Temperature correction method and correction system for a hot press

Abstract

The application discloses a temperature correction method and correction system of a hot press, and particularly relates to the technical field of hot press temperature control, and comprises the following steps: acquiring and preprocessing multi-dimensional operation data, including temperature time sequence, pressure setting and material properties; constructing a coupling model based on a multi-layer neural network to predict temperature distribution in different process stages; performing sensor drift compensation, analyzing temperature deviation characteristics, judging whether there is sensor drift or material thermal diffusion anomaly, and generating a local distortion mark; determining a compensation mode according to the distortion mark, and starting a local compensation mode or a global balance control mode if the distortion is identified as an out-of-limit distortion; comparing the output of the coupling model to verify the temperature correction effect, and outputting a stable temperature correction result if the deviation is reduced and the distortion mark converges. The application effectively solves the problems of temperature distortion and sensor drift, and improves the long-term precision and stability of the hot press under complex working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hot press temperature control, and more particularly, to a hot press temperature correction method and system. BACKGROUND

[0002] Hot presses are widely used in the manufacturing process of multilayer composites. Its key function is to achieve uniform heating and close bonding of materials by precisely controlling the temperature, pressure and pressing time of the heating plate, so as to meet the requirements of different processes on strength, dimensional accuracy and performance. In actual production, multiple temperature sensors are arranged in the heating area of the hot press to monitor the temperature distribution on the surface of the heating plate in real time, and dynamic adjustment of heating power or cooling is made through feedback mechanism, so as to realize temperature control. However, with the increase of process complexity and the diversification of material types, higher requirements are put forward for the accuracy and response speed of temperature control. The traditional single sensor monitoring and static adjustment strategy is difficult to adapt to the dynamic needs in complex scenarios.

[0003] The existing hot press temperature control technology has multiple problems in long-term operation or complex working conditions. On the one hand, sensor drift may cause measurement error accumulation, which in turn affects the accuracy of temperature feedback; on the other hand, in the process of hot pressing of multilayer composites, there is a high coupling between temperature, pressure and material properties. If only constant parameter control is used, it is often difficult to deal with local temperature distortion and sensor drift problems. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the present application provides a hot press temperature correction method and system to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a hot press temperature correction method, comprising the following steps:

[0006] Step one, obtaining multi-dimensional running data to be corrected: obtaining multi-dimensional running data generated by the hot press in the process of pressing multilayer composites, the multi-dimensional running data including hot press temperature time series data, pressure setting parameters and material attribute information; denoising and interpolating the multi-dimensional running data to output a preprocessed data set;

[0007] Step two, building a coupling model: identifying the coupling relationship between temperature, pressure and material in the preprocessed data set, building a coupling model based on a multilayer neural network, the coupling model outputting temperature prediction distribution in different process stages;

[0008] Step three, sensor drift compensation and local distortion marking: according to the difference between the temperature prediction distribution of the coupling model and the actual temperature distribution, the deviation feature is extracted, and it is judged whether there is sensor drift or material thermal diffusion abnormality by analyzing the deviation feature, the corrected temperature value after compensation is fed back to the coupling model to correct the sensor drift, and the local distortion marking is carried out based on the sensor drift compensation result, the local distortion marking includes position coordinates, duration and deviation amplitude;

[0009] Step four, determining the temperature correction mode according to the local distortion marking: analyzing the local distortion marking, determining whether to start the compensation mode or the global balance control mode; if it is identified as an out-of-limit distortion, the compensation mode is started, and power adjustment or auxiliary cooling is implemented in the corresponding heating section; otherwise, the global balance control mode is entered, and the pressure and temperature parameters of the hot press are cooperatively fine-tuned;

[0010] Step five, temperature correction verification: after the local compensation mode or the global balance control mode, the new temperature measurement value is compared with the coupling model output, if the deviation is reduced and the distortion marking is in a convergent state, it is considered that the compensation in step four is effective, and the temperature correction result is output; otherwise, it indicates that the compensation is invalid.

[0011] Preferably, the coupling model is constructed in the following manner:

[0012] The pressure, material feature vector and start-stop time sequence are selected as inputs to construct the coupling model of the hot press;

[0013] Based on the mutual information theory, the key variables with high coupling degree to temperature change are screened;

[0014] The key variables are nonlinearly fitted by using a multilayer neural network to output the predicted temperature;

[0015] The accuracy of temperature prediction is improved and overfitting is suppressed by minimizing the loss function;

[0016] The loss function L(θ) of the coupling model satisfies the following formula: Wherein, θ is the coupling model parameter, specifically the set of internal weights or coefficients to be trained of the coupling model, including the weight matrix of the neural network and the kernel function coefficient of the support vector machine, which is updated by iterative training algorithm (such as gradient descent or Adam); M is the number of data points, indicating the number of samples or measurement times contained in the training set, i is the sequential number; x i Refers to the input vector of the i-th sample or measurement time; Indicates the predicted temperature, which is the output result under the given input vector x i And the coupling model parameter θ. Indicates the reference temperature, which is the output result under the given input vector x ithe output results from the temperature sensor network, thermal imager measurements or experimental calibration values; a balance factor, balancing the model fitting error and the model complexity, controlling the importance of the regularization term ; a regularization term to suppress overfitting.

