Electric heating cooperative control system in hot-pressing tee joint manufacturing process

Abstract

The invention discloses an electric heating cooperative control system in a hot-pressing tee joint manufacturing process, and the system comprises an obtaining module which is used for obtaining and collecting real-time operation data in the hot-pressing tee joint manufacturing process; the preprocessing module is used for preprocessing the real-time operation data to obtain preprocessed data; the prediction module is used for inputting the preprocessed data into a multi-field coupling prediction model to obtain a multi-parameter prediction result; the strategy generation module is used for generating a target control strategy based on the multi-parameter prediction result; the cooperative control module is used for controlling the manufacturing process of the hot-pressing tee joint based on the target control strategy; and the actual requirement of hot-pressing tee joint manufacturing is met, and the manufacturing accuracy and the product quality are improved.

Description

Technical Field

[0001] This invention relates to the field of electrothermal synergistic control, and in particular to an electrothermal synergistic control system for the hot-pressing tee manufacturing process. Background Technology

[0002] Hot-pressed tees are core connectors in pipeline transportation systems, widely used in major engineering projects such as oil and gas. Their forming quality directly affects the reliability and service life of the pipeline system. The manufacturing of hot-pressed tees involves a complex dynamic process of high-temperature heating and high-pressure forming. Multiple physical fields, such as temperature and stress fields, are coupled together, and multiple process parameters, such as heating power and pressing pressure, influence each other. Fluctuations in any parameter can lead to defects such as uneven wall thickness and cracks. Currently, domestic hot-pressed tee manufacturing mostly adopts traditional experience-based or single-parameter feedback control modes, which are difficult to adapt to the complex characteristics of multi-field coupling and multi-parameter coordination, resulting in obvious technical bottlenecks: incomplete data acquisition and poor real-time performance, relying heavily on manual recording or single-parameter acquisition, failing to cover multi-source data throughout the molding process, leading to blind spots in process perception and a lack of data support for precise control; lack of multi-parameter coupling prediction capabilities, making it impossible to predict the impact of parameter changes on molding quality, only able to passively adjust, making it difficult to achieve forward-looking control; insufficient coordination of electrothermal control, with independent control of heating and pressurization systems and no coordinated scheduling mechanism, easily leading to defects such as stress concentration and cracks due to parameter mismatch, affecting product qualification rate. Therefore, there is an urgent need for an electrothermal synergistic control system in the hot-pressing tee manufacturing process. Summary of the Invention

[0003] This invention aims to at least partially solve one of the technical problems in the aforementioned technologies. Therefore, the objective of this invention is to propose an electrothermal synergistic control system for the hot-pressed tee manufacturing process, which meets the actual needs of hot-pressed tee manufacturing and improves manufacturing accuracy and product quality.

[0004] To achieve the above objectives, embodiments of the present invention propose an electrothermal coordinated control system for the hot-pressed tee manufacturing process, comprising: The acquisition module is used to acquire real-time operating data during the hot-pressed tee manufacturing process; The preprocessing module is used to preprocess the real-time running data to obtain preprocessed data; The prediction module is used to input the preprocessed data into the multi-field coupled prediction model to obtain multi-parameter prediction results; The strategy generation module is used to generate target control strategies based on multi-parameter prediction results. The collaborative control module is used to control the hot-pressed tee manufacturing process based on the target control strategy.

[0005] Preferably, the acquisition module includes: The first acquisition submodule is used to collect the regional temperatures of the inner arc side, neutral zone and outer arc side in the hot-pressed tee manufacturing process in real time based on temperature sensors, as the first data; The second acquisition submodule is used to collect dynamic stress data during the hot pressing process based on strain gauge sensors, and monitor the tensile stress on the inner arc side and the compressive stress on the outer arc side as the second data. The third acquisition submodule is used to collect real-time power, current and voltage data of each heating module as third data. The fourth acquisition submodule is used to collect radial deformation data from the top of the branch pipe and both sides of the main pipe based on the laser displacement sensor, as the fourth data. The fifth acquisition submodule is used to collect ambient temperature and humidity data as the fifth data. The sixth acquisition submodule is used to collect the hot air temperature and delivery flow rate of the waste heat recovery system based on the waste heat pipeline temperature sensor and flow meter, as the sixth data. The determination submodule is used to use the first data, second data, third data, fourth data, fifth data, and sixth data as real-time operating data in the hot-pressed tee manufacturing process.

[0006] Preferably, the preprocessing module includes: The alignment submodule is used to timestamp-align all real-time running data based on the PLC controller clock to obtain aligned data. The data cleaning submodule is used to clean the aligned data, remove abnormal data, and obtain cleaned data. The matching submodule is used to associate and match temperature, stress, deformation, environmental and residual heat data in the cleaned data based on timestamps to construct an electrothermal coupling dataset; The feature extraction submodule is used to extract features from the electrothermal coupling dataset to obtain preprocessed data.

