Forming die service life monitoring method and system based on powder metallurgy metal material

Abstract

The invention discloses a forming die service life monitoring method and system based on a powder metallurgy metal material, and the method comprises the steps: carrying out the multiple fitting and updating of an initial data set composed of pressing process data and die service life change data, and obtaining a prediction model of the die service life change data under a normal condition, obtaining an estimation model of the mold life change data under the extreme condition according to the pressing process data, the duration and the mold life change data of a data segment formed by the data points at the adjacent moments in the data set under the extreme condition, and the data set under the normal condition and the data set under the extreme condition; and after the to-be-evaluated forming mold completes the current pressing process, according to the pressing process data belonging to the normal condition and the extreme condition, a mold life change value is obtained in combination with the prediction model of the mold life change data, and the mold life change value is superposed with the mold life value after the previous pressing process to obtain the mold life value after the current pressing process. According to the invention, the service life change of the mold can be comprehensively and accurately estimated.

Description

Technical Field

[0001] The present invention relates to the technical field of powder metallurgy. More specifically, the present invention relates to a method and system for monitoring the service life of a forming die based on powder metallurgy metal materials. Background Art

[0002] In the production process of powder metallurgy parts, the forming die is a key process equipment; traditional die life monitoring methods often rely on regular manual inspections, such as observing the wear condition of the die surface and measuring dimensional changes to determine whether the die needs to be replaced.

[0003] However, this method has obvious limitations: on the one hand, the frequency of manual inspection is usually low, and it is impossible to capture the subtle changes of the die during use in a timely manner, which may lead to excessive wear of the die between two inspections and affect the product quality; on the other hand, the subjectivity of manual inspection is strong, and the judgment results of different inspectors may vary, reducing the accuracy of the inspection.

[0004] Therefore, the die life prediction method based on data analysis has important practical significance for monitoring the service life of the forming die.

[0005] However, the existing die life prediction models are single and cannot comprehensively reflect the die life change rules under different conditions, with limited adaptability and accuracy. Summary of the Invention

[0006] To solve the above technical problems that the existing die life prediction models are single and cannot comprehensively reflect the die life change rules under different conditions, with limited adaptability and accuracy, the present invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a method for monitoring the service life of a forming die based on powder metallurgy metal materials, including: when pressing a powder metallurgy part with the forming die, collecting the pressing process data at each moment during each pressing process and the die parameter data after each pressing ends; performing multi-source data weighted fusion on the die parameter data through the analytic hierarchy process to obtain the die service life data after each pressing ends; forming an initial data set with the pressing process data and the die service life change data, performing multiple fittings and updates on the initial data set to obtain a prediction model for the die service life change data under normal conditions, and determining the data set under normal conditions and the data set under extreme conditions. Through the maximum inter-class variance method, determine the optimal threshold for distinguishing whether the pressing process data belongs to the normal situation or the extreme situation; form data segments from the data points at adjacent moments in the data set under the extreme situation; construct equations for each pressing process based on the pressing process data, the duration, and the die service life change data of all data segments in each pressing process; solve the system of equations composed of the equations of all pressing processes through the least squares method to obtain a prediction model for the die service life change data under the extreme situation; after the forming die to be evaluated completes the current pressing process, through the optimal threshold, obtain the pressing process data belonging to the normal situation and the pressing process data belonging to the extreme situation in all the pressing process data collected during the current pressing process, and input them into the prediction models for the die service life change data under the normal situation and the extreme situation respectively, to obtain the die service life change value after the current pressing process of the forming die to be evaluated, and add it to the die service life value after the previous pressing process to obtain the die service life value after the current pressing process of the forming die to be evaluated.

[0008] Preferably, the pressing process data includes the pressing pressure, the pressing speed, the die temperature distribution, and the die temperature change rate, and the die parameter data includes: die deformation data, die surface roughness data, and die crack data. Among them, the die deformation data includes dimensional deformation data, parallelism deformation data, and perpendicularity deformation data, and the die crack data includes the maximum crack length, the maximum crack depth, and the number of cracks.

[0009] Preferably, the method for obtaining the initial data set is: calculating the difference between the die service life data after each pressing ends and the die service life data after the previous pressing ends as the die service life change data after each pressing ends; using the pressing process data as the first dimension and the die service life change data as the second dimension to construct a two-dimensional rectangular coordinate system. Then, the pressing process data at a moment during a pressing process and the die service life change data after this pressing end together form a data point in the two-dimensional rectangular coordinate system; and then, according to the pressing process data at all moments during all pressing processes and the die service life change data after all pressing ends, obtain the initial data set composed of all data points.

