nonlinear correspondence is used: the greater the residual non-uniformity, the more significant the thickness and distribution of the insulation layer formed by the difference in density inside the material, and the greater the extension correction applied to the basic lag time and the basic thermal inertia time constant increases with an accelerating trend; the smaller the residual non-uniformity, the greater the extension correction decreases with a decelerating trend and approaches zero; the corresponding extension correction is superimposed on the basic lag time and the basic thermal inertia time constant respectively to obtain the reconstructed dynamic lag time and thermal inertia time constant, which are determined synchronously based on the same residual non-uniformity and the same nonlinear correspondence.

[0029] Specifically, for each material type, a standard heating power input is applied in advance under the calibration conditions of uniform spatial distribution and zero residual non-uniformity. The pure lag time and heating development time required for the heat to reach the preset observation position are measured and recorded as the basic lag time and basic thermal inertia time constant of the material type, respectively. They are stored by type as the reconstruction benchmark.

[0030] In each combination for obtaining and calibrating the non-uniformity index, in addition to measuring the degree of attenuation, the lag time and heating development time of the material actually reaching the preset observation position under that combination are simultaneously measured. The increment is obtained by comparing it with the basic value of the corresponding material type. Based on the process of obtaining the non-uniformity index, the residual non-uniformity of the combination is calculated from the degree of attenuation of that combination. The residual non-uniformity and the corresponding lag time increment and thermal inertia time constant increment are recorded one by one to form a non-linear correspondence.

[0031] Calibration data shows that both increments exhibit an accelerating growth trend with increasing residual non-uniformity: when residual non-uniformity is low, areas with lower density are dispersed and unconnected, allowing heat to still bypass the dominant heat transfer channels formed by the surrounding denser materials and reach the core of the cross-section. At this stage, the increase in non-uniformity mainly manifests as an increase in the number or range of dispersed areas, with limited obstruction to the heat conduction path. As non-uniformity further increases, areas with lower density become interconnected and form a continuous thermal blocking layer. The dominant heat transfer channels that can be bypassed decrease or even disappear. Subsequently, each increase in non-uniformity directly translates to an increase in the thickness or connectivity of this blocking layer, correspondingly accelerating and strengthening the obstruction to the heat conduction path. The accelerating growth trend of these two increments is precisely the manifestation of this physical process of transitioning from dispersed bypass to continuous blocking in terms of time parameters.

[0032] For the interval distribution of each calibration point in the nonlinear correspondence, according to the aforementioned method of determining the proximity threshold, based on the interval density of the residual non-uniformity values ​​between adjacent calibration points in the relationship, a smaller proportion of the interval is re-selected as the proximity threshold specifically for the relationship. This is used to determine whether the current residual non-uniformity coincides with the calibrated points, avoiding the direct use of the threshold determined for the cumulative shear exposure during the non-uniformity index acquisition process, which would lead to a mismatch with the actual range and calibration density of the residual non-uniformity.

[0033] In real-time operation, using the current residual non-uniformity as input, the corresponding increment is obtained by looking up or transitioning values ​​in the nonlinear correspondence through a proximity threshold judgment method. The lag time increment is then superimposed on the base lag time of the current material type to obtain the reconstructed dynamic lag time. The thermal inertia time constant increment is then superimposed on the base thermal inertia time constant of the material type to obtain the reconstructed thermal inertia time constant. Both are determined collaboratively based on the same residual non-uniformity, the same nonlinear correspondence, and the same base value for the same material type, avoiding mismatch caused by independent adjustments.

