Self-adaptive control optimization method of surface grinding machine for mold component manufacturing

Abstract

The invention discloses a surface grinding machine self-adaptive control optimization method for mold component manufacturing, and belongs to the technical field of intelligent manufacturing. The method comprises the steps that machining prediction modeling is conducted on manufacturing design information through the digital twinning technology; a mold component manufacturing data set of the surface grinding machine is collected and obtained, and retrieval matching is conducted on the basis of the manufacturing design information and the mold component manufacturing data set; carrying out processing production and real-time monitoring on the target mold component based on the target grinding machine processing parameters; performing processing prediction on the real-time processing data of the mold component by using the digital twin model of the mold component; and adaptive optimization is carried out on the target grinding machine machining parameters based on the mold component prediction manufacturing information, and adaptive optimization control is carried out on the target mold component through the grinding machine optimization machining parameters. The technical problems that in the prior art, die component machining parameters cannot be dynamically optimized, and machining precision and efficiency are difficult to meet at the same time, so that machining quality is unstable, and trial and error cost is high are solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to an adaptive control optimization method for a surface grinder used in the manufacture of mold parts. Background Art

[0002] Mold parts have extensive and important applications in modern manufacturing, including the automotive, electronics, home appliances and other industries, which have put forward stringent requirements for the high precision and high stability of mold parts. In order to meet the functional requirements and appearance quality of the product, the mold manufacturing process often involves complex multi-process processing links. Surface grinding, as a common process in mold finishing, must not only ensure high-precision shape and size, but also take into account multiple indicators such as material properties, surface roughness and processing efficiency.

[0003] With the continuous advancement of intelligent manufacturing and digital transformation, the industry has gradually paid attention to improving the processing quality and efficiency of mold parts through digital twin technology, big data analysis, advanced prediction algorithms and other means. Existing surface grinder controls mostly use fixed empirical parameters or semi-automatic parameter compensation methods, which cannot be adaptively adjusted according to the real-time processing status, and it is difficult to timely predict and correct dynamic changing factors such as grinding wheel loss, grinding temperature and part deformation. Therefore, there is an urgent need for an optimization method that integrates digital twin technology and intelligent algorithms to dynamically monitor and adjust processing parameters to comprehensively improve the efficiency and quality of mold component manufacturing. Summary of the invention

[0004] This application provides an adaptive control optimization method for a surface grinder used in mold component manufacturing, aiming to solve the technical problems in the prior art that mold component processing parameters cannot be dynamically optimized, and processing accuracy and efficiency are difficult to meet simultaneously, resulting in unstable processing quality and high trial and error costs.

[0005] In view of the above problems, the present application provides an adaptive control optimization method for a surface grinder used in mold component manufacturing.

[0006] The present application provides an adaptive control optimization method for a surface grinder used for mold component manufacturing, the method comprising obtaining manufacturing design information of a target mold component, performing processing prediction modeling on the manufacturing design information using digital twin technology, and generating a digital twin model of the mold component; acquiring a mold component manufacturing data set of a surface grinder, performing retrieval and matching based on the manufacturing design information and the mold component manufacturing data set, and determining target grinder processing parameters; performing processing and real-time monitoring of the target mold component based on the target grinder processing parameters, and obtaining real-time processing data of the mold component; performing processing prediction on the real-time processing data of the mold component using the digital twin model of the mold component, and obtaining predicted manufacturing information of the mold component; adaptively optimizing the target grinder processing parameters based on the predicted manufacturing information of the mold component, outputting the optimized grinder processing parameters, and adaptively optimizing and controlling the target mold component through the optimized grinder processing parameters.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: By adopting a technical solution that obtains mold manufacturing design information and uses digital twin technology to generate a mold digital twin model, collects surface grinder manufacturing data and performs matching optimization, monitors processing data in real time and performs adaptive parameter optimization based on prediction results, it solves the technical problems in the existing technology that mold component processing parameters cannot be dynamically optimized, and processing accuracy and efficiency are difficult to meet at the same time, resulting in unstable processing quality and high trial and error costs. It achieves the technical effect of real-time prediction of processing results, dynamic adjustment of processing parameters, and improvement of processing accuracy and efficiency, providing an efficient and intelligent optimization method for the manufacturing of complex mold components.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flow chart of an adaptive control optimization method for a surface grinder used in mold component manufacturing is provided for an embodiment of the present application.