[0010] Preferably, the method of performing multiple fittings and updates on the initial data set to obtain a prediction model for the die life change data under normal conditions includes: using the pressing process data of the data points as independent variables and the die life change data of the data points as dependent variables, constructing an equation based on the pressing process data and the die life change data of one data point, and denoting it as the equation of one data point; solving the system of equations composed of the equations of all data points in the data set by the least squares method to obtain the -th fitting result; calculating the -th fitting error of each data point in the data set based on the -th fitting result; determining the outer limit of the -th fitting error of all data points through a box plot; deleting the data points outside the outer limit from the data set to obtain the data set ; where the value range of is , stop after obtaining the fourth fitting result, and use the fourth fitting result as the prediction model for the pressing process data and the die life change data under normal conditions, where when , the data set is the initial data set.

[0011] Preferably, the method of determining the data set under normal conditions and the data set under extreme conditions includes: using the data set as the data set under normal conditions, and using the difference set between the initial data set and the data set as the data set under extreme conditions. Preferably, the method of determining the optimal threshold for distinguishing whether the pressing process data belongs to the normal situation or the extreme situation by the maximum inter-class variance method includes: obtaining the upper quartile of the pressing process data in the data set under normal conditions, and obtaining the lower quartile of the pressing process data in the data set under extreme conditions; using each data within the range as a threshold to classify the pressing process data of all data points in the initial data set, classifying the pressing process data greater than or equal to the threshold into one category and the pressing process data less than the threshold into another category, and calculating the inter-class variance between the two categories; using the threshold with the maximum inter-class variance as the optimal threshold.

[0012] Preferably, the steps of constructing the equations for each pressing process include: for any pressing process, input the pressing process data and the duration of each data segment in this pressing process into the expression of the model to obtain the simulated values of each data segment; construct the equations for each pressing process by making the simulated values of all data segments in this pressing process equal to the die life change data after the end of this pressing process.

[0013] Preferably, the steps of obtaining the die life change value of the to-be-evaluated forming die after the current pressing process include: for any pressing process data: input each pressing process data belonging to the normal situation into the pressing process data under normal conditions, and take the mean value of the pre-estimated die life change values corresponding to all the pressing process data belonging to the normal situation as the pre-estimated die life change value of the current pressing process under normal conditions; form data segments from the pressing process data at adjacent time points among all the pressing process data belonging to the extreme situation, input the pressing process data and the duration of each data segment into the pre-estimation model of the die life change data under the extreme situation, and take the sum value of the pre-estimated die life change values corresponding to all the data segments as the pre-estimated die life change value of the current pressing process under the extreme situation; take the maximum value between the pre-estimated die life change value of the current pressing process under normal conditions and the pre-estimated die life change value of the current pressing process under the extreme situation as the die life change value corresponding to this pressing process data; take the mean value of the die life change values corresponding to all the pressing process data as the die life change value of the to-be-evaluated forming die after the current pressing process.

[0014] Preferably, the pressing process data of the data segment is equal to the mean value of the pressing process data of all data points in the data segment; the method for obtaining the duration of the data segment is: calculate the product of the number of all pressing process data in the data segment and the acquisition time interval, and take the ratio of the product to as the duration of the data segment, where is equal to the average value of the durations of all pressing processes.

[0015] In a second aspect, the present invention provides a forming die life monitoring system based on powder metallurgy metal materials, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned forming die life monitoring method based on powder metallurgy metal materials is implemented.

[0016] By adopting the above technical solution, the above-mentioned forming die life monitoring method based on powder metallurgy metal materials is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0017] The beneficial effects of the present invention are as follows: The present invention combines the influence of the numerical values of the pressing process data under normal conditions on the die life, and the superimposed influence of the numerical values and duration of the pressing process data under extreme conditions on the die life, obtains the optimal threshold for distinguishing whether the pressing process data belongs to normal conditions or extreme conditions, and solves the prediction models of the die life change data under normal conditions and extreme conditions; for the molding die to be evaluated, by means of classification processing and model calculation, dynamically and real-time track the change of the die life, so as to more comprehensively and accurately predict the change of the die life, provide a reliable basis for production decision-making, effectively avoid excessive wear or premature replacement of the die, reduce production costs, and improve production efficiency and product quality. Brief Description of the Drawings

[0018] By reading the following detailed description with reference to the drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown by way of example and not limitation, and the same or corresponding reference numerals denote the same or corresponding parts, wherein: Figure 1 is a flowchart schematically showing the method for monitoring the life of a molding die based on powder metallurgy metal materials in the present invention. Detailed Embodiments

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] The following will describe in detail the specific embodiments of the present invention with reference to the drawings.

