Metal casting cooling control method and system

Abstract

The invention discloses a metal casting cooling control method and system, relates to the technical field of data processing, and solves the problem that firstly, it is difficult to combine metal casting data with thermal imaging data to construct a three-dimensional temperature field model; secondly, the temperature distribution of the surface and the interior of the metal casting is difficult to analyze, and the cooling uniformity of the metal casting is difficult to analyze; then, a cooling medium is difficult to automatically adjust by using a PID control algorithm; and finally, defect detection is difficult to carry out on the metal casting. According to the invention, the cooling process of the metal casting is monitored in real time by using a sensor and a thermal imaging technology, and a three-dimensional temperature field model is constructed to analyze temperature distribution and evaluate the cooling effect; and adjusting cooling medium parameters through a PID control algorithm, and carrying out defect detection and visualization on the metal casting.

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and in particular relates to a metal casting cooling control method and system. Background Art

[0002] Metal castings are formed metal objects obtained through the casting process and are widely used in a variety of fields, including machinery manufacturing, automotive, shipbuilding, chemical engineering, and mining. With the advancement of information technology, sensor technology and thermal imaging are being used to monitor the cooling process of metal castings in real time. Based on the monitored data, a three-dimensional temperature field model of the metal casting is constructed to analyze the temperature distribution of the metal casting. Furthermore, a PID control algorithm is used to automatically adjust the cooling medium to ensure precise control of the cooling process. Furthermore, deep learning models are being used to detect and visualize defects in castings, enabling intelligent optimization and quality improvement of the casting process.

[0003] The existing technology has the following problems: first, it is difficult to combine metal casting data and thermal imaging data to construct a three-dimensional temperature field model; second, it is difficult to analyze the temperature distribution on the surface and inside of the metal casting and difficult to analyze the cooling uniformity of the metal casting; then, it is difficult to use the PID control algorithm to automatically adjust the cooling medium; finally, it is difficult to detect defects in metal castings. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a metal casting cooling control method and system for solving the above-mentioned problem. To this end, a first aspect of the present invention provides a metal casting cooling control method, comprising the following steps: S1: Use sensors to collect real-time information data including: metal casting data, cooling data and ambient temperature; use thermal imaging cameras to collect thermal imaging data of the casting surface; S2: Preprocessing of real-time information data by filtering, denoising, data cleaning and standardization; Preprocessing of thermal imaging data by filtering, denoising, calibration, image enhancement and grayscale conversion; S3: Construct a three-dimensional temperature field model of the metal casting based on real-time information data and thermal imaging data; analyze the temperature distribution on the surface and inside of the metal casting based on the constructed three-dimensional temperature field model and calculate the temperature gradient, cooling rate, and cooling uniformity; comprehensively evaluate the cooling effect index of the metal casting based on the calculated results of temperature gradient, cooling rate, and cooling uniformity, and calculate the cooling uniformity deviation of the metal casting; S4: Use PID control algorithm to automatically adjust the flow, pressure and temperature of the cooling medium; S5: Defect detection and visualization of metal castings based on deep learning models.

[0005] Furthermore, the real-time information data collected by sensors in step S1 includes: metal casting data, cooling data and ambient temperature, including the following steps: wherein the metal casting data includes: density, thermal conductivity, specific heat capacity, mass, volume, shape, thickness, length, width and temperature of the metal casting; the cooling data includes: type, flow, pressure, temperature and cooling time of the cooling medium.

[0006] Furthermore, in step S3, constructing a three-dimensional temperature field model of the metal casting based on the real-time information data and the thermal imaging data includes the following steps: Based on the collected and pre-processed real-time information data, a 3D geometric model of the metal casting is constructed, including: the external contour and internal structure of the metal casting; the constructed 3D geometric model is divided into finite element meshes; Setting the physical parameters of the three-dimensional geometric model includes: inputting the thermal conductivity, specific heat capacity, and density of the metal casting; setting the boundary conditions of the three-dimensional geometric model includes: setting a temperature boundary condition on the mold surface in contact with the cooling medium; setting the outer boundary of the mold as a convection heat transfer boundary condition, setting the convection heat transfer coefficient and ambient temperature on the outer surface of the mold; setting the radiation heat transfer boundary condition, setting the radiation coefficient and ambient temperature on the casting surface; setting the inner boundaries of the casting and the mold as third-class boundary conditions, and inputting the interface heat transfer coefficient; Based on Fourier's law of heat conduction, a heat conduction equation for metal castings is established; the temperature field of metal castings is numerically simulated using the finite element analysis method; by solving the heat conduction equation, the temperature distribution of the metal casting during the cooling process is calculated to obtain a three-dimensional temperature field model; By setting marker points on the casting surface, the thermal imaging data is spatially aligned with the 3D temperature field model; the timestamp of the thermal imaging data is synchronized with the time step of the numerical simulation; and image processing technology is used to match the thermal imaging data with the surface mesh of the 3D temperature field model. The thermal imaging data are smoothly integrated into the three-dimensional temperature field model using interpolation algorithm and color coding technology, and the surface temperature values in the thermal imaging data are mapped to the surface grid nodes of the three-dimensional temperature field model.