[0007] The preferred multi-field coupled prediction model includes: an input layer, a feature encoding layer, a cross-modal fusion layer, a multi-task prediction layer, and an output layer; Input layer: Receives the preprocessed feature set, including the initial temperature of the billet, ambient temperature, environmental compensation factor, real-time heating power sequence of three regions, temperature time series sequence, stress time series sequence, and deformation time series sequence; Feature encoding layer: includes 4 encoders, namely time-series temperature encoder, stress time-series encoder, power sequence encoder, and deformation time-series encoder. Each encoder is a 3-layer LSTM network, and the output dimension is 256-dimensional feature vector. Cross-modal fusion layer: The feature vectors output by the four encoders are mapped to the same dimension to construct a feature matrix. Attention weights are calculated through four cross attention heads, normalized by Softmax, and then weighted and summed. Finally, the dimensionality is reduced to a 128-dimensional cross-modal fusion feature vector through one convolutional layer. Multi-task prediction layer: includes a main task prediction branch and three auxiliary task prediction branches. The main task branch is a three-layer MLP network, and the auxiliary task branches are temperature gradient trend prediction branch, stress trend prediction branch and deformation trend prediction branch, respectively. Output layer: Outputs the temperature, stress, and radial deformation values ​​for the three future regions.

[0008] Preferred training methods for multi-field coupled prediction models include: Historical production data of hot-pressed tees were obtained and divided into training and validation datasets according to a preset ratio. The training feature set is input into the model, and the Adam optimization algorithm is used for training. The total loss function is constructed to iteratively train the model until the training result meets the requirements, thus forming the initial multi-field coupled prediction model. The initial multi-field coupled prediction model is validated based on the validation dataset. When the temperature prediction error, stress prediction error, and deformation prediction error of the validation dataset all meet the requirements, training is stopped, and the trained multi-field coupled prediction model is obtained.

[0009] Preferably, the strategy generation module includes: The calculation submodule is used for: Based on the multi-parameter prediction results, the core indicators corresponding to the temperature parameter layer, stress parameter layer, and deformation parameter layer are calculated respectively; the core indicators include the global temperature gradient, the maximum stress value, and the deformation deviation. The construction submodule is used to build a three-dimensional scenario level table and risk labeling table; The comparison submodule is used to compare the global temperature gradient, maximum stress value, and deformation deviation with the temperature gradient threshold, maximum stress threshold, and deformation deviation threshold, respectively. Based on the comparison results, it queries the three-dimensional scene level table to determine the target three-dimensional scene level. Determine the submodule, used for: Based on the three-dimensional target level query risk labeling table, determine the target risk level; Obtain the risk control strategy corresponding to the target risk level and determine the target control strategy.

[0010] Preferably, the sub-modules include: The division unit is used to classify the temperature gradient dimension, stress dimension and deformation dimension into levels respectively, so as to obtain the dimension level corresponding to each dimension. The first generation unit is used to obtain a three-dimensional scene level table based on the full permutation of the three-dimensional levels. The second generation unit is used to classify risks based on the degree of impact of the control scenario and obtain a risk labeling table.

[0011] Preferred risk control strategies include: Level 1 Risk Scenario Strategy: Use PWM pulse modulation mode to reduce power in the stress concentration area, reduce power in the first area, and simultaneously start the first and second level cooling to directionally transport waste heat to the periphery of the stress concentration area; Level 2 risk scenario strategy: Reduce power in the first zone in a preset step size, increase power in the second zone in a preset step size, start the first-level cooling, and use the residual heat for compensation in the second zone and preheating of the billet; Level 3 risk scenario strategy: Maintain current power and fine-tune fluctuations step by step, keep the first-level cooling at minimum flow rate on standby, and transfer waste heat to the preheating box of the next batch of billets.

[0012] Preferably, based on the multi-parameter prediction results, the core indicators corresponding to the temperature parameter layer, stress parameter layer, and deformation parameter layer are calculated respectively, including: The core indicator corresponding to the temperature parameter stratification is the global temperature gradient; calculate the temperature gradient evaluation value and take the maximum value as the global temperature gradient. ; ; ;in, This is the predicted temperature value for the inner arc side region; This is the predicted temperature value for the neutral zone. This is the predicted temperature value for the outer arc side region; The core indicator corresponding to the stress parameter stratification is the maximum stress: ; This represents the predicted stress value for the inner arc side region; This represents the predicted stress value for the neutral region. This represents the predicted stress value for the outer arc side region; The core indicator corresponding to the deformation parameter layer is the deformation deviation, which determines whether the predicted deformation is within the target forming range.

[0013] Preferably, the collaborative control module includes: The conversion submodule is used to convert the target control strategy into a PLC-recognizable format, including instruction type, target module, parameter value, execution time and feedback requirements. The execution submodule is used for coordinated control of the hot-pressed tee manufacturing process based on instruction type, target module, parameter value, execution time and feedback requirements; The feedback submodule is used to collect actual temperature, stress, and deformation data after the strategy is executed, and store them in the historical database for online model updates.

