Intelligent temperature control system and method for casting system

Abstract

The present invention belongs to, but is not limited to, the field of low-pressure casting technology and discloses an intelligent temperature control system and method for a casting system. Integrated sensing modules are distributed according to characteristic regions of the casting mold to monitor temperature changes at different locations within the mold in real time. A data acquisition and processing module receives data from the integrated sensing modules, preprocesses it, and stores it. A central control module receives data provided by the data acquisition and processing module and dynamically adjusts the casting equipment and the intelligent temperature control system execution module based on preset casting process parameters and real-time data. The execution module includes a heating device and a cooling device, and adjusts the process parameters of the heating and cooling channels according to instructions from the central control system to intelligently control the mold temperature. The present invention improves temperature control accuracy and enables timely temperature adjustment, thereby optimizing the casting process, reducing defects, and ensuring high-quality manufacturing of aluminum alloy wheels.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the technical field of low-pressure casting processes, and in particular relates to an intelligent temperature control system and method for a casting system. Background Art

[0002] Lightweighting is one of the important strategies for energy conservation and emission reduction of fuel vehicles and efficiency improvement of new energy vehicles. Among them, the use of aluminum alloy materials to replace traditional steel materials is a key means to achieve lightweighting of vehicles. The low-pressure casting process has become the mainstream process for producing automotive aluminum alloy castings due to its smooth filling, good shrinkage compensation effect and high degree of automation. In the low-pressure casting process, the influence of mold temperature on casting quality is particularly important. The alloy melt is in direct contact with the mold during the filling and solidification process, and is cooled and solidified through heat exchange. Therefore, the temperature distribution of the mold directly determines the solidification sequence and cooling rate of the alloy melt, which has a key influence on the solidification structure and performance of the casting. By controlling the mold temperature within a reasonable range through a suitable cooling process, it can ensure that the alloy solidifies rapidly in the expected order, thereby significantly reducing casting defects such as shrinkage cavities and porosity, shortening the production cycle, and achieving a dual improvement in casting quality and production efficiency.

[0003] Currently, in low-pressure casting production of aluminum alloy castings, mold temperature control primarily relies on an open-loop control method. Real-time mold temperature detection and monitoring are typically not performed, relying instead on on-site technicians to directly adjust cooling process parameters based on casting defects. While this method is simple to operate, it has significant drawbacks: cooling process adjustments are significantly influenced by the technicians' experience and subjective judgment, lacking a scientific basis. This results in unstable process improvements and makes it difficult to achieve automated and intelligent mold temperature control. With the advancement of lightweighting in automobiles and the increasing complexity of casting structures, achieving reasonable mold temperature control becomes even more challenging. Traditional temperature control methods are no longer able to meet modern production demands.

[0004] Temperature measurement and control of low-pressure die-cast aluminum alloy wheel molds is a critical task in the manufacturing industry. This involves acquiring and analyzing temperature data from the mold's operating status to achieve precise temperature control. Effective temperature measurement and control methods can help manufacturers reduce production costs, improve product quality, and increase production efficiency. During the low-pressure die-casting process, die temperature has a direct impact on the mold quality and production stability of the aluminum alloy wheel. Therefore, accurate temperature measurement and timely temperature adjustment are particularly important. Existing temperature control methods suffer from poor accuracy and delayed temperature adjustment. Summary of the Invention

[0005] In response to the problems existing in the prior art, the present invention provides a casting system temperature intelligent control system and method.

[0006] The present invention is implemented as follows: a casting system temperature intelligent control system, comprising:

[0007] Integrated sensing module: Consists of multiple temperature sensors, distributed according to the characteristic areas of the casting mold, used to monitor temperature changes at different locations within the mold in real time; the sensors are connected to the central control system via wireless or wired means;

[0008] Data acquisition and processing module: This module receives data from the integrated sensor module and performs pre-processing and storage. It also has data analysis capabilities, predicting potential problems in the casting process based on historical and current data and issuing warnings when the set range is exceeded.

[0009] Central Control Module: This module is the core of the system and is used to receive data from the data acquisition and processing module. It dynamically adjusts the casting equipment and temperature intelligent control system execution modules based on preset casting process parameters and real-time data. The central control system also has remote monitoring and fault diagnosis functions, which can display various parameters of the casting process in real time and automatically diagnose and handle faults when they occur.

[0010] Execution module: includes heating and cooling devices, which are used to adjust the process parameters of the heating channel and the cooling channel according to the instructions of the central control system, and intelligently control the mold temperature.

[0011] Furthermore, the system uses data storage and tracing methods to uniquely identify all data of each casting process (including temperature, time, process parameters, etc.) and store them in the database; when necessary, the casting process data of the product can be quickly traced and queried by entering the batch number or unique identifier of the cast product, providing strong support for product quality control and problem analysis.

[0012] Furthermore, the temperature sensors of the integrated sensing module are arranged in characteristic areas of the mold based on the structural characteristics of the casting. A random forest model is established based on the temperature measurement results of the temperature sensors and the casting quality inspection results, and recursive feature elimination is performed to determine the correlation between the temperature measurement data of each temperature sensor and the casting quality, thereby optimizing the number of temperature sensors and obtaining the temperature measurement positions of the temperature sensors corresponding to various defects of the casting, thereby screening the temperature sensors.

[0013] The data acquisition and processing module analyzes the temperature data of the screened temperature sensor measurement points, cooling process parameters and corresponding casting quality, and constructs the relationship between the cooling process parameters, the starting temperature of each mold temperature sensor and the casting quality through a gradient boosting decision tree model. Based on the constructed relationship, the casting system temperature is controlled by the central control module.

[0014] Furthermore, the acquisition of the mold feature area includes:

[0015] Collect historical process data of castings and casting molds, and determine the characteristic areas of casting molds based on the shrinkage volume results of simulated castings based on X-ray inspection and historical process data;

[0016] The collected historical process data of castings and casting molds include: cooling process plan data and casting quality data;

[0017] The cooling process plan data includes cooling channel configuration, switching time and / or coolant flow rate at each production stage;

[0018] The casting quality data includes casting quality indicators, and the casting quality indicators include defect type, defect location, characteristic area yield strength and / or tensile strength.

