Industrial reaction kettle temperature control method and system

Abstract

This application relates to the field of industrial reaction process control technology, and discloses a method and system for temperature control of an industrial reactor. The method includes: acquiring multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and constructing an initial temperature field; calculating the distribution of temperature non-uniformity gradient intensity to determine whether there are risk areas in the reactor; if so, analyzing potential abnormal signals of temperature changes within a preset short time period; performing feedforward compensation power correction on the jacket medium flow synchronization parameters to generate a heating power adjustment trajectory; determining whether there is a potential heating power overshoot risk in the reactor; if so, generating a steady-state deviation correction vector, updating the actuator's control pulse sequence, and obtaining a real-time temperature adjustment command set; monitoring abnormal fluctuations in the local thermal change rate; if the abnormal fluctuation increases, triggering an abnormal power reduction interlock mechanism to achieve stable overall operation and output of the reactor. This application improves the accuracy of temperature control in industrial reactors.

Description

Technical Field

[0001] This application relates to the field of industrial reaction process control technology, and in particular to a method and system for temperature control of an industrial reactor. Background Technology

[0002] As the core equipment for realizing chemical reactions in industrial production, the temperature control accuracy of the reactor directly determines the reaction efficiency, material conversion rate, product quality stability, and production safety, making it a key link in chemical process control. In complex reaction systems, the material inside the reactor exhibits significant spatial non-uniform heat distribution along the axial and radial directions. Affected by multiple factors such as stirring effect, jacket medium flow rate, and reactor wall heat transfer resistance, local overheating or cold spots can easily form, leading to problems such as material decomposition, exacerbation of side reactions, and even safety hazards.

[0003] Existing temperature control methods often rely on single-point or limited temperature measurement data, making it difficult to comprehensively characterize the heat distribution within the vessel. Furthermore, they fail to adequately consider key factors such as cumulative deviations in heat transfer impedance and dynamic heat load changes, resulting in lag in control response. Simultaneously, traditional methods lack coordination between heating power and stirring speed adjustments, have inadequate overshoot suppression mechanisms, and cannot effectively address localized temperature fluctuations, making it difficult to achieve overall thermal equilibrium and stability within the vessel. Therefore, there is an urgent need for a temperature control method that can accurately sense spatial heat distribution, respond in real-time to dynamic operating conditions, and coordinate heat input and material mixing to address the pain points of uneven temperature distribution, low control accuracy, and high safety risks inherent in existing technologies. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a method and system for controlling the temperature of an industrial reactor, which improves the accuracy of temperature control in industrial reactors.

[0005] In a first aspect, this application provides a method for temperature control of an industrial reactor, the method comprising: Acquire multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and construct the initial temperature field; Based on the initial temperature field, the rate of thermal change at each local location is obtained in real time, the distribution of temperature non-uniform gradient intensity is calculated, and the presence of a risk area in the reactor is determined based on the distribution of temperature non-uniform gradient intensity. If a risk area exists, local data where the stirring speed and shear force distribution do not match are extracted from the risk area, and combined with the rapid heating control logic, potential abnormal signals of temperature change within a preset short period of time are analyzed. Based on the potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, the jacket medium flow synchronization parameters are fed forward compensation power correction to generate a heating power adjustment trajectory. Based on the heating power adjustment trajectory, overshoot risk signal assessment data is obtained. Combined with the overshoot suppression feedback mechanism and the weighted processing of the cumulative deviation of heat transfer impedance, it is determined whether there is a heating power overshoot risk in the reactor. If there is, a steady-state deviation correction vector is generated based on the local thermal change rate data of the drastic fluctuation area, the control pulse sequence of the actuator is updated, and a real-time temperature adjustment command set is obtained. Based on the real-time temperature regulation instruction set, the actuator that drives the heating power and stirring speed is activated to monitor abnormal fluctuations in the local thermal change rate. If the abnormal fluctuation increases, the abnormal power reduction interlock mechanism is triggered to achieve stable overall operation and output of the reactor.

[0006] Secondly, this application provides an industrial reactor temperature control system, the system comprising: The data acquisition unit is used to acquire multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and to construct the initial temperature field; Anomaly analysis unit is used to acquire the rate of thermal change at each local location in real time based on the initial temperature field, calculate the temperature non-uniform gradient intensity distribution, determine whether there is a risk area in the reactor based on the temperature non-uniform gradient intensity distribution, and if so, extract local data where the stirring speed and shear force distribution do not match in the risk area, and analyze potential abnormal signals of temperature change within a preset short period of time in combination with rapid heating control logic. The correction unit is used to perform feedforward compensation power correction on the jacket medium flow synchronization parameters based on the potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, and generate a heating power adjustment trajectory. The risk assessment unit is used to obtain overshoot risk signal assessment data based on the heating power adjustment trajectory, and combine the overshoot suppression feedback mechanism with the weighted processing of the cumulative deviation of heat transfer impedance to determine whether there is a potential heating power overshoot in the reactor. If there is, a steady-state deviation correction vector is generated based on the local thermal change rate data of the drastic fluctuation area, the control pulse sequence of the actuator is updated, and a real-time temperature adjustment instruction set is obtained. The temperature control unit is used to drive the actuator that coordinates the heating power and stirring speed according to the real-time temperature adjustment command set, monitor abnormal fluctuations in the local thermal change rate, and trigger the abnormal power reduction interlock mechanism if the abnormal fluctuation increases, so as to achieve stable operation output of the reactor as a whole.

[0007] Compared with the prior art, the beneficial effects of the present invention are at least as follows: First, by acquiring multiple temperature data points along the axial and radial directions, dynamic parameters of the jacket medium flow rate and stirring speed are simultaneously obtained. An initial temperature field is constructed using inverse distance weighted interpolation and dynamic time warping algorithms, accurately characterizing the non-uniform heat distribution within the vessel. This overcomes the limitations of traditional single-point or limited-scale temperature measurement methods, which struggle to comprehensively perceive the thermal state, providing reliable data support for subsequent temperature control. Simultaneously, based on the initial temperature field, the rate of thermal change and the intensity of temperature gradient at various local locations are calculated. By integrating a thermal disturbance model and weighted analysis methods, local overheating or cold spot risk areas are accurately identified, enabling directional positioning of abnormal areas and providing clear targets for targeted control, thus improving the precision of temperature control. Then, data on the mismatch between stirring speed and shear force distribution within the risk areas are extracted. Potential abnormal temperature change signals are analyzed using an exponential weighted average method. Combined with the accumulated deviation of the inner wall heat transfer impedance, a feedforward compensation power correction coefficient is generated, dynamically adjusting the jacket medium flow rate parameters to precisely match the heating power with the dynamic heat load, effectively mitigating the temperature imbalance problem caused by accumulated heat transfer impedance. Next, by integrating an overshoot suppression feedback mechanism with weighted processing of accumulated thermal impedance deviation, potential heating power overshoot is accurately identified. An adaptive control algorithm optimizes the steady-state deviation correction vector, and pulse width modulation (PWM) technology updates the actuator control pulse sequence, significantly improving overshoot suppression capabilities and preventing localized temperature runaway. Finally, the actuators for synchronized operation of the driving heating power and stirring speed are coupled with an abnormal power reduction interlock mechanism to respond in real-time to abnormal fluctuations in localized thermal change rates. This achieves dynamic equilibrium control of the internal thermal state, significantly improving reaction efficiency, material conversion rate, and product quality stability, reducing safety risks caused by localized overheating, and expanding the applicability of this control method in complex reaction systems. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart of an industrial reactor temperature control method according to an embodiment of this application; Figure 2 This is a schematic diagram of the three-dimensional distribution of the initial temperature field in an embodiment of this application. Figure 3 This is a schematic diagram comparing the heating power adjustment trajectory and temperature response curve of an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an industrial reactor temperature control system according to an embodiment of this application. Detailed Implementation

[0010] This application provides a method and system for temperature control of an industrial reactor. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0011] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of a method for controlling the temperature of an industrial reactor in this application includes: Step S1: Obtain multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and construct the initial temperature field.

[0012] The initial temperature field is constructed by including: A multi-point temperature acquisition layout is distributed along the axial and radial directions inside the reactor to collect temperature sequence data at different heights and radial positions, which are then aggregated to form multi-point temperature sequence data. Simultaneously, the jacket medium flow rate information and the dynamic adjustment parameters of the stirring speed are acquired and set as synchronous dynamic parameters. The multi-point temperature sequence data and synchronous dynamic parameters are weighted and averaged using inverse distance weighted interpolation and spatial interpolation algorithms to generate temperature data for each acquisition point inside the reactor. Based on a dynamic time warping algorithm, the temperature data within different time periods are time-series aligned to obtain the temperature field mapping result inside the reactor. The temperature field mapping result is corrected and optimized by integrating environmental thermal disturbance and heat flow transfer characteristic models, and considering heat transfer impedance and stirring effects, an initial temperature field containing spatially non-uniform heat distribution is obtained.

