Cyclic heating system of glue mixing and preparing reaction kettle

Abstract

The invention relates to the technical field of glue mixing and preparing equipment, and discloses a cyclic heating system of a glue mixing and preparing reaction kettle. The system comprises a temperature acquisition module for acquiring and processing temperature data to generate a standardized temperature sequence; the heat distribution module generates a heat distribution proportion of each heating area according to a heat conduction equation; the dynamic adjusting module is used for optimizing the heat distribution proportion and dividing a real-time adjusting interval; the loop optimization module is used for constructing a loop control unit to generate a steady-state heating control scheme; and the feedback control module outputs a regulation and control instruction. And the early warning module is used for comparing the regulation and control instruction with a thermal stress threshold value, outputting an alarm signal and triggering an emergency protection mechanism. The system can accurately control the temperature of the reaction kettle, improve the temperature uniformity, dynamically respond to the material change, reduce the energy consumption, guarantee the equipment safety and improve the glue mixing and making production quality and efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of glue mixing and glue making equipment, and specifically to a circulating heating system for a glue mixing and glue making reactor. Background Art

[0002] In the industrial production process of glue mixing and glue making, the heating control of the reactor is crucial, which directly affects the quality of the glue and the production efficiency. There are many problems in the traditional heating system of the glue mixing and glue making reactor, and it is difficult to meet the growing high-quality production requirements of modern industry.

[0003] From the aspect of temperature control accuracy, the traditional heating system usually adopts a relatively simple temperature control method. For example, it only performs heating control based on the single-point temperature measurement value inside the reactor. This method ignores the temperature differences in different regions inside the reactor and cannot accurately adjust the heat supply to each heating region. In actual production, the chemical reaction of the materials inside the reactor is extremely sensitive to temperature. Even a tiny temperature fluctuation or uneven distribution may cause changes in the chemical structure of the glue, thereby affecting the performance of the glue, such as bonding strength, curing time, etc. For example, when producing high-performance electronic adhesives, if the temperature control accuracy inside the reactor is insufficient, it may cause abnormalities in the curing process of the adhesive, unable to meet the high-precision requirements of electronic products for adhesives, and resulting in problems such as component detachment during the use of electronic products.

[0004] In response to the changes in material properties, the traditional heating system performs rather poorly. During the production of different types of glues, there are significant differences in the physical and chemical properties of the materials, such as viscosity and specific heat capacity. Moreover, during the reaction process, these properties of the materials will also change dynamically as the reaction progresses. The traditional system often cannot adjust the heating strategy in a timely and accurate manner according to the changes in material properties. Taking the production of high-viscosity glue as an example, due to the poor fluidity of the material and the difficulty of heat transfer, the traditional heating system may not be able to ensure that the heat is evenly transferred to all parts of the material, resulting in local overheating or overcooling phenomena, affecting the quality stability and production efficiency of the glue. If the specific heat capacity of the material changes during the reaction process, it is also difficult for the traditional system to automatically adapt to this change and reasonably adjust the heating power, thereby causing waste of energy and an increase in production costs.

[0005] The energy utilization efficiency is also a major pain point of traditional heating systems. Their heating methods lack refined analysis of the heat distribution in the reactor and the material requirements, and usually adopt fixed-power or extensive heating modes. This means that when a large amount of heat is not required, the system may still continue to heat at a high power, resulting in ineffective energy consumption; while when rapid temperature rise is needed, the slow temperature rise may be caused due to insufficient heating power, prolonging the production cycle and further increasing energy consumption. Against the backdrop of the global advocacy for energy conservation and emission reduction, this high-energy-consuming heating method not only increases the production costs of enterprises, but also does not conform to the concept of sustainable development, restricting the market competitiveness of enterprises.

[0006] In addition, there are also potential hazards in the equipment safety and stability of traditional heating systems. Due to the lack of effective monitoring and control of the thermal stress of the equipment, during long-term and high-intensity heating processes, the reactor and its related components are prone to problems such as fatigue damage, deformation, and even rupture due to excessive thermal stress. This will not only affect the normal service life of the equipment, increase the equipment maintenance cost, but also may trigger safety accidents, posing a threat to the lives and safety of production personnel. Once equipment failures occur, production will be interrupted, bringing huge economic losses to enterprises.

[0007] With the continuous improvement of the quality requirements of glue mixing and glue making products in modern industry, as well as the increasing attention to aspects such as production efficiency, energy utilization efficiency, and equipment safety, it is extremely urgent to develop a more advanced, efficient, and intelligent circulating heating system for glue mixing and glue making reactors. This new type of system needs to be able to accurately control the temperature, flexibly adapt to changes in material characteristics, improve energy utilization efficiency, and ensure the safe and stable operation of the equipment to meet the diverse needs of industrial production. Summary of the Invention

[0008] The purpose of the present invention is to provide a circulating heating system for a glue mixing and glue making reactor to solve the problems raised in the above background technology.