[0010] Figure 2 A schematic flow chart of determining target grinding machine processing parameters in an adaptive control optimization method for a surface grinder used in mold component manufacturing is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0011] The overall idea of ​​the technical solution provided by this application is as follows: The embodiment of the present application provides an adaptive control optimization method for a surface grinder used in mold component manufacturing. First, the mold manufacturing design information is obtained and a digital twin model is established. Then, the surface grinder processing data is collected, matching optimization is performed, and initial processing parameters are determined. During the processing, the data is monitored in real time and predicted using the digital twin model. The processing parameters are adjusted based on the prediction results to achieve adaptive optimization control.

[0012] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0013] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a surface grinder adaptive control optimization method for mold component manufacturing, the method comprising: Step S100: Acquire manufacturing design information of a target mold component, use digital twin technology to perform processing prediction modeling on the manufacturing design information, and generate a digital twin model of the mold component.

[0014] Specifically, manufacturing design information refers to various data used to describe the design requirements of mold components, including but not limited to geometric structure, dimensional tolerance, material properties, heat treatment requirements, functional requirements, etc. The digital twin model of mold components refers to a virtual model built using digital twin technology, which not only contains the three-dimensional geometric shape of the mold components, but also integrates its material properties, dynamic data of the processing process, time series prediction and other multi-dimensional information to achieve prediction and feedback of the actual processing process.

[0015] First, it is necessary to obtain the three-dimensional structural design drawings of the mold parts from the CAD software, and read the relevant manufacturing design information such as part processing requirements, material grades, heat treatment data, etc. from the company's PLM (product lifecycle management) system or MES (manufacturing execution system). Then, based on these design information, the three-dimensional geometric structure and material properties of the parts are integrated in the three-dimensional modeling tool to build a preliminary virtual model. Subsequently, the digital twin technology is introduced to establish a processing prediction model by matching the above virtual model with the historical processing parameters collected by the sensor. Here, finite element analysis software can be used to numerically simulate the stress and deformation in the grinding process, or the deep learning-based algorithm in Python or MATLAB can be used to carry out time series prediction training on the historical processing data, thereby realizing the fusion between the virtual model and the dynamic processing features. After the digital twin model is established, the effectiveness of the current processing strategy can be judged based on the real-time or predicted processing status, and the foundation for the adaptive optimization of subsequent process parameters is laid.

[0016] Through this step, the digital twin model of the mold components constructed greatly shortens the processing parameter debugging cycle and significantly improves the predictability of processing.

[0017] Step S200: Acquire a mold component manufacturing data set of a surface grinder, perform retrieval and matching based on the manufacturing design information and the mold component manufacturing data set, and determine target grinder processing parameters.

[0018] Specifically, the mold component manufacturing data set refers to the historical data set recorded and stored during the processing of mold components, including processing parameters (such as cutting depth, grinding wheel speed, feed speed), processing status (such as vibration amplitude, temperature), processing results (such as surface roughness, dimensional tolerance) and related equipment operation data. The target grinder processing parameters refer to the surface grinder operating parameters recommended for actual processing after retrieval matching and optimization, such as grinding wheel speed, feed speed, cutting depth, etc., which can meet the mold design requirements.

[0019] First, collect mold component manufacturing data sets from production databases or sensor systems. These data include parameters and results recorded in historical processing processes and are stored in relational databases (such as MySQL) or big data platforms (such as Hadoop). Next, the acquired manufacturing design information is used as input, including the mold geometry, material properties, and target surface quality requirements. Then, based on the manufacturing design information and historical data, a retrieval matching algorithm is used for analysis. According to the design information, the most similar processing case to the current demand is found in the historical data, and the preliminary matching processing parameters are output. After the matching is completed, the preliminary matching parameters are optimized in combination with the specific mold processing requirements.

[0020] Through this step, accurate matching and optimization of processing parameters from historical data to actual needs are achieved, which significantly reduces the trial and error time in the parameter setting process and improves processing efficiency. At the same time, the retrieval matching method combined with the optimization algorithm can recommend the optimal parameter combination for different design requirements to ensure that the processing quality and accuracy meet the design requirements.

[0021] Step S300: Processing and real-time monitoring the target mold component based on the target grinder processing parameters to obtain real-time data of mold component processing.

[0022] Specifically, the optimized target grinding machine processing parameters are input into the surface grinder control system, such as setting the grinding wheel speed to 3000RPM, the cutting depth to 0.5mm, and the feed speed to 800mm / min. Then, the grinder is started to process the target mold parts.

[0023] During the processing, real-time data is collected through a real-time monitoring system. For example, a vibration sensor (such as an accelerometer) installed on a grinder can collect vibration signals and measure the contact status between the grinding wheel and the workpiece during processing; an infrared temperature sensor can monitor the temperature change of the processing surface; a laser measuring instrument or a 3D scanner can capture the size and surface roughness of mold parts. The data is transmitted to the central processing unit in real time for analysis through an industrial data acquisition system (such as PLC or SCADA).

