Preheating control method and system for polyurethane raw material

Abstract

The invention relates to the technical field of industrial production control, in particular to a polyurethane raw material preheating control method and system. The method comprises the following steps: acquiring tank body temperature, environment temperature, batch temperature rise, tank body structure and raw material liquid data, constructing a thermal resistance-thermal capacitance series-parallel model to represent system thermal inertia characteristics, then identifying an initial value drift resistance section, a main temperature rise section and a disturbance absorption section in a segmented manner, and implementing power back-stepping based on segmented data, so as to obtain the thermal inertia characteristic of the system. Dynamically generating a heating power track; and a heating track is replanned in combination with real-time change data, so that precise modeling and self-adaptive control of a temperature rise process are realized, and the temperature control precision and energy efficiency response capability of the preheating system under different raw material batches and different working conditions are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of industrial production control technology, and in particular to a method and system for controlling the preheating of polyurethane raw materials. Background Technology

[0002] In the production of polyurethane materials, the preheating of raw materials is a crucial step in ensuring the stability of subsequent mixing, reaction, and foaming quality. Traditional preheating control methods often rely on a fixed temperature setpoint combined with simple PID feedback control. However, when faced with differences in thermal inertia between different batches of materials, fluctuations in ambient temperature, and nonlinear heating characteristics, problems such as overshoot, reaction delay, or high energy consumption often occur, failing to balance temperature control accuracy and energy efficiency.

[0003] In scenarios with complex tank structures and significant differences in the specific heat of raw materials, traditional strategies struggle to accurately describe the system's thermal response characteristics, leading to a mismatch between power settings and temperature trajectories. This reduces the consistency of production cycle time and equipment thermal efficiency. Furthermore, existing methods are ill-suited for adjusting preheating strategies in the event of raw material replenishment, ambient temperature drops, or actuator performance fluctuations, exhibiting a certain robustness bottleneck. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method and system for controlling the preheating of polyurethane raw materials, thereby resolving at least one of the aforementioned technical problems.

[0005] This application provides a method for controlling the preheating of polyurethane raw materials, including the following steps: Step S1: Obtain tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; construct a thermal resistance-thermal capacity series-parallel model based on the tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data to obtain the thermal resistance-thermal capacity series-parallel model. Step S2: Perform initial value drift resistance processing on the thermal resistance-capacity series-parallel model to obtain initial value drift resistance data; perform main heating segment processing on the thermal resistance-capacity series-parallel model to obtain main heating segment data; perform disturbance absorption segment processing on the thermal resistance-capacity series-parallel model to obtain disturbance absorption segment data. Step S3: Based on the data of the anti-initial-value drift section, the main heating section, and the disturbance absorption section, perform power back-calculation to obtain the heating power data; Step S4: Obtain real-time change data, and perform trajectory changes based on heating power data and real-time change data to obtain heating change data.

[0006] This invention constructs a thermal resistance-thermal capacity series-parallel model that integrates tank temperature, ambient temperature, historical batch temperature rise data, and physical structural parameters. This model accurately represents the thermal inertia characteristics during the raw material preheating process, improving the fitting accuracy of the temperature rise response. An anti-initial-value drift segment is employed to effectively compensate for initial temperature uncertainty and system hysteresis, enhancing the model's stability during the preheating start-up phase. Through main heating segment modeling and disturbance absorption segment processing, precise control of the target temperature rise trajectory and dynamic adaptation to mid-process disturbances are achieved. Power back-calculation based on this segmented trajectory allows for the acquisition of heating power data matching the physical process, ensuring coordinated optimization of energy consumption control and heating timeliness. Combining real-time change data with trajectory modification enhances the system's rapid response to external disturbances and dynamic production adjustments, thereby achieving high-precision, highly robust intelligent control of the polyurethane raw material preheating process.

[0007] Preferably, the thermal resistance-thermal capacity series-parallel model is constructed as follows: Based on tank data and raw material liquid data, a heat capacity-thermal resistance network model is constructed to obtain the heat capacity-thermal resistance network model. Based on the tank temperature data and the liquid temperature data in the raw material liquid data, the state space model data of the heat capacity-thermal resistance network model is processed to obtain the state space model data. The tank temperature data, ambient temperature data, and batch temperature rise data are labeled on the state-space model data to obtain the thermal resistance-thermal capacity series-parallel model.

[0008] This invention utilizes a thermal capacity-thermal resistance network model to model tank structural parameters and raw material liquid thermal properties in a series-parallel manner. This enables structured analysis of heat conduction paths and heat storage capacity, enhancing the physical representation of the heat exchange process. By combining tank temperature and liquid temperature data with state-space model processing, the model acquires dynamic response characteristics and time-series prediction capabilities, helping to capture the evolution of states at each node during temperature rise. Annotated training of the state-space model using batch temperature rise data and ambient temperature data not only enhances the model's adaptability to different operating scenarios but also improves the accuracy of identifying trends in thermal inertia. This establishes a thermal resistance-thermal capacity series-parallel model that integrates physical constraints, dynamic modeling capabilities, and scenario adaptability, providing a high-quality modeling foundation for segmented control and power planning.

[0009] Preferably, the anti-initial-value drift section processing specifically includes: Initial drift data is obtained by performing initial drift discrimination based on the thermal resistance-thermal capacity series-parallel model. The initial drift data is subjected to temperature rise rate deviation activation determination to obtain drift activation data; The drift activation data is processed to resist initial value drift segmentation to obtain preliminary segment data; The initial section data is dynamically corrected by a gradual increase phase to obtain the section data resistant to initial value drift.

[0010] This invention employs initial drift detection in the thermal resistance-thermal capacity series-parallel model to identify response drift caused by deviations in the initial temperature measurement, lag due to tank thermal inertia, or unstable heat exchange during startup, ensuring the model reflects the true heating state. Temperature rise rate deviation activation detection accurately identifies sections where drift intensity exceeds the normal range, triggering differentiated compensation strategies. Anti-initial drift section processing reconstructs the temperature rise trajectory within the activated interval, eliminating the cumulative impact of early model deviations and preventing erroneous trajectories from being used for power back-calculation. Dynamic feedback correction during the gradual rise phase ensures a smooth transition in model output, effectively improving the response stability and reliability of power planning in the early stages of the preheating process, and enhancing the accuracy and robustness of the entire temperature control process.

[0011] Preferably, the temperature rise rate deviation activation determination specifically includes: A rate offset curve is constructed from the initial drift data to obtain the rate offset curve data; Slope activation data is obtained by performing slope activation judgment based on the rate offset curve data. Threshold surface triggering is performed based on the activation data to obtain drift activation data.

[0012] This invention constructs a rate offset curve from the initial drift data, dynamically extracting the actual rate change trend during the temperature rise process and quantifying the difference with the theoretical temperature rise trajectory, thereby accurately representing the response offset behavior in the initial stage. Based on this rate offset curve, the system can identify abnormally steep points or hysteretic flat sections in the rate change in real time through slope activation judgment, thus achieving early perception of potential drift situations. By activating a data-driven threshold surface triggering mechanism, compared with the traditional fixed threshold judgment method, this method can construct dynamic activation conditions by combining multi-dimensional features such as slope amplitude and duration, achieving adaptive triggering of temperature rise deviation modes and effectively avoiding misjudgment or omission. The above strategies enhance the system's sensitivity and discrimination accuracy to temperature control deviations in the startup stage, providing highly reliable and high-resolution input for drift segment modeling and feedback correction, thereby improving the overall temperature control system's adaptability to initial state anomalies and the stability of the temperature rise trajectory.

[0013] Preferably, the anti-initial-value drift section processing specifically includes: The drift activation data and the preset prior thermal inertia data are compared differentially to obtain the drift error data; The disturbance intensity integral data is obtained by performing disturbance intensity integral calculation on the drift error data; The drift risk interval data is divided into drift risk intervals based on the integral data of disturbance intensity. Based on the drift risk interval data, anti-drift correction is performed on the drift activation data to obtain preliminary segment data.

[0014] This invention constructs a personalized drift error curve that conforms to the current equipment structure and raw material thermal properties by differentially comparing drift activation data with preset prior thermal inertia data, achieving precise quantification of the magnitude and direction of the initial stage deviation. Through disturbance intensity integral calculation, not only the instantaneous extreme value of the drift error is considered, but also the cumulative impact of its duration on system stability is integrated, thus obtaining a more comprehensive drift intensity evaluation index. By dividing the drift risk interval, different levels of offset states can be dynamically segmented for management, clearly defining the key sections requiring correction and avoiding misadjustment of the normal response interval. Anti-drift correction based on the risk interval ensures the smoothness and consistency of the temperature rise trajectory during the startup phase, providing reliable initial data support for energy efficiency control and trajectory planning during the heating phase. This invention effectively improves the adaptability of the model response under non-ideal initial conditions and enhances the robustness and control accuracy of the temperature control system under thermal inertia conditions.

[0015] Preferably, the main heating section is specifically processed as follows: The boundary of the main heating section is determined for the thermal resistance-thermal capacity series-parallel model, and the boundary data of the main heating section are obtained. Based on the boundary data of the main heating section, the shortest heating time is calculated by back-calculating the thermal resistance-thermal capacity series-parallel model to obtain the shortest heating time data. Based on the shortest heating time data, the constant temperature rise rate slope of the thermal resistance-thermal capacity series-parallel model is calculated to obtain constant temperature rise rate slope data. The main heating segment function is generated based on the slope of the constant temperature rise rate, and the main heating segment data is obtained.