[0009] To achieve the above purpose, the present invention provides the following technical solution: A circulating heating system for a glue mixing and glue making reactor, the system includes:

[0010] A temperature acquisition module, which is used to obtain the temperature data set within a preset time range in the reactor, and perform noise reduction processing on the temperature data set according to the time series filtering strategy to generate a standardized temperature sequence;

[0011] A heat distribution module, which is used to input the standardized temperature sequence into a dynamic weight distribution model, and generate the heat distribution ratio of each heating area based on the heat conduction equation, and the heat distribution ratio includes the multi-dimensional correlation relationship of temperature gradient, material viscosity, and heating power;

[0012] A dynamic adjustment module, configured to perform non-linear programming optimization on the heat distribution ratio according to preset temperature control constraints, divide a real-time adjustment interval through a gradient descent algorithm, and extract a set of temperature deviation indicators within each interval;

[0013] A cyclic optimization module, configured to construct a cyclic control unit including a hierarchical optimization strategy, and iteratively update the hierarchical optimization strategy by using the set of temperature deviation indicators to generate a steady-state heating control scheme;

[0014] A feedback control module, configured to perform multivariable coupling analysis on the steady-state heating control scheme according to a model predictive control framework, and output a reaction kettle cyclic heating regulation instruction.

[0015] Preferably, generating the heat distribution ratio of each heating area based on the heat conduction equation includes:

[0016] Extracting timestamps, heating area numbers, and temperature change rates in the standardized temperature sequence to construct a multi-dimensional heat conduction matrix;

[0017] Using a moving average algorithm to smooth the multi-dimensional heat conduction matrix to generate a set of steady-state heat conduction sub-matrices;

[0018] Adopting the least squares method to perform parameter fitting on the set of steady-state heat conduction sub-matrices and removing abnormal fluctuation points;

[0019] Performing weight allocation on the remaining sub-matrices through a fuzzy inference algorithm to generate a heat distribution ratio with temperature equilibrium.

[0020] Preferably, the non-linear programming optimization includes:

[0021] Calculating the extreme values and fluctuation ranges of temperature deviation indicators within each interval according to the historical temperature distribution of the real-time adjustment interval;

[0022] Based on a rolling horizon optimization algorithm, dynamically correcting the extreme values and fluctuation ranges to generate optimization constraint conditions;

[0023] Using the Lagrange multiplier method to decouple the optimization constraint conditions to generate a real-time adjustment interval division scheme.

[0024] Preferably, the gradient descent algorithm divides the real-time adjustment interval, including:

[0025] Defining the objective function as the weighted minimization of temperature deviation and energy consumption within the adjustment interval;

[0026] Setting the constraint conditions as the heating power upper limit, the material phase change threshold, and the equipment safe temperature range;

[0027] Solve the objective function by the conjugate gradient method and output the optimal adjustment interval division result.

[0028] Preferably, the hierarchical optimization strategy is iteratively updated, including:

[0029] Input the temperature deviation index set into the input layer of the hierarchical optimization strategy, and use a recurrent neural network to extract the cyclic control feature vector;

[0030] Perform priority sorting on the cyclic control feature vector by the entropy weight method to generate a regional regulation sequence;

[0031] Adopt a particle swarm optimization algorithm combined with a simulated annealing algorithm to optimize the strategy parameters and update the regulation logic of the cyclic control unit.

[0032] Preferably, the time series filtering strategy includes:

[0033] Identify the high-frequency noise segments in the temperature data set and smooth them based on the moving window mean rule;

[0034] Detect missing temperature points and interpolate the missing temperature points by the cubic spline interpolation method;

[0035] Normalize the interpolated data to generate a standardized temperature sequence with a mean of zero and a variance of one.

[0036] Preferably, the model predictive control framework includes:

[0037] Construct a multivariable coupling analysis model based on the state space equation to quantify the sensitivity of the regulation instructions to each heating parameter;

[0038] Generate a parameter perturbation sample set by Latin hypercube sampling and calculate the coupling contribution degree of each parameter;

[0039] Select the parameters with a contribution degree higher than the preset threshold as the core regulation variables.

[0040] Preferably, the multivariable coupling analysis includes:

[0041] Integrate the temperature sensor data, material property data, and heater status data to construct a multi-dimensional coupling data matrix;

[0042] Use the principal component analysis method to extract the features of the multi-dimensional coupling data matrix and generate a dimensionality-reduced feature vector;

[0043] Input the dimensionality-reduced feature vector into the feedback control module to generate an optimized regulation instruction.