[0013] Specifically, before constructing the initial temperature field, the deployment of the temperature acquisition layout within the reactor must be completed. The initial temperature field refers to a set of temperature data that comprehensively reflects the temperature distribution at different locations along the axial and radial directions inside the reactor, and includes spatially non-uniform heat distribution characteristics, providing basic data for subsequent thermal state analysis. Specifically, multiple layers of temperature acquisition points are uniformly arranged along the axial direction on the inner wall of the vertical cylindrical industrial reactor. Each layer has multiple temperature sensing elements arranged at equal angular intervals along the radial direction, forming a multi-point temperature acquisition layout covering different axial heights and radial positions within the reactor. This layout can comprehensively capture temperature differences in the bottom liquid accumulation area, the central main reaction area, the upper gas phase area, and the near-wall area, ensuring the spatial integrity of the acquired data. Thermocouples with fast response speed and high measurement accuracy are preferred as the temperature sensing elements, enabling rapid capture of dynamic temperature changes and avoiding control delays caused by static data.

[0014] During the data acquisition phase, temperature data from each acquisition point is simultaneously collected using the aforementioned multi-point temperature acquisition layout. This data is recorded chronologically to form a temperature sequence. Summarizing the temperature sequence data from all acquisition points yields the multi-point temperature sequence data, which reflects the temperature status at different locations within the vessel at different times. Simultaneously, flow detection devices installed on the inlet and outlet pipelines of the jacket medium collect real-time jacket medium flow information. The frequency converter control system reads the dynamic adjustment parameters of the stirring speed. These two types of parameters are set as synchronous dynamic parameters. The jacket medium flow information is directly related to the vessel wall heat transfer efficiency, while the dynamic adjustment parameter of the stirring speed reflects the synergistic effect of material mixing and heat transfer. Together, they serve as key auxiliary data for constructing the initial temperature field.

[0015] After data acquisition, inverse distance weighted interpolation and spatial interpolation algorithms were used to process the multi-point temperature sequence data and synchronous dynamic parameters. Inverse distance weighted interpolation is a distance-based spatial interpolation method. Its core principle is that the attribute value of any point in space is influenced by surrounding known points, and the degree of influence is proportional to the inverse of the distance between the two points. The closer the known point is, the greater its contribution to the unknown point. The input data for this algorithm includes the temperature sequence data of each acquisition point, the synchronous dynamic parameters, and the spatial coordinates of each acquisition point. The output data is the estimated temperature value for locations within the vessel where no acquisition points are located. This algorithm can fill in the gaps in temperature data between acquisition points, forming a preliminary continuous temperature distribution. Spatial interpolation algorithms are used to further optimize the continuity of temperature distribution. The principle is to use the temperature data of known collection points to fit the temperature distribution function of the entire reactor space through mathematical modeling. The input data are the temperature estimate after processing by inverse distance weighted interpolation and the spatial position relationship of each collection point. The output data is the temperature distribution data of the entire reactor. By combining the two algorithms for weighted averaging and fitting calculation, temperature data of each collection point and key locations in the entire reactor can be generated, ensuring the accuracy of temperature distribution characterization and generating preliminary temperature distribution data of continuous spatial locations in the reactor.

[0016] Dynamic time warping is an algorithm used to handle the alignment problem of time series data. Its core function is to solve the problems of inconsistent temperature data collection frequencies and timestamp deviations in different time periods. The input data of this algorithm are temperature data collected in different time periods and their corresponding timestamp information. By finding the optimal matching path between two time series, the time axis is flexibly adjusted to make the temperature data in different time periods accurately correspond in the time dimension. The output data is the unified time axis temperature series data after time alignment, that is, the temperature field mapping result inside the reactor. The temperature field mapping result refers to the visualized or digital representation result that can intuitively reflect the temperature change law of the reactor inside the reactor with time and space after data processing. It can clearly show the temperature state of different locations inside the reactor at different times, providing a direct basis for subsequent correction and optimization of the initial temperature field.

[0017] The time-aligned temperature field mapping results, along with synchronously acquired dynamic parameters such as jacket medium flow rate and stirring speed, are input into the aforementioned environmental thermal disturbance model and heat flow transfer characteristic model.

[0018] Environmental thermal disturbance refers to the interference of external factors such as changes in the external ambient temperature and heat dissipation from equipment on the temperature field inside the reactor. The environmental thermal disturbance model is a mathematical model established by fitting experimental data to quantify the degree of this type of disturbance. The environmental thermal disturbance model can be quantified using the following simplified calculation method: Let the heat exchange coefficient between the outer wall of the reactor and the environment be... This coefficient can be obtained through equipment parameter tables or steady-state heat dissipation experiments. The ambient temperature is... The temperature of the outer wall of the reactor is (This can be estimated from the average temperature of the jacket medium or measured by sensors on the outer wall). Then, the additional heat flux density per unit area caused by environmental thermal disturbance is... for: This heat flux density is conducted through the vessel wall to the inner wall, and its effect can be determined according to Fourier's law and the thermal resistance of the vessel wall. After calculation, the temperature at a certain local location inside the vessel is finally obtained. Perturbation correction amount The calculation formula is: The positional geometry factor describes the relationship between the location and the heat transfer path of the outer wall, and can be determined by the geometry of the reactor. This represents a functional relationship. Therefore, the input data includes parameters such as the external ambient temperature and the equipment's heat dissipation coefficient, while the output data is the influence of environmental thermal disturbances on the temperature at various locations inside the vessel, i.e., the disturbance correction amount.

[0019] The heat transfer characteristic model is a model based on the principles of heat transfer, characterizing the heat transfer processes such as conduction, convection, and radiation within a reactor. This model can be established based on the macroscopic energy balance equation of the reactor. For the entire reactor system, within a small time interval... Internally, its energy balance can be expressed as: The system's internal energy changes, with the input heat mainly including heat transferred through the jacket. ,in, For the mass flow rate of the jacketed medium, For its specific heat capacity, and The inlet and outlet temperatures represent the output heat, which mainly includes the heat carried away by the outflowing material and the heat dissipated to the environment through the vessel. The change in internal energy is reflected as the product of the change in the average temperature of the material inside the vessel and the total heat capacity. Using this equation, combined with real-time collected jacket flow rate, temperature, and stirring power (partially converted into heat), the overall heat flow distribution trend inside the vessel can be inferred or verified. For spatial distribution, the stirring speed can be further combined to estimate the heat transfer intensity in different regions, thereby distributing the total heat flow to different axial and radial regions and generating an estimated result of the heat flow intensity distribution. Therefore, the input data includes parameters such as the thermal conductivity of the vessel material, the specific heat capacity of the material, and the jacket medium temperature; the output data is the intensity distribution of heat flow transfer inside the vessel and the temperature change trend.

[0020] Thermal resistance refers to the obstruction of heat transfer caused by scale, fouling, and the material itself on the vessel wall. Its magnitude directly affects heat transfer efficiency. The stirring effect refers to the promoting effect of material mixing caused by stirring on the uniform distribution of heat. Both factors jointly influence the spatial distribution of temperature within the vessel. By combining the outputs of the environmental thermal disturbance model and the heat flow transfer characteristic model with the temperature field mapping results, and comprehensively considering both thermal resistance and stirring effect for correction, an initial temperature field containing spatially non-uniform heat distribution is finally obtained. (Cumulative deviation of inner wall thermal resistance) This can be considered as a time-varying additional thermal resistance. In practical calculations, it can be determined by comparing the actual total heat transfer under long-term operation. The theoretical heat transfer calculated under clean conditions and the same operating conditions (jacket temperature difference, flow rate) This is used for indirect evaluation. The relationship between the two can be approximated as: ,in, and These are the overall heat transfer coefficients under actual and clean conditions, respectively. Calculated in real-time or periodically. and with Comparison allows for dynamic updates. The value is then incorporated into the temperature field correction.

[0021] By combining the above methods, the influence of environmental thermal disturbance and heat flow transfer characteristics can be quantified into correction values ​​for the temperature data at each spatial location in the temperature field mapping result, thereby completing the optimization and obtaining the initial temperature field.

[0022] For example, Figure 2 To construct a three-dimensional schematic diagram of the initial temperature field, the figure uses grayscale gradients to represent temperature levels, with bright areas corresponding to high-temperature regions and dark areas corresponding to low-temperature regions. The coordinate axes represent the radial position (X, Y) and axial height (Z) of the reactor, respectively. Figure 2 The effectiveness of the multi-point temperature acquisition layout and the inverse distance weighted interpolation method was verified.