[0017] Preferably, the coupling model compares the predicted temperature with the measured temperature point by point in the training stage, and if the prediction error or mutual information deviation of a local area continuously exceeds the preset threshold, it is marked as deviation, and the deviation feature is extracted; the preset threshold represents the standard for determining whether the temperature deviation exceeds the allowed range;

[0018] Analyzing the deviation feature, if the temperature prediction distribution of the coupling model and the measured temperature distribution at a single type interface show deviation for more than a threshold time, it can be determined that there is an abnormal problem of material interface heat diffusion;

[0019] If there is an abnormal material interface heat diffusion, a user is warned, and process abnormalities are prompted, and process parameters are prompted to be adjusted;

[0020] If the local temperature deviation analysis and continuous monitoring find that the temperature reading of a specific sensor and the difference between the adjacent normal sensor and the predicted temperature value of the coupling model continuously exceed the normal variation range, it is recorded as an abnormal drift sensor;

[0021] Calculate the weighted difference between the adjacent normal sensor reading and the coupling model prediction value to obtain the compensation coefficient k j of the jth sensor, the compensation coefficient ranges from 0 to 1; then apply k j to the abnormal drift sensor data: wherein, represents the corrected temperature value of the jth sensor, which is used to replace the original measured temperature; represents the original measured temperature of the jth sensor; represents the predicted temperature of the coupling model; represents the complementary factor corresponding to k j , which is used to balance the weighted proportion of the measured temperature and the predicted temperature.

[0022] Preferably, the process of analyzing the local distortion mark to determine whether to start the compensation or global balance control mode includes:

[0023] Setting a dynamic distortion threshold: based on the deviation amplitude provided by the local distortion mark, defining a dynamic distortion threshold that decays over time to distinguish between short-term fluctuations and persistent distortion, the dynamic distortion threshold satisfies the following formula: wherein, is the initial threshold, denotes the rate of attenuation, is the cumulative time since the last compensation is completed; if the deviation of the heating area exceeds at time t, it is determined that there is local distortion;

[0024] Identify the distortion beyond the limit: if the local distortion persists within the time window, whether the spatial range exceeds the preset proportion, mark it as distortion beyond the limit;

[0025] Determine the temperature correction mode: if the distortion beyond the limit is detected, start the compensation mode, implement rapid power adjustment or auxiliary cooling in the corresponding heating section; otherwise, start the global balance control mode, and make linkage fine-tuning of the pressure and temperature of the entire hot press.

[0026] Preferably, the local compensation mode comprises:

[0027] Obtain the local compensation control signal: for the area marked by the distortion beyond the limit, define the proportional gain coefficient and the integral gain coefficient, calculate the local compensation control signal through the local compensation control signal model, and use s to represent the sequential number of the distortion beyond the limit area, the local compensation control signal model satisfies the following formula. wherein, denotes the local compensation control signal of the heating or cooling adjustment amount of the s-th distortion beyond the limit area at time k;

[0028] the temperature deviation of the s-th distortion beyond the limit area at time k, if is positive, it indicates that the actual temperature is higher than the target, and cooling needs to be performed; if is negative, it indicates that the actual temperature is lower than the target, and heating needs to be performed;

[0029] is the proportional gain coefficient of the s-th distortion beyond the limit area, is the integral gain coefficient of the s-th distortion beyond the limit area, integrates the temperature deviation in the time interval , and τ is the integral variable.

[0030] Preferably, the starting global balance control mode comprises:

[0031] Coupling input correction: taking the initial input of the coupling model as the benchmark, increase the temperature correction amount , the temperature correction amount satisfies the following formula: wherein, is the balance coefficient, controlling the influence degree of the corresponding global correction amount on the heating or cooling of the entire system; is the temperature deviation at the j-th sensor; N is the number of sensors, and η is the regularization constant, avoiding the numerical instability phenomenon caused by the denominator tending to zero.

[0032] The temperature correction amount is used to adjust the power and the pressure in linkage:

[0033] The temperature correction amount is decomposed into the adjustment signals of the heating power W and the pressure P, the adjustment amount of the heating power ΔW, and the adjustment amount of the pressure ΔP. Wherein, k W , k P , are the correction coefficients of the heating power and the pressure respectively, which are adjusted according to experiments or process requirements; if the temperature correction amount , the heating power is increased, and the pressure is also increased, so as to accelerate the temperature rise and improve the thermal contact efficiency; if the temperature correction amount , the heating power is reduced, and the pressure is also reduced, so as to delay the temperature rise and reduce the heat accumulation.

[0034] After the above adjustment is completed, the real-time temperature distribution is acquired again through the sensor, the new temperature is input into the coupling model for verification, and if the deviation is not significantly reduced, the correction coefficients of the heating power and the pressure are further increased in the next iteration period.

[0035] Preferably, when the hot press is in a frequent start-stop state, the iteration period of the local compensation mode and the global balance control is automatically shortened, so that the global adjustment and the local compensation are alternately performed at a higher frequency to adapt to sudden temperature fluctuations or material replacement.

[0036] Preferably, the method further comprises the step of adaptive optimization iteration correction, if the local deviation feature still continuously exceeds the dynamic distortion threshold after compensation, the regularization term and the compensation gain parameter of the coupling model are adjusted, the compensation gain parameter includes a proportional gain coefficient and an integral gain coefficient; adaptive optimization is performed to update the coupling model; the effectiveness of the compensation strategy is re-determined according to the optimized coupling model; if the distortion is not exceeded after multiple iterations, the stable temperature correction result is output.

