A method for evaluating residual life of a forging die and warning die replacement

Abstract

This invention discloses a method for assessing the remaining life of forging dies and providing early warning of die replacement. The method includes: collecting and binding corresponding cycle data of the die based on the boundary of a single forging cycle; generating a state vector for a single forging cycle and identifying the working condition category; configuring a penalty coefficient, setting a dominant damage factor, and calculating an equivalent damage coefficient; calculating the single-cycle equivalent damage increment based on the equivalent damage coefficient and the basic damage amount of the standard cycle; recursively calculating the overall cumulative damage amount of the corresponding die; correcting the cumulative damage amount based on records of die repair, welding, polishing, cavity rework, or maintenance; performing a one-time retrospective compensation on the historical damage amount within the window; calculating the remaining life ratio based on the cumulative damage amount after retrospective compensation, and outputting a backup die warning, a die replacement warning, or a forced die replacement command. This method provides a life assessment that more closely approximates the actual damage process of the die and solves the timing misalignment problem between quality inspection lag and real-time damage recursion.

Description

Technical Field

[0001] This invention relates to the fields of forging die condition management, industrial data acquisition, predictive maintenance, and closed-loop control of production quality, and particularly to a method for assessing the remaining life of forging dies and providing early warning of die replacement. Background Technology

[0002] Forging dies are subjected to long-term high-temperature, high-load, repeated impact, and alternating lubrication / cooling environments. Typical failures include cavity wear, thermal fatigue cracks, localized collapse, adhesive wear, changes in flash clearance, demolding scratches, and dimensional deviations in forgings. When die failures are not identified in a timely manner, they can easily lead to batch forging dimensional deviations, insufficient filling, abnormal flash, surface indentations, difficulty in demolding, and insufficient subsequent machining allowances.

[0003] Current management methods primarily rely on fixed forging cycles, manual visual inspection, or experience-based mold replacement. However, the fixed counting method assumes that each forging cycle causes the same damage, failing to distinguish between normal cycles and high-damage cycles such as overload, overtemperature, prolonged holding pressure, cold mold impact, and abnormal demolding. Manual judgment is heavily influenced by experience and shift differences, making it difficult to identify early, latent damage in a timely manner. Some solutions attempt to use finite element simulation, wear models, or historical data fitting for life prediction, but in actual production, molds undergo multiple repairs, welding, polishing, cavity rework, and abnormal impact conditions, resulting in discrepancies between theoretical lifespan and actual on-site conditions. Simply recounting after mold repair overestimates the usable lifespan of the mold, while replacing molds solely based on historical average lifespan leads to underutilization of the mold.

[0004] Therefore, there is a need for an on-site early warning method that can identify abnormal working conditions online using a single forging cycle as the evaluation unit, calculate equivalent damage according to the type of abnormality, and continuously correct the remaining service life by combining local sensitive areas, mold repair records, and product quality feedback. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for assessing the remaining life of forging dies and providing early warning of die replacement based on forging cycle condition identification, equivalent damage calculation of abnormal cycles, accumulation of local damage fields in the cavity, correction of non-zeroing of die repair records, and backward traceability compensation based on quality feedback. This method avoids the deviation caused by simply estimating the die life based on the number of forging passes, enabling the die life assessment to truly reflect the load, temperature, holding pressure, preheating, demolding, die repair, and product quality status during the forging process. Furthermore, it can trigger backup die, die replacement, or forced die replacement commands in advance when there is rapid failure in local areas or concentrated occurrence of abnormal cycles.

[0006] This invention is achieved using the following technical solution: A method for assessing the remaining life of forging dies and providing early warning of die replacement includes the following steps: Step S1: Using the forging equipment completion signal or forging equipment cycle signal as the boundary of a single forging cycle, collect the cycle data corresponding to the current mold, and bind the mold information with the data of each cycle; Step S2: Generate a single forging cycle state vector based on the cycle data, and identify the working condition category based on a preset process window, statistical threshold, or trained anomaly detection model; Step S3: Configure penalty coefficients for different abnormal working condition types, extract the one with the largest value from the penalty coefficients of the current cycle as the dominant damage factor, normalize and saturate the penalty coefficients corresponding to the other abnormalities to obtain the equivalent damage coefficient, and calculate the single equivalent damage increment based on the equivalent damage coefficient and the basic damage amount of the standard cycle. Step S4: Based on the single equivalent damage increment, calculate the overall cumulative damage of the corresponding mold, divide the mold cavity into multiple local sensitive areas according to the differences in stress concentration and temperature distribution, and map and distribute the overall damage increment to each local sensitive area to obtain the cumulative damage of the local sensitive area. Step S5: Correct the cumulative damage amount based on the mold repair, welding repair, polishing, cavity rework or maintenance records; Step S6: After receiving a quality anomaly record with a timestamp, automatically retrieve a preset number of forging cycle windows before that time point, and perform a one-time retrospective supplementary compensation for the historical damage amount within that window; Step S7: Calculate the remaining lifetime ratio based on the cumulative damage amount after retrospective compensation and the lifetime damage threshold, and output a backup mold warning, a mold replacement warning, or a forced mold replacement command.