[0023] Through the above process, it is possible to accurately construct the temperature field inside the industrial reactor and quantitatively analyze its heat distribution characteristics, providing reliable initial data support for the entire temperature control method and ensuring the pertinence and effectiveness of subsequent temperature regulation commands.

[0024] Step S2: Based on the initial temperature field, the rate of thermal change at each local location is acquired in real time, the distribution of temperature gradient intensity is calculated, and the presence of a risk area in the reactor is determined based on the distribution of temperature gradient intensity. If a risk area exists, local data where the stirring speed and shear force distribution do not match are extracted, and potential abnormal signals of temperature change within a preset short period of time are analyzed in conjunction with the rapid heating control logic.

[0025] The determination of whether a risk area exists in the reactor includes: Based on the initial temperature field, gradient calculations are performed on the temperature data at various local locations within the reactor to obtain the thermal change rate at each local location. Using an online thermal disturbance estimator, the impact of external thermal disturbances on each local location is estimated based on the real-time ambient temperature, the average jacket temperature, and pre-calibrated equipment coefficients. The initial temperature field is then corrected based on the impact of the thermal disturbances to obtain the corrected temperature field. The temperature non-uniform gradient intensity at each local location is calculated based on the corrected temperature field and weighted according to the disturbance impact to obtain the temperature non-uniform gradient intensity distribution. Based on preset intensity values, areas of temperature non-uniformity are identified in the temperature non-uniform gradient intensity distribution and designated as risk areas. Risk areas include local overheating areas or local cold spots.

[0026] The analysis includes potential abnormal signals of temperature changes within a preset short time period, including: Local data includes local stirring speed data and temperature change sequence data. The local temperature change rate is calculated based on the local data. Based on the rapid heating control logic, a temperature change prediction model is constructed. The prediction model uses an exponential weighted average method to weight the local temperature change rate within the historical time period to obtain a temperature change prediction curve for a preset short period. Based on the preset temperature rise rate, potential abnormal signals with rapid temperature changes in the temperature change prediction curve are identified.

[0027] Specifically, the rate of thermal change refers to the amount of temperature change at a local location per unit time. It is calculated by selecting the temperature data of two consecutive moments at that local location, dividing the difference between the temperature value at the later moment and the temperature value at the earlier moment by the time interval. This calculation can capture the dynamic trend of temperature change at each local location in real time.

[0028] An online thermal disturbance estimator is a device used to estimate the impact of external thermal disturbances on the temperature of various local locations within a reactor in real time. Its core function is to separate the temperature changes caused by external environmental disturbances from the reactor's own reactions and heat transfer, avoiding temperature field distortion caused by external thermal disturbances and ensuring the accuracy of subsequent risk area assessments. The online thermal disturbance estimator operates according to the law of conservation of energy in heat transfer. By collecting real-time environmental and equipment temperature parameters and combining them with pre-calibrated equipment thermal characteristic coefficients, a thermal disturbance transmission model is established. This model estimates the transmission and attenuation process of external thermal disturbances in the reactor body, jacket medium, and materials in real time, thereby obtaining the specific quantitative value of the disturbance's impact on each local location. The thermal disturbance transmission model analyzes the impact of external thermal disturbances on the temperature of a specific local location *i* within the reactor. The specific calculation formula is as follows: ,in, This represents the estimated temperature effect (in °C) of an external thermal disturbance on local location i at time t. A positive value indicates that the external environment causes a temperature increase at that point, while a negative value indicates a temperature decrease. The ambient temperature (°C) at time t is measured by an ambient temperature sensor installed around the reactor. The average temperature (°C) of the jacket medium at time t is calculated from the jacket inlet and outlet temperatures and represents the main thermal boundary temperature of the vessel. The effective thermal disturbance coefficient at local location i represents the intensity of the influence of ambient temperature changes on that location and is obtained through offline calibration.

[0029] The pre-calibrated equipment coefficients refer to parameters characterizing the thermal properties of the reactor equipment, obtained through multiple sets of comparative experiments in the early stages. These parameters include thermal resistance, heat transfer area, and heat exchange efficiency. These coefficients are determined by collecting heat transfer data under different temperature conditions in ideal operating conditions without external disturbances, and then optimizing the data through fitting. Once calibration is complete, these coefficients can be fine-tuned periodically during subsequent operation based on equipment aging. Real-time ambient temperature refers to the real-time temperature of the environment surrounding the reactor, collected in real-time by temperature sensors installed around the equipment. The jacket average temperature refers to the average temperature of the medium inside the reactor jacket, obtained by taking the arithmetic mean of the inlet and outlet temperatures of the jacket medium. The input data for this online thermal disturbance estimator includes the real-time ambient temperature, the jacket average temperature, and the pre-calibrated equipment coefficients. The output data is the amount of thermal disturbance impact at each local location due to external thermal disturbances. This impact is a quantified value; the larger the value, the stronger the interference from external thermal disturbances at that local location, and the greater the possibility of temperature data distortion.

[0030] Thermal disturbance influence is a quantified value of local temperature deviation inside the vessel caused by external factors such as changes in ambient temperature and equipment heat dissipation. The core purpose of correcting the initial temperature field is to eliminate the interference of external thermal disturbance and restore the true temperature distribution state brought by the reaction and heat transfer inside the vessel. The specific correction process is as follows: subtract the thermal disturbance influence corresponding to each local location from the temperature data of each local location in the initial temperature field to obtain the true temperature data of each local location after correction. Arrange all the corrected temperature data in sequence according to their corresponding spatial locations to form the corrected temperature field. The corrected temperature field can more accurately reflect the actual spatial distribution of temperature inside the vessel, avoid misjudgment of temperature unevenness caused by external interference, and provide a reliable data basis for subsequent calculation of temperature unevenness gradient intensity.

[0031] Temperature gradient intensity is a physical quantity that characterizes the degree of non-uniformity in the spatial distribution of temperature within a reactor. Its calculation principle is as follows: Based on the real temperature data of each local location in the corrected temperature field, the temperature difference between two adjacent local locations is calculated. Then, the temperature difference is divided by the spatial distance between the two local locations to obtain the axial temperature gradient and radial temperature gradient of each local location. The axial temperature gradient refers to the rate of temperature change along the height of the reactor, and the radial temperature gradient refers to the rate of temperature change along the radius of the reactor. Subsequently, the gradient vector magnitude of the axial and radial temperature gradients is obtained through vector synthesis. This gradient vector magnitude is the temperature gradient intensity of that local location. The larger the value, the more non-uniform the temperature distribution around that local location, and the easier it is to form overheating or cold spot hazards. Since different local locations are affected by external thermal disturbances to varying degrees, the calculated temperature non-uniformity gradient intensity needs to be weighted and corrected based on the amount of thermal disturbance at each local location. The core logic of the weighted correction is: assign lower weights to local locations with larger thermal disturbance impacts to reduce the influence of external interference on the temperature non-uniformity judgment; assign higher weights to local locations with smaller thermal disturbance impacts to highlight their true temperature non-uniformity. The weighting coefficients used in the weighted correction are obtained after normalizing the thermal disturbance impact, ensuring that the weight allocation is negatively correlated with the degree of disturbance impact. For example, the weighted correction formula is: ,in, This represents the weighted correction for the temperature non-uniformity gradient intensity. Let be the initial temperature gradient intensity at local location i. This is the weighting adjustment coefficient, which is a preset positive number. The larger the value, the stronger the suppression of gradient intensity in highly perturbated regions. Let be the temperature influence at local location i mentioned above. This weighted correction further improves the accuracy of the temperature gradient intensity. By arranging the weighted corrected temperature gradient intensities at all local locations according to their spatial positions, the temperature gradient intensity distribution can be obtained. The temperature gradient intensity distribution refers to the set of distributions of temperature gradient intensities at all local locations within the reactor according to spatial coordinates, which clearly shows the specific location and degree of non-uniformity of the temperature non-uniform region within the reactor.

[0032] The preset intensity value is a critical value for the intensity of temperature non-uniformity gradient, pre-set based on the reaction process requirements of the reactor, the thermal stability of the materials, and the safety operation standards of the equipment. This value is determined through prior experiments combined with engineering experience. It ensures that when the intensity of the temperature non-uniformity gradient in a certain area exceeds this preset value, the temperature non-uniformity in that area may affect reaction efficiency, product quality, or equipment safety, and must be included in the scope of key monitoring and control. The identification process is as follows: All local locations within the temperature non-uniformity gradient intensity distribution are traversed, and areas with a temperature non-uniformity gradient intensity greater than the preset intensity value are selected. These areas are designated as high-intensity temperature non-uniformity areas and set as risk areas. Further comparison is made between the temperature data of each local location within this risk area and the average temperature data of the entire reactor. If the average temperature of this area is higher than the preset difference between the average temperature and the entire area, the risk area is determined to be a local overheating area; if the average temperature of this area is lower than the preset difference between the average temperature and the entire area, the risk area is determined to be a local cold spot area. This method can accurately locate the type and specific location of risk areas, providing a clear target for subsequent targeted extraction of local data and analysis of abnormal signals.