[0021] The embodiments of the present invention disclose a method for monitoring the life of a molding die based on powder metallurgy metal materials, referring to Figure 1 , including steps S1 to S5: In the actual production process of pressing powder metallurgy parts through a forming die, data collection mainly involves two types of data, namely, pressing process data and die parameter data. Among them, the collection of pressing process data is relatively simple and efficient, usually completed synchronously during the pressing process, without significant interference to the normal production process. Specifically, by pre-installing sensors on the forming die and related equipment, key data such as pressing pressure, pressing speed, die temperature distribution, and die temperature change rate can be collected in real time during each pressing of powder metallurgy parts. These sensors can quickly convert various physical quantities into electrical signals and transmit them to the data acquisition system, thereby realizing the dynamic monitoring of the pressing process. This not only ensures the real-time and efficiency of data collection but also does not affect the normal production of powder metallurgy parts at all, ensuring the balance between production efficiency and data acquisition.

[0022] However, the collection of die parameter data is completely different. The process is relatively complex and time-consuming, mainly relying on a variety of precision instruments and detection methods. After each pressing is completed, the die needs to be removed from the production equipment, and then professional equipment such as optical measurement equipment, surface roughness detectors, and ultrasonic detectors are used to carefully detect the dimensional deformation of the die (covering deformation indicators such as dimensions, parallelism, and perpendicularity), surface roughness, and crack conditions (including the maximum length, depth, and number of cracks). This series of operations not only requires professional detection equipment but also has high requirements for the detection environment and the technical level of operators.

[0023] More importantly, since the die cannot be used for production during the detection period, this will undoubtedly have a direct impact on the production progress of powder metallurgy parts, increasing the production waiting time and reducing the overall production efficiency.

[0024] Therefore, although die parameter data can more intuitively reflect the die life, from the perspectives of production efficiency and cost control, it is expected to estimate the die life through pressing process data, and establish an association model between the pressing process data and the die life change data, that is, the die life prediction model, by analyzing the pressing process data and die life data in the test data.

[0025] S1. Collect the life test data of the forming die. The test data includes pressing process data and die parameter data, and perform multi-source data weighted fusion on the die parameter data through the analytic hierarchy process to obtain the die life data after each pressing.

[0026] It should be noted that first, the analytic hierarchy process is used to perform multi-source data weighted fusion on the die parameter data of each forming die after each pressing in the life test data, so as to obtain the die life data of each forming die after each pressing, providing strong data support for subsequent analysis of the pressing process data and die life data in the test data, and establishing an association model between the pressing process data and the die life change data, that is, the die life prediction model.

[0027] Specifically, powder metallurgy parts are pressed by a forming die. During the process of pressing powder metallurgy parts by the forming die each time, the pressing process data at each moment during each pressing is collected; after each pressing of powder metallurgy parts by the forming die, the die parameter data after each pressing is collected; the pressing process data at each moment during each pressing and the die parameter data after each pressing are used as the life test data of each forming die during each pressing process; the multi-source data weighted fusion is performed on the die parameter data of each forming die after each pressing by the analytic hierarchy process to obtain the die life data of each forming die after each pressing, realizing the life test of each forming die.

[0028] Among them, the pressing process data includes pressing pressure, pressing speed, die temperature distribution, and die temperature change rate. The collection method of each pressing process data and its influence on the life of the forming die are as follows: (1) For the pressing pressure, a high-precision pressure sensor is installed at the key position of the pressure system of the powder metallurgy forming machine to sense the pressure magnitude during pressing in real time. After being processed by the signal conversion and acquisition device, it is transmitted to the data acquisition system, and continuous and accurate pressure data during the pressing process can be obtained; and if the pressure is too high, it will cause local stress concentration in the forming die, accelerating the wear and deformation of the forming die and affecting the life.

[0029] (2) For the pressing speed, a laser displacement sensor is assembled at the key position of the transmission system of the powder metallurgy forming machine to monitor the displacement change of the pressing head in real time and calculate the speed. Both of them transmit the speed data to the data acquisition system; and too fast pressing speed will cause the forming die to receive a large impact force, increasing the wear and fatigue damage of the forming die, and thus affecting the life.

[0030] (3) For the die temperature distribution, the temperature data of the forming die is collected non-contact by an infrared thermometer and transmitted to the data acquisition system in real time to realize multi-point and real-time temperature monitoring; calculate the range and variance of the temperature data at all positions of the forming die, and take the product of the range and variance as the die temperature distribution; uneven temperature distribution will cause uneven thermal expansion of the forming die, generating thermal stress, and thus causing problems such as deformation and cracking of the forming die, affecting the life.

[0031] (4)Regarding the mold temperature change rate, the mold temperature change rate refers to the ratio of the difference in the temperature data of the molding die at adjacent times to the time interval between adjacent times, which can reflect the dynamic change characteristics of the temperature of the molding die; too fast temperature change will cause the molding die to be thermally shocked, resulting in cracks on the surface and inside of the molding die, and reducing the service life of the molding die.

