An intelligent parameter recommendation method and system for a precision injection mold production process

Abstract

The present invention relates to the technical field of injection molds, and specifically to an intelligent parameter recommendation method and system for the production process of precision injection molds, including: constructing and training a mold deformation prediction model based on the historical heat treatment data, mold geometric feature data, material property data, environmental factor data of the precision injection mold, and the real-time mold deformation amount during the injection mold production process to predict the deformation amount of the precision injection mold during the heat treatment process, and obtaining the predicted mold deformation amount; calculating a first compensation coefficient by combining the historical heat treatment data, mold geometric feature data, material property data, environmental factor data, and the predicted mold deformation amount; optimizing the first compensation coefficient by comparing the real-time mold deformation amount and the predicted mold deformation amount to obtain a second compensation coefficient; optimizing the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and performing parameter recommendation through the optimized intelligent parameter recommendation scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of injection molds, and in particular to an intelligent parameter recommendation method and system for a precision injection mold production process. Background Art

[0002] During the production process of precision injection molds, the dimensional accuracy of the mold is one of the key factors affecting the quality of the final product. However, during the heat treatment process, the mold often produces deformation, warping, and dimensional errors due to factors such as uneven temperature fields, differences in cooling rates, and material properties. At the same time, since precision injection molds require extremely high dimensional accuracy and stability, any slight deformation may seriously affect the quality and performance of the final product. With the continuous development of precision injection mold technology, how to effectively control the deformation and dimensional errors of precision injection molds during the production process has become a key issue.

[0003] The existing precision injection mold production process mainly relies on traditional experience and manual adjustment of process parameters. This method is not only difficult to accurately control the deformation and dimensional error of the mold, but is also easily affected by human factors, resulting in greater difficulty in controlling mold deformation, and usually requires repeated adjustments, making it difficult to ensure the stability and consistency of the production process. Therefore, there is an urgent need for an intelligent parameter recommendation method to accurately control the deformation and dimensional error of precision injection molds and ensure the stability and reliability of the dimensional accuracy of precision injection molds.

[0004] Therefore, an intelligent parameter recommendation method and system for precision injection mold production process is proposed. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent parameter recommendation method and system for a precision injection mold production process, which is suitable for controlling the deformation and dimensional error of precision injection molds with extremely high dimensional accuracy and stability requirements. The method includes: based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation in the injection mold production process, constructing and training a mold deformation prediction model to predict the deformation of the precision injection mold during the heat treatment process to obtain the predicted mold deformation; combining the historical heat treatment data, the mold geometry feature data, the material property data, the environmental factor data and the predicted mold deformation to calculate the first compensation coefficient; optimizing the first compensation coefficient by comparing the real-time mold deformation and the predicted mold deformation to obtain the second compensation coefficient; optimizing the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and performing parameter recommendation through the optimized intelligent parameter recommendation scheme.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent parameter recommendation method for a precision injection mold production process, comprising:

[0008] Based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation in the injection mold production process, a mold deformation prediction model is constructed and trained to predict the deformation of the precision injection mold during the heat treatment process to obtain the predicted mold deformation;

[0009] Calculating a first compensation coefficient by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount;

[0010] The first compensation coefficient is optimized by comparing the real-time mold deformation amount and the predicted mold deformation amount to obtain a second compensation coefficient; the optimization process is:

[0011] Acquire an initial intelligent parameter recommendation scheme, perform intelligent parameter recommendation according to the initial intelligent parameter recommendation scheme based on the first compensation coefficient, compare the real-time mold deformation variable with the predicted mold deformation variable, and calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, perform N heat treatments on the precision injection mold and iteratively update the mold size error; determine the error downward trend, and when the determination result of the error downward trend meets the stop condition, stop the heat treatment and obtain the second compensation coefficient;

[0012] The initial intelligent parameter recommendation scheme is optimized according to the second compensation coefficient, and parameter recommendation is performed through the optimized intelligent parameter recommendation scheme.

[0013] Preferably, the historical heat treatment data includes: heating temperature, holding time, tempering method and heat treatment deformation;

[0014] The mold geometric feature data includes: mold wall thickness, mold key structure length, mold volume and mold surface area;

[0015] The material property data include: thermal expansion coefficient of different materials, quenching temperature, ambient temperature, residual stress inside the mold and material yield degree;

[0016] The environmental factor data include: furnace temperature uniformity, cooling rate and cooling method.