[0007] Specifically, the forging equipment completion signals in step S1 include forging completion signal, slide stroke completion signal, and hydraulic press main cylinder positioning signal; the cycle data includes, but is not limited to, peak pressure, holding time, die working surface temperature, die preheating temperature, demolding status, number of forgings, and forging quality feedback data; The bound mold information includes mold number, cavity number, product model, process card number, and production batch; the binding of mold information with each cycle data also includes: for a multi-cavity mold or a separate upper and lower mold structure, establishing the overall mold damage amount, cavity local damage amount, and mold component damage amount respectively, so that the data generated in each cycle belongs to the corresponding mold, corresponding cavity, or corresponding mold component.

[0008] Specifically, the state vector of a single forging cycle in step S2 is represented as follows: ; in, For the first Peak pressure of the next forging cycle For the first Holding time of each forging cycle For the first Temperature of the working surface of the die during each forging cycle For the first Die preheating temperature for the next forging cycle For the first The demolding status or demolding resistance index of the second forging cycle. For the first Forging quality feedback indicators for each forging cycle; The identified operating condition categories include, but are not limited to, normal cycle, overload cycle, long pressure holding cycle, over-temperature cycle, cold mold impact cycle, demolding abnormal cycle, and quality abnormality associated cycle.

[0009] Specifically, the penalty coefficient configured in step S3 includes: ; ; ; ; ; in, This is the pressure penalty coefficient. This is the pressure holding penalty coefficient. This is the temperature penalty coefficient. This is the cold mold impact penalty coefficient. The penalty coefficients are for the demolding state. Each penalty coefficient is a monotonic function of its corresponding independent variable parameter. The larger the deviation of the independent variable from the process reference value, the larger the value of the penalty coefficient. The functional form of each penalty coefficient is determined by regression calibration of historical damage events and cyclic data. As the process reference pressure, Standard holding time, and These are the upper and lower limits of the working face temperature, respectively. The lower limit of the preheating temperature. This represents the theoretical upper limit for demolding. , , , , , , , , , , These are the coefficients calibrated using historical data, trial model data, or simulation data for the corresponding abnormal operating condition types; The single-instance equivalent damage increment is expressed as: The calculation formula is: ; ; ; ; in, As the dominant damage factor, The equivalent damage coefficient, Indicates the standard cyclic baseline damage amount; This refers to the design life or historical qualified life of the mold under standard operating conditions. This is the threshold for lifespan impairment. The maximum gain coefficient of the secondary factor, 0 ≤ ≤1; These are the secondary factor weights.

[0010] Specifically, the total cumulative damage to the mold in step S4 is expressed as follows: ; The cumulative damage in the localized sensitive area is expressed as: ; ; in, This indicates locally sensitive areas, including but not limited to the edge bridge area, chamfered corner stretching area, and bottom extrusion area; , , These represent the load concentration, temperature concentration, and demolding friction influence coefficients of the local sensitive area, respectively; when the cumulative damage in any area first reaches the preset limit damage threshold corresponding to that area, an early warning or mold change action is triggered.

[0011] Specifically, the correction of the cumulative damage amount in step S5 is expressed as follows: ; ; in, This indicates the percentage of damage retained after mold repair. This indicates the cumulative damage amount before mold repair. This represents the cumulative damage amount after mold repair. The cumulative damage amount after mold repair is reduced by the proportion of remaining damage after mold repair, and the proportion of remaining damage increases non-linearly with the increase of the cumulative number of mold repairs in history. This is a truncation function. To accumulate the number of mold repairs, For the depth of this mold repair, To allow for the maximum mold repair depth, This refers to the critical dimensional deviations after mold repair. The upper limit of allowable deviation; the smaller the dimensional deviation between the critical dimension after mold repair and the original state, the lower the proportion of retained damage; the more times the mold is repaired or the greater the dimensional deviation, the higher the proportion of retained damage.