[0033] Specifically, the local data includes local stirring speed data and temperature change sequence data within the risk area. Local stirring speed data refers to the real-time monitoring data of the stirring speed at the corresponding location in the risk area, while temperature change sequence data refers to the set of temperature data recorded in chronological order at each collection point within the risk area. Based on these two types of data, the local temperature change rate can be calculated. This local temperature change rate is consistent with the calculation logic of the thermal change rate mentioned earlier, that is, it is obtained by the ratio of the temperature difference between adjacent moments to the time interval, which can accurately reflect the dynamic changes in temperature within the risk area.

[0034] Rapid temperature rise control logic refers to the temperature rise rate control rules preset based on reaction process requirements. Its core function is to provide a benchmark for predicting temperature change trends, ensuring that the prediction results meet actual production needs. A temperature change prediction model is constructed based on this rapid temperature rise control logic. This model uses an exponentially weighted average method to weight the local temperature change rate over a historical time period to obtain a temperature change prediction curve for a preset short-term period. The exponentially weighted average method is an algorithm used for trend analysis of time series data. Its principle is to assign higher weights to recent historical data and lower weights to distant historical data. By weighting the data, the influence of recent data on the prediction results is highlighted, thereby improving the accuracy of trend prediction. The input data for this algorithm are the local temperature change rate over a historical time period, a preset weight decay coefficient, and a preset short-term duration. The weight decay coefficient is used to adjust the weight ratio between recent and distant data. The output data is the predicted temperature change rate at each moment within the preset short-term period. Arranging these predicted values ​​in chronological order forms a temperature change prediction curve, which can intuitively present the expected temperature change trend in the risk area over a short period in the future.

[0035] The preset temperature rise rate is the maximum allowable temperature rise rate determined based on the safety requirements of the reaction process and the thermal stability of the materials. The potential anomaly signal refers to a signal characterizing the abnormal development direction of local temperature fluctuations, which can provide early warning of potential temperature runaway within the risk area. Specifically, the predicted temperature change rate at each moment in the temperature change prediction curve is compared with the preset temperature rise rate. If the predicted value at a certain moment exceeds the preset temperature rise rate, or if the curvature of the prediction curve shows an accelerating temperature change rate, then the corresponding temperature change is determined to be an abnormal change, and a corresponding potential anomaly signal is generated. The signal strength can be quantified based on the degree of overshoot or the steepness of the acceleration trend, providing accurate signal input for subsequent feedforward compensation power correction.

[0036] Through the above-mentioned risk area judgment and potential abnormal signal analysis process, the abnormal temperature area inside the reactor can be accurately located and its temperature change trend can be predicted. This provides a reliable basis for subsequent targeted adjustment of jacket medium flow rate, heating power and stirring speed, ensuring the accuracy and timeliness of reactor temperature control and effectively avoiding production safety hazards caused by local overheating or cold spots.

[0037] Step S3: Based on potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, perform feedforward compensation power correction on the jacket medium flow synchronization parameters to generate a heating power adjustment trajectory.

[0038] The generation of the heating power adjustment trajectory includes: The temperature fluctuation amplitude of each local area is calculated based on local data and potential abnormal signals. Based on the heat transfer condition of the reactor inner wall, the cumulative deviation of the inner wall heat transfer impedance is calculated by integrating the difference between the actual total heat transfer and the theoretical heat transfer, combined with the running time. The potential abnormal signals, the cumulative deviation of the inner wall heat transfer impedance, and the temperature fluctuation amplitude are normalized and synthesized using a weighted average method to generate a feedforward compensation power correction coefficient. Based on the feedforward compensation power correction coefficient, the jacket medium flow synchronization parameters are dynamically adjusted to generate a heating power adjustment trajectory that matches the dynamic heat load.

[0039] Specifically, the core of generating the heating power adjustment trajectory lies in comprehensively considering the local temperature fluctuation state, the development trend of temperature anomalies, and the long-term cumulative deviation of the heat transfer system. The synergistic effect of these three factors ensures the pertinence and accuracy of the correction strategy.

[0040] Local data refers to the local stirring speed data and temperature change sequence data within the risk area. Local stirring speed data reflects the dynamic state of material mixing in that area, while temperature change sequence data is a collection of temperatures from various collection points in that area, recorded chronologically, providing a clear view of temperature changes over time. Potential anomaly signals characterize the direction of abnormal local temperature fluctuations. These signals are generated by fitting historical temperature sequence trends using the exponential weighted average method, as described earlier, and are used to warn of potential temperature runaway risks in the short term. Temperature fluctuation amplitude refers to the degree to which temperature data deviates from the average temperature within a specific local area. It is calculated by selecting the difference between the maximum and minimum temperature values ​​within a preset time period for that local area, or by quantifying it through the standard deviation of the temperature sequence data. A larger standard deviation indicates more severe temperature fluctuations, and the temperature fluctuation amplitude directly reflects the degree of temperature instability in a local area.

[0041] Internal wall heat transfer resistance refers to the obstruction to heat transfer caused by scale, fouling, and the material itself on the inner wall of the reactor. Its magnitude directly affects the heat transfer efficiency between the reactor wall and the material, and is one of the important factors leading to uneven temperature distribution inside the reactor. The cumulative deviation of internal wall heat transfer resistance is a parameter used to quantify the long-term cumulative effect of this obstruction. Its calculation process is as follows: First, obtain the actual total heat transfer and theoretical heat transfer at the current moment. The actual total heat transfer is calculated by the flow rate, specific heat capacity, and inlet / outlet temperature difference of the jacket medium. The specific formula is: Actual total heat transfer = Jacket medium flow rate × Medium specific heat capacity × (Jacket medium outlet temperature - Jacket medium inlet temperature). This value reflects the heat actually transferred to the material inside the reactor through the jacket per unit time. The theoretical heat transfer refers to the heat that should be transferred through the jacket per unit time under ideal operating conditions with no heat transfer resistance accumulation. Its calculation is based on the overall heat transfer coefficient, heat transfer area, and heat transfer temperature difference (the difference between the average temperature of the jacket medium and the average temperature of the material inside the reactor) under the design conditions. The formula is: Theoretical heat transfer = Design overall heat transfer coefficient × Heat transfer area × Heat transfer temperature difference. Next, the difference between the actual total heat transfer and the theoretical heat transfer is calculated. This difference reflects the degree of deviation between the current heat transfer state and the ideal state. A positive difference indicates that the actual heat transfer efficiency is lower than the ideal state, meaning that the heat transfer resistance has a negative impact. Then, this difference is integrated over a period of time. The significance of integration is to accumulate the heat transfer deviation at different times and capture the changing trend of heat transfer resistance during long-term operation. Finally, the integral result is divided by the operating time over this period to obtain the cumulative deviation of the inner wall heat transfer resistance. The larger the value of the cumulative deviation of the inner wall heat transfer resistance, the more serious the problems such as scaling and fouling on the inner wall, and the stronger the interference with temperature control.

[0042] The weighted average method is a mathematical approach that assigns weights to various influencing factors based on their importance and performs comprehensive calculations. Its core principle is to assign corresponding weight coefficients based on the degree of influence of different parameters on heating power adjustment, making the correction coefficients more closely match actual control needs. The weight coefficients are determined based on the correlation between each parameter and temperature equilibrium control. For example, potential abnormal signals directly warn of abnormal temperature trends, so their weight can be set to a higher value; the cumulative deviation of the inner wall heat transfer resistance reflects long-term heat transfer characteristics, so its weight is secondary; the temperature fluctuation amplitude reflects the current temperature stability, and its weight is fine-tuned according to actual operating conditions. The sum of all weight coefficients is 1, and the weights can be determined through fitting multiple sets of experimental data or optimization based on engineering experience. For instance, during the rapid heating phase, the weight of potential abnormal signals can be set to a higher value (e.g., 0.5~0.7) to prioritize suppressing overheating risks; during the isothermal reaction phase, the weight of the cumulative deviation of the inner wall heat transfer resistance can be set to a higher value (e.g., 0.4~0.6) to compensate for the decrease in heat transfer efficiency; the weight of the temperature fluctuation amplitude is usually treated as a stabilizing term, with a value between 0.1 and 0.3.