[0032] Among them, the mold parameter data includes mold deformation data, mold surface roughness data, and mold crack data. Among them, the mold deformation data includes dimensional deformation data, parallelism deformation data, and perpendicularity deformation data. The mold crack data includes the maximum crack length, the maximum crack depth, and the number of cracks. Then the mold parameter data includes dimensional deformation data, parallelism deformation data, perpendicularity deformation data, mold surface roughness data, the maximum crack length, the maximum crack depth, and the number of cracks. The acquisition method for each type of mold parameter data is as follows: (1)Regarding the mold deformation data, use a coordinate measuring machine or an optical measuring device to measure the dimensions, parallelism, and perpendicularity of the molding die; measure the unused molding die to obtain the initial dimensions, initial parallelism, and initial perpendicularity of the molding die, and measure the molding die after each pressing to obtain the dimensions, parallelism, and perpendicularity of the molding die after each pressing. Calculate the absolute value of the difference between the dimensions, parallelism, and perpendicularity of the molding die after each pressing and the initial dimensions, initial parallelism, and initial perpendicularity of the molding die to obtain the dimensional deformation data, parallelism deformation data, and perpendicularity deformation data of the molding die after each pressing.

[0033] (2)Regarding the mold surface roughness data, use a surface roughness measuring instrument to detect the roughness value of the molding die surface after each pressing to obtain the mold surface roughness data after each pressing.

[0034] (3)Regarding the mold crack data, use non-destructive testing methods such as ultrasonic testing, magnetic particle testing, or penetrant testing to conduct a comprehensive crack detection on the molding die after each pressing; for all detected cracks, record their lengths and depths, and take the maximum value among the lengths of all cracks as the maximum crack length, and take the maximum value among the depths of all cracks as the maximum crack depth.

[0035] Among them, through the analytic hierarchy process, multi-source data weighted fusion is performed on the mold parameter data of each molding die after each pressing to obtain the mold life data of each molding die after each pressing, including: (1)Establish a hierarchical structure model. The hierarchical structure model consists of three levels, namely the goal level, the criterion level, and the index level: The goal level is the expected result of the problem. In this embodiment, the die life assessment is taken as the goal level; The criterion level includes the main factors affecting the achievement of the goal. In this embodiment, the die deformation, the die surface roughness, and the die crack are taken as the criterion level; The index level is the specific refinement of the criterion level factors. In this embodiment, the die deformation corresponds to the dimensional deformation data, the parallelism deformation data, and the perpendicularity deformation data, the die surface roughness corresponds to the die surface roughness data, and the die crack corresponds to the maximum crack length, the maximum crack depth, and the number of cracks.

[0036] (2)Construct a judgment matrix: Through expert scoring, compare the factors at the criterion level and the index level pairwise, assign scales, and use the 1-9 scale method to quantify the comparison results. 1 means the two are equally important, 3 means one is slightly more important, 5 means one is significantly more important, 7 means one is very important, 9 means one is absolutely important, and 2, 4, 6, 8 are intermediate values; Construct a judgment matrix to determine the relative importance of each factor.

[0037] (3)Calculate the weight vector, including: Normalize the judgment matrix to obtain the weight vector of each factor.

[0038] (4)Conduct hierarchical single sorting and hierarchical total sorting, including: Hierarchical single sorting refers to calculating the weight vector for each judgment matrix to obtain the sorting of each factor at this level relative to a certain factor at the previous level; Hierarchical total sorting refers to calculating the total sorting weight of each factor relative to the goal level by synthesizing the single sorting results of each level, so as to obtain the final weight of each factor.

[0039] (5)Utilize each die parameter data and its weight, and complete data fusion through the weighted average method to obtain the die life data after each pressing.

[0040] It should be noted that the analytic hierarchy process is a multi-criteria decision-making analysis method that combines qualitative and quantitative methods, which can help determine the weights of each data. The analytic hierarchy process is a well-known technology and will not be elaborated here.

[0041] S2. Construct an initial data set from the pressing process data and the die life change data, perform multiple fittings and updates on the initial data set, obtain a prediction model for the die life change data under normal conditions, and determine the data set under normal conditions and the data set under extreme conditions.

[0042] It should be noted that the life test data of the forming die includes the pressing process data under normal conditions and the pressing process data under extreme conditions: for the pressing process data under normal conditions, the degree of influence on the die life changes with the change of the value; for the pressing process data under extreme conditions, the degree of influence on the die life is not only related to the value, but also increases with the duration of the extreme condition; therefore, the influence of the pressing process data under extreme conditions on the die life is different from that of the pressing process data under normal conditions. So in this embodiment, through multiple fittings and updates of the initial data set composed of the pressing process data and the die life change data, by continuously updating the initial data set, the pressing process data under extreme conditions is deleted from the initial data set, and only the pressing process data under normal conditions is retained and composed into a data set. By fitting the data set composed of the pressing process data under normal conditions, a prediction model of the die life change data under normal conditions is obtained.