[0012] Specifically, the retrospective supplementary compensation for the loss in step S6 is expressed as follows: ; in, To retrospectively calculate the cumulative damage after additional compensation, The quality is abnormal. The traceability window length is determined based on the quality anomaly level. Preset value, quality anomaly level The higher the value, the longer the traceability window. The larger; For the first in the window The time decay weight in the next cycle, and the time decay weight By cycle number Decreasing from the nearest to the furthest point in time, i.e. ; This is the compensation coefficient for quality anomalies.

[0013] Specifically, the formula for calculating the remaining lifetime ratio in step S7 is: ; Among them, the warning threshold calls up different threshold configuration parameter tables according to mold type, forging value, line stop cost, product size accuracy requirements and mold number; Step S7 also includes the calculation of the proportion of recent abnormal cycles, expressed as: ; in, The length of the sliding window. For the first The operating condition category of the next cycle. This is a normal cycle; when the proportion of recent abnormal cycles... When the proportion exceeds the preset high-frequency threshold or rises rapidly within a unit of time, even if the remaining lifespan proportion... Even if the mold replacement threshold has not been reached, it is determined that the mold is at risk of rapid failure and a reinforcement warning or forced mold replacement instruction is issued in advance.

[0014] The beneficial effects of this invention are as follows: This solution starts with the identification of forging cycle boundaries, comparing the pressure, holding pressure, temperature, preheating, demolding, and quality signals during each forging process with the process window to automatically distinguish between normal cycles and various abnormal cycles. For identified abnormal cycles, a penalty coefficient is generated according to the abnormality type, and an equivalent damage coefficient is obtained through the extraction of the dominant cause and the saturation synthesis of secondary factors, thereby calculating the single-cycle equivalent damage increment. The equivalent damage increment is accumulated as the total damage of the entire mold on the one hand, and as the local damage amount according to the local sensitive areas of the cavity on the other hand. When the mold undergoes mold repair, welding repair, polishing, or cavity rework, the accumulated damage amount is non-zeroed based on the mold repair depth, the number of mold repairs, and the critical dimension deviation. When feedback of forging quality abnormalities or subsequent hidden defects is received, damage compensation is added to the historical cycle window before the abnormality occurred through a backward traceability compensation mechanism, and the remaining life ratio is recalculated. Finally, combining the remaining life ratio, the local area damage threshold, and the recent abnormal cycle ratio, a warning for mold preparation, mold replacement, or forced mold replacement is output. For the life assessment of forging dies, this solution also has the following advantages: (1) Compared with simply counting by the number of forgings, this solution identifies the working condition category of each forging cycle and amplifies the damage increment according to the abnormality type, making the life assessment closer to the actual damage process of the mold. (2) This scheme uses the method of extracting the dominant cause and saturating the secondary factor to calculate the equivalent damage coefficient, avoiding the damage value divergence caused by multiplication when multiple anomalies occur concurrently, and the algorithm stability is significantly improved. (3) This scheme divides the mold cavity into multiple local sensitive areas and accumulates the local damage amount separately, so that the life evaluation is upgraded from the whole mold homogeneous assumption to the multi-danger area damage field evaluation, and the evaluation results are more accurate. (4) The cumulative damage after mold repair is not cleared in this scheme, and the proportion of damage retained increases non-linearly with the cumulative number of mold repairs, so as to avoid the risk of failure caused by managing the life of a new mold after multiple repairs. (4) This scheme adopts a quality feedback backward traceability compensation mechanism, which solves the problem of timing misalignment between quality detection lag and real-time damage recursion; (5) This solution is also configured to determine that the mold is at risk of rapid failure when the remaining life ratio has not reached the mold replacement threshold but abnormal cycles occur in a concentrated manner, and to output a strengthening warning or forced mold replacement command in advance to reduce the risk of sudden shutdown and batch quality. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0016] Figure 1 This is an overall flowchart of a method for assessing the remaining life of a forging die and providing early warning of die replacement, as described in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0018] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0019] The following is in conjunction with the appendix Figure 1 The following describes some embodiments of the present invention in detail. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0020] This invention proposes a method for assessing the remaining life of forging dies and providing early warning of die replacement. In a preferred embodiment, the overall process of this method is as follows: Figure 1 As shown, it includes: Step S1: Using the forging equipment completion signal or forging equipment cycle signal as the boundary of a single forging cycle, collect the cycle data corresponding to the current mold, and bind the mold information with the data of each cycle; Step S2: Generate a single forging cycle state vector based on the cycle data, and identify the working condition category based on a preset process window, statistical threshold, or trained anomaly detection model; Step S3: Configure penalty coefficients for different abnormal working condition types, extract the one with the largest value from the penalty coefficients of the current cycle as the dominant damage factor, normalize and saturate the penalty coefficients corresponding to the other abnormalities to obtain the equivalent damage coefficient, and calculate the single equivalent damage increment based on the equivalent damage coefficient and the basic damage amount of the standard cycle. Step S4: Based on the single equivalent damage increment, calculate the overall cumulative damage of the corresponding mold, divide the mold cavity into multiple local sensitive areas according to the differences in stress concentration and temperature distribution, and map and distribute the overall damage increment to each local sensitive area to obtain the cumulative damage of the local sensitive area. Step S5: Correct the cumulative damage amount based on the mold repair, welding repair, polishing, cavity rework or maintenance records; Step S6: After receiving a quality anomaly record with a timestamp, automatically retrieve a preset number of forging cycle windows before that time point, and perform a one-time retrospective supplementary compensation for the historical damage amount within that window; Step S7: Calculate the remaining lifetime ratio based on the cumulative damage amount after retrospective compensation and the lifetime damage threshold, and output a backup mold warning, a mold replacement warning, or a forced mold replacement command.