[0024] Through processing production and real-time monitoring based on target processing parameters, dynamic changes in the processing process can be effectively captured to ensure processing accuracy and stability. The real-time monitoring system can detect potential problems in time and respond quickly to avoid further spread of processing defects.

[0025] Step S400: Utilize the mold component digital twin model to perform processing prediction on the real-time processing data of the mold component to obtain mold component prediction manufacturing information.

[0026] Specifically, mold component predictive manufacturing information refers to the processing result data predicted by the digital twin model, including information such as the quality, shape characteristics, and error distribution of the processed surface, providing a basis for real-time optimization of the processing process.

[0027] First, the sensor system is used to collect real-time data of mold component processing, such as obtaining vibration data through acceleration sensors, recording surface temperature changes through infrared sensors, and capturing geometric information of the processing surface in real time through laser scanners. These data are transmitted to the central control system through industrial data acquisition systems (such as PLC or IoT gateways).

[0028] The collected real-time data is then input into the digital twin model of the mold component. The digital twin model simulates and predicts the current processing status by integrating the 3D solid model, the processing status prediction network and the algorithm constructed by historical data. For example, the model can predict the surface roughness or thermal deformation risk through real-time temperature and cutting depth data.

[0029] The prediction process of the model can be implemented using a deep learning framework. For example, a processing prediction network trained based on historical data can output the processing results (such as surface quality or dimensional deviation) at the next moment after inputting real-time processing parameters.

[0030] By predicting real-time processing data through the digital twin model, potential problems in processing can be identified in advance, defects can be avoided from expanding, and processing quality and efficiency can be improved. In addition, the prediction results can also provide data support for real-time optimization of parameters, further reducing scrap rate and processing time.

[0031] Step S500: adaptively optimizing the target grinder processing parameters based on the predicted manufacturing information of the mold component, outputting the optimized grinder processing parameters, and adaptively optimizing and controlling the target mold component through the optimized grinder processing parameters.

[0032] Specifically, the grinding machine optimized processing parameters are the final processing parameters output after adaptive optimization, which can solve the deviation problem and ensure the processing quality. The optimized parameters are applied to the actual processing again to form a closed-loop control. Adaptive optimization control is a control method that dynamically adjusts the processing process. According to real-time feedback data and optimization results, the processing parameters are automatically adjusted to ensure the stability and efficiency of the entire processing process.

[0033] First, the predicted manufacturing information of the mold parts is obtained from the digital twin model, including the predicted value of surface quality, dimensional deviation or processing efficiency index. Then, the predicted information is compared with the manufacturing design requirements, the deviation is calculated, and it is input into the optimization algorithm. For example, if the predicted value of surface roughness is too high, it can be improved by adjusting the grinding wheel speed or reducing the cutting depth. This optimization process can use particle swarm optimization (PSO), genetic algorithm (GA) or gradient descent method. The SciPy library of Python is used to implement parameter optimization, and the optimization goal is set to minimize the surface roughness deviation, while constraining the processing efficiency to be not less than the set value.

[0034] After the optimization is completed, the adjusted grinding machine optimization processing parameters are output, such as the grinding wheel speed is adjusted from 3000RPM to 3200RPM, and the cutting depth is reduced from 0.5mm to 0.4mm. The optimized parameters are applied to the grinder again, and the target mold parts are processed. During the processing, new processing data is obtained through the real-time monitoring system and predicted and optimized again to form a closed-loop adaptive control.

[0035] Through adaptive optimization based on prediction information, processing parameters can be adjusted dynamically to ensure that the processing results meet the design requirements and significantly improve the processing quality and efficiency. Closed-loop control of the processing process is achieved, reducing trial and error costs and manual intervention.

[0036] Furthermore, the generation of a digital twin model of a mold component includes: obtaining mold component structural design information and mold component processing requirement attribute information based on the manufacturing design information; performing three-dimensional modeling and attribute assignment based on the mold component structural design information and the mold component processing requirement attribute information to generate a three-dimensional solid model of the mold component; collecting and obtaining a historical processing data set of the mold component, and performing processing prediction training on the historical processing data set of the mold component using a recurrent neural network structure to obtain a mold component processing state prediction network; and using digital twin technology to perform twin fusion modeling on the mold component three-dimensional solid model and the mold component processing state prediction network to generate the mold component digital twin model.