[0007] Furthermore, the step S3 analyzes the temperature distribution on the surface and inside of the metal casting according to the constructed three-dimensional temperature field model and calculates the temperature gradient, cooling rate and cooling uniformity respectively, including the following steps: According to the constructed three-dimensional temperature field model, the surface temperature distribution of the metal casting is obtained as follows: ,in, represents the coordinate, t is the cooling time; through reverse heat conduction analysis, the internal temperature distribution of the metal casting is inferred from the surface temperature: ; Formula for calculating the temperature gradient on the surface of metal castings: Formula for calculating the cooling rate of metal casting surface: Formula for calculating cooling uniformity on the surface of metal castings: The temperature gradient on the surface of the metal casting is obtained respectively , cooling rate and cooling uniformity ;in, represents the surface temperature on the i-th network node; n is the total number of network nodes on the surface of the metal casting; Indicates the average value of the surface temperature; Similarly, the temperature gradient formula, cooling rate formula and cooling uniformity formula on the surface of the metal casting are used to calculate the temperature gradient inside the casting. , cooling rate , cooling uniformity .

[0008] Furthermore, the step S3 comprehensively evaluates the cooling effect index of the metal casting and calculates the cooling uniformity deviation of the metal casting based on the calculation results of the temperature gradient, the cooling rate, and the cooling uniformity, including the following steps: The formula for calculating and evaluating the cooling effect index of the metal casting surface is: Get the cooling effect index of the metal casting surface ;in, It is a mold for the surface temperature gradient of metal castings; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes on the surface of metal casting respectively; and They represent the maximum and minimum cooling rates at n network nodes on the surface of the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity on the surface of metal castings; 、 and are weight coefficients respectively, and the sum is 1; The formula for calculating and evaluating the cooling effect index inside metal castings is: Get the cooling effect index inside the metal casting ;in, It is a mold for the temperature gradient inside the metal casting; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes inside the metal casting; and They represent the maximum and minimum cooling rates at n network nodes inside the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity inside the metal casting; 、 and are weight coefficients respectively, and the sum is 1; The cooling effect index of the metal casting surface is compared with the cooling effect index of the metal casting interior to obtain the cooling uniformity deviation of the metal casting; when the cooling uniformity deviation is higher than the preset threshold When it is 0.15, it means that the metal casting is cooling unevenly.

[0009] Furthermore, the step S4 includes the following steps: When metal castings cool unevenly, a PID control algorithm is used to automatically adjust the flow, pressure, and temperature of the cooling medium. The control formula of the PID algorithm is: in, is the output of the controller at time t, is the error, 、 and are proportional, integral, and differential coefficients respectively; Set the target flow rate, target pressure and target temperature of the cooling medium, and monitor the actual flow rate, actual pressure value and actual temperature of the cooling medium in real time; calculate the control signal through the PID algorithm based on the difference between the target flow rate and the actual flow rate, and adjust the actual flow rate, actual pressure value and actual temperature of the cooling medium to the set target flow rate, target pressure and target temperature of the cooling medium; During the cooling process, the target flow rate, target pressure and target temperature of the cooling medium are dynamically adjusted according to the cooling effect index on the surface and inside of the metal casting.

[0010] Furthermore, the step S5 includes the following steps: Collecting thermal imaging data and three-dimensional temperature field model data corresponding to metal castings containing various defect types, including pores, shrinkage, and cracks; and preprocessing the collected thermal imaging data and three-dimensional temperature field model data; Label the presence of defects in metal castings and the type and location of the defects; use the labeled information as training labels for supervised learning to train deep learning models; Generate a data set from labeled thermal imaging data and 3D temperature field model data and divide it into a training set with a ratio of 70% for training, a validation set with a ratio of 15% for validation, and a test set with a ratio of 15% for testing; Use the convolutional neural network model and the recurrent neural network model in the deep learning model for combined training; input the labeled training set into the deep learning model that combines the convolutional neural network model and the recurrent neural network model for training; use the validation set and test set to verify and test the deep learning model respectively; When the deep learning model training is completed, the real-time collected and pre-processed thermal imaging data and three-dimensional temperature field model data of the metal casting are input into the trained deep learning model for defect detection, and the detected defects are annotated and visualized in the three-dimensional temperature field model.