[0021] The following is a detailed explanation of each step: (1) Data acquisition and object binding In this embodiment, this step uses the completion signal of the forging equipment (including forging completion signal, slide stroke completion signal, and hydraulic press main cylinder positioning signal) or the equipment cycle signal as the boundary of a single forging cycle, and collects cycle data corresponding to the current mold. The cycle data includes at least several of the following: peak pressure, holding time, mold working surface temperature, mold preheating temperature, demolding state, number of forging cycles, and forging quality feedback data. In this embodiment, the mold preheating temperature measurement point can be obtained by infrared thermal imaging scanning of the cavity working surface, or by arranging armored thermocouples at a depth of 5-15 mm near the mold base / cavity, preferably using a combination of multi-point averaging and minimum point dual constraints.

[0022] After collecting the cyclic data, the mold number, cavity number, product model, process card number, material batch, production batch, and shift information are bound to the data for each cycle, so that the data generated in each cycle belongs to the corresponding mold, cavity, or mold assembly. In this embodiment, for a multi-cavity mold or a split upper / lower mold structure, the overall mold damage, local cavity damage, and mold assembly damage can be established separately.

[0023] (2) Identification of working condition categories In this embodiment, this step generates a single forging cycle state vector based on the collected cyclic data, and identifies the working condition category according to a preset process window, moving statistical threshold, or trained anomaly detection model. The working condition categories include at least normal cycle, overload cycle, long holding pressure cycle, overtemperature cycle, cold die impact cycle, demolding anomaly cycle, and quality anomaly-related cycle. The single forging cycle state vector is represented as follows: in, For the first The peak pressure of each forging cycle is obtained through pressure sensors, press PLC, and main cylinder pressure curve. For the first The holding time for each forging cycle is obtained by measuring the duration during which the pressure is higher than the peak pressure by a set percentage. For the first The working surface temperature of the die during each forging cycle was obtained through infrared thermometry, thermal imager, and thermocouples near the cavity. For the first The preheating temperature of the die for each forging cycle is obtained by thermocouple inside the die holder or by scanning temperature on the working surface. For the first The demolding status or demolding resistance index of each forging cycle is obtained through the demolding mechanism current, demolding pressure, and demolding time. For the first The forging quality feedback indicators for each forging cycle are obtained through dimensional inspection, flash thickness inspection, visual inspection, or subsequent feedback.

[0024] In this embodiment, overload cycles can be identified by a set percentage in which the peak pressure exceeds the process reference pressure or the moving average pressure of the same batch; long holding pressure cycles can be identified by a duration in which the pressure is higher than the set percentage in which the peak pressure is exceeded for more than the standard holding time; overheating cycles can be identified by a mold working surface temperature exceeding the upper limit temperature or a heating rate per unit time exceeding a threshold; cold mold impact cycles can be identified by a mold preheating temperature lower than the preheating lower limit and a forging cycle has already occurred; abnormal demolding cycles can be identified by a demolding mechanism current, demolding pressure, or demolding time exceeding the corresponding threshold.