[0016] This invention, by determining the boundary of the main heating segment using a thermal resistance-heat capacity series-parallel model, clarifies the start and end conditions for the system to enter the steady-state heating range, avoiding interference from initial fluctuations and terminal disturbances on the temperature control strategy and improving the phased accuracy of trajectory planning. By back-calculating the shortest heating time using boundary conditions, it not only combines the tank's thermal inertia and the raw material's heat capacity to assess the heating limit efficiency but also provides time constraints for power allocation. Calculating the constant temperature rise rate slope based on the shortest heating time ensures a dynamic balance between power output and thermal response, avoiding the risk of thermal shock or overheating caused by sudden power changes. By driving the generation of the main heating segment function using slope parameters, the temperature control trajectory is transformed from physical modeling to functional expression, giving the heating process adjustable, predictable, and schedulable characteristics. This enhances the system's response efficiency and energy consumption control capabilities during critical heating stages, providing strong model support for efficient and stable preheating of polyurethane raw materials within a specific time window.

[0017] Preferably, the disturbance absorption section processing specifically includes: The absorption section trigger judgment is performed on the thermal resistance-thermal capacity series-parallel model to obtain the absorption section data; The time window for the disturbance absorption section is determined based on the absorption section data, and the time window data for the disturbance absorption section is obtained. The thermal resistance-thermal capacity series-parallel model is processed with a slow convergence structure based on the time window data of the disturbance absorption segment to obtain the disturbance absorption segment data.

[0018] This invention utilizes absorption segment triggering judgment in the thermal resistance-thermal capacity series-parallel model to identify response deviations caused by environmental fluctuations, batch differences, or external power disturbances during the heating process, achieving dynamic perception and boundary judgment of disturbance events. By determining the time window of the disturbance absorption segment using absorption segment data, it considers not only the start and end points of the disturbance but also the hysteresis and buffering capacity of the system's thermal response, making the disturbance handling process time-adaptive and response-redundant. The use of a slow-convergence structure to stage-controlled the thermal capacity-thermal resistance coupling channel effectively suppresses slope abrupt changes or trajectory nonlinear fluctuations caused by disturbances, allowing the temperature rise process to gradually return to the preset trajectory. This enhances the model's recovery capability against unsteady-state disturbances, improves the robustness and fault tolerance of trajectory planning under real-world conditions, and helps achieve rapid absorption of local disturbances and continuous maintenance of global stability during the preheating process of polyurethane raw materials.

[0019] Preferably, step S3 specifically includes: Trajectory stitching is performed based on the data from the anti-initial-value drift section, the main heating section, and the disturbance absorption section to obtain trajectory stitching data. Temperature derivatives are calculated from the trajectory stitched data to obtain temperature derivative data; Heating power data is obtained by back-calculating the power based on the temperature derivative data and the thermal resistance-thermal capacity series-parallel model.

[0020] This invention employs trajectory stitching by combining data from the initial value drift segment, the main heating segment, and the disturbance absorption segment. This allows for the construction of a complete, smooth, and physically consistent heating target trajectory while ensuring the continuity of temperature control at each stage. This effectively avoids power fluctuations or thermal inertia mismatches caused by segment disconnections. The temperature derivative calculation method is used to perform first-order dynamic extraction on the stitched trajectory, representing the rate of temperature change and providing a time-series signal reflecting the intensity of the heating demand. Combining derivative information with a thermal resistance-thermal capacity series-parallel model for power back-calculation not only achieves a dynamic mapping from the desired temperature behavior to the physical heating input but also adjusts the power output curve according to the system's actual thermal response capability. This ensures that the heating process meets efficiency requirements while avoiding energy overload and thermal shock risks. This invention improves the system's interpretability and execution consistency regarding heating behavior, enhances the matching accuracy between power planning and actual thermal inertia, and provides solid support for high-performance temperature control strategies.

[0021] Preferably, step S4 specifically includes: Get real-time change data; The trajectory change trigger conditions are determined based on real-time change data to obtain trajectory change data; The remaining trajectory segment is trimmed based on the trajectory change data and heating power data to obtain the remaining trajectory segment data; A change window is generated based on the remaining segment data of the trajectory, and the change window data is obtained. The trajectory function is replanned based on the changed window data to obtain the heating change data.

[0022] This invention, by acquiring real-time change data, can promptly perceive dynamic demand changes caused by fluctuations in environmental conditions, changes in raw material status, or adjustments in upper-level control strategies, ensuring the system's ability to respond to external inputs in real time. Through trajectory change trigger condition judgment, it can accurately determine whether the trajectory needs reconstruction based on multi-dimensional indicators such as temperature deviation, rate mutation, or energy consumption threshold, avoiding frequent interference with stable system operation due to minor disturbances. The trajectory remaining segment trimming mechanism can perform phased reorganization of the current heating process, adjusting only the incomplete parts while preserving the continuity and stability of existing heating data. Generating a change window and combining it with the current power state and thermal response model, the trajectory function is replanned, enabling the system to quickly switch to a new heating target trajectory, ensuring dynamic consistency between energy consumption control, temperature rise rhythm, and target scheduling. This enhances the adaptability and robustness of the heating process to non-ideal operating conditions, improving the intelligent control capability and practicality of the entire temperature control system in application scenarios.

[0023] Preferably, this application also provides a polyurethane raw material preheating control system for executing the polyurethane raw material preheating control method described above, the polyurethane raw material preheating control system comprising: The thermal resistance-heat capacity modeling module is used to acquire tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; and to construct a thermal resistance-heat capacity series-parallel model based on the tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data to obtain the thermal resistance-heat capacity series-parallel model. The segmented response identification and processing module is used to process the anti-initial-value drift segment of the thermal resistance-thermal-capacity series-parallel model to obtain the anti-initial-value drift segment data; process the main heating segment of the thermal resistance-thermal-capacity series-parallel model to obtain the main heating segment data; and process the disturbance absorption segment of the thermal resistance-thermal-capacity series-parallel model to obtain the disturbance absorption segment data. The reverse power calculation module is used to back-calculate the heating power data based on the data of the anti-initial drift section, the main heating section, and the disturbance absorption section. The heating trajectory change control module is used to acquire real-time change data and change the trajectory based on the heating power data and the real-time change data to obtain heating change data.

[0024] The beneficial effects of this invention are as follows: By constructing a series-parallel model of thermal resistance and thermal capacity, multi-source fusion modeling of tank structural parameters, raw material thermal properties, ambient temperature, and historical batch temperature rise behavior is achieved. This accurately characterizes the thermal inertial response characteristics of each stage during raw material preheating, providing a high-fit physical basis for dynamic temperature control. Based on this, the phased introduction of anti-initial-value drift processing, main heating segment modeling, and disturbance absorption mechanisms not only effectively suppresses the impact of temperature drift during the start-up phase on subsequent control accuracy but also ensures the stability and timeliness of the heating process through slope back-calculation and boundary reconstruction. Simultaneously, the disturbance absorption segment processing can achieve gradual convergence adjustment when the tank is disturbed or raw material properties fluctuate, enhancing the system's adaptability to non-ideal operating conditions. By uniformly splicing the above trajectory segments and calculating the temperature derivative, combined with power back-calculation using the thermal model, a high-precision power output sequence conforming to the system's thermal dynamic characteristics can be obtained, ensuring a reasonable distribution of energy consumption. Through real-time trajectory changes, it supports the reconstruction of control targets and power adjustment under sudden scenarios, improving the robustness and intelligent response capability of the temperature control system and forming a closed-loop optimized control system. Attached Figure Description

[0025] Other features, objects, and advantages of this application will become more apparent from the following detailed description of the non-limiting embodiments, taken with reference to the accompanying drawings: Figure 1 A flowchart illustrating the steps of a polyurethane raw material preheating control method according to an embodiment is shown. Figure 2 A flowchart illustrating the steps of a method for constructing a thermal resistance-thermal capacity series-parallel model according to an embodiment is shown. Figure 3 A flowchart illustrating the steps of an embodiment of a method for handling anti-initial value drift sections is shown. Figure 4 A flowchart illustrating the steps of a heating power back-calculation method according to one embodiment is shown. Figure 5 A flowchart illustrating the steps of a heating trajectory modification method according to an embodiment is shown. Detailed Implementation

[0026] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0027] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. Functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0028] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] Please see Figures 1 to 5 This application provides a method for controlling the preheating of polyurethane raw materials, comprising the following steps: Step S1: Obtain tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; construct a thermal resistance-thermal capacity series-parallel model based on the tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data to obtain the thermal resistance-thermal capacity series-parallel model. Specifically, the system collects multi-source thermophysical data related to the heating process of polyurethane raw materials, including data on tank temperature changes over time; ambient temperature data; initial temperature data of the raw material liquid; historical temperature rise curve data for multiple batches; tank structural parameters, including tank wall thickness, thermal conductivity, and tank heat capacity; and data on the heat capacity and thermal conductivity of the raw material liquid. Based on the heat transfer path during the tank heating process, the tank wall layer, the raw material liquid layer, and the outer layer in contact with air are modeled in layers according to the heat flux direction, establishing an equivalent thermal resistance and heat capacity series-parallel network structure. This includes equivalent thermal resistance calculation (i.e., by dividing the layer thickness by the product of thermal conductivity and heat exchange area) and equivalent heat capacity (i.e., mass multiplied by specific heat capacity). Spatially, this forms a heat capacity-thermal resistance series-parallel combined network, realistically reflecting the thermal inertia characteristics of the tank and raw material liquid during dynamic heating. Based on the heat balance equation, the above thermal network structure is mathematically modeled and transformed into a standard state-space equation form: the rate of change of each temperature node (i.e., the derivative of temperature with time) is jointly determined by the system matrix of the heat capacity-thermal resistance structure and the heating power input. Formally, the temperature change rate can be expressed as: Temperature Change Rate = System State Matrix × Current Temperature State + Control Matrix × Heating Power Input. Here, the system state matrix is ​​determined by the topology of the thermal resistance-heat capacity network; the heating power input represents the actual output behavior of the heater. To ensure the constructed thermal resistance-heat capacity state-space model has good practical representativeness, historical batch temperature rise response curves are used as supervisory data for model fitting training or parameter calibration. By minimizing the mean square error between the model's output temperature and historical measured temperature rise curves, the model achieves adaptive adjustment to reproduce the thermal response characteristics of the tank-raw material system.