[0011] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses sensors and thermal imaging technology to monitor the cooling process of metal castings in real time. Real-time information data and thermal imaging data are combined to construct a three-dimensional temperature field model of the metal casting. This model is used to analyze the temperature distribution on the surface and inside of the metal casting, and further obtain the temperature gradient, cooling rate, and cooling uniformity. This allows for a comprehensive assessment of the cooling effect of the metal casting. The present invention automatically adjusts the flow rate, pressure and temperature of the cooling medium by adopting a PID control algorithm, thereby achieving precise control of the cooling process and improving cooling efficiency. The present invention uses a deep learning model combined with thermal imaging data and three-dimensional temperature field model data to perform defect detection and visualization, making defect detection of metal castings more intuitive and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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 paying any creative work.

[0013] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0014] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0015] See also Figure 1 As shown, the first embodiment of the present invention provides a metal casting cooling control method, comprising the following steps: S1: Use sensors to collect real-time information data including: metal casting data, cooling data and ambient temperature; use thermal imaging cameras to collect thermal imaging data of the casting surface; S2: Preprocessing of real-time information data by filtering, denoising, data cleaning and standardization; Preprocessing of thermal imaging data by filtering, denoising, calibration, image enhancement and grayscale conversion; S3: Construct a three-dimensional temperature field model of the metal casting based on real-time information data and thermal imaging data; analyze the temperature distribution on the surface and inside of the metal casting based on the constructed three-dimensional temperature field model and calculate the temperature gradient, cooling rate, and cooling uniformity; comprehensively evaluate the cooling effect index of the metal casting based on the calculated results of temperature gradient, cooling rate, and cooling uniformity, and calculate the cooling uniformity deviation of the metal casting; S4: Use PID control algorithm to automatically adjust the flow, pressure and temperature of the cooling medium; S5: Defect detection and visualization of metal castings based on deep learning models.

[0016] Specifically, during the cooling process of metal castings, a variety of sensors are installed, including temperature sensors, flow sensors, pressure sensors, etc., as well as high-resolution thermal imaging cameras, to achieve real-time data collection of metal casting temperature, cooling medium parameters, ambient temperature and thermal imaging data. After the collected data undergoes pre-processing steps such as filtering and denoising, data cleaning, standardization, and calibration, enhancement and grayscale of thermal imaging data, a three-dimensional temperature field model of the metal casting is constructed in combination with three-dimensional modeling software, and the temperature distribution, temperature gradient, cooling rate and cooling uniformity are analyzed to comprehensively evaluate the cooling effect. At the same time, the PID controller parameters are configured, and the PID control algorithm is used to automatically adjust the flow, pressure and temperature of the cooling medium according to real-time data to ensure the ideal cooling effect. In addition, by collecting a large number of samples to train a deep learning model, defect detection of metal castings is achieved, and the defect information is visualized.

[0017] In this embodiment, the real-time information data collected by sensors in step S1 includes: metal casting data, cooling data and ambient temperature, including the following steps: wherein the metal casting data includes: density, thermal conductivity, specific heat capacity, mass, volume, shape, thickness, length, width and temperature of the metal casting; cooling data includes: type, flow, pressure, temperature and cooling time of the cooling medium.

[0018] Specifically, during the cooling process, sensors are used to collect real-time information data of metal castings, and metal casting data, including but not limited to: density, thermal conductivity, specific heat capacity, mass, volume, shape, thickness, length, width and temperature, as well as cooling data of the cooling process, including but not limited to: type, flow, pressure, temperature and cooling time of the cooling medium, are comprehensively monitored; the types of the cooling medium include but are not limited to: water, air, cooling oil and refrigerant; and temperature sensors are used to collect real-time ambient temperature around the metal castings.

