An industrial internet of things-based dynamic optimization system for precision casting die casting process

Abstract

The application discloses a kind of based on industrial internet of things's precision casting die casting process dynamic optimization system, specifically related to die casting process field, including sensing device deployment module, casting initialization module, die casting process extraction module, casting image acquisition module, casting quality evaluation module, die casting process analysis module, process optimization module, parameter feedback module.The application calculates the degree of deviation by calling initial process parameters, and carries out casting surface and internal sensing after casting die casting, and then constructs casting quality optimization model, so as to obtain the second demand process parameters optimized, and when verifying model deviation based on the landing effect of optimized parameters, process parameter updating and model adjustment are carried out, so as to form the complete closed loop of production, optimization, feedback, which can avoid the influence of artificial detection error on the consistency of quality determination, and is beneficial to reduce market after-sales failure risk and maintenance cost under the premise of guaranteeing the quality of casting leaving factory.

Description

Technical Field

[0001] This invention relates to the field of die casting technology, and more specifically, to a dynamic optimization system for precision die casting process based on the Industrial Internet of Things. Background Technology

[0002] With the rapid development of automotive lightweighting, medical devices and other fields, the demand for precision castings has surged, driving die casting technology towards thinner walls, greater complexity and integration. Traditional precision die casting processes are generally judged by senior process engineers by listening to the machine's sound, looking at the product surface, and breaking the sprue. Under cost pressure and high-quality requirements, traditional manufacturing industries are using the latest digital, networked and intelligent technologies to optimize processes. The die casting machine itself is controlled by a PLC program and can accurately and repeatedly execute the set sequence of actions.

[0003] However, it still has some drawbacks in actual use. First, the existing traditional precision die casting process is prone to quality fluctuations during production. In particular, when switching from simple structural parts to large, complex, thin-walled parts, more extreme process parameters will inevitably be used, which will put a greater load on the mold life and equipment condition. Based on this situation, it is necessary to adjust the process parameters before and after production. However, the existing operations to deal with quality fluctuations are mostly limited to fine-tuning the injection speed and pressure based on experience, which limits the process optimization effect and poses a risk of not being able to meet the quality requirements of high-end products. In addition, excessive adjustment of process parameters will also interfere with the production cycle to a certain extent, resulting in the inability to complete orders on time. At the same time, frequent changes in parameters mean repeated start-ups and shutdowns and trial molding of the production process, which will increase the thermal fatigue damage of the mold and the wear and tear of the equipment, and shorten the life of key assets.

[0004] Secondly, it is difficult to control the scrap rate when judging product quality after die casting. Traditional manual visual inspection and sampling dissection methods are greatly affected by human factors, and the judged product qualification range does not match the actual situation. This leads to the need for additional full inspection of the entire batch of products or the flow of quality defects to the customer, which increases quality and after-sales costs. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a dynamic optimization system for precision die casting process based on the Industrial Internet of Things. By extracting die casting process parameters and evaluating casting quality during the die casting process, the system enables refined and flexible adjustment of the die casting process, maximizing both the die casting process and the adjustment effect. Furthermore, after process optimization, parameter feedback is performed to identify model deviations, effectively solving the problems raised in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The sensing device deployment module is used to set up visual sensing terminals, die casting terminals, and sensor terminals in the die casting area of ​​the precision casting production line.

[0008] The casting initialization module is used to allow the die-casting terminal to retrieve the initial process parameters of the target casting model from the die-casting database for die-casting production after the operator inputs the model number.

[0009] The die casting process extraction module is used to extract the real-time process parameters of the target casting after the molten metal is die-cast to form the target casting, and compare them with the initial process parameters to obtain the degree of die casting deviation. Then, the real-time process parameters are corrected to obtain the first required process parameters.

[0010] The casting image acquisition module is used to perceive the surface and interior of the target casting through a visual perception terminal after the target casting is conveyed to the die casting area via the production line, and to extract the initial casting process features of the target casting.

[0011] The casting quality assessment module is used to analyze the casting quality of the target casting based on the initial casting process characteristics. The casting quality perception objects are casting integrity and casting accuracy, thereby calculating the comprehensive casting quality evaluation coefficient corresponding to the target casting.