[0043] The purpose of normalization synthesis is to eliminate the influence of dimensional differences between different parameters, ensuring that all parameters are on the same order of magnitude and guaranteeing the rationality of the comprehensive evaluation. The specific process is as follows: First, divide the actual value of each parameter by its preset maximum value to obtain a normalized value (ranging from 0 to 1). The preset maximum value can be determined based on potential abnormal signals, cumulative deviation of internal wall heat transfer resistance, temperature fluctuation amplitude, etc. Then, multiply each normalized value by its corresponding weighting coefficient. Finally, sum all the products to obtain the feedforward compensation power correction coefficient. Therefore, the input data for this weighted average method includes the quantized values ​​of potential abnormal signals, cumulative deviation of internal wall heat transfer resistance, and temperature fluctuation amplitude, as well as the corresponding weighting coefficients for each parameter. The output data is the feedforward compensation power correction coefficient, which is the core basis for subsequent adjustment of the jacket medium flow rate; its magnitude directly reflects the required correction level.

[0044] The jacket medium flow synchronization parameter refers to the jacket medium flow information acquired synchronously during temperature data acquisition. This parameter is directly related to the rate at which the jacket transfers heat to the reactor and is a key control variable for regulating the reactor temperature. Dynamic heat load refers to the real-time heat demand generated by the reactor during operation due to the combined effects of various factors such as reaction exothermics, material mixing, and environmental heat dissipation. Its value fluctuates dynamically with changes in reaction stage and operating conditions. The specific method for adjusting the jacket medium flow synchronization parameter is as follows: multiply the current jacket medium flow setpoint by the feedforward compensation power correction coefficient to obtain the corrected jacket medium flow target value. During the correction process, an upper limit for the flow adjustment amplitude needs to be set to avoid excessive impact on the heat transfer system caused by sudden flow changes. The upper limit can be set according to the rated flow rate of the equipment and the reaction process requirements. Subsequently, based on the corrected jacket medium flow target value, the convective heat transfer coefficient on the jacket side is recalculated. The calculation of the convective heat transfer coefficient is based on the flow state of the jacket medium (laminar or turbulent), the physical properties of the medium (viscosity, thermal conductivity, etc.), and the jacket structural parameters (pipe diameter, flow velocity, etc.), reflecting the heat transfer capacity between the jacket medium and the reactor wall. By combining the recalculated convective heat transfer coefficient with the distribution of the initial temperature field constructed above, the target heating power required for different local areas in the reactor can be deduced. For example, for local areas with drastic temperature fluctuations and large heat transfer resistance, a higher heating power is allocated to compensate for heat transfer losses; for local areas with stable temperatures, the heating power is appropriately reduced to avoid overshoot, ultimately forming a heating power adjustment trajectory segmented along the axial or radial direction.

[0045] Specifically, a power allocation model based on simplified dynamic thermal equilibrium is used to generate target heating power values ​​for each local region. This model divides the reactor into N local locations (or control volumes) corresponding to the initial temperature field. For any local location i (i=1,...,N), its thermal equilibrium relationship in the next control cycle can be simplified as follows: ,in, The target heating power (in W) for the local location i to be solved is the output of the power distribution model. Heat flow (in W) is the heat flow (in W) that is transferred to local location i through the jacket wall. The net heat flow (in W) between local location i and adjacent locations through material conduction and convection exchange. The total heat capacity (J / K) of materials and equipment within local location i is estimated based on the region's volume, material density, and specific heat capacity. Let i be the desired temperature at local position i at the end of the next control cycle. Let be the current temperature (°C) at local location i in the initial temperature field. To control the cycle duration (s), all local locations are... Arranged according to their spatial position, they form a heating power adjustment trajectory that matches the dynamic heat load.

[0046] The heating power adjustment trajectory refers to the optimized heat input distribution scheme that can match the current dynamic heat load demand. It presents the target heating power values ​​for different times and areas, providing precise instructions for the actuator's actions, ensuring that the heat input inside the vessel is dynamically adapted to the real-time heat load, and effectively suppressing the further development of local overheating or cold spots.

[0047] Step S4: Based on the heating power adjustment trajectory, obtain overshoot risk signal assessment data, combine the overshoot suppression feedback mechanism and the weighted processing of the cumulative deviation of heat transfer impedance, determine whether there is a potential heating power overshoot in the reactor. If so, generate a steady-state deviation correction vector based on the local thermal change rate data of the drastic fluctuation area, update the actuator's control pulse sequence, and obtain a real-time temperature regulation command set.

[0048] Among them, determining whether the reactor has a potential heating power overshoot includes: Based on real-time monitoring of heating power adjustment trajectory and the trend of heating power change, overshoot risk signal assessment data is extracted. The overshoot risk signal assessment data includes the temperature rise rate and the deviation between the peak and average heating power. Based on the overshoot suppression feedback mechanism, the deviation between heating power and temperature is collected in real time and feedback adjustment is performed through closed-loop control to generate an overshoot risk suppression signal. The overshoot risk suppression signal and the cumulative deviation of the inner wall heat transfer resistance are weighted and processed, and combined with the overshoot risk signal assessment data to generate a comprehensive assessment value. If the comprehensive assessment value exceeds the preset overshoot threshold, it is determined that there is a potential overshoot risk in the reactor, triggering the overshoot suppression feedback mechanism to adjust the input of heating power.

[0049] The real-time temperature adjustment instruction set obtained includes: The difference between the temperature data and the target temperature in the region of drastic temperature fluctuation is calculated based on the local thermal change rate data, and a steady-state deviation correction vector is generated based on the difference. Through an adaptive control algorithm, the weights of the steady-state deviation correction vector are adjusted in real time, combining the current temperature change and the preset temperature target range, to optimize the steady-state deviation correction vector. The optimized steady-state deviation correction vector is input into the control system to update the actuator's control pulse sequence. The duty cycle of the updated control pulse sequence is adjusted using pulse width modulation technology to control the heating power and stirring speed. A dynamic feedback adjustment mechanism is introduced into the updated control pulse sequence to generate a real-time temperature regulation command set.

[0050] Specifically, the overshoot risk signal assessment data is a set of key parameters used to quantify the degree of heating power overshoot risk. The temperature rise rate and the deviation between the peak and average heating power are core assessment indicators. The temperature rise rate refers to the increase in temperature in a localized area within the vessel per unit time. It is calculated by selecting temperature data from two consecutive moments, dividing the difference between the two times by the time interval. The temperature rise rate directly reflects the swiftness of temperature change; a larger value indicates a higher overshoot risk. The deviation between the peak and average heating power refers to the difference between the maximum heating power and the average power within a preset monitoring period. It is calculated by subtracting the average power from the peak heating power. A positive deviation indicates excessive power; a larger absolute value indicates a more significant overshoot probability and impact. When extracting these two types of data, the corresponding temperature change data must be simultaneously correlated to ensure that the assessment data accurately reflects the matching relationship between heat input and temperature response.

[0051] The overshoot suppression feedback mechanism is a closed-loop control process based on real-time temperature response. Its core principle is to dynamically adjust the heating power input by continuously comparing the difference between the actual temperature and the target temperature, thereby suppressing temperature runaway caused by heating power overshoot. The core function of this mechanism is to promptly correct heat input deviations and avoid risk accumulation. Its workflow is as follows: First, real-time temperature data of various local areas within the reactor is collected and compared with the preset target temperature to calculate the temperature deviation. The temperature deviation reflects the difference between the heating power and the temperature. A positive deviation value indicates that the actual temperature is higher than the target temperature, requiring a reduction in heating power; a negative deviation value requires an appropriate increase in heating power. Then, this temperature deviation is input into the closed-loop control algorithm, which uses proportional-integral-derivative (PID) control logic. The principle is that the proportional element responds to the current deviation, the integral element eliminates the accumulated deviation, and the derivative element predicts the deviation change trend. These three elements work together to calculate the power adjustment amount, which is the overshoot risk suppression signal. The input data for the overshoot suppression feedback mechanism are the actual temperature data, the target temperature, and the heating power adjustment trajectory. The output data is the overshoot risk suppression signal, which is a quantified power adjustment command used to indicate the magnitude of the heating power that needs to be increased or decreased.