[0043] Specifically, calculate the difference between the die life data after each pressing and the die life data after the previous pressing as the die life change data after each pressing; for any kind of pressing process data: form an initial data set with the pressing process data and the die life change data, perform multiple fittings and updates on the initial data set to obtain a prediction model of the die life change data under normal conditions, and determine the data set under normal conditions and the data set under extreme conditions. The specific process is as follows: 1. Take the pressing process data as the first dimension and the die life change data as the second dimension to construct a two-dimensional rectangular coordinate system. Then, for a moment in a pressing process, the pressing process data and the die life change data after this pressing together form a data point in the two-dimensional rectangular coordinate system. Furthermore, based on the pressing process data at all moments in all pressing processes and the die life change data after all pressings, obtain the initial data set composed of all data points, and denote it as the data set .

[0044] 2. Take the pressing process data of the data point as the independent variable and the die life change data of the data point as the dependent variable. According to the pressing process data and the die life change data of a data point, construct an equation, and denote it as the equation of a data point. Solve the system of equations composed of the equations of all data points in the data set by the least squares method to obtain the first fitting result. Based on the first fitting result, calculate the first fitting error of each data point in the data set . Determine the outer limit of the first fitting error of all data points through a box plot. Delete the data points located outside the outer limit from the data set to obtain the data set 。

[0045] Among them, when performing fitting, linear fitting of one variable, quadratic fitting of one variable, and cubic fitting of one variable are respectively carried out, that is, a linear equation of one variable, a quadratic equation of one variable, and a cubic equation of one variable are respectively constructed. After obtaining the fitting results of different degrees, the fitting effect is evaluated through residual analysis: if the residuals are randomly distributed around zero, it indicates that the fitting result is good; if the residuals show a systematic trend, a higher-degree fitting needs to be tried.

[0046] 3. Take the pressing process data of the data points as the independent variable and the die life change data of the data points as the dependent variable. According to the pressing process data and die life change data of one data point, construct an equation, and record it as the equation of one data point; solve the system of equations composed of the equations of all data points in the data set by the least squares method to obtain the fitting result of the second time; based on the fitting result of the second time, calculate the second fitting error of each data point in the data set ; determine the outer limit of the second fitting error of all data points through a box plot; delete the data points located outside the outer limit from the data set to obtain the data set 。

[0047] 4. Take the pressing process data of the data points as the independent variable and the die life change data of the data points as the dependent variable. According to the pressing process data and die life change data of one data point, construct an equation, and record it as the equation of one data point; solve the system of equations composed of the equations of all data points in the data set by the least squares method to obtain the fitting result of the third time; based on the fitting result of the third time, calculate the third fitting error of each data point in the data set ; determine the outer limit of the third fitting error of all data points through a box plot; delete the data points located outside the outer limit from the data set to obtain the data set 。

[0048] 5. Take the pressing process data of the data points as the independent variable and the die life change data of the data points as the dependent variable, and perform the fourth linear fitting on all data points in the data set by the least squares method to obtain the fitting result of the fourth time, which is used as the prediction model for the pressing process data and die life change data under normal conditions.

[0049] It should be noted that when performing the second to fourth linear fittings, the degree of the fitting result is the same as the degree of the fitting result obtained from the first linear fitting.

[0050] 6. The data set As the dataset under normal conditions, the difference set between the initial dataset and the dataset is used as the dataset under extreme conditions.

[0051] S3. According to the dataset under normal conditions and the dataset under extreme conditions, by using the Otsu method, determine the optimal threshold for distinguishing whether the data in the pressing process belongs to the normal situation or the extreme situation.

[0052] It should be noted that in order to find a boundary in the data of the pressing process that can effectively distinguish between the normal situation and the extreme situation, in this embodiment, based on the statistical characteristics of the data, the Otsu method is used to determine the optimal threshold, which can maximize the difference between the classified normal situation and extreme situation data, thereby improving the accuracy of classification.

[0053] Specifically, for any kind of data in the pressing process: According to the dataset under normal conditions and the dataset under extreme conditions, by using the Otsu method, determine the optimal threshold for distinguishing whether the data in the pressing process belongs to the normal situation or the extreme situation. The specific process is as follows: 1. Obtain the upper quartile of the data in the pressing process in the dataset under normal conditions , and obtain the lower quartile of the data in the pressing process in the dataset under extreme conditions .

[0054] Among them, the upper quartile and the lower quartile refer to the values at the 25% and 75% positions after sorting a set of data from small to large.

[0055] 2. Take each data within the range as the threshold, classify the data of all data points in the initial dataset in the pressing process, classify the data of the pressing process that is greater than or equal to the threshold into one category, and classify the data of the pressing process that is less than the threshold into another category, and calculate the between-class variance of the two categories; take the threshold with the largest between-class variance as the optimal threshold for distinguishing whether the data in the pressing process belongs to the normal situation or the extreme situation.