[0017] Preferably, the specific process of constructing and training the mold deformation prediction model to predict the deformation amount of the precision injection mold during the heat treatment process is:

[0018] Taking the historical heat treatment data, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation amount during the injection mold production process as input data;

[0019] The input data is trained by a random forest algorithm to establish a prediction model describing the mapping relationship between the mold deformation amount and the various types of input data during the heat treatment process, thereby obtaining the mold deformation prediction model;

[0020] The mold deformation prediction model is used to predict the deformation amount of the precision injection mold in real time according to the heat treatment data obtained in real time, so as to obtain the predicted mold deformation amount.

[0021] Preferably, the first compensation coefficient includes: heat treatment compensation coefficient, mold geometry compensation coefficient, material property compensation coefficient, environmental factor compensation coefficient, real-time error feedback compensation coefficient and deformation prediction error compensation coefficient; the specific formula is:

[0022] ;

[0023] in, is the first compensation coefficient; is the weight of historical heat treatment data; is the number of historical heat treatments; is the deformation of the ith heat treatment; is the nominal size of the mold; is the weight of the mold geometry feature; is the first normalization coefficient; is the mold wall thickness; is the key structural length of the mold; is the second normalization coefficient; is the mold volume; is the mold surface area; is the material property weight; is the thermal expansion coefficient of the material; is the quenching temperature; is the ambient temperature; is the maximum deformation; is the residual stress inside the mold; is the yield degree of the material; is the weight of environmental factors; For furnace temperature uniformity; is the maximum furnace temperature; is the cooling rate; is the maximum cooling rate; For cooling method; is the highest cooling mode value; is the deformation prediction error weight; is the error adjustment coefficient; To predict the mold deformation; is the target mold size.

[0024] Preferably, the real-time mold deformation amount is compared with the predicted mold deformation amount to calculate the mold size error; the specific process includes: matching the real-time mold deformation amount with the corresponding predicted mold deformation amount;

[0025] The key dimension points are extracted according to the mold geometric feature data, and the error between the real-time mold deformation amount of the key dimension points and the predicted mold deformation amount is calculated to obtain the mold dimension error.

[0026] Preferably, the error downward trend is determined, and when the determination result of the error downward trend meets the stop condition, the heat treatment is stopped to obtain the second compensation coefficient; the specific process is:

[0027] After each heat treatment, the real-time mold deformation amount of the precision injection mold is obtained, the updated mold size error is calculated, and the value of the mold size error after N consecutive heat treatments is recorded to form an error value sequence;

[0028] Determining the error decreasing trend through linear regression fitting based on the error value sequence;

[0029] When the error downward trend determination result satisfies the stop condition, the heat treatment operation is terminated, and the final mold size error value is output as the error convergence value; the stop condition is: the updated mold size error does not decrease after 5 consecutive heat treatments;

[0030] Based on the error convergence value, in combination with the predicted mold deformation amount, the real-time mold deformation amount and the first compensation coefficient, the optimized second compensation coefficient is obtained by utilizing the least squares optimization method.

[0031] Preferably, an intelligent parameter recommendation system for a precision injection mold production process comprises:

[0032] The mold deformation prediction module is used to build and train a mold deformation prediction model to predict the deformation of the precision injection mold during the heat treatment process based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation during the injection mold production process, and obtain the predicted mold deformation;

[0033] A first compensation coefficient calculation module, used for calculating a first compensation coefficient by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount;

[0034] The first compensation coefficient optimization module is used to optimize the first compensation coefficient by comparing the real-time mold deformation variable and the predicted mold deformation variable to obtain a second compensation coefficient; the optimization process is:

[0035] Acquire an initial intelligent parameter recommendation scheme, perform intelligent parameter recommendation according to the initial intelligent parameter recommendation scheme based on the first compensation coefficient, compare the real-time mold deformation variable with the predicted mold deformation variable, and calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, perform heat treatment on the precision injection mold and iteratively update the mold size error; determine the error downward trend, and when the determination result of the error downward trend meets the stop condition, stop the heat treatment and obtain the second compensation coefficient;