[0052] In the above, the larger the cumulative deviation of the inner wall heat transfer resistance, the more severe the heat transfer obstruction, and the higher the risk of local overheating caused by heat input overshoot. Weighted processing is a mathematical method that assigns weights to each parameter based on their impact on the overshoot risk and performs a comprehensive calculation. Its core principle is to assign higher weights to parameters with greater impact, ensuring that the comprehensive assessment value accurately reflects the actual risk. The weighting coefficients here are determined by fitting multiple sets of industrial experimental data. For example, the overshoot risk suppression signal directly reflects the current temperature deviation and power adjustment requirements, so its weighting coefficient is set to a higher value. The cumulative deviation of the inner wall heat transfer resistance reflects the amplification effect of long-term heat transfer characteristics on the risk, so its weighting coefficient is next. The temperature rise rate in the overshoot risk signal assessment data has a higher weight than the peak deviation of the heating power. The sum of all weighting coefficients is 1. The specific process of weighted processing is as follows: First, the overshoot risk suppression signal, the cumulative deviation of the inner wall heat transfer resistance, the temperature rise rate, and the deviation between the peak and average values ​​of the heating power are normalized separately. This involves dividing the actual value of each parameter by its preset maximum value to eliminate dimensional differences. Then, each normalized parameter is multiplied by its corresponding weighting coefficient. Finally, all the products are summed, and the resulting total is the comprehensive evaluation value. The input data for this weighted processing are the overshoot risk suppression signal, the cumulative deviation of the inner wall heat transfer resistance, the temperature rise rate, and the deviation between the peak and average values ​​of the heating power; the output data is the comprehensive evaluation value.

[0053] The preset overshoot threshold is a critical risk value set based on the reaction process requirements of the reactor, the thermal stability of the materials, and the safe operation standards of the equipment. Its value is determined through industrial experiments. When the comprehensive evaluation value exceeds this threshold, it indicates that the risk of heating power overshoot has reached a level requiring intervention. If the comprehensive evaluation value exceeds the preset overshoot threshold, it is determined that there is a potential risk of heating power overshoot in the reactor. At this time, the overshoot suppression feedback mechanism is triggered, and the input of heating power is adjusted according to the overshoot risk suppression signal to initially reduce the overshoot risk and reserve buffer space for the subsequent generation of precise control commands.

[0054] Once a potential heating power overshoot is identified, a steady-state deviation correction vector needs to be generated based on the local thermal change rate data of the drastically fluctuating region. This vector is then used to update the actuator's control pulse sequence, resulting in a real-time temperature regulation command set. The drastically fluctuating region refers to a localized area where the temperature standard deviation within a preset time period exceeds a preset multiple of the overall average temperature standard deviation. This region exhibits the worst temperature stability and is the core area affected by the overshoot risk. The local thermal change rate data represents the temperature change per unit time at each sampling point within this region, accurately reflecting the dynamic characteristics of temperature fluctuations. The target temperature is the ideal temperature value inside the reactor set according to the reaction process requirements, serving as the benchmark for temperature control. The steady-state deviation correction vector is a vector data used to quantify the temperature deviation between the drastically fluctuating region and the target temperature, guiding the actuator to make corrections. Each element corresponds to a quantified temperature deviation value at a local location, and the vector dimension is consistent with the number of sampling points within the drastically fluctuating region. The specific process for generating the steady-state deviation correction vector is as follows: calculate the difference between the real-time temperature data and the target temperature at each acquisition point in the area of ​​severe fluctuation, arrange all the differences in order of spatial location of the acquisition points to form an initial deviation vector, and then eliminate the interference of instantaneous fluctuations through data smoothing to finally obtain the steady-state deviation correction vector. Therefore, the input data are the local thermal change rate data, real-time temperature data and target temperature in the area of ​​severe fluctuation, and the output data is a quantization vector that characterizes the temperature correction requirements of each local location.

[0055] By employing an adaptive control algorithm, combining the current temperature change with the preset temperature target range, the weights of the steady-state deviation correction vector are adjusted in real time to optimize the steady-state deviation correction vector. The adaptive control algorithm is one that dynamically adjusts control parameters based on the system's operating state. Its core principle is to continuously monitor the fit between the current temperature change trend and the preset temperature target range, automatically adjusting the weights of each element in the steady-state deviation correction vector to better match the actual temperature control needs and improve correction accuracy. The preset temperature target range is an allowable fluctuation interval set around the target temperature. Its width is determined based on material characteristics and reaction requirements. For example, for heat-sensitive materials, the interval width is set narrower to ensure accurate temperature control. Specifically, the adaptive control algorithm includes: first, collecting the current temperature change rate and temperature value to determine if it is within the preset temperature target range; if the temperature at a certain local location is close to the upper limit of the target range, increasing the weight of the corresponding vector element at that location and prioritizing a reduction in heating power in that area; if the temperature is close to the lower limit of the target range, appropriately reducing the weight to avoid over-correction leading to a lower temperature; then, optimizing the weight allocation using gradient descent to ensure that the power adjustment corresponding to the correction vector allows the temperature to quickly return to the target range without overshoot. The input data consists of the steady-state deviation correction vector, the current temperature change data, and the preset temperature target range. The output data is the optimized steady-state deviation correction vector.

[0056] For example, the weight of each local position i Based on its current temperature (Taken from real-time temperature field) and preset temperature target range The dynamic calculation of the relationship is as follows: ,in, The deviation of local position i The weight it occupies in the optimized vector is a value between 0 and... Positive numbers between The base weighting coefficient is a preset normal value that determines the baseline range for weight adjustment. This is the weight decay adjustment coefficient, which is a preset positive number. The larger the value, the faster the weight decays as the temperature deviates from the target. Let i be the current temperature at local location i. The target temperature for the process, This is the half-width of the preset temperature target range, i.e., the maximum absolute value of the allowable deviation. It is an exponential function.

[0057] The optimized steady-state deviation correction vector is input into the control system to update the actuator's control pulse sequence. Pulse width modulation (PWM) technology is used to adjust the duty cycle of the updated control pulse sequence to control the heating power and stirring speed. The actuator, including a heating element and a stirring motor, is a device that receives control commands and performs adjustments to heating power and stirring speed. The control pulse sequence is an electrical signal sequence used to drive the actuator's actions; the presence and width of the pulses directly determine the actuator's operating state. PWM technology is a technique that adjusts output power by changing the duty cycle of the pulse sequence. Its core principle is to adjust the duration (duty cycle) of the high-level pulse within a fixed pulse period; a higher duty cycle results in greater output power, and vice versa. After the optimized steady-state deviation correction vector is input into the control system, the system converts the quantized deviation values ​​in the vector into corresponding power and speed adjustment values. Then, using pulse width modulation (PWM), these adjustment values ​​are converted into changes in the duty cycle of the control pulse sequence. For example, for areas requiring reduced heating power, the duty cycle of the corresponding heating element control pulse is decreased; for areas requiring increased stirring speed to enhance heat transfer, the duty cycle of the stirring motor control pulse is increased. This achieves precise control of heating power and stirring speed. The input data for this process is the optimized steady-state deviation correction vector, and the output data is the updated actuator control pulse sequence.

[0058] A dynamic feedback adjustment mechanism is introduced into the updated control pulse sequence to generate a real-time temperature regulation command set. This dynamic feedback adjustment mechanism is a closed-loop mechanism that monitors the control effect and corrects control commands in real time. Based on the PWM duty cycle commands generated from feedforward and steady-state deviation, it introduces a fast proportional feedback fine-tuning loop to compensate for model errors and instantaneous disturbances in real time. Its core principle is to continuously collect in-vessel temperature data and actuator operating status data during the execution of the control pulse sequence to determine whether the actual temperature change meets expectations. If the temperature drop rate at a certain local location is too slow, it indicates insufficient power adjustment, requiring a further increase in the control pulse duty cycle adjustment amplitude. If the temperature drop is too fast, the adjustment amplitude needs to be reduced to avoid overcorrection. The final heating duty cycle commands and stirring duty cycle commands for all regions are packaged according to actuator address and timing, thus forming a real-time temperature regulation command set that can directly drive the actuators in joint debugging. The input data for this mechanism includes real-time temperature data, actuator operating status data, and a preset temperature target range. The output data is a corrected control pulse sequence. This corrected control pulse sequence is categorized and organized according to the controlled object (heating element, stirring motor) and the controlled area to form a real-time temperature adjustment instruction set. The real-time temperature adjustment instruction set is a set of instructions that includes all local area heating power adjustment parameters and stirring speed adjustment parameters. It can directly drive the actuator to perform coordinated actions, thereby achieving precise suppression of heating power overshoot and balanced control of the temperature inside the vessel.

[0059] For example, Figure 3 This is a schematic diagram comparing the heating power adjustment trajectory and the temperature response curve. Figure 3 In the middle, the upper subfigure (a) is a comparison diagram of heating power adjustment trajectory, showing the dynamic heating power adjustment trajectory generated by the feedforward compensation power correction coefficient of this method. The power trajectory of the traditional method (gray dashed line) is used as a comparison to show the power change process without the overshoot suppression mechanism of this application. After time t2, the power continues to rise and fluctuates significantly. The lower subfigure (b) is a comparison diagram of temperature response curves, showing the temperature change process of the key area (the area of ​​violent fluctuation) of the reactor when the control method of this application is used. The temperature change under the traditional control method shows obvious overshoot after time t2, the temperature peak exceeds the safe range, and the subsequent fluctuation is violent. Figure 3 This fully verifies the effectiveness of the overshoot suppression mechanism in this application.