[0019] Furthermore, a random forest model is established based on the temperature sensor measurement results and the casting quality inspection results, and recursive feature elimination is performed to determine the correlation between the temperature measurement data of each temperature sensor and the casting quality, thereby optimizing the number of temperature sensors and obtaining the temperature sensor measurement positions corresponding to various defects of the casting, thereby screening the temperature sensors, including:

[0020] A sensitivity test was conducted on the mold, and the temperature data that met the mold closing time was selected as the independent variable, and the casting quality was selected as the dependent variable for calculation. A random forest model was established and the importance of each feature was calculated. According to the feature importance, temperature sensors that reflect the casting quality were selected as standard temperature sensors in the top mold, bottom mold and side mold of the mold.

[0021] Furthermore, the random forest model is established based on the temperature sensor measurement results and the casting quality inspection results and recursive feature elimination is performed, including:

[0022] Step 1: Use the acquired features to train the random forest model and calculate the weight or coefficient of each feature of the random forest model;

[0023] Step 2: Sort the features according to their weights or coefficients;

[0024] Step 3: Delete one or more features with the smallest weight or coefficient in the ranking and retrain the random forest model with the remaining features;

[0025] Step 4: Repeat steps 2 and 3 until the desired number of features is reached or no further features can be eliminated.

[0026] Furthermore, the importance of the temperature feature of the temperature sensor is calculated, including:

[0027] Calculate the node Gini impurity;

[0028] Calculate the contribution of each feature;

[0029] Cumulative feature contribution;

[0030] Calculate the feature importance of all trees in the random forest and average the feature importance values ​​across all trees to get the final importance of the feature.

[0031] Furthermore, the selected temperature sensor measurement point temperature data, cooling process parameters, and corresponding casting quality are analyzed. The relationship between the cooling process parameters, the starting temperature of each mold temperature sensor, and the casting quality is constructed through a gradient boosting decision tree model. Based on the constructed relationship, the casting system temperature is controlled, including:

[0032] Obtaining the selected temperature sensor temperature measurement point temperature data, cooling process parameters and parameter data sets corresponding to casting quality;

[0033] While ensuring the quality of the casting, the relationship model between the initial temperature of the mold feature area and the cooling process is output based on the parameter data set to predict the performance strength of the wheel hub;

[0034] Optimize and iterate the relationship model based on the difference between the casting quality and performance strength and the actual value;

[0035] The relationship model is evaluated based on the mean square error and the coefficient of determination. When the prediction success rate is greater than the set value, the optimization iteration is stopped.

[0036] Furthermore, the cooling process plan data and casting quality data are used to train a gradient boosting decision tree model and dynamically update the mold cooling process plan and control rules;

[0037] Use the gradient boosting decision tree model to predict cooling process solutions that meet inspection standards;

[0038] The learning rate of the gradient boosting decision tree model is 0.01-0.3, and the tree depth is 3-10.

[0039] Another object of the present invention is to provide a casting system temperature intelligent control method of a casting system temperature intelligent control system, comprising the following steps:

[0040] Step 1: Determine the mold feature area and arrange the temperature sensors;

[0041] Step 2: Collect historical process data of target products and establish a database;

[0042] Step 3: Establish a temperature sensor weight coefficient model and optimize the quantity;

[0043] Step 4: Establish the cooling process, mold initial temperature and casting quality model.

[0044] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0045] First, the intelligent temperature control method for the casting system provided by the present invention improves the accuracy of temperature control and timely adjusts the temperature by setting a temperature sensor and based on the random forest model and the gradient boosting decision tree model, thereby achieving the purpose of optimizing the casting process, reducing defects and ensuring high-quality manufacturing of aluminum alloy wheels.

[0046] Due to the adoption of the above technical solution, the present invention has the following advantages compared with the existing method:

[0047] 1. Combining on-site production data with simulation, the mold's characteristic areas are determined based on X-ray inspection of wheel hub defects. Thermocouples are set in the mold's characteristic areas for temperature measurement based on the wheel hub's structural characteristics. Based on the temperature measurement results and casting quality inspection, a random forest model is established and recursive feature elimination is performed to determine the correlation between each thermocouple's temperature measurement data and the wheel hub quality. The number of thermocouples is optimized, and the thermocouple temperature measurement locations corresponding to various casting defects are found.

[0048] 2. A gradient boosting decision tree model is established based on cooling process parameters, mold temperature, casting quality, and characteristic region performance. This model determines casting quality based on the cooling process and mold temperature. Compared to manual control, this model can detect casting defects during production and then adjust them, changing cooling parameters and controlling mold temperature, achieving automation and intelligence.

[0049] 3. Predict casting quality based on mold temperature range, abandon the time delay of traditional cooling process. When defects occur in castings, on-site workers rely on experience to change cooling parameters for regulation, which cannot improve mold temperature in a short time. Control according to mold temperature range can determine the opening and closing time of cooling process according to thermocouple temperature in each casting production process, control within a reasonable temperature range, and ensure casting quality.

[0050] Second, as auxiliary evidence for the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:

[0051] (1) The expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0052] The technical solution of the present invention can significantly reduce production costs and enhance market competitiveness by optimizing the casting process and improving product quality. Specifically, by precisely controlling mold temperature and cooling parameters, casting defects can be reduced, thereby reducing scrap and rework rates. This not only directly saves material and labor costs, but also improves production efficiency and product delivery speed. Furthermore, high-quality products can win more customer trust and orders, further expanding market share. Therefore, the technical solution of the present invention has enormous commercial value and broad market prospects.

[0053] (2) The technical solution of the present invention fills the technical gap in the industry at home and abroad:

[0054] The technical solution of the present invention introduces advanced machine learning algorithms (random forest model and gradient boosting decision tree model) in the field of casting, realizing intelligent control of the casting process. This innovation not only improves the precision and efficiency of the casting process, but also provides a new solution to many problems existing in traditional casting processes. At present, the casting industry at home and abroad generally faces problems such as unstable product quality and low production efficiency, and the technical solution of the present invention is designed to address these problems. Therefore, its emergence undoubtedly fills the technical gap in the industry at home and abroad, and promotes the technological progress and development of the entire industry.