[0044] Preferably, the system further includes:

[0045] Build an early warning module that includes a device thermal stress evaluation function, compare the control instruction with the thermal stress threshold for analysis, and output a graded alarm signal;

[0046] Trigger a preset emergency protection mechanism according to the graded alarm signal to generate a power limit or cooling start instruction.

[0047] Preferably, the present invention further includes an electronic device, and the device includes:

[0048] At least one processor;

[0049] And a memory communicatively connected to the at least one processor;

[0050] Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the operations of the circulating heating system of the glue mixing and glue making reactor as described above.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] In terms of temperature control accuracy, the temperature acquisition module can effectively remove the noise in the temperature data through a time series filtering strategy, and at the same time use the cubic spline interpolation method to accurately interpolate the missing temperature points, and then obtain a standardized temperature sequence after normalization processing. This makes the subsequent heat distribution and regulation more accurate, greatly improving the reliability of temperature monitoring. Based on this, the heat distribution module generates a heat distribution ratio according to the standardized temperature sequence, combined with the heat conduction equation, considering the multi-dimensional correlation relationship of temperature gradient, material viscosity and heating power, which can ensure that the temperature in each area of the reactor is more uniform. For example, when synthesizing a specific high-performance glue, the requirement for temperature uniformity is extremely high, and it is difficult for traditional systems to meet. However, through precise heat distribution, the temperature deviation in the reactor of this system is controlled within a very small range, ensuring the consistency of chemical reactions during the glue synthesis process, effectively improving the quality of the glue and reducing the defective rate.

[0053] In terms of dynamic regulation ability, the dynamic regulation module divides the real-time regulation interval according to the preset temperature control constraint conditions, using the nonlinear programming optimization method and combining the gradient descent algorithm. This method can quickly and accurately adjust the heating strategy according to the real-time changes of the materials and temperature in the reactor. When the material viscosity changes due to the reaction process, the system can sense it in time and redistribute the heat to ensure that the heating effect is not affected. During the entire production process, no matter what sudden situation causes the temperature to change, the system can respond quickly and always control the temperature within a suitable range, ensuring the stability and continuity of the production process, and avoiding production interruptions and product quality problems caused by temperature runaway.

[0054] The loop control unit of the loop optimization module, which constructs a hierarchical optimization strategy, is iteratively updated using the set of temperature deviation metrics to generate a steady-state heating control scheme. This process continuously optimizes the heating control logic, further enhancing the system's adaptability. As the number of production runs increases, the system can automatically adjust the control strategy based on historical data and real-time feedback to better suit the actual production requirements. For example, during long-term production, the performance of the equipment may gradually change, and this system can continuously optimize to maintain good heating control effects, extend the service life of the equipment, and reduce equipment maintenance costs.

[0055] In terms of energy consumption control, by defining the objective function as the weighted minimization of temperature deviation and energy consumption within the adjustment interval and setting reasonable constraint conditions, the conjugate gradient method is used to solve the objective function to divide the real-time adjustment interval, enabling the system to minimize energy consumption to the greatest extent while meeting the production temperature requirements. This not only conforms to the environmental protection concept of energy conservation and emission reduction but also saves a large amount of production costs for enterprises. In the current context of rising energy prices, this energy-saving effect is of great significance for enhancing the economic benefits of enterprises.

[0056] In addition, the warning module added to the system uses the equipment thermal stress evaluation function to compare and analyze the control instructions with the thermal stress threshold, output a graded alarm signal, and can trigger a preset emergency protection mechanism. This provides comprehensive safety protection for the reactor equipment, effectively avoiding the risk of equipment damage caused by abnormal temperature, ensuring the safety of production personnel, and reducing economic losses and production delays caused by equipment failures. Overall, the circulating heating system of the present invention comprehensively improves the performance of the glue mixing and glue making reactor, providing strong support for the efficient, stable, and sustainable development of the industry. Brief Description of the Drawings

[0057] Figure 1 It is the working principle diagram of the circulating heating system of the glue mixing and glue making reactor described in the present invention;

[0058] Figure 2 It is the working principle diagram of the non-linear programming optimization;

[0059] Figure 3 It is the working principle diagram of dividing the real-time adjustment interval by the gradient descent algorithm;

[0060] Figure 4 It is the working principle diagram of the time series filtering strategy. Detailed Implementation Modes

[0061] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0062] Please refer to Figures 1-4 , the present invention provides a circulating heating system for an adhesive mixing and preparation reactor, aiming to achieve precise and efficient control of the temperature inside the reactor to ensure the stability of the adhesive preparation process and product quality. The overall implementation solution is as follows:

[0063] The temperature acquisition module is responsible for obtaining the temperature data set within a preset time range inside the reactor. In practical applications, it can be achieved by installing multiple high-precision temperature sensors at different positions inside the reactor. These sensors collect temperature data in real time and transmit the data to the temperature acquisition module. The collected temperature data set may contain noise. Therefore, the temperature acquisition module will perform denoising processing on the temperature data set according to the time series filtering strategy. Its specific operation is to identify the high-frequency noise segments in the temperature data set and smooth them based on the moving window mean rule; detect missing temperature points and use cubic spline interpolation to interpolate the missing temperature points; normalize the interpolated data to generate a standardized temperature sequence with a mean of zero and a variance of one, thereby providing a reliable data basis for subsequent precise analysis and control.