[0060] Step S5: Based on the real-time temperature regulation instruction set, drive the joint actuator of heating power and stirring speed to operate, monitor abnormal fluctuations in local thermal change rate, and if the abnormal fluctuation increases, trigger the abnormal power reduction interlock mechanism to achieve stable overall operation output of the reactor.

[0061] Among these, achieving stable operation output that brings the overall thermal state of the reactor to a near-equilibrium includes: The real-time temperature control instruction set includes heating power adjustment parameters and stirring speed adjustment parameters; it monitors the local thermal change rate of the reactor in real time, calculates the temperature change rate at each local location, and, in conjunction with a preset fluctuation threshold, determines whether there are abnormal fluctuations at local locations. Abnormal fluctuations are manifested as a temperature change rate exceeding the preset threshold. If abnormal fluctuations exist, the heating power and stirring speed are adjusted in conjunction with the abnormal power reduction interlock mechanism to ensure that the temperature data fluctuates smoothly within a preset safety range. If the temperature data at all local locations fluctuates smoothly, the overall thermal state of the reactor is determined to be stable, and the local thermal change rate is continuously monitored to perform precise temperature control adjustments, thereby achieving stable overall operation and output of the reactor.

[0062] Specifically, the real-time temperature control command set is a complete set of commands including heating power adjustment parameters and stirring speed adjustment parameters. The heating power adjustment parameters are generated based on the previously optimized steady-state deviation correction vector and are used to specify the output power of the heating elements in each local area, directly determining the distribution of heat input within the vessel. The stirring speed adjustment parameters guide the adjustment of the stirring motor speed, affecting heat transfer efficiency by changing the material mixing intensity. The two work together to achieve precise temperature control within the vessel. The co-operation actuator, as the core device for receiving and executing this command set, integrates the heating power control module and the stirring speed control module. Its core function is to synchronously respond to both types of parameter commands, avoiding temperature distribution imbalances caused by adjusting only one parameter. The working principle of the joint-adjustment actuator is as follows: after receiving the real-time temperature adjustment command set, it converts the heating power adjustment parameters into the driving signal of the heating element (such as voltage or current signal), and at the same time converts the stirring speed adjustment parameters into the frequency conversion control signal of the motor, ensuring that the heating power change and stirring speed adjustment are synchronized, so that the material can achieve uniform heat diffusion through the adaptation of mixing intensity while the heat input changes. The input data of the actuator are two types of parameters in the real-time temperature adjustment command set, and the output data are the actual power output of the heating element and the actual speed of the stirring motor.

[0063] While the actuators are being tested, the local thermal change rate of the reactor needs to be monitored in real time. The local thermal change rate refers to the temperature change per unit time at each sampling point within the reactor. Its calculation method is the same as described above: the temperature difference between two consecutive moments is divided by the time interval. This parameter can capture the dynamic temperature changes at various local locations in real time and is the core basis for judging whether the temperature is stable. During monitoring, the temperature change rate at each local location needs to be calculated simultaneously and combined with a preset fluctuation threshold to determine if there are any abnormal fluctuations. The preset fluctuation threshold is a critical value for the temperature change rate set according to the reaction process requirements, material thermal stability, and equipment safety standards. For example, for easily decomposable materials, this threshold is set to a lower value to avoid rapid temperature changes that could trigger side reactions. Abnormal fluctuations are specifically manifested when the temperature change rate at a local location exceeds the preset rate threshold, meaning the rate of temperature rise or fall exceeds the safe control range, which may lead to local overheating or overcooling.

[0064] If abnormal fluctuations are detected, the heating power and stirring speed must be adjusted immediately based on the abnormal power reduction interlock mechanism to ensure that the temperature fluctuates smoothly within a preset safety range. The abnormal power reduction interlock mechanism is a closed-loop protection mechanism based on safety priority. Its core principle is that when abnormal temperature fluctuations are detected, the heating power is reduced first to decrease heat input, while the stirring speed is adjusted to enhance material mixing, accelerate heat diffusion, and prevent local heat accumulation. This mechanism is built based on multiple sets of safety experimental data, and the timeliness and accuracy of the action are ensured through preset interlock adjustment rules. Its specific workflow is as follows: First, an abnormal fluctuation detection signal is received, which includes the local location of the abnormal fluctuation, the rate of temperature change, and the current heating power and stirring speed data. Then, according to preset rules, the heating power is quickly reduced to a preset safety ratio of the rated value (this ratio is determined experimentally to ensure that the reaction continuity is not affected while reducing heat input). Simultaneously, the stirring speed is adjusted according to the location and range of the abnormal area. If the abnormal area is a localized accumulation zone, the speed is appropriately increased to enhance material convection in that area and accelerate heat transfer. If the abnormal area is widespread, the speed is maintained or slightly adjusted to avoid excessive heat generation due to over-stirring. The input data for this interlocking mechanism includes abnormal fluctuation detection signals, current heating power, stirring speed, and preset safety range. The output data includes heating power adjustment commands and stirring speed adjustment commands. After execution by the interlocking actuator, the temperature in the abnormal area quickly returns to the preset safety range and maintains stable fluctuations.

[0065] The preset safety range is the allowable temperature fluctuation interval set around the target temperature. Its width is determined according to the material characteristics and reaction requirements. For example, for precision chemical reactions, the interval width is set narrower to ensure the consistency of product quality. This range is the benchmark for judging whether the temperature is stable; temperature fluctuations within this range are considered stable fluctuations. If the temperature at all local locations is in a stable fluctuation state, that is, the rate of temperature change at each local location is less than the preset rate threshold, and the temperature value is within the preset safety range, then the overall thermal state of the reactor is considered stable. At this time, it is still necessary to continuously monitor the local thermal change rate and maintain temperature stability through precise adjustment. Precise adjustment adopts a proportional-integral-derivative (PID) control algorithm, the principle of which is consistent with the closed-loop control logic mentioned above. The proportional element responds to the small deviation between the current temperature and the target temperature, the integral element eliminates long-term accumulated deviations, and the derivative element predicts the trend of deviation changes. The three elements work together to calculate the small power and speed adjustments, thereby achieving fine temperature control. The input data for this PID control algorithm are the real-time local thermal change rate, the current temperature data, and the target temperature. The output data are the small adjustments to the heating power and the stirring speed. After execution by the joint-tuning actuator, it ensures that the temperature inside the reactor remains stable near the target temperature for a long time, and finally achieves a stable operation output with the overall thermal state of the reactor tending to be balanced.

[0066] Through a closed-loop logic of "command execution - real-time monitoring - abnormal interlocking - precise fine-tuning", the reactor is ensured to maintain a constant temperature balance under dynamic heat load. This not only avoids local overheating caused by overshooting of heating power, but also enhances heat diffusion through coordinated adjustment of stirring speed, effectively solving the problem of uneven heat distribution in the reactor space and ensuring the stability and safety of the reaction process.

[0067] The above describes a method for controlling the temperature of an industrial reactor according to an embodiment of this application. The following describes a method for controlling the temperature of an industrial reactor according to an embodiment of this application. Please refer to [link / reference]. Figure 4 One embodiment of an industrial reactor temperature control system in this application includes: The data acquisition unit is used to acquire multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and to construct the initial temperature field.

[0068] The anomaly analysis unit is used to acquire the rate of thermal change at each local location in real time based on the initial temperature field, calculate the temperature gradient intensity distribution, determine whether there is a risk area in the reactor based on the temperature gradient intensity distribution, and if so, extract local data where the stirring speed and shear force distribution do not match within the risk area, and combine it with the rapid heating control logic to analyze potential abnormal signals of temperature changes within a preset short period of time.

[0069] The correction unit is used to perform feedforward compensation power correction on the jacket medium flow synchronization parameters based on potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, and to generate a heating power adjustment trajectory.

[0070] The risk assessment unit is used to obtain overshoot risk signal assessment data based on the heating power adjustment trajectory. Combined with the overshoot suppression feedback mechanism and the weighted processing of the cumulative deviation of heat transfer impedance, it determines whether there is a potential overshoot of heating power in the reactor. If so, it generates a steady-state deviation correction vector based on the local thermal change rate data of the area with severe fluctuations, updates the control pulse sequence of the actuator, and obtains a real-time temperature adjustment command set.

[0071] The temperature control unit is used to drive the actuator that coordinates the heating power and stirring speed according to the real-time temperature adjustment command set, monitor abnormal fluctuations in the local thermal change rate, and trigger the abnormal power reduction interlock mechanism if the abnormal fluctuation increases, so as to achieve stable operation output of the reactor as a whole.