[0025] (3) Calculation of single equivalent damage increment First, penalty coefficients are configured for different anomaly types; for loops without corresponding anomalies, the penalty coefficient is set to 1; the higher the anomaly severity, the larger the corresponding penalty coefficient. To avoid damage value divergence caused by simple multiplication when multiple anomalies occur concurrently, this invention adopts a "dominant cause extraction and secondary factor saturation synthesis" mechanism; in this embodiment, the penalty coefficient corresponding function is calibrated by regression testing of pressure, holding pressure, temperature, preheating, demolding data, and mold failure / quality anomaly records, or it can be initially assigned by finite element simulation or process experts; the penalty coefficients configured for each anomaly type are as follows: Stress penalty coefficient: ; Pressure holding penalty coefficient: ; Temperature penalty coefficient: ; Cold mold impact penalty coefficient: ; Demolding resistance penalty coefficient: ; in, As the process reference pressure, Standard holding time, and These are the upper and lower limits of the working face temperature, respectively. The lower limit of the preheating temperature. This represents the theoretical upper limit for demolding. , , , , , , , , , , These are the coefficients calibrated using historical data, trial data, or simulation data for the corresponding abnormal operating condition types.

[0026] In one embodiment, each penalty coefficient is a monotonic function of its corresponding independent variable parameter. The larger the deviation of the independent variable from the process baseline, the larger the value of the penalty coefficient. The functional form of each penalty coefficient is determined through regression calibration of historical damage events and cyclic data. Each penalty coefficient can be used as an initial model in the following form: pressure penalty coefficient, , here The pressure overload sensitivity coefficient is used, and the other penalty coefficients adopt a similar structure. Each coefficient is calibrated by regression analysis of the historical failure events of the mold and the corresponding cyclic data. If the production line lacks sufficient historical data, the initial values ​​can be assigned using finite element simulation results, and then iteratively corrected according to the actual failure events after production.

[0027] In the "dominant cause extraction and secondary factor saturation synthesis" mechanism of this scheme, the calculation formula for the dominant damage factor is: ; The formula for calculating the equivalent damage coefficient is: ; This leads to the standard cyclic baseline damage amount. Increment of single equivalent damage The formula for calculation is: ; ; in, The standard cycle basic damage amount is represented in this embodiment. The standard cycle basic damage amount is calculated by back-calculating the design life, historical qualified life, or trial mold to failure data. If there is no historical data, it is first set according to the design life and then corrected by actual failure events after production. This refers to the design life or historical qualified life of the mold under standard operating conditions. This is the threshold for lifespan impairment. The maximum gain coefficient of the secondary factor, 0 ≤ ≤1; The secondary factor weights are used. In this embodiment, the secondary factors adopt a saturated fractional structure rather than a multiplicative structure, so that the additional contribution of concurrent anomalies to the damage has diminishing marginal returns, avoiding numerical divergence.

[0028] (4) Cumulative damage amount and cumulative damage amount in local sensitive areas In this embodiment, the total cumulative damage to the mold is calculated using a recursive formula: ; After calculating the overall cumulative damage to the mold, the mold cavity is divided into multiple locally sensitive areas based on stress concentration and temperature distribution differences, including at least the flash bridge area, the chamfer and corner stretching area, and the bottom extrusion area. Based on the finite element stress concentration factor, thermal imaging data, empirical weighting of the cavity structure, or historical failure location statistics, the overall damage increment is mapped and distributed to each local area to obtain the cumulative damage to the locally sensitive areas. Cumulative damage in localized sensitive areas The calculation of the total cumulative damage corresponds to the calculation of the overall cumulative damage amount, both of which are calculated using a recursive formula, expressed as: ; ; in, Indicates a localized sensitive area; , , These represent the load concentration, temperature concentration, and demolding friction influence coefficients of the localized sensitive area, respectively. In this embodiment, these coefficients are determined through finite element analysis of the mold structure, thermal image distribution, and statistical analysis of historical crack / wear locations. Coefficients are created for the edge bridge section, corner area, and bottom extrusion area, respectively. When the cumulative damage in any area first reaches the corresponding limit damage threshold, an early warning or mold change action is triggered.

[0029] (5) Modification of mold repair records without clearing In this step, the cumulative damage amount or lifespan threshold is corrected based on records of mold repair, welding, polishing, cavity rework, or maintenance. The cumulative damage amount after mold repair is not reset to zero; instead, it is reduced by the proportion of damage retained after repair. This proportion of retained damage increases non-linearly with the historical cumulative number of mold repairs. The corrected cumulative damage amount after mold repair is expressed as: ; ; in, This indicates the percentage of damage retained after mold repair. This indicates the cumulative damage amount before mold repair. This represents the cumulative damage amount after mold repair. The cumulative damage amount after mold repair is reduced by the proportion of remaining damage after mold repair, and the proportion of remaining damage increases non-linearly with the increase of the cumulative number of mold repairs in history. This is a truncation function. To accumulate the number of mold repairs, For the depth of this mold repair, To allow for the maximum mold repair depth, This refers to the critical dimensional deviations after mold repair. The upper limit of allowable deviation is set. In this embodiment, the closer the critical dimension is to the original state after mold repair, the lower the proportion of retained damage can be (that is, the smaller the dimensional deviation between the critical dimension after mold repair and the original state, the lower the proportion of retained damage); the more times the mold is repaired or the greater the dimensional deviation, the higher the proportion of retained damage.