[0036] The parameter intelligent recommendation module is used to optimize the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and perform parameter recommendation through the optimized intelligent parameter recommendation scheme.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention proposes a mold deformation prediction model to predict the deformation of precision injection molds during heat treatment. The mold deformation prediction model is constructed by using multi-dimensional data such as historical heat treatment data, mold geometric characteristics, material properties, environmental factors, and real-time mold deformation, and a random forest algorithm is used to construct the mold deformation prediction model, thereby realizing accurate prediction of mold deformation during heat treatment. Compared with the traditional method of adjusting parameters based on a single experience or partial data, the present invention has significantly improved prediction accuracy, comprehensive data utilization rate, and process automation, and can effectively avoid dimensional deviations caused by factors such as temperature or residual stress, ensure mold processing quality and stability, and provide reliable prediction data for subsequent calculation of the first compensation coefficient.

[0039] 2. The present invention proposes a method for intelligent parameter recommendation of the production process of precision injection molding process by optimizing the compensation coefficient. By comprehensively analyzing the historical heat treatment data, mold geometry characteristics, material properties, environmental factors and predicted mold deformation, the influence of various factors on mold deformation can be accurately evaluated. This comprehensive evaluation method can effectively identify the main factors affecting mold deformation and provide a scientific basis for the subsequent optimization of the compensation coefficient. By accurately calculating the first compensation coefficient, the mold deformation in the production process can be effectively predicted and compensated, thereby improving the dimensional accuracy and consistency of the product, laying the foundation for the subsequent optimization of the compensation coefficient and intelligent parameter recommendation through the optimized compensation coefficient.

[0040] 3. The present invention proposes a method for optimizing the compensation coefficient through an error feedback mechanism, and uses error feedback and the least square method to iteratively optimize the first compensation coefficient to obtain the second compensation coefficient, thereby realizing dynamic control of the mold size error. This method determines the error downward trend through data comparison, error sequence analysis and linear regression after each heat treatment, and terminates the operation when there is no obvious decrease after 5 consecutive heat treatments, thereby ensuring the reliability and stability of parameter recommendations. Compared with traditional processes, this method greatly improves the response speed and process consistency of parameter adjustment, and accurately controls the deformation and dimensional errors of precision injection molds. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flow chart of an intelligent parameter recommendation method for a precision injection mold production process provided by an embodiment of the present invention;

[0042] Figure 2 A structural diagram of an intelligent parameter recommendation system for a precision injection mold production process provided by an embodiment of the present invention;

[0043] Figure 3 A flowchart of optimizing the first compensation coefficient provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] During the production process of precision injection molds, the dimensional accuracy of the mold is one of the key factors affecting the quality of the final product. However, during the heat treatment process, the mold often produces deformation, warping, and dimensional errors due to factors such as uneven temperature fields, differences in cooling rates, and material properties. At the same time, since precision injection molds require extremely high dimensional accuracy and stability, any slight deformation may seriously affect the quality and performance of the final product. With the continuous development of precision injection mold technology, how to effectively control the deformation and dimensional errors of precision injection molds during the production process has become a key issue.

[0046] The present invention proposes an intelligent parameter recommendation method for the production process of precision injection molds, which realizes the precise control of the deformation and dimensional error of precision injection molds. This method is applied to an intelligent parameter recommendation system for the production process of precision injection molds. For the specific method flow chart and system structure diagram, please refer to Figure 1 and Figure 2In order to illustrate that the method and system of the present invention can play a role in accurately controlling the deformation and dimensional error of a precision injection mold, the effectiveness of the present invention will be illustrated from two embodiments below.

[0047] Embodiment 1

[0048] In the embodiment of the present application, the method and system proposed in the present invention are used to describe in detail the precise control process of the deformation and dimensional error of the precision injection mold. In the embodiment of the present application, the precise control of the deformation and dimensional error of the precision injection mold is aimed at the precise control of the deformation and dimensional error of the precision injection mold A. Figure 1 and Figure 2 The content describes in detail the precise control process of deformation and dimensional error of these three different precision injection molds; among them, Figure 1 The specific process of the method proposed in the present invention includes: obtaining historical heat treatment data, mold geometric feature data, material property data, environmental factor data and real-time mold deformation of the precision injection mold during the production process of the injection mold; constructing and training a mold deformation prediction model to predict the deformation of the precision injection mold during the heat treatment process to obtain the predicted mold deformation; calculating a first compensation coefficient by combining the historical heat treatment data, mold geometric feature data, material property data, environmental factor data and the predicted mold deformation; optimizing the first compensation coefficient by comparing the real-time mold deformation and the predicted mold deformation to obtain a second compensation coefficient; optimizing the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and performing parameter recommendation through the optimized intelligent parameter recommendation scheme. Figure 2 The structure diagram of the system proposed by the present invention includes: a mold deformation prediction module, a first compensation coefficient calculation module, a first compensation coefficient optimization module and a parameter intelligent recommendation module; Figure 1 and Figure 2 The following is a description of the contents:

[0049] Obtain historical heat treatment data, mold geometry data, material property data, environmental factor data, and real-time mold deformation during the production process of precision injection molds;

[0050] The historical heat treatment data include: heating temperature, holding time, tempering method and heat treatment deformation;

[0051] The mold geometric feature data includes: mold wall thickness, mold key structure length, mold volume and mold surface area;

[0052] The material property data include: thermal expansion coefficient of different materials, quenching temperature, ambient temperature, residual stress inside the mold and material yield degree;

[0053] The environmental factor data include: furnace temperature uniformity, cooling rate and cooling method.

[0054] The embodiment of the present application clarifies the specific composition of the collected historical heat treatment data, mold geometry data, material property data and environmental factor data to ensure that the data collection is comprehensive and accurate. Various types of data complement each other, making the subsequent deformation prediction and compensation calculation more accurate and reliable.

[0055] Preferably, based on historical heat treatment data, mold geometry feature data, material property data, environmental factor data and real-time mold deformation during the injection mold production process, a mold deformation prediction model is constructed and trained to predict the deformation of the precision injection mold during the heat treatment process to obtain the predicted mold deformation; the specific process is:

[0056] Taking the historical heat treatment data, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation amount during the injection mold production process as input data;

[0057] The input data is trained by a random forest algorithm to establish a prediction model describing the mapping relationship between the mold deformation amount and the various types of input data during the heat treatment process, thereby obtaining the mold deformation prediction model;

[0058] The mold deformation prediction model is used to predict the deformation amount of the precision injection mold in real time according to the heat treatment data obtained in real time, so as to obtain the predicted mold deformation amount.

[0059] Table 1 shows the comparison results of the prediction of the deformation of precision injection molds by random forest algorithm and other algorithms.

[0060] Table 1 Comparison of shape variable prediction errors

[0061]

[0062] The embodiment of this application proposes to use the random forest algorithm to train the mold deformation prediction model to make high-precision predictions on real-time data during the heat treatment process. Through multi-factor data fusion, the model can accurately depict the mapping relationship between mold deformation and various input parameters, ensure the reliability of real-time prediction results, and provide solid data support for subsequent compensation coefficient calculation and parameter optimization.

[0063] Preferably, the first compensation coefficient is calculated by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount;

[0064] The first compensation coefficient includes: heat treatment compensation coefficient, mold geometry compensation coefficient, material property compensation coefficient, environmental factor compensation coefficient, real-time error feedback compensation coefficient and deformation prediction error compensation coefficient; the specific formula is:

[0065] ;

[0066] in, is the first compensation coefficient; is the weight of historical heat treatment data; is the number of historical heat treatments; is the deformation of the ith heat treatment; is the nominal size of the mold; is the weight of the mold geometry feature; is the first normalization coefficient; is the mold wall thickness; is the key structural length of the mold; is the second normalization coefficient; is the mold volume; is the mold surface area; is the material property weight; is the thermal expansion coefficient of the material; is the quenching temperature; is the ambient temperature; is the maximum deformation; is the residual stress inside the mold; is the yield degree of the material; is the weight of environmental factors; For furnace temperature uniformity; is the maximum furnace temperature; is the cooling rate; is the maximum cooling rate; For cooling method; is the highest cooling mode value; is the deformation prediction error weight; is the error adjustment coefficient; To predict the mold deformation; is the target mold size.