[0034] Furthermore, referring to Figure 3The spatiotemporal reshaping process of the feedforward power trajectory is as follows: The total heat required for compensation is determined based on external disturbances; the initial power change rate, duration, and peak shape of the feedforward power trajectory at each time node are determined by combining the dynamic lag time and the thermal inertia time constant; the power output dynamic response envelope is obtained by pre-applying a change test to the temperature control execution link and recording its actual achievable upper limit of power change rate; the required power change rate at each time node in the initial shape is compared point-by-point with the power output dynamic response envelope; if the required change rate at a certain time node does not exceed the envelope, the initial power value of that node is retained; if it exceeds the envelope, the output power value of that node is adjusted according to the envelope. The upper limit of the network is defined, and the difference between the power values ​​before and after the limit is accumulated at the time node as the compensation gap. The time axis is traced backward to the time node where the required rate of change is lower than the upper limit of the envelope. The accumulated compensation gap is added to the output power value of the subsequent node so that the accumulated value of the total output heat catches up with the original total before the limit is set, thus completing the limit. If the compensation gap is not caught up before the predicted heat arrival time window approaches, the remaining gap is converted into an increment of the output power sequence to the adjacent heating area. The adjacent area synchronously increases its output power in the corresponding time period to make up for the gap, thus completing the coordinated allocation. In this way, spatiotemporal reshaping is completed while keeping the total compensation heat unchanged.

[0035] Specifically, during the commissioning phase, compensation power is applied and continuously adjusted for external disturbances of different magnitudes until the controlled temperature returns to stability and does not exceed the preset allowable deviation range. The total area covered by the feedforward power trajectory on the time axis during this process (i.e., the total compensation heat) is recorded, and the corresponding disturbance magnitude is recorded as the compensation relationship. During real-time operation, the required total compensation heat is obtained by looking up or transitioning the value according to the current disturbance magnitude.

[0036] Using the total compensation heat as a constraint, the starting delay position of the initial power trajectory relative to the time of disturbance is determined by combining the dynamic lag time. The climbing speed, duration and peak value are determined by combining the thermal inertia time constant: the larger the time constant, the smoother the trajectory climb, the longer the duration and the lower the peak value, and vice versa. During the adjustment process, the total area covered by the trajectory is always kept equal to the required total compensation heat.

[0037] During the commissioning phase, different power change test commands were applied to the temperature control execution links of each heating zone in advance. The actual maximum power change rate that can be achieved was recorded when the temperature control execution links were at different current power levels. The results were compiled into a power output dynamic response envelope covering the entire power range and stored by region.

[0038] The required rate of change for each time node of the initial trajectory is compared point by point with the upper limit of the envelope under the corresponding region and power level: if it does not exceed the limit, the original value is retained; if it exceeds the limit, the output power of that node is limited to the value corresponding to the upper limit, and the difference is included in the cumulative registered value of the gap to be compensated. For subsequent nodes, it is determined whether their required rate of change is lower than the upper limit of the corresponding envelope, that is, whether there is an unfilled response margin. If so, without exceeding the response margin of that node, the portion of the cumulative registered value of the gap to be compensated that does not exceed the margin is added to the output power of that node, and the registered value is reduced accordingly; this limiting and sequential addition process continues node by node along the time axis.

[0039] Before the predicted heat arrival time window arrives, the cumulative registered value of the compensation gap at this time is compared with the upper limit of the heat that this region can release in the remaining time according to its own envelope: if it does not exceed the upper limit, the heat released by this region will continue to be released in the remaining time; if it exceeds the upper limit, the excess part will be converted into the output power increment of the adjacent heating region in the corresponding time period, and the adjacent region will make up for it synchronously within its own envelope. The above-mentioned limiting, extension and adjacent region coordinated allocation is essentially to extend or transfer the heat deducted by the limiting along the time axis rather than discard it. Therefore, the total heat released by the feedforward power trajectory after reshaping is consistent with the total compensation heat corresponding to the initial trajectory before limiting, only the release rhythm is adjusted accordingly due to the limitation of the actuator response capability.