[0030] Step S2: Perform initial value drift resistance processing on the thermal resistance-capacity series-parallel model to obtain initial value drift resistance data; perform main heating segment processing on the thermal resistance-capacity series-parallel model to obtain main heating segment data; perform disturbance absorption segment processing on the thermal resistance-capacity series-parallel model to obtain disturbance absorption segment data. Specifically, the system compares the slope changes of the actual temperature rise curve (obtained based on the actual measured temperature) and the model prediction curve (obtained by multivariate regression fitting of the curves in the model) to calculate the real-time temperature rise rate deviation. When the absolute value of the difference between the measured temperature rise rate and the model prediction rate per unit time exceeds a preset threshold, it is considered an initial value drift. The system integrates the rate deviation during the initial heating period to obtain the disturbance intensity integral. Based on the integral value, the system classifies the drift level into three levels—high, medium, and low—through clustering calculation or a preset threshold, and then determines the corresponding response correction intensity. To prevent excessively rapid heating during the initial heating process from causing dynamic system deviation, the system applies a slope limiting strategy to the initial heating trajectory. This involves comparing the heating slope calculated by the model with the set maximum starting slope, taking the smaller one as the actual heating rate, and obtaining the anti-initial value drift segment data. This processing helps prevent thermal response deviation caused by large initial temperature differences or model errors during the start-up phase, ensuring a smooth transition of the system in the initial heating stage. By limiting the actual temperature rise rate, the temperature control system can be prevented from over-responding or overshooting.

[0031] After identifying the endpoint of the anti-drift phase (i.e., rate recovery stabilization), the system sets the starting point of the main heating phase; the endpoint is set to 95% of the target temperature for the current batch, reserving adjustment space for the disturbance absorption phase. Based on the current temperature, target temperature, and the maximum allowable heating slope, the theoretical shortest heating time for this stage is estimated: Shortest heating time = (Target temperature − Current temperature) / Maximum allowable heating slope; The main heating phase uses a constant slope for heating control, and the trajectory is expressed as: Current temperature = Current temperature + Constant heating slope × (Current time − Starting time of the main heating phase). The control of the main heating phase aims to ensure the linear controllability of the heating process, enabling the system to maintain an efficient and stable energy input mode during the main heating phase, while reserving buffer space for subsequent disturbance absorption, improving overall temperature control accuracy and robustness.

[0032] During the main heating phase, the system continuously monitors for temperature curvature anomalies caused by external disturbances (such as liquid level fluctuations, external feed, sudden changes in heat load, etc.). If the second derivative of the heating curve (i.e., the acceleration of temperature change) exceeds a preset threshold, it is designated as a disturbance duration interval, triggering disturbance absorption. This includes setting an absorption time window and implementing a slow-convergence structure. Based on the disturbance duration (obtained from the disturbance duration interval) and the system's thermal inertia constant (characterizing the system's thermal response delay time, which can be estimated through a state-space model or step response), the system sets a minimum duration window for the disturbance absorption phase. ,in This is the minimum sustaining time window for the disturbance absorption section. It is a function with maximum value. This refers to the actual duration of the detected disturbance. This is an empirical adjustment factor to ensure that the absorption window sufficiently covers the disturbance process. The system's thermal inertia constant is the product of its equivalent thermal resistance and equivalent heat capacity. In the disturbance absorption phase, the system employs curvature suppression on the original heating slope, using a smoothing function to adjust the current slope to match the target slope of the main heating phase. This achieves a flexible return to the heating trajectory, preventing control oscillations or system overshoot. The disturbance absorption phase addresses external disturbances or temperature fluctuations during the heating process, avoiding power bounces or control system oscillations caused by severe disturbances. A gentle convergence mechanism smoothly transitions back to the main trajectory, enhancing system robustness and temperature control flexibility.

[0033] Step S3: Based on the data of the anti-initial-value drift section, the main heating section, and the disturbance absorption section, perform power back-calculation to obtain the heating power data; Specifically, the system performs time alignment and logical merging (including replacing the corresponding main heating segment data with the data from the disturbance absorption segment) on the aforementioned three types of characteristic heating segments to construct the target heating trajectory function. The temperature change data of the anti-initial-value drift segment, the main heating segment, and the disturbance absorption segment are spliced ​​together according to a time series to form a continuous and complete heating function. After obtaining the complete temperature trajectory, the system uses numerical differentiation methods (such as forward difference, center difference, etc.) or analytical derivative methods to solve for the first derivative of the temperature curve to obtain the real-time temperature rise rate. Based on the constructed thermal resistance-heat capacity series-parallel model and combined with the temperature derivative data, the system uses the following energy balance formula to back-calculate the power: ,in This represents the heating power at time t. This represents the equivalent heat capacity of the system, which includes the total heat capacity of all heat capacity units such as tank material and raw liquid. Indicates that the system is in The actual temperature at that moment, express The ambient temperature at any given time This represents the various thermal resistance units involved in the heat conduction process from the system to the outside. This inverse model includes two core thermodynamic terms: one is the heat capacity term. The first term represents the internal energy input required for the system temperature to rise; the second term is the heat loss term. This represents the power loss compensation due to heat conduction. The sum of these two values ​​is the actual heating power required for the system to reach the target trajectory.

[0034] Step S4: Obtain real-time change data, and perform trajectory changes based on heating power data and real-time change data to obtain heating change data.

[0035] Specifically, the system continuously collects and monitors real-time changes in the environment and equipment during the heating process. When preset change trigger conditions are met, the system adjusts and replans the current heating trajectory to achieve system stability and target orientation. The system continuously monitors various dynamic disturbances or state change events during operation, including but not limited to interruptions of external heating sources; sudden changes in liquid heat capacity due to replenishment of raw material tanks; sudden drops or rises in ambient temperature; and discrepancies between the actual output power of the heater and the planned power. These events constitute real-time change data, used to determine whether adjustments to the original planned trajectory are necessary. The system sets multi-dimensional trigger threshold conditions to determine whether the current heating process deviates from the predetermined trajectory. Trigger conditions include the absolute deviation between the current measured temperature and the target temperature exceeding a set tolerance; and the current actual output power of the heater being inconsistent with the originally planned power. When any condition is met, the system triggers the trajectory change process. Once a trajectory change is triggered, the system deletes all planned segments after the current time in the original temperature trajectory to avoid subsequent control relying on outdated predictions. The system sets an update window interval forward from the current time t. ,in This is the dynamically set time length based on system response inertia, disturbance duration, or controller adjustment capability. This window serves as the definition range for the new trajectory function. Within the newly generated change window, the system constructs a new temperature control trajectory, satisfying the following two requirements: the beginning of the trajectory must smoothly connect with the current temperature state to avoid abrupt changes in the derivative; the end of the trajectory should guide to the new target temperature or correct the planned point under the target curve. This trajectory function uses a smooth interpolation function (such as a cubic spline), a time-optimal control function, or an optimized solution based on model predictive control. This invention can dynamically respond and reconstruct the heating trajectory when sudden disturbances or input deviations occur during the temperature control process, avoiding the continuous execution of unreasonable historical trajectories, effectively improving the system's response flexibility, heating accuracy, and energy consumption control capability. It is particularly suitable for production processes where polyurethane raw materials are highly sensitive to temperature changes.

[0036] Preferably, the thermal resistance-thermal capacity series-parallel model is constructed as follows: Step S11: Construct a heat capacity-thermal resistance network model based on tank data and raw material liquid data to obtain the heat capacity-thermal resistance network model; Specifically, the tank data includes tank wall thickness, material thermal conductivity, heat exchange area (the effective heat transfer area of ​​the tank's outer wall in contact with the heater), and equivalent heat capacity of the tank (tank mass multiplied by the tank material's specific heat capacity). The raw material liquid data includes liquid density, liquid volume, liquid specific heat capacity, liquid thermal conductivity, liquid layer height, and the heat transfer area of ​​the liquid layer in contact with the tank wall. The system performs equivalent thermal resistance calculations, including tank wall thermal resistance calculations (wall thickness divided by the product of material thermal conductivity and heat exchange area) and liquid layer thermal resistance calculations (liquid height divided by the product of liquid thermal conductivity and the heat transfer area of ​​the liquid layer in contact with the tank wall). The system also performs liquid heat capacity calculations: liquid density multiplied by liquid volume multiplied by liquid specific heat capacity. Based on the physical topology of the heat transfer path, a series-parallel thermal capacity-thermal resistance network is constructed, including a series structure: the path from the heater input to the liquid center is sequentially heater → tank wall thermal resistance → tank equivalent thermal capacity → liquid layer thermal resistance → liquid thermal capacity. If the liquid is stratified (e.g., due to temperature differences between different material components), the liquid layer can be divided into multiple equivalent thermal capacity nodes, forming a parallel structure to reflect the uneven heat distribution within the liquid. The resulting thermal capacity-thermal resistance network model can be viewed as a graphical structure composed of multiple thermal capacity nodes and thermal resistance channels.