[0019] In this embodiment, the three-dimensional temperature field model of the metal casting is constructed based on the real-time information data and the thermal imaging data in step S3, including the following steps: Based on the collected and pre-processed real-time information data, a three-dimensional geometric model of the metal casting is constructed, including: the external contour and internal structure of the metal casting; the constructed three-dimensional geometric model is divided into a finite element mesh; Setting the physical parameters of the three-dimensional geometric model includes: inputting the thermal conductivity, specific heat capacity, and density of the metal casting; setting the boundary conditions of the three-dimensional geometric model includes: setting a temperature boundary condition on the mold surface in contact with the cooling medium; setting the outer boundary of the mold as a convection heat transfer boundary condition, setting the convection heat transfer coefficient and ambient temperature on the outer surface of the mold; setting the radiation heat transfer boundary condition, setting the radiation coefficient and ambient temperature on the casting surface; setting the inner boundaries of the casting and the mold as third-class boundary conditions, and inputting the interface heat transfer coefficient; Based on Fourier's law of heat conduction, a heat conduction equation for metal castings is established; the temperature field of metal castings is numerically simulated using the finite element analysis method; by solving the heat conduction equation, the temperature distribution of the metal casting during the cooling process is calculated to obtain a three-dimensional temperature field model; By setting marker points on the casting surface, the thermal imaging data is spatially aligned with the 3D temperature field model; the timestamp of the thermal imaging data is synchronized with the time step of the numerical simulation; and the thermal imaging data is matched with the surface mesh of the 3D temperature field model using image processing technology. The thermal imaging data are smoothly integrated into the three-dimensional temperature field model using interpolation algorithm and color coding technology, and the surface temperature values in the thermal imaging data are mapped to the surface grid nodes of the three-dimensional temperature field model.

[0020] Specifically, a 3D geometric model of the metal casting is constructed using CAD software or specialized geometric modeling tools based on preprocessed real-time data, such as the metal casting's density, mass, volume, shape, and dimensions. This model should accurately reflect the casting's external contours and internal structure. Finite element analysis software, such as Ansys or Abaqus, is used to divide the constructed 3D geometric model into a finite element mesh. The meshing should take into account the casting's geometric characteristics, material properties, and expected temperature gradient distribution to ensure the accuracy and efficiency of the numerical simulation. The casting's thermal conductivity, specific heat capacity, and density are input; these parameters are used to establish and solve the heat conduction equation. Temperature boundary conditions are set on the mold surfaces that contact the coolant, typically based on the coolant temperature and cooling time. Convection heat transfer boundary conditions are set on the mold's outer boundary, and the convection heat transfer coefficient and ambient temperature are input. The convection heat transfer coefficient depends on factors such as the mold surface roughness, the flow rate, and temperature of the coolant. The emissivity coefficient and ambient temperature are set on the casting surface to account for radiative heat exchange between the casting and the surrounding environment. Third-class boundary conditions were set for the inner boundaries of the casting and the mold, and the interface heat transfer coefficient was input to account for heat conduction between the casting and the mold. Based on Fourier's law of heat conduction, a heat conduction equation for the metal casting was established. This equation describes the temporal and spatial variations of the internal temperature of the metal casting. Finite element analysis was used to numerically simulate the temperature field of the metal casting. By solving the heat conduction equation, the temperature distribution of the metal casting during cooling was calculated. The numerical simulation yielded temperature distribution data for the cooling process and used it to construct a 3D temperature field model. Using visualization tools such as the Ansys Workbench Postprocessing Toolbox and ParaView, the temperature distribution data was mapped onto the finite element mesh of the 3D geometric model, thereby constructing a 3D temperature field model of the metal casting. Marker points were set on the casting surface to spatially align the thermal imaging data with the 3D temperature field model. The timestamps of the thermal imaging data were synchronized with the time step of the numerical simulation to ensure data synchronization. Image processing techniques such as image registration and image fusion were used to match the thermal imaging data to the surface mesh of the 3D temperature field model. Using interpolation algorithms and color coding techniques, the thermal imaging data is smoothly integrated into the 3D temperature field model, and the surface temperature values in the thermal imaging data are mapped to the surface mesh nodes of the 3D temperature field model. A 3D temperature field model of the metal casting is constructed based on real-time information data and thermal imaging data, accurately reflecting the temperature distribution and changes of the metal casting during the cooling process.

[0021] In this embodiment, step S3 analyzes the temperature distribution on the surface and inside of the metal casting according to the constructed three-dimensional temperature field model and calculates the temperature gradient, cooling rate and cooling uniformity respectively, including the following steps: According to the constructed three-dimensional temperature field model, the surface temperature distribution of the metal casting is obtained as follows: ,in, represents the coordinate, t is the cooling time; through reverse heat conduction analysis, the internal temperature distribution of the metal casting is inferred from the surface temperature: ; Formula for calculating the temperature gradient on the surface of metal castings: Formula for calculating the cooling rate of metal casting surface: Formula for calculating cooling uniformity on the surface of metal castings: The temperature gradient on the surface of the metal casting is obtained respectively , cooling rate and cooling uniformity ;in, represents the surface temperature on the i-th network node; n is the total number of network nodes on the surface of the metal casting; Indicates the average value of the surface temperature; Similarly, the temperature gradient formula, cooling rate formula and cooling uniformity formula on the surface of the metal casting are used to calculate the temperature gradient inside the casting. , cooling rate , cooling uniformity .