[0014] This invention provides an electrothermal synergistic control system for the hot-pressed tee manufacturing process. A preprocessing module preprocesses real-time operational data from the hot-pressed tee manufacturing process, effectively improving data quality and providing accurate data for subsequent predictions, reducing manufacturing problems caused by data errors. A multi-field coupled prediction model derives multi-parameter prediction results based on the preprocessed data, enabling early prediction of potential situations during the hot-pressed tee manufacturing process and allowing relevant personnel to develop contingency plans in advance. A strategy generation module generates a target control strategy based on the multi-parameter prediction results. This strategy is highly targeted and meets the actual needs of hot-pressed tee manufacturing, improving manufacturing accuracy and product quality. The synergistic control module controls the hot-pressed tee manufacturing process based on the target control strategy, achieving synergistic optimization of electrothermal processes, improving manufacturing efficiency, and reducing energy consumption and production costs. The overall system operation mechanism helps improve the stability and reliability of the hot-pressed tee manufacturing process, reducing the defect rate and improving product quality.

[0015] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a block diagram of an electrothermal synergistic control system in the hot-pressing tee manufacturing process according to an embodiment of the present invention; Figure 2 This is a block diagram of a preprocessing module according to an embodiment of the present invention; Figure 3 This is a block diagram of a strategy generation module according to an embodiment of the present invention. Detailed Implementation

[0018] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0019] Example 1: As Figure 1 As shown, an electrothermal coordinated control system for the hot-pressed tee manufacturing process includes: The acquisition module is used to acquire real-time operating data during the hot-pressed tee manufacturing process; The preprocessing module is used to preprocess the real-time running data to obtain preprocessed data; The prediction module is used to input the preprocessed data into the multi-field coupled prediction model to obtain multi-parameter prediction results; The strategy generation module is used to generate target control strategies based on multi-parameter prediction results. The collaborative control module is used to control the hot-pressed tee manufacturing process based on the target control strategy.

[0020] In this embodiment, the real-time operating data includes temperature data, stress data, current / voltage data, deformation data, ambient temperature and humidity data, and waste heat system data.

[0021] In this embodiment, the multi-field coupling prediction model is a pre-trained model used to predict the temporal changes of physical quantities, and does not directly output the ranking results.

[0022] The working principle and beneficial effects of the above technical solution are as follows: Preprocessing the real-time operating data of hot-pressed tee manufacturing by the preprocessing module effectively improves data quality, providing accurate basis for subsequent predictions and reducing manufacturing problems caused by data errors. The multi-field coupled prediction model derives multi-parameter prediction results based on the preprocessed data, enabling early prediction of potential situations during hot-pressed tee manufacturing and allowing relevant personnel to develop contingency plans in advance. The strategy generation module generates a target control strategy based on the multi-parameter prediction results. This strategy is highly targeted, meets the actual needs of hot-pressed tee manufacturing, and improves manufacturing accuracy and product quality. The collaborative control module controls the hot-pressed tee manufacturing process based on the target control strategy, achieving synergistic optimization of electrothermal processes, improving manufacturing efficiency, and reducing energy consumption and production costs. The overall system operation mechanism helps improve the stability and reliability of the hot-pressed tee manufacturing process, reduces the defect rate, and improves product quality.

[0023] Example 2: Acquisition module, including: The first acquisition submodule is used to collect the regional temperatures of the inner arc side, neutral zone and outer arc side in the hot-pressed tee manufacturing process in real time based on temperature sensors, as the first data; The second acquisition submodule is used to collect dynamic stress data during the hot pressing process based on strain gauge sensors, and monitor the tensile stress on the inner arc side and the compressive stress on the outer arc side as the second data. The third acquisition submodule is used to collect real-time power, current and voltage data of each heating module as third data. The fourth acquisition submodule is used to collect radial deformation data from the top of the branch pipe and both sides of the main pipe based on the laser displacement sensor, as the fourth data. The fifth acquisition submodule is used to collect ambient temperature and humidity data as the fifth data. The sixth acquisition submodule is used to collect the hot air temperature and delivery flow rate of the waste heat recovery system based on the waste heat pipeline temperature sensor and flow meter, as the sixth data. The determination submodule is used to use the first data, second data, third data, fourth data, fifth data, and sixth data as real-time operating data in the hot-pressed tee manufacturing process.

[0024] The working principle and beneficial effects of the above technical solution are as follows: Real-time data collection from multiple dimensions such as temperature, stress, power, deformation, environmental conditions, and waste heat recovery can comprehensively reflect the status of the hot-pressed tee manufacturing process, providing rich and comprehensive data for subsequent precise control; Monitoring the temperature, stress, and deformation of key parts such as the inner arc side, neutral zone, and outer arc side of the hot-pressed tee helps to promptly identify potential problems in the manufacturing process and ensure product quality; Collecting power, current, and voltage data of each heating module, as well as relevant data from the waste heat recovery system, facilitates the assessment of energy usage and provides data support for optimizing energy consumption and improving energy efficiency; Collecting ambient temperature and humidity data can reduce the interference of environmental factors on the manufacturing process, making the entire manufacturing process more stable and reliable.