[0012] The die casting process analysis module is used to perform correlation analysis between the comprehensive evaluation coefficient of casting quality corresponding to the target casting and the first required process parameters, thereby constructing a casting quality optimization model between casting quality and the first required process parameters;

[0013] The process optimization module optimizes the first required process parameters based on the casting quality optimization model between the casting quality and the first required process parameters, thereby obtaining the second required process parameters.

[0014] The parameter feedback module is used to transmit the second required process parameters to the die-casting terminal via the Industrial Internet of Things. The die-casting terminal updates the second required process parameters to the initial process parameters corresponding to the new die-casting process and performs model deviation identification.

[0015] The technical effects and advantages of this invention are as follows:

[0016] 1. This invention retrieves initial process parameters and calculates the degree of deviation based on a comparison between the initial process parameters and real-time process parameters, thereby obtaining the first required process parameters. It is not limited to static production with fixed process parameters. On the one hand, it can maximize the satisfaction of the working condition adaptation requirements of different types of castings, and on the other hand, it can avoid batch defects and waste of raw materials caused by continuous deviation of parameters.

[0017] 2. This invention perceives the surface and interior of the casting after die casting, thereby extracting the initial casting process characteristics. Then, it conducts quality assessment based on the integrity and precision of the casting, which is not limited to the single inspection by human eyes. This ensures the stability of casting quality to a certain extent, and avoids the impact of human inspection errors on the consistency of quality judgment. It is beneficial to reduce the risk of after-sales failure and maintenance costs in the market while ensuring the quality of castings leaving the factory.

[0018] 3. This invention constructs a casting quality optimization model, outputs feasible second-requirement process parameters for production, and updates and adjusts the process parameters and the model when verifying the model deviation based on the implementation effect of the optimization parameters. This forms a complete closed loop of production, optimization, and feedback, which can maximize the satisfaction of the production needs of multiple types of castings and the continuous upgrading of processes. At the same time, it can avoid optimization failure caused by the disconnect between the model and production, which is conducive to improving the stability of casting quality and production efficiency while ensuring production stability. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0020] Figure 2 This is a flowchart of the initial casting process feature extraction for the present invention.

[0021] Figure 3 This is a flowchart illustrating the construction of the casting quality optimization model for this invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] As attached Figure 1-3 The system shown is a dynamic optimization system for precision die casting process based on the Industrial Internet of Things, including a sensing device deployment module, a casting initialization module, a die casting process extraction module, a casting image acquisition module, a casting quality assessment module, a die casting process analysis module, a process optimization module, and a parameter feedback module.

[0024] The specific embodiments of the present invention include the following:

[0025] The sensing device deployment module is used to set up visual sensing terminals, die-casting terminals, and sensor terminals in the die-casting area of ​​the precision casting production line.

[0026] In a more specific application of this invention, the visual perception terminal is used to achieve visual monitoring of surface and internal defects of castings through image acquisition and analysis, making up for the limitations of traditional sensors in morphological detection. Specifically, it can be one of the following cameras: high-definition camera, infrared thermal imager, and industrial CT. When determining the camera selection, it is necessary to select one based on the actual suitability of the casting production monitoring and the camera. Among them, the high-definition camera can identify small defects such as scratches, dents, and bubbles on the surface of the casting; the infrared thermal imager can cover the entire mold surface and identify defects based on the temperature field distribution; and the industrial CT can monitor defects such as porosity, shrinkage, and density inside the casting.

[0027] The sensor terminal is used to sense the target casting during the casting production process. Specifically, it includes a photoelectric sensor, which is used to sense the casting on the casting production line. When the front end of the casting triggers the photoelectric sensor, it sends a signal.

[0028] The die-casting terminal is integrated into each die-casting equipment through the Industrial Internet of Things (IIoT) to acquire die-casting process parameters and analyze monitoring data and control related parameters.

[0029] The casting initialization module is used to allow the die-casting terminal to retrieve the initial process parameters of the target casting model from the die-casting database for die-casting production after the operator inputs the model number.

[0030] It should be added that operators can input the target casting model manually through the touch screen of the die-casting terminal, or they can select the model through the die-casting database of the die-casting terminal.

