Optimization method of heat treatment process for carbon dioxide cylinders
By deploying cellular networks and intelligent decision-making systems, the heat treatment process of carbon dioxide cylinders is monitored and optimized in real time, and the problems of low control accuracy and lack of real-time monitoring in traditional methods are solved, and the quality and efficiency of cylinder heat treatment are improved.
Patent Information
- Application Number
- CN202411311616.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-20
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-20
AI Technical Summary
The traditional carbon dioxide cylinder heat treatment method has low control accuracy and lacks real-time monitoring and feedback, which affects the consistency of its performance and quality.
Deploy cellular networks, connect sensor groups, intelligent decision-making and control centers and heat treatment control modules, perform gradient descent compensation through real-time data monitoring and feature extraction, optimize heat treatment processes, and improve process accuracy through timing feedback verification.
The quality and efficiency of carbon dioxide cylinder heat treatment have been improved, ensuring the efficient, accurate and stable process.
Smart Images

Figure CN118861582B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of heat treatment, and in particular to a method for optimizing the heat treatment process of a carbon dioxide cylinder. Background Art
[0002] With the rapid development of industrial production and transportation, CO2 cylinders are increasingly used in various industries. These cylinders are often used to store and transport gases, and play an important role in the fields of medical, chemical, food processing, etc. However, with the continuous expansion of production scale and the complexity of process flow, the heat treatment process of CO2 cylinders is also facing increasing challenges.
[0003] Traditional heat treatment methods for CO2 cylinders often have problems such as low control accuracy and lack of real-time monitoring and feedback. These deficiencies make it difficult to ensure the performance and quality consistency of cylinders during the heat treatment process, which in turn affects their safety and reliability in practical applications. Modern industrial production has increasingly higher quality requirements for CO2 cylinders, requiring more efficient, precise and stable heat treatment processes to ensure product quality and performance. Summary of the invention
[0004] The present application provides a method for optimizing the heat treatment process of carbon dioxide cylinders, which is used to solve the technical problems of low control accuracy and lack of real-time monitoring and feedback in the prior art.
[0005] In view of the above problems, the present application provides a method for optimizing the heat treatment process of carbon dioxide cylinders.
[0006] In a first aspect of the present application, a method for optimizing the heat treatment process of a carbon dioxide cylinder is provided, the method comprising:
[0007] A cellular network is deployed, wherein the cellular network is used to connect a sensor group, an intelligent decision and control center, and a heat treatment control module in series, wherein the sensor group is used to monitor the real-time data of a carbon dioxide cylinder, and the intelligent decision and control center is used to perform prediction and compensation decisions for heat treatment; the intelligent decision and control center calls a preprocessing layer to perform feature extraction of the carbon dioxide cylinder, and synchronously sends the feature extraction results and the calibrated heat treatment process to a key node analysis network to establish a timing key node; the sensor group is configured based on the feature extraction results, and the sensor group is activated at the timing key node to perform data acquisition and establish a timing data set; the timing data set and the calibrated heat treatment process are sent to the intelligent decision and control center, and the loss analysis network is called by the intelligent decision and control center to perform control loss analysis, and gradient descent compensation is performed based on the control loss analysis result to generate a gradient descent compensation result; feedback parameters are established according to the gradient descent compensation result, and the heat treatment control module is called through the feedback parameters to perform heat treatment process optimization, and a timing feedback window is constructed according to the feedback parameters; timing feedback verification is performed through the timing feedback window, and the timing feedback verification result is reported.