[0025] Example 3: As Figure 2 As shown, the preprocessing module includes: The alignment submodule is used to timestamp-align all real-time running data based on the PLC controller clock to obtain aligned data. The data cleaning submodule is used to clean the aligned data, remove abnormal data, and obtain cleaned data. The matching submodule is used to associate and match temperature, stress, deformation, environmental and residual heat data in the cleaned data based on timestamps to construct an electrothermal coupling dataset; The feature extraction submodule is used to extract features from the electrothermal coupling dataset to obtain preprocessed data.

[0026] In this embodiment, the initial temperature of the billet is obtained from historical batch data.

[0027] In this embodiment, the real-time heating power of the three zones is calculated using third data.

[0028] In this embodiment, the temperature-stress-deformation time series is extracted from the most recent 5-second window of data.

[0029] In this embodiment, the environmental compensation factor = 1 + 0.01 × (ambient humidity - 50%).

[0030] In this embodiment, the output feature set includes: billet material code, billet size, ambient temperature, environmental compensation factor, three-zone real-time heating power sequence, temperature time sequence, stress time sequence, and deformation time sequence.

[0031] The working principle and beneficial effects of the above technical solution are as follows: The alignment submodule timestamps the real-time running data based on the PLC controller clock, ensuring that data from different sources are consistent in the time dimension, providing an accurate and synchronized data foundation for subsequent analysis, and avoiding analysis errors caused by time differences; The data cleaning submodule removes abnormal data, effectively removing noise and erroneous information, improving the reliability and accuracy of the data, and enabling subsequent processing and analysis to be based on high-quality data; The matching submodule associates and matches the cleaned data to construct an electrothermal coupling dataset, which can reflect the intrinsic relationship between various parameters and helps to deeply analyze the electrothermal interaction mechanism in the hot-pressed tee manufacturing process; The feature extraction submodule extracts key features from the electrothermal coupling dataset, reduces data redundancy, highlights important information in the data, and improves the efficiency and accuracy of subsequent prediction and control.

[0032] Example 4: Multi-field coupled prediction model, including: input layer, feature encoding layer, cross-modal fusion layer, multi-task prediction layer and output layer; Input layer: Receives the preprocessed feature set, including the initial temperature of the billet, ambient temperature, environmental compensation factor, real-time heating power sequence of three regions, temperature time series sequence, stress time series sequence, and deformation time series sequence; Feature encoding layer: includes 4 encoders, namely time-series temperature encoder, stress time-series encoder, power sequence encoder, and deformation time-series encoder. Each encoder is a 3-layer LSTM network, and the output dimension is 256-dimensional feature vector. Cross-modal fusion layer: The feature vectors output by the four encoders are mapped to the same dimension to construct a feature matrix. Attention weights are calculated through four cross attention heads, normalized by Softmax, and then weighted and summed. Finally, the dimensionality is reduced to a 128-dimensional cross-modal fusion feature vector through one convolutional layer. Multi-task prediction layer: includes a main task prediction branch and three auxiliary task prediction branches. The main task branch is a three-layer MLP network, and the auxiliary task branches are temperature gradient trend prediction branch, stress trend prediction branch and deformation trend prediction branch, respectively. Output layer: Outputs the temperature, stress, and radial deformation values ​​for the three future regions.

[0033] The working principle and beneficial effects of the above technical solution are as follows: The input layer can receive various types of preprocessed feature sets, covering information such as initial temperature of billet, environmental factors, power sequence, temperature and stress time series, etc., realizing the comprehensive utilization of multi-source data, reflecting the physical characteristics of the hot-pressed tee manufacturing process more comprehensively, and providing rich basis for accurate prediction; The feature encoding layer uses multiple LSTM network encoders to process different time series data, which can capture the time dependence in the data and output fixed-dimensional feature vectors, transforming complex time series data into more representative and analyzable feature expressions, which helps the subsequent model learning and prediction; The cross-modal fusion layer maps the feature vectors output by different encoders and calculates attention weights. The recombination and dimensionality reduction processing effectively integrates data information from different modes, uncovers potential correlations between various physical quantities, and enables the model to understand and analyze the manufacturing process from multiple perspectives, enhancing the model's comprehensive analytical capabilities. The multi-task prediction layer sets up a main task prediction branch and multiple auxiliary task prediction branches, which can simultaneously predict multiple key indicators such as temperature values, stress values, and radial deformation in the three regions in the future. Furthermore, the auxiliary tasks can provide additional constraints and information to the main task, improving the accuracy and stability of the prediction. The output layer directly outputs the temperature values, stress values, and radial deformation in the three regions in the future, providing key information for real-time monitoring and precise control of the hot-pressed tee manufacturing process, which helps to adjust process parameters in a timely manner and ensure product quality.