[0037] Step S12: Based on the tank temperature data and the liquid temperature data in the raw material liquid data, perform state-space model processing on the heat capacity-thermal resistance network model data to obtain state-space model data; Specifically, based on the heat capacity-thermal resistance network model, a state-space dynamic response model of the system is constructed by collecting tank temperature and liquid temperature data. The raw material liquid data also includes liquid temperature data, specifically the following time-series temperature data collected by the system: tank outer wall temperature data: acquired in real time by tank sensors; liquid temperature data: which can be single-point thermometer measurements or average values ​​from multiple measuring points. The system's state variables are set as a set of temperature nodes. According to the laws of heat conduction and energy conservation, each layer of nodes in the heat capacity-thermal resistance network model is constructed as a system of differential equations for dynamic temperature changes. ,in For the state variable vector, For time variables, This is the system matrix (state transition matrix). To control the input matrix, The input power is the heating power. The system dynamics can be modeled as a state equation of the following form: the rate of change of the system state is determined by the system matrix, the control input matrix, the current state (i.e., the state variables), and the heating input power; the system matrix is ​​determined by the thermal resistance and heat capacity parameters, specifically: the first row indicates that the derivative of the tank layer temperature is affected by its own heat capacity and thermal resistance, and is also affected by the heat flow back from the liquid layer; the second row indicates that the derivative of the liquid layer temperature is affected by the coupling effect of heat transfer from the tank layer and heat dissipation from the liquid itself; the control input matrix indicates that the heating input only acts on the heat capacity nodes of the tank. Thus, a complete continuous-time state-space model is established to describe the thermal inertia dynamics of the entire heat transfer process. The state-space model data includes the system matrix, the control input matrix, and the state variables.

[0038] ; in For the system matrix, The thermal resistance between the tank and the liquid. For the heat capacity of the tank, For the heat capacity of the tank, For the heat capacity of the liquid, To control the input matrix.

[0039] Step S13: Label the state-space model data with the tank temperature data, ambient temperature data, and batch temperature rise data to obtain the thermal resistance-thermal capacity series-parallel model.

[0040] Specifically, the aforementioned state-space model is subjected to supervised fitting and label enhancement processing using historical batch temperature control data, outputting a thermal resistance-heat capacity series-parallel model that can be used for heating segment identification and power planning. The system collects temperature control data from multiple historical batches as the target output for supervised fitting, including but not limited to the tank temperature change sequence corresponding to each batch; the ambient temperature change sequence within the corresponding time period; and the actual heating power input data during the batch heating process. Each batch of data constitutes a triplet: tank temperature sequence, ambient temperature sequence, and power input sequence. This sample set is input into the state-space model to supervise the model's temperature prediction capability and parameter fitting accuracy. The system sets a fitting loss function: for the tank temperature sequence predicted by the model, the difference between it and the corresponding measured temperature is integrated; the summation is applied to all batches of data to form the total loss function, with the weighted average mean square error as the index; the goal is to minimize this loss. For the system parameter matrix (such as thermal resistance, heat capacity, coupling coefficient, etc.) in the state-space model, the following optimization methods are adopted: least squares fitting method; intelligent search optimization strategies such as genetic algorithms; and posterior distribution update methods such as Bayesian estimation. In the above process, parameters are finely tuned under the premise of maintaining thermal physical constraints to ensure that the model has stable predictive ability under different environments and initial tank conditions. After multiple batches of supervised learning and error control, if the model output shows good predictive ability of heating trajectory on all samples, the current state space model can be upgraded to a thermal resistance-thermal capacity series-parallel model with engineering adaptability.

[0041] Preferably, the anti-initial-value drift section processing specifically includes: Step S21: Perform initial drift discrimination based on the thermal resistance-thermal capacity series-parallel model to obtain initial drift data; Specifically, the temperature of the target tank is predicted using a pre-constructed thermal resistance-thermal capacity series-parallel model, resulting in a model output temperature sequence. Combined with real-time measured temperature data of the tank, a temperature deviation function is calculated between the two. The system sets a fixed detection time interval to capture the initial deviation trend during the heating process. This detection interval is preferably the first two minutes after system heating starts. Within this detection interval, the deviation function is integrated and averaged to obtain the initial drift deviation average, which is the average value of the deviation function within the set time period, reflecting the systematic error between the model and the actual response. If the calculated average deviation absolute value exceeds a preset deviation tolerance threshold, the system determines that the current batch has an initial drift problem. Once the drift is activated and the determination is established, the system automatically extracts data from the time period with the largest deviation as input data for subsequent temperature rise trajectory optimization and anti-drift control, and marks it as initial drift data.

[0042] Step S22: Perform temperature rise rate deviation activation determination on the initial drift data to obtain drift activation data; Specifically, based on the difference between the measured temperature change rate and the model-predicted temperature change rate, a rate deviation curve is constructed and an activation threshold is determined. Within the identified initial drift time segment, the system calculates the instantaneous rate of change of the measured temperature (i.e., the temperature rise rate) and the instantaneous rate of change of the model-predicted temperature. The system constructs a rate deviation function based on the difference between the two. The system analyzes the aforementioned rate deviation curve and sets a rate activation threshold and a cumulative time threshold. That is, if the cumulative time value of the rate deviation exceeding the threshold exceeds the preset threshold time, it indicates that there is a significant rate drift response in the current stage. Once the above activation conditions are met, the system identifies the segment as a drift activation segment and outputs the data for the corresponding time period as drift activation data.

[0043] Step S23: Perform anti-initial-value drift segment processing on the drift activation data to obtain preliminary segment data; Specifically, based on the start time and initial temperature of the drift activation period, the system constructs a linear heating trajectory with a controlled slope to replace the segment with large rate fluctuations in the original trajectory. This control function can be expressed as: ,in The corrected trajectory temperature function, The starting temperature, To control the slope for temperature rise and satisfy the constraints ,in This is the upper limit of the system's nominal heating rate, used to limit the heating speed and prevent thermal jump. For the current moment, This is the starting temperature of the trajectory segment. Based on the aforementioned control function, the system replaces the drift activation time periods in the original temperature trajectory, generating a preliminary trajectory segment for subsequent power back-calculation. This process is equivalent to dynamically correcting the original heating path with significant initial disturbances to achieve a smooth heating transition. When the heat capacity of the raw material liquid fluctuates, the system uses a second-order derivative constraint of temperature change to suppress abrupt changes in the curvature of the heating trajectory and prevent nonlinear acceleration effects caused by sudden changes in heat capacity. This constraint is stated as follows: ,in It is a function of temperature. For time variables, The preset smoothing threshold parameter is used. Through this method, the system can effectively correct abnormal heating trends caused by initial disturbances, generating stable, controllable, and structurally stable preliminary section heating data.

[0044] Step S24: Perform dynamic feedback correction on the initial segment data to obtain segment data resistant to initial value drift.

[0045] Specifically, the system continuously monitors the actual temperature rise rate at the current moment and compares it with the temperature rise rate of the slow-start trajectory generated by the model. The actual temperature rise rate is measured by sensors; the expected temperature rise rate is calculated from the initial slow-rise trajectory. If the actual rate is found to be significantly lower than the expected rate, it indicates that there is hysteresis in the heating system or that environmental heat dissipation is affecting the system. The system will automatically lower the target temperature rise rate to avoid the risk of thermal hysteresis or overshoot due to excessively strong temperature control commands. The system sets a time decay for the current temperature rise slope parameter, that is, dynamically adjusts the original constant slope parameter in the temperature trajectory into a time decay expression: ,in The adjusted constant slope parameter, To initially set the slope, It is an exponential function. This is the slope decay factor, used to control the decay rate. For the current moment, This represents the starting point of the gradual temperature rise phase. Based on the dynamic slope correction mechanism, the system employs a sliding time window approach to locally smooth and optimize the continuity of the current temperature trajectory. This approach combines the temperature change trend of the previous time period with the target trajectory's connection requirements, correcting for abrupt changes or discontinuities in the curve through average slope correction within the window. The system outputs data for the anti-initial-value drift segment after dynamic feedback optimization.

[0046] Preferably, the temperature rise rate deviation activation determination specifically includes: A rate offset curve is constructed from the initial drift data to obtain the rate offset curve data; Specifically, the system extracts the following temperature sequences from the initial drift period: the measured tank temperature sequence, i.e., the temperature change data obtained from real-time monitoring; and the model-predicted temperature sequence, i.e., the temperature rise trend data output by the thermal resistance-heat capacity model. The system uses a sliding time window method to estimate the first derivative of the temperature rise. The duration of the sliding window is set (e.g., ...). =10 seconds), for each time t, calculate the difference between that time and the endpoint of the previous window. The system calculates the average temperature rise rate between measured and model temperatures; it also calculates the actual temperature rise rate at a given moment based on measured temperature data; and it calculates the model temperature rise rate at the same moment based on model predicted temperature data. This method approximates the temperature rise slope numerically by dividing the temperature difference between adjacent time points by the time interval, balancing real-time performance and noise resistance. At each moment, the system calculates the difference between the actual and model temperature rise rates, obtaining a temperature rise rate offset value, and constructs a complete rate offset curve in the form of a time series. If the actual temperature rise is faster than the model, the offset value is positive; if the actual temperature rise is slower than the model, the offset value is also positive; if they are completely equal, the offset value is zero. The system outputs a complete rate offset curve, which is a sequence of offset values ​​changing over time.

[0047] Slope activation data is obtained by performing slope activation judgment based on the rate offset curve data. Specifically, the system sets a slope deviation threshold, for example, 1.5°C / min. If the following conditions are met: ,in To monitor the start time, For the end time of monitoring, For indicator functions, The temperature rise derivative deviation function, This is the slope deviation threshold. For time variables, The activation duration threshold (e.g., 30 seconds) is used. The time intervals that meet the conditions constitute the slope activation data interval.

[0048] Threshold surface triggering is performed based on the activation data to obtain drift activation data.