[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0009] The present application deploys a cellular network, which is used to connect a sensor group, an intelligent decision and control center, and a heat treatment control module in series. The sensor group is used to monitor the real-time data of the carbon dioxide cylinder, and the intelligent decision and control center is used to execute prediction and compensation decisions for the heat treatment; the intelligent decision and control center calls the preprocessing layer to perform feature extraction of the carbon dioxide cylinder, and synchronously sends the feature extraction results and the calibrated heat treatment process to the key node analysis network to establish a timing key node; configures the sensor group based on the feature extraction results, activates the sensor group at the timing key node, performs data collection, and establishes a timing data set; sends the timing data set and the calibrated heat treatment process to the intelligent decision and control center, calls the loss analysis network to perform control loss analysis through the intelligent decision and control center, and performs gradient descent compensation based on the control loss analysis results to generate gradient descent compensation results; establishes feedback parameters based on the gradient descent compensation results, calls the heat treatment control module through the feedback parameters to perform heat treatment process optimization, and constructs a timing feedback window based on the feedback parameters; performs timing feedback verification through the timing feedback window, and reports the timing feedback verification results. The present invention solves the technical problems of low control accuracy and lack of real-time monitoring and feedback in the prior art, and achieves the technical effect of improving the quality and efficiency of heat treatment of carbon dioxide cylinders through real-time monitoring and feedback verification. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0011] Figure 1 A schematic flow chart of a method for optimizing the heat treatment process of a carbon dioxide cylinder provided in an embodiment of the present application;
[0012] Figure 2 A schematic flow chart of obtaining a calibration heat treatment process in the heat treatment process optimization method for a carbon dioxide cylinder provided in an embodiment of the present application. DETAILED DESCRIPTION
[0013] The present application provides a method for optimizing the heat treatment process of carbon dioxide cylinders, which is used to solve the technical problems of low control accuracy and lack of real-time monitoring and feedback in the prior art. Through real-time monitoring and feedback verification, the technical effect of improving the quality and efficiency of the heat treatment of carbon dioxide cylinders is achieved.
[0014] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0015] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.
[0016] Embodiment 1
[0017] like Figure 1 As shown, the present application provides a method for optimizing the heat treatment process of a carbon dioxide cylinder, the method comprising:
[0018] Step S100: deploying a cellular network, wherein the cellular network is used to connect in series a sensor group, an intelligent decision and control center, and a heat treatment control module, wherein the sensor group is used to monitor real-time data of a carbon dioxide cylinder, and the intelligent decision and control center is used to execute prediction and compensation decisions for heat treatment.
[0019] In the embodiment of the present application, when deploying a cellular network, the cellular network hardware is first installed, including base stations and wireless transmission equipment, and then the network parameters are configured to ensure that the network coverage can cover the working areas of all sensor groups, intelligent decision and control centers, and heat treatment control modules. The sensor group, intelligent decision and control center, and heat treatment control module are connected in series through the cellular network. The sensor group includes an ambient temperature sensor, etc., which is used to monitor real-time data of the heat treatment environment. The intelligent decision and control center performs prediction and compensation decisions for heat treatment based on real-time data.
[0020] Step S200: The intelligent decision and control center calls the preprocessing layer to perform feature extraction of the carbon dioxide cylinder, and synchronously sends the feature extraction results and the calibrated heat treatment process to the key node analysis network to establish a timing key node.
[0021] In an embodiment of the present application, the intelligent decision and control center calls the preprocessing layer to perform preliminary processing on the data received from the sensor group, including data cleaning and normalization, to ensure the accuracy and consistency of the data. Next, the preprocessed data is feature extracted by a pre-trained convolutional neural network to obtain a feature extraction result. The feature extraction result is synchronized with the pre-calibrated heat treatment process, and the time consistency of the data and process steps is ensured by a time synchronization protocol such as NTP. Next, the synchronized feature extraction results and the calibrated heat treatment process are sent to the key node analysis network. The key node analysis network runs on a graph database for storing and analyzing these time series data. By applying the graph neural network algorithm, the relationship between the features is analyzed and the key nodes in the heat treatment process are identified. These key nodes represent important time points in the heat treatment process. Finally, the key nodes are arranged in the time sequence of the heat treatment process to obtain the time sequence key nodes.
[0022] Step S300: configuring a sensor group based on the feature extraction result, activating the sensor group at a critical timing node, performing data collection, and establishing a timing data set.
[0023] In the embodiment of the present application, the specific parameters to be monitored, such as temperature, stress distribution, etc., are determined based on the key feature results extracted by the intelligent decision and control center. The appropriate sensor type is selected according to the parameters to be monitored. For example, a temperature sensor is used to monitor the ambient temperature, etc.
[0024] According to the previously identified key timing nodes, the corresponding sensor groups are activated at these key time points. The intelligent decision and control center sends instructions to start data collection at key nodes. After activation, the sensor group starts to collect real-time data about the environment around the carbon dioxide cylinder. These data include key parameters such as temperature and stress distribution. The collected data is timestamped and formatted. The processed data is stored in a time series order to form a time series data set.