[0055] (3) The technical solution of the present invention solves the technical problems that people have been eager to solve but have never been able to solve successfully:

[0056] The technical solution of the present invention successfully solves long-standing technical problems such as the difficulty in controlling mold temperature and the untimely adjustment of the cooling process during the casting process. In traditional casting processes, due to the lack of effective real-time monitoring and control methods, it is often difficult to accurately control the mold temperature and cooling rate, resulting in unstable casting quality. However, the present invention realizes real-time control of mold temperature and cooling process by setting up thermocouples for precise temperature measurement and combining random forest models and gradient boosting decision tree models for data analysis and prediction. This intelligent control method not only improves the quality stability of castings, but also shortens the production cycle and reduces production costs.

[0057] (4) The technical solution of the present invention overcomes technical prejudice:

[0058] The technical solution of this invention breaks through some of the technical biases and limitations of traditional casting processes. For example, while traditionally believed that cooling process adjustments rely on worker experience and judgment, this invention achieves automated and intelligent control by introducing machine learning algorithms. Another example is the traditional belief that mold temperature control is difficult to precisely control in every characteristic region, while this invention achieves precise mold temperature control by setting thermocouples and establishing models. These innovations not only overcome the limitations of traditional technical biases but also open up new avenues for the development of casting processes.

[0059] Third, the lightweight, anti-occlusion UAV target tracking method of the present invention has been applied in industry to solve the problems of computational burden, occlusion effects, and insufficient real-time performance in the practical application of traditional target tracking technology in UAVs, achieving significant technological progress. Specific performance is as follows:

[0060] 1. Existing technical problems solved

[0061] Excessive computing resource requirements: Existing target tracking algorithms typically rely on complex deep neural network models (such as ResNet), which require powerful computing resources and equipment support, making real-time operation difficult for resource-limited devices such as drones. This paper significantly reduces the number of model parameters and computational complexity by replacing the SiamRPN++ backbone network with the AlexNet network. This enables drones to track targets efficiently and in real time with limited computing resources, thus solving the problem of excessive computational burden in practical applications.

[0062] Tracking failure under occlusion: Traditional target tracking methods are prone to tracking failure or mistracking when the target is occluded, making them inadequate for applications in complex environments. This invention, by introducing a weighted neighborhood index and an extended Kalman filter algorithm, automatically switches to a Kalman filter for predictive compensation when occlusion or tracking failure occurs. This effectively addresses the tracking failure issue under occlusion and enhances the method's robustness in complex environments.

[0063] Poor real-time performance: Existing algorithms are slow to process on embedded devices, making them difficult to meet the real-time requirements of drone target tracking. This invention significantly improves tracking speed and real-time performance through a lightweight network structure and simplified indicator calculations, ensuring the drone's real-time responsiveness in dynamic scenarios and resolving the real-time performance issue.

[0064] 2. Significant technological advancements

[0065] Lightweight design improves applicability: By using a lightweight AlexNet network, the computational and storage burdens are reduced, enabling resource-constrained devices such as drones to efficiently perform target tracking tasks. This design significantly improves the algorithm's applicability in environments with limited computing power, meeting industry demand for small, lightweight drones.

[0066] Enhanced occlusion tolerance: The proposed method, combining a weighted neighborhood index and an extended Kalman filter, automatically switches to Kalman filter prediction when the target is partially occluded or tracking fails, making the tracking process more robust and enhancing tracking continuity in complex scenarios. This improved occlusion tolerance enables drones to perform better in scenarios such as rescue and surveillance, reducing tracking failures caused by occlusion.

[0067] Dual Improvements in Real-Time Performance and Accuracy: This invention achieves a balance between computational efficiency and tracking accuracy. The average weighted neighborhood index and threshold determination mechanism effectively reduce unnecessary computational work, while the extended Kalman filter prediction improves tracking accuracy. This dual improvement ensures the stability of the drone in highly dynamic scenarios and enhances the real-time and accuracy of tracking, making it suitable for a variety of real-time applications.

[0068] Adaptable to complex and diverse application scenarios: Due to its anti-occlusion and lightweight design, this invention is not only suitable for drone applications but can also be expanded to other mobile devices or small robots for use in monitoring, environmental surveys, logistics and distribution, and other scenarios. In these fields, the target tracking solution of this invention can achieve robust tracking in complex and dynamic environments, meeting the industry's demand for diverse scenarios.

[0069] In summary, the lightweight anti-occlusion UAV target tracking method of the present invention has made significant progress in solving the high computing resource requirements, the large impact of occlusion problems and the lack of real-time performance. It provides reliable technical support and broader application prospects for UAV target tracking in complex and dynamic environments, and has important industrial application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 This is a structural diagram of an intelligent temperature control system for a casting system provided by an embodiment of the present invention;

[0071] Figure 2 A schematic diagram of the specific positions of the mold thermocouple temperature measurement position, cooling channel, and heating channel on the mold provided in an embodiment of the present disclosure;

[0072] Figure 3 A schematic diagram of thermocouple weight coefficients and feature importance ranking based on random forests provided in an embodiment of the present disclosure;

[0073] Figure 4A schematic diagram of a method for constructing a gradient boosting decision tree model provided in an embodiment of the present disclosure;

[0074] Figure 5 A schematic diagram of mold temperature measurement results of actual casting production after optimizing process parameters according to an embodiment of the present disclosure;

[0075] Figure 6a and Figure 6b An X-ray inspection image of the wheel core of an actual casting produced after optimizing the process parameters according to an embodiment of the present disclosure;

[0076] Figure 7 This is a flow chart of the casting system temperature intelligent control method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0078] like Figure 1 As shown, an embodiment of the present invention provides an intelligent temperature control system for a casting system, comprising:

[0079] Integrated sensing module: Consists of multiple temperature sensors, distributed according to the characteristic areas of the casting mold, used to monitor temperature changes at different locations within the mold in real time; the sensors are connected to the central control system via wireless or wired means;

[0080] Data acquisition and processing module: This module receives data from the integrated sensor module and performs pre-processing and storage. It also has data analysis capabilities, predicting potential problems in the casting process based on historical and current data and issuing warnings when the set range is exceeded.