[0049] Specifically, constructing a three-dimensional threshold surface ,in For time, For the temperature rise derivative deviation, The threshold surface represents the current temperature. This threshold surface can be obtained through empirical data fitting, historical simulation optimization, or expert parameter tuning, and represents the assessment result of the drift risk intensity under the combined effect of the above three parameters. In the embodiment, the threshold surface can be modeled using a polynomial function, for example, set as a weighted sum of the square term of the temperature rise derivative deviation, the first term of the temperature value, the first term of time, and a constant term, representing the contribution intensity of different influencing factors to the drift risk. ,in This is the value of the disturbance intensity assessment function. For the temperature rise derivative deviation, These are the derivative deviation weighting coefficients. This is the temperature weighting factor. The time-weighted coefficient, This is a bias term (constant term), representing the base value or initial threshold for the disturbance intensity score, used to adjust the overall level. The system performs threshold value judgment on the 3D feature vector at each time point: if the threshold surface output value is greater than the preset drift activation threshold, a significant drift risk is determined at that moment, triggering the anti-drift mechanism; the threshold value can be adjusted empirically to adapt to different heating power scenarios or raw material batch differences. To avoid misjudgments or sudden changes caused by the "critical value effect," the system can set fuzzy judgment for the threshold trigger area: for judgment points close to the threshold (such as threshold values ​​fluctuating by a certain percentage, e.g., ±10%), a fuzzy weighting factor is set; this factor represents the intensity or confidence of the current drift activation, with a value range between 0 and 1; it is dynamically adjusted according to the activation degree to enhance the system's robustness and response smoothness. The system integrates all time periods that meet the threshold surface trigger conditions into a drift activation data segment.

[0050] Preferably, the anti-initial-value drift section processing specifically includes: The drift activation data and the preset prior thermal inertia data are compared differentially to obtain the drift error data; Specifically, to identify whether the current temperature rise process significantly deviates from normal thermal inertia behavior, the system performs a differential comparison between the current temperature rise rate and the prior thermal inertia model within the drift activation zone. Within the drift activation zone, the system calculates the rate of temperature change per unit time based on measured tank temperature data. This rate can be achieved by calculating the first derivative of the temperature time series at each moment or by using the sliding window difference quotient method, reflecting the current actual temperature rise rate, and is denoted as the real-time temperature rise rate. The system pre-constructs and stores a set of standard thermal inertia response data, called the prior thermal inertia model. This model can be obtained statistically based on stable operating data from multiple historical batches of temperature rise processes, or established through experimental calibration, representing the reference temperature rise rate under standard conditions with the same structure and material properties, and is denoted as the reference temperature rise rate. The system calculates the difference between the current actual temperature rise rate and the reference temperature rise rate, forming drift error data in time series form. This error quantity represents the degree of deviation of the current system temperature rise behavior from the ideal thermal inertial response at each time point, and can be used for subsequent disturbance intensity assessment, quantitative analysis of drift risk, or generation of compensation control strategies. The output drift error data is a sequence of error signals that varies with time, serving as an important input for drift intensity integral calculation and risk interval division.

[0051] The disturbance intensity integral data is obtained by performing disturbance intensity integral calculation on the drift error data; Specifically, the disturbance intensity is defined as the absolute area of ​​the error curve, representing the cumulative offset intensity: ,in For the integral data of disturbance intensity, This is the start time of the drift activation period. This is the end time of the drift activation period. Let be the drift error function. The time variable is used. This integration process can be implemented using numerical methods, such as the trapezoidal integral method, Simpson's integral method, or the midpoint method, to ensure accurate acquisition of the total disturbance even with short sampling times. The output is a single-valued integral index of disturbance intensity.

[0052] The drift risk interval data is divided into drift risk intervals based on the integral data of disturbance intensity. Specifically, the system classifies drift activation data into risk levels based on the disturbance intensity integral index to achieve dynamic assessment and response adjustment of abnormal temperature control trends. The system sets tiered thresholds for the disturbance intensity integral value to classify the disturbance level within the drift activation period into different risk levels: Low-risk zone: If the disturbance intensity integral value within the corresponding time period is less than the first threshold, it is judged as low risk; Medium-risk zone: If the disturbance intensity integral value is between the first and second thresholds, it is judged as medium risk; High-risk zone: If the disturbance intensity integral value is greater than or equal to the second threshold, it is judged as high risk. The first and second thresholds can be set based on historical data analysis, domain experience, or dynamic adaptive strategies, possessing engineering adjustability, with the second threshold being greater than the first threshold. The system divides the drift activation time period into multiple fixed-length sub-intervals (e.g., every 30 seconds), and applies a sliding window mechanism to calculate the corresponding local disturbance intensity integral within each time period. Based on the above classification rules, each sub-interval is labeled with a corresponding risk level label. The output is a "time period – risk level" mapping table, which indicates the risk status of each time segment. For example, t1 is the start parameter of the time period. Time period [t1, t1+30 seconds]: medium risk; time period [t1+30 seconds, t1+60 seconds]: high risk; time period [t1+60 seconds, t1+90 seconds]: low risk; and so on.

[0053] Based on the drift risk interval data, anti-drift correction is performed on the drift activation data to obtain preliminary segment data.

[0054] Specifically, based on the aforementioned drift risk interval data, the system dynamically corrects the temperature rise trajectory within the corresponding drift activation segment, thereby generating preliminary trajectory segment data after anti-drift processing. According to the risk level of the current drift activation segment, a corresponding correction control strategy is implemented on the temperature rise trajectory to achieve a layered response and control temperature control offset caused by system overshoot and thermal inertia. Table 1: Risk Level Control Strategy Table

[0055] For medium- to high-risk areas, the system introduces a dynamic trajectory correction function to generate temperature change curves. Taking the high-risk area as an example, an exponential heating function with slow-start characteristics can be used as an alternative trajectory: ,in To correct the temperature function value, To correct the initial temperature, For the temperature increment, is the base of the natural logarithm. The gradual increase adjustment coefficient, For the current time, To correct the start time, the corrected trajectory segment is output as the initial segment data, replacing the original active interval.

[0056] Preferably, the main heating section is specifically processed as follows: The boundary of the main heating section is determined for the thermal resistance-thermal capacity series-parallel model, and the boundary data of the main heating section are obtained. Specifically, the system receives the following models and parameters as input: thermal resistance-thermal capacity series-parallel model, the target temperature value set by the system (manually input or pre-input), the initial heating temperature value (obtained through a sensor preset on the terminal), and the maximum applicable heating power value (manually input or pre-input). The start time (start of the main heating segment) is the time point after the initial value drift section processing is completed. The end time (end of the main heating segment) is triggered when any of the following conditions are met: the real-time temperature reaches η times the target temperature, where η is a preset proportionality coefficient, ranging from 0.90 to 0.98, used to reserve a disturbance absorption window; the current temperature rise rate (i.e., the first derivative of temperature) drops below a specified rate threshold, indicating that the system's thermal inertial response tends to stabilize and the heating power weakens. The system, based on a thermal resistance-thermal capacity series-parallel model, simulates the heating trajectory under maximum input power, obtaining a fitted curve of temperature change over time. Combining the derivative trend of the simulated temperature curve with the absolute temperature value, and according to the aforementioned boundary definition conditions, it automatically identifies the start and end times of the main heating segment. It outputs structured time interval data to label the time window range of the main heating segment. The system outputs the boundary time data of the main heating segment, i.e., time pairs.

[0057] Based on the boundary data of the main heating section, the shortest heating time is calculated by back-calculating the thermal resistance-thermal capacity series-parallel model to obtain the shortest heating time data. Specifically, based on the thermal resistance-capacity series-parallel model, under the constraint of a limited maximum heating power, the system reverse-engineers the theoretical shortest heating time required for the system to rise from the current initial temperature to the target temperature, which is used to set the energy efficiency baseline or optimize the objective function. The system determines the heating objective based on the initial and target temperatures recorded at the boundary of the main heating segment, combined with the current maximum heating power limit: to solve for the shortest time required for the temperature to rise from the current initial temperature to the target temperature without exceeding the maximum heatable power. The thermal resistance-capacity series-parallel structure model is transformed into a first- or second-order linear differential equation system to make it solvable in the state space; the power input function is set as a unit step function, meaning the system continuously heats at maximum power from time zero; numerical simulation methods (such as the fourth-order Runge-Kutta method) are used to integrally calculate the temperature response trajectory over time; during the simulation, the time point when the temperature curve first satisfies that the current temperature is greater than or equal to the target temperature is determined in real time, and this time is defined as the shortest heating time. The system outputs the theoretical shortest heating time. This value is a single scalar variable that reflects the shortest time required for the system to reach the target temperature under the current model structure and power constraints.

[0058] Based on the shortest heating time data, the constant temperature rise rate slope of the thermal resistance-thermal capacity series-parallel model is calculated to obtain constant temperature rise rate slope data. Specifically, based on the theoretical shortest heating time obtained from the aforementioned solution, the system calculates the constant temperature rise rate slope required for the system to rise from the initial temperature to the target temperature under ideal conditions. This slope is then corrected using upper and lower limits to obtain a constant rate parameter that can be used to construct the temperature control trajectory. The system calculates the constant temperature rise rate slope under ideal conditions based on the difference between the initial temperature and the target temperature in the main heating segment, as well as the theoretical shortest heating time. This slope represents the temperature increase that the system should achieve per unit time, forming the basic reference parameter for the subsequent heating control function. To adapt to the control capabilities and safety limitations of the actual system, the system performs upper and lower limit constraints on the aforementioned constant temperature rise rate slope. Specifically, if the system has a preset maximum temperature rise rate limit, i.e., the maximum allowable temperature change rate per unit time, the calculated constant slope is compared with this upper limit, and the smaller one is taken as the current effective heating slope. If the system has a minimum rate requirement for the heating response (such as ensuring a certain thermal drive efficiency or time constraint), the aforementioned slope is compared with the minimum rate lower limit, and the larger one is taken as the heating slope. The output slope parameter is a scalar variable representing the constant heating rate slope required to raise the temperature to the target temperature in the shortest time, taking into account the upper and lower limits of the heating rate allowed by the system.