[0031] It should be explained that the initial process parameters are the die casting process parameters corresponding to the target casting model, including temperature parameters, pressure parameters, speed parameters, and mold parameters. Among them, the temperature parameters include the molten metal temperature and the mold preheating temperature, the injection parameters include the injection pressure, holding pressure, and injection speed, and the time parameters include the injection time, holding time, and cooling time.

[0032] The die casting process extraction module is used to extract the real-time process parameters of the target casting after the molten metal is die-cast to form the target casting, and compare them with the initial process parameters to obtain the degree of die casting deviation. Then, the real-time process parameters are corrected to obtain the first required process parameters.

[0033] In this embodiment, it should be specifically explained that the die-casting deviation degree is obtained through the following operations:

[0034] Real-time process parameters are extracted during the die-casting process using a die-casting terminal, and the corresponding die-casting time is extracted based on the time parameters of the real-time process parameters. Then, a die-casting time sequence is constructed according to the order of the die-casting times.

[0035] It should be added that the die casting time series is specifically constructed by dividing the time axis according to the order of die casting time, and the time axis is divided into injection stage, holding pressure stage and cooling stage according to the die casting process, which are respectively denoted as [0, ts), [ts, ts+th) and [ts+th, ts+th+tc), where ts is the injection end time, th is the holding pressure time and tc is the cooling time.

[0036] Based on the die-casting time, the real-time process parameters are compared with the initial process parameters to obtain the die-casting deviation of each parameter corresponding to the real-time process parameters. Specifically, the difference between the real-time process parameters and the initial process parameters is calculated, and the difference is compared with the allowable deviation value to obtain the die-casting deviation of each parameter corresponding to the real-time process parameters. The allowable deviation value is obtained by subtracting the allowable deviation range of each parameter during the die-casting process. The allowable deviation range is preset before the casting is die-cast.

[0037] The die-casting deviation of each real-time process parameter is weighted and averaged to obtain the comprehensive deviation, which is then compared with the deviation threshold. For example, the deviation threshold is 1. If the comprehensive deviation is outside the deviation threshold, the real-time process parameters are corrected. If the comprehensive deviation is within the deviation threshold, it means that the real-time process parameters have no obvious deviation, and no parameter correction is required.

[0038] The corrected real-time process parameters are saved to the die-casting terminal via the Industrial Internet of Things as the primary required process parameters.

[0039] Furthermore, the real-time process parameter correction includes: taking the real-time process parameters with a comprehensive deviation greater than the deviation threshold as the correction object, extracting the corresponding stage of the die casting process where the correction object is located, and calculating the correction amount according to the corresponding stage of the die casting process. The specific calculation of the correction amount is obtained by subtracting the correction object from the initial process parameter and multiplying it by the correction coefficient. The correction coefficient is dynamically adjusted according to the magnitude of the deviation. The larger the deviation, the larger the correction coefficient. For example, the correction coefficient ranges from 0.3 to 0.8. Then, the correction object and the correction amount are added together to obtain the real-time process parameter correction value.

[0040] In the example of calculating the real-time process parameter correction value above, assuming the correction object is the injection pressure in the injection stage, the real-time injection pressure is 70, and the initial injection pressure is extracted as 90, the die casting deviation of the injection pressure is 1.2, and the correction coefficient is 0.5. Then the correction amount is (90-70)×0.5=10, so the correction value is 70+10=80.

[0041] The casting image acquisition module is used to perceive the surface and interior of the target casting through a visual perception terminal after the target casting is conveyed to the die casting area via the production line, and to extract the initial casting process features of the target casting.

[0042] It should be added that, after the target casting is conveyed to the die-casting area via the production line, the specific execution of the perception by the vision perception terminal can be achieved by installing photoelectric sensors before the target casting enters the die-casting area. When the photoelectric sensors detect that the target casting has entered the die-casting area, the production line control center determines whether the start-up conditions are met. If the start-up conditions are met, the control center sends a start-up command to the vision perception terminal, thereby realizing the perception of the surface and interior of the target casting. The start-up conditions include sending the start-up command when the target casting reaches the first detection position, the second detection position, and the third detection position, respectively.

[0043] It should be explained that detecting the position of the target casting and determining the start-up conditions through photoelectric sensors is to ensure that the visual perception terminal scans in advance to obtain clear and accurate images. This helps to accurately identify the initial casting process characteristics on the surface and inside of the target casting.