[0064] The heat distribution module inputs the standardized temperature sequence into the dynamic weight distribution model. This module generates the heat distribution ratio for each heating area based on the heat conduction equation, and this heat distribution ratio contains the multi-dimensional correlation relationship of temperature gradient, material viscosity, and heating power. In actual implementation, first, extract the time stamps, heating area numbers, and temperature change rates in the standardized temperature sequence to construct a multi-dimensional heat conduction matrix; then use the moving average algorithm to smooth the multi-dimensional heat conduction matrix to generate a set of steady-state heat conduction sub-matrices; then use the least squares method to perform parameter fitting on the set of steady-state heat conduction sub-matrices to eliminate abnormal fluctuation points; finally, use the fuzzy inference algorithm to assign weights to the remaining sub-matrices to generate a heat distribution ratio with temperature balance. In this way, it is possible to reasonably distribute the heating heat according to the actual temperature conditions in different areas of the reactor, improving the heating efficiency and temperature uniformity.

[0065] The dynamic adjustment module performs nonlinear programming optimization on the heat distribution ratio according to the preset temperature control constraints. Specifically, first, based on the historical temperature distribution in the real-time adjustment interval, the extreme values and fluctuation ranges of the temperature deviation indexes in each interval are calculated; then, based on the rolling horizon optimization algorithm, the dynamic boundary correction is performed on the extreme values and fluctuation ranges to generate the optimization constraints; finally, the Lagrange multiplier method is used to decouple the optimization constraints to generate the real-time adjustment interval division scheme. At the same time, this module divides the real-time adjustment interval through the gradient descent algorithm, defines the objective function as the weighted minimization of the temperature deviation and energy consumption in the adjustment interval, sets the constraints as the upper limit of the heating power, the material phase change threshold, and the equipment safety temperature range, and solves the objective function through the conjugate gradient method to output the optimal adjustment interval division result. Through these operations, it is possible to achieve dynamic optimization adjustment of the heating process while satisfying various constraints, improving the control accuracy and stability of the system.

[0066] The loop optimization module constructs a loop control unit containing a hierarchical optimization strategy, and uses the set of temperature deviation indexes to iteratively update the hierarchical optimization strategy to generate a steady-state heating control scheme. The specific implementation steps are as follows: input the set of temperature deviation indexes into the input layer of the hierarchical optimization strategy, and use a recurrent neural network to extract the loop control feature vector; perform priority sorting on the loop control feature vector through the entropy weight method to generate a regional regulation sequence; use the particle swarm algorithm combined with the simulated annealing algorithm to optimize the strategy parameters and update the regulation logic of the loop control unit. Through continuous iterative updates, the system can better adapt to the complex temperature changes in the reactor, achieving more stable and efficient heating control.

[0067] The feedback control module performs multivariable coupling analysis on the steady-state heating control scheme according to the model predictive control framework, and outputs the reactor loop heating regulation instructions. In actual operation, first, a multivariable coupling analysis model based on the state space equation is constructed to quantify the sensitivity of the regulation instructions to each heating parameter; a parameter perturbation sample set is generated through Latin hypercube sampling, and the coupling contribution degree of each parameter is calculated; the parameters with a contribution degree higher than the preset threshold are selected as the core regulation variables. Then, integrate the temperature sensor data, material property data, and heater status data to construct a multi-dimensional coupling data matrix; use the principal component analysis method to extract the features of the multi-dimensional coupling data matrix to generate a reduced-dimensional feature vector; input the reduced-dimensional feature vector into the feedback control module to generate optimized regulation instructions. Through this multivariable coupling analysis and feedback control mechanism, it is possible to achieve precise control of the reactor loop heating system and ensure that the glue-making process is carried out under the best temperature conditions.

[0068] The technical solution of the present invention will be further described below through specific embodiments.