[0067] The embodiment of the present application proposes a compensation coefficient formula, which realizes the comprehensive quantification of the influence of multiple factors by introducing the weight distribution of heat treatment history data, mold geometric characteristics, material properties, environmental factors and predicted deformation. This formula can accurately reflect the comprehensive influence of various factors on mold deformation, and provides a theoretical basis for the subsequent iterative optimization of the compensation coefficient and intelligent parameter recommendation through the compensation coefficient, effectively reducing the dimensional deviation caused by improper parameter setting.

[0068] Preferably, the first compensation coefficient is optimized by comparing the real-time mold deformation amount and the predicted mold deformation amount to obtain a second compensation coefficient, referring to Figure 3 ; The optimization process is:

[0069] Acquire an initial intelligent parameter recommendation scheme, perform intelligent parameter recommendation according to the initial intelligent parameter recommendation scheme based on the first compensation coefficient, compare the real-time mold deformation variable with the predicted mold deformation variable, and calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, perform N heat treatments on the precision injection mold and iteratively update the mold size error; determine the error downward trend, and when the determination result of the error downward trend meets the stop condition, stop the heat treatment and obtain the second compensation coefficient;

[0070] The real-time mold deformation amount is compared with the predicted mold deformation amount to calculate the mold size error; the specific process includes: matching the real-time mold deformation amount with the corresponding predicted mold deformation amount;

[0071] The key dimension points are extracted according to the mold geometric feature data, and the error between the real-time mold deformation amount of the key dimension points and the predicted mold deformation amount is calculated to obtain the mold dimension error.

[0072] In this embodiment, the preset mold dimension error threshold is set in combination with expert experience.

[0073] Specifically, the real-time mold deformation amount of the precision injection mold during the heat treatment process is obtained, and the real-time mold deformation amount is matched with the corresponding predicted mold deformation amount;

[0074] Obtain measurement data of precision injection molds, select key dimension points based on the mold geometric feature data and processing requirements, and obtain real-time mold deformation of key dimension points through online measuring equipment; the selection of key dimension points includes: important feature parts and key positions that have a greater impact on mold service life and product molding accuracy;

[0075] The error between the real-time mold deformation variable of the key dimension point obtained and the predicted mold deformation variable is calculated to obtain the mold dimension error; the formula of the mold dimension error is:

[0076] ;

[0077] in, is the mold size error; is the total number of key size points; For the Real-time mold deformation of key dimension points; For the The predicted mold deformation at the key dimension points.

[0078] The embodiment of the present application achieves quantitative analysis of mold size error by comparing the real-time mold deformation variable with the predicted mold deformation variable, extracting key size points and calculating errors. This method can not only accurately reflect the current process status, but also provide real-time data support for iterative optimization of compensation coefficients, thereby making process control more accurate and reducing quality risks caused by error accumulation.

[0079] Preferably, the error downward trend is determined, and when the determination result of the error downward trend meets the stop condition, the heat treatment is stopped to obtain the second compensation coefficient; the specific process is:

[0080] After each heat treatment, the real-time mold deformation amount of the precision injection mold is obtained, the updated mold size error is calculated, and the value of the mold size error after N consecutive heat treatments is recorded to form an error value sequence;

[0081] Determining the error decreasing trend through linear regression fitting based on the error value sequence;

[0082] When the error downward trend determination result satisfies the stop condition, the heat treatment operation is terminated, and the final mold size error value is output as the error convergence value; the stop condition is: linear regression fitting is performed on the most recent 5 error values, and if the fitting slope is close to 0 or positive, it is determined that the error no longer decreases, and the heat treatment operation is terminated;

[0083] Based on the error convergence value, combined with the predicted mold deformation, the real-time mold deformation and the first compensation coefficient, the optimized second compensation coefficient is obtained by using the least squares optimization method, see Table 2.

[0084] Table 2 Compensation coefficient iterative optimization data table

[0085]

[0086] The embodiment of the present application adopts the convergence analysis method of the mold size error after continuous heat treatment, determines the error downward trend through linear regression, and stops the process when the error no longer decreases after 5 consecutive heat treatments. This method realizes the dynamic adjustment and optimization control of the compensation coefficient, ensures that the parameter recommendation reaches the best state, improves the stability and reliability of the production process, and provides reliable data for the subsequent intelligent recommendation of parameters through the compensation coefficient.

[0087] The initial intelligent parameter recommendation scheme is optimized according to the second compensation coefficient, and parameter recommendation is performed through the optimized intelligent parameter recommendation scheme.