[0081] Central Control Module: This module is the core of the system and is used to receive data from the data acquisition and processing module. It dynamically adjusts the casting equipment and temperature intelligent control system execution modules based on preset casting process parameters and real-time data. The central control system also has remote monitoring and fault diagnosis functions, which can display various parameters of the casting process in real time and automatically diagnose and handle faults when they occur.

[0082] Execution module: includes heating and cooling devices, which are used to adjust the process parameters of the heating channel and the cooling channel according to the instructions of the central control system, and intelligently control the mold temperature.

[0083] The working principle of the casting system temperature intelligent control system of the present invention is as follows:

[0084] First, the integrated sensing module utilizes multiple temperature sensors distributed across distinct mold areas to monitor the temperature at every location within the mold in real time. These temperature sensors are strategically positioned based on the mold's regional characteristics to ensure comprehensive and accurate temperature monitoring of all mold locations. The monitored temperature data is transmitted wirelessly or wired to a central control system to ensure real-time and accurate data.

[0085] Next, the data acquisition and processing module receives temperature data from the integrated sensing module, performs preprocessing, and stores it. This module not only performs basic temperature data processing but also analyzes it. By analyzing historical and current data, it can predict temperature anomalies or other potential problems during the casting process. If the temperature exceeds a preset safety range, the system automatically issues a warning signal, allowing for proactive intervention.

[0086] The central control module, the core of the system, then receives temperature information from the data acquisition and processing module and compares it with preset casting process parameters. Based on this real-time monitoring data, the central control module automatically adjusts the casting equipment's process parameters and temperature control scheme to maintain the ideal temperature range. Furthermore, the central control system features remote monitoring and fault diagnosis capabilities, displaying key casting process parameters in real time. If the system detects a fault, it automatically diagnoses and addresses it, ensuring smooth production.

[0087] Finally, the execution module, comprising heating and cooling devices, adjusts the operating parameters of the heating and cooling channels according to instructions from the central control system, thereby precisely controlling the mold temperature. This intelligent control automatically adjusts the temperature according to the requirements of different mold positions, ensuring casting quality and reducing defects caused by temperature fluctuations during the casting process, ultimately achieving intelligent and precise control of the casting process.

[0088] Furthermore, the system uses data storage and tracing methods to uniquely identify all data of each casting process (including temperature, time, process parameters, etc.) and store them in the database; when necessary, the casting process data of the product can be quickly traced and queried by entering the batch number or unique identifier of the cast product, providing strong support for product quality control and problem analysis.

[0089] Furthermore, the temperature sensors of the integrated sensing module are arranged in characteristic areas of the mold based on the structural characteristics of the casting. A random forest model is established based on the temperature measurement results of the temperature sensors and the casting quality inspection results, and recursive feature elimination is performed to determine the correlation between the temperature measurement data of each temperature sensor and the casting quality, thereby optimizing the number of temperature sensors and obtaining the temperature measurement positions of the temperature sensors corresponding to various defects of the casting, thereby screening the temperature sensors.

[0090] The data acquisition and processing module analyzes the temperature data of the screened temperature sensor measurement points, cooling process parameters and corresponding casting quality, and constructs the relationship between the cooling process parameters, the starting temperature of each mold temperature sensor and the casting quality through a gradient boosting decision tree model. Based on the constructed relationship, the casting system temperature is controlled by the central control module.

[0091] Furthermore, the acquisition of the mold feature area includes:

[0092] Collect historical process data of castings and casting molds, and determine the characteristic areas of casting molds based on the shrinkage volume results of simulated castings based on X-ray inspection and historical process data;

[0093] The collected historical process data of castings and casting molds include: cooling process plan data and casting quality data;

[0094] The cooling process plan data includes cooling channel configuration, switching time and / or coolant flow rate at each production stage;

[0095] The casting quality data includes casting quality indicators, and the casting quality indicators include defect type, defect location, characteristic area yield strength and / or tensile strength.

[0096] Furthermore, a random forest model is established based on the temperature sensor measurement results and the casting quality inspection results, and recursive feature elimination is performed to determine the correlation between the temperature measurement data of each temperature sensor and the casting quality, thereby optimizing the number of temperature sensors and obtaining the temperature sensor measurement positions corresponding to various defects of the casting, thereby screening the temperature sensors, including:

[0097] A sensitivity test was conducted on the mold, and the temperature data that met the mold closing time was selected as the independent variable, and the casting quality was selected as the dependent variable for calculation. A random forest model was established and the importance of each feature was calculated. According to the feature importance, temperature sensors that reflect the casting quality were selected as standard temperature sensors in the top mold, bottom mold and side mold of the mold.

[0098] Furthermore, the random forest model is established based on the temperature sensor measurement results and the casting quality inspection results and recursive feature elimination is performed, including:

[0099] Step 1: Use the acquired features to train the random forest model and calculate the weight or coefficient of each feature of the random forest model;

[0100] Step 2: Sort the features according to their weights or coefficients;

[0101] Step 3: Delete one or more features with the smallest weight or coefficient in the ranking and retrain the random forest model with the remaining features;

[0102] Step 4: Repeat steps 2 and 3 until the desired number of features is reached or no further features can be eliminated.

[0103] Furthermore, the importance of the temperature feature of the temperature sensor is calculated, including:

[0104] Calculate the node Gini impurity;

[0105] Calculate the contribution of each feature;

[0106] Cumulative feature contribution;

[0107] Calculate the feature importance of all trees in the random forest and average the feature importance values ​​across all trees to get the final importance of the feature.

