Parameter optimization method for hard alloy split mold demolding process

Abstract

The present application belongs to the technical field of computer-aided demolding process, and relates to a parameter optimization method for the demolding process of a hard alloy split mold. By obtaining the geometric structure of the mold, the material properties and the working parameters of the demolding equipment, a finite element simulation model of the demolding process is constructed. A plurality of candidate parameter combinations are generated for the demolding parameters, the simulation model is called to perform transient dynamics simulation, and the demolding force change curve and the internal stress distribution field of the mold are obtained. The demolding force peak value and fluctuation amplitude, as well as the maximum stress value and stress gradient of each split module, are extracted, the demolding force fluctuation coefficient and the stress concentration coefficient are calculated, the non-dominated solution set is screened, and the comprehensive evaluation value is calculated to select the optimal parameter combination. The present application can effectively balance the demolding force stability and the mold stress uniformity, reduce the demolding impact and local stress concentration, reduce the trial and error cost, improve the mold service life and the workpiece demolding quality, and provide an efficient and reliable computer-aided optimization means for the demolding process of a hard alloy split mold.

Description

Technical Field

[0001] This invention belongs to the field of computer-aided demolding technology, and relates to a parameter optimization method for the demolding process of cemented carbide segmented molds. Background Technology

[0002] Carbide segmented dies are widely used in the pressing and forming processes of carbide products such as bars, tubes, and irregularly shaped parts. The demolding process refers to the operation of removing the pressed carbide workpiece from the die. Due to the high hardness and brittleness of carbide materials, and because segmented dies are composed of multiple circumferentially distributed segmented modules, the contact state between each segmented module and the workpiece during demolding is complex and variable. The magnitude and fluctuation of the demolding force directly affect the surface quality and dimensional accuracy of the workpiece. Simultaneously, the stress distribution within the die determines its service life and reliability. In the field of computer-aided engineering, finite element simulation technology has been widely used to simulate the mechanical behavior during demolding. However, existing simulation methods typically only analyze single parameter combinations, lacking a systematic approach to synergistically optimizing multiple demolding process parameters, making it difficult to achieve the optimal balance between demolding force stability and die stress uniformity.

[0003] Currently, the determination of process parameters for the demolding process of segmented molds mainly relies on trial and error adjustments by technicians based on experience, or on a single-factor rotation method to adjust each parameter one by one and then observe the demolding effect. The shortcomings of this approach are: there are complex coupling relationships between multiple parameters during the demolding process, such as ejection speed, demolding temperature, holding time, segmentation sequence, and ejector rod action sequence. Single-factor adjustments cannot capture the interaction effects between parameters, and the trial-and-error process is lengthy and difficult to obtain globally optimal parameter configurations. Furthermore, existing technologies use relatively singular evaluation indicators for demolding effect, typically relying solely on whether the peak demolding force exceeds the equipment's capacity, ignoring the time-domain fluctuation characteristics of the demolding force and the degree of localized stress concentration within the mold. This results in a one-sided optimization direction and fails to effectively guide parameter setting in actual production.

[0004] With the continuous development of computer simulation technology and multi-objective optimization methods, it has become possible to optimize the parameters of the demolding process using simulation models. However, some existing multi-objective optimization methods, when applied to demolding parameter optimization, often require the construction of surrogate models or extensive iterative training, resulting in high computational costs and sensitivity to the quality of initial samples, making them unsuitable for rapid application in engineering settings. Furthermore, these methods typically treat the peak demolding force and maximum stress value as separate optimization objectives, failing to fully consider the demolding stability reflected by the amplitude of demolding force fluctuations and the stress distribution uniformity characterized by stress gradients. This can lead to problems such as excessive demolding impact or localized stress exceeding limits in actual demolding operations. Therefore, there is an urgent need for a demolding parameter optimization method that can directly leverage transient dynamic simulation results while considering both the characteristics of demolding force fluctuations and the uniformity of stress distribution, providing efficient and reliable parameter decision support for the demolding process of cemented carbide segmented molds. Summary of the Invention

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] A parameter optimization method for the demolding process of a cemented carbide segmented mold includes:

[0007] S1, obtain the geometric structure parameters, hard alloy material parameters, and demolding equipment operating parameters of the target segmented mold;

[0008] S2, construct a finite element simulation model of the demolding process based on the geometric structure parameters and the material parameters, and set the demolding direction, constraint boundaries and contact conditions;

[0009] S3, obtain the set of demolding parameters to be optimized, and generate multiple candidate parameter combinations within a preset value range for each parameter in the set of demolding parameters;

[0010] S4. Using each candidate parameter combination as input conditions, the finite element simulation model is called to perform transient dynamic simulation of the demolding process, and the curve of demolding force changing with time corresponding to each candidate parameter combination and the stress distribution field inside each segment module of the mold are obtained.

[0011] S5. Based on the curve of the demolding force changing with time, extract the maximum demolding force value as the peak value of the demolding force, and extract the maximum change amplitude of the demolding force value per unit time as the demolding force fluctuation amplitude.

[0012] S6. Based on the stress distribution field, extract the maximum stress value inside each segmented module, and extract the stress change rate along the demolding direction inside each segmented module as the stress gradient.

[0013] S7. Calculate the demolding force fluctuation coefficient and stress concentration coefficient for each candidate parameter combination. With minimizing the demolding force fluctuation coefficient and stress concentration coefficient as the optimization objectives, compare the demolding force fluctuation coefficient and stress concentration coefficient of all candidate parameter combinations, and select the non-dominated solution set.

[0014] S8. From the non-dominated solution set, calculate the comprehensive evaluation value based on the demolding force fluctuation coefficient and stress concentration coefficient of each candidate parameter combination, and select the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter.

[0015] Further, S3 includes:

[0016] Candidate parameter combinations are generated within the preset value range using the Latin hypercube sampling method. The preset value range is dynamically determined based on the upper limit of ejection stroke and upper limit of ejection force in the working parameters of the demolding equipment, as well as the tilt angle of the segmented contact surface in the geometric structure parameters.

[0017] Furthermore, the contact conditions in the finite element simulation model include the variation of the friction coefficient of the contact surface between each segmented module with the demolding temperature;

[0018] During transient dynamic simulation, the friction coefficient at each time step is obtained by interpolation from the friction characteristic curve based on the current simulation temperature. Furthermore, when a relative sliding tendency occurs between the segmented module and the workpiece, the dynamic friction coefficient is used instead of the static friction coefficient.

[0019] Further, S5 includes:

[0020] The curve of the demolding force changing with time is divided into multiple continuous time intervals. Within each time interval, all local maxima and all local minima of the demolding force value are detected.