[0040] Further, the process of obtaining the temperature control command is as follows: based on the real-time conveying rate of the material in the heating area, the axial position of the material in the heating area corresponding to each time node on the feedforward power trajectory after spatiotemporal reshaping is determined, and then the power value of the time node should be output by the specific heating area that the material actually passes through at that moment; the feedforward power trajectory after reshaping is segmented according to the control cycle of each heating area, and the power values ​​falling in the same area and the same cycle are converted into the target output power of the area in the cycle. Each heating area obtains the corresponding target output power accordingly, which together constitute the temperature control command; the process of obtaining the predicted heat arrival time window is as follows: taking the time node corresponding to the temperature control command as the starting time, based on the real-time conveying rate and the axial distance between the heating area corresponding to the time node and the preset observation position in the heating area, the conveying time required for the material to be conveyed from the heating area to the preset observation position is determined; the dynamic lag time is added to the conveying time to obtain the total delay required from the application of power to the arrival of heat at the preset observation position; the total delay is superimposed on the starting time to determine the predicted heat arrival time window.

[0041] Specifically, based on the real-time conveying rate of the material within the heating zone, the actual axial position of the material corresponding to each time node of the reshaped power trajectory is determined, thereby determining the specific heating zone to which the power value of that node belongs. The start and end times of the control cycle for each heating zone are preset and fixed before the equipment is run. When segmenting, the start and end times of each control cycle are used as the dividing points, and the power trajectory is divided into time segments corresponding to each control cycle of each zone. If the material enters an adjacent heating zone from the current heating zone before the end of a certain control cycle, the power trajectory within that control cycle is further divided into two segments belonging to the two heating zones, with the time when the material crosses the boundary between the two zones as the boundary. The segment before crossing the boundary is included in the power value of the original heating zone for that cycle, and the segment after crossing is included in the power value of the corresponding cycle of the adjacent heating zone, so that the power value of each time segment belongs only to the single heating zone in which the material is actually located within that time segment. The trajectory is segmented according to the preset control cycle for each region. Within the same region and cycle, the average power value of each node is taken as the target output power for that region in that cycle. The average value is used instead of just the endpoint values ​​because that region can only output at a single power level within the same cycle. Taking only the endpoint values ​​would cause the actual total heat released in that cycle to deviate from the total amount that the reshaped trajectory should release in that cycle. The average value keeps the two consistent. The target output power obtained sequentially for each region and cycle together constitutes the temperature control command.

[0042] The predicted heat arrival time window starts at a specific time node in the temperature control command. The delivery time is determined by combining the real-time delivery rate with the axial distance between the area of ​​that node and the preset observation position. The dynamic lag time is added to this delivery time to obtain the total delay. The window position is determined by superimposing this total delay on the starting time. The window width is determined based on the thermal inertia time constant: the dynamic lag time represents the pure delay before heat arrival and determines the starting position of the window; the thermal inertia time constant represents the time required for the temperature to rise and unfold after heat arrival, meaning the effect is not concentrated at a certain moment but gradually appears within a certain range. The larger the time constant, the wider the range of manifestation, and the wider the window accordingly. The preset observation position is bound one-to-one with the actual axial position of the heat flow sensing element, ensuring that the predicted results and subsequent measured results point to the same physical location.

[0043] Furthermore, referring to Figure 4The process of obtaining the corrected mapping relationship is as follows: At the preset observation position within the heating area, at a location flush with the inner surface of the conveying channel and not protruding from the material conveying path, continuously acquire near-inner wall heat flux density change information; based on whether the currently processed material exhibits phase change endothermic characteristics, use either the phase change endothermic characteristic interval where the heat flux density change rate changes from fast to slow and then back to fast, or the gradual softening characteristic interval where the heat flux density change rate shows a phased slowdown, as characteristic inflection point intervals, and extract the timestamp corresponding to these characteristic inflection point intervals; compare these characteristic inflection point timestamps with the predicted heat arrival time window to obtain the phase offset. If the inflection point timestamp lags behind the predicted heat arrival time window, it is determined that the attenuation degree of the mapping relationship in the region corresponding to the current cumulative shear exposure is too high, and the attenuation degree in the region is adjusted downward according to the phase deviation amplitude; if the timestamp is earlier than the predicted heat arrival time window, it is determined that the attenuation degree of the mapping relationship in the region corresponding to the current cumulative shear exposure is too low, and the attenuation degree in the region is adjusted upward according to the phase deviation amplitude; after the adjustment, the attenuation degree between the region and the adjacent cumulative shear exposure region is processed to obtain the corrected mapping relationship.