[0059] The main heating segment function is generated based on the slope of the constant temperature rise rate, and the main heating segment data is obtained.

[0060] Specifically, based on the calculated constant temperature rise rate slope, a temperature trajectory function for the main heating stage is constructed and discretized for actual control execution, thereby obtaining the temperature setpoint data for the main heating segment. The system can model the main heating segment using the following two types of functions according to different control objectives and thermal inertia response characteristics: When the system has stable heating capability and low thermal inertia or has already smoothly transitioned through previous anti-drift control, a linear temperature function is used to construct the trajectory of the main heating segment. The temperature function is: the temperature of the main heating segment equals the initial temperature plus the product of the effective heating slope and the time difference (current time - initial time). If the system has strong thermal inertia or high requirements for the stability of the temperature control process, an exponentially increasing function is generated to avoid temperature oscillations. Specifically, the temperature function equals the target temperature minus an exponential decay term, ensuring that the heating rate gradually slows down, i.e. ,in This is the current temperature value. For the target temperature, The initial temperature, is the base of the natural logarithm. This is the temperature response adjustment coefficient. For the current time, The system calculates the start time of the main heating phase. It performs numerical discretization and control command conversion on the aforementioned temperature control trajectory function. Specifically, it discretizes the continuous trajectory function into a sequence of temperature setpoints at time intervals (e.g., per second), forming a fixed-length temperature control frame, which can be transmitted as input to a PID controller or other temperature control execution module. Based on the fitting relationship between the temperature trajectory function and the thermal resistance-heat capacity model, a calculated power reference value can be attached to each setpoint, assisting the power back-calculation module in achieving precise heating control. The output is a complete dataset of the main heating phase temperature control function, specifically including a temperature setpoint time sequence, representing the time node corresponding to each control frame; a temperature setpoint numerical sequence, representing the set temperature the system should reach at each moment; and the corresponding analytical expression of the heating trajectory function.

[0061] Preferably, the disturbance absorption section processing specifically includes: The absorption section trigger judgment is performed on the thermal resistance-thermal capacity series-parallel model to obtain the absorption section data; Specifically, the input data includes: real-time temperature curves: these are real-time temperature change sequences collected from the tank or system temperature sensors, reflecting the actual heating behavior of the system under current operating conditions; model-predicted temperature curves: these are theoretical heating curves derived from the thermal resistance-capacity series-parallel model, representing the temperature rise trend under ideal, undisturbed conditions; and heating rate deviation sequences: these are the differences between the two at the time derivative level, calculated as the absolute value of the difference between the real-time temperature change rate and the predicted temperature change rate. The system can identify a disturbance event and trigger absorption segment identification if any of the following conditions are met: If the heating rate deviation continuously exceeds a set threshold and the duration exceeds a preset duration threshold τ seconds, it indicates that the system is experiencing a sudden heating disturbance. For example, the heating derivative deviation is greater than the threshold and the duration is >τ seconds. If the system temperature curve shows a plateau or irregular fluctuations over a period of time, manifested as the first derivative after low-pass filtering being lower than a set threshold, and the temperature standard deviation exceeding a set upper limit during that time period, it is determined that the system has temperature oscillations. The system employs a sliding window analysis strategy to continuously monitor temperature change data, with the sliding window time range set to 20 to 30 seconds. Within each sliding window, the system calculates the indicators involved in the two aforementioned criteria and performs joint analysis. The specific fusion strategy can be optimized and adjusted according to the set weights or logical relationships to achieve sensitive identification of different types of disturbances. When either criterion is met, the system outputs the following information as the absorption segment judgment result: an absorption segment trigger timestamp, i.e., the start time when the disturbance occurs and the trigger condition is met, used to mark the starting point of the absorption segment; and preliminary label information for the absorption segment, including but not limited to "sudden temperature rise," "sudden temperature drop," or "temperature plateau" (i.e., the system temperature curve shows a plateau period or irregular fluctuations over a period of time).

[0062] The time window for the disturbance absorption section is determined based on the absorption section data, and the time window data for the disturbance absorption section is obtained. Specifically, based on the absorption segment trigger information, the effective duration interval of the disturbance absorption segment on the time axis is determined, i.e., the absorption segment time window, to support temperature control trajectory switching and disturbance type identification. The system sets a preliminary time window based on the identified disturbance absorption segment trigger time, specifically including the start time, set as the absorption segment trigger time point determined in the aforementioned steps, i.e., the time when the system identifies the temperature rise disturbance signal; and an initial estimate of the end time, which by default extends the trigger time by a fixed period, denoted as Δt, set to 60–120 seconds, to cover most of the disturbance absorption cycle; these two are combined to form a coarse-grained disturbance absorption segment duration estimate. The system performs temperature recovery trend judgment and dynamically corrects the absorption segment end time, specifically as follows: The absolute error between the real-time temperature and the prediction model is calculated, defined as the residual function, i.e. ,in For residuals, This is the measured temperature value. The system predicts temperature values ​​for the thermal resistance-capacity model. If the system detects that the residual is less than the safety error threshold for a continuous period of τ seconds, it considers the current disturbance to be nearing its end. At this point, the current time is taken as the true end time of the disturbance segment, locking the effective time window of the disturbance absorption segment. The system outputs the time window data and attribute label information of the disturbance absorption segment, specifically including the disturbance absorption segment time window: represented in the form of a time interval, used to clarify the duration range of the disturbance; disturbance attribute type label: combined with the temperature change pattern and rate characteristics, the absorption segment is assigned one of the following type labels: plateau-type disturbance: the temperature curve remains approximately constant for a period of time; oscillating disturbance: the temperature curve shows obvious up and down oscillations; slope reversal-type disturbance: the direction of the temperature derivative changes in the opposite direction.

[0063] The thermal resistance-thermal capacity series-parallel model is processed with a slow convergence structure based on the time window data of the disturbance absorption segment to obtain the disturbance absorption segment data.

[0064] Specifically, for the time window corresponding to the disturbance absorption phase, a smooth transition temperature control function is constructed to avoid sudden changes in the temperature control trajectory after the disturbance ends, thereby reducing the risk of system thermal shock and enhancing the stability and safety of the thermal control system. A temperature buffer transition mechanism is introduced after the disturbance phase, with specific objectives including avoiding trajectory jumps: not directly returning to the main heating trajectory in the early stages of disturbance recovery to prevent drastic fluctuations in power or temperature slope; and introducing a smooth transition period: delaying convergence to the target temperature by smoothing the trajectory to mitigate the impact of the disturbance on the system's thermal inertia. For different types of disturbances, the following two types of temperature control functions can be used to construct the transition trajectory: an exponential transition function (suitable for disturbances with abrupt slope changes), which constructs an exponential function whose temperature gradually converges to the target temperature after the disturbance trigger time: the current temperature equals the temperature at the time of disturbance triggering plus the target temperature difference multiplied by an exponential decay factor; ,in This is the current temperature value. The measured temperature at the moment the disturbance was triggered. The target temperature for the main heating section. is the base of the natural logarithm. The convergence adjustment coefficient, whose value is positively correlated with the disturbance intensity, is typically adjusted between 0.01 and 0.1 to control the transition speed. For the current time, This refers to the disturbance trigger time. For temperature plateau periods or oscillating disturbances, a piecewise linear function can be used for transition, with the following structure: First stage: Temperature remains constant for a delay of 5–20 seconds to buffer the system; Second stage: Temperature rises slowly at a slower ramp rate (typically 50%–80% of the normal ramp rate); Third stage: A smooth connection curve is constructed between the current temperature and the main heating segment to ensure continuity and slope consistency, seamlessly connecting to the main heating trajectory. Based on the generated disturbance absorption segment temperature curve, the heating power required for each control moment is recalculated; a power change rate constraint is set to limit the power adjustment amplitude per unit time to reduce equipment response pressure and avoid energy consumption surges or thermal inertia buildup caused by "power jumps". Based on the above strategy, the system generates the following output data: Absorption segment temperature trajectory function data: containing the target temperature value corresponding to each moment in the absorption segment, i.e., the time-temperature pair sequence, which can be directly used as the controller input; Power correction suggestions: including the suggested power adjustment value for each sampling moment, or the constraint conditions for the power change slope; Absorption segment end marker: clearly indicating the end time of the absorption segment, used to splice it with the main heating segment trajectory to ensure the temporal continuity and physical consistency of the overall trajectory.

[0065] Preferably, step S3 specifically includes: Step S31: Based on the data of the anti-initial-value drift section, the main heating section, and the disturbance absorption section, perform trajectory stitching to obtain trajectory stitching data; Specifically, after modeling the temperature trajectories for each stage, they are uniformly processed and sequentially spliced ​​to generate a complete temperature setpoint trajectory that is continuous in time, physically reasonable, and numerically differentiable, used to guide the execution of the heating control system. The system receives temperature trajectory data from the following three stages: trajectory data for the anti-initial-value drift section, denoted as the first temperature trajectory, corresponding to the time interval from the start time to the first turning point; trajectory data for the main heating section, denoted as the second temperature trajectory, corresponding to the time interval from the turning point to when the main target temperature zone is nearly completed; and trajectory data for the disturbance absorption section, denoted as the third temperature trajectory, corresponding to the time interval from after the main heating section to when the system enters the stable transition period. Each trajectory segment is a time-temperature sequence and can be described in function form. The system adopts the following strategy for splicing: the difference between the endpoint temperature and the starting temperature of the next segment at the connection point between any two adjacent trajectory segments should not exceed 1 degree Celsius to prevent temperature jumps; if the endpoint temperature difference between two trajectory segments exceeds the limit, a transition trajectory is inserted between them. The transition section can employ linear interpolation or Hermite interpolation to achieve a smooth slope transition while maintaining continuous differentiability. The system samples three trajectory segments uniformly, sampling each segment at fixed time intervals (e.g., once per second) to obtain a set of temperature points on a standard time axis. The three sampled trajectories are then concatenated in chronological order to form a unified time-temperature trajectory sequence. Each sampling point in this sequence contains the sampling time and the corresponding target temperature value. The output is a complete temperature control trajectory with the following characteristics: from the start time of the anti-initial drift section to the end time of the disturbance absorption section, the entire trajectory is uninterrupted; the temperature change between any adjacent sampling points is smooth, supporting the calculation of the first derivative (temperature rise rate) and the second derivative (thermal inertia); each moment contains a data pair of "time point - target temperature value".