[0021] Compare the demolding force values ​​corresponding to all local maxima within all time intervals, and take the maximum value as the peak demolding force;

[0022] For each time interval, the average value of the demolding force at all local maxima is calculated as the upper peak mean of the interval, and the average value of the demolding force at all local minima in the time interval is calculated as the lower valley mean of the interval. The difference between the upper peak mean and the lower valley mean is taken as the interval fluctuation amplitude of the time interval.

[0023] The arithmetic mean of the interval fluctuation amplitudes across all time intervals is calculated as the base fluctuation value. The standard deviation of the interval fluctuation amplitudes across all time intervals is calculated as the fluctuation dispersion. The sum of the base fluctuation value and the fluctuation dispersion is taken as the demolding force fluctuation amplitude.

[0024] Further, S6 includes:

[0025] For each segmented module, the stress tensor components on all finite element nodes inside the segmented module are read from the stress distribution field, and the Mises equivalent stress value of each node is calculated based on the stress tensor components.

[0026] Compare the Mises equivalent stress values ​​of all nodes within the segmented module, and take the largest Mises equivalent stress value as the maximum stress value of the segmented module.

[0027] For each segmented module, multiple parallel feature paths are selected within the segmented module along the demolding direction, and each feature path is evenly distributed on the cross-section of the segmented module.

[0028] Multiple sampling points are selected at equal intervals on each feature path. The Mises equivalent stress value at each sampling point is read. The ratio of the difference between the Mises equivalent stress values ​​between adjacent sampling points to the distance between the sampling points is calculated as the local stress change rate of the adjacent point pair.

[0029] The root mean square of the local stress change rate of all adjacent point pairs on a feature path is taken as the path stress change rate of that feature path, and the arithmetic mean of the path stress change rates of all feature paths within that lobe module is taken as the stress gradient of that lobe module.

[0030] Furthermore, the stress concentration factor is calculated based on the maximum stress value and stress gradient of each segment module according to a preset weighting rule, specifically including:

[0031] For each segmented module, the arithmetic mean of the Mises equivalent stress values ​​of all finite element nodes within the segmented module is taken as the average stress value of the segmented module.

[0032] The ratio of the maximum stress value of the segmented module to the average stress value of the segmented module is used as the local concentration factor of the segmented module.

[0033] Calculate the local concentration coefficients of all segmented modules, the maximum value and the arithmetic mean of all local concentration coefficients, and the sample standard deviation of all local concentration coefficients.

[0034] For each segmented module, the angle between the stress gradient direction vector and the demolding direction vector of the segmented module is obtained. When the angle is less than a preset angle threshold, a direction amplification factor is applied to the local concentration coefficient of the segmented module. The local concentration coefficient of the segmented module is multiplied by the direction amplification factor to obtain the adjusted local concentration coefficient.

[0035] When the included angle is greater than or equal to the angle threshold, no direction amplification factor is applied, and the adjusted local concentration coefficient is the original local concentration coefficient.

[0036] Recalculate the maximum value, arithmetic mean, and sample standard deviation of the local concentration coefficients for all segmented modules after adjustment;

[0037] The difference between the maximum value and the arithmetic mean of the adjusted local concentration coefficients is taken as the first difference term, and the sample standard deviation of the adjusted local concentration coefficients is taken as the second difference term. The first difference term is assigned a first weight, and the second difference term is assigned a second weight. The product of the first difference term and the first weight is added to the product of the second difference term and the second weight, and the sum is taken as the basic concentration coefficient.

[0038] Next, check the difference between the adjusted local concentration coefficients of any two adjacent lobe modules in all lobe modules. If there is a difference between adjacent modules that exceeds the preset difference threshold, then take the average of all the differences between adjacent modules that exceed the difference threshold as the neighborhood difference term. Multiply the neighborhood difference term by the neighborhood weight and add it to the basic concentration coefficient. The result is used as the stress concentration coefficient of the candidate parameter combination.

[0039] Furthermore, after selecting the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter, the method further includes calling the finite element simulation model again as the input condition to perform verification simulation, thereby obtaining the verification curve of the demolding force changing with time and the verification stress distribution field.

[0040] Extract the peak value of the demolding force from the verification curve, and extract the global maximum stress value from all the segmented modules from the verification stress distribution field;

[0041] The peak value of the demolding force is compared with the upper limit of the ejection force of the demolding equipment, and the global maximum stress value is compared with the yield strength of the cemented carbide material.

[0042] If the peak value of the demolding force is less than or equal to the upper limit of the ejection force and the global maximum stress value is less than or equal to the yield strength, then the optimized output parameters are confirmed as the final effective parameters.

[0043] If the peak value of the demolding force is greater than the upper limit of the ejection force or the global maximum stress value is greater than the yield strength, then the preset value range is narrowed, candidate parameter combinations are regenerated, and subsequent steps are repeated until optimized output parameters that meet the comparison conditions are obtained.

[0044] Furthermore, the demolding parameter set includes the ejection sequence of the ejector rods and the time interval between adjacent ejection actions. The ejection sequence specifies the starting order of the ejector rods corresponding to each segment module during the demolding process, and the time interval specifies the waiting time between adjacent ejector rods.

[0045] When generating candidate parameter combinations, the ejection order is determined by the full permutation of the ejection rod numbers corresponding to all segment modules, and the time interval is determined by the minimum response time and maximum allowable interval of the demolding equipment.

[0046] Furthermore, the finite element simulation model uses an explicit dynamics solver for transient dynamics simulation. The explicit dynamics solver uses the central difference method to discretize the time domain and updates the node velocity and displacement according to the current node acceleration in each time step, while calculating the contact force according to the contact conditions.

[0047] During the simulation, a mass scaling factor is set to control the settling time step. The mass scaling factor is determined based on the minimum element size and the material wave velocity.

[0048] A contact penetration tolerance is set. When the penetration amount between the segmented module and the workpiece exceeds the contact penetration tolerance, a penalty force is applied to the corresponding contact node to correct the penetration.

[0049] This invention belongs to the field of computer-aided demolding technology, and relates to a parameter optimization method for the demolding process of cemented carbide segmented molds. By acquiring the mold geometry, material properties, and demolding equipment operating parameters, a finite element simulation model of the demolding process is constructed. Multiple candidate parameter combinations are generated for the demolding parameters, and the simulation model is called to perform transient dynamic simulation to obtain the demolding force variation curve and the internal stress distribution field of the mold. The peak value and fluctuation amplitude of the demolding force, as well as the maximum stress value and stress gradient of each segmented module, are extracted, and the demolding force fluctuation coefficient and stress concentration coefficient are calculated. Non-dominated solution sets are screened, and a comprehensive evaluation value is calculated to select the optimal parameter combination. This invention can effectively balance the stability of the demolding force and the uniformity of mold stress, reduce demolding impact and local stress concentration, reduce trial and error costs, improve mold life and workpiece demolding quality, and provide an efficient and reliable computer-aided optimization method for the demolding process of cemented carbide segmented molds. Attached Figure Description

[0050] Figure 1 A flowchart illustrating a parameter optimization method for the demolding process of a cemented carbide segmented mold, as claimed in an embodiment of the present invention.