[0069] Example 1:

[0070] This embodiment elaborates in detail the specific process by which the heat distribution module generates the heat distribution ratio of each heating area based on the heat conduction equation. In an actual glue mixing and glue making reactor, it is assumed that the reactor is divided into heating areas, and each area has an independent heating device and temperature sensor. First, obtain the standardized temperature sequence from the temperature acquisition module, and extract the timestamps , heating area numbers ( ), and temperature change rates to construct a multi-dimensional heat conduction matrix . Among them, the matrix element ( represents different moments in the time series) contains information such as the timestamp, area number, and temperature change rate of the corresponding area at the moment .

[0071] Use the moving average algorithm to smooth the multi-dimensional heat conduction matrix . The window size of the moving average algorithm is set to . For each column in the matrix (corresponding to all area data at a time point), calculate the average value of elements centered on the elements of this column to generate a new matrix, that is, the steady-state heat conduction sub-matrix set ( ). In this way, the noise fluctuations in the data can be effectively removed, and the data can better reflect the steady-state characteristics of heat conduction.

[0072] Use the least squares method to perform parameter fitting on the steady-state heat conduction sub-matrix set . Let the fitting model be (here represents temperature-related parameters, represents other related variables, which are specifically determined according to the actual heat conduction model). Determine the values of the parameters and by minimizing the sum of the squares of the errors ( is the number of data points). During the fitting process, some abnormal fluctuation points with large deviations from the overall trend will be identified and removed to ensure the accuracy of the fitting results.

[0073] Finally, the remaining sub-matrices are weighted by the fuzzy inference algorithm to generate a heat distribution ratio with temperature balance. The fuzzy inference algorithm first determines the fuzzy sets and membership functions of the input variables (such as temperature difference, temperature change rate, etc.) and output variable (heat distribution ratio) based on practical experience and heat conduction theory. For example, the temperature difference is divided into fuzzy sets such as "small", "medium", "large", etc., and reasoning calculations are carried out through fuzzy rules (such as "if the temperature difference is small and the temperature change rate is small, then the heat distribution ratio is a certain value"), and finally a suitable heat distribution ratio is determined for each heating area. The heat distribution ratio generated in this way can comprehensively consider the temperature conditions of each area and achieve balanced temperature control.

[0074] Example 2:

[0075] In the actual glue mixing and production environment, the real-time adjustment range of the reaction kettle will change continuously during the glue production process. The non-linear programming optimization of the heat distribution ratio by the dynamic adjustment module is the key link to ensure precise temperature control.

[0076] First, calculate the relevant indicators according to the historical temperature distribution of the real-time adjustment range. Suppose we divide the past hour into multiple sub-intervals with a length of 10 minutes to record temperature data, which is used as the historical temperature distribution data. For each sub-interval, based on the ideal temperature set for the reaction kettle as the benchmark, calculate the temperature deviation index . , where is the actually measured temperature within the sub-interval. By statistically calculating the maximum value and the minimum value of in each sub-interval, the extreme values of the temperature deviation index are obtained. The fluctuation range is obtained through the formula , which reflects the degree of temperature fluctuation within the sub-interval.

[0077] Then, dynamic boundary correction is performed based on the rolling horizon optimization algorithm. The rolling horizon optimization algorithm will continuously update the optimization range according to the newly obtained data. For example, after every 10 minutes of new measurement data is obtained, the oldest 10-minute data is removed, new data is added, and the extreme values and fluctuation range are recalculated. Let the current time be , considering the requirements of the glue production process for temperature stability and the characteristics of the equipment, the upper and lower limits of the temperature deviation index are adjusted according to the rolling horizon optimization algorithm to form optimization constraints. For example, set the upper limit of the temperature deviation and the lower limit . When When this occurs, the heating power needs to be increased. These upper and lower limits are not fixed values, but are continuously adjusted according to the rolling horizon optimization algorithm due to factors such as changes in the material properties during the glue-making process and fluctuations in the ambient temperature.

[0078] Finally, the Lagrange multiplier method is used to decouple the optimization constraints. Assume the objective function Comprehensively considers factors such as temperature stability and heating efficiency. The constraint conditions include the above-mentioned upper and lower limits of temperature deviation and equipment heating power limits, etc. Let the constraint conditions be (such as equality constraints related to power limits) and (such as inequality constraints of temperature deviation), and construct the Lagrangian function , where is a vector containing decision variables such as heating power and heating time, and are the Lagrange multipliers corresponding to the equality and inequality constraints respectively, and are the numbers of equality constraints and inequality constraints respectively. By taking the partial derivatives of the Lagrangian function with respect to , , and setting these partial derivatives to 0, solve the system of equations. During the solution process, use numerical calculation methods such as Newton's iteration method to gradually approach the optimal solution that satisfies the constraint conditions, thereby determining the division scheme of the real-time adjustment interval, realizing the refined adjustment of the heating process, and ensuring that the temperature in the reactor always remains within the ideal range.