[0022] Specifically, the constructed three-dimensional temperature field model is used to obtain the temperature values of each node on the surface and inside the metal casting, as well as the corresponding time points. Based on the obtained temperature values, the temperature gradients at each node on the surface and inside the metal casting are calculated using numerical differentiation methods. Based on the obtained temperature values at different time points, the cooling rate of each node on the surface and inside the metal casting is calculated using the cooling rate formula. The cooling uniformity of the surface and inside the metal casting is calculated using the cooling uniformity formula described above.

[0023] In this embodiment, the step S3 comprehensively evaluates the cooling effect index of the metal casting and calculates the cooling uniformity deviation of the metal casting based on the calculation results of the temperature gradient, the cooling rate, and the cooling uniformity, including the following steps: The formula for calculating and evaluating the cooling effect index of the metal casting surface is: Get the cooling effect index of the metal casting surface ;in, It is a mold for the surface temperature gradient of metal castings; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes on the surface of metal casting respectively; and They represent the maximum and minimum cooling rates at n network nodes on the surface of the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity on the surface of metal castings; 、 and are weight coefficients respectively, and the sum is 1; The formula for calculating and evaluating the cooling effect index inside metal castings is: Get the cooling effect index inside the metal casting ;in, It is a mold for the temperature gradient inside the metal casting; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes inside the metal casting; and They represent the maximum and minimum cooling rates at n network nodes inside the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity inside the metal casting; 、 and are weight coefficients respectively, and the sum is 1; The cooling effect index of the metal casting surface is compared with the cooling effect index of the metal casting interior to obtain the cooling uniformity deviation of the metal casting; when the cooling uniformity deviation is higher than the preset threshold When it is 0.15, it means that the metal casting is cooling unevenly.

[0024] Specifically, based on the constructed three-dimensional temperature field model, the temperature values, temperature gradients, cooling rates, and cooling uniformity of each node on the surface and inside of the metal casting are obtained. Based on the obtained data, the maximum and minimum values of the model of the temperature gradient on the surface and inside of the metal casting, the maximum and minimum values of the cooling rate, and the maximum and minimum values of the cooling uniformity are calculated. Based on the formula and the calculated parameter values and weight coefficients, the cooling effect index of the surface and inside of the metal casting is calculated respectively; among them, the weight coefficient when calculating the cooling effect index of the metal casting surface is 、 and 0.4, 0.3 and 0.3 respectively; weight coefficients when calculating the cooling effect index inside metal castings 、 and The weight coefficients are 0.3, 0.35 and 0.35 respectively. The weight coefficients are automatically adjusted based on actual production experience and process requirements or through experimental data and mathematical optimization methods. The cooling uniformity deviation of the metal casting is calculated based on the formula and the cooling effect index value of the surface and interior of the metal casting. If the cooling uniformity deviation is higher than the preset threshold When the threshold is 0.15, it indicates that the metal casting is cooled unevenly. Otherwise, it indicates that the metal casting is cooled relatively evenly. The preset threshold should be adjusted and determined based on the actual production requirements of metal castings, experimental data and industry experience.

[0025] In this embodiment, step S4 includes the following steps: When metal castings cool unevenly, a PID control algorithm is used to automatically adjust the flow, pressure, and temperature of the cooling medium. The control formula of the PID algorithm is: in, is the output of the controller at time t, is the error, 、 and are proportional, integral, and differential coefficients respectively; Set the target flow rate, target pressure and target temperature of the cooling medium, and monitor the actual flow rate, actual pressure value and actual temperature of the cooling medium in real time; calculate the control signal through the PID algorithm based on the difference between the target flow rate and the actual flow rate, and adjust the actual flow rate, actual pressure value and actual temperature of the cooling medium to the set target flow rate, target pressure and target temperature of the cooling medium; During the cooling process, the target flow rate, target pressure and target temperature of the cooling medium are dynamically adjusted according to the cooling effect index on the surface and inside of the metal casting.

[0026] Specifically, when uneven cooling occurs on a metal casting, the PID control algorithm's control formula automatically adjusts the cooling medium. Here, the error, e(t), represents the difference between the set value and the actual value. The proportional, integral, and differential coefficients determine the PID controller's response speed and stability to the error. The target flow rate, target pressure, and target temperature of the cooling medium are set based on the material, size, and cooling requirements of the metal casting. Sensors are used to monitor the actual flow rate, pressure, and temperature of the cooling medium in real time. The error, e(t), is calculated based on the set value and the actual monitored value. This error, e(t), is substituted into the PID control formula to generate a control signal. The magnitude and direction of this control signal depend on the error, e(t), and the values of Kp, Ki, and Kd. Based on the control signal, the actual flow rate, pressure, and temperature of the cooling medium are adjusted to gradually approach the set value. During the cooling process, the cooling effect of the metal casting is evaluated based on the cooling effect index on the surface and interior of the metal casting, and the set target flow rate, pressure, and temperature of the cooling medium are dynamically adjusted accordingly. If a part of the metal casting cools too quickly, the cooling medium flow or temperature in that area can be appropriately reduced; if a part of the metal casting cools too slowly, the flow rate or temperature can be appropriately increased.