[0025] Step S400: Send the time series data set and the calibrated heat treatment process to the intelligent decision and control center, call the loss analysis network through the intelligent decision and control center to perform control loss analysis, and perform gradient descent compensation based on the control loss analysis result to generate a gradient descent compensation result.
[0026] In an embodiment of the present application, after the intelligent decision and control center receives the time series data set and the calibrated heat treatment process, it calls the loss analysis network to start the control loss analysis. The loss analysis network analyzes the various losses generated during the heat treatment process through a preset algorithm. These losses include temperature loss, stress loss, and component loss. Next, gradient descent compensation is performed according to the control loss analysis results to optimize the heat treatment process. The gradient descent algorithm minimizes the total loss value by continuously adjusting the heat treatment process parameters, such as heating power, cooling rate, and holding time. During the gradient descent compensation process, the intelligent decision and control center continuously adjusts the heat treatment process parameters until the loss function converges to the minimum value, and finally generates the gradient descent compensation result.
[0027] Step S500: establishing feedback parameters according to the gradient descent compensation result, calling the heat treatment control module through the feedback parameters to perform heat treatment process optimization, and constructing a timing feedback window according to the feedback parameters.
[0028] In an embodiment of the present application, according to the result of gradient descent compensation, the heat treatment process parameter values that need to be adjusted are determined, and these parameter values are used as feedback parameters. The feedback parameters are sent to the heat treatment control module, and the module is called to perform heat treatment process optimization. The heat treatment control module performs heat treatment process optimization according to the received feedback parameters, and adjusts the operating parameters of the heat treatment equipment, such as the temperature setting of the heater, the control of the holding time, etc. In the intelligent decision-making and control center, a timing feedback window is constructed based on the feedback parameters and the real-time data of the heat treatment process. The timing feedback window is used to display the state and performance of the heat treatment process at different time points, as well as the effect after adjustment by the feedback parameters.
[0029] Step S600: performing timing feedback verification through the timing feedback window, and reporting the timing feedback verification result.
[0030] In an embodiment of the present application, key parameter data are first collected in real time during the heat treatment process by a sensor group installed around the carbon dioxide cylinder. These data include temperature, stress and composition changes, etc. Next, statistical analysis is performed to verify whether the change in the parameter is within the expected range. Statistical analysis software such as MATLAB is used to calculate the real-time mean and variance of the key parameters and compare them with the preset target values. Through these statistical analyses, the change pattern and abnormal conditions of the parameters are identified. In the timing feedback window, the parameter value monitored in real time is compared with the preset target parameter value, and the actual parameter value and target parameter value of each time node are recorded. Finally, the timing feedback verification result is generated, including the actual parameter value, target parameter value and deviation in each timing feedback window.
[0031] Furthermore, step S400 in the method provided in the application embodiment further includes:
[0032] Establish a feedback cycle, the feedback cycle includes A critical timing node, is an integer greater than 1; configure the loss analysis network as follows:
[0033] ;
[0034] in, Characterize the total loss, is the total number of critical timing nodes in the current feedback cycle, For any critical timing node, Characterizing the key timing nodes The temperature loss, Characterizing the key timing nodes The stress loss, Characterizing the key timing nodes The loss of ingredients, , , Timing critical node The corresponding weighted coefficients of temperature loss, stress loss and component loss; and completing the control loss analysis according to the loss analysis network.
[0035] In the embodiment of the present application, when performing the control loss analysis, the technical experts first determine the feedback cycle based on the demand, and the determined feedback cycle includes: A critical timing node.
[0036] Next, the control loss analysis is performed through the loss analysis network, and the time series data set is input into the loss analysis network, and the total loss within the feedback cycle is calculated through the formula. The time series data set includes temperature, stress and composition change data. , , Timing critical node The corresponding weighting coefficients of temperature loss, stress loss, and component loss are pre-set by technical experts based on historical experience.