[0044] Specifically, the heat flow sensing element is embedded inside the wall thickness of the channel corresponding to the preset inspection position in the heating area. The surface facing the material is integrated with the channel body and is flush with the inner hole surface, without protruding from the material conveying path. This achieves continuous monitoring of the heat flow density near the inner wall without changing the original fit clearance or introducing mechanical interference.

[0045] For each material type, the rate of change of heat flux density is continuously collected during its normal heating and non-transformation phase, and its fluctuation range is statistically analyzed as the background fluctuation range for that material. Historical heating data of the material is statistically analyzed to identify the typical duration range of instantaneous fluctuations caused by random interference, and the upper limit of this range is taken as a preset duration threshold to filter out instantaneous interference. During real-time monitoring, the rate of change is continuously compared with the background fluctuation range, and intervals exceeding the background fluctuation range by a certain amplitude and with a duration exceeding the preset duration threshold are uniformly determined as characteristic inflection point intervals: for materials mainly undergoing phase change endothermic, the rate of change during phase change shows a significant slowdown from fast to slow and then back to fast, satisfying this criterion; for materials blended from virgin and recycled materials, with crystallization characteristics between phase change endothermic and gradual softening, the rate of change during softening shows a phased slowdown, and as long as the amplitude and duration also satisfy this criterion, it is also identified; the same detection process can adaptively cover the characteristic inflection point morphology of the above-mentioned different materials without the need for pre-classification.

[0046] Extract the timestamp corresponding to the interval and compare it with the predicted heat arrival time window to obtain the phase deviation. If the characteristic inflection point timestamp lags behind the window, the reason is: the attenuation degree corresponding to the current cumulative shear exposure in the mapping relationship is too high, resulting in an underestimation of the calculated residual non-uniformity, which in turn makes the reconstructed dynamic lag time and thermal inertia time constant too small. The feedforward power trajectory generated accordingly completes the reshaping at a faster heat penetration rhythm, making the prediction window earlier than the actual situation; while the actual residual non-uniformity of the material is higher than this underestimation result, the actual thermal resistance is greater, and the time when the heat actually arrives and causes the rate of change to slow down is later than the prediction window, which is manifested as lag. Based on this, it is determined that the attenuation degree value is too high, and it is adjusted downward according to the phase deviation magnitude; conversely, if the timestamp is earlier than the window, the reason is that the attenuation degree value is too low, resulting in an overestimation of the residual non-uniformity, a larger dynamic parameter, and the prediction window lagging behind the actual arrival time. It is determined that the attenuation degree value is too low, and it is adjusted upward according to the deviation magnitude.

[0047] After adjustment, the effective range of the attenuation value is verified: residual non-uniformity is the remaining part after the non-uniformity index is subtracted according to the proportion represented by the attenuation degree. Therefore, this proportion should be within the range that does not cause the residual non-uniformity to be negative, nor does it equal or exceed the value of the non-uniformity index itself. If it exceeds the range, it is limited to the boundary value. After the verification is passed, the adjusted value and the attenuation degree corresponding to the adjacent cumulative shear exposure area are weighted according to their proximity to avoid step-like abrupt changes in the mapping relationship, thus obtaining the corrected mapping relationship.