[0027] In this embodiment, step S5 includes the following steps: Collecting thermal imaging data and three-dimensional temperature field model data corresponding to metal castings containing various defect types, including pores, shrinkage, and cracks; and preprocessing the collected thermal imaging data and three-dimensional temperature field model data; Label the presence of defects in metal castings and the type and location of the defects; use the labeled information as training labels for supervised learning to train deep learning models; Generate a data set from labeled thermal imaging data and 3D temperature field model data and divide it into a training set with a ratio of 70% for training, a validation set with a ratio of 15% for validation, and a test set with a ratio of 15% for testing; Use the convolutional neural network model and the recurrent neural network model in the deep learning model for combined training; input the labeled training set into the deep learning model that combines the convolutional neural network model and the recurrent neural network model for training; use the validation set and test set to verify and test the deep learning model respectively; When the deep learning model training is completed, the real-time collected and pre-processed thermal imaging data and three-dimensional temperature field model data of the metal casting are input into the trained deep learning model for defect detection, and the detected defects are annotated and visualized in the three-dimensional temperature field model.

[0028] Specifically, thermal imaging data and three-dimensional temperature field model data are collected for metal castings containing various defect types, including pores, shrinkage, and cracks. The three-dimensional temperature field model data includes, but is not limited to, temperature distribution, temperature gradient, thermal conductivity, and cooling time. The data covers castings of different casting processes, materials, and sizes to increase data diversity and generalization. The thermal imaging data undergoes noise removal and image enhancement to improve image quality. The three-dimensional temperature field model data undergoes smoothing to reduce the impact of measurement errors and noise. Professional inspectors manually annotate the thermal imaging data and the three-dimensional temperature field model data of the metal castings, labeling defect types and locations. Defect types include, but are not limited to, pores, shrinkage, and cracks. The accuracy of the annotated defect types and locations is ensured, providing reliable training labels for supervised learning. The annotated thermal imaging data and three-dimensional temperature field model data are divided into training, validation, and test sets in a ratio of 70%, 15%, and 15%, respectively. The data distribution of the training, validation, and test sets is ensured to be consistent to avoid data leakage and overfitting. A convolutional neural network model is selected to extract image features from the thermal imaging data. A recurrent neural network model was selected to capture the time series features in the 3D temperature field model data. A convolutional neural network model and a recurrent neural network model were combined to form a deep learning model capable of processing both image and time series data. The labeled training set was fed into the combined deep learning model for training. Appropriate loss functions and optimization algorithms, such as the cross-entropy loss function and the Adam optimization algorithm, were set. During training, metrics such as loss and accuracy were monitored, and model parameters and learning rates were adjusted accordingly. The deep learning model was validated using a validation set to evaluate its generalization and performance. Based on the validation results, the model was fine-tuned to improve its accuracy and robustness. The deep learning model was tested using a test set to evaluate its final performance. During the actual inspection process, thermal imaging data and 3D temperature field model data of metal castings were collected and preprocessed in real time. This preprocessed real-time data was then fed into the trained deep learning model for defect detection. The deep learning model output information such as defect type, location, and confidence level. The detected defects are marked in the 3D temperature field model, and different colors or shapes are used to represent different types of defects, so that users can intuitively view and analyze the defect detection results and realize the visualization of defects.

[0029] See also Figure 2 As shown, the present invention is a metal casting cooling control system, comprising the following modules: Data acquisition module: uses sensors to collect real-time information data including: metal casting data, cooling data and ambient temperature; uses thermal imaging cameras to collect thermal imaging data of the casting surface; Data preprocessing module: performs filtering, denoising, data cleaning and standardization preprocessing on the collected real-time information data; performs filtering, denoising, calibration, image enhancement and grayscale preprocessing on the collected thermal imaging data; Model building and data analysis module: Builds a three-dimensional temperature field model of the metal casting based on real-time information data and thermal imaging data; analyzes the temperature distribution on the surface and inside the metal casting based on the constructed three-dimensional temperature field model and calculates the temperature gradient, cooling rate, and cooling uniformity; comprehensively evaluates the cooling effect index of the metal casting based on the calculated results of temperature gradient, cooling rate, and cooling uniformity, and calculates the cooling uniformity deviation of the metal casting; Cooling control module: uses PID control algorithm to automatically adjust the flow, pressure and temperature of the cooling medium; Defect Detection Module: Detects and visualizes defects in metal castings based on deep learning models.