[0108] Furthermore, the selected temperature sensor measurement point temperature data, cooling process parameters, and corresponding casting quality are analyzed. The relationship between the cooling process parameters, the starting temperature of each mold temperature sensor, and the casting quality is constructed through a gradient boosting decision tree model. Based on the constructed relationship, the casting system temperature is controlled, including:

[0109] Obtaining the selected temperature sensor temperature measurement point temperature data, cooling process parameters and parameter data sets corresponding to casting quality;

[0110] While ensuring the quality of the casting, the relationship model between the initial temperature of the mold feature area and the cooling process is output based on the parameter data set to predict the performance strength of the wheel hub;

[0111] Optimize and iterate the relationship model based on the difference between the casting quality and performance strength and the actual value;

[0112] The relationship model is evaluated based on the mean square error and the coefficient of determination. When the prediction success rate is greater than the set value, the optimization iteration is stopped.

[0113] Furthermore, the cooling process plan data and casting quality data are used to train a gradient boosting decision tree model and dynamically update the mold cooling process plan and control rules;

[0114] Use the gradient boosting decision tree model to predict cooling process solutions that meet inspection standards;

[0115] The learning rate of the gradient boosting decision tree model is 0.01-0.3, and the tree depth is 3-10.

[0116] Another object of the present invention is to provide a casting system temperature intelligent control method of a casting system temperature intelligent control system, comprising the following steps:

[0117] Step 1: Determine the mold feature area and arrange the thermocouples;

[0118] Step 2: Collect historical process data of target products and establish a database;

[0119] Step 3: Establish a thermocouple weight coefficient model and optimize the quantity;

[0120] Step 4: Establish the cooling process, mold initial temperature and casting quality model.

[0121] In a specific application scenario, take the wheel hub as an example:

[0122] Step 1: Determine the mold feature area and arrange the thermocouples.

[0123] First, historical process data for the target product and mold must be collected. This data includes numerical simulation results and trial production results. Characteristic areas of the casting mold are determined based on X-ray inspection and simulated wheel hub shrinkage volume results.

[0124] Next, a detailed database will be established based on the collected process data. This database should include the entire production process of each product, covering the following key contents:

[0125] Cooling process plan data: Records detailed information such as cooling channel configuration, switching time, coolant flow rate, etc. at each production stage.

[0126] Casting quality data: record the corresponding casting quality indicators, such as defect type, defect location, characteristic area yield strength, tensile strength, etc.

[0127] Each database record corresponds to the production process of a specific product. By systematically organizing and storing this data, it can provide a reliable foundation for subsequent process optimization and model training.

[0128] Step 2: Establish a thermocouple weight coefficient model and optimize the quantity.

[0129] Mold sensitivity testing was performed: A standard cooling process was selected as a reference. During continuous production, the first three wheels were tested using this standard cooling process to ensure the mold temperature remained normal and the casting quality was good. The next five wheels were tested using this standard cooling process with one cooling channel closed. This process was repeated until each cooling channel was tested for closure. Thermocouple data was processed, and the temperature data at the 1st and 100th seconds of mold closing time were selected as independent variables, with wheel quality as the dependent variable. A random forest model was built in Python and the importance of each feature was calculated. Based on this feature importance, a thermocouple reflecting casting quality was selected from each of the three major mold structures: the top mold, the bottom mold, and the side mold. This thermocouple served as the standard thermocouple for subsequent calculations and analysis. The process is as follows.

[0130] Step 1. Initialization: First, a model is trained using all available features and the weights or coefficients for each feature are calculated.

[0131] Step 2. Feature Sorting: The features are then sorted based on their weights or coefficients, and usually the feature with the smallest weight or coefficient is considered the least important.

[0132] Step 3: Feature Elimination: Next, remove the feature or features with the smallest weight or coefficient in the ranking and retrain the model with the remaining features.

[0133] Step 4. Iterative process: Repeat steps 2 and 3 until the required number of features is reached or no further features can be eliminated.

[0134] The importance of each thermocouple temperature characteristic is calculated as follows:

[0135] 1. Training the Random Forest Model

[0136] Constructing a random forest: A random forest is an integrated model consisting of multiple decision trees. During training, each decision tree randomly selects a subset (sample and feature subset) from the original data and then trains on these subsets.

[0137] 2. Calculate Gini Importance

[0138] Gini importance measures the contribution of each feature in reducing the impurity of the model (Gini impurity).

[0139] Calculate node Gini impurity: For each node j, calculate the Gini impurity Gj:

[0140]

[0141] in. p jk is the probability of class k in node j, where k is the number of classes.

[0142] Calculate the contribution of each feature: The contribution of feature i on node j is:

[0143]

[0144] Among them, G parent is the Gini impurity of the parent node, G left and G right are the Gini impurities of the left and right child nodes, N parent is the number of samples of the parent node, N left and N right are the number of samples of the left child node and the right child node respectively.

[0145] Cumulative feature contribution: For each tree t, the importance of feature i Ii(t) is the sum of the contributions of all nodes:

[0146]

[0147] Calculate the feature importance of all trees in the random forest: average the feature importance values ​​across all trees to get the final importance of feature i:

[0148]

[0149] Where T is the number of trees.

[0150] Step 3: Establish cooling process, mold initial temperature and casting quality model.

[0151] After eliminating the thermocouple positions, the temperature data of the selected thermocouple measurement points, the cooling process parameters, and the corresponding casting quality are analyzed. The relationship between the cooling process parameters and the starting temperature of each mold thermocouple and the casting quality is constructed through the gradient boosting decision tree model. The specific operations are as follows:

[0152] The temperature data of each thermocouple temperature measuring point is extracted and the data of the first second of mold closing is T starti , where i represents different positions, arranged in order according to the importance of thermocouple characteristics. The air flow rate of the cooling channel is 80m 3 / h, the water cooling channel is 5L / min. The cooling channel opening time is t openj , the closing time is t closej , where j represents different cooling channels. The parameter encoding for each casting is:

[0153] x=(T start1 , T peak1 ,…,T startn , T peakn , t open1 , t close1 ,…,t openm , t closem ,)

[0154] y=(q1,q2,q3,…,q n , Rm1,…,Rm n , Q1,…,Q n ,)

[0155] y is the casting quality inspection, q iIt represents whether the i-th type 3 defect occurs, represented by 0 and 1, 0 represents no defect, and 1 represents a defect. Rm1 represents the test result of the tensile strength performance of the hub in the first characteristic area, and Q1 represents the test result of the yield strength performance of the hub in the first characteristic area. The characteristic area includes the outer rim, inner rim, spokes, wheel core and rim of the hub and the finite element simulation results.