[0079] Example 3:

[0080] During the glue mixing and making process, to achieve efficient and precise control of the heating process, the gradient descent algorithm is used to divide the real-time adjustment interval.

[0081] First, define the objective function. Within the adjustment interval, temperature deviation and energy consumption are two key considerations. Assume there are temperature sampling points within the adjustment interval, and the temperature deviation at the th sampling point is , , where is the actual temperature at the th sampling point, and is the set target temperature. At the same time, record the energy consumption corresponding to the th sampling point as . To balance temperature control accuracy and energy consumption, set the weights and , and construct the objective function . For example, if the current glue-making process has extremely high requirements for the quality of the glue and strict temperature control is needed, can be appropriately increased.Value; if more attention is paid to production costs and it is desired to reduce energy consumption, then the weight of can be increased.

[0082] Then set the constraint conditions. The upper limit of the heating power is determined by the hardware specifications of the heating equipment. If it exceeds this upper limit, it may not only damage the heating equipment but also pose a safety hazard. The material phase change threshold is the critical temperature at which the material undergoes a phase change (such as solidification, liquefaction, etc.). During the glue-making process, it is necessary to avoid the material from undergoing unexpected physical changes due to the temperature reaching the phase change threshold, which may affect the quality of the glue. The equipment safe temperature range is used to ensure the normal operation of the equipment. is the lowest temperature at which the equipment can operate normally. is the highest temperature. Exceeding this range may cause equipment failure.

[0083] Finally, solve the objective function through the conjugate gradient method. The conjugate gradient method is an iterative optimization algorithm. First, a given initial adjustment interval division scheme is given, and the gradient of the objective function at this point is calculated , and the initial search direction . In each iteration , calculate the step size to minimize , and update the adjustment interval division scheme . Calculate the new gradient , and calculate the parameter for updating the search direction through , and then update the search direction . Continue to iterate until the convergence condition is met, such as the norm of the gradient is less than a pre-set threshold . At this time, the obtained is the optimal adjustment interval division result, realizing the optimal control of the heating process and reducing energy consumption while meeting the temperature control requirements.

[0084] In the loop optimization module, the iterative update of the hierarchical optimization strategy is crucial for improving the system control performance.

[0085] Input the temperature deviation index set into the input layer of the hierarchical optimization strategy. Assume that the temperature deviation index set consists of temperature deviation data at different positions and different times in the reactor, denoted as , , , and ( (indicating the time step), the hidden layer state is , and the output is . At each time step, the hidden layer state is updated through the formula , where is the weight matrix input to the hidden layer, is the weight matrix from the hidden layer to the hidden layer, is the bias vector of the hidden layer. The output is then calculated by , is the weight matrix from the hidden layer to the output layer, is the bias vector of the output layer. The RNN is trained with a large amount of historical temperature deviation data to enable it to accurately extract the vector reflecting the characteristics of cyclic control.

[0086] The priority of the cyclic control feature vectors is sorted by the entropy weight method to generate the regional regulation sequence. The entropy weight method is an objective weighting method used to measure the importance of each feature vector in reflecting the system information. Suppose there are cyclic control feature vectors . For the -th feature vector, its information entropy is first calculated. When calculating, first calculate the proportion of this feature vector in each sample according to , where is the value of the -th sample of the -th feature vector, is the number of samples. Then calculate the information entropy through the formula . Then calculate the entropy weight . The larger the entropy weight, the more important the system information carried by this feature vector. Sort the cyclic control feature vectors according to the entropy weight size, determine the priority of each region in the regulation, and generate the regional regulation sequence.

[0087] The particle swarm algorithm combined with the simulated annealing algorithm optimization strategy parameters is used to update the regulation logic of the cyclic control unit. In the particle swarm algorithm, each particle represents a set of strategy parameters. Let the position of particle be , is the parameter dimension, and the velocity is . In each iteration, the particle updates its velocity and position according to its own historical optimal position and the global optimal position of the group . The velocity update formula is , where is the inertia weight, used to balance the global search and local search capabilities of the particle; , is the learning factor, usually set as a positive number, which controls the degree of the particle learning from its own historical optimal position and the global optimal position; , is a random number between . The position update formula is . The simulated annealing algorithm introduces a control parameter (temperature ), and in the iterative process, it accepts worse solutions with a certain probability to avoid falling into local optimality. As the iteration progresses, the temperature gradually decreases, and the probability of accepting worse solutions also gradually decreases. Combining the two algorithms, on the basis of the particle swarm algorithm exploring the solution space, the simulated annealing algorithm helps to jump out of the local optimal solution, more effectively optimize the strategy parameters, update the control logic of the loop control unit, and improve the control performance of the system for the temperature of the reaction kettle.