[0037] Furthermore, in the method provided in the embodiment of the application, the step of performing gradient descent compensation based on the control loss analysis result and generating the gradient descent compensation result further includes:
[0038] Obtain the cycle response time limit constraint of the current feedback cycle; configure the learning rate according to the cycle response time limit constraint and the total loss value, and perform gradient descent compensation based on the learning rate, as follows:
[0039] ;
[0040] ;
[0041] ;
[0042] ;
[0043] ;
[0044] ;
[0045] in, Characterize the temperature gradient, Characterize the cooling rate gradient, Characterize the holding time gradient, Characterizing the key timing nodes The rate at which the temperature difference changes with the heating power, Characterizing the key timing nodes The rate at which the stress difference changes with heating power, Characterizing the key timing nodes The rate at which the composition difference of changes with heating power, Characterizing the key timing nodes The rate at which the temperature difference changes with the cooling rate, Characterizing the key timing nodes The rate at which the stress difference changes with cooling rate, Characterizing the key timing nodes The rate at which the composition difference of changes with cooling rate, Characterizing the key timing nodes The rate at which the temperature difference changes with the holding time, Characterizing the key timing nodes The rate at which the stress difference changes with the holding time, Characterizing the key timing nodes The rate at which the composition difference changes with the holding time, is the learning rate, , , are the original heating power, original cooling rate and original holding time, respectively. , , To update heating power, update cooling rate and update holding time.
[0046] In the embodiment of the present application, the cycle response time limit constraint refers to the time response requirement that the process must meet in each feedback cycle. The specific value of the cycle response time limit is obtained through documents such as system specifications, production standards or safety specifications. According to the cycle response time limit constraint and the total loss value, an adaptive learning rate algorithm such as AdaGrad or Adam is used to configure the learning rate of gradient descent. , ensuring dynamic adjustment of learning rate in different feedback cycles.
[0047] Next, the gradients of heating power, cooling rate, and holding time are calculated by formula decomposition. The core formula of the gradient descent algorithm is used to update the parameters. For example, for the update of heating power, the formula can be used ,in is the learning rate, and the temperature gradient is the partial derivative of the loss function with respect to the heating power. At the end of each feedback cycle, the gradient is recalculated and the parameters are updated. This process is repeated until the performance requirements are met or the maximum number of iterations is reached. Finally, the gradient descent compensation result is generated, including the optimized heating power, cooling rate, and holding time.
[0048] Furthermore, step S500 in the method provided in the application embodiment also includes:
[0049] Perform parameter analysis on the feedback parameters to establish a first feedback stage and a second feedback stage; and establish a timing feedback window based on time nodes of the first feedback stage and the second feedback stage.
[0050] In the embodiment of the present application, the feedback parameters are analyzed using statistical indicators such as mean and variance, and the mean and variance of each parameter at different stages are calculated. Based on the parameter analysis results, the feedback cycle is divided into a first feedback stage and a second feedback stage. Each stage corresponds to a specific parameter change characteristic. According to the parameter change trend and the process, the start and end time of each feedback stage are determined. For example, the first feedback stage corresponds to the initial heating and cooling process, and the second feedback stage corresponds to the insulation and cooling end process.
[0051] After determining the time nodes of the first feedback stage and the second feedback stage, establish a timing feedback window based on these time nodes. First, according to the divided stages, use time management tools, such as Gantt charts or project management software to mark key time nodes and make a detailed time plan. Then use time series analysis methods, such as Python's Pandas library, to establish a timing feedback window and define the monitoring and feedback mechanism within the window. Ensure that the timing feedback window can cover the key time nodes of the first feedback stage and the second feedback stage, so as to achieve real-time monitoring and adjustment of process parameters. Through the above process, the establishment of the timing feedback window is completed.
[0052] Furthermore, the method provided in the application embodiment also includes:
[0053] The feedback phase identifier of the timing feedback verification result is called; the compensation strategy library is called with the feedback phase identifier to construct a penalty factor of the learning rate; the learning rate is optimized by the penalty factor, and the gradient descent compensation is continued according to the optimized learning rate.
[0054] In the embodiment of the present application, key parameters are extracted from the real-time data collected in the timing feedback window to determine the current feedback stage identifier. The feedback stage identifier is classified into different process stages, such as the initial heating stage, the insulation stage, the cooling stage, etc., and the corresponding time nodes and parameter states are recorded.