[0030] (6) Quality feedback and backward traceability compensation The quality inspection results of forgings are usually lagging behind the forging cycle itself. Therefore, this solution does not directly incorporate the quality feedback coefficient into the real-time damage extrapolation of a single cycle. Instead, after receiving a quality anomaly record with a timestamp, it automatically retrieves a preset number of forging cycle windows before that time point and performs a one-time retrospective compensation for the historical damage amount within that window.

[0031] The backward traceability compensation for quality anomalies is represented as follows: ; in, To retrospectively calculate the cumulative damage after additional compensation, The quality is abnormal. The traceability window length is determined based on the quality anomaly level. Preset value, quality anomaly level The higher the value, the longer the traceability window. The larger; For the first in the window The time decay weight in the next cycle, and the time decay weight By cycle number Decreasing from the nearest to the furthest point in time, i.e. ; This is the compensation coefficient for quality anomalies.

[0032] In this embodiment, quality anomalies include dimensional deviations, abnormal flash thickness, surface defects, non-conforming samples, or latent defects revealed by subsequent machining / assembly. Sources of quality feedback include flash thickness detection, coordinate measuring machine (CMM) inspection, visual surface defect detection, and tool wear or assembly clearance trends during subsequent machining. Traceability window length. The configuration is based on the following: For dimensional quality anomalies (such as excessive burr thickness or critical dimension deviations), the traceability window length is set to 5-20 cycles to cover the typical hysteresis period for anomalies to propagate to detectable quality deviations; for surface quality anomalies (such as cracks or indentations), the length is set to 3-10 cycles. Time decay weight. Linear attenuation can be used Or, in an exponential decay form, the cycle closer to the quality anomaly detection time point receives higher damage compensation. (7) Remaining life calculation and graded early warning In this step, the formula for calculating the remaining lifetime percentage is: ; Based on the proportion of remaining lifespan Outputs pre-warning messages for mold standby, mold change, or forced mold change. The warning thresholds are not fixed percentages, but rather configured using different threshold parameter tables based on mold type, forging value, line stoppage cost, product dimensional accuracy requirements, and mold number.

[0033] Recent Abnormal Cycle Ratio ; in, The length of the sliding window. For the first The operating condition category of the next cycle. This is a normal cycle; when the proportion of recent abnormal cycles... When the frequency exceeds the high-frequency threshold or rises rapidly within a unit of time, even the remaining lifetime ratio Even if the mold replacement threshold has not been reached, it is determined that the mold is at risk of rapid failure and a reinforcement warning or forced mold replacement instruction is issued in advance.

[0034] This solution also provides specific application examples. In this embodiment, the forging object is a hot forging die for an automotive differential gear ring. The press slide is equipped with a pressure acquisition unit, and a thermocouple or infrared thermometer is placed near the die cavity. After discharge, a flash thickness detection and dimensional sampling station is set up. The controller uses the press forging completion signal as the cycle boundary, and records the peak pressure, holding time, die working surface temperature, die preheating temperature, demolding status, and quality feedback data in each forging cycle.

[0035] Before production begins, read the mold number, product model, process card number, and production batch, and bind these identifiers to the data for each forging cycle. For a mold containing multiple cavities, establish cumulative damage amounts according to the cavity number, and simultaneously establish local damage amounts for the flash bridge, corner stretching area, and bottom extrusion area.

[0036] During normal cycles, all parameters are within the process window, and the cumulative damage is calculated based on the standard cycle baseline damage. If the peak pressure of a cycle exceeds the set percentage of the process reference pressure, it is identified as an overload cycle and the pressure penalty coefficient is increased. If the pressure is maintained above the set percentage of the peak pressure for more than the standard holding time, it is identified as a long holding cycle and the holding penalty coefficient is increased. If the die working surface temperature exceeds the process upper limit or the heating rate exceeds the threshold, it is identified as an overheat cycle and the temperature penalty coefficient is increased. If forging begins before the die reaches the lower limit of the preheating temperature, it is identified as a cold die impact cycle and the cold die impact penalty coefficient is increased. If the current, pressure, or demolding time of the demolding mechanism exceeds the threshold, it is identified as an abnormal demolding cycle and the demolding resistance penalty coefficient is increased.