[0030] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for controlling cooling of metal castings, characterized in that: The following steps are involved: S1: Use sensors to collect real-time information data including: metal casting data, cooling data and ambient temperature; use thermal imaging cameras to collect thermal imaging data of the casting surface; S2: Preprocessing of real-time information data by filtering, denoising, data cleaning and standardization; Preprocessing of thermal imaging data by filtering, denoising, calibration, image enhancement and grayscale conversion; S3: Construct a three-dimensional temperature field model of the metal casting based on real-time information data and thermal imaging data; analyze the temperature distribution on the surface and inside of the metal casting based on the constructed three-dimensional temperature field model and calculate the temperature gradient, cooling rate, and cooling uniformity; comprehensively evaluate the cooling effect index of the metal casting based on the calculated results of temperature gradient, cooling rate, and cooling uniformity, and calculate the cooling uniformity deviation of the metal casting; S4: Use PID control algorithm to automatically adjust the flow, pressure and temperature of the cooling medium; S5: Defect detection and visualization of metal castings based on deep learning models.

2. A metal casting cooling control method according to claim 1, characterized in that: The real-time information data collected by sensors in step S1 include: metal casting data, cooling data and ambient temperature, including the following steps: wherein the metal casting data includes: density, thermal conductivity, specific heat capacity, mass, volume, shape, thickness, length, width and temperature of the metal casting; cooling data includes: type, flow rate, pressure, temperature and cooling time of the cooling medium.

3. A metal casting cooling control method according to claim 1, characterized in that: The step S3 constructs a three-dimensional temperature field model of the metal casting based on the real-time information data and the thermal imaging data, including the following steps: Based on the collected and pre-processed real-time information data, a 3D geometric model of the metal casting is constructed, including: the external contour and internal structure of the metal casting; the constructed 3D geometric model is divided into finite element meshes; Setting the physical parameters of the three-dimensional geometric model includes: inputting the thermal conductivity, specific heat capacity, and density of the metal casting; setting the boundary conditions of the three-dimensional geometric model includes: setting a temperature boundary condition on the mold surface in contact with the cooling medium; setting the outer boundary of the mold as a convection heat transfer boundary condition, setting the convection heat transfer coefficient and ambient temperature on the outer surface of the mold; setting the radiation heat transfer boundary condition, setting the radiation coefficient and ambient temperature on the casting surface; setting the inner boundaries of the casting and the mold as third-class boundary conditions, and inputting the interface heat transfer coefficient; Based on Fourier's law of heat conduction, a heat conduction equation for metal castings is established; the temperature field of metal castings is numerically simulated using the finite element analysis method; by solving the heat conduction equation, the temperature distribution of the metal casting during the cooling process is calculated to obtain a three-dimensional temperature field model; By setting marker points on the casting surface, the thermal imaging data is spatially aligned with the 3D temperature field model; the timestamp of the thermal imaging data is synchronized with the time step of the numerical simulation; and image processing technology is used to match the thermal imaging data with the surface mesh of the 3D temperature field model. The thermal imaging data are smoothly integrated into the three-dimensional temperature field model using interpolation algorithm and color coding technology, and the surface temperature values in the thermal imaging data are mapped to the surface grid nodes of the three-dimensional temperature field model.

4. A metal casting cooling control method according to claim 3, characterized in that: The step S3 analyzes the temperature distribution on the surface and inside of the metal casting according to the constructed three-dimensional temperature field model and calculates the temperature gradient, cooling rate and cooling uniformity respectively, including the following steps: According to the constructed three-dimensional temperature field model, the surface temperature distribution of the metal casting is obtained as follows: ,in, represents the coordinate, t is the cooling time; through reverse heat conduction analysis, the internal temperature distribution of the metal casting is inferred from the surface temperature: ; Formula for calculating the temperature gradient on the surface of metal castings: Formula for calculating the cooling rate of metal casting surface: Formula for calculating cooling uniformity on the surface of metal castings: The temperature gradient on the surface of the metal casting is obtained respectively , cooling rate and cooling uniformity ;in, represents the surface temperature on the i-th network node; n is the total number of network nodes on the surface of the metal casting; Indicates the average value of the surface temperature; Similarly, the temperature gradient formula, cooling rate formula and cooling uniformity formula on the surface of the metal casting are used to calculate the temperature gradient inside the casting. , cooling rate , cooling uniformity .