[0055] Next, based on the feedback stage identifier, the corresponding compensation strategy is called from the compensation strategy library. The compensation strategy library contains optimization strategies and parameter adjustment schemes for different feedback stages. According to the feedback stage identifier, the corresponding compensation strategy is retrieved from the strategy library. The compensation strategy contains a method for constructing a learning rate penalty factor based on the current feedback stage. The penalty factor is a multiplier used to adjust the learning rate, and its value depends on the current feedback stage and the performance of the system in that stage. For example, in a stage with poor performance, the penalty factor may be large to reduce the learning rate and reduce the step size of parameter updates to avoid further instability of the system. According to the formula or rule in the compensation strategy, combined with the specific parameters of the current feedback stage, such as time length, performance indicators, etc., the specific value of the penalty factor is calculated. After the penalty factor is calculated, it is applied to the current learning rate to optimize the learning rate. The specific optimization method is to multiply the original learning rate by the penalty factor to obtain a new learning rate. After obtaining the optimized learning rate, the gradient descent compensation process is continued.
[0056] Further, such as Figure 2 As shown, in the method provided in the embodiment of the application, obtaining the calibration heat treatment process also includes:
[0057] The historical control data of the heat treatment control module is obtained, and the historical control data is preprocessed, wherein the data preprocessing includes complete data identification and abnormal data screening; the data preprocessing result is matched with the feature extraction result to establish a similar solution space; the similar solution space is fitted and optimized, and a calibration heat treatment process is constructed according to the fitting and optimization result.
[0058] In an embodiment of the present application, historical control data is extracted from a control system database, including parameters such as temperature, pressure, time, stress and composition. Data preprocessing is performed on the historical control data, and data processing tools such as Excel or MATLAB are used to check and identify the data to ensure that each record contains all required parameter values and there are no missing values. Statistical methods such as Z scores and interquartile range methods are used to detect and remove outliers and noise in the data set, which may be caused by sensor failure or recording errors. Next, the preprocessed historical data and the current feature extraction results are matched by similarity calculation methods such as Euclidean distance and cosine similarity to identify similar historical records. According to the matching results, a similar solution space is established.
[0059] In the similar solution space, optimization algorithms such as genetic algorithm and particle swarm optimization are used for fitting and optimization. The goal is to find a set of optimal heat treatment process parameters that can meet specific performance indicators such as material properties and production efficiency while being as close to or reaching the best performance in historical data as possible. Based on the results of fitting and optimization, a calibration heat treatment process is constructed. This process includes specific parameters such as temperature curve, heating / cooling rate, and holding time.
[0060] Furthermore, in the method provided in the embodiment of the application, the data preprocessing of the historical control data further includes:
[0061] Establish a time sequence priority, perform similar data authentication of historical control data through the time sequence priority, and generate a first elimination coefficient; perform data clustering on the historical control data, generate a group smoothing evaluation result under the clustering group based on the data clustering result, and construct a second elimination coefficient based on the group smoothing evaluation result; complete abnormal data screening based on the first elimination coefficient and the second elimination coefficient.
[0062] In the embodiment of the present application, a technical expert assigns a priority to each time node according to the process steps and importance of the heat treatment process. Key process nodes, such as the start of heating, the start of insulation and the start of cooling, have a higher priority. For each data point in the historical control data, according to its timing priority, find a data point with a similar timing and similar process conditions. Use an appropriate similarity measurement method, such as Euclidean distance, cosine similarity, etc. to calculate the similarity between the target data point and similar data points. At the same time, a similarity threshold is set. When the similarity is lower than the threshold, it is considered that the target data point may be abnormal data. Based on the similarity results, a first elimination coefficient is generated for each data point, that is, 1-similarity is used as the first elimination coefficient. The lower the similarity of the data point, the higher its first elimination coefficient, indicating that it is more likely to be abnormal data.
[0063] Next, cluster analysis is performed on the historical control data. Clustering algorithms such as K-means clustering and hierarchical clustering are used to cluster the data, and similar process conditions are divided into the same group to obtain different cluster populations. For each group, the statistics of its internal data points, such as the standard deviation, are calculated. A smoothness threshold is set according to actual needs. This threshold is determined by technical experts. If the standard deviation of the group exceeds the smoothness threshold, the elimination coefficient is the difference between the standard deviation and the threshold minus the number of group data points; if it does not exceed the threshold, the elimination coefficient is 0. Through this process, the second elimination coefficient is determined.