[0034] Example 5: Training method for multi-field coupled prediction model, including: Historical production data of hot-pressed tees were obtained and divided into training and validation datasets according to a preset ratio. The training feature set is input into the model, and the Adam optimization algorithm is used for training. The total loss function is constructed to iteratively train the model until the training result meets the requirements, thus forming the initial multi-field coupled prediction model. The initial multi-field coupled prediction model is validated based on the validation dataset. When the temperature prediction error, stress prediction error, and deformation prediction error of the validation dataset all meet the requirements, training is stopped, and the trained multi-field coupled prediction model is obtained.

[0035] In this embodiment, the total loss function is: ; Predict MSE loss based on temperature; For stress prediction, MSE loss is calculated. Predict MSE loss for deformation.

[0036] The working principle and beneficial effects of the above technical solution are as follows: Historical production data of hot-pressed tee manufacturing is divided into training and validation sets according to a preset ratio, ensuring the model has sufficient data to learn patterns. Simultaneously, the validation set can be used to evaluate generalization ability, avoiding overfitting and improving the model's performance on unknown data. The Adam optimization algorithm is used to train the model, which can adaptively adjust the learning rate, accelerate convergence, improve training efficiency, and enable the model to find optimal parameters more quickly. A total loss function is constructed for iterative training, continuously adjusting model parameters based on the loss, making the model output closer to the true value and gradually improving prediction accuracy. The initial model is validated using a validation set, with temperature, stress, and deformation prediction errors as evaluation indicators. Training stops only when all errors meet the requirements, ensuring the model has high prediction accuracy and reliability in practical applications.

[0037] Example 6: As Figure 3 As shown, the strategy generation module includes: The calculation submodule is used for: Based on the multi-parameter prediction results, the core indicators corresponding to the temperature parameter layer, stress parameter layer, and deformation parameter layer are calculated respectively; the core indicators include the global temperature gradient, the maximum stress value, and the deformation deviation. The construction submodule is used to build a three-dimensional scenario level table and risk labeling table; The comparison submodule is used to compare the global temperature gradient, maximum stress value, and deformation deviation with the temperature gradient threshold, maximum stress threshold, and deformation deviation threshold, respectively. Based on the comparison results, it queries the three-dimensional scene level table to determine the target three-dimensional scene level. Determine the submodule, used for: Based on the three-dimensional target level query risk labeling table, determine the target risk level; Obtain the risk control strategy corresponding to the target risk level and determine the target control strategy.

[0038] In this embodiment, the objective function is: ;in, The maximum temperature gradient; The maximum stress; This represents the total power consumption.

[0039] The working principle and beneficial effects of the above technical solution are as follows: The calculation submodule calculates key core indicators through multi-parameter prediction results, quantifying complex hot-pressed tee manufacturing data to facilitate subsequent analysis and decision-making; the construction submodule constructs a three-dimensional scenario level table and risk labeling table, establishing a standardized assessment and management system, making risk identification and response systematic; the comparison submodule determines the target three-dimensional scenario level by comparing key indicators with thresholds, enabling rapid and accurate identification of the current manufacturing scenario status; the determination submodule determines the risk level based on the scenario level and obtains corresponding strategies, fine-tuning strategy parameters under specific risk scenarios to achieve precise and dynamic risk control, ensuring the quality and safety of hot-pressed tee manufacturing.

[0040] Example 7: Constructing a submodule, including: The division unit is used to classify the temperature gradient dimension, stress dimension and deformation dimension into levels respectively, so as to obtain the dimension level corresponding to each dimension. The first generation unit is used to obtain a three-dimensional scene level table based on the full permutation of the three-dimensional levels. The second generation unit is used to classify risks based on the degree of impact of the control scenario and obtain a risk labeling table.

[0041] In this embodiment, the temperature gradient dimension A: A1 ( ≤30℃, optimal), A2 (30℃ < ≤40℃, good), A3 (40℃< ≤50℃, critical), A4 ( (Temperature >50℃, exceeding the standard).

[0042] In this embodiment, stress dimension B: B1 ( ≤150MPa, low stress), B2 (150MPa< ≤200MPa, medium stress), B3 (200MPa< ≤250MPa, high stress), B4 ( >250MPa, overstress); In this embodiment, the deformation dimension C: C1 ( ≤0.5mm, qualified), C2 (0.5mm < ≤1.0mm, warning), C3 ( >1.0mm, exceeding the standard).

[0043] In this embodiment, control scenarios are generated based on the level combinations of the three dimensions A, B, and C, as shown in the following examples: Scenario 1: A1+B1+C1 (optimal scenario); Scenario 8: A3+B3+C2 (critical warning scenario); Scenario 24: A4+B4+C3 (severe exceedance scenario).

[0044] In this embodiment, the scenarios are divided into three levels of risk based on their impact on the hot pressing quality and equipment safety: Level 1 risk (high safety hazard): scenarios containing A4, B4, or C3 (11 categories in total); Level 2 risk (quality warning): scenarios containing A3, B3, or C2 but not containing Level 1 risk factors (9 categories in total); Level 3 risk (stable and controllable): scenarios containing only A1, A2, B1, B2, and C1 (4 categories in total).