[0044] It needs to be further explained that when the target casting reaches the first detection position, the high-definition camera is triggered to perform surface inspection of the target casting; when the target casting reaches the second detection position, the infrared thermal imager is triggered to perform near-surface inspection of the target casting; and when the target casting reaches the third detection position, the industrial CT is triggered to perform internal inspection of the target casting.

[0045] In this embodiment, it should be specifically explained that the initial casting process features of the target casting are extracted as follows:

[0046] When the target casting reaches the first detection position, a high-definition camera is used to scan the image of the target casting and extract the surface defect features of the casting from the image. The surface defect features of the casting include the surface defect type and the surface defect area. For example, the surface features of the casting include, but are not limited to, scratches, dents, bubbles, etc.

[0047] It should be explained that the surface defect features of the casting reflect the defects in the casting during die casting. The initial casting process features of the target casting are extracted from the casting image to reflect the defects in the casting during die casting. This is because the first detection position is the first quality inspection node after the casting is demolded. Extracting surface features at this time can obtain quality feedback in the shortest time. These features can reflect the surface quality of the casting and provide accurate data support for subsequent casting quality assessment.

[0048] When the target casting reaches the second detection position, the temperature field of the target casting is identified using an infrared imager, and the surface temperature anomaly features and temperature field distribution features are extracted based on the temperature field identification results. The near-surface temperature anomaly features of the casting include the area of ​​the abnormal temperature region and the number of abnormal temperature distributions. The temperature field distribution features include the temperature field temperature and the isotherm spacing. For example, the near-surface temperature anomaly features of the casting include, but are not limited to, local low temperature points, local high temperature points, and temperature gradient abrupt change points.

[0049] It should be explained that the infrared imager generates a real-time temperature field thermogram by covering the entire surface of the target casting. First, it identifies abnormal temperature areas through a temperature threshold algorithm and counts the area of ​​each abnormal area. Then, it extracts the temperature field distribution characteristics through temperature field uniformity analysis. The temperature field temperature reflects the temperature distribution of the target casting, and the isotherm spacing reflects the uniformity of the corresponding molten metal filling. These characteristics are directly related to the mold temperature control, cooling water circuit design, and molten metal flow state in the die casting process, providing a basis for the near-surface quality dimension for subsequent process optimization.

[0050] When the target casting reaches the third detection position, the internal structure of the target casting is scanned using industrial CT, and the internal defect features of the casting are extracted based on the scanned structure. The internal defect features include the type of internal defect and the volume ratio of the defect. For example, the internal defect features of the casting include, but are not limited to, internal porosity, shrinkage porosity, and non-metallic inclusions.

[0051] It should be added that when extracting the initial casting process features of the target casting, preprocessing operations such as image enhancement, noise reduction, and background segmentation of the acquired images are required, which can significantly improve the accuracy and reliability of subsequent feature extraction.

[0052] The casting quality assessment module is used to analyze the casting quality of the target casting based on the initial casting process characteristics. The casting quality perception objects are casting integrity and casting accuracy, thereby calculating the comprehensive casting quality evaluation coefficient corresponding to the target casting.

[0053] In this embodiment, it should be specifically noted that the casting integrity perception is as follows:

[0054] The surface defect types are extracted from the surface defect features of the casting, and the surface defect areas corresponding to the same surface defect type are summed to obtain the cumulative area of ​​each surface defect type.

[0055] Influence factors are assigned to each surface defect type. The cumulative area of ​​each surface defect type is divided by the total area of ​​the casting inspection and multiplied by the corresponding influence factor for each surface defect type. The results of the multiplication are then added together to obtain the surface defect degree of the casting.

[0056] It should be added that the influence factor allocation for each surface defect type can be set based on experience, taking into account the degree of influence of different surface defect types on the casting quality in different scenarios. For example, in the production of decorative castings, appearance is the priority consideration for casting quality. Therefore, defects that affect aesthetics, such as scratches, have a higher weight, with the weight of scratches being 0.7.

[0057] Similarly, the volume of internal defects is extracted based on the characteristics of internal defects in the casting, and then the internal defect degree of the casting is calculated by adding the weights of different internal defect types.