5. A metal casting cooling control method according to claim 4, characterized in that: The step S3 comprehensively evaluates the cooling effect index of the metal casting and calculates the cooling uniformity deviation of the metal casting based on the calculation results of the temperature gradient, the cooling rate and the cooling uniformity, including the following steps: The formula for calculating and evaluating the cooling effect index of the metal casting surface is: Get the cooling effect index of the metal casting surface ;in, It is a mold for the surface temperature gradient of metal castings; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes on the surface of metal casting respectively; and They represent the maximum and minimum cooling rates at n network nodes on the surface of the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity on the surface of metal castings; 、 and are weight coefficients respectively, and the sum is 1; The formula for calculating and evaluating the cooling effect index inside metal castings is: Get the cooling effect index inside the metal casting ;in, It is a mold for the temperature gradient inside the metal casting; and They represent the maximum and minimum values of the module of temperature gradient at n network nodes inside the metal casting; and They represent the maximum and minimum cooling rates at n network nodes inside the metal casting; and Respectively represent the maximum and minimum values of cooling uniformity inside the metal casting; 、 and are weight coefficients respectively, and the sum is 1; The cooling effect index of the metal casting surface is compared with the cooling effect index of the metal casting interior to obtain the cooling uniformity deviation of the metal casting; when the cooling uniformity deviation is higher than the preset threshold When it is 0.15, it means that the metal casting is cooling unevenly.

6. A metal casting cooling control method according to claim 1, characterized in that: The step S4 comprises the following steps: When metal castings cool unevenly, a PID control algorithm is used to automatically adjust the flow, pressure, and temperature of the cooling medium. The control formula of the PID algorithm is: in, is the output of the controller at time t, is the error, 、 and are proportional, integral, and differential coefficients respectively; Set the target flow rate, target pressure and target temperature of the cooling medium, and monitor the actual flow rate, actual pressure value and actual temperature of the cooling medium in real time; calculate the control signal through the PID algorithm based on the difference between the target flow rate and the actual flow rate, and adjust the actual flow rate, actual pressure value and actual temperature of the cooling medium to the set target flow rate, target pressure and target temperature of the cooling medium; During the cooling process, the target flow rate, target pressure and target temperature of the cooling medium are dynamically adjusted according to the cooling effect index on the surface and inside of the metal casting.

7. A metal casting cooling control method according to claim 1, characterized in that: The step S5 comprises the following steps: Collecting thermal imaging data and three-dimensional temperature field model data corresponding to metal castings containing various defect types, including pores, shrinkage, and cracks; and preprocessing the collected thermal imaging data and three-dimensional temperature field model data; Label the presence of defects in metal castings and the type and location of the defects; use the labeled information as training labels for supervised learning to train deep learning models; Generate a data set from labeled thermal imaging data and 3D temperature field model data and divide it into a training set with a ratio of 70% for training, a validation set with a ratio of 15% for validation, and a test set with a ratio of 15% for testing; Use the convolutional neural network model and the recurrent neural network model in the deep learning model for combined training; input the labeled training set into the deep learning model that combines the convolutional neural network model and the recurrent neural network model for training; use the validation set and test set to verify and test the deep learning model respectively; When the deep learning model training is completed, the real-time collected and pre-processed thermal imaging data and three-dimensional temperature field model data of the metal casting are input into the trained deep learning model for defect detection, and the detected defects are annotated and visualized in the three-dimensional temperature field model.

8. A metal casting cooling control system, using a metal casting cooling control method according to any one of claims 1 to 7, characterized in that: Includes the following modules: Data acquisition module: uses sensors to collect real-time information data including: metal casting data, cooling data and ambient temperature; uses thermal imaging cameras to collect thermal imaging data of the casting surface; Data preprocessing module: performs filtering, denoising, data cleaning and standardization preprocessing on the collected real-time information data; performs filtering, denoising, calibration, image enhancement and grayscale preprocessing on the collected thermal imaging data; Model building and data analysis module: Builds a three-dimensional temperature field model of the metal casting based on real-time information data and thermal imaging data; analyzes the temperature distribution on the surface and inside the metal casting based on the constructed three-dimensional temperature field model and calculates the temperature gradient, cooling rate, and cooling uniformity; comprehensively evaluates the cooling effect index of the metal casting based on the calculated results of temperature gradient, cooling rate, and cooling uniformity, and calculates the cooling uniformity deviation of the metal casting; Cooling control module: uses PID control algorithm to automatically adjust the flow, pressure and temperature of the cooling medium; Defect Detection Module: Detects and visualizes defects in metal castings based on deep learning models.

Images