[0048] Further, the linkage update process is as follows: the corrected mapping relationship is fed back to the process of obtaining the residual nonuniformity; based on the current cumulative shear exposure, the attenuation degree is re-searched in the corrected mapping relationship, and the nonuniformity index is reduced proportionally again to obtain the updated residual nonuniformity; the updated residual nonuniformity is substituted back into the nonlinear correspondence along with the basic lag time and the basic thermal inertia time constant to obtain the updated dynamic lag time and thermal inertia time constant; the updated dynamic lag time and thermal inertia time constant are used again to determine the initial power change rate, duration, and peak shape of the feedforward power trajectory, and the limiting and coordinated allocation are completed sequentially to obtain the updated feedforward power trajectory; based on the real-time transmission rate, the updated feedforward power trajectory is converted into an updated temperature control command, replacing the original temperature control command output, forming a complete loop from phase deviation detection to temperature control command update. The corrected mapping relationship is fed back to the shear attenuation correction unit, replacing the original mapping relationship. The system uses the same lookup table and weighted transition method to re-find or transition the attenuation level in the corrected mapping relationship based on the current cumulative shear exposure, and then proportionally reduces the inhomogeneity index to obtain the updated residual inhomogeneity.

[0049] The updated residual nonuniformity, based on the same nonlinear correspondence and superposition method, is re-obtained or transitionally valued to obtain the corresponding increment, which is then superimposed onto the base lag time and base thermal inertia time constant to obtain the updated dynamic lag time and thermal inertia time constant. Both, based on the same spatiotemporal reshaping process, redetermine the initial trajectory's starting delay position, climb rate, duration, and peak value, and sequentially complete the same set of limiting, extension, and coordinated allocation processes to obtain the updated feedforward power trajectory. This trajectory, based on the same conversion and calculation process, redetermines the region to which each node belongs and averages the results segmented according to the control cycle to obtain the updated temperature control command, which replaces the original command output. Simultaneously, the updated predicted heat arrival time window is redefined for subsequent phase deviation comparison.

[0050] During the aforementioned coordinated update process, the updates to residual non-uniformity, dynamic lag time, thermal inertia time constant, feedforward power trajectory, and temperature control commands simply replace the inputs with the corrected mapping relationships and the resulting downstream values, rather than employing new, unexplained processing methods. The mapping relationship correction is triggered by phase deviation detection, and through the aforementioned step-by-step coordinated updates, it ultimately manifests as a change in the temperature control commands. The system continuously repeats this detection and update process during operation, ensuring that the mapping relationships and downstream parameters are continuously adjusted according to changes in the actual characteristics of the materials.

[0051] Example 2: Based on the collaborative operation of the gap field sensing unit, shear attenuation correction unit, waveform reshaping unit and heat flow anchor point closed loop unit described in Example 1, this system can further be equipped with a mass flow feedforward reference correction unit and an overheating interlock protection unit to improve the temperature control command from two aspects: matching the actual material processing volume and intercepting the risk of local overmelting.

[0052] Furthermore, the mass flow feedforward reference correction unit is used to obtain the deviation ratio between the real-time mass flow rate and the preset reference mass flow rate in the material conveying channel; based on the deviation ratio, the target output power of each time node in the temperature control command output after the spatiotemporal reshaping of the feedforward power trajectory is scaled proportionally, wherein when the real-time mass flow rate is higher than the reference mass flow rate, the target output power is amplified according to the deviation ratio, and when it is lower than the reference mass flow rate, the target output power is reduced according to the deviation ratio; the scaled target output power replaces the original temperature control command output, and the width of the predicted heat arrival time window is corrected synchronously according to the deviation ratio.