[0072] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0073] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0074] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for temperature control in an industrial reactor, characterized in that, The method includes: Acquire multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and construct the initial temperature field; Based on the initial temperature field, the rate of thermal change at each local location is obtained in real time, the distribution of temperature non-uniform gradient intensity is calculated, and the presence of a risk area in the reactor is determined based on the distribution of temperature non-uniform gradient intensity. If a risk area exists, local data where the stirring speed and shear force distribution do not match are extracted from the risk area, and combined with the rapid heating control logic, potential abnormal signals of temperature change within a preset short period of time are analyzed. Based on the potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, the jacket medium flow synchronization parameters are fed forward compensation power correction to generate a heating power adjustment trajectory. Based on the heating power adjustment trajectory, overshoot risk signal assessment data is obtained. Combined with the overshoot suppression feedback mechanism and the weighted processing of the cumulative deviation of the inner wall heat transfer impedance, it is determined whether there is a heating power overshoot risk in the reactor. If there is, a steady-state deviation correction vector is generated based on the local thermal change rate data of the drastic fluctuation area, the control pulse sequence of the actuator is updated, and a real-time temperature adjustment command set is obtained. Based on the real-time temperature regulation instruction set, the actuator that drives the heating power and stirring speed is activated to monitor abnormal fluctuations in the local thermal change rate. If the abnormal fluctuation increases, the abnormal power reduction interlock mechanism is triggered to achieve stable overall operation and output of the reactor.

2. The method according to claim 1, characterized in that, Constructing the initial temperature field includes: A multi-point temperature acquisition layout is distributed along the axial and radial directions inside the reactor to collect temperature sequence data at different heights and radial positions and summarize them to form the multi-point temperature sequence data. The jacket medium flow rate information and the dynamic adjustment parameters of the stirring speed are acquired synchronously and set as the synchronous dynamic parameters; The multi-point temperature sequence data and the synchronous dynamic parameters are weighted and averaged and fitted using inverse distance weighted interpolation and spatial interpolation algorithms to generate temperature data for each collection point inside the reactor. The temperature data in different time periods are time-series aligned based on the dynamic time warping algorithm to obtain the temperature field mapping result inside the reactor. The temperature field mapping results are corrected and optimized by integrating the environmental thermal disturbance and heat flow transfer characteristic model, and the heat transfer impedance and stirring effect are combined to obtain the initial temperature field containing spatial non-uniform heat distribution.

3. The method according to claim 1, characterized in that, Determining whether there are risk areas in the reactor includes: Gradient calculations are performed on the temperature data at various local locations within the reactor based on the initial temperature field to obtain the rate of thermal change at each local location. Using an online thermal disturbance estimator, based on real-time ambient temperature, jacket average temperature and pre-calibrated equipment coefficients, the amount of external thermal disturbance affecting each local location is estimated. The initial temperature field is then corrected based on the amount of thermal disturbance, resulting in the corrected temperature field. The temperature non-uniform gradient intensity at each local location is calculated based on the corrected temperature field, and a weighted correction is performed according to the amount of disturbance to obtain the temperature non-uniform gradient intensity distribution. Based on preset intensity values, identify intensity regions with uneven temperature gradients in the temperature unevenness gradient intensity distribution and set them as risk regions. Risk regions include local overheating regions or local cold spot regions.

4. The method according to claim 1, characterized in that, Analyze potential abnormal signals of temperature changes within a preset short time period, including: The local data includes local stirring speed data and temperature change sequence data, and the local temperature change rate is calculated based on the local data. Based on the rapid heating control logic, a temperature change prediction model is constructed. The prediction model uses an exponential weighted average method to weight the local temperature change rate within a historical time period to generate a temperature change prediction curve for a preset short time period. The potential abnormal signals of the temperature change prediction curve are identified based on a preset temperature rise rate.

5. The method according to claim 1, characterized in that, Generate a heating power adjustment trajectory, including: Based on the local data and the potential abnormal signals, the temperature fluctuation amplitude of each local area is calculated. Based on the heat transfer situation of the inner wall of the reactor, the cumulative deviation of the heat transfer resistance of the inner wall is calculated by integrating the difference between the actual total heat transfer and the theoretical heat transfer, combined with the running time. The potential abnormal signal, the cumulative deviation of the inner wall heat transfer impedance, and the temperature fluctuation amplitude are normalized and synthesized using a weighted average method to generate a feedforward compensation power correction coefficient. Based on the feedforward compensation power correction coefficient, the jacket medium flow synchronization parameters are dynamically adjusted to generate a heating power adjustment trajectory that matches the dynamic heat load.

6. The method according to claim 1, characterized in that, Determining whether a reactor has a potential heating power overshoot risk includes: Based on real-time monitoring of the heating power adjustment trajectory, the overshoot risk signal assessment data is extracted, which includes the temperature rise rate and the deviation between the peak heating power and the average heating power. Based on the overshoot suppression feedback mechanism, the deviation between heating power and temperature data is collected in real time, and feedback adjustment is performed through closed-loop control to generate an overshoot risk suppression signal. The overshoot risk suppression signal and the cumulative deviation of the inner wall heat transfer impedance are weighted and combined with the overshoot risk signal evaluation data to generate a comprehensive evaluation value. If the comprehensive evaluation value exceeds the preset overshoot threshold, it is determined that there is a potential overshoot risk in the reactor, triggering the overshoot suppression feedback mechanism to adjust the input of heating power.

7. The method according to claim 1, characterized in that, Obtain the real-time temperature adjustment instruction set, including: Based on the local thermal change rate data, the difference between the temperature data and the target temperature in the area of ​​violent fluctuation is calculated, and the steady-state deviation correction vector is generated based on the difference. By using an adaptive control algorithm, combining the current temperature change and the preset temperature target range, the weight of the steady-state deviation correction vector is adjusted in real time to optimize the steady-state deviation correction vector. The optimized steady-state deviation correction vector is input into the control system to update the control pulse sequence of the actuator. The duty cycle of the updated control pulse sequence is adjusted by pulse width modulation technology to control the heating power and stirring speed. In the updated control pulse sequence, a dynamic feedback adjustment mechanism is introduced to generate the real-time temperature regulation instruction set.

8. The method according to claim 1, characterized in that, To achieve stable operation and output with the overall thermal state of the reactor approaching equilibrium, including: The real-time temperature control instruction set includes heating power adjustment parameters and stirring speed adjustment parameters; The local thermal change rate of the reactor is monitored in real time, the temperature change rate of each local location is calculated, and combined with a preset fluctuation threshold, it is determined whether there is abnormal fluctuation at the local location. The abnormal fluctuation is manifested as the temperature change rate being greater than the preset rate threshold. If abnormal fluctuations occur, the heating power and stirring speed will be adjusted in conjunction with the abnormal power reduction interlock mechanism to ensure that the temperature data fluctuates stably within the preset safety range. If the temperature data at all local locations fluctuate steadily, the overall thermal state of the reactor is considered stable. The rate of local thermal change is continuously monitored, and precise temperature control is implemented.

9. An industrial reactor temperature control system, used to implement the industrial reactor temperature control method as described in any one of claims 1-8, characterized in that, The system includes: The data acquisition unit is used to acquire multi-point temperature sequence data and synchronous dynamic parameters of the reactor, and to construct the initial temperature field; Anomaly analysis unit is used to acquire the rate of thermal change at each local location in real time based on the initial temperature field, calculate the temperature non-uniform gradient intensity distribution, determine whether there is a risk area in the reactor based on the temperature non-uniform gradient intensity distribution, and if so, extract local data where the stirring speed and shear force distribution do not match in the risk area, and analyze potential abnormal signals of temperature change within a preset short period of time in combination with rapid heating control logic. The correction unit is used to perform feedforward compensation power correction on the jacket medium flow synchronization parameters based on the potential abnormal signals and the cumulative deviation of the inner wall heat transfer impedance, and generate a heating power adjustment trajectory. The risk assessment unit is used to obtain overshoot risk signal assessment data based on the heating power adjustment trajectory, and combine the overshoot suppression feedback mechanism with the weighted processing of the cumulative deviation of heat transfer impedance to determine whether there is a potential heating power overshoot in the reactor. If there is, a steady-state deviation correction vector is generated based on the local thermal change rate data of the drastic fluctuation area, the control pulse sequence of the actuator is updated, and a real-time temperature adjustment instruction set is obtained. The temperature control unit is used to drive the actuator that coordinates the heating power and stirring speed according to the real-time temperature adjustment command set, monitor abnormal fluctuations in the local thermal change rate, and trigger the abnormal power reduction interlock mechanism if the abnormal fluctuation increases, so as to achieve stable operation output of the reactor as a whole.

Images