[0088] The embodiment of the present application provides an intelligent parameter recommendation method for a precision injection mold production process, which improves the manufacturing accuracy of precision injection molds, reduces the number of heat treatments, and improves production consistency through mold deformation prediction, compensation coefficient optimization, and intelligent parameter recommendation. First, a mold deformation prediction model is constructed using a random forest algorithm, and the deformation prediction accuracy is improved by combining historical heat treatment data, mold geometry, material properties, and real-time deformation variables, so that the dimensional error is reduced from 0.55mm-0.72mm to 0.20mm-0.30mm, and the accuracy is improved by about 75%. Secondly, by calculating the first compensation coefficient and optimizing it as the second compensation coefficient in combination with real-time error feedback, unnecessary heat treatment iterations are reduced. Compared with the traditional fixed compensation method, the number of heat treatments is reduced from 10-13 times to 5-7 times, and energy consumption is reduced by about 45%. In addition, the compensation strategy is optimized by combining the least squares method to make the final mold more accurate. Finally, through the optimized intelligent parameter recommendation scheme, the dimensional error of the same batch of molds is reduced by 63%, production consistency is improved, and the final finished product qualification rate is increased from 85.2% to 96.5%. The present invention comprehensively improves the production process level of precision injection molds through accurate deformation prediction, intelligent compensation optimization and efficient parameter recommendation, achieves higher molding accuracy, lower energy consumption and better production consistency, and provides an efficient and intelligent optimization solution for the mold manufacturing industry.

[0089] Embodiment 2

[0090] In Example 1, the method proposed in the present invention successfully achieved accurate control of the deformation and dimensional error of the precision injection mold A. To further verify the effectiveness of the present invention, the present embodiment also proposed an intelligent parameter recommendation system for the production process of precision injection molds to accurately control the deformation and dimensional error of the precision injection mold B; the system includes: a mold deformation prediction module, a first compensation coefficient calculation module, a first compensation coefficient optimization module and a parameter intelligent recommendation module.

[0091] The mold deformation prediction module is used to build and train a mold deformation prediction model to predict the deformation of the precision injection mold during the heat treatment process based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation during the injection mold production process, and obtain the predicted mold deformation;

[0092] The historical heat treatment data include: heating temperature, holding time, tempering method and heat treatment deformation;

[0093] The mold geometric feature data includes: mold wall thickness, mold key structure length, mold volume and mold surface area;

[0094] The material property data include: thermal expansion coefficient of different materials, quenching temperature, ambient temperature, residual stress inside the mold and material yield degree;

[0095] The environmental factor data include: furnace temperature uniformity, cooling rate and cooling method.

[0096] Preferably, the specific process of constructing and training the mold deformation prediction model to predict the deformation amount of the precision injection mold during the heat treatment process is:

[0097] Taking the historical heat treatment data, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation amount during the injection mold production process as input data;

[0098] The input data is trained by a random forest algorithm to establish a prediction model describing the mapping relationship between the mold deformation amount and the various types of input data during the heat treatment process, thereby obtaining the mold deformation prediction model;

[0099] The mold deformation prediction model is used to predict the deformation amount of the precision injection mold in real time according to the heat treatment data obtained in real time, so as to obtain the predicted mold deformation amount.

[0100] Preferably, a first compensation coefficient calculation module is used to calculate a first compensation coefficient by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount;

[0101] The first compensation coefficient includes: heat treatment compensation coefficient, mold geometry compensation coefficient, material property compensation coefficient, environmental factor compensation coefficient, real-time error feedback compensation coefficient and deformation prediction error compensation coefficient; the specific formula is:

[0102] ;

[0103] in, is the first compensation coefficient; is the weight of historical heat treatment data; is the number of historical heat treatments; is the deformation of the ith heat treatment; is the nominal size of the mold; is the weight of the mold geometry feature; is the first normalization coefficient; is the mold wall thickness; is the key structural length of the mold; is the second normalization coefficient; is the mold volume; is the mold surface area; is the material property weight; is the thermal expansion coefficient of the material; is the quenching temperature; is the ambient temperature; is the maximum deformation; is the residual stress inside the mold; is the yield degree of the material; is the weight of environmental factors; For furnace temperature uniformity; is the maximum furnace temperature; is the cooling rate; is the maximum cooling rate; For cooling method; is the highest cooling mode value; is the deformation prediction error weight; is the error adjustment coefficient; To predict the mold deformation; is the target mold size.