[0053] Specifically, the mass flow rate feedforward reference correction unit is located at the downstream metering position of the material conveying channel to acquire the material mass flow rate under the current operating conditions in real time. The reference mass flow rate is determined by conducting multiple trial runs of the currently used conveying drive device at its rated speed and with consistent material types before the system is officially put into operation. The arithmetic mean of the stable section readings of the mass flow rate during each trial run is calculated and stored in the system as the default reference value for that material type and drive device configuration. When the material type or the rated parameters of the conveying drive device change, the reference mass flow rate is recalibrated in the above manner. The mass flow rate feedforward reference correction unit compares the real-time mass flow rate with the reference mass flow rate to obtain a deviation ratio. This deviation ratio is then applied to the target output power at each time node in the temperature control command output by the waveform reshaping unit, causing the target output power to adjust in the same direction as the real-time mass flow rate deviates from the reference mass flow rate. When the deviation in mass flow rate originates from a change in the conveyor drive speed, the material conveying speed changes accordingly. In this case, the deviation ratio is also used to synchronously correct the width of the predicted heat arrival time window used by the heat flow anchor closed-loop unit, ensuring the window width matches the actual conveying time altered by the change in conveying speed. When the deviation in mass flow rate originates from a change in the material accumulation state but the conveyor drive speed remains unchanged, the conveying time required for the material to reach the preset observation position is still based on the real-time conveying rate calculation, and the width of the predicted heat arrival time window is not adjusted. To avoid the adjusted target output power exceeding the actual output capacity of the temperature control execution stage, the mass flow rate feedforward reference correction unit, after scaling, compares the scaling result with the power output dynamic response envelope already acquired in the waveform reshaping unit. If the scaled value exceeds the upper limit of the envelope, the upper limit is applied, and the excess is handled according to the collaborative allocation method.

[0054] Furthermore, the overheating interlock protection unit is used to collect the wall temperature and the surface temperature of the material adjacent to the wall in real time, and calculate the temperature difference gradient between the two; a temperature difference gradient threshold associated with the melting point of the currently processed material is preset; when the temperature difference gradient exceeds the threshold for a preset duration, it is determined that there is a risk of local overmelting, triggering an interlock action, forcibly reducing the target output power in the temperature control command of the corresponding heating area to a preset safe power, until the temperature difference gradient falls back to a preset margin below the threshold, after which the interlock is released and the original output of the temperature control command is restored.

[0055] Specifically, the overheat interlock protection unit is installed in the heating zone to collect the wall temperature and the surface temperature of the material in direct contact with the wall in real time, and calculate the ratio of the temperature difference between the two as a function of their distance, as the temperature gradient. The temperature gradient threshold is determined based on the melting point of the material being processed, combined with the historical statistical range of the temperature difference between the wall and the material surface under normal process conditions in the heating zone. During equipment commissioning, for this material type, the fluctuation range of the temperature gradient is continuously recorded under normal process parameters that do not trigger local overmelting. The upper boundary of this range, with a safety margin, is taken as the temperature gradient threshold. The preset time required to determine the risk of local overmelting is determined based on the shortest time required for the material to reach its melting point from a localized temperature rise when the temperature gradient exceeds the threshold, minus the power response delay of the temperature control actuator. Both the shortest time and the response delay are obtained through heating test records during equipment commissioning. After the interlock is triggered, the overheat interlock protection unit will forcibly reduce the target output power in the temperature control command of the corresponding heating area to the preset safe power. The preset safe power is determined by gradually reducing the power of the heating area during the equipment commissioning phase and recording the results to ensure that the temperature of the area can still be maintained within the minimum power level allowed by the temperature control command. The interlock is released and the original output of the temperature control command is restored after the temperature difference gradient falls below the threshold and maintains the preset margin.

[0056] Furthermore, after the interlock action is released, the heat gap caused by the forced reduction of the target output power during the interlock period is converted into an incremental output power sequence to be compensated for in a coordinated allocation manner. However, the premise for this compensation action is that the system determines, through a safety assessment model, that the real-time wall temperature and material state of the current adjacent heating areas have sufficient safe heat absorption margin. If the assessment determines that there is a risk of cascading overheating, the heat gap is actively discarded and forced compensation is prohibited, so as to establish the control principle of prioritizing safety protection. At the same time, the duration of the interlock trigger is associated with the cumulative shear exposure of the corresponding area and recorded. In the next mapping relationship correction process, the attenuation degree of the interval where the cumulative shear exposure is located is given a correction weight related to the interlock trigger frequency.