[0088] Example 5:

[0089] During the process of collecting the temperature data of the glue mixing and making reaction kettle, the collected temperature data often contains noise and may have missing values. The time series filtering strategy can effectively handle these problems.

[0090] First, identify the high-frequency noise segments in the temperature data set and smooth them based on the moving window mean rule. Assume that the collected temperature data sequence is , , is the total number of data points. Set the moving window size as , for the data point , calculate the average value of the data within the window centered on . When is odd, ; when is even, (when exceeds the data range, boundary data is used for filling). Use to replace , so that high-frequency noise can be effectively suppressed and the temperature data can better reflect the actual temperature change trend.

[0091] Detect the missing temperature points and use the cubic spline interpolation method to interpolate the missing temperature points. Assume that the known temperature data points are , where represents the time point and represents the temperature value corresponding to the time point. If there is a missing temperature point ( ), the cubic spline interpolation method requires that on each sub-interval , the interpolation function is a cubic polynomial . At the same time, it satisfies that the function values are equal at the nodes, that is , ; the first derivative is continuous; the second derivative is continuous, etc. By establishing and solving the system of equations composed of these conditions, the polynomial coefficients , , , on each subinterval are determined, and then the temperature value at the missing temperature point is calculated to fill in the missing data.

[0092] The interpolated data is normalized to generate a standardized temperature sequence with a mean of zero and a variance of one. Let the interpolated data be . First, calculate the mean and variance of the data. Then, through the formula , the normalization process is carried out to obtain the standardized temperature sequence . After the normalization process, the influence of data with different magnitudes is eliminated, which is convenient for subsequent algorithms to analyze and process the temperature data more accurately, providing reliable data support for the precise control of the reactor temperature.

[0093] Example 6:

[0094] In the feedback control module, the model predictive control framework and multivariable coupling analysis are the core steps to achieve precise control.

[0095] Build a multivariable coupling analysis model based on the state - space equation. Let the state variables of the reactor include temperature , pressure , material flow rate , etc., and the control variables are heating power , cooling flow rate , etc. The state - space equation can be expressed as , where is the state vector, containing the values of each state variable at time; is the control vector, containing the values of each control variable at time; is the output vector, usually containing measurable variables (such as temperature); is the state - transition matrix, describing the relationship of the state variables changing with time; is the input matrix, reflecting the influence of the control variables on the state variables; is the output matrix, determining the relationship between the state variables and the output variables; and They are process noise and measurement noise respectively. Through this model, the sensitivities of the quantization control instructions to each heating parameter are quantified, and the influence degrees of different control variables on state variables such as temperature are analyzed.

[0096] Generate a parameter perturbation sample set through Latin hypercube sampling, and calculate the coupling contribution degree of each parameter. Latin hypercube sampling is an efficient sampling method. It divides the value range of each parameter into several non-overlapping intervals, and randomly selects a sample in each interval, so as to ensure the uniform distribution of samples in the parameter space. Suppose we want to sample parameters, each parameter is divided into intervals, and sample points are obtained through Latin hypercube sampling. For each sample point, substitute it into the state space equation for simulation calculation to obtain the corresponding output result. According to the change situation of the output result, calculate the coupling contribution degree of each parameter. For example, measure the influence size of each parameter on the system output by calculating indicators such as the variance contribution rate of the output variable change caused by the parameter change.

[0097] Screen the parameters with contribution degrees higher than the preset threshold as the core control variables. The preset threshold is determined according to the actual requirements of the glue-making process and the system characteristics. Suppose the coupling contribution degrees of each parameter are calculated as , , the set threshold is , when , determine the corresponding parameter as the core control variable. These core control variables have a greater impact on the system output and are focused on and adjusted in the subsequent control process.

[0098] When performing multivariable coupling analysis, integrate the temperature sensor data, material property data and heater status data to construct a multi-dimensional coupling data matrix. For example, the temperature sensor data contains temperature values at different positions and times; the material property data includes the viscosity, specific heat capacity, etc. of the material; the heater status data has information such as heating power and heating time. Arrange these data into a multi-dimensional matrix according to certain rules. Use the principal component analysis method to extract the features of the multi-dimensional coupling data matrix and generate a dimensionality-reduced feature vector. The principal component analysis method transforms multiple related original variables into a few uncorrelated principal components through operations such as singular value decomposition of the data matrix. These principal components are linear combinations of the original variables and can retain most of the information of the original data. Input the dimensionality-reduced feature vector into the feedback control module, and combine the analysis results of the model predictive control framework to generate optimized control instructions to achieve precise control of the reaction kettle circulating heating system and ensure the stable and efficient progress of the glue-making process.