[0058] The casting integrity coefficient is calculated by adding the surface defect rate and the internal defect rate of the casting and then averaging the results. Specifically, it is expressed as follows:

[0059] ,

[0060] Where Hd represents the casting integrity coefficient of the target casting, and Md and Sd represent the surface defect degree and internal defect degree of the target casting, respectively. The smaller the surface defect degree and internal defect degree of the target casting, the better the quality of the target casting, and the larger the casting integrity coefficient.

[0061] It should be further noted that the accuracy perception of the castings is as follows:

[0062] Extract the area of ​​abnormal temperature region from the surface temperature anomaly characteristics of the target casting, and compare it with the area of ​​temperature distribution region to obtain the proportion of abnormal temperature.

[0063] Extract the number of abnormal temperature distributions and compare them with the number of temperature distribution areas to obtain the proportion of abnormal areas;

[0064] The degree of temperature distribution anomaly is obtained by weighting the proportion of abnormal temperatures and the proportion of abnormal areas. For example, the weights of the proportion of abnormal temperatures and the proportion of abnormal areas are 0.6 and 0.4, respectively.

[0065] The temperature field temperature is extracted from the temperature field distribution characteristics of the target casting, and then the maximum and minimum temperatures in the temperature field are extracted. The difference between the two is then used to obtain the maximum temperature difference value.

[0066] The isotherm spacing is extracted to extract the temperature field distribution characteristics, and quantified by comparing the numerical value 1 with the isotherm spacing, thereby obtaining the isotherm density.

[0067] Based on the density of each isotherm corresponding to the temperature field, the average density and standard deviation of the density are obtained using the average value calculation formula and the standard deviation calculation formula, respectively. The average density and the standard deviation of the density are then compared exponentially to obtain the temperature distribution concentration, which is specifically expressed as follows:

[0068] ,

[0069] Where Cm represents the temperature distribution concentration, Tb and Ta represent the standard deviation and average concentration of the density, respectively. This formula shows that when the standard deviation of the density is larger, it indicates that the temperature distribution is more dispersed and the concentration is lower. At this time, the temperature distribution of the target casting is more uniform.

[0070] The casting accuracy coefficient is calculated by averaging the exponential sums of temperature distribution anomaly and temperature distribution concentration, as follows:

[0071] ,

[0072] Where Qd represents the casting accuracy coefficient of the target casting, and Tm and Cm represent the temperature distribution anomaly and temperature distribution concentration of the target casting, respectively. The smaller the temperature distribution anomaly and temperature distribution concentration of the target casting, the more uniform the temperature distribution of the target casting, and the larger the casting accuracy coefficient.

[0073] Furthermore, the comprehensive evaluation coefficient of casting quality is obtained by multiplying the casting integrity coefficient and casting accuracy coefficient of the target casting.

[0074] The die casting process analysis module is used to perform correlation analysis between the comprehensive evaluation coefficient of casting quality corresponding to the target casting and the first required process parameters, thereby constructing a casting quality optimization model between casting quality and the first required process parameters.

[0075] In this embodiment, it should be specifically noted that the construction of the casting quality optimization model is as follows:

[0076] The correlation coefficient between the comprehensive evaluation coefficient of the casting quality corresponding to the target casting and the first required process parameter is calculated, thereby obtaining the correlation between the first required process parameter and the comprehensive evaluation of casting quality.

[0077] It is important to understand that the correlation coefficient can specifically be the Pearson correlation coefficient. When calculating the correlation coefficient between the comprehensive evaluation coefficient of casting quality and the first required process parameter, the comprehensive evaluation coefficient of casting quality and the first required process parameter can be treated as five variables, thus dividing the five variables into five arrays. Then, the covariance and standard deviation of each of the five arrays are calculated. Finally, the Pearson correlation coefficient is calculated using the covariance and standard deviation. The correlation coefficient obtained through the calculation can provide the correlation between the comprehensive evaluation coefficient of casting quality and the first required process parameter, thereby providing a basis for the subsequent construction of the casting quality optimization model.

[0078] Set a relevant threshold, for example, the relevant threshold is 0.4. When the absolute value of the correlation is greater than or equal to the relevant threshold, it is determined that the comprehensive evaluation coefficient of casting quality is strongly correlated with the first required process parameter; otherwise, it is determined that the correlation is weak.

[0079] The primary process parameters with strong correlation are screened, while those with weak correlation are removed. Then, the screened primary process parameters are standardized.