[0057] Specifically, after the interlock is released, the overheat interlock protection unit compensates for the heat gap caused by the forced reduction of the target output power during the interlock period by converting it into an increment of the output power sequence to the adjacent heating area according to the collaborative distribution method in the waveform reshaping unit. At the same time, it establishes a correlation record between the duration of the current interlock trigger and the cumulative shear exposure of the corresponding area at the time of triggering. When the heat flow anchor closed-loop unit corrects the mapping relationship next time, it increases the correction weight of the attenuation degree of the cumulative shear exposure interval according to the preset grade interval of the historical interlock trigger frequency of the area. The higher the grade of the trigger frequency, the larger the corresponding correction weight increment, so that the attenuation degree estimation of the mapping relationship in the interval gradually approaches the actual attenuation level under the operating condition.

[0058] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A constant temperature control system for polymer granules based on feedforward compensation, characterized in that, include: The porosity field sensing unit acquires the spatial distribution of porosity at the material conveying cross section and generates a non-uniformity index. The shear attenuation correction unit establishes a mapping relationship between the cumulative shear exposure and the attenuation degree of the non-uniformity index, performs attenuation correction on the non-uniformity index to obtain the residual non-uniformity, and reconstructs the dynamic hysteresis time and thermal inertia time constant. The waveform reshaping unit reshapes the feedforward power trajectory in time and space according to the dynamic lag time and thermal inertia time constant, and limits and coordinates the reshaped waveform according to the dynamic response envelope of the power output. The reshaped feedforward power trajectory is output as a temperature control command, and the corresponding predicted heat arrival time window is output at the same time. The heat flow anchor closed-loop unit collects the inflection point timestamp of the near-surface heat flow density change rate characteristic of the material at the corresponding location, compares it with the predicted heat arrival time window to obtain the phase deviation, corrects the mapping relationship accordingly, and updates the residual non-uniformity, dynamic lag time, thermal inertia time constant and temperature control command in conjunction with the data.

2. The constant temperature control system for polymer granules based on feedforward compensation according to claim 1, characterized in that: The process of obtaining the non-uniformity index is as follows: In the non-heated section upstream of the heating zone in the material conveying channel, detection signals reflecting the local density differences of the material are simultaneously acquired in multiple directions along the circumference of the conveying cross section; the conveying cross section is divided into multiple sub-regions, and based on the initial assumption that the density of each sub-region is the same, the theoretical signal values ​​that should be obtained in each direction under the initial assumption are calculated according to the propagation path of the detection signals in each direction; the theoretical values ​​are compared with the actual detection signal values ​​obtained in the corresponding directions to obtain the deviation in each direction; Based on the range of sub-regions traversed by the propagation path of the detection signals in each direction, the corresponding deviation is back-allocated to each sub-region according to the path coverage ratio, and the density value of each sub-region is adjusted to obtain the spatial distribution of porosity of each sub-region within the conveying cross section; the degree of difference in porosity between the central region and the edge region of the cross section is extracted as the non-uniformity index.

3. The constant temperature control system for polymer granules based on feedforward compensation according to claim 2, characterized in that: The process of establishing the mapping relationship is as follows: obtain the non-uniformity index of the material at the conveying section under different working conditions, and obtain a reference non-uniformity index for the same batch of materials when the material reaches the inlet position of the heating zone. The degree of difference between the non-uniformity index and the reference non-uniformity index is determined as the attenuation degree under the corresponding working condition. The cumulative shear exposure amount determined by the material conveying drive speed and the corresponding residence time under each working condition is associated with the attenuation degree one by one to establish a mapping relationship. The current residence time is determined based on the relationship between the current conveying rate and the distance from the conveying section to the entrance of the heating zone; the current cumulative shear exposure is then determined. The attenuation level corresponding to the current cumulative shear exposure is found in the mapping relationship. If the current cumulative shear exposure does not coincide with the corresponding point established in the mapping relationship, the attenuation level of the adjacent established corresponding point is used as the transition value according to the degree of proximity. The non-uniformity index is reduced proportionally according to the attenuation level obtained by searching or transitioning the value to obtain the residual non-uniformity.