[0045] The working principle and beneficial effects of the above technical solution are as follows: By dividing temperature gradient, stress, and deformation into levels through unit division, continuous numerical data can be discretized, reducing data complexity and improving the efficiency of subsequent data processing and analysis; the first generation unit generates a three-dimensional scene level table based on the full permutation of the three-dimensional levels, which can comprehensively consider various scenarios under different combinations of dimension levels, providing a more comprehensive basis for subsequent risk assessment and decision-making; the second generation unit divides risks based on the degree of influence of control scenarios, generating a risk labeling table, which helps to accurately assess the risks under different control scenarios, thereby taking corresponding measures to reduce risks; the results generated by these sub-modules can intuitively show the relationship between different dimension levels and risk levels, providing decision-makers with clear information and facilitating the formulation of reasonable decision-making and control strategies.

[0046] Example 8: Risk control strategy, including: Level 1 Risk Scenario Strategy: Use PWM pulse modulation mode to reduce power in the stress concentration area, reduce power in the first area, and simultaneously start the first and second level cooling to directionally transport waste heat to the periphery of the stress concentration area; Level 2 risk scenario strategy: Reduce power in the first zone in a preset step size, increase power in the second zone in a preset step size, start the first-level cooling, and use the residual heat for compensation in the second zone and preheating of the billet; Level 3 risk scenario strategy: Maintain current power and fine-tune fluctuations step by step, keep the first-level cooling at minimum flow rate on standby, and transfer waste heat to the preheating box of the next batch of billets.

[0047] In this embodiment, the waste heat path switching rule is as follows: when the risk level changes, the waste heat allocation ratio transitions linearly, and the transition time is equal to the difference between the old and new risk levels, which is 2 seconds.

[0048] In this embodiment, the temperature of the first region is greater than the temperature of the second region, that is, the first region is a high-temperature region and the second region is a low-temperature region.

[0049] In this embodiment, the electric heating power adjustment strategy is as follows: power in the stress concentration area is immediately reduced by 40% using PWM pulse modulation; power in the first area is reduced by 30% to activate power limiting protection; power in the second area is slightly increased by 5% to alleviate the gradient; and power in the neutral zone is adjusted accordingly. For every 10°C increase, power is adjusted by -0.3kW; cooling system control strategy: total liquid cooling flow rate is increased, and phase change heat absorbers are fully activated; if The rate of descent is <2°C / s or If the descent rate is <15MPa / s, the flow rate will be increased by 1L / min per second; Heat utilization path: the billet preheating channel is closed, and 100% of the residual heat is directionally transported to the stress concentration area, with a flow rate of... m³ / h; Cycle: Adjust parameters every 1 second, evaluate scene level every 1 second.

[0050] In this embodiment, the secondary risk scenario strategy, taking scenario A3+B3+C2 as an example, is as follows: Electric heating power adjustment: the power step decreases in the first region, and the power step increases in the second region; if the stress concentration area... ≥220MPa, power reduction of 0.4kW; neutral zone adjustment step ±0.2kW; cooling system control: start primary cooling, liquid cooling flow rate... (High-temperature zone accounts for 70%); Waste heat utilization path: Waste heat distribution ratio = f( ),like If the thickness is less than -0.5mm (too small), then the second zone compensation accounts for 60%, and the billet preheating accounts for 40%; if If the thickness is >0.5mm (too large), then the second region should be compensated by 30% and the billet preheated by 70%; Optimization objective: Under the premise of meeting risk degradation requirements, fine-tune the power distribution to make... Minimize; Cycle: Adjust parameters every 1 second, evaluate scene level every 2 seconds.

[0051] In this embodiment, the three-level risk scenario strategy, taking scenario A1+B1+C1 as an example, is as follows: Electric heating power adjustment: power fluctuation compensation step size for the three zones ≤ 0.1kW; based on Real-time values ​​will allocate redundant power to areas with high waste heat recovery efficiency; Cooling system control: secondary cooling will be shut down, and primary cooling flow will be adjusted. L / min; Waste heat utilization path: Waste heat allocation ratio = 85% preheating of the next batch of billets + 15% workshop heating; If the waiting time for the new batch is >60s, the preheating ratio is reduced to 50%; Cycle: Monitor the prediction results every 5s, and maintain the strategy if there are no abnormalities.

[0052] The working principle and beneficial effects of the above technical solution are as follows: power reduction and multi-stage cooling reduce stress and high-temperature damage, extending equipment life; directional waste heat transport improves stress distribution and realizes energy reuse; power step adjustment balances the power of equipment areas, ensuring uniform operation; cooling combined with waste heat utilization controls temperature and preheats the billet, improving efficiency; fine-tuning power maintains equipment stability and ensures production continuity; waste heat preheats the billet, and cooling is ready to respond to abnormalities, reducing costs.