[0156] While ensuring the quality of the casting, the relationship model between the initial temperature of the mold feature area and the cooling process is output, and the performance strength of the wheel hub is predicted.

[0157] The model is optimized and iterated based on the difference between the casting quality and performance strength and the actual value: First, the maximum deviation Emax and average deviation Eavg between the target value of the casting quality inspection and the actual inspection value under the current cooling process conditions are calculated. Then, based on these deviations, the residual value is determined, and the cooling channel opening time adjustment value t is calculated by the residual. openm and closing time adjustment value t closem The cooling process plan is modified according to these adjustment values ​​to obtain new cooling process settings. This optimization process is repeated until the actual quality test value of the casting falls completely within the target range and the deviation from the target value is reduced to the preset range, and then the optimization is stopped.

[0158] The model is iterated based on the residuals and evaluated based on the mean square error and the coefficient of determination. When the prediction success rate is greater than 90%, the model can be used.

[0159] In step 1, the database used for model training contains at least one set of trial production data; at least 10 sets of cooling process plans are output through the gradient boosting decision tree model prediction for on-site production verification.

[0160] In step 1, the GBDT model is used to predict a cooling process solution that meets the inspection standards and is then verified on-site. For example, according to the ASTM E155 standard, a cooling process solution corresponding to castings with a defect level of no more than level 2 is selected for on-site production testing.

[0161] The newly acquired field verification data (including cooling process plan and casting quality) in step 1 are added to the database for training the GBDT model and dynamically updating the mold cooling process plan and control rules.

[0162] In step 1, the upper mold, lower mold and side mold in the mold must all have thermocouple measurements and be arranged at different positions.

[0163] In step 1, at least 90% of the data in the established database is used as a training set for model training, and the remaining data is used as a test set for evaluating the model prediction performance.

[0164] In step 3, the learning rate (learning_rate) of the gradient boosted decision tree (GBDT) model is between 0.01 and 0.3, and the tree depth (max_depth) is between 3 and 10.

[0165] In step three, after establishing the casting quality defect prediction model, when the accuracy of the model on the test set reaches or exceeds the set value, the model is used to predict the cooling process plan and conduct on-site production verification.

[0166] The feature importance ranking in step 2 is based on the criteria of different areas of the mold, and the thermocouples arranged at different positions of the upper mold, lower mold and side mold are ranked separately.

[0167] Specific as Figure 7 As shown, the process is as follows:

[0168] Step 101: Determine the mold feature area.

[0169] Taking the low-pressure casting mold of a certain wheel hub as an example, X-ray inspection was performed on the production wheel hubs among 100 trial-produced wheel hubs. Shrinkage cavities and porosity defects were found in the outer rim, the junction between the spokes and the outer rim, and the wheel center. The corresponding mold locations according to the wheel hub defect locations are the mold feature areas.

[0170] like Figure 2 As shown in Table 1, thermocouples were placed in the mold according to the aforementioned mold feature areas: Ten thermocouples were placed in the top mold: at the mold ejector pin opening, next to the T4 cooling channel, near the outer side of the inner rim, inside the inner rim, on the edge of the top mold, on the inner side of the top mold rim, on the inner side of the top mold rim, on the outer side of the top mold rim, on the outer side of the top mold rim, and at the inner rim ejector pin opening. The side mold consisted of four side molds, labeled 1, 2, 3, and 4. Twelve thermocouples were placed on each side mold: near the inner rim, near the outer rim, and on the rim. One thermocouple was placed on each side mold, inside and outside the rim. Six thermocouples were placed in the bottom mold: next to bottom mold B1, between B2 and B3, between B2 and B3, next to bottom mold B3, between bottom molds B4 and B5, and between bottom molds B4 and B5. 16 cooling channels and two heating channels are arranged on the mold, among which B1, B2, B3, B4 and B5 cooling channels are arranged at the bottom mold. The cooling medium is air cooling and the flow rate is set to 80m 3 / h, according to the size of the number, from B1 closest to the wheel center to B5 on the outside of the bottom mold; the top mold is arranged with T1, T2, T3, T4, T5, T6, and T7 cooling channels, the cooling medium is air cooling, and the flow rate is set to 80m 3 / h, according to the numbering, from T1 closest to the ejector pin opening to T7 near the outer side of the top mold. Each of the four side molds has a cooling channel, the cooling medium is water-cooled, and the flow rate is set at 5L / min. The numbers S1, S2, S3, and S4 correspond to the side mold numbers. The side mold cooling channel is located near the outer wheel rim of the side mold. Two heating channels, numbered H1 and H2, are located near the wheel core of the top mold. The heating resistor wire is a nickel-chromium alloy with a diameter of 0.5mm.

[0171] Table 1 Thermocouple arrangement positions

[0172] Mold cooling pipe code Cooling channel location Die casting machine temperature collection location code Mold temperature collection location B5 (Bottom mold air cooling) TC1 Ejector hole position B4 (Bottom mold air cooling) TC2 Lateral T4 B3 (Bottom mold air cooling) TC3 Top mold (inside of wheel rim) B2 (Bottom mold air cooling) TC4 Top mold (outside of wheel rim) B1 (Bottom mold air cooling) TC5 Top mold edge T1 (Top mold air cooling) TC6 On the inner side of the top mold rim T2 (Top mold air cooling) TC7 Lower inner side of top mold rim T3 (Top mold air cooling) TC8 On the inner side of the top mold rim T4 (Top mold air cooling) TC9 Lower inner side of top mold rim T5 (Top mold air cooling) TC10 Inner rim ejector port T6 (Top mold air cooling) TC11 Next to bottom mold B1 T7 (Top mold air cooling) TC12 Between bottom molds B2 and B3 S1 (Side mold water cooling) TC13 Between bottom molds B2 and B3 S2 (Side mold water cooling) TC14 Next to bottom mold B3 S3 (Side mold water cooling) TC15 Between bottom molds B4 and B5 S4 (Side mold water cooling) TC16 Between bottom molds B4 and B5 H1 (Top mold heating) TC17 1# side mold inner rim H2 (Top mold heating) TC18 1# side mold rim TC19 1# side mold outer rim TC20 2# side mold inner rim TC21 2# side mold rim TC22 2# side mold outer rim TC23 3# side mold inner rim TC24 3# side mold rim TC25 3# side mold outer rim TC26 4# side mold inner rim TC27 4# side mold rim TC28 4# side mold outer rim TC29 1# side mold outer side TC30 1# inner side of side mold