[0037] Specifically, the structural design information of mold parts refers to the geometric structure data of mold parts, such as shape, size, positional relationship, etc. It is usually generated by computer-aided design (CAD) software and directly reflects the physical shape and spatial layout of the mold. The attribute information of mold parts processing requirements refers to the technical requirements that need to be met during the mold manufacturing process, including material properties (such as hardness, thermal conductivity), surface roughness, tolerance range, and processing efficiency. The three-dimensional solid model is a three-dimensional visualization model of the mold parts, which contains geometric information and attribute information. The mold parts historical processing data set is a data set recorded in the past processing process of the mold parts, including processing parameters and environmental conditions. These data can be used as training data for machine learning models. The processing state prediction network is a prediction model trained based on a machine learning algorithm, which is used to predict the state or result of the processing process, such as vibration amplitude, temperature change, or surface roughness, based on the input processing parameters.

[0038] First, extract the structural design information and processing requirement attribute information from the manufacturing design information of the target mold parts. Use AutoCAD software to extract the three-dimensional geometric information of the mold. Next, use the three-dimensional modeling tool to combine the extracted information for three-dimensional modeling. The three-dimensional solid model needs to contain not only the geometric features of the mold, but also the material properties, tolerance information, etc.

[0039] After obtaining the 3D solid model, collect the historical processing data set of the mold parts, such as the historical data recorded by industrial sensors or production databases, including feed speed, cutting depth, surface roughness after processing, etc. Use recurrent neural networks to train these historical data and build a processing state prediction network through machine learning frameworks such as Keras or PyTorch.

[0040] Finally, the 3D solid model is integrated with the processing state prediction network by combining digital twin technology. In this process, the mold geometry and processing behavior can be bound in the same twin model with the help of platforms such as Siemens MindSphere or Ansys Twin Builder.

[0041] By generating a digital twin model of the mold parts, potential processing problems can be predicted before processing, such as excessive vibration, substandard surface roughness, etc., and the processing parameters can be optimized through simulation to reduce the number of trial and error. During the processing, the twin model can monitor the processing status in real time, compare it with the actual processing results, and dynamically adjust the processing parameters to ensure that the accuracy and quality meet the design requirements. The analysis feedback after processing can also be used to improve the efficiency of the next processing. It improves the manufacturing efficiency and accuracy of mold parts.

[0042] Furthermore, the method of obtaining the mold component processing state prediction network includes: performing data cleaning and normalization on the mold component historical processing data set to obtain a standard mold component historical processing data set; arranging, integrating and marking the processing state of the standard mold component historical processing data set according to timing information to obtain a mold component timing processing sample set; performing processing prediction training and loss assessment calculation on the mold component timing processing sample set using a recurrent neural network structure to obtain an initial processing state prediction network and prediction network loss data; and using a back propagation algorithm to optimize and update the model parameters of the initial processing state prediction network based on the prediction network loss data to obtain the mold component processing state prediction network.

[0043] Specifically, time series information refers to the arrangement of processing data in chronological order, reflecting the dynamic changes in the processing process. Processing state identification is a label or classification of a specific state in the processing process, such as "surface roughness is qualified" or "cutting vibration is too large". The identification can be used to guide the model to learn the relationship between processing state and parameters. The back propagation algorithm is an optimization algorithm used in machine learning to adjust model parameters. It calculates the error (loss) between the predicted result and the true value, and back-propagates the gradient to update the model parameters, so that its performance is gradually optimized.

[0044] First, collect the historical processing data set of mold parts, including processing parameters, processing status and processing results. These data contain noise or outliers, and need to be cleaned, such as deleting abnormal data during equipment failure. Then, normalize the data of each dimension and convert the values ​​to a unified range (such as 0 to 1) so that the model can process more efficiently.

[0045] Then, the standardized data is arranged and integrated in chronological order to form a time series processing sample set; at the same time, the corresponding status identifier (such as "qualified" or "unqualified") is marked according to the processing results. Next, the processing sample set is trained using a recurrent neural network. Through deep learning frameworks such as PyTorch or TensorFlow, an RNN model is constructed, with the input being the time series processing sample and the output being the predicted processing state. During the training process, the error between the model's predicted value and the actual value is calculated, and the mean square error (MSE) is used as the loss function.

[0046] Finally, the back propagation algorithm is used to optimize the model parameters, and the loss value is gradually reduced through the gradient descent method, thereby improving the accuracy of model prediction. After the training is completed, the final mold component processing state prediction network is obtained. The network can predict the processing state based on the input processing parameters and provide optimization suggestions.

[0047] Through this step, the generated processing state prediction network significantly improves the controllability and quality prediction ability of the processing process. The network can predict abnormal conditions that occur during the processing process in real time and provide a basis for parameter adjustment, reducing the processing defect rate.