[0037] Preferably, the coupling model is trained based on the adjusted loss function, and the optimized coupling model is output, and the proportional gain coefficient and the integral gain coefficient are re-set through test data.

[0038] The optimized coupling model and the compensation gain parameter will affect the temperature prediction and compensation operation in the next step.

[0039] In order to achieve the purpose of the present application, a temperature correction system of a hot press is provided, comprising:

[0040] A multi-dimensional data acquisition and preprocessing module is used to acquire multi-dimensional running data of the hot press, to perform denoising processing on the acquired data, to use an interpolation algorithm to complete missing values, and to format the processed data into a time sequence and output a preprocessed data set.

[0041] Coupling model construction module: based on the preprocessed data set, the coupling relationship between temperature, pressure and material properties is identified; a coupling model capable of predicting temperature distribution at different process stages is generated by training a multi-layer neural network; the preprocessed data set is input into the coupling model, and the corresponding temperature prediction distribution is output;

[0042] Deviation feature extraction and labeling module: according to the temperature prediction distribution and the actually measured temperature distribution, the difference is calculated and the local deviation feature is extracted, including deviation amplitude, duration and spatial distribution; it is judged whether the distortion source is related to sensor drift or material thermal diffusion anomaly, if it is sensor drift, the sensor output is corrected by calculating the compensation coefficient; local distortion labels are generated, including position coordinates and key information of deviation attributes;

[0043] Compensation strategy determination and execution module: analyze the local distortion label, if it is identified as an out-of-limit distortion, start the compensation mode, adjust the heating power or auxiliary cooling of the distortion area; otherwise, enter the global balance control mode, and make linkage fine tuning to the pressure and temperature parameters of the hot press;

[0044] Correction result verification and iteration module: record the temperature measurement value after compensation, and compare it with the predicted value of the coupling model; if the deviation is reduced and the local distortion label converges, output the stable temperature correction result; otherwise, return to the deviation feature extraction and labeling module to generate the local distortion label again, and iteratively optimize the coupling model and the compensation strategy.

[0045] Technical effects and advantages of the present application:

[0046] (1) The temperature correction method of the hot press provided by the present application is based on the coupling relationship of temperature, pressure and material properties, a corresponding coupling model is constructed by using a multi-layer neural network, real-time temperature prediction distribution can be realized, local deviation features are extracted combined with measured temperature data, sensor drift and material interface thermal diffusion anomalies are accurately determined, data reliability and deviation tracing ability are significantly improved; based on the compensation strategy combining dynamic distortion threshold and compensation mode, local rapid temperature rise and fall and global collaborative balance control can be realized, temperature uniformity and production efficiency are effectively improved, and the problem of inaccurate temperature control caused by sensor drift and material thermal diffusion anomaly of traditional hot press is effectively solved.

[0047] (2) The temperature correction method of the hot press provided by the present application, under the condition of frequent start-stop, by shortening the control iteration period and real-time feedback optimization, the adaptability to sudden temperature fluctuations and material changes is enhanced, through sensor drift correction and inertia correction strategy, the long-term stability of the correction result is ensured, and reliable guarantee is provided for high-precision temperature control of hot press in complex scenes. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1A temperature correction method flowchart of the hot press of the present application.

[0049] Figure 2 An adaptive optimization iterative correction flowchart. DETAILED DESCRIPTION

[0050] Example embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. While example embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood, and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0051] It should be understood, however, that the sizes of the various portions shown in the drawings are chosen for convenience only, and are not necessarily to scale.

[0052] The following description of at least one example embodiment is merely illustrative in nature and is in no way intended to limit the application or its application or uses.

[0053] Techniques, methods, and apparatus known to those of ordinary skill in the relevant art can not be discussed in detail herein, but should be considered part of the specification.

[0054] Embodiment 1, refer to Figure 1 A temperature correction method flowchart of the hot press of the present application, the present application provides a temperature correction method for a hot press, comprising the following steps:

[0055] Step one, obtain multi-dimensional running data to be corrected: obtain multi-dimensional running data generated by the hot press during the pressing of multi-layer composite materials, the multi-dimensional running data including hot press temperature time series data, pressure setting parameters and material attribute information; denoise and interpolate the multi-dimensional running data to form a preprocessed data set for temperature coupling analysis;

[0056] Step two, build a coupling model: identify the coupling relationship between temperature, pressure, and material in the preprocessed data set, and build a coupling model based on a multi-layer neural network, the coupling model outputting temperature prediction distribution at different process stages;

[0057] Step three, perform sensor drift compensation and obtain local distortion markers: according to the difference between the temperature prediction distribution of the coupling model and the actual temperature distribution, extract the deviation feature, determine whether there is sensor drift or material thermal diffusion anomaly by analyzing the deviation feature, feed back the corrected temperature value after compensation to the coupling model to correct the sensor drift, and perform local distortion marking based on the sensor drift compensation result, the local distortion marking including position coordinates, duration, and deviation amplitude;

[0058] It is explained that the bias feature refers to the multi-dimensional characteristics of the difference between the temperature prediction distribution and the actual temperature distribution, including not only the numerical value of the bias (i.e. the bias amplitude), but also the persistence of the bias (such as whether the bias is long-term or continuously changing), the spatial distribution (such as whether the bias is concentrated in a certain heating area or presents a specific spatial pattern), and the timing change;