[0051] Figure 2 The second flowchart is shown for a parameter optimization method for the demolding process of a cemented carbide segmented mold, as claimed in an embodiment of the present invention.

[0052] Figure 3 The third flowchart is a parameter optimization method for the demolding process of a cemented carbide segmented mold, as claimed in an embodiment of the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0054] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of those features. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications in the embodiments of this application, such as up, down, left, right, front, back, etc., are only used to explain the relative positional relationships and movements between components in a specific orientation as shown in the accompanying drawings. If the specific orientation changes, the directional indications will change accordingly. Furthermore, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0055] References to embodiments herein mean that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0056] According to a first embodiment of the present invention, the present invention claims protection for a parameter optimization method for the demolding process of a cemented carbide segmented mold, referring to... Figure 1 ,include:

[0057] S1, obtain the geometric structure parameters, hard alloy material parameters, and demolding equipment operating parameters of the target segmented mold;

[0058] S2, construct a finite element simulation model of the demolding process based on the geometric structure parameters and the material parameters, and set the demolding direction, constraint boundaries and contact conditions;

[0059] S3, obtain the set of demolding parameters to be optimized, and generate multiple candidate parameter combinations within a preset value range for each parameter in the set of demolding parameters;

[0060] S4. Using each candidate parameter combination as input conditions, the finite element simulation model is called to perform transient dynamic simulation of the demolding process, and the curve of demolding force changing with time corresponding to each candidate parameter combination and the stress distribution field inside each segment module of the mold are obtained.

[0061] S5. Based on the curve of the demolding force changing with time, extract the maximum demolding force value as the peak value of the demolding force, and extract the maximum change amplitude of the demolding force value per unit time as the demolding force fluctuation amplitude.

[0062] S6. Based on the stress distribution field, extract the maximum stress value inside each segmented module, and extract the stress change rate along the demolding direction inside each segmented module as the stress gradient.

[0063] S7. Calculate the demolding force fluctuation coefficient and stress concentration coefficient for each candidate parameter combination. With minimizing the demolding force fluctuation coefficient and stress concentration coefficient as the optimization objectives, compare the demolding force fluctuation coefficient and stress concentration coefficient of all candidate parameter combinations, and select the non-dominated solution set.

[0064] S8. From the non-dominated solution set, calculate the comprehensive evaluation value based on the demolding force fluctuation coefficient and stress concentration coefficient of each candidate parameter combination, and select the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter.

[0065] In this embodiment, a six-segment mold used in a cemented carbide product manufacturing workshop is taken as the implementation object. The mold is composed of six segmented modules evenly distributed along the circumference. The radial dimension of each segmented module is the distance from the central axis of the mold to the outer wall of the module, and the axial dimension is the length along the demolding direction, i.e., the direction of the central axis of the mold. The inclination angle of the segmented contact surface is the angle between the inclined surface of two adjacent segmented modules in contact with each other and the radial direction of the mold. Regarding the cemented carbide material parameters, the cemented carbide grade used in this mold is YG15. Its measured elastic modulus value is determined according to the material inspection report. Poisson's ratio is obtained by tensile testing on standard specimens. The coefficient of thermal expansion is obtained by measuring the thermal deformation of the specimens within the demolding temperature range. The demolding equipment is a four-column hydraulic press. Its upper limit of ejection stroke is determined by the maximum displacement of the piston of the hydraulic cylinder, and the upper limit of ejection force is determined by the product of the rated working pressure of the hydraulic system and the piston area.

[0066] The operator first inputs the above-mentioned geometric parameters, material parameters, and working parameters into the preprocessing module of the computer-aided engineering software. Then, the operator constructs a finite element simulation model of the demolding process in the software: the demolding direction is set to the direction extending from the inside of the workpiece outward along the mold axis, the fixed connection between the outer cylindrical surface of the mold and the worktable of the equipment is set as the constraint boundary, the contact surface between each segment module is set as the friction contact type and assigned a corresponding friction coefficient value, and the contact between the inner surface of the mold segment module and the outer surface of the cemented carbide workpiece is set as the adhesive contact type and assigned an adhesive strength value.

[0067] Next, the operator specifies the set of demolding parameters to be optimized, including ejection speed, demolding temperature, holding time, and segmentation sequence. Ejection speed refers to the rate at which the hydraulic press ejector rod pushes the segmentation modules upward. Demolding temperature refers to the temperature value to which the entire mold is heated before the demolding operation. Holding time refers to the length of time the demolding temperature is maintained before the ejection action begins. Segmentation sequence refers to the order in which the six segmentation modules separate from the workpiece during the demolding process. For each of the above parameters, the operator sets its own value range based on equipment capabilities and process experience, and then uses a comprehensive combination method to generate a large number of candidate parameter combinations.

[0068] For each candidate parameter combination, the operator inputs the specific values ​​of ejection speed, demolding temperature, holding time, and segmentation sequence as boundary and initial conditions into the established finite element simulation model and starts the transient dynamics solution calculation. After the solution is completed, the software outputs the curve data of the demolding force changing with time corresponding to the combination, as well as the stress distribution field data inside each segment module of the mold.

[0069] The operator reads the maximum demolding force value on the entire curve from the demolding force curve data as the peak value of demolding force, and then reads the maximum value of the difference between demolding force values ​​between adjacent sampling points on the curve as the demolding force fluctuation amplitude; from the stress distribution field data, the operator reads the stress values ​​on all nodes inside each segment module, finds the maximum stress value in each module, and calculates the rate of change of stress values ​​between adjacent nodes along the demolding direction in each module as the stress gradient.

[0070] Then, the operator divides the amplitude of the demolding force fluctuation by the peak value of the demolding force to obtain the demolding force fluctuation coefficient; for the stress concentration coefficient, the operator calculates the maximum stress value and stress gradient of each segment module according to a pre-set weighting rule; the weighting rule is: first, assign a weight factor proportional to the stress gradient of the module to the maximum stress value of each module, and then sum the weighted maximum stress values ​​of each module and divide by the sum of the weight factors.

[0071] The operators used minimizing the demolding force fluctuation coefficient and the stress concentration coefficient as two optimization objectives, and adopted an optimization strategy based on Pareto dominance to screen all candidate parameter combinations. Specifically, they compared the two objective values ​​of any two candidate parameter combinations one by one. If the two objective values ​​of one combination are not inferior to the other combination and at least one objective value is strictly superior to the other combination, then the former is considered to dominate the latter. All dominated combinations were deleted, and the remaining combinations constituted the non-dominated solution set.