[0056] It should be noted that by determining the optimal threshold, the normal data and extreme data in the pressing process can be more accurately identified, providing a reliable data classification basis for separately constructing prediction models and performing life prediction in the subsequent process, and improving the credibility of the prediction results.

[0057] S4. According to the data of the pressing process, the duration, and the data of the change in the die life of the data segments composed of adjacent time points in the dataset under extreme conditions, obtain the prediction model of the change in the die life under extreme conditions.

[0058] It should be noted that for the pressing process data belonging to extreme cases, the degree of influence on the die life is not only related to the numerical value, but also accumulates with the duration of the extreme case. Therefore, in this embodiment, the data points at adjacent times in the data set under extreme cases are combined into data segments, and the die life change data after one pressing is equal to the sum of the die life change evaluation values corresponding to all data segments during one pressing process. Accordingly, based on the pressing process data, duration, and die life change data after pressing for all data segments in each pressing process, an equation for each pressing process is constructed, and the least squares method is used to solve the system of equations composed of the equations for all pressing processes to obtain a prediction model for the die life change data under extreme cases.

[0059] Specifically, for any kind of pressing process data: combine the data points at adjacent times in the data set under extreme cases into data segments; construct an equation for each pressing process based on the pressing process data, duration, and die life change data after pressing for all data segments in each pressing process; use the least squares method to solve the system of equations composed of the equations for all pressing processes to obtain a prediction model for the die life change data under extreme cases. The specific process is as follows: 1. For all data points in the data set under extreme cases, divide all data points into multiple data segments according to the pressing process and time corresponding to the data points, requiring that all data points in each data segment correspond to the same pressing process and the times corresponding to all data points in each data segment are continuous.

[0060] Exemplarily, when the data set under extreme cases is where represents the data point jointly composed of the pressing process data at the th moment in the th pressing process and the die life change data after the th pressing, then represents the data point jointly composed of the pressing process data at the 18th moment in the 1st pressing process and the die life change data after the 1st pressing; divide all data points into multiple data segments according to the pressing process and time corresponding to the data points, and a total of 4 data segments are obtained, which are respectively , , , .

[0061] 2. For any one data segment, calculate the mean of the pressing process data of all data points in the data segment as the pressing process data of the data segment; calculate the product of the number of all pressing process data in the data segment and the acquisition time interval, and take the ratio of the product to as the duration of the data segment, where Equal to the average duration of all pressing processes, used to normalize the duration of the data segment.

[0062] 3. Taking the pressing process data and duration of the data segment as independent variables, and the die life change data as the dependent variable. For any one pressing process: According to the pressing process data, duration of all data segments in this pressing process, and the die life change data after the end of this pressing, input the pressing process data and duration of each data segment in this pressing process into the expression of the model to obtain the simulated value of each data segment; By making the simulated values of all data segments in this pressing process equal to the die life change data after the end of this pressing, construct the equations of each pressing process, specifically: ; In the formula, is the pressing process data of the th data segment in this pressing process, is the duration of the th data segment in this pressing process, is the expression of the model, is the number of all data segments in this pressing process, is the die life change data after the end of this pressing.

[0063] 4. Solve the system of equations composed of the equations of all pressing processes by the least squares method to obtain the prediction model of the pressing process data and die life change data in extreme cases.

[0064] Among them, when fitting, perform linear - binary fitting, quadratic - binary fitting, cubic - binary fitting, and quartic - binary fitting respectively, that is, construct linear - binary equations, quadratic - binary equations, cubic - binary equations, and quartic - binary equations respectively. After obtaining the fitting results of models with different degrees, evaluate the fitting effect of the model through residual analysis: If the residuals are randomly distributed around zero, it indicates that the fitting effect of the model is good; If the residuals show a systematic trend, a model with a higher degree needs to be tried.

[0065] It should be noted that according to the continuity and relevance of data in extreme cases, by constructing and solving the system of equations, a prediction model of die life change specifically for extreme cases can be established, which makes up for the prediction deviation that may be caused by only using the normal - case model and improves the prediction ability of die life under various complex conditions.

[0066] S5. After the current pressing process of the forming die to be evaluated is completed, based on the pressing process data belonging to normal and extreme conditions, combined with the prediction model of the die life change data, the die life change value is obtained, and it is superimposed on the die life value after the previous pressing process to obtain the die life value after the current pressing process.

[0067] It should be noted that in this embodiment, the prediction model established above is applied to the process of actual die life prediction. Through classification processing and model calculation, the change of die life can be tracked dynamically and in real time.