[0059] Step four, determining the temperature correction mode according to the local distortion mark: analyzing the local distortion mark to determine whether to start the compensation mode or the global balance control mode; if it is identified as an out-of-limit distortion, the compensation mode is started, and power adjustment or auxiliary cooling is implemented in the corresponding heating section; otherwise, the global balance control mode is entered, and the pressure and temperature parameters of the hot press are cooperatively fine-tuned;

[0060] Step five, temperature correction verification: after the local compensation mode or the global balance control mode, the new temperature measurement value Tnew is obtained, which is compared with the coupling model output again. If the bias is reduced and the distortion mark is in a convergent state, it is considered that the compensation in step four is effective, and the temperature correction result is output; otherwise, it indicates that the compensation is invalid.

[0061] In the embodiments of the present application, it needs to be further explained that the coupling model is constructed in the following way:

[0062] The pressure, material characteristic vector (including the thermal conductivity, specific heat capacity and thickness of each layer, which are obtained based on material attribute information) and start-stop time sequence are selected as inputs to construct the coupling model of the hot press;

[0063] Based on the mutual information theory, the key variables with high coupling degree to temperature change are screened;

[0064] It is explained that the key variables with high coupling degree to temperature change screened based on the mutual information theory means that when analyzing the relationship between temperature change and multiple operating parameters (such as pressure, material characteristics and start-stop cycle), the mutual information is used to quantify the dependence degree or correlation strength of each operating parameter and temperature; when the mutual information value between a variable and temperature is larger, it means that the corresponding variable has a higher influence degree or coupling degree on temperature change; through this method, the key variables that have the most decisive or significant influence on temperature are first screened from numerous potential influencing factors, and the key variables will be focused on and processed in the subsequent coupling modeling or compensation analysis, thereby improving the efficiency and accuracy of the model;

[0065] A multi-layer neural network is used to perform nonlinear fitting on the key variables, and the predicted temperature is output;

[0066] The accuracy of temperature prediction is improved and overfitting is suppressed by minimizing the loss function;

[0067] The loss function L(θ) of the coupling model satisfies the following formula: Wherein θ is the coupling model parameter, specifically the set of weights or coefficients inside the coupling model to be trained, including the weight matrix of the neural network, the kernel function coefficient of the support vector machine, and θ is updated through the iterative training algorithm (such as gradient descent or Adam); M is the number of data points, indicating the number of samples or measurement times contained in the training set, and i is the sequential number;

[0068] x i Refers to the input vector of the i-th sample or measurement time, such as pressure, time, material properties (such as thermal conductivity, specific heat capacity), and other temperature-related features (such as ambient temperature, start-stop state);

[0069] Indicates the predicted temperature, which is the output result under the given input vector x i And the coupling model parameter θ;

[0070] Indicates the reference temperature, which is the output result under the given input vector x i , from the temperature sensor network, thermal imager measurement or experimental calibration value;

[0071] Is a balance factor to balance the fitting error and model complexity of the model, and to control the importance of the regularization term ;

[0072] Is a regularization term to suppress overfitting, which is the sum of the squares of the weights in the embodiment of the present application.

[0073] Further explained in the embodiment of the present application is that the coupling model compares the predicted temperature with the measured temperature point by point in the training stage, and if the prediction error or mutual information deviation of the local area continuously exceeds the preset threshold, it is marked as deviation, and the deviation feature is extracted; The preset threshold value represents the standard for determining whether the temperature deviation exceeds the allowed range;

[0074] Analyze the deviation feature, if the temperature prediction distribution of the coupling model and the measured temperature distribution at a single type interface show deviation for more than a threshold value, it can be determined that there is an abnormal problem of material interface heat diffusion; If there is an abnormal material interface heat diffusion, a user is warned, prompting process abnormality, prompting to adjust process parameters;

[0075] If the local temperature deviation analysis and continuous monitoring find that the temperature reading of a specific sensor and the difference between the adjacent normal sensor and the predicted temperature value of the coupling model continuously exceed the normal variation range, it is recorded as an abnormal drift sensor;

[0076] Calculate the weighted difference between the adjacent normal sensor reading and the coupling model predicted value to obtain the compensation coefficient k of the jth sensor j The compensation coefficient ranges between 0 and 1; the larger the value, the more trust in the measured data, and the smaller the value, the more dependent on the model prediction; then apply k j to the abnormal drift sensor data: Wherein, represents the corrected temperature value of the jth sensor, which is used to replace the original measured temperature; represents the original measured temperature of the jth sensor; represents the coupling model predicted temperature; represents the corresponding complementary factor of k j , which is used to balance the weighted proportion of the measured temperature and the predicted temperature.

[0077] It is explained that the compensation coefficient k j of the jth sensor is obtained in the following way:

[0078] Calculate the difference between the actual temperature of the jth sensor and the coupling model predicted temperature, denoted as ;

[0079] Calculate the difference between the temperature of the jth sensor and the reading of its adjacent normal sensor, denoted as ;

[0080] The compensation coefficient is calculated by the following formula, Wherein, and are weight factors that control the relative influence of the difference between the actual temperature and the coupling model predicted temperature.