[0072] Finally, the operator multiplies the demolding force fluctuation coefficient and stress concentration coefficient of each candidate parameter combination in the non-dominated solution set by their respective weighting coefficients and then adds them together to obtain the comprehensive evaluation value of the combination. The operator compares all the comprehensive evaluation values ​​and selects the candidate parameter combination with the smallest value as the final optimized output parameter for use in actual demolding production.

[0073] Further, S3 includes:

[0074] Candidate parameter combinations are generated within the preset value range using the Latin hypercube sampling method. The preset value range is dynamically determined based on the upper limit of ejection stroke and upper limit of ejection force in the working parameters of the demolding equipment, as well as the tilt angle of the segmented contact surface in the geometric structure parameters.

[0075] In this embodiment, when the operator generates candidate parameter combinations for the four parameters in the demolding parameter set—ejection speed, demolding temperature, holding time, and segmentation sequence—the Latin hypercube sampling method is used. Specifically, the operator first determines the preset value range for each parameter. For ejection speed, its upper limit is not a fixed value, but is dynamically calculated based on the upper limit of the hydraulic press's ejection force and the total friction area between the mold segment modules and the workpiece. When the ejection speed exceeds a certain value, the required demolding force will exceed the upper limit of the equipment's ejection force; this value is the upper limit of the ejection speed. For demolding temperature, its lower limit is determined by the phase transformation temperature of the cemented carbide material. If the demolding temperature is lower than this phase transformation temperature, the microstructure of the cemented carbide will undergo an unfavorable transformation; therefore, this phase transformation temperature is used as the lower limit of the demolding temperature. For holding time, its value range is set according to the working cycle time of the hydraulic press; that is, the holding time cannot be shorter than the shortest time required for the hydraulic system to complete one pressure build-up, nor can it be longer than the longest time allowed by the entire production cycle. For segmentation sequence, its value range is all possible arrangements of the six segment modules, that is, six factorial arrangements.

[0076] The operator discretizes the value ranges of the four parameters into the same number of levels; for example, the value range of ejection speed is divided into several equally spaced levels, the value range of demolding temperature is divided into the same number of equally spaced levels, and the value range of holding time is also divided into the same number of equally spaced levels. For the segmentation order, since this parameter is discrete, the operator randomly selects an arrangement from all possible arrangements with the same number of levels as the sampling level. Then, the operator randomly arranges each level of each parameter, forming four sets of randomly arranged level sequences. Next, the operator follows these four random arrangement sequences... The corresponding positions are combined, that is, the first level of the first parameter random permutation sequence is combined with the first level of the second parameter random permutation sequence, the first level of the third parameter random permutation sequence, and the first level of the fourth parameter random permutation sequence to form a candidate parameter combination, and so on, until the corresponding positions of all levels are taken, thus generating a number of candidate parameter combinations equal to the number of levels. This sampling method ensures that each value level of each parameter appears exactly once in the entire set of candidate parameter combinations, so that the candidate parameter combinations are evenly distributed in the parameter space, avoiding a large number of combinations being concentrated in a local area of ​​a certain parameter value.

[0077] Furthermore, the contact conditions in the finite element simulation model include the variation of the friction coefficient of the contact surface between each segmented module with the demolding temperature;

[0078] During transient dynamic simulation, the friction coefficient at each time step is obtained by interpolation from the friction characteristic curve based on the current simulation temperature. Furthermore, when a relative sliding tendency occurs between the segmented module and the workpiece, the dynamic friction coefficient is used instead of the static friction coefficient.

[0079] In this embodiment, during the construction of the aforementioned finite element simulation model, the operator sets friction coefficients that vary with temperature for the contact surfaces between each segmented module and the contact surfaces between the segmented module and the cemented carbide workpiece. Specifically, the operator obtains the friction characteristic curve between the cemented carbide grade and the mold steel from the technical data provided by the cemented carbide material supplier. This friction characteristic curve records the static friction coefficient and dynamic friction coefficient values ​​measured at certain temperature intervals from room temperature to the upper limit of the demolding temperature. The static friction coefficient refers to the ratio of the maximum frictional resistance to the normal pressure when there is no relative sliding between the two contact surfaces, and the dynamic friction coefficient refers to the ratio of the frictional resistance to the normal pressure when there is relative sliding between the two contact surfaces.

[0080] During transient dynamics simulation, the simulation software calculates the temperature value of each contact area at each time step, and then interpolates the temperature value onto the friction characteristic curve. The specific interpolation method is as follows: if the current temperature is exactly equal to a recorded temperature point on the friction characteristic curve, the friction coefficient value corresponding to that recorded point is directly used; if the current temperature is between two adjacent recorded temperature points, the friction coefficient value corresponding to that temperature is calculated using linear interpolation; for contact surfaces that have not yet experienced relative sliding, the interpolated static friction coefficient is used for calculation; for contact surfaces that have already experienced relative sliding, the interpolated dynamic friction coefficient is used for calculation.

[0081] During the simulation, the software also needs to determine whether a relative sliding trend has appeared in each contact area at each time step. The determination method is as follows: calculate the ratio of tangential friction force to normal pressure on the contact surface at the current time step. If the ratio is less than the static friction coefficient at the current temperature, it is determined that no sliding trend has appeared, and the static friction state continues. If the ratio reaches or exceeds the static friction coefficient at the current temperature, it is determined that a sliding trend has appeared, and the contact state switches from static friction to dynamic friction. Subsequent time steps use the dynamic friction coefficient at the current temperature for calculation. This processing method enables the simulation model to realistically reflect the changes in the friction state of each contact surface with temperature and relative motion state during the demolding process, thus improving the simulation accuracy.

[0082] Furthermore, referring to Figure 2 S5 includes:

[0083] The curve of the demolding force changing with time is divided into multiple continuous time intervals. Within each time interval, all local maxima and all local minima of the demolding force value are detected.

[0084] Compare the demolding force values ​​corresponding to all local maxima within all time intervals, and take the maximum value as the peak demolding force;

[0085] For each time interval, the average value of the demolding force at all local maxima is calculated as the upper peak mean of the interval, and the average value of the demolding force at all local minima in the time interval is calculated as the lower valley mean of the interval. The difference between the upper peak mean and the lower valley mean is taken as the interval fluctuation amplitude of the time interval.

[0086] The arithmetic mean of the interval fluctuation amplitudes across all time intervals is calculated as the base fluctuation value. The standard deviation of the interval fluctuation amplitudes across all time intervals is calculated as the fluctuation dispersion. The sum of the base fluctuation value and the fluctuation dispersion is taken as the demolding force fluctuation amplitude.