[0037] The maximum value is extracted from the aforementioned coefficients as the dominant damage factor, and the remaining coefficients are used as secondary damage interference variables for saturation synthesis to obtain the equivalent damage coefficient. After multiplying by the standard cycle basic damage amount, the single-cycle equivalent damage increment is obtained and simultaneously accumulated to the overall model damage amount and the local area damage amount.

[0038] When the forging quality inspection results show dimensional deviations, abnormal flash thickness, or failure to pass the spot check, the historical damage amount within a preset number of cyclic windows before the time point of the quality anomaly will be retrospectively compensated and the remaining service life will be recalculated. If quality anomalies occur continuously, a mold replacement warning will be issued in advance even if the remaining service life ratio has not yet reached the mold replacement threshold.

[0039] After mold repair, operators record the repair time, repair depth, repair location, number of repairs, and the results of critical dimension inspections after repair. The cumulative damage is reduced based on the retained damage percentage, but not reset to zero. The retained damage percentage increases non-linearly with the cumulative number of repairs; for example, the retained damage percentage after the first repair may be lower than that after the third repair.

[0040] The remaining lifespan percentage is calculated based on the corrected cumulative damage, and a tiered warning is output. The warning threshold can be configured according to the mold type; for example, a higher backup and replacement threshold can be used for precision molds than for ordinary molds. The system also dynamically tracks the proportion of abnormal cycles in the most recent cycles. When overload, overtemperature, or demolding abnormalities occur frequently within a short window, a reinforced warning is triggered even if the remaining lifespan percentage has not reached the replacement threshold.

[0041] For the foregoing embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.

[0042] The above embodiments describe the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the invention should be within the protection scope of the appended claims.

Claims

1. A method for assessing the remaining life of forging dies and providing early warning of die replacement, characterized in that, Includes the following steps: Step S1: Using the forging equipment completion signal or forging equipment cycle signal as the boundary of a single forging cycle, collect the cycle data corresponding to the current mold, and bind the mold information with the data of each cycle; Step S2: Generate a single forging cycle state vector based on the cycle data, and identify the working condition category based on a preset process window, statistical threshold, or trained anomaly detection model; Step S3: Configure penalty coefficients for different abnormal working condition types, extract the one with the largest value from the penalty coefficients of the current cycle as the dominant damage factor, normalize and saturate the penalty coefficients corresponding to the other abnormalities to obtain the equivalent damage coefficient, and calculate the single equivalent damage increment based on the equivalent damage coefficient and the basic damage amount of the standard cycle. Step S4: Based on the single equivalent damage increment, calculate the overall cumulative damage of the corresponding mold, divide the mold cavity into multiple local sensitive areas according to the differences in stress concentration and temperature distribution, and map and distribute the overall damage increment to each local sensitive area to obtain the cumulative damage of the local sensitive area. Step S5: Correct the cumulative damage amount based on the mold repair, welding repair, polishing, cavity rework or maintenance records; Step S6: After receiving a quality anomaly record with a timestamp, automatically retrieve a preset number of forging cycle windows before that time point, and perform a one-time retrospective supplementary compensation for the historical damage amount within that window; Step S7: Calculate the remaining lifetime ratio based on the cumulative damage amount after retrospective compensation and the lifetime damage threshold, and output a backup mold warning, a mold replacement warning, or a forced mold replacement command.

2. The method for assessing the remaining life of forging dies and providing early warning of die replacement as described in claim 1, characterized in that, The forging equipment completion signals in step S1 include forging completion signal, slide stroke completion signal, and hydraulic press main cylinder positioning signal; the cycle data includes, but is not limited to, peak pressure, holding time, die working surface temperature, die preheating temperature, demolding status, number of forgings, and forging quality feedback data; The bound mold information includes mold number, cavity number, product model, process card number, and production batch. Binding mold information to data for each cycle also includes: for multi-cavity molds or separate upper and lower mold structures, establishing the overall mold damage amount, local cavity damage amount, and mold component damage amount respectively, so that the data generated in each cycle belongs to the corresponding mold, corresponding cavity, or corresponding mold component.

3. The method for assessing the remaining life of forging dies and providing early warning of die replacement as described in claim 2, characterized in that, The state vector for a single forging cycle in step S2 is represented as follows: ； in, For the first Peak pressure of the next forging cycle For the first Holding time of each forging cycle For the first Temperature of the working surface of the die during each forging cycle For the first Die preheating temperature for the next forging cycle For the first The demolding status or demolding resistance index of the second forging cycle. For the first Forging quality feedback indicators for each forging cycle; The identified operating condition categories include, but are not limited to, normal cycle, overload cycle, long pressure holding cycle, over-temperature cycle, cold mold impact cycle, demolding abnormal cycle, and quality abnormality associated cycle.