[0081] It needs to be further explained in the embodiments of the present application that in step four, the process of analyzing the local distortion marker to determine whether to start the compensation or global balance control mode includes:

[0082] Set a dynamic distortion threshold: based on the deviation amplitude provided by the local distortion marker, define a dynamic distortion threshold that decays over time to distinguish between short-term fluctuations and persistent distortion, the dynamic distortion threshold satisfies the following formula: Wherein, is the initial threshold, represents the decay rate, which is initially set as an empirical coefficient, and Δtl is the cumulative time since the last compensation was completed; if the deviation of the heating area exceeds at time t, it is determined to be local distortion;

[0083] Identify the out-of-limit distortion: if the local distortion is continuous in the time window, whether the spatial range exceeds the preset proportion (spatial distribution is concentrated and the area is larger than the preset value), mark it as out-of-limit distortion, that is, only when the corresponding area reaches the preset conditions in amplitude, time and spatial range, it is determined to be in the "out-of-limit distortion" state;

[0084] Determine the temperature correction mode: if the out-of-limit distortion is detected, start the compensation mode, implement rapid power adjustment or auxiliary cooling in the corresponding heating section; otherwise, start the global balance control mode, and make linkage fine adjustment to the pressure and temperature of the whole hot press.

[0085] The compensation mode selection information will be synchronously transmitted to the subsequent calculation unit to calculate the correction amount and update the coupling model.

[0086] Further explained in the embodiments of the present application is that the starting compensation mode comprises:

[0087] Obtaining a local compensation control signal: for the region marked by the out-of-limit distortion, define a proportional gain coefficient and an integral gain coefficient, calculate the local compensation control signal through a local compensation control signal model, and use s to represent the sequential number of the out-of-limit distortion region, the local compensation control signal model satisfies the following formula, Wherein, represents the local compensation control signal of the heating or cooling adjustment amount of the s-th out-of-limit distortion region at time k;

[0088] The temperature deviation of the s-th out-of-limit distortion region at time k, if is positive, it indicates that the actual temperature is higher than the target, and cooling needs to be performed; if is negative, it indicates that the actual temperature is lower than the target, and heating needs to be performed;

[0089] is the proportional gain coefficient of the s-th out-of-limit distortion region, used to amplify the temperature deviation at the current time, and determines the influence of the instant deviation on the compensation control; is the integral gain coefficient of the s-th out-of-limit distortion region, which determines the influence of the accumulated deviation on the compensation control; Integrate the temperature deviation on the time interval , which is used to eliminate the persistent or slow accumulated temperature error; τ is the integral variable.

[0090] Further explained in the embodiments of the present application is that the starting global balance control mode comprises:

[0091] Coupling input correction: taking the initial input of the coupling model as the benchmark, increase the temperature correction amount , which satisfies the following formula: Wherein, is a balance coefficient, controlling the influence degree of the corresponding global correction on the overall heating or cooling of the system; is the temperature deviation at the jth sensor; N is the number of sensors, and η is a regularization constant to avoid numerical instability caused by the denominator tending to zero;

[0092] The temperature correction is used to adjust the power and pressure in linkage:

[0093] The temperature correction is decomposed into adjustment signals for heating power W and pressure P, ΔW is the adjustment amount of heating power, and ΔP is the adjustment amount of pressure. wherein k W , k P are the correction coefficients of heating power and pressure respectively, which are adjusted according to experiments or process requirements; if the temperature correction , the heating power is increased, and the pressure is also increased, so as to accelerate the temperature rise and improve the thermal contact efficiency; if the temperature correction , the heating power is reduced, and the pressure is also reduced, so as to delay the temperature rise and reduce the heat accumulation.

[0094] Real-time feedback and iterative optimization: after the above adjustment is completed, the real-time temperature distribution is obtained again through the sensor, the new temperature is input into the coupling model for verification, and if the deviation is not significantly reduced, the correction coefficients of heating power and pressure are further increased in the next iteration period.

[0095] It needs to be further explained in the embodiments of the present application that the starting of the global balance control mode includes:

[0096] When the material has poor thermal conductivity (such as a polymer composite material): the heating power is preferentially adjusted, and the potential impact of pressure change on the material structure is reduced;

[0097] When the material has high thermal sensitivity (such as a thin layer of metal): the pressure is preferentially adjusted to affect the thermal contact efficiency, so as to avoid local overheating caused by large power change.

[0098] It needs to be further explained in the embodiments of the present application that when the hot press is in a frequent start-stop state, the iteration period of the local compensation mode and the global balance control is automatically shortened, so that the global adjustment and the local compensation are alternately performed at a higher frequency to adapt to sudden temperature fluctuations or material replacement.

[0099] Summary: The hot press temperature correction method provided by the embodiment of the application adopts a multilayer neural network and a coupling model based on mutual information theory, can accurately analyze and correct temperature deviations in the operation of a hot press, can effectively identify and process problems such as local distortion, sensor drift, and abnormal thermal diffusion at a material interface by monitoring temperature, pressure, and material attribute data in real time and combining a dynamic compensation strategy, and not only improves the real-time performance of temperature correction but also enhances the adaptability of the system to different materials and production processes.