[0087] In this embodiment, the operator exports a complete curve of demolding force changing over time from the simulation software. This data contains several data points sampled at equal time intervals, and each data point records a time value and its corresponding demolding force value.

[0088] The operator first determines the duration of an ejection step cycle based on the working characteristics of the hydraulic press's ejection mechanism. This step cycle is the time required for the ejector rod to advance by a small displacement. The operator then divides the entire demolding force curve into multiple consecutive time intervals at equal intervals according to the duration of this step cycle. For each time interval, the operator searches for local maxima and local minima among all sampling points within that interval. The criteria for determining a local maximum are: the demolding force value at that point is greater than the demolding force value at the preceding sampling point and greater than the demolding force value at the following sampling point. The criteria for determining a local minimum are: the demolding force value at that point is less than the demolding force value at the preceding sampling point and less than the demolding force value at the following sampling point.

[0089] After completing the detection of extreme points in all time intervals, the operator compares the demolding force values ​​corresponding to all the local maxima detected in all time intervals and selects the one with the largest value as the peak value of the demolding force of the entire curve.

[0090] Next, for each time interval, the operator calculates the average of the demolding force values ​​at all local maxima within that interval, which is taken as the peak mean of that interval; the operator also calculates the average of the demolding force values ​​at all local minima within that interval, which is taken as the valley mean of that interval; the operator subtracts the valley mean from the peak mean of that interval, and the difference is taken as the interval fluctuation amplitude of that time interval; for the interval fluctuation amplitudes of all time intervals, the operator first calculates their arithmetic mean as the base fluctuation value, then calculates their sample standard deviation as the fluctuation dispersion, and finally adds the base fluctuation value and the fluctuation dispersion to obtain the result as the demolding force fluctuation amplitude of the entire curve.

[0091] Furthermore, referring to Figure 3 S6 includes:

[0092] For each segmented module, the stress tensor components on all finite element nodes inside the segmented module are read from the stress distribution field, and the Mises equivalent stress value of each node is calculated based on the stress tensor components.

[0093] Compare the Mises equivalent stress values ​​of all nodes within the segmented module, and take the largest Mises equivalent stress value as the maximum stress value of the segmented module.

[0094] For each segmented module, multiple parallel feature paths are selected within the segmented module along the demolding direction, and each feature path is evenly distributed on the cross-section of the segmented module.

[0095] Multiple sampling points are selected at equal intervals on each feature path. The Mises equivalent stress value at each sampling point is read. The ratio of the difference between the Mises equivalent stress values ​​between adjacent sampling points to the distance between the sampling points is calculated as the local stress change rate of the adjacent point pair.

[0096] The root mean square of the local stress change rate of all adjacent point pairs on a feature path is taken as the path stress change rate of that feature path, and the arithmetic mean of the path stress change rates of all feature paths within that lobe module is taken as the stress gradient of that lobe module.

[0097] In this embodiment, for the post-processing of the stress distribution field, the operator exports the stress tensor data of all finite element nodes within each segmented module from the simulation software. The stress tensor of each node contains six independent components, namely three normal stress components and three shear stress components. The operator calculates the Mises equivalent stress value of the node based on these six components, using the standard definition of Mises equivalent stress in mechanics of materials. For each segmented module, the operator compares the Mises equivalent stress values ​​of all nodes within the module and selects the largest value as the maximum stress value of the module.

[0098] To calculate the stress gradient, the operator selects multiple feature paths along the demolding direction within each segmented module. These feature paths are uniformly distributed across the cross-section of the segmented module. For example, for a segmented module with a fan-shaped cross-section, the operator can select several radial lines at equal intervals between the inner and outer arcs of the fan as feature paths. On each feature path, the operator selects multiple sampling points at equal intervals from one end to the other and reads the Mises equivalent stress value at each sampling point. Then, for two adjacent sampling points on the path, the operator calculates the difference between their Mises equivalent stress values ​​and divides it by the distance between the two sampling points. The resulting ratio is taken as the local stress change rate of that adjacent point pair. The operator squares the local stress change rates of all adjacent point pairs on a feature path, averages them, and then takes the square root to obtain the path stress change rate of that feature path. Finally, the operator takes the arithmetic mean of the path stress change rates of all feature paths within a segmented module as the stress gradient of that segmented module.

[0099] Furthermore, the stress concentration factor is calculated based on the maximum stress value and stress gradient of each segment module according to a preset weighting rule, specifically including:

[0100] For each segmented module, the arithmetic mean of the Mises equivalent stress values ​​of all finite element nodes within the segmented module is taken as the average stress value of the segmented module.

[0101] The ratio of the maximum stress value of the segmented module to the average stress value of the segmented module is used as the local concentration factor of the segmented module.

[0102] Calculate the local concentration coefficients of all segmented modules, the maximum value and the arithmetic mean of all local concentration coefficients, and the sample standard deviation of all local concentration coefficients.

[0103] For each segmented module, the angle between the stress gradient direction vector and the demolding direction vector of the segmented module is obtained. When the angle is less than a preset angle threshold, a direction amplification factor is applied to the local concentration coefficient of the segmented module. The local concentration coefficient of the segmented module is multiplied by the direction amplification factor to obtain the adjusted local concentration coefficient.

[0104] When the included angle is greater than or equal to the angle threshold, no direction amplification factor is applied, and the adjusted local concentration coefficient is the original local concentration coefficient.

[0105] Recalculate the maximum value, arithmetic mean, and sample standard deviation of the local concentration coefficients for all segmented modules after adjustment;

[0106] The difference between the maximum value and the arithmetic mean of the adjusted local concentration coefficients is taken as the first difference term, and the sample standard deviation of the adjusted local concentration coefficients is taken as the second difference term. The first difference term is assigned a first weight, and the second difference term is assigned a second weight. The product of the first difference term and the first weight is added to the product of the second difference term and the second weight, and the sum is taken as the basic concentration coefficient.

[0107] Next, check the difference between the adjusted local concentration coefficients of any two adjacent lobe modules in all lobe modules. If there is a difference between adjacent modules that exceeds the preset difference threshold, then take the average of all the differences between adjacent modules that exceed the difference threshold as the neighborhood difference term. Multiply the neighborhood difference term by the neighborhood weight and add it to the basic concentration coefficient. The result is used as the stress concentration coefficient of the candidate parameter combination.

[0108] In this embodiment, the operator first takes the arithmetic mean of the Mises equivalent stress values ​​at all finite element nodes within each segmented module to obtain the average stress value of the segmented module. Then, the operator divides the maximum stress value of the segmented module by its average stress value, and the resulting ratio is used as the local concentration coefficient of the segmented module. The larger the local concentration coefficient, the more uneven the stress distribution within the module, and the more concentrated the stress is in a certain local area.