[0066] Step S32: Calculate the temperature derivative of the trajectory stitching data to obtain the temperature derivative data; Specifically, the derivative of the complete temperature trajectory generated in the previous stage is calculated to obtain a sequence of temperature change rate (i.e., temperature rise rate) over time, providing a basis for power estimation and control response analysis. For each time point in the temperature trajectory, the first derivative of the temperature at that point is calculated, representing the temperature rise rate per unit time. This is defined as the derivative of the target temperature with respect to time, i.e., temperature rise rate = temperature change after a certain time point divided by the time interval, expressed as the first derivative of temperature with respect to time. An approximate calculation is performed using the moving difference method (central difference). For the i-th time point in the trajectory, its temperature derivative is approximately expressed as: the current temperature rise rate ≈ the temperature difference between two adjacent time points divided by the total time interval between them. This involves using the data from the previous and next time points to estimate the derivative of the current time point, thus obtaining a smooth and relatively accurate temperature change rate. The time step is set to 1 second, but can be adjusted according to the actual sampling period of the trajectory data. To reduce high-frequency noise interference caused by sensor errors or sampling fluctuations, it is recommended to add one of the following smoothing strategies before and after the difference calculation: use a Savitzky-Golay smoothing filter to process the original temperature sequence or derivative sequence; or use a sliding median filter to suppress anomalous jumps. The system performs targeted processing in the following scenarios: For the disturbance absorption segment, to preserve the disturbance transition characteristics, the sliding window width can be appropriately reduced (e.g., only one point before and after the disturbance) to improve the derivative response sensitivity; for boundary point processing: at the starting point, since there is no previous data point, the forward difference method (using the current point and the next point) is used for estimation; at the ending point, since there is no subsequent data point, the backward difference method (using the current point and the previous data point) is used for estimation; to avoid derivative loss or error offset at the trajectory boundaries. The system output is a sequence of time-stamped temperature rise rate data points, each data point including the current time point and the corresponding temperature rise rate (i.e., the first derivative value of temperature).

[0067] Step S33: Based on the temperature derivative data and the thermal resistance-thermal capacity series-parallel model, perform power back-calculation to obtain the heating power data.

[0068] Specifically, based on previously obtained temperature derivative data and parameters of the thermal resistance-heat capacity series-parallel model, the real-time heating power during the tank heating process is back-calculated to evaluate the energy consumption level and the effectiveness of the thermal control strategy. In the thermal resistance-heat capacity series-parallel network model (mainly using a first-order approximation), the expression for heating power at any given time includes two terms: one is the power consumption due to temperature rise caused by heat capacity, and the other is the heat flow loss caused by the temperature difference with the environment. Its overall structure is as follows: the first term represents the heat capacity effect: equivalent heat capacity multiplied by the rate of temperature change; the second term represents the heat dissipation effect: the difference between the current tank temperature and the ambient temperature divided by the equivalent thermal resistance. That is, the total power is jointly determined by the system's heat storage response and the environmental heat dissipation load. Input temperature derivative data, i.e., the temperature rise rate sequence per unit time; temperature trajectory data, i.e., the tank temperature at each moment; ambient temperature data, collected in real time or obtained through sensors. Equivalent heat capacity is estimated by factors such as tank material, liquid volume, and specific heat capacity; equivalent thermal resistance is obtained by fitting historical data, system calibration, or thermal simulation. If a multi-node serial-parallel structure model is adopted, it is converted into a state-space model, and the relationship between power input and state derivative is solved through matrix operations. The system calls the thermal resistance-thermal capacity model constructed in step S1 to extract the required equivalent thermal capacity and equivalent thermal resistance parameters. For each time point in the temperature derivative sequence, the current derivative value, real-time temperature, and ambient temperature are substituted into the power model for calculation, generating heating power values ​​point by point. If the back-calculated power value at a certain time point shows a non-physical jump (such as a sudden increase or decrease or a negative value), the system determines it as a sensor malfunction; external disturbance (such as opening the lid or liquid fluctuation); or model fitting error; such points will be marked with an anomaly label for the trajectory adjustment module to refer to; the system sets upper and lower power limit thresholds, and exceeding these limits can trigger an alarm mechanism or switch back to the predicted power value. The output is a set of heating power value pairs arranged by time, including timestamps and corresponding power values.

[0069] Preferably, step S4 specifically includes: Step S41: Obtain real-time change data; Specifically, the system continuously monitors and collects data on real-time changes affecting the accuracy of the heating trajectory and the system's response behavior during operation, providing input for dynamic trajectory adjustment. The real-time change data collected by the system may include the deviation between the current actual tank temperature and the set temperature; the external ambient temperature is collected and compared with the previous cycle temperature; if the temperature difference exceeds a set threshold, it is considered an environmental disturbance trigger condition; raw material replenishment or sudden event detection is also included; if a sudden change is detected in the liquid level sensor (such as a sharp increase or decrease in liquid level), it is considered a replenishment event; or if the internal pressure of the tank experiences a momentary drop, it can also be considered a raw material disturbance. The system also collects the feedback power of the heating actuator in real time; if the feedback power continuously exceeds a set upper limit threshold (such as excessive load), the system will record it as an abnormal energy consumption event or equipment malfunction. The real-time data sampling period is recommended to be set to 1 to 5 seconds. The system uses a sliding window caching mechanism to perform dynamic trend analysis on continuous data from the past 10 to 30 seconds to avoid misjudgments due to instantaneous fluctuations. All types of real-time change data will be uniformly packaged into a structured dataset as input to the dynamic control module. The real-time change data set output by the system includes, but is not limited to, temperature deviation, temperature error at the current moment; ambient temperature difference, current ambient temperature change; replenishment event identifier, Boolean flag or liquid level change value; power anomaly flag or feedback power value; and other disturbance-related data fields (such as tank pressure, heater status, etc.).

[0070] Step S42: Determine the trajectory change trigger condition based on the real-time change data to obtain the trajectory change data; Specifically, based on real-time collected system operating status data, the system determines whether the triggering conditions for trajectory adjustment are met, thereby achieving dynamic response and trajectory updates during temperature control. The system identifies the following typical disturbance types and sets corresponding triggering rules: Table 2: Trigger Types for Trajectory Changes

[0071] When any of the above triggering conditions is met, the system will generate a trajectory change flag and trigger time information. For example, if the change flag is set to "on", the trigger time is recorded as the current system clock time, i.e., trajectory change flag: change flag = on; trigger timestamp: change trigger time = current time. The output trajectory change data structure includes a disturbance type field, such as "temperature difference disturbance", "environmental disturbance", "liquid level disturbance", etc.; a trigger time field, recording the timestamp of the first time the triggering condition is met; and additional information, such as trigger threshold, deviation amplitude, power error value, etc.

[0072] Step S43: Based on the trajectory change data and heating power data, trim the remaining trajectory segment to obtain the remaining trajectory segment data; Specifically, based on the triggered trajectory change information, the original temperature setting trajectory is truncated, and the key features and power trend data of the remaining segment are extracted for use as input to the reconstruction logic. The system uses the trigger time in the current trajectory change data as the boundary to trim the original temperature setting trajectory: the original temperature setting trajectory is a set of ordered temperature-time pairs with corresponding heating power setting sequences; the system defines the portion of the original setting trajectory where the current time is greater than the trajectory change trigger time as the remaining trajectory segment. The system analyzes the thermodynamic characteristics of the remaining trajectory segment, including whether the remaining trajectory segment sequence is in a heating state, an isothermal state, or exhibits temperature overshoot behavior at the trimming start point; based on the joint evolution trend of the temperature-time pairs and heating power setting sequences of the remaining trajectory segment sequence, the system calculates the heating slope of the current control stage (e.g., the temperature rise rate of linear heating); and extracts the target power value or actual predicted power in the current segment to evaluate the actuator control stability and power boundary constraints. The system will output the following data: Temperature remaining trajectory segment: the sequence of temperature setpoints in the original trajectory where t > trajectory change trigger time; Power remaining trajectory segment: the heating power data matching the above time series; Trend parameter set: including temperature trend labels (heating / constant temperature / overshoot), current control slope value, power estimation information, etc.

[0073] Step S44: Generate a change window based on the remaining segment data of the trajectory to obtain change window data; Specifically, based on the data characteristics of the remaining trajectory segment and the response capability of the control system, an effective time window for trajectory adjustment, namely the change window, is determined, which serves as the initial deployment range for subsequent interpolation trajectories or correction functions. The change window refers to the minimum control duration during which trajectory change operations can be safely implemented under the current thermal and power conditions of the system. This window must satisfy both the system's physical feasibility and provide sufficient control margin to accommodate the slowly changing trajectory function. The system sets the length of the change window based on the following constraints: power change rate limit: to avoid abrupt changes in heating power during trajectory reconstruction, it must be ensured that the subsequent power change rate does not exceed the maximum power slope that the system can withstand, i.e., the power change per unit time after trajectory adjustment must not exceed the preset safety limit; thermal inertia response delay: considering the physical hysteresis characteristics of thermal resistance-thermal capacity systems, the system estimates the typical thermal response delay time under the current state (e.g., 10~30 seconds) as a reference for the delayed effectiveness of trajectory changes; actuator control cycle limit: the system also needs to consider the minimum control cycle of temperature control actuators (such as heaters or solenoid valves) to ensure that control commands during trajectory adjustment can be stably executed at the hardware level. The start time is set to the trajectory change trigger time; the end time is set to the trajectory change trigger time plus a dynamically adjusted window duration, ranging from 20 to 60 seconds, depending on the three constraints mentioned above. This window segment is the change window. The system will output the following structured data, where the change window data includes two fields: start time and end time, used to define the deployment range of the new function segment that is allowed to be inserted into the trajectory.