[0100] Embodiment 2, the difference between the embodiment of the application and embodiment 1 is that the method further includes a step of adaptive optimization iterative correction, specifically including:

[0101] Referring to Figure 2 the adaptive optimization iterative correction flowchart, the method further includes a step of adaptive optimization iterative correction, if the local deviation feature still continuously exceeds the dynamic distortion threshold after compensation, the regularization term and the compensation gain parameter of the coupling model are adjusted, the compensation gain parameter includes a proportional gain coefficient and an integral gain coefficient; adaptive optimization is performed to update the coupling model; the effectiveness of the compensation strategy is re-determined according to the optimized coupling model; if no over-limit distortion occurs after multiple iterations, a stable temperature correction result is output to further guarantee the long-term accuracy and stability of the hot press in a high complexity scenario, specifically including the following steps:

[0102] Step 101, compensation effect evaluation and local distortion mark update: after each compensation is completed, the temperature measurement value after compensation is recorded and compared with the predicted value of the coupling model; if the deviation amplitude is reduced and the local distortion mark converges, a stable temperature correction result is output; if the deviation is not significantly reduced, the next step is entered;

[0103] Step 102, deviation feature analysis and compensation strategy adjustment: analyze the deviation feature of the local distortion area and determine whether the compensation effect meets:

[0104] In the continuous k sampling periods, the local area deviation value is less than the set dynamic distortion threshold; the area ratio of the distortion area is lower than the set threshold;

[0105] If the conditions are not met and the deviation feature still exists after multiple compensations, the corresponding area is judged to be a long-term over-limit distortion area, and the coupling model needs to be further optimized;

[0106] Step 103, optimization of the coupling model and adjustment of the compensation gain: the coupling model is retrained by adjusting the regularization factor, and the compensation gain parameter is adjusted;

[0107] Step 104, dynamically correcting the dynamic distortion threshold and the final temperature correction: adjusting the decay rate of the dynamic distortion threshold to ensure adaptability to temperature changes; if no long-term over-limit distortion region is detected after continuous multiple rounds of optimization, and the difference between the temperature prediction distribution and the actual temperature is significantly reduced, then output the final stable temperature correction result.

[0108] It needs to be further explained in the embodiments of the present application that the coupled model is trained based on the adjusted loss function, and an optimized coupled model is output, and the proportional gain coefficient and the integral gain coefficient are reset through test data;

[0109] The optimized coupled model and the compensation gain parameter will affect the next step of temperature prediction and compensation operation.

[0110] It is explained that the adjustment mode of the loss function includes adding the sum of squares of temperature deviations of all local distortion regions in the loss function, and through the punishment of local distortion temperature deviation, the coupled model not only pays attention to the overall temperature error, but also tries to reduce the temperature deviation in the local distortion region, so as to ensure more accurate temperature control in the local region.

[0111] Further, after optimizing the coupled model and adjusting the compensation gain, if the local distortion still exists and shows temperature rising lag or temperature change lag;

[0112] The inertia correction factor μ is taken as part of the compensation gain, which plays a role in correcting the temperature change lag effect, and the optimized local compensation control signal is output through the following formula; When the dynamic distortion threshold is corrected, the inertia correction factor μ will help to further adjust the compensation strategy, so that the temperature change is more stable, and over-compensation or delayed compensation caused by thermal inertia is avoided.

[0113] It is explained that the inertia correction factor μ takes a value between 0 and 1, the smaller the value, the lighter the correction of the inertia effect of temperature change, and the larger the value, the more the system relies on inertia correction to cope with slower temperature rise or lag response.

[0114] Summary: The embodiments of the present application can realize accurate correction of temperature control of the hot press under complex working conditions by continuously optimizing the coupled model, the compensation gain parameter and the dynamic distortion threshold; through multiple rounds of compensation and optimization, the system can analyze the local distortion characteristics in real time, dynamically adjust the compensation strategy, and continuously improve the correction accuracy in each iteration, the adjusted coupled model can effectively cope with material property changes, sensor drift and other temperature distortion problems, and ensure the long-term stability and accuracy of the temperature control system.

[0115] Embodiment 3, the present application provides a temperature correction system of a hot press, comprising:

[0116] A multi-dimensional data acquisition and preprocessing module is configured to acquire multi-dimensional operation data of the hot press, including temperature time series data, pressure setting parameters and material attribute information; the acquired data is denoised, missing values are completed using an interpolation algorithm, and the processed data is formatted into a time series, and a preprocessed data set is output;

[0117] A coupling model construction module is configured to identify the coupling relationship between temperature, pressure and material attributes based on the preprocessed data set; a coupling model is trained through a multi-layer neural network to generate a coupling model that can predict the temperature distribution at different process stages; the preprocessed data set is input into the coupling model, and a corresponding temperature prediction distribution is output;

[0118] A deviation feature extraction and labeling module is configured to calculate the difference and extract local deviation features, including deviation amplitude, duration and spatial distribution, according to the temperature prediction distribution and the actually measured temperature distribution; it is determined whether the distortion source is related to sensor drift or material thermal diffusion anomaly, and if it is sensor drift, the sensor output is corrected by calculating a compensation coefficient; local distortion labels are generated, including position coordinates and key information of deviation attributes;

[0119] A compensation strategy determination and execution module is configured to analyze the local distortion labels, and if it is identified as an out-of-limit distortion, a compensation mode is started to adjust the heating power or auxiliary cooling of the distorted area; otherwise, a global balance control mode is entered to make linkage fine tuning of the pressure and temperature parameters of the hot press;

[0120] A correction result verification and iteration module is configured to compare the temperature measurement value after compensation with the predicted value of the coupling model; if the deviation is reduced and the local distortion label converges, a stable temperature correction result is output; otherwise, the deviation feature extraction and labeling module is returned to regenerate the local distortion label, and the coupling model and the compensation strategy are iteratively optimized.