[0173] Step 102: Establish a thermocouple weight coefficient model and optimize the quantity.

[0174] A sensitivity test of the castings was conducted. A standard cooling process was first selected as a baseline. This process was applied to the first three wheels to ensure proper mold temperature and good casting quality. Subsequently, the remaining five wheels were tested, each time with one cooling channel closed. Temperature data was recorded at the 1st and 100th seconds of mold closing time. These data were used as independent variables, and wheel quality was used as the dependent variable. A random forest model in Python was used to analyze the importance of each feature. Based on the analysis, one key thermocouple was selected from each of the top, bottom, and side molds to reflect casting quality for subsequent analysis. Thermocouple temperature measurement points with an importance greater than 90% were selected for subsequent data analysis.

[0175] Figure 2 Figure 3 As shown in Table 3,

[0176] Table 2 Specific parameters of random forest model

[0177] Random Forest Model Conditions parameter Number of trees 100 to 1000 The number of features considered by each tree 'sqrt' Minimum number of samples required for internal node repartitioning 2 to 10 The number of features removed in each recursion 1 Split Quality 'gini' Minimum number of samples required for a leaf node 1-10 Metrics for evaluating model performance 'precision'

[0178] Table 3 Codes and specific locations of optimized thermocouple temperature measurement positions

[0179]

[0180] Step 103: Establish cooling process, mold initial temperature and casting quality model

[0181] like Figure 4 As shown, the construction method of the gradient boosting decision tree model is demonstrated.

[0182] Data from process simulations and actual production trials was collected to create a database. This database included cooling process data for 16 cooling channels, temperature data from 10 mold temperature measurement points, and mechanical properties of characteristic regions of the corresponding castings. Each data record covered the entire production process for a single product. Cooling process and mold temperature data were collected in 1-second increments, with cooling channel status represented by 1 (on) or 0 (off). The database contained at least 100 data records.

[0183] Based on the database established in step one, the gradient boosting decision tree (GBDT) method is used to establish a prediction model for the cooling process scheme. First, 90% of the data in the database is used for model training, and the remaining 10% of the data is used for performance evaluation. A gradient boosting decision tree model is established based on 10 thermocouple temperature measurement data, cooling process and casting quality. The initial model training settings include a learning rate of 0.1, the number of trees is 100, the maximum depth is 3, and GBDT is used for training. During the training process, we continuously adjust the learning rate, the number of trees and the maximum depth until the average prediction accuracy of the model on the test set reaches or exceeds 90%. In this embodiment, the average prediction accuracy of the final model on the test set is 90.3%. The cooling process parameters at different mold temperatures are output according to the model. The process optimization is shown in Table 4: This process was used for trial production, and no scrap was generated during the 2-hour uninterrupted trial production. Some wheel hub mold temperature data are as follows Figure 5 As shown in the figure, the temperature remains stable, and the difference between the maximum temperature and the minimum temperature at the same time does not exceed 5°C. Figure 6a and Figure 6b As shown in Table 5, X-ray inspection of the wheel core before and after process optimization shows that some shrinkage and porosity exist before optimization. A comparison of the mechanical properties of the hub's characteristic areas before and after process optimization shows improved performance at all locations, as shown in Table 5. A comparison of defects in characteristic areas of actual castings before and after the optimized process parameters are applied, as shown in Table 6, shows that defects in characteristic areas of the hub are significantly improved after the optimized process parameters are applied.

[0184] Table 4 Process optimization table

[0185]

[0186]

[0187] Table 5 Comparison of mechanical properties of the hub characteristic area before and after process optimization, performance of each position of the hub

[0188]

[0189] Table 6 Comparison of characteristic regional defects in actual production castings before and after using optimized process parameters

[0190] Characteristic area defects Initial process example Optimization process example Spoke shrinkage many Very few Rim shrinkage many Very few Wheel core air hole More none Pinholes exceed the standard few none Insufficient watering few none

[0191] The control method disclosed in this embodiment has the following advantages:

[0192] 1. Combining on-site production data with simulation, the mold's characteristic areas are determined based on X-ray inspection of wheel hub defects. Thermocouples are set in these characteristic areas for temperature measurement according to the wheel hub's structural characteristics. Based on the temperature measurement results and casting quality inspection, a random forest model is established and recursive feature elimination is performed to determine the correlation between each thermocouple's temperature measurement data and the wheel hub quality. The number of thermocouples is optimized, and the thermocouple measurement locations corresponding to various casting defects are found.

[0193] 2. A gradient boosting decision tree model is established based on cooling process parameters, mold temperature, casting quality, and characteristic region performance. This model determines casting quality based on the cooling process and mold temperature. Compared to manual control, this model can detect casting defects during production and then adjust them, changing cooling parameters and controlling mold temperature, achieving automation and intelligence.

[0194] 3. Predict casting quality based on mold temperature range, abandon the time delay of traditional cooling process. When defects occur in castings, on-site workers rely on experience to change cooling parameters for regulation, which cannot improve mold temperature in a short time. Control according to mold temperature range. In each casting production process, the opening and closing time of the cooling process can be determined according to the thermocouple temperature, and the temperature can be controlled within a reasonable range to ensure the quality of the casting.