[0048] Further, such as Figure 2 As shown, the determination of the target grinder processing parameters includes: using a classification algorithm to retrieve, analyze and match the manufacturing design information and the mold component manufacturing data set to obtain matching grinder processing parameters; determining matching parameter manufacturing information based on the matching grinder processing parameters; using the difference between the matching parameter manufacturing information and the manufacturing design information as a manufacturing design deviation parameter; optimizing the matching grinder processing parameters based on the manufacturing design deviation parameters to determine the target grinder processing parameters.

[0049] Specifically, the classification algorithm is a machine learning algorithm that aims to classify input data into predefined categories based on its characteristics. Here, the classification algorithm is used to analyze the degree of match between the manufacturing design information and the mold component manufacturing data set, and recommend appropriate processing parameters. Matching parameter manufacturing information refers to the processing characteristic information associated with the matching grinder processing parameters, such as material properties, surface roughness, tolerance range, etc. in historical processing. It is used to evaluate the gap between the current design requirements and the matching parameters. The manufacturing design deviation parameters are a set of parameters obtained by comparing the difference between the current manufacturing design information and the matching parameter manufacturing information, which represents the deviation between the two in terms of size, accuracy, material, etc. The target grinder processing parameters are the final processing parameters that can meet the current design requirements after optimization and adjustment, and are used to guide the actual grinder processing.

[0050] First, obtain the manufacturing design information of the target mold parts and the mold parts manufacturing data set of historical processing records. Then, use a classification algorithm (such as KNN or random forest) to analyze the manufacturing design information, retrieve similar historical processing cases in the mold parts manufacturing data set, and recommend preliminary matching grinder processing parameters. Next, extract the matching parameter manufacturing information associated with the matching processing parameters, such as the surface roughness and dimensional deviation achieved in the historical processing. Compare the matching parameter manufacturing information with the current manufacturing design information, calculate the difference and generate the manufacturing design deviation parameters. Finally, optimize the matching grinder processing parameters according to the manufacturing design deviation parameters. Use an optimization algorithm (such as particle swarm optimization or genetic algorithm) to adjust the preliminary matching parameters to make them more in line with the current design requirements. After the optimization is completed, the determined parameters are the target grinder processing parameters.

[0051] By using classification algorithms for retrieval and matching, and optimizing and adjusting the processing parameters in combination with manufacturing design deviation parameters, the target processing parameters can be recommended quickly and accurately, reducing parameter debugging time and improving processing efficiency and quality.

[0052] Furthermore, the determination of the target grinder processing parameters includes: extracting elements from the manufacturing design information to obtain a set of key manufacturing design elements; using the grinder processing parameters as independent variables and each key element in the set of key manufacturing design elements as dependent variables; performing multi-dimensional fitting on the independent variables and the dependent variables based on the mold component manufacturing data set to generate a processing parameter-manufacturing design association model; and utilizing the processing parameter-manufacturing design association model to optimize the matching grinder processing parameters based on the manufacturing design deviation parameters to determine the target grinder processing parameters.

[0053] Specifically, the manufacturing design key element set refers to the core parameters extracted from the manufacturing design information that play a decisive role in the processing process. The grinding machine processing parameters are adjustable variables in the actual processing process, including grinding wheel speed, feed speed, cutting depth, etc. These parameters directly affect the processing accuracy, surface quality and efficiency. The independent variable is the input parameter that can be controlled (such as processing parameters), and the dependent variable is the output result affected by the input parameter (such as surface roughness or processing time). In this step, the independent variable is the grinding machine processing parameter, and the dependent variable is each element in the manufacturing design key element set. Multidimensional fitting is a mathematical modeling method used to represent the complex nonlinear relationship between independent variables and dependent variables. Through fitting, an association model is generated to predict the impact of different processing parameter combinations on the manufacturing design key elements. The processing parameter-manufacturing design association model is a mathematical model established based on historical data and fitting analysis to describe the relationship between processing parameters and manufacturing design requirements. For example, it can predict the impact of different grinding wheel speeds on surface roughness. The manufacturing design deviation parameter refers to the difference between the current matching parameters and the manufacturing design requirements, and is an important basis for optimizing the processing parameters.

[0054] First, extract key elements from manufacturing design information and construct a set of key elements of manufacturing design. Next, define the machining parameters of the grinder as independent variables and the key elements of manufacturing design as dependent variables. By analyzing historical machining data sets, multidimensional fitting is performed using regression analysis or machine learning methods to generate a machining parameter-manufacturing design association model. In the generated association model, input the current manufacturing design deviation parameters, optimize the dependent variables by adjusting the independent variables, and determine the target machining parameters that meet the design requirements.