[0068] Specifically, for the forming die to be evaluated, powder metallurgy parts are pressed by the forming die to be evaluated. During the process of pressing powder metallurgy parts by the forming die to be evaluated this time, the pressing process data at each moment during each pressing process is collected.

[0069] Furthermore, after pressing powder metallurgy parts by the forming die to be evaluated this time, for any kind of pressing process data: through the optimal threshold, the pressing process data belonging to normal conditions and the pressing process data belonging to extreme conditions in all the pressing process data collected during the current pressing process are obtained, and are respectively input into the prediction models of die life change data under normal conditions and extreme conditions to obtain the die life change values corresponding to the pressing process data. The specific process is as follows: 1. Through the optimal threshold, the pressing process data belonging to normal conditions and the pressing process data belonging to extreme conditions in all the pressing process data collected during the current pressing process are obtained: if the pressing process data is less than the optimal threshold, then this pressing process data belongs to normal conditions; if the pressing process data is greater than or equal to the optimal threshold, then this pressing process data belongs to extreme conditions.

[0070] 2. Input each pressing process data belonging to normal conditions in all the pressing process data collected during the current pressing process into the pressing process data under normal conditions to obtain the predicted die life change values corresponding to each pressing process data belonging to normal conditions; calculate the mean value of the predicted die life change values corresponding to all the pressing process data belonging to normal conditions as the predicted die life change value of the current pressing process under normal conditions.

[0071] 3. Divide all the pressing process data belonging to extreme conditions in all the pressing process data collected during the current pressing process into multiple data segments, requiring that the moments corresponding to all the data points in each data segment are continuous, and record the data segment as the data segment belonging to extreme conditions; calculate the mean value of the pressing process data of all the data points in each data segment as the pressing process data of each data segment; calculate the product of the number of all the pressing process data in each data segment and the collection time interval, and take the ratio of the product to as the duration of each data segment.

[0072] 4. Input the pressing process data and duration of each data segment belonging to the extreme case into the prediction model of the die life change data under the extreme case, and obtain the predicted die life change value corresponding to each data segment belonging to the extreme case; calculate the sum of the predicted die life change values corresponding to all data segments belonging to the extreme case as the predicted die life change value of the current pressing process under the extreme case.

[0073] 5. Take the maximum value between the predicted die life change value of the current pressing process under normal conditions and the predicted die life change value of the current pressing process under the extreme case as the die life change value corresponding to the data of this pressing process.

[0074] Finally, take the average value of the die life change values corresponding to all pressing process data as the die life change value of the to-be-evaluated forming die after the current pressing process; superimpose the die life change value of the to-be-evaluated forming die after the current pressing process with the die life value after the previous pressing process to obtain the die life value of the to-be-evaluated forming die after the current pressing process.

[0075] It should be noted that through the real-time and accurate prediction of the forming die life, production personnel can make decisions in a timely manner according to the prediction results, such as arranging the maintenance or replacement of the die in advance, effectively avoiding the production quality problems and production interruption risks caused by excessive die wear, improving the production efficiency and product quality, and at the same time reducing the production cost.

[0076] The embodiment of the present invention also discloses a forming die life monitoring system based on powder metallurgy metal materials, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the forming die life monitoring method based on powder metallurgy metal materials according to the present invention is implemented.

[0077] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface, and their settings and functions are known in the art, so they will not be elaborated here.

Claims

1. A method for monitoring the service life of a forming die based on powder metallurgy metal materials, characterized in that, Including: When pressing a powder metallurgy part through a forming die, collect the pressing process data at each moment during each pressing process and the die parameter data after each pressing ends; perform multi-source data weighted fusion on the die parameter data through the analytic hierarchy process to obtain the die life data after each pressing ends; Construct an initial data set from the pressing process data and the die life change data, perform multiple fittings and updates on the initial data set to obtain a prediction model for the die life change data under normal conditions, and determine the data set under normal conditions and the data set under extreme conditions. Through the maximum inter-class variance method, determine the optimal threshold for distinguishing whether the pressing process data belongs to the normal situation or the extreme situation; Form data segments from the data points at adjacent moments in the data set under extreme conditions; construct equations for each pressing process based on the pressing process data, duration, and die life change data of all data segments in each pressing process; solve the system of equations composed of the equations of all pressing processes through the least squares method to obtain a prediction model for the die life change data under extreme conditions; After the forming die to be evaluated completes the current pressing process, through the optimal threshold, obtain the pressing process data belonging to the normal situation and the pressing process data belonging to the extreme situation in all the pressing process data collected during the current pressing process, and input them into the prediction models for the die life change data under normal conditions and extreme conditions respectively to obtain the die life change value of the forming die to be evaluated after the current pressing process, and add it to the die life value after the previous pressing process to obtain the die life value of the forming die to be evaluated after the current pressing process.

2. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 1, wherein, The pressing process data includes pressing pressure, pressing speed, die temperature distribution, and die temperature change rate. The die parameter data includes: die deformation amount data, die surface roughness data, and die crack data. Among them, the die deformation amount data includes dimensional deformation amount data, parallelism deformation amount data, and perpendicularity deformation amount data. The die crack data includes the maximum crack length, the maximum crack depth, and the number of cracks.

3. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 1, wherein, The method for obtaining the initial data set is: Calculate the difference between the die life data after each pressing ends and the die life data after the previous pressing ends as the die life change data after each pressing ends; Take the pressing process data as the first dimension and the die life change data as the second dimension to construct a two-dimensional rectangular coordinate system. Then, the pressing process data at a moment during a pressing process and the die life change data after this pressing ends together form a data point in the two-dimensional rectangular coordinate system; furthermore, according to the pressing process data at all moments during all pressing processes and the die life change data after all pressing ends, obtain the initial data set composed of all data points.

4. The method for monitoring the service life of a forming die based on powder metallurgy metal materials according to claim 1, characterized in that, The multiple fittings and updates of the initial data set to obtain a prediction model for the die life change data under normal conditions include: Taking the pressing process data of the data points as the independent variable and the die life change data of the data points as the dependent variable, an equation is constructed according to the pressing process data and the die life change data of a data point, and it is denoted as the equation of a data point; by using the least squares method to solve the system of equations composed of the equations of all data points in the data set to obtain the fitting result of the th time; based on the fitting result of the th time, calculate the th fitting error of each data point in the data set; determine the outer limit of the th fitting error of all data points through a box plot; delete the data points outside the outer limit from the data set to obtain the data set ; ; Among them, The value range of is, stop after obtaining the fitting result of the fourth time, and use the fitting result of the fourth time as the prediction model for the data of the pressing process and the data of the change of the die life under normal conditions. Among them, When the data set is the initial data set.

5. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 4, characterized in that, The determination of the data set under normal conditions and the data set under extreme conditions includes: Take the data set as the data set under normal circumstances, and take the difference set between the initial data set and the data set as the data set under extreme circumstances.

6. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 1, wherein, The determination of the optimal threshold for distinguishing whether the pressing process data belongs to the normal situation or the extreme situation through the maximum inter-class variance method includes: Obtain the upper quartile of the pressing process data in the dataset under normal conditions , obtain the lower quartile of the pressing process data in the dataset under extreme conditions ; Take each data within the range as a threshold, classify the suppression process data of all data points in the initial dataset. Classify the suppression process data greater than or equal to the threshold into one category, and classify the suppression process data less than the threshold into another category, and calculate the between-class variance of the two categories; take the threshold with the largest between-class variance as the optimal threshold.

7. The method for monitoring the service life of a forming die based on powder metallurgy metal materials according to claim 1, characterized in that, The equations for constructing each pressing process include: For any pressing process, input the pressing process data and duration of each data segment in this pressing process into the expression of the model to obtain the simulated values of each data segment; Construct the equations for each pressing process by making the simulated values of all data segments in this pressing process equal to the die life change data after the end of this pressing.

8. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 1, characterized in that, The obtaining of the die life change value of the to-be-evaluated molding die after the current pressing process includes: For any pressing process data: Input each pressing process data belonging to the normal situation into the pressing process data under normal conditions, and take the mean value of the pre-estimated die life change values corresponding to all the pressing process data belonging to the normal situation obtained as the pre-estimated die life change value of the current pressing process under normal conditions; Form data segments from the pressing process data at adjacent times among all the pressing process data belonging to the extreme situation, input the pressing process data and duration of each data segment into the pre-estimation model of the die life change data under the extreme situation, and take the sum value of the pre-estimated die life change values corresponding to all the data segments obtained as the pre-estimated die life change value of the current pressing process under the extreme situation; Take the maximum value between the pre-estimated die life change value of the current pressing process under normal conditions and the pre-estimated die life change value of the current pressing process under the extreme situation as the die life change value corresponding to this pressing process data; Take the mean value of the die life change values corresponding to all the pressing process data as the die life change value of the to-be-evaluated molding die after the current pressing process.

9. The method for monitoring the service life of a forming die based on a powder metallurgy metal material according to claim 1 or 8, characterized in that The pressing process data of the data segment is equal to the mean value of the pressing process data of all data points in the data segment; the method for obtaining the duration of the data segment is: calculate the product of the number of all pressing process data in the data segment and the acquisition time interval, and use the ratio of the product to as the duration of the data segment, where is equal to the average value of the durations of all pressing processes.

10. A forming die life monitoring system based on powder metallurgy metal materials, characterized in that, Including: A processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for monitoring the life of a molding die based on powder metallurgy metal materials according to any one of claims 1-9 is implemented.

Images