[0195] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware portion can be implemented using dedicated logic; the software portion can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art will appreciate that the above-mentioned devices and methods can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0196] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A casting system temperature intelligent control system, characterized in that: include: Integrated sensing module: Consists of multiple temperature sensors, distributed according to the characteristic areas of the casting mold, used to monitor temperature changes at different locations within the mold in real time; the sensors are connected to the central control system via wireless or wired means; Data acquisition and processing module: This module receives data from the integrated sensor module and performs pre-processing and storage. It also has data analysis capabilities, predicting potential problems in the casting process based on historical and current data and issuing warnings when the set range is exceeded. Central Control Module: This module receives data from the Data Acquisition and Processing Module and dynamically adjusts the casting equipment and temperature intelligent control system execution modules based on preset casting process parameters and real-time data. The central control system also features remote monitoring and fault diagnosis capabilities, displaying various casting process parameters in real time and automatically diagnosing and addressing any faults. Execution module: includes heating and cooling devices, which are used to adjust the process parameters of the heating channel and the cooling channel according to the instructions of the central control system, and intelligently control the mold temperature; The temperature sensor of the integrated sensing module is arranged in the characteristic area of ​​the mold based on the structural characteristics of the casting; Based on the temperature sensor measurement results and casting quality inspection results, a random forest model is established and recursive feature elimination is performed to determine the correlation between the temperature sensor measurement data and the casting quality. This optimizes the number of temperature sensors and obtains the temperature sensor measurement positions corresponding to various casting defects, thereby screening the temperature sensors. The data acquisition and processing module analyzes the temperature data of the selected temperature sensor measuring points, the cooling process parameters, and the corresponding casting quality, and constructs the relationship between the cooling process parameters, the starting temperature of each mold temperature sensor, and the casting quality through a gradient boosting decision tree model. Based on the constructed relationship, the casting system temperature is controlled through the central control module; The acquisition of the mold feature area includes: Collect historical process data of castings and casting molds, and determine the characteristic areas of casting molds based on the shrinkage volume results of simulated castings based on X-ray inspection and historical process data; The collected historical process data of castings and casting molds include: cooling process plan data and casting quality data; The cooling process plan data includes cooling channel configuration, switching time and / or coolant flow rate at each production stage; The casting quality data includes casting quality indicators, and the casting quality indicators include defect type, defect location, characteristic area yield strength and / or tensile strength.

2. The casting system temperature intelligent control system according to claim 1, characterized in that: The system uses data storage and tracing methods to uniquely identify the temperature, time, and process parameters of each casting process and store them in a database; when necessary, by entering the batch number or unique identifier of the cast product, the casting process data of the product can be quickly traced and queried, providing strong support for product quality control and problem analysis.

3. The casting system temperature intelligent control system according to claim 1, characterized in that: The random forest model is established based on the temperature sensor measurement results and the casting quality inspection results, and recursive feature elimination is performed to determine the correlation between the temperature measurement data of each temperature sensor and the casting quality, thereby optimizing the number of temperature sensors and obtaining the temperature sensor measurement positions corresponding to various defects of the casting, thereby screening the temperature sensors, including: A sensitivity test was conducted on the mold, and the temperature data that met the mold closing time was selected as the independent variable, and the casting quality was selected as the dependent variable for calculation. A random forest model was established and the importance of each feature was calculated. According to the feature importance, temperature sensors that reflect the casting quality were selected as standard temperature sensors in the top mold, bottom mold and side mold of the mold.

4. The casting system temperature intelligent control system according to claim 1, characterized in that: The method of establishing a random forest model based on the temperature sensor measurement results and the casting quality inspection results and performing recursive feature elimination includes: Step 1: Use the acquired features to train the random forest model and calculate the weight or coefficient of each feature of the random forest model; Step 2: Sort the features according to their weights or coefficients; Step 3: Delete one or more features with the smallest weight or coefficient in the ranking and retrain the random forest model with the remaining features; Step 4: Repeat steps 2 and 3 until the desired number of features is reached or no further features can be eliminated.

5. The casting system temperature intelligent control system according to claim 1, characterized in that: Calculation of the importance of temperature characteristics of temperature sensors, including: Calculate the node Gini impurity; Calculate the contribution of each feature; Cumulative feature contribution; Calculate the feature importance of all trees in the random forest and average the feature importance values ​​across all trees to get the final importance of the feature.

6. The casting system temperature intelligent control system according to claim 1, characterized in that: The selected temperature sensor measurement point temperature data, cooling process parameters, and corresponding casting quality are analyzed. The relationship between cooling process parameters, the starting temperature of each mold temperature sensor, and casting quality is constructed using a gradient boosting decision tree model. Based on this relationship, the casting system temperature is controlled, including: Obtaining the selected temperature sensor temperature measurement point temperature data, cooling process parameters and parameter data sets corresponding to casting quality; While ensuring the quality of the casting, the relationship model between the initial temperature of the mold feature area and the cooling process is output based on the parameter data set to predict the performance strength of the wheel hub; Optimize and iterate the relationship model based on the difference between the casting quality and performance strength and the actual value; The relationship model is evaluated based on the mean square error and the coefficient of determination. When the prediction success rate is greater than the set value, the optimization iteration is stopped.

7. The casting system temperature intelligent control system according to claim 1, characterized in that: The cooling process plan data and casting quality data are used to train the gradient boosting decision tree model and dynamically update the mold cooling process plan and control rules; Use the gradient boosting decision tree model to predict cooling process solutions that meet inspection standards; The learning rate of the gradient boosting decision tree model is 0.01-0.3, and the tree depth is 3-10.

8. A casting system temperature intelligent control method of the casting system temperature intelligent control system according to any one of claims 1 to 7, characterized in that: The following steps are involved: Step 1: Determine the mold feature area and arrange the temperature sensors; Step 2: Collect historical process data of target products and establish a database; Step 3: Establish a temperature sensor weight coefficient model and optimize the quantity; Step 4: Establish the cooling process, mold initial temperature and casting quality model.

Images