4. The constant temperature control system for polymer granules based on feedforward compensation according to claim 3, characterized in that: The process of obtaining the dynamic lag time and thermal inertia time constant is as follows: Based on the type of material being processed, retrieve the pre-calibrated basic lag time and basic thermal inertia time constant; use the residual non-uniformity, basic lag time, and basic thermal inertia time constant as inputs, and substitute them into the pre-established nonlinear correspondence: the larger the residual non-uniformity, the more significant the thickness and distribution of the insulation layer formed by the difference in density inside the material, and the extension correction applied to the basic lag time and basic thermal inertia time constant increases at an accelerating rate; the smaller the residual non-uniformity, the extension correction decreases at a decelerating rate and approaches zero; superimpose the corresponding extension correction onto the basic lag time and basic thermal inertia time constant respectively to obtain the reconstructed dynamic lag time and thermal inertia time constant.

5. The constant temperature control system for polymer granules based on feedforward compensation according to claim 4, characterized in that: The spatiotemporal reshaping process of the feedforward power trajectory is as follows: the total heat to be compensated is determined according to the external disturbance, and the initial power change rate, duration and peak shape of the feedforward power trajectory at each time node are determined by combining the dynamic lag time and the thermal inertia time constant. By applying a change test to the temperature control execution link in advance and recording the upper limit of the actual achievable power change rate, the power output dynamic response envelope is obtained; the initial power change rate required at each time node is compared with the power output dynamic response envelope point by point, and if the change rate at each time node does not exceed the envelope, the initial power value of the node is retained. If the output power of a node exceeds the envelope, the output power value of the node is limited according to the upper limit of the envelope, and the difference between the power values ​​before and after the limit is accumulated at the time node as the gap to be compensated. The time axis is traced backward to the time node where the required rate of change is lower than the upper limit of the envelope, and the gap to be compensated is added to the output power value of the subsequent node so that the cumulative value of the total output heat is equal to the total heat before the limit is set, thus completing the limit. If the gap to be compensated is not caught up before the predicted heat arrival time window approaches, the remaining gap will be converted into an increment of the power sequence output to the adjacent heating area to complete the coordinated allocation; the spatiotemporal reshaping will be completed while the total compensation heat remains unchanged.

6. The constant temperature control system for polymer granules based on feedforward compensation according to claim 5, characterized in that: The process of obtaining the temperature control command is as follows: the reshaped feedforward power trajectory is segmented according to the control cycle of each heating zone, and the power values ​​falling within the same zone and the same cycle are converted into the target output power of the zone in the cycle. The corresponding target output power is obtained to form the temperature control command. The process of obtaining the predicted heat arrival time window is as follows: taking the corresponding time node in the temperature control command as the starting time, the conveying time required for the material to be conveyed from the heating zone to the preset observation position is determined based on the real-time conveying rate and the axial distance between the heating zone corresponding to the time node and the preset observation position within the heating zone. The dynamic lag time is added to the conveying time to obtain the total delay required from the application of power to the arrival of heat at the preset observation position. The predicted heat arrival time window is determined by superimposing the total delay on the starting time.

7. The constant temperature control system for polymer granules based on feedforward compensation according to claim 1, characterized in that: The process of correcting the mapping relationship is as follows: based on whether the material being processed has phase change endothermic characteristics, a characteristic inflection point interval is used as the characteristic inflection point interval, and the timestamp corresponding to the characteristic inflection point interval is extracted; the characteristic inflection point timestamp is compared with the predicted heat arrival time window to obtain the phase deviation; If the feature inflection point timestamp is different from the predicted heat arrival time window, determine the attenuation level of the mapping relationship in the region corresponding to the current cumulative shear exposure, and adjust the attenuation level in the region according to the phase deviation amplitude. After adjustment, the attenuation degree between the region and the adjacent cumulative shear exposure region is processed to obtain the corrected mapping relationship.

Images