4. The method for assessing the remaining life of forging dies and providing early warning of die replacement as described in claim 3, characterized in that, The penalty coefficient configured in step S3 includes: ； ； ； ； ； in, This is the pressure penalty coefficient. This is the pressure holding penalty coefficient. This is the temperature penalty coefficient. This is the cold mold impact penalty coefficient. The penalty coefficients are for the demolding state. Each penalty coefficient is a monotonic function of its corresponding independent variable parameter. The larger the deviation of the independent variable from the process reference value, the larger the value of the penalty coefficient. The functional form of each penalty coefficient is determined by regression calibration of historical damage events and cyclic data. As the process reference pressure, Standard holding time, and These are the upper and lower limits of the working face temperature, respectively. The lower limit of the preheating temperature. This represents the theoretical upper limit for demolding. , , , , , , , , , , These are the coefficients calibrated using historical data, trial model data, or simulation data for the corresponding abnormal operating condition types; The single-instance equivalent damage increment is expressed as: The calculation formula is: ； ； ； ； in, As the dominant damage factor, The equivalent damage coefficient, Indicates the standard cyclic baseline damage amount; This refers to the design life or historical qualified life of the mold under standard operating conditions. This is the threshold for lifespan impairment. The maximum gain coefficient of the secondary factor, 0 ≤ ≤1; These are the secondary factor weights.

5. The method for assessing the remaining life of a forging die and providing early warning of die replacement as described in claim 4, characterized in that, The total cumulative damage to the corresponding mold in step S4 is expressed as follows: ； The cumulative damage in the localized sensitive area is expressed as: ； ； in, This indicates locally sensitive areas, including but not limited to the edge bridge area, chamfered corner stretching area, and bottom extrusion area; , , These represent the load concentration, temperature concentration, and demolding friction influence coefficients of the local sensitive area, respectively; when the cumulative damage in any area first reaches the preset limit damage threshold corresponding to that area, an early warning or mold change action is triggered.

6. The method for assessing the remaining life of a forging die and providing early warning of die replacement as described in claim 5, characterized in that, The correction for the cumulative damage in step S5 is expressed as follows: ； ； in, This indicates the percentage of damage retained after mold repair. This indicates the cumulative damage amount before mold repair. This represents the cumulative damage amount after mold repair. The cumulative damage amount after mold repair is reduced by the proportion of remaining damage after mold repair, and the proportion of remaining damage increases non-linearly with the increase of the cumulative number of mold repairs in history. This is a truncation function. To accumulate the number of mold repairs, For the depth of this mold repair, To allow for the maximum mold repair depth, This refers to the critical dimensional deviations after mold repair. The upper limit of allowable deviation; the smaller the dimensional deviation between the critical dimension after mold repair and the original state, the lower the proportion of retained damage; the more times the mold is repaired or the greater the dimensional deviation, the higher the proportion of retained damage.

7. The method for assessing the remaining life of a forging die and providing early warning of die replacement as described in claim 6, characterized in that, The retrospective supplementary compensation for the loss in step S6 is expressed as follows: ； in, To retrospectively calculate the cumulative damage after additional compensation, The quality is abnormal. The traceability window length is determined based on the quality anomaly level. Preset value, quality anomaly level The higher the value, the longer the traceability window. The larger; For the first in the window The time decay weight in the next cycle, and the time decay weight By cycle number Decreasing from the nearest to the furthest point in time, i.e. ; This is the compensation coefficient for quality anomalies.

8. The method for assessing the remaining life of a forging die and providing early warning of die replacement as described in claim 7, characterized in that, The formula for calculating the remaining lifespan percentage in step S7 is as follows: ； Among them, the warning threshold calls up different threshold configuration parameter tables according to mold type, forging value, line stop cost, product size accuracy requirements and mold number; Step S7 also includes the calculation of the proportion of recent abnormal cycles, expressed as: ； in, The length of the sliding window. For the first The operating condition category of the next cycle. This is a normal cycle; when the proportion of recent abnormal cycles... When the proportion exceeds the preset high-frequency threshold or rises rapidly within a unit of time, even if the remaining lifespan proportion... Even if the mold replacement threshold has not been reached, it is determined that the mold is at risk of rapid failure and a reinforcement warning or forced mold replacement instruction is issued in advance.

Images