[0099] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0100] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A circulating heating system for a glue mixing and making reaction kettle, characterized in that: The system comprises: The temperature acquisition module is used to obtain a set of temperature data within a preset time range in the reactor, and to perform denoising on the temperature data set according to a time series filtering strategy to generate a standardized temperature series; A heat distribution module, for inputting the standardized temperature sequence into a dynamic weight distribution model, and generating a heat distribution ratio of each heating area based on a heat conduction equation, wherein the heat distribution ratio includes a multi-dimensional correlation relationship among temperature gradient, material viscosity and heating power; A dynamic adjustment module is used to perform nonlinear programming optimization on the heat distribution ratio according to preset temperature control constraints, divide the real-time adjustment intervals by a gradient descent algorithm, and extract a set of temperature deviation indicators in each interval; A loop optimization module, used to construct a loop control unit including a hierarchical optimization strategy, iteratively update the hierarchical optimization strategy using the temperature deviation index set, and generate a steady-state heating control scheme; A feedback control module is used to perform multivariable coupling analysis on the steady-state heating control scheme according to a model predictive control framework and output a reactor cycle heating control instruction; The nonlinear programming optimization comprises: According to the historical temperature distribution of the real-time adjustment interval, calculate the extreme value and fluctuation range of the temperature deviation index in each interval; Based on the rolling time domain optimization algorithm, the extreme value and fluctuation range are dynamically corrected to generate optimization constraints; The optimization constraints are decoupled using the Lagrange multiplier method to generate a real-time adjustment interval division scheme.

2. The circulating heating system for the glue mixing and making reaction kettle according to claim 1, characterized in that: The heat distribution ratio of each heating area is generated based on the heat conduction equation, including: Extracting the timestamp, heating area number and temperature change rate in the standardized temperature sequence to construct a multi-dimensional heat conduction matrix; The multidimensional heat conduction matrix is ​​smoothed by using a sliding average algorithm to generate a steady-state heat conduction submatrix set; The least square method is used to perform parameter fitting on the steady-state heat conduction submatrix set to eliminate abnormal fluctuation points; The remaining sub-matrices are weighted by fuzzy inference algorithm to generate a heat distribution ratio with temperature balance.

3. The circulating heating system for the glue mixing and making reaction kettle according to claim 1, characterized in that: The gradient descent algorithm divides the real-time adjustment interval into: The objective function is defined as the weighted minimization of temperature deviation and energy consumption within the adjustment interval; Set the constraints as the upper limit of heating power, the material phase change threshold, and the equipment safe temperature range; The objective function is solved by the conjugate gradient method, and the optimal adjustment interval division result is output.

4. The circulating heating system for the glue mixing and making reaction kettle according to claim 1, characterized in that: The layered optimization strategy is iteratively updated, including: Inputting the temperature deviation index set into the input layer of the hierarchical optimization strategy, and extracting the loop control feature vector using a recursive neural network; Prioritizing the cyclic control feature vectors by an entropy weight method to generate a regional regulation sequence; The particle swarm algorithm is combined with the simulated annealing algorithm to optimize the strategy parameters and update the control logic of the cyclic control unit.

5. The circulating heating system for the glue mixing and making reaction kettle according to claim 1, characterized in that: The time series filtering strategy includes: Identifying high frequency noise segments in the temperature data set and smoothing based on a sliding window mean rule; Detecting missing temperature points and interpolating the missing temperature points using a cubic spline interpolation method; The interpolated data are normalized to generate a standardized temperature series with a mean of zero and a variance of one.

6. The circulating heating system for the glue mixing and making reaction kettle according to claim 1, characterized in that: The model predictive control framework includes: Construct a multivariable coupling analysis model based on state-space equations to quantify the sensitivity of control instructions to various heating parameters; Generate parameter perturbation sample sets through Latin hypercube sampling and calculate the coupling contribution of each parameter; Parameters with contributions higher than the preset threshold are selected as core control variables.

7. The circulating heating system for the glue mixing and making reaction kettle according to claim 6, characterized in that: The multivariate coupling analysis includes: Integrate temperature sensor data, material attribute data and heater status data to build a multi-dimensional coupled data matrix; Using principal component analysis to extract features from the multidimensional coupling data matrix to generate a reduced-dimensional feature vector; The dimension-reduced feature vector is input into the feedback control module to generate optimized control instructions.

8. The circulating heating system for the glue mixing and making reaction kettle according to any one of claims 1 to 7, characterized in that: The system further comprises: Constructing an early warning module including an equipment thermal stress assessment function, comparing and analyzing the control instructions with the thermal stress threshold, and outputting a graded alarm signal; A preset emergency protection mechanism is triggered according to the graded alarm signal to generate a power limitation or cooling start instruction.

9. An electronic device, characterized in that: include: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the operation of the circulating heating system of the glue mixing and making reactor as described in any one of claims 1 to 8.

Images