[0109] After calculating the local concentration coefficients for all segmented modules, the operator tallies the maximum value, arithmetic mean, and sample standard deviation of all local concentration coefficients. Simultaneously, for each segmented module, the operator obtains the stress gradient direction vector and calculates the angle between this vector and the demolding direction vector. The operator pre-sets an angle threshold, determined based on the anisotropy of the cemented carbide material. For each segmented module, if the angle between its stress gradient direction and the demolding direction is less than this threshold, it indicates that the stress within the module changes drastically along the demolding direction. In this case, the operator applies a directional amplification factor to the local concentration coefficient of the module. This directional amplification factor is calculated by dividing the angle threshold by the actual angle; the resulting ratio is used as the amplification factor. The local concentration coefficient of the module is then multiplied by this amplification factor to obtain the adjusted local concentration coefficient. If the angle is greater than or equal to the angle threshold, no directional amplification factor is applied, and the adjusted local concentration coefficient is equal to the original local concentration coefficient.

[0110] The operator recalculates the maximum value, arithmetic mean, and sample standard deviation of the local concentration coefficients after adjustment for all segmented modules. Then, the maximum value of the adjusted local concentration coefficients is subtracted from the arithmetic mean, and the difference is taken as the first difference term, which reflects the degree of deviation of the module with the most severe stress concentration from the overall average level. The sample standard deviation of the adjusted local concentration coefficients is taken as the second difference term, which reflects the degree of dispersion of stress concentration among modules. The operator assigns a first weight to the first difference term and a second weight to the second difference term, where the first weight is greater than the second weight, indicating that the maximum deviation is more important than the overall dispersion. The first difference term is multiplied by the first weight, and the second difference term is multiplied by the second weight, and the two are added together to obtain the basic concentration coefficient.

[0111] Finally, the operator checks the difference between the adjusted local concentration coefficients of any two adjacent segmented modules in all segmented modules. Adjacent segmented modules refer to two modules that are in direct contact with each other in the circumferential direction. If the difference between any pair of adjacent modules exceeds the preset difference threshold, the arithmetic mean of these differences is taken as the neighborhood difference term. This neighborhood difference term is multiplied by the neighborhood weight and added to the basic concentration coefficient. The result is used as the stress concentration coefficient of the candidate parameter combination. The purpose of this neighborhood correction term is to punish cases where the stress concentration between adjacent modules changes too abruptly, because in the actual demolding process, drastic changes in the stress state of adjacent modules may cause scratches on the workpiece surface or abnormal wear on the mold segmented surface.

[0112] Furthermore, after selecting the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter, the method further includes calling the finite element simulation model again as the input condition to perform verification simulation, thereby obtaining the verification curve of the demolding force changing with time and the verification stress distribution field.

[0113] Extract the peak value of the demolding force from the verification curve, and extract the global maximum stress value from all the segmented modules from the verification stress distribution field;

[0114] The peak value of the demolding force is compared with the upper limit of the ejection force of the demolding equipment, and the global maximum stress value is compared with the yield strength of the cemented carbide material.

[0115] If the peak value of the demolding force is less than or equal to the upper limit of the ejection force and the global maximum stress value is less than or equal to the yield strength, then the optimized output parameters are confirmed as the final effective parameters.

[0116] If the peak value of the demolding force is greater than the upper limit of the ejection force or the global maximum stress value is greater than the yield strength, then the preset value range is narrowed, candidate parameter combinations are regenerated, and subsequent steps are repeated until optimized output parameters that meet the comparison conditions are obtained.

[0117] In this embodiment, after obtaining the candidate parameter combination with the smallest comprehensive evaluation value through the aforementioned steps, the operator does not directly use this combination in actual production, but first performs a verification simulation. Specifically, the specific values ​​of ejection speed, demolding temperature, holding time, and segmentation sequence in the optimized output parameters are used as input conditions. The previously established finite element simulation model is called again to perform transient dynamic simulation calculations to obtain the verification demolding force variation curve and verification stress distribution field under this set of parameters.

[0118] After the simulation is completed, the operator extracts the maximum demolding force value from the demolding force curve as the peak value of the demolding force; from the verification stress distribution field, the operator traverses the Mises equivalent stress values ​​on all finite element nodes inside all segmented modules and finds the global maximum value as the global maximum stress value.

[0119] The operator will compare the peak demolding force with the upper limit of the ejection force of the demolding equipment; this upper limit of the ejection force is a capability parameter of the equipment clearly given in the hydraulic press technical specifications, representing the maximum ejection force that the hydraulic press can provide within its safe operating range; at the same time, the operator will compare the global maximum stress value with the yield strength of the cemented carbide material; this yield strength is measured by performing a uniaxial compression test on a standard specimen of this grade of cemented carbide, representing the maximum stress that the material can withstand before permanent plastic deformation occurs.

[0120] If the peak demolding force is verified to be less than or equal to the upper limit of the ejection force, and the global maximum stress value is less than or equal to the yield strength, it means that the optimized output parameters will neither exceed the working capacity of the equipment nor cause plastic deformation or damage to the mold material. The operator confirms that the optimized output parameters are the final effective parameters and can be used for actual demolding production.

[0121] If the peak demolding force exceeds the upper limit of the ejection force, or the global maximum stress exceeds the yield strength, or neither condition is met, it indicates that the optimized output parameter poses a risk in practical applications. In this case, the operator narrows the preset value range of each parameter, converging the range towards reducing the demolding force and stress concentration factor. Then, within the narrowed range, candidate parameter combinations are regenerated, and the entire process from simulation calculation to Pareto screening to comprehensive evaluation value calculation is repeated until an optimized output parameter is obtained that simultaneously satisfies the requirement that the peak demolding force does not exceed the upper limit of the ejection force and the global maximum stress does not exceed the yield strength.

[0122] Furthermore, the demolding parameter set includes the ejection sequence of the ejector rods and the time interval between adjacent ejection actions. The ejection sequence specifies the starting order of the ejector rods corresponding to each segment module during the demolding process, and the time interval specifies the waiting time between adjacent ejector rods.

[0123] When generating candidate parameter combinations, the ejection order is determined by the full permutation of the ejection rod numbers corresponding to all segment modules, and the time interval is determined by the minimum response time and maximum allowable interval of the demolding equipment.

[0124] In this embodiment, in the demolding equipment of the six-segment mold, each segment module corresponds to an independent ejector rod, and the six ejector rods are driven by six independent control valves of the hydraulic system. When setting the demolding parameter set, the operator will include the ejection sequence of the ejector rods and the time interval between adjacent ejection actions in the optimization scope, in addition to the ejection speed, demolding temperature, holding time and segmentation sequence.