[0055] By extracting the key elements of manufacturing design and establishing a processing parameter-manufacturing design association model, an efficient mapping relationship between processing parameters and design requirements is achieved. Based on the optimization adjustment of deviation parameters, the optimal target processing parameters can be quickly recommended, significantly improving processing efficiency and quality. At the same time, using historical data for multi-dimensional fitting not only improves the accuracy of parameter matching, but also enables the association model to have a strong generalization ability to adapt to different mold design requirements.

[0056] Furthermore, the output of the optimized processing parameters of the grinder includes: taking the deviation value between the predicted manufacturing information of the mold component and the manufacturing design information as the manufacturing processing information to be optimized; performing an optimization strategy analysis on the target grinder processing parameters based on the manufacturing processing information to be optimized to obtain a processing parameter optimization strategy; using the processing parameter optimization strategy to adaptively optimize the target grinder processing parameters, and outputting the optimized processing parameters of the grinder.

[0057] Specifically, the deviation value of manufacturing design information refers to the difference between the predicted manufacturing information and the manufacturing design information, such as the predicted value of surface roughness is higher than the design requirements, the dimensional deviation exceeds the tolerance range, etc. The deviation value indicates the insufficiency between the processing parameters and the design requirements. The manufacturing processing information to be optimized is a data set that integrates the predicted manufacturing information of mold parts and the deviation value of manufacturing design information, which serves as the input for further optimizing the processing parameters. The processing parameter optimization strategy is a specific operation plan generated based on the optimization analysis, which is used to guide the adjustment of the target processing parameters. The optimized processing parameters of the grinder are the final processing parameter set obtained after applying the optimization strategy, which is used to guide the actual processing process and ensure the processing quality and efficiency.

[0058] First, obtain the deviation values ​​of mold component prediction manufacturing information and manufacturing design information. Then, input these deviation values ​​into the optimization strategy analysis module as the manufacturing processing information to be optimized. Use data analysis methods (such as principal component analysis and sensitivity analysis) to identify key deviation items and their correlation with processing parameters. For example, surface roughness deviation is mainly affected by grinding wheel speed and cutting depth, and dimensional deviation is related to feed speed.

[0059] On this basis, the optimization strategy of machining parameters is generated based on the optimization algorithm. The optimization goal is set to minimize the deviation value, while constraining other indicators (such as machining efficiency and machining time) to remain within a reasonable range. For example, increase the grinding wheel speed by 10% to reduce the surface roughness deviation; reduce the cutting depth by 0.1mm to correct the dimensional deviation. Apply the generated optimization strategy to adjust the target machining parameters and output the final grinder optimized machining parameters.

[0060] Through this step, the use of predicted manufacturing information and design deviations for optimization can significantly improve the adaptability of processing parameters and the accuracy of processing results. The optimized processing parameters not only meet the design requirements, but also find the best balance between efficiency and quality.

[0061] Furthermore, the output grinder optimized processing parameters include: using the processing parameter optimization strategy to adaptively optimize the target grinder processing parameters to obtain a processing parameter optimization threshold; randomly selecting multiple processing optimization parameters within the processing parameter optimization threshold, and using the mold component digital twin model to globally optimize the multiple processing optimization parameters to obtain the grinder optimized processing parameters.

[0062] Specifically, the processing parameter optimization threshold refers to the effective range set for the processing parameters after adaptive optimization, which is used to constrain the parameter adjustment range. Randomly selecting multiple processing optimization parameters means randomly generating several groups of processing parameter combinations within the optimization threshold range as candidate parameters for further optimization.

[0063] First, the target grinding machine processing parameters are adaptively optimized based on the processing parameter optimization strategy, and the optimization threshold of each parameter is set. For example, by analyzing the prediction results and deviations, the threshold of the grinding wheel speed is set to 3100-3300RPM, the cutting depth is set to 0.4-0.6mm, and the feed speed is set to 700-800mm / min.

[0064] Next, several groups of machining parameter combinations are randomly generated within the optimization threshold range. These candidate parameters can be generated by Monte Carlo method or Latin hypercube sampling. For example, 5 groups of parameter combinations are randomly generated within the above range, such as: Group 1: grinding wheel speed 3150RPM, cutting depth 0.45mm, feed speed 750mm / min; Group 2: grinding wheel speed 3200RPM, cutting depth 0.5mm, feed speed 780mm / min; Group 3: grinding wheel speed 3250RPM, cutting depth 0.4mm, feed speed 700mm / min.

[0065] Subsequently, the digital twin model of the mold component is used to simulate and predict each set of candidate parameters and evaluate their processing results (such as surface roughness, processing time, and dimensional deviation). The performance indicators of each set of parameters are obtained through simulation calculation as input for subsequent global optimization. A global optimization algorithm (such as genetic algorithm or particle swarm optimization) is applied to find the optimal solution among the candidate parameters. For example, surface roughness, processing time, and dimensional deviation are used as weighted combinations of the objective function, and a set of parameters with the best performance indicators is selected after optimization. Finally, the optimized processing parameters of the grinder after global optimization are output to guide the actual processing operation.