[0080] It should be added that the purpose of screening the first requirement process parameters is to simplify the model dimensions and avoid irrelevant or weakly influential parameters from interfering with the optimization direction. The purpose of standardizing the screened first requirement process parameters is to eliminate the dimensions of each process parameter, thereby ensuring that the casting quality optimization model can directly optimize the parameters and avoid the problem of misjudgment due to large dimensions.

[0081] Using the comprehensive evaluation coefficient of casting quality as the dependent variable and the selected primary requirement process parameters as independent variables, a casting quality optimization model is constructed, specifically expressed as: Y=k1*X1+k2*X2+…+k n *X n Where Y represents the comprehensive evaluation coefficient of casting quality, k1, k2, ..., k n The regression coefficients, X1, X2, ..., X, represent the influence of various process parameters on the quality of the casting. n This represents the standardized values ​​of the first required process parameters after screening, X1, X2, ..., X... n All are greater than 0 and less than or equal to 1. The positive and negative values ​​of the influence coefficients reflect the direction of influence. Positive coefficients indicate that the comprehensive evaluation coefficient of casting quality increases when the parameter increases, while negative coefficients have the opposite effect. The regression coefficients can be solved by fitting using the least squares method.

[0082] The process optimization module optimizes the first required process parameters based on the casting quality optimization model between the casting quality and the first required process parameters, thereby obtaining the second required process parameters.

[0083] In this embodiment, it should be specifically explained that the second requirement process parameters are obtained as follows: Based on the constructed casting quality optimization model, a linear programming solution algorithm is used to traverse and calculate all standardized process parameter combinations within the feasible region of the model, thereby calculating the Y value corresponding to each combination, and comparing each Y value to select the standardized parameter combination with the maximum Y value.

[0084] Based on the selected standardized parameter combinations, the required process parameters are reverse-converted, and the optimal process parameter combination is calculated as the second required process parameter.

[0085] It should be added that, since the selected standardized parameter combinations are standardized values, they cannot directly guide production. By using a standardized reverse conversion formula, they are restored to the original process parameter units to obtain the final executable second requirement process parameters.

[0086] The parameter feedback module is used to transmit the second required process parameters to the die-casting terminal via the Industrial Internet of Things. The die-casting terminal updates the second required process parameters to the initial process parameters corresponding to the new die-casting process and performs model deviation identification.

[0087] In this embodiment, it should be specifically explained that the model deviation identification involves re-analyzing the integrity and accuracy of the casting after the second required process parameters are updated to the initial process parameters corresponding to the new die-casting process. This analysis is then compared with the integrity and accuracy of the casting before the process parameters were optimized to obtain the difference in casting integrity and accuracy. If the difference in casting integrity and accuracy is positive, it indicates that the casting quality has not improved. In this case, the casting quality optimization model needs to be rebuilt, and an early warning notice should be issued to the operator to check the die-casting equipment for faults. If the difference in casting integrity and accuracy is negative, it indicates that the casting quality has improved, and the second required process parameters are continued to be used as the initial process parameters for die-casting.