[0125] The ejection sequence specifies the order in which the six ejector pins are activated during the demolding process. For example, one possible ejection sequence is: first activate the ejector pin corresponding to the first segment module, then activate the ejector pin corresponding to the third segment module, then activate the ejector pin corresponding to the fifth segment module, then activate the ejector pin corresponding to the second segment module, then activate the ejector pin corresponding to the fourth segment module, and finally activate the ejector pin corresponding to the sixth segment module. The ejection sequence can be any of the six possible permutations of the ejector pin numbers, which is 720 possible arrangements.

[0126] The time interval between adjacent ejection actions specifies the waiting time between two ejection rods that start adjacently; for example, after the ejection rod of the first segment module starts and before the ejection rod of the third segment module starts, a specific waiting time is required; the range of this time interval is determined based on the minimum response time and the maximum allowable interval of the demolding equipment; the minimum response time is the shortest time required for the hydraulic control valve to actually start outputting hydraulic oil from receiving the start signal, and the maximum allowable interval is the upper limit set considering the total time limit of the entire demolding process.

[0127] When generating candidate parameter combinations, operators incorporate the ejection sequence and time interval along with the existing ejection speed, demolding temperature, and holding time into a Latin hypercube sampling or comprehensive combination generation framework. For the discrete permutation parameter of ejection sequence, operators select several permutations from 720 possible permutations as sampling samples. For the continuous parameter of time interval, operators divide its value range into levels and then sample. In this way, each candidate parameter combination generated simultaneously includes the specific values ​​of the six parameters: ejection speed, demolding temperature, holding time, segmentation sequence, ejection sequence, and time interval, enabling the optimization process to simultaneously consider the impact of the ejector rod action sequence on the demolding process.

[0128] Furthermore, the finite element simulation model uses an explicit dynamics solver for transient dynamics simulation. The explicit dynamics solver uses the central difference method to discretize the time domain and updates the node velocity and displacement according to the current node acceleration in each time step, while calculating the contact force according to the contact conditions.

[0129] During the simulation, a mass scaling factor is set to control the settling time step. The mass scaling factor is determined based on the minimum element size and the material wave velocity.

[0130] A contact penetration tolerance is set. When the penetration amount between the segmented module and the workpiece exceeds the contact penetration tolerance, a penalty force is applied to the corresponding contact node to correct the penetration.

[0131] In this embodiment, the operator selects an explicit dynamics solver as the solution module in the simulation software. This solver uses the central difference method to discretize the time domain. The basic implementation of the central difference method is as follows: at the beginning of each time step, the solver calculates the nodal velocity of the next time step based on the nodal acceleration data of the current time step using the central difference format, and then calculates the nodal displacement of the next time step based on the updated nodal velocity. After completing the nodal displacement update, the solver calculates the element strain based on the displacement field, then calculates the element stress based on the material constitutive model, and then calculates the nodal internal force based on the element stress. At the same time, the solver calculates the contact force on each contact interface based on the contact state of the current time step, and applies the contact force as an external force to the corresponding node. After combining the nodal internal force and the nodal external force, the solver calculates the nodal acceleration of the next time step, preparing for the iteration of the next time step.

[0132] Before the simulation begins, the operator needs to set a mass scaling factor to control the settling time step. The settling time step of the explicit dynamics solver is limited by the minimum element size and material wave velocity in the model. The operator first checks the characteristic dimensions of all elements in the finite element mesh to find the minimum element size. Then, the operator calculates the stress wave propagation velocity in the material based on the density and elastic modulus of the cemented carbide material. The operator sets a mass scaling factor, which artificially increases the mass density of nodes corresponding to certain small-sized elements in the model, thereby reducing the material wave velocity of these elements and allowing the settling time step to be scaled up to an acceptable level. The specific value of the mass scaling factor is determined based on a combination of the minimum element size and the material wave velocity, so that the scaled-up settling time step can meet the computational efficiency requirements without affecting the accuracy of the simulation results due to excessive mass increase.

[0133] Operators also need to set contact penetration tolerance. In explicit dynamic simulation, when the contact surfaces between the segmented module and the workpiece penetrate each other, the solver needs to detect the penetration amount and compare it with the preset contact penetration tolerance. If the penetration amount between a pair of contact nodes does not exceed the contact penetration tolerance, the solver considers the penetration amount to be within the allowable range and does not perform special processing. If the penetration amount exceeds the contact penetration tolerance, the solver applies a penalty force to the pair of contact nodes. The direction of the penalty force is to push the penetrating node back outside the contact surface, and its magnitude is proportional to the extent by which the penetration amount exceeds the tolerance. By applying the penalty force, the solver corrects the penetration amount back to within the tolerance range, thereby maintaining the geometric consistency of the contact interface and avoiding distortion of simulation results due to excessive penetration.

[0134] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0135] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0136] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A parameter optimization method for the demolding process of a cemented carbide segmented mold, characterized in that, include: S1, obtain the geometric structure parameters, hard alloy material parameters, and demolding equipment operating parameters of the target segmented mold; S2, construct a finite element simulation model of the demolding process based on the geometric structure parameters and the material parameters, and set the demolding direction, constraint boundaries and contact conditions; S3, obtain the set of demolding parameters to be optimized, and generate multiple candidate parameter combinations within a preset value range for each parameter in the set of demolding parameters; S4. Using each candidate parameter combination as input conditions, the finite element simulation model is called to perform transient dynamic simulation of the demolding process, and the curve of demolding force changing with time corresponding to each candidate parameter combination and the stress distribution field inside each segment module of the mold are obtained. S5. Based on the curve of the demolding force changing with time, extract the maximum demolding force value as the peak value of the demolding force, and extract the maximum change amplitude of the demolding force value per unit time as the demolding force fluctuation amplitude. S6. Based on the stress distribution field, extract the maximum stress value inside each segmented module, and extract the stress change rate along the demolding direction inside each segmented module as the stress gradient. S7. Calculate the demolding force fluctuation coefficient and stress concentration coefficient for each candidate parameter combination. With minimizing the demolding force fluctuation coefficient and stress concentration coefficient as the optimization objectives, compare the demolding force fluctuation coefficient and stress concentration coefficient of all candidate parameter combinations, and select the non-dominated solution set. S8. From the non-dominated solution set, calculate the comprehensive evaluation value based on the demolding force fluctuation coefficient and stress concentration coefficient of each candidate parameter combination, and select the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter.