[0121] Finally, the above-described only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included within the scope of protection of the present application.

Claims

1. A temperature calibration method for a hot press, characterized in that, Includes the following steps: Step 1: Obtain the multidimensional operational data to be corrected: Obtain the multidimensional operational data generated by the hot press during the pressing of multilayer composite materials, perform noise reduction and interpolation on the multidimensional operational data, and output the preprocessed dataset; Step 2: Construct a coupling model: Identify the coupling relationship between temperature, pressure, and materials in the preprocessed dataset, and construct a coupling model based on a multi-layer neural network. The coupling model outputs the predicted temperature distribution under different process stages. During the training phase, the coupled model compares the predicted temperature with the measured temperature point by point. If the prediction error or mutual information deviation in a local area continues to exceed a preset threshold, it is marked as a deviation, and deviation features are extracted. The preset threshold represents the standard for determining whether the temperature deviation exceeds the allowable range. Analyzing the deviation features, if the cumulative time for the temperature prediction distribution of the coupled model and the measured temperature distribution to deviate at a single type of interface exceeds the threshold, it can be determined that there is an abnormal thermal diffusion problem at the material interface. If there is an abnormal thermal diffusion problem at the material interface, an alert is issued to the user, indicating a process abnormality and prompting adjustment of process parameters. If, through local temperature deviation analysis and continuous monitoring, it is found that the difference between the temperature reading of a specific sensor and the predicted temperature value of adjacent normal sensors and the coupled model continuously exceeds the normal variation range, it is recorded as an abnormal drift sensor. The weighted difference between the readings of adjacent normal sensors and the predicted value of the coupled model is calculated to obtain the compensation coefficient k of the j-th sensor. j The compensation coefficient ranges from 0 to 1; then k j Application to abnormal drift sensor data: ;in, This represents the corrected temperature value from the j-th sensor, used to replace the original measured temperature. This represents the original measured temperature of the j-th sensor; This indicates that the coupled model predicts the temperature; Indicates the relationship with k j The corresponding complementary factor is used to balance the weighting ratio of the measured temperature and the predicted temperature. Step 3: Perform sensor drift compensation and obtain local distortion markers: Based on the difference between the temperature prediction distribution and the actual temperature distribution of the coupled model, extract the deviation features, and analyze the deviation features to determine whether there is sensor drift or abnormal material thermal diffusion. Feed back the compensated temperature correction value to the coupled model to correct the sensor drift. Based on the sensor drift compensation results, perform local distortion markers, which include position coordinates, duration and deviation amplitude. Step 4: Determine the temperature correction mode based on the local distortion markers: Analyze the local distortion markers to determine whether it is necessary to activate the compensation mode or the global balance control mode; if it is identified as excessive distortion, activate the compensation mode and implement power adjustment or auxiliary cooling in the corresponding heating section; otherwise, enter the global balance control mode to perform coordinated fine-tuning of the pressure and temperature parameters of the hot press. The process of analyzing local distortion markers to determine whether compensation or global balance control mode needs to be activated includes: Define a dynamic distortion threshold: Based on the deviation magnitude provided by the local distortion markers, define a dynamic distortion threshold that decays over time. To distinguish between short-term fluctuations and persistent distortion, the dynamic distortion threshold... Satisfy the following formula: in, As the initial threshold, Indicates the decay rate. This is the cumulative time since the last compensation was completed; if the deviation of the heating area exceeds [time t]... If so, it is judged as local distortion; If local distortion persists within a time window and its spatial range exceeds a preset percentage, it is marked as over-limit distortion. That is, only when the corresponding area meets the preset conditions in terms of amplitude, time, and spatial range is it considered to be in an over-limit distortion state. Determine the temperature correction mode: If excessive distortion is detected, activate the compensation mode to implement rapid power adjustment or auxiliary cooling in the corresponding heating section; otherwise, activate the global balance control mode to make linkage fine adjustments to the overall pressure and temperature of the hot press. Step 5: Verification after temperature correction: After obtaining the new temperature measurement value in local compensation mode or global balance control mode, compare it with the output of the coupled model. If the deviation decreases and the distortion marker is in a convergent state, the compensation in step 4 is considered effective, and the temperature correction result is output; otherwise, the compensation is considered ineffective. The method further includes an adaptive optimization iterative correction step. If the local deviation characteristics continue to exceed the dynamic distortion threshold after compensation, the regularization term and compensation gain parameter of the coupled model are adjusted. The compensation gain parameter includes the proportional gain coefficient and the integral gain coefficient. Adaptive optimization is performed to update the coupled model. The effectiveness of the compensation strategy is re-determined based on the optimized coupled model. If no over-limit distortion occurs after multiple iterations, a stable temperature correction result is output.