[0088] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0089] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An industrial internet of things based dynamic optimization system for precision casting process of die casting, characterized in that, include: The sensing device deployment module is used to set up visual sensing terminals, die casting terminals, and sensor terminals in the die casting area of ​​the precision casting production line. The casting initialization module is used to allow the die-casting terminal to retrieve the initial process parameters of the target casting model from the die-casting database for die-casting production after the operator inputs the model number. The die casting process extraction module is used to extract the real-time process parameters of the target casting after the molten metal is die-cast to form the target casting, and compare them with the initial process parameters. Based on the order of die casting time, a die casting time series is constructed, the die casting deviation of each real-time process parameter is calculated, and the weighted average is used to obtain the comprehensive deviation. If the overall deviation exceeds the deviation threshold, the correction coefficient is dynamically adjusted based on the die-casting stage where the real-time process parameters are located to obtain the first required process parameters. The casting image acquisition module is used to sense the surface and interior of the target casting after it is conveyed to the die casting area through the production line. The module uses a high-definition camera at the first detection position, an infrared thermal imager at the second detection position, and an industrial CT at the third detection position to extract the initial casting process features, including surface defect features, surface temperature anomaly features, temperature field distribution features, and internal defect features. The casting quality assessment module is used to perform casting quality analysis on target castings based on the initial casting process characteristics, including calculating the casting integrity coefficient and casting accuracy coefficient, and multiplying them to obtain the comprehensive casting quality evaluation coefficient; The casting integrity coefficient is calculated based on the cumulative area of ​​surface defect types and the distribution of influencing factors, while the casting accuracy coefficient is calculated based on the proportion of abnormal temperatures, the proportion of abnormal areas, and the concentration of temperature distribution. The die casting process analysis module is used to perform correlation analysis between the comprehensive evaluation coefficient of casting quality corresponding to the target casting and the first required process parameters. The correlation is calculated by Pearson correlation coefficient, parameters with strong correlation are selected, and they are standardized to construct a casting quality optimization model. The process optimization module, based on the casting quality optimization model, uses a linear programming algorithm to traverse all standardized process parameter combinations within the feasible region, selects the combination that maximizes the comprehensive evaluation coefficient of casting quality, and obtains the second required process parameters after reverse transformation. The parameter feedback module is used to transmit the second required process parameters to the die-casting terminal via the Industrial Internet of Things. The die-casting terminal updates the second required process parameters to the new initial process parameters, and re-analyzes the integrity and accuracy of the casting, calculates the comparison difference, and identifies model deviations. If the comparison difference is positive, the model is rebuilt and an early warning is issued; if it is negative, production continues to use the same parameters.

2. The system for dynamic optimization of precision die casting process based on industrial internet of things according to claim 1, characterized in that: The specific steps for obtaining the first required process parameters are as follows: Real-time process parameters are extracted during the die-casting process through the die-casting terminal, and the corresponding die-casting time is extracted based on the time parameters of the real-time process parameters. Then, a die-casting time sequence is constructed according to the order of the die-casting times. Based on the die-casting time, the real-time process parameters are compared with the initial process parameters to obtain the die-casting deviation of each parameter corresponding to the real-time process parameters. Specifically, the difference between the real-time process parameters and the initial process parameters is calculated, and the difference is compared with the allowable deviation value to obtain the die-casting deviation of each parameter corresponding to the real-time process parameters. The weighted average of the die-casting deviation of each real-time process parameter is calculated to obtain the comprehensive deviation, which is then compared with the deviation threshold. If the overall deviation is outside the deviation threshold, then real-time process parameter correction is performed. If the overall deviation is within the deviation threshold, it means that there is no significant deviation in the real-time process parameters, and no parameter correction is required. The corrected real-time process parameters are saved to the die-casting terminal via the Industrial Internet of Things as the primary required process parameters.

3. The dynamic optimization system for precision die casting process based on industrial Internet of Things as described in claim 1, characterized in that: The initial casting process features of the target casting are extracted when the target casting reaches each detection position, wherein the detection positions include the first detection position, the second detection position, and the third detection position. The specific operation is as follows: When the target casting reaches the first detection position, a high-definition camera is used to scan the image of the target casting and extract the surface defect features of the casting from the image. The surface defect features of the casting include the surface defect type and the surface defect area. When the target casting reaches the second detection position, the infrared imager is used to identify the temperature field of the target casting, and the surface temperature anomaly features and temperature field distribution features are extracted based on the temperature field identification results. The near-surface temperature anomaly features of the casting include the area of ​​the abnormal temperature region and the number of abnormal temperature distributions, and the temperature field distribution features include the temperature field temperature and the isotherm spacing. When the target casting reaches the third detection position, the internal structure of the target casting is scanned using industrial CT, and the internal defect features of the casting are extracted based on the scanned structure. The internal defect features include the type of internal defect and the volume ratio of the defect.

4. The system of claim 1, wherein the system is configured to: The casting integrity perception is described below: ​ Surface defect types are extracted from the surface defect features of the casting, and the surface defect areas corresponding to the same surface defect type in the surface defect features of the casting are summed to obtain the cumulative area of ​​each surface defect type. Influence factors are assigned to each surface defect type. The cumulative area of ​​each surface defect type is divided by the total area of ​​the casting inspection and multiplied by the corresponding influence factor of each surface defect type. The results of the multiplication are then added together to obtain the surface defect degree of the casting. Similarly, the volume of internal defects is extracted based on the characteristics of internal defects in the casting, and then the internal defect degree of the casting is calculated by adding the weights of different internal defect types. The casting integrity coefficient is calculated by adding the indices of the surface defects and internal defects of the casting and taking the average value.