2. The method according to claim 1, characterized in that, S3 includes: Candidate parameter combinations are generated within the preset value range using the Latin hypercube sampling method. The preset value range is dynamically determined based on the upper limit of ejection stroke and upper limit of ejection force in the working parameters of the demolding equipment, as well as the tilt angle of the segmented contact surface in the geometric structure parameters.

3. The method according to claim 1, characterized in that, The contact conditions in the finite element simulation model include the variation of the friction coefficient of the contact surface between each segmented module with the demolding temperature. During transient dynamic simulation, the friction coefficient at each time step is obtained by interpolation from the friction characteristic curve based on the current simulation temperature. Furthermore, when a relative sliding tendency occurs between the segmented module and the workpiece, the dynamic friction coefficient is used instead of the static friction coefficient.

4. The method according to claim 1, characterized in that, S5 includes: The curve of the demolding force changing with time is divided into multiple continuous time intervals. Within each time interval, all local maxima and all local minima of the demolding force value are detected. Compare the demolding force values ​​corresponding to all local maxima within all time intervals, and take the maximum value as the peak demolding force; For each time interval, the average value of the demolding force at all local maxima is calculated as the upper peak mean of the interval, and the average value of the demolding force at all local minima in the time interval is calculated as the lower valley mean of the interval. The difference between the upper peak mean and the lower valley mean is taken as the interval fluctuation amplitude of the time interval. The arithmetic mean of the interval fluctuation amplitudes across all time intervals is calculated as the base fluctuation value. The standard deviation of the interval fluctuation amplitudes across all time intervals is calculated as the fluctuation dispersion. The sum of the base fluctuation value and the fluctuation dispersion is taken as the demolding force fluctuation amplitude.

5. The method according to claim 4, characterized in that, S6 includes: For each segmented module, the stress tensor components on all finite element nodes inside the segmented module are read from the stress distribution field, and the Mises equivalent stress value of each node is calculated based on the stress tensor components. Compare the Mises equivalent stress values ​​of all nodes within the segmented module, and take the largest Mises equivalent stress value as the maximum stress value of the segmented module. For each segmented module, multiple parallel feature paths are selected within the segmented module along the demolding direction, and each feature path is evenly distributed on the cross-section of the segmented module. Multiple sampling points are selected at equal intervals on each feature path. The Mises equivalent stress value at each sampling point is read. The ratio of the difference between the Mises equivalent stress values ​​between adjacent sampling points to the distance between the sampling points is calculated as the local stress change rate of the adjacent point pair. The root mean square of the local stress change rate of all adjacent point pairs on a feature path is taken as the path stress change rate of that feature path, and the arithmetic mean of the path stress change rates of all feature paths within that lobe module is taken as the stress gradient of that lobe module.

6. The method according to claim 1, characterized in that, The stress concentration factor is calculated based on the maximum stress value and stress gradient of each segment module according to a preset weighting rule, specifically including: For each segmented module, the arithmetic mean of the Mises equivalent stress values ​​of all finite element nodes within the segmented module is taken as the average stress value of the segmented module. The ratio of the maximum stress value of the segmented module to the average stress value of the segmented module is used as the local concentration factor of the segmented module. Calculate the local concentration coefficients of all segmented modules, the maximum value and the arithmetic mean of all local concentration coefficients, and the sample standard deviation of all local concentration coefficients. For each segmented module, the angle between the stress gradient direction vector and the demolding direction vector of the segmented module is obtained. When the angle is less than a preset angle threshold, a direction amplification factor is applied to the local concentration coefficient of the segmented module. The local concentration coefficient of the segmented module is multiplied by the direction amplification factor to obtain the adjusted local concentration coefficient. When the included angle is greater than or equal to the angle threshold, no direction amplification factor is applied, and the adjusted local concentration coefficient is the original local concentration coefficient. Recalculate the maximum value, arithmetic mean, and sample standard deviation of the local concentration coefficients for all segmented modules after adjustment; The difference between the maximum value and the arithmetic mean of the adjusted local concentration coefficients is taken as the first difference term, and the sample standard deviation of the adjusted local concentration coefficients is taken as the second difference term. The first difference term is assigned a first weight, and the second difference term is assigned a second weight. The product of the first difference term and the first weight is added to the product of the second difference term and the second weight, and the sum is taken as the basic concentration coefficient. Next, check the difference between the adjusted local concentration coefficients of any two adjacent lobe modules in all lobe modules. If there is a difference between adjacent modules that exceeds the preset difference threshold, then take the average of all the differences between adjacent modules that exceed the difference threshold as the neighborhood difference term. Multiply the neighborhood difference term by the neighborhood weight and add it to the basic concentration coefficient. The result is used as the stress concentration coefficient of the candidate parameter combination.

7. The method according to claim 1, characterized in that, After selecting the candidate parameter combination with the smallest comprehensive evaluation value as the optimized output parameter, the process also includes calling the finite element simulation model again as the input condition to perform verification simulation, thereby obtaining the verification curve of the demolding force changing with time and the verification stress distribution field. Extract the peak value of the demolding force from the verification curve, and extract the global maximum stress value from all the segmented modules from the verification stress distribution field; The peak value of the demolding force is compared with the upper limit of the ejection force of the demolding equipment, and the global maximum stress value is compared with the yield strength of the cemented carbide material. If the peak value of the demolding force is less than or equal to the upper limit of the ejection force and the global maximum stress value is less than or equal to the yield strength, then the optimized output parameters are confirmed as the final effective parameters. If the peak value of the demolding force is greater than the upper limit of the ejection force or the global maximum stress value is greater than the yield strength, then the preset value range is narrowed, candidate parameter combinations are regenerated, and subsequent steps are repeated until optimized output parameters that meet the comparison conditions are obtained.

8. The method according to claim 1, characterized in that, The demolding parameter set includes the ejection sequence of the ejector rods and the time interval between adjacent ejection actions. The ejection sequence specifies the starting order of the ejector rods corresponding to each segment module during the demolding process, and the time interval specifies the waiting time between adjacent ejector rods. When generating candidate parameter combinations, the ejection order is determined by the full permutation of the ejection rod numbers corresponding to all segment modules, and the time interval is determined by the minimum response time and maximum allowable interval of the demolding equipment.

9. The method according to claim 1, characterized in that, The finite element simulation model uses an explicit dynamics solver for transient dynamics simulation. The explicit dynamics solver uses the central difference method to discretize the time domain and updates the node velocity and displacement according to the current node acceleration in each time step, while calculating the contact force according to the contact conditions. During the simulation, a mass scaling factor is set to control the settling time step. The mass scaling factor is determined based on the minimum element size and the material wave velocity. A contact penetration tolerance is set. When the penetration amount between the segmented module and the workpiece exceeds the contact penetration tolerance, a penalty force is applied to the corresponding contact node to correct the penetration.

Images