2. The temperature correction method for a hot press according to claim 1, characterized in that, The coupling model is constructed as follows: A coupled model of the hot press is constructed by selecting pressure, material feature vectors, and start-up / shutdown time series as inputs. Based on mutual information theory, key variables with high coupling to temperature changes are selected; a multi-layer neural network is used to perform nonlinear fitting on the key variables and output the predicted temperature. The accuracy of temperature prediction is improved and overfitting is suppressed by minimizing the loss function; the loss function L(θ) of the coupled model satisfies the following formula: Where θ represents the coupled model parameters, specifically the set of weights or coefficients to be trained within the coupled model, which is updated through an iterative training algorithm; M represents the number of data points, indicating the number of samples or measurement times included in the training set; and i represents the sequential number; x i This refers to the input vector of the i-th sample; This represents the predicted temperature, given an input vector x. i The output results under the coupling model parameter θ; Represents the reference temperature, given the input vector x. i The output results come from temperature sensor networks, thermal imager measurements, or experimental calibration values. As a balancing factor, it controls the effect on the regularization term. The degree of importance attached to it; This is a regularization term to suppress overfitting.

3. The temperature correction method for the hot press according to claim 2, characterized in that, The compensation coefficient k of the j-th sensor j The method of obtaining it is: Calculate the difference between the actual temperature of the j-th sensor and the temperature predicted by the coupled model, denoted as . ; Calculate the difference between the temperature of the j-th sensor and the readings of its adjacent normal sensors, denoted as . ; The compensation coefficient is calculated using the following formula. in, and It is a weighting factor that controls the relative impact of the difference between the actual temperature and the temperature predicted by the coupled model.

4. The temperature correction method for a hot press according to claim 3, characterized in that, The local compensation mode includes: Obtaining the local compensation control signal: For the region where the over-limit distortion marker is located, define the proportional gain coefficient and the integral gain coefficient, and calculate the local compensation control signal through the local compensation control signal model. Let 's' represent the sequential number of the over-limit distortion region. The local compensation control signal model satisfies the following formula: in, This represents a local compensation control signal for adjusting the heating or cooling amount of the s-th over-limit distortion region at time k. Let be the temperature deviation of the s-th out-of-limit distortion region at time k. A positive value indicates that the actual temperature is higher than the target temperature, and cooling is required; if... A negative value indicates that the actual temperature is lower than the target temperature, and heating is required. Let be the proportional gain coefficient for the s-th region of excessive distortion. Let be the integral gain coefficient for the s-th region of excessive distortion. In the time interval The integral over the temperature deviation is given by τ, where τ is the integral variable.

5. The temperature correction method for a hot press according to claim 3, characterized in that, The activation of the global balance control mode includes: Coupled input correction: Based on the initial input of the coupled model, add a temperature correction amount. The temperature correction amount satisfies the following formula: in, The balance coefficient controls the impact of the corresponding global correction on the overall heating or cooling of the system. Let N be the temperature deviation at the j-th sensor; N is the number of sensors; and η is a regularization constant to avoid numerical instability caused by the denominator approaching zero. Power and pressure are adjusted in tandem using temperature correction: The temperature correction is decomposed into adjustment signals for heating power W and pressure P, where ΔW is the adjustment amount for heating power and ΔP is the adjustment amount for pressure. Where, k W k P These are correction factors for heating power and pressure, respectively, which are adjusted according to experimental or process requirements; If temperature correction >0, increases heating power, increases pressure, accelerates heating and improves thermal contact efficiency; If temperature correction <0, reducing heating power, while also reducing pressure, slowing down the temperature rise and reducing heat accumulation.

6. The temperature correction method for a hot press according to claim 1, characterized in that, When the hot press is in a state of frequent start-stop, the iteration cycle of local compensation mode and global balance control is automatically shortened, so that global adjustment and local compensation are alternated at a higher frequency to adapt to sudden temperature fluctuations or material changes.

7. A temperature correction system for a hot press, used to implement the temperature correction method for the hot press described in claim 1, characterized in that, include: The multidimensional data acquisition and preprocessing module is used to acquire multidimensional operating data of the hot press, perform noise reduction on the acquired data, use interpolation algorithms to fill in missing values, format the processed data into a time series, and output the preprocessed dataset. Coupled model construction module: Based on the preprocessed dataset, it identifies the coupling relationship between temperature, pressure, and material properties; it trains the coupled model through a multi-layer neural network to generate a coupled model that can predict the temperature distribution at different process stages; the preprocessed dataset is input into the coupled model, and the corresponding temperature prediction distribution is output. Deviation Feature Extraction and Labeling Module: Based on the predicted temperature distribution and the actual measured temperature distribution, calculate the difference and extract local deviation features, including deviation amplitude, duration and spatial distribution; determine whether the source of distortion is related to sensor drift or abnormal material thermal diffusion; if it is sensor drift, correct the sensor output by calculating the compensation coefficient. Generate local distortion markers, which include location coordinates and key information about deviation attributes; Compensation strategy determination and execution module: Analyze local distortion markers. If the distortion is identified as exceeding the limit, the compensation mode is activated to adjust the heating power or auxiliary cooling of the distorted area. Otherwise, it enters the global balance control mode and makes linkage fine adjustments to the pressure and temperature parameters of the hot press; Correction result verification and iteration module: Record the compensated temperature measurement value and compare it with the predicted value of the coupled model; If the deviation decreases and the local distortion markers converge, a stable temperature correction result is output; otherwise, the deviation feature extraction and marking module is returned to regenerate the local distortion markers and iteratively optimize the coupled model and compensation strategy.

Images