[0066] By randomly selecting parameters within the optimization threshold range and using the digital twin model for global optimization, this method ensures that the selection of processing parameters is more scientific and reasonable, avoiding the problem of local optimality. The final output of the grinder's optimized processing parameters can simultaneously meet multiple requirements such as surface quality, dimensional accuracy and processing efficiency, and improve the stability and consistency of the processing results.

[0067] Furthermore, obtaining the optimized processing parameters of the grinder includes: using the digital twin model of the mold component to perform processing prediction on the multiple processing optimization parameters to obtain multiple parameter prediction manufacturing information; performing cross-mutation expansion on the multiple processing optimization parameters based on the multiple parameter prediction manufacturing information to obtain a processing parameter update population; performing global parameter optimization within the processing parameter update population to obtain the optimized processing parameters of the grinder.

[0068] Specifically, parameter prediction manufacturing information is the processing result data obtained by simulating and predicting the processing optimization parameter combination through the digital twin model of the mold component, such as surface roughness, dimensional accuracy, processing time, etc. Crossover mutation expansion is an operation in the optimization algorithm that aims to generate new candidate parameters (called "population") through crossover and random mutation of parameter combinations to expand the search space and improve the optimization effect. The processing parameter update population is a new set of processing parameters generated by the crossover mutation operation, in which each set of parameters is an optimization candidate.

[0069] First, the digital twin model of the mold component is used to simulate and predict the multiple processing optimization parameter combinations initially generated to obtain the corresponding parameter prediction manufacturing information. Then, based on the predicted information, a new processing parameter update population is generated through crossover mutation expansion. For example: Crossover operation: Partially exchange two sets of parameter combinations (A: grinding wheel speed 3200RPM, cutting depth 0.5mm; B: grinding wheel speed 3100RPM, cutting depth 0.6mm) to generate a new parameter combination (C: grinding wheel speed 3200RPM, cutting depth 0.6mm). Mutation operation: Introduce a small disturbance to the randomly selected parameters, such as changing the cutting depth from 0.5mm to 0.55mm and the feed speed from 750mm / min to 760mm / min. The generated new population is simulated and predicted again to evaluate its manufacturing information (such as surface roughness, processing time, dimensional accuracy), and the entire updated population is globally optimized in combination with an optimization algorithm (such as genetic algorithm or particle swarm optimization). The optimization goal can be to minimize surface roughness, maximize processing efficiency, or a weighted combination of the two. Finally, the optimized machining parameters of the grinder after global optimization are output.

[0070] The candidate parameters are simulated and predicted through the digital twin model of mold components, and a new population is generated by combining cross-mutation. Global optimization ensures the comprehensiveness and optimality of processing parameter selection.

[0071] In summary, the surface grinder adaptive control optimization method for mold component manufacturing provided by the embodiment of the present application has the following technical effects: 1. By constructing an adaptive control optimization method for surface grinders, a closed-loop optimization of the entire process from mold design information to processing parameters is achieved. Combining digital twin technology, real-time monitoring, and processing parameter optimization, the processing strategy can be dynamically adjusted to ensure that the processing results meet the design requirements. This method significantly improves the accuracy, efficiency, and stability of mold processing, reduces trial and error and resource waste, and provides a scientific basis for intelligent mold manufacturing.

[0072] 2. By extracting the key elements of manufacturing design and building a processing parameter-manufacturing design association model, the precise matching between processing parameters and design requirements is achieved. Multidimensional fitting modeling effectively explores the complex influence relationship between processing parameters and design goals, providing a scientific basis for parameter optimization. This technology improves the scientificity and adaptability of the processing plan, and significantly improves the precision and quality of complex mold manufacturing.

[0073] 3. Use the digital twin model to simulate and predict candidate parameters, and combine cross-mutation to expand the population, and finally obtain the optimal combination of processing parameters through global optimization. This technology greatly improves the efficiency and accuracy of processing parameter optimization, ensuring that the processing results meet the design requirements. The global optimization method further reduces the cost of trial and error and significantly improves the level of intelligent mold manufacturing.

[0074] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0075] Furthermore, the first or second mentioned above not only represents an order relationship, but also represents a specific concept, and / or refers to the selection of multiple elements individually or in whole. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application intends to include these modifications and variations.