[0104] Preferably, the first compensation coefficient optimization module is used to optimize the first compensation coefficient by comparing the real-time mold deformation amount and the predicted mold deformation amount to obtain a second compensation coefficient; the optimization process is:

[0105] Acquire an initial intelligent parameter recommendation scheme, perform intelligent parameter recommendation according to the initial intelligent parameter recommendation scheme based on the first compensation coefficient, compare the real-time mold deformation variable with the predicted mold deformation variable, and calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, perform heat treatment on the precision injection mold and iteratively update the mold size error; determine the error downward trend, and when the determination result of the error downward trend meets the stop condition, stop the heat treatment and obtain the second compensation coefficient;

[0106] The real-time mold deformation amount is compared with the predicted mold deformation amount to calculate the mold size error; the specific process includes:

[0107] Matching the real-time mold deformation amount with the corresponding predicted mold deformation amount;

[0108] The key dimension points are extracted according to the mold geometric feature data, and the error between the real-time mold deformation amount of the key dimension points and the predicted mold deformation amount is calculated to obtain the mold dimension error.

[0109] Preferably, the error downward trend is determined, and when the determination result of the error downward trend meets the stop condition, the heat treatment is stopped to obtain the second compensation coefficient; the specific process is:

[0110] After each heat treatment, the real-time mold deformation amount of the precision injection mold is obtained, the updated mold size error is calculated, and the value of the mold size error after N consecutive heat treatments is recorded to form an error value sequence;

[0111] Determining the error decreasing trend through linear regression fitting based on the error value sequence;

[0112] When the error downward trend determination result satisfies the stop condition, the heat treatment operation is terminated, and the final mold size error value is output as the error convergence value; the stop condition is: the updated mold size error does not decrease after 5 consecutive heat treatments;

[0113] Based on the error convergence value, combined with the predicted mold deformation, the real-time mold deformation and the first compensation coefficient, the optimized second compensation coefficient is obtained by using the least squares optimization method, see Table 3.

[0114] Table 3 Compensation coefficient iterative optimization data table

[0115]

[0116] Preferably, the parameter intelligent recommendation module is used to optimize the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and perform parameter recommendation through the optimized intelligent parameter recommendation scheme.

[0117] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent parameter recommendation method for precision injection mold production process, characterized in that: include: Based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation in the injection mold production process, a mold deformation prediction model is constructed and trained to predict the deformation of the precision injection mold during the heat treatment process to obtain the predicted mold deformation; Calculating a first compensation coefficient by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount; The first compensation coefficient is optimized by comparing the real-time mold deformation amount and the predicted mold deformation amount to obtain a second compensation coefficient; the optimization process is: Obtaining an initial intelligent parameter recommendation scheme and performing intelligent parameter recommendation in combination with the first compensation coefficient, and comparing the real-time mold deformation variable with the predicted mold deformation variable to calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, performing N heat treatments on the precision injection mold and iteratively updating the mold size error; determination error downward trend, when the determination result of the error downward trend satisfies the stop condition, stopping the heat treatment and obtaining the second compensation coefficient; The initial intelligent parameter recommendation scheme is optimized according to the second compensation coefficient, and parameter recommendation is performed through the optimized intelligent parameter recommendation scheme.

2. The intelligent parameter recommendation method for a precision injection mold production process according to claim 1, characterized in that: The historical heat treatment data includes: heating temperature, holding time, tempering method and heat treatment deformation; The mold geometric feature data includes: mold wall thickness, mold key structure length, mold volume and mold surface area; The material property data include: thermal expansion coefficient of different materials, quenching temperature, ambient temperature, residual stress inside the mold and material yield degree; The environmental factor data include: furnace temperature uniformity, cooling rate and cooling method.

3. The intelligent parameter recommendation method for a precision injection mold production process according to claim 1, characterized in that: The specific process of building and training the mold deformation prediction model to predict the deformation of the precision injection mold during heat treatment is as follows: Taking the historical heat treatment data, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation amount during the injection mold production process as input data; The input data is trained by a random forest algorithm to establish a prediction model describing the mapping relationship between the mold deformation amount and different input data during the heat treatment process, thereby obtaining the mold deformation prediction model; The mold deformation prediction model is used to predict the deformation amount of the precision injection mold in real time according to the heat treatment data obtained in real time, so as to obtain the predicted mold deformation amount.