[0064] Combine the first elimination coefficient and the second elimination coefficient to calculate a comprehensive elimination coefficient for each data point. The comprehensive elimination coefficient is the weighted sum of the two coefficients, where the first elimination coefficient and the second elimination coefficient have the same weight. At the same time, technical experts set the comprehensive elimination threshold. Based on the comprehensive elimination coefficient and threshold, eliminate the data with a comprehensive elimination coefficient higher than the comprehensive elimination threshold to complete the abnormal data screening.
[0065] Furthermore, the method provided in the application embodiment also includes:
[0066] A self-check cycle is established, and a control self-check is performed on the heat treatment control module during the self-check cycle to generate a control self-check result; and control compensation of the heat treatment control module is performed based on the control self-check result.
[0067] In the embodiment of the present application, the inspection cycle of the heat treatment equipment in the industry standard is 1 year. However, due to the importance and real-time nature of the heat treatment control module, the self-test cycle is set to a shorter time, such as quarterly or monthly. Self-test hardware, such as temperature sensors, pressure sensors, etc., are integrated in the heat treatment control system to monitor and record key parameters in real time. According to the preset self-test process, various detection tasks, such as temperature stability test, pressure response test, etc., are automatically performed. Key data in the detection process, such as temperature fluctuation range, pressure response time, etc., are recorded in real time. The recorded data is statistically analyzed, and the qualifiedness of each test is judged according to the preset evaluation criteria, and a self-test result report is generated. In the case of slight abnormalities, the control parameters, such as heating rate, holding time, etc., are automatically adjusted according to the preset compensation algorithm. The compensation algorithm adopts control algorithms such as PID control and fuzzy control to improve control accuracy and stability. For severe abnormalities, an alarm is issued and the staff is regulated.
[0068] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:
[0069] The present application deploys a cellular network, which is used to connect a sensor group, an intelligent decision and control center, and a heat treatment control module in series. The sensor group is used to monitor the real-time data of the carbon dioxide cylinder, and the intelligent decision and control center is used to execute prediction and compensation decisions for the heat treatment; the intelligent decision and control center calls the preprocessing layer to perform feature extraction of the carbon dioxide cylinder, and synchronously sends the feature extraction results and the calibrated heat treatment process to the key node analysis network to establish a timing key node; configures the sensor group based on the feature extraction results, activates the sensor group at the timing key node, performs data collection, and establishes a timing data set; sends the timing data set and the calibrated heat treatment process to the intelligent decision and control center, calls the loss analysis network to perform control loss analysis through the intelligent decision and control center, and performs gradient descent compensation based on the control loss analysis results to generate gradient descent compensation results; establishes feedback parameters based on the gradient descent compensation results, calls the heat treatment control module through the feedback parameters to perform heat treatment process optimization, and constructs a timing feedback window based on the feedback parameters; performs timing feedback verification through the timing feedback window, and reports the timing feedback verification results. The present invention solves the technical problems of low control accuracy and lack of real-time monitoring and feedback in the prior art, and achieves the technical effect of improving the quality and efficiency of heat treatment of carbon dioxide cylinders through real-time monitoring and feedback verification.
[0070] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0071] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
[0072] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. A method for optimizing the heat treatment process of a carbon dioxide cylinder, characterized in that: The method comprises: Deploy a cellular network, wherein the cellular network is used to connect in series a sensor group, an intelligent decision and control center, and a heat treatment control module, wherein the sensor group is used to monitor the real-time data of the carbon dioxide cylinder, and the intelligent decision and control center is used to perform prediction and compensation decisions for heat treatment; The intelligent decision and control center calls the preprocessing layer to perform feature extraction of the carbon dioxide cylinder, and synchronously sends the feature extraction results and the calibrated heat treatment process to the key node analysis network to establish the timing key nodes; Configure a sensor group based on the feature extraction result, activate the sensor group at a critical timing node, perform data collection, and establish a timing data set; Sending the time series data set and the calibrated heat treatment process to an intelligent decision and control center, calling a loss analysis network through the intelligent decision and control center to perform control loss analysis, and performing gradient descent compensation based on the control loss analysis result to generate a gradient descent compensation result; Establishing feedback parameters according to the gradient descent compensation result, calling the heat treatment control module to perform heat treatment process optimization through the feedback