Claims

1. The adaptive control optimization method of surface grinder for mold parts manufacturing is characterized by: The method comprises: Obtaining manufacturing design information of target mold components, using digital twin technology to perform processing prediction modeling on the manufacturing design information, and generating a digital twin model of the mold components; Acquire a mold component manufacturing data set of a surface grinder, perform retrieval and matching based on the manufacturing design information and the mold component manufacturing data set, and determine a target grinder processing parameter; Based on the target grinding machine processing parameters, the target mold component is processed and produced and monitored in real time to obtain real-time data of mold component processing; Using the mold component digital twin model to perform processing prediction on the real-time processing data of the mold component to obtain mold component prediction manufacturing information; Based on the predicted manufacturing information of the mold component, the target grinder processing parameters are adaptively optimized, the grinder optimized processing parameters are output, and the target mold component is adaptively optimized and controlled by the grinder optimized processing parameters.

2. The surface grinder adaptive control optimization method for mold parts manufacturing according to claim 1, characterized in that: The generating of the digital twin model of the mold component includes: According to the manufacturing design information, the mold component structure design information and the mold component processing requirement attribute information are obtained; Based on the mold component structure design information and the mold component processing requirement attribute information, three-dimensional modeling and attribute assignment are performed to generate a three-dimensional solid model of the mold component; Acquire a historical processing data set of mold parts, perform processing prediction training on the historical processing data set of mold parts using a recurrent neural network structure, and obtain a mold part processing state prediction network; The digital twin technology is used to perform twin fusion modeling on the three-dimensional solid model of the mold component and the mold component processing state prediction network to generate the digital twin model of the mold component.

3. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 2, characterized in that: The method of obtaining a mold component processing state prediction network includes: Performing data cleaning and normalization processing on the mold component historical processing data set to obtain a standard mold component historical processing data set; Arrange and integrate the historical processing data set of the standard mold parts according to the time sequence information and mark the processing status to obtain the mold parts time sequence processing sample set; Using a recurrent neural network structure to perform processing prediction training and loss assessment calculation on the mold component time series processing sample set, to obtain an initial processing state prediction network and prediction network loss data; The back propagation algorithm is used to optimize and update the model parameters of the initial processing state prediction network based on the prediction network loss data to obtain the mold component processing state prediction network.

4. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 1, characterized in that: Determining the target grinding machine processing parameters includes: Using a classification algorithm to perform retrieval, analysis and matching based on the manufacturing design information and the mold component manufacturing data set to obtain matching grinding machine processing parameters; Determining matching parameter manufacturing information according to the matching grinding machine processing parameters; Using the difference between the matching parameter manufacturing information and the manufacturing design information as a manufacturing design deviation parameter; The matching grinding machine processing parameters are optimized based on the manufacturing design deviation parameters to determine target grinding machine processing parameters.

5. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 4, characterized in that: Determining the target grinding machine processing parameters includes: Extracting elements from the manufacturing design information to obtain a set of key elements of manufacturing design; The grinding machine processing parameters are used as independent variables, and each key element in the manufacturing design key element set is used as a dependent variable; Perform multi-dimensional fitting on the independent variables and the dependent variables based on the mold component manufacturing data set to generate a processing parameter-manufacturing design association model; The matching grinding machine processing parameters are optimized based on the manufacturing design deviation parameters using the processing parameter-manufacturing design association model to determine the target grinding machine processing parameters.

6. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 1, characterized in that: The output grinder optimizes processing parameters, including: Using the deviation between the predicted manufacturing information of the mold component and the manufacturing design information as the manufacturing processing information to be optimized; Performing an optimization strategy analysis on the target grinding machine processing parameters based on the manufacturing processing information to be optimized to obtain a processing parameter optimization strategy; The target grinding machine machining parameters are adaptively optimized using the machining parameter optimization strategy, and the optimized grinding machine machining parameters are output.

7. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 6, characterized in that: The output grinder optimizes processing parameters, including: Adopting the processing parameter optimization strategy to adaptively optimize the target grinding machine processing parameters to obtain a processing parameter optimization threshold; A plurality of processing optimization parameters are randomly selected within the processing parameter optimization threshold, and the mold component digital twin model is used to perform global optimization on the plurality of processing optimization parameters to obtain optimized processing parameters of the grinder.

8. The surface grinder adaptive control optimization method for mold component manufacturing according to claim 7, characterized in that: The method of obtaining the optimized processing parameters of the grinding machine comprises: Using the mold component digital twin model to perform processing prediction on the multiple processing optimization parameters to obtain multiple parameter prediction manufacturing information; Based on the plurality of parameter prediction manufacturing information, the plurality of processing optimization parameters are cross-mutated and expanded to obtain a processing parameter update population; A global parameter optimization is performed within the processing parameter update population to obtain the optimized processing parameters of the grinding machine.

Images