[0053] Example 9: Based on the multi-parameter prediction results, the core indicators corresponding to the temperature parameter layering, stress parameter layering, and deformation parameter layering are calculated respectively, including: The core indicator corresponding to the temperature parameter stratification is the global temperature gradient; calculate the temperature gradient evaluation value and take the maximum value as the global temperature gradient. ; ; ;in, This is the predicted temperature value for the inner arc side region; This is the predicted temperature value for the neutral zone. This is the predicted temperature value for the outer arc side region; The core indicator corresponding to the stress parameter stratification is the maximum stress: ; This represents the predicted stress value for the inner arc side region; This represents the predicted stress value for the neutral region. This represents the predicted stress value for the outer arc side region; The core indicator corresponding to the deformation parameter layer is the deformation deviation, which determines whether the predicted deformation is within the target forming range.

[0054] In this embodiment, the global temperature gradient is: .

[0055] The working principle and beneficial effects of the above technical solution are as follows: By calculating the core indicators of temperature, stress, and deformation parameters at different levels, the state of equipment or objects can be quantitatively assessed from multiple key dimensions, providing a comprehensive and accurate understanding of their actual operating status. The global temperature gradient reflects differences in temperature distribution, the maximum stress reflects stress concentration, and the deformation deviation can detect whether it deviates from the forming target. These indicators help to promptly identify potential thermal stress, mechanical stress problems, and forming anomalies, providing early warnings of potential risks. They also provide clear quantitative basis for equipment maintenance and process adjustments, enabling decision-makers to formulate more targeted and scientific strategies based on specific indicator conditions, ensuring stable equipment operation and product quality.

[0056] Example 10: Cooperative control module, including: The conversion submodule is used to convert the target control strategy into a PLC-recognizable format, including instruction type, target module, parameter value, execution time and feedback requirements. The execution submodule is used for coordinated control of the hot-pressed tee manufacturing process based on instruction type, target module, parameter value, execution time and feedback requirements; The feedback submodule is used to collect actual temperature, stress, and deformation data after the strategy is executed, and store them in the historical database for online model updates.

[0057] The working principle and beneficial effects of the above technical solution are as follows: The conversion submodule converts the target control strategy into a PLC-recognizable format, ensuring that different strategies can be seamlessly integrated with the PLC system in the hot-pressed tee manufacturing process, thereby enhancing the system's compatibility and versatility; The execution submodule performs collaborative control of the manufacturing process based on multi-dimensional information from instructions, enabling precise regulation of each stage, ensuring the efficient and orderly progress of the hot-pressed tee manufacturing process, and improving product quality stability; The feedback submodule collects and stores actual data for online model updates, allowing the control strategy to be continuously optimized according to actual production conditions, adapting to complex and ever-changing manufacturing environments, and enhancing the system's adaptability and intelligence level.

[0058] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An electrothermal coordinated control system for the hot-pressing tee manufacturing process, characterized in that, include: The acquisition module is used to acquire real-time operating data during the hot-pressed tee manufacturing process; The preprocessing module is used to preprocess the real-time running data to obtain preprocessed data; The prediction module is used to input the preprocessed data into the multi-field coupled prediction model to obtain multi-parameter prediction results; The strategy generation module is used to generate target control strategies based on multi-parameter prediction results. The collaborative control module is used to control the hot-pressed tee manufacturing process based on the target control strategy.

2. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 1, characterized in that, The acquisition module includes: The first acquisition submodule is used to collect the regional temperatures of the inner arc side, neutral zone and outer arc side during the hot-pressed tee manufacturing process based on temperature sensors, and use them as the first data. The second acquisition submodule is used to collect dynamic stress data during the hot pressing process based on strain gauge sensors, and monitor the tensile stress on the inner arc side and the compressive stress on the outer arc side as the second data. The third acquisition submodule is used to collect real-time power, current and voltage data of each heating module as third data. The fourth acquisition submodule is used to collect radial deformation data from the top of the branch pipe and both sides of the main pipe based on the laser displacement sensor, as the fourth data. The fifth acquisition submodule is used to collect ambient temperature and humidity data as the fifth data. The sixth acquisition submodule is used to collect the hot air temperature and delivery flow rate of the waste heat recovery system based on the waste heat pipeline temperature sensor and flow meter, as the sixth data. The determination submodule is used to use the first data, second data, third data, fourth data, fifth data, and sixth data as real-time operating data in the hot-pressed tee manufacturing process.

3. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 1, characterized in that, The preprocessing module includes: The alignment submodule is used to timestamp-align all real-time running data based on the PLC controller clock to obtain aligned data. The data cleaning submodule is used to clean the aligned data, remove abnormal data, and obtain cleaned data. The matching submodule is used to associate and match temperature, stress, deformation, environmental and residual heat data in the cleaned data based on timestamps to construct an electrothermal coupling dataset; The feature extraction submodule is used to extract features from the electrothermal coupling dataset to obtain preprocessed data.

4. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 1, characterized in that, The multi-field coupled prediction model includes: an input layer, a feature encoding layer, a cross-modal fusion layer, a multi-task prediction layer, and an output layer; Input layer: Receives the preprocessed feature set, including the initial temperature of the billet, ambient temperature, environmental compensation factor, real-time heating power sequence of three regions, temperature time series sequence, stress time series sequence, and deformation time series sequence; Feature encoding layer: includes 4 encoders, namely time-series temperature encoder, stress time-series encoder, power sequence encoder, and deformation time-series encoder. Each encoder is a 3-layer LSTM network, and the output dimension is 256-dimensional feature vector. Cross-modal fusion layer: The feature vectors output by the four encoders are mapped to the same dimension to construct a feature matrix. Attention weights are calculated through four cross attention heads, normalized by Softmax, and then weighted and summed. Finally, the dimensionality is reduced to a 128-dimensional cross-modal fusion feature vector through one convolutional layer. Multi-task prediction layer: includes a main task prediction branch and three auxiliary task prediction branches. The main task branch is a three-layer MLP network, and the auxiliary task branches are temperature gradient trend prediction branch, stress trend prediction branch, and deformation trend prediction branch. Output layer: Outputs the temperature, stress, and radial deformation values ​​for the three future regions.

5. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 4, characterized in that, Training methods for multi-field coupled prediction models include: Historical production data of hot-pressed tees were obtained and divided into training and validation datasets according to a preset ratio. The training feature set is input into the model, and the Adam optimization algorithm is used for training. The total loss function is constructed to iteratively train the model until the training result meets the requirements, thus forming the initial multi-field coupled prediction model. The initial multi-field coupled prediction model is validated based on the validation dataset. When the temperature prediction error, stress prediction error, and deformation prediction error of the validation dataset all meet the requirements, training is stopped, and the trained multi-field coupled prediction model is obtained.

6. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 1, characterized in that, The strategy generation module includes: The calculation submodule is used for: Based on the multi-parameter prediction results, the core indicators corresponding to the temperature parameter layer, stress parameter layer, and deformation parameter layer are calculated respectively; the core indicators include the global temperature gradient, the maximum stress value, and the deformation deviation. The construction submodule is used to build a three-dimensional scenario level table and risk labeling table; The comparison submodule is used to compare the global temperature gradient, maximum stress value, and deformation deviation with the temperature gradient threshold, maximum stress threshold, and deformation deviation threshold, respectively. Based on the comparison results, it queries the three-dimensional scene level table to determine the target three-dimensional scene level. Determine the submodule, used for: Based on the three-dimensional target level query risk labeling table, determine the target risk level; Obtain the risk control strategy corresponding to the target risk level and determine the target control strategy.

7. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 6, characterized in that, Build submodules, including: The division unit is used to classify the temperature gradient dimension, stress dimension and deformation dimension into levels respectively, so as to obtain the dimension level corresponding to each dimension. The first generation unit is used to obtain a three-dimensional scene level table based on the full permutation of the three-dimensional levels. The second generation unit is used to classify risks based on the degree of impact of the control scenario and obtain a risk labeling table.

8. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 6, characterized in that, Risk control strategies include: Level 1 Risk Scenario Strategy: Use PWM pulse modulation mode to reduce power in the stress concentration area, reduce power in the first area, and simultaneously start the first and second stage cooling to directionally transport the waste heat to the periphery of the stress concentration area; Level 2 risk scenario strategy: Reduce power in the first zone in a preset step size, increase power in the second zone in a preset step size, start the first-level cooling, and use the residual heat for compensation in the second zone and preheating of the billet; Level 3 risk scenario strategy: Maintain current power and fine-tune fluctuations step by step, keep the first-level cooling at minimum flow rate on standby, and transfer waste heat to the preheating box of the next batch of billets.

9. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 6, characterized in that, Based on the multi-parameter prediction results, the core indices corresponding to the temperature parameter layer, stress parameter layer, and deformation parameter layer are calculated respectively, including: The core indicator corresponding to the temperature parameter stratification is the global temperature gradient; calculate the temperature gradient evaluation value and take the maximum value as the global temperature gradient. ; ; ;in, This is the predicted temperature value for the inner arc side region; This is the predicted temperature value for the neutral zone. This is the predicted temperature value for the outer arc side region; The core indicator corresponding to the stress parameter stratification is the maximum stress: ; This represents the predicted stress value for the inner arc side region; This represents the predicted stress value for the neutral region. This represents the predicted stress value for the outer arc side region; The core indicator corresponding to the deformation parameter layer is the deformation deviation, which determines whether the predicted deformation is within the target forming range.

10. The electrothermal coordinated control system in the hot-pressing tee manufacturing process as described in claim 1, characterized in that, The collaborative control module includes: The conversion submodule is used to convert the target control strategy into a PLC-recognizable format, including instruction type, target module, parameter value, execution time and feedback requirements. The execution submodule is used for coordinated control of the hot-pressed tee manufacturing process based on instruction type, target module, parameter value, execution time and feedback requirements; The feedback submodule is used to collect actual temperature, stress, and deformation data after the strategy is executed, and store them in the historical database for online model updates.

Images