5. The system for dynamic optimization of precision die casting process based on industrial internet of things according to claim 1, wherein: The accuracy perception of the casting is as follows: Extract the area of ​​abnormal temperature region from the surface temperature anomaly characteristics of the target casting, and compare it with the area of ​​temperature distribution region to obtain the proportion of abnormal temperature. Extract the number of abnormal temperature distributions and compare them with the number of temperature distribution areas to obtain the proportion of abnormal areas; The degree of temperature distribution anomaly is obtained by weighting the proportion of abnormal temperatures and the proportion of abnormal areas. The temperature field temperature is extracted from the temperature field distribution characteristics of the target casting, and then the maximum and minimum temperatures in the temperature field are extracted. The difference between the two is then used to obtain the maximum temperature difference value. The isotherm spacing is extracted to extract the temperature field distribution characteristics, and quantified by comparing the numerical value 1 with the isotherm spacing, thereby obtaining the isotherm density. The average density and standard deviation of the density are obtained by using the average value calculation formula and the standard deviation calculation formula respectively based on the density of each isotherm corresponding to the temperature field. The average density and the standard deviation of the density are then compared and exponentially calculated to obtain the temperature distribution concentration. The casting accuracy coefficient is calculated by adding the exponents of temperature distribution anomaly and temperature distribution concentration and taking the average value.

6. The system of claim 1, wherein the system is configured to: determine a set of process parameters for a new casting process based on the set of process parameters for the previous casting process and the set of process parameters for the current casting process. The specific construction of the casting quality optimization model is as follows: The correlation coefficient between the comprehensive evaluation coefficient of the casting quality corresponding to the target casting and the first required process parameter is calculated, thereby obtaining the correlation between the first required process parameter and the comprehensive evaluation of casting quality. Set a relevant threshold. When the absolute value of the correlation is greater than or equal to the relevant threshold, it is determined that the comprehensive evaluation coefficient of casting quality is strongly correlated with the first required process parameter; otherwise, it is determined that the correlation is weak. The primary demand process parameters with strong correlation are screened, and the primary demand process parameters with weak correlation are removed. Then, the screened primary demand process parameters are standardized. With the comprehensive evaluation coefficient of casting quality as the dependent variable and the screened first demand process parameters as the independent variable, a casting quality optimization model is constructed, which is specifically expressed as: Y=k1*X1+k2*X2+…+k n *X n , wherein Y represents the comprehensive evaluation coefficient of casting quality, k1, k2, …, k n are regression coefficients, reflecting the influence coefficients of various process parameters on casting quality, X1, X2, …, X n represent the standardized values of the screened first demand process parameters, and X1, X2, …, X n are all greater than 0 and less than or equal to 1.

7. The system for dynamic optimization of precision die casting process based on industrial internet of things according to claim 1, wherein: The second required process parameters are obtained as follows: Based on the constructed casting quality optimization model, a linear programming algorithm is used to traverse and calculate all standardized process parameter combinations within the feasible region of the model, thereby calculating the Y value corresponding to each combination, and comparing each Y value to select the standardized parameter combination with the maximum Y value. Based on the selected standardized parameter combinations, the required process parameters are reverse-converted, and the optimal process parameter combination is calculated as the second required process parameter.

8. The system of claim 1, wherein the system is configured to: determine a set of process parameters for a new casting process based on the set of process parameters for the previous casting process and the set of process parameters for the current casting process. The model deviation identification involves re-analyzing the integrity and accuracy of the casting after updating the second required process parameters to the initial process parameters corresponding to the new die-casting process. This is then compared with the integrity and accuracy of the casting before process parameter optimization to obtain the difference in casting integrity and accuracy. If the difference in casting integrity and accuracy is positive, it indicates that the casting quality has not improved. In this case, the casting quality optimization model needs to be reconstructed, and an early warning notice is issued to the operator to check the die-casting equipment for faults. If the difference in casting integrity and accuracy is negative, it indicates that the casting quality has improved, and the second required process parameters are continued to be used as the initial process parameters for die-casting.

Images