parameters, and constructing a timing feedback window according to the feedback parameters; Performing timing feedback verification through the timing feedback window and reporting the timing feedback verification result; The method of calling the loss analysis network to perform control loss analysis through the intelligent decision and control center further includes: Establish a feedback cycle, the feedback cycle includes A critical timing node, is an integer greater than 1; Configure the loss analysis network as follows: ; in, Characterize the total loss, is the total number of critical timing nodes in the current feedback cycle, For any critical timing node, Characterizing the key timing nodes The temperature loss, Characterizing the key timing nodes The stress loss, Characterizing the key timing nodes The loss of ingredients, , , Timing critical node The corresponding weighting coefficients of temperature loss, stress loss, and composition loss; completing a control loss analysis according to the loss analysis network; The step of performing gradient descent compensation based on the control loss analysis result to generate a gradient descent compensation result further includes: Get the cycle response time limit constraint of the current feedback cycle; The learning rate is configured according to the cycle response time limit constraint and the total loss value, and the gradient descent compensation is performed based on the learning rate as follows: ; ; ; ; ; ; in, Characterize the temperature gradient, Characterize the cooling rate gradient, Characterize the holding time gradient, Characterizing the key timing nodes The rate at which the temperature difference changes with the heating power, Characterizing the key timing nodes The rate at which the stress difference changes with heating power, Characterizing the key timing nodes The rate at which the composition difference of changes with heating power, Characterizing the key timing nodes The rate at which the temperature difference changes with the cooling rate, Characterizing the key timing nodes The rate at which the stress difference changes with cooling rate, Characterizing the key timing nodes The rate at which the composition difference of changes with cooling rate, Characterizing the key timing nodes The rate at which the temperature difference changes with the holding time, Characterizing the key timing nodes The rate at which the stress difference changes with the holding time, Characterizing the key timing nodes The rate at which the composition difference changes with the holding time, is the learning rate, , , are the original heating power, original cooling rate and original holding time, respectively. , , To update heating power, update cooling rate and update holding time.
2. The method for optimizing the heat treatment process of a carbon dioxide cylinder according to claim 1, characterized in that: The constructing the timing feedback window according to the feedback parameters further includes: Perform parameter analysis on the feedback parameters to establish a first feedback stage and a second feedback stage; A timing feedback window is established based on time nodes of the first feedback stage and the second feedback stage.
3. The method for optimizing the heat treatment process of a carbon dioxide cylinder according to claim 2, characterized in that: The method further comprises: Calling the feedback phase identifier of the timing feedback verification result; Calling the compensation strategy library with the feedback stage identifier to construct a penalty factor of the learning rate; The learning rate is optimized by the penalty factor, and the gradient descent compensation is continued according to the optimized learning rate.
4. The method for optimizing the heat treatment process of a carbon dioxide cylinder according to claim 1, characterized in that: Get a certified heat treatment process, including: Acquire historical control data of the heat treatment control module, and perform data preprocessing on the historical control data, wherein the data preprocessing includes complete data identification and abnormal data screening; Match the data preprocessing results with the feature extraction results to establish a similar solution space; The similar solution space is fitted and optimized, and a calibration heat treatment process is constructed according to the fitting and optimization results.
5. The method for optimizing the heat treatment process of a carbon dioxide cylinder according to claim 4, characterized in that: The data preprocessing of the historical control data further includes: Establishing a time sequence priority, performing similar data authentication of historical control data through the time sequence priority, and generating a first elimination coefficient; Performing data clustering on the historical control data, generating a group smoothing evaluation result under the cluster group according to the data clustering result, and constructing a second elimination coefficient based on the group smoothing evaluation result; The abnormal data screening is completed according to the first elimination coefficient and the second elimination coefficient.
6. The method for optimizing the heat treatment process of a carbon dioxide cylinder according to claim 1, characterized in that: The method further comprises: Establish a self-check cycle, perform control self-check on the heat treatment control module under the self-check cycle, and generate control self-check results; Control compensation of the heat treatment control module is performed based on the control self-test result.
Citation Information
Patent Citations
Cold and hot temperature combined system for steel part thermal forming
CN114985623A
Graphite crucible forming method and system
CN116680987A
Additive manufacturing method for multi-joint ultra-long stainless steel petroleum pipeline
CN117358944A
Balanced heating method for chip capacitor processing
CN118448182A
Intelligent control of spunlace production line using classification of current production state of real-time production line data
US11853019B1