[0074] Step S45: Replan the trajectory function based on the changed window data to obtain the heating change data.

[0075] Specifically, based on the generated trajectory change window, the temperature setting trajectory is reconstructed using a function to achieve dynamic adjustment of the thermal control strategy. The goal of the reconstruction is to ensure good continuity and smoothness of the system temperature curve before and after the change point, avoiding temperature control overshoot or control breakage, thereby maintaining the stability of the system's thermal inertial response. Within the change window, i.e., the start time is the trajectory change trigger time and the end time is the trigger time plus the window duration, the system replaces the original setting trajectory with the newly fitted temperature function. Based on the control objective and thermal response characteristics, one of the following three trajectory function types is used: piecewise linear function (suitable for scenarios requiring rapid reset of the target temperature, where the temperature rise slope remains constant), exponentially rising function (used for scenarios requiring suppression of system inertia or avoidance of oscillations, where the trajectory exhibits a smooth change trend from slow to fast and then gradually stabilizes), and third-order spline interpolation function (if continuous interpolation transition is required between the starting temperature and the original setting trajectory, a cubic spline function can be used to ensure that the fitted curve maintains continuity on the first derivative, thereby avoiding temperature curve breakpoints or abrupt slope changes). The replanned temperature function consists of three parts: The first segment uses the original trajectory before the change is triggered; the second segment uses the replanned temperature trajectory function within the change window; and the third segment connects to the remaining part of the original trajectory after the change window ends, and can be shifted or have its slope corrected depending on the adjusted target. The overall trajectory function should meet the requirements of continuity and differentiability, meaning the start and end temperature values ​​of each segment must be continuous, and short smoothing segments should be inserted at the connection points if necessary. Based on the replanning results of the temperature trajectory, the system will synchronously generate a new power demand trajectory. Using a thermal resistance-thermal capacity model, the required heating power at each moment is estimated as follows: First, the first derivative of the reconstructed temperature curve at each moment is calculated, i.e., the temperature rise rate; then, the required power is calculated based on the equivalent thermal capacity, equivalent thermal resistance, and real-time ambient temperature. This power value will be used to control the actuator output and also to determine the energy consumption trend. The system outputs the following heating change data: a newly set temperature trajectory function, composed of three segments: the original trajectory (before the change trigger point); the replanning function (during the change window); the adjusted remaining trajectory (after the change window); a synchronously generated heating power data sequence, used for control and energy consumption analysis; and function parameter information, such as the slope coefficient of the exponential function and the control points of the spline function, output to further explain the fitting behavior.

[0076] Preferably, this application also provides a polyurethane raw material preheating control system for executing the polyurethane raw material preheating control method described above, the polyurethane raw material preheating control system comprising: The thermal resistance-heat capacity modeling module is used to acquire tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; and to construct a thermal resistance-heat capacity series-parallel model based on the tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data to obtain the thermal resistance-heat capacity series-parallel model. The segmented response identification and processing module is used to process the anti-initial-value drift segment of the thermal resistance-thermal-capacity series-parallel model to obtain the anti-initial-value drift segment data; process the main heating segment of the thermal resistance-thermal-capacity series-parallel model to obtain the main heating segment data; and process the disturbance absorption segment of the thermal resistance-thermal-capacity series-parallel model to obtain the disturbance absorption segment data. The reverse power calculation module is used to back-calculate the heating power data based on the data of the anti-initial drift section, the main heating section, and the disturbance absorption section. The heating trajectory change control module is used to acquire real-time change data and change the trajectory based on the heating power data and the real-time change data to obtain heating change data.

[0077] Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended application documents rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of the equivalents of the application documents be incorporated into the invention.

[0078] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for controlling the preheating of polyurethane raw materials, characterized in that, Includes the following steps: Step S1: Obtain tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; Based on tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data, a series-parallel thermal resistance-thermal capacity model was constructed to obtain the series-parallel thermal resistance-thermal capacity model. Step S2: Perform initial value drift resistance processing on the thermal resistance-capacity series-parallel model to obtain initial value drift resistance data; perform main heating segment processing on the thermal resistance-capacity series-parallel model to obtain main heating segment data; perform disturbance absorption segment processing on the thermal resistance-capacity series-parallel model to obtain disturbance absorption segment data. Step S3: Based on the data of the anti-initial-value drift section, the main heating section, and the disturbance absorption section, perform power back-calculation to obtain the heating power data; Step S4: Obtain real-time change data, and perform trajectory changes based on heating power data and real-time change data to obtain heating change data.

2. The method according to claim 1, characterized in that, The thermal resistance-thermal capacity series-parallel model is constructed as follows: Based on tank data and raw material liquid data, a heat capacity-thermal resistance network model is constructed to obtain the heat capacity-thermal resistance network model. Based on the tank temperature data and the liquid temperature data in the raw material liquid data, the state space model data of the heat capacity-thermal resistance network model is processed to obtain the state space model data. The tank temperature data, ambient temperature data, and batch temperature rise data are labeled on the state-space model data to obtain the thermal resistance-thermal capacity series-parallel model.

3. The method according to claim 1, characterized in that, The specific handling of the anti-initial-value drift section is as follows: Initial drift data is obtained by performing initial drift discrimination based on the thermal resistance-thermal capacity series-parallel model. The initial drift data is subjected to temperature rise rate deviation activation determination to obtain drift activation data; The drift activation data is processed to resist initial value drift segmentation to obtain preliminary segment data; The initial section data is dynamically corrected by a gradual increase phase to obtain the section data resistant to initial value drift.

4. The method according to claim 3, characterized in that, The specific determination of temperature rise rate deviation activation is as follows: A rate offset curve is constructed from the initial drift data to obtain the rate offset curve data; Slope activation data is obtained by performing slope activation judgment based on the rate offset curve data. Threshold surface triggering is performed based on the activation data to obtain drift activation data.

5. The method according to claim 3, characterized in that, The specific handling of the anti-initial-value drift section is as follows: The drift activation data and the preset prior thermal inertia data are compared differentially to obtain the drift error data; The disturbance intensity integral data is obtained by performing disturbance intensity integral calculation on the drift error data; The drift risk interval data is divided into drift risk intervals based on the integral data of disturbance intensity. Based on the drift risk interval data, anti-drift correction is performed on the drift activation data to obtain preliminary segment data.

6. The method according to claim 1, characterized in that, The main heating section is specifically processed as follows: The boundary of the main heating section is determined for the thermal resistance-thermal capacity series-parallel model, and the boundary data of the main heating section are obtained. Based on the boundary data of the main heating section, the shortest heating time is calculated by back-calculating the thermal resistance-thermal capacity series-parallel model to obtain the shortest heating time data. Based on the shortest heating time data, the constant temperature rise rate slope of the thermal resistance-thermal capacity series-parallel model is calculated to obtain constant temperature rise rate slope data. The main heating segment function is generated based on the slope of the constant temperature rise rate, and the main heating segment data is obtained.

7. The method according to claim 1, characterized in that, The specific processing of the disturbance absorption section is as follows: The absorption section trigger judgment is performed on the thermal resistance-thermal capacity series-parallel model to obtain the absorption section data; The time window for the disturbance absorption section is determined based on the absorption section data, and the time window data for the disturbance absorption section is obtained. The thermal resistance-thermal capacity series-parallel model is processed with a slow convergence structure based on the time window data of the disturbance absorption segment to obtain the disturbance absorption segment data.

8. The method according to claim 1, characterized in that, Step S3 is as follows: Trajectory stitching is performed based on the data from the anti-initial-value drift section, the main heating section, and the disturbance absorption section to obtain trajectory stitching data. Temperature derivatives are calculated from the trajectory stitched data to obtain temperature derivative data; Heating power data is obtained by back-calculating the power based on the temperature derivative data and the thermal resistance-thermal capacity series-parallel model.

9. The method according to claim 1, characterized in that, Step S4 is as follows: Get real-time change data; The trajectory change trigger conditions are determined based on real-time change data to obtain trajectory change data; The remaining trajectory segment is trimmed based on the trajectory change data and heating power data to obtain the remaining trajectory segment data; A change window is generated based on the remaining segment data of the trajectory, and the change window data is obtained. The trajectory function is replanned based on the changed window data to obtain the heating change data.

10. A polyurethane raw material preheating control system, characterized in that, For performing the polyurethane raw material preheating control method as described in claim 1, the polyurethane raw material preheating control system includes: The thermal resistance-heat capacity modeling module is used to acquire tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data; and to construct a thermal resistance-heat capacity series-parallel model based on the tank temperature data, ambient temperature data, batch temperature rise data, tank data, and raw material liquid data to obtain the thermal resistance-heat capacity series-parallel model. The segmented response identification and processing module is used to process the anti-initial-value drift segment of the thermal resistance-thermal-capacity series-parallel model to obtain the anti-initial-value drift segment data; process the main heating segment of the thermal resistance-thermal-capacity series-parallel model to obtain the main heating segment data; and process the disturbance absorption segment of the thermal resistance-thermal-capacity series-parallel model to obtain the disturbance absorption segment data. The reverse power calculation module is used to back-calculate the heating power data based on the data of the anti-initial drift section, the main heating section, and the disturbance absorption section. The heating trajectory change control module is used to acquire real-time change data and change the trajectory based on the heating power data and the real-time change data to obtain heating change data.

Images