4. The intelligent parameter recommendation method for a precision injection mold production process according to claim 1, characterized in that: The first compensation coefficient includes: heat treatment compensation coefficient, mold geometry compensation coefficient, material property compensation coefficient, environmental factor compensation coefficient, real-time error feedback compensation coefficient and deformation prediction error compensation coefficient; the specific formula is: ； in, is the first compensation coefficient; is the weight of historical heat treatment data; is the number of historical heat treatments; is the deformation of the ith heat treatment; is the nominal size of the mold; is the weight of the mold geometry feature; is the first normalization coefficient; is the mold wall thickness; is the key structural length of the mold; is the second normalization coefficient; is the mold volume; is the mold surface area; is the material property weight; is the thermal expansion coefficient of the material; is the quenching temperature; is the ambient temperature; is the maximum deformation; is the residual stress inside the mold; is the yield degree of the material; is the weight of environmental factors; For furnace temperature uniformity; is the maximum furnace temperature; is the cooling rate; is the maximum cooling rate; For cooling method; is the highest cooling mode value; is the deformation prediction error weight; is the error adjustment coefficient; To predict the mold deformation; is the target mold size.

5. The intelligent parameter recommendation method for a precision injection mold production process according to claim 1, characterized in that: The real-time mold deformation amount is compared with the predicted mold deformation amount to calculate the mold size error; the specific process includes: Matching the real-time mold deformation amount with the corresponding predicted mold deformation amount; The key dimension points are extracted according to the mold geometric feature data, and the error between the real-time mold deformation amount of the key dimension points and the predicted mold deformation amount is calculated to obtain the mold dimension error.

6. The intelligent parameter recommendation method for a precision injection mold production process according to claim 1, characterized in that: Determine the downward trend of the error, and when the determination result of the downward trend of the error meets the stop condition, stop the heat treatment and obtain the second compensation coefficient; the specific process is: After each heat treatment, the real-time mold deformation amount of the precision injection mold is obtained, the updated mold size error is calculated, and the value of the mold size error after N consecutive heat treatments is recorded to form an error value sequence; Determining the error decreasing trend through linear regression fitting based on the error value sequence; When the error downward trend determination result satisfies the stop condition, the heat treatment operation is terminated, and the final mold size error value is output as the error convergence value; the stop condition is: the updated mold size error does not decrease after 5 consecutive heat treatments; Based on the error convergence value, in combination with the predicted mold deformation amount, the real-time mold deformation amount and the first compensation coefficient, the optimized second compensation coefficient is obtained by utilizing the least squares optimization method.

7. An intelligent parameter recommendation system for precision injection mold production process, characterized in that: include: The mold deformation prediction module is used to build and train a mold deformation prediction model to predict the deformation of the precision injection mold during the heat treatment process based on the historical heat treatment data of the precision injection mold, the mold geometry feature data, the material property data, the environmental factor data and the real-time mold deformation during the injection mold production process, and obtain the predicted mold deformation; A first compensation coefficient calculation module, used for calculating a first compensation coefficient by combining the historical heat treatment data, the mold geometric feature data, the material property data, the environmental factor data and the predicted mold deformation amount; The first compensation coefficient optimization module is used to optimize the first compensation coefficient by comparing the real-time mold deformation variable and the predicted mold deformation variable to obtain a second compensation coefficient; the optimization process is: Acquire an initial intelligent parameter recommendation scheme, perform intelligent parameter recommendation according to the initial intelligent parameter recommendation scheme based on the first compensation coefficient, compare the real-time mold deformation variable with the predicted mold deformation variable, and calculate the mold size error; if the mold size error is greater than and / or equal to a preset mold size error threshold, heat treat the precision injection mold and iteratively update the mold size error; determination error downward trend, when the determination result of the error downward trend satisfies the stop condition, stopping the heat treatment and obtaining the second compensation coefficient; The parameter intelligent recommendation module is used to optimize the initial intelligent parameter recommendation scheme according to the second compensation coefficient, and perform parameter recommendation through the optimized intelligent parameter recommendation scheme.

Images