Search

CN-121980544-A - Assessment method and system for service life of die

CN121980544ACN 121980544 ACN121980544 ACN 121980544ACN-121980544-A

Abstract

The invention relates to the technical field of die life management, in particular to an evaluation method and system for die life, comprising the steps of collecting die multi-source data, and generating a structured multi-source data set after light-weight pretreatment; combining the service life characteristics and the maintenance state of the die, calculating the early warning proportion through curve fitting and a parameter fine adjustment algorithm, calculating the service life increasing parameter through a constraint matching formula, eliminating low-efficiency data, generating a high-precision evaluation parameter, constructing a double-time-domain prediction model, outputting the prediction results of short-term early warning risk and mid-term service life consumption trend, adopting a dynamic threshold algorithm to conduct abnormal judgment, correcting the core parameter through a data calibration mechanism according to the judgment result, generating an evaluation task adjustment instruction, generating system optimization data according to an instruction execution result, model calibration information and die operation data, and adapting management strategies of different scenes through a collaborative strategy model. The scheme can realize the accurate control of the service life of the die, and improves the production efficiency.

Inventors

  • LI XIAOCHUN

Assignees

  • 杭州友成科技有限公司

Dates

Publication Date
20260505
Application Date
20260408

Claims (9)

  1. 1. A method for evaluating die life, the method comprising: collecting multi-source data of a die, and generating a structured multi-source data set after light-weight pretreatment; based on the structured multi-source data set, combining the service life characteristics and the maintenance state of the die, calculating the early warning proportion through curve fitting and parameter fine tuning algorithm, calculating the service life increasing parameter through constraint matching formula, eliminating low-efficiency data, and generating high-precision evaluation parameter; constructing a double-time-domain prediction model based on the high-precision evaluation parameters, and outputting prediction results of short-term early warning risks and medium-term life consumption trends; performing abnormality judgment by adopting a dynamic threshold algorithm according to the prediction result, correcting core parameters through a data calibration mechanism according to the judgment result, generating an evaluation task adjustment instruction, and generating system optimization data according to an instruction execution result, model calibration information and mold operation data; and inputting the high-precision evaluation parameters, the system optimization data and the die use scene labels into a collaborative strategy model, and outputting management strategies adapting to different scenes.
  2. 2. The method for evaluating die life according to claim 1, wherein the specific acquisition process of the structured multi-source dataset is as follows: Collecting die operation, basic attribute, maintenance record, production order and scrapping data from production equipment, system ledgers and business records; The method comprises the steps of screening collected original data, removing invalid data caused by unnatural abrasion, unifying and regulating data formats, ensuring consistent unit and field naming of data from different sources, and reasonably completing missing key data to form a structured multi-source data set.
  3. 3. The method for evaluating the service life of a mold according to claim 2, wherein the specific obtaining process of the early warning ratio is as follows: The method comprises the steps of extracting data such as a historical accumulated combination module, a rated life, a unit die closing frequency abrasion loss, maintenance record and the like of a die from a structured multi-source data set, and determining the current life proportion and maintenance effectiveness state of the die; and combining the maintenance records, judging the replacement condition and the maintenance missing condition of the core vulnerable part, and carrying out fine adjustment on the initial early warning proportion through a preset algorithm to obtain the early warning proportion.
  4. 4. A method for evaluating die life according to claim 3, wherein the specific obtaining process of the high-precision evaluation parameter is as follows: based on the obtained early warning proportion and the current estimated life of the die, calculating the estimated life and the early warning proportion through a threshold calculation formula to obtain the minimum cumulative total modulus for triggering the die life estimation, namely the early warning trigger threshold; Determining a safe life-increasing upper limit through a material limit constraint logic by combining the rated life of the material and the accumulated sum modulus; And after unreasonable life-increasing demand data exceeding the upper limit of the safe life-increasing and invalid die closing data not conforming to the weight rule of the working condition are removed, integrating the early warning trigger threshold, the temporary life-increasing value and the next evaluation proportion to form high-precision evaluation parameters.
  5. 5. The method for evaluating the life of a mold according to claim 4, wherein the specific obtaining process of the prediction result is as follows: Extracting an early warning trigger threshold value, a temporary life-increasing value, a next evaluation proportion, the service life of the die after current evaluation and a real-time accumulated sum modulus from the high-precision evaluation parameters to be used as input data of a double-time-domain LSTM prediction model; constructing a short-term and medium-term double-time-domain LSTM prediction model based on the historical service life evaluation data, the mold closing trend data and the wear change data of the mold; And inputting the prepared input data into a trained double-time-domain LSTM prediction model, outputting a short-term early-warning triggering risk by a short-term time-domain model based on real-time accumulated mould closing acceleration and early-warning triggering threshold, and outputting a mid-term life consumption trend by a mid-term time-domain model based on the mould life consumption rate and the temporary life increasing value to obtain a complete prediction result.
  6. 6. The method for evaluating die life according to claim 5, wherein the specific acquisition process of the system optimization data is as follows: Extracting short-term early warning triggering risk and mid-term life consumption trend data from a prediction result, and simultaneously retrieving real-time die closing data, wear detection data and historical operation reference data of a die; Comparing the real-time die closing data and the wear data with a preset normal fluctuation threshold by adopting a dynamic threshold algorithm, and identifying abnormal data exceeding a threshold range; And extracting an early warning proportion calibration value and a working condition weight optimization value by combining an execution result of the evaluation task adjustment instruction, model month calibration information and the latest operation data of the die, and integrating to form system optimization data.
  7. 7. The method for evaluating the life of a mold according to claim 6, wherein the specific obtaining process of the management policy is as follows: Extracting an early warning trigger threshold value, a temporary life-increasing value and a next evaluation proportion from high-precision evaluation parameters, and calling an early warning proportion calibration value and a working condition weight optimization value in system optimization data, and simultaneously integrating an input data set of a collaborative strategy model by combining a die use field Jing Biaoqian; importing an input data set into a collaborative strategy model, performing weight distribution and logic operation on input parameters by the model based on a preset scene adaptation rule, and calculating personalized control coefficients adapted to different scenes; And generating a differentiated management strategy by combining the model life assessment flow specification and the early warning response mechanism framework according to the personalized control coefficient obtained by the model calculation.
  8. 8. An evaluation system for mold life, characterized in that the system is adapted to perform an evaluation method for mold life according to any one of claims 1-7.
  9. 9. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program, which is executed by a processor to implement a method for evaluating die life according to any of the preceding claims 1-7.

Description

Assessment method and system for service life of die Technical Field The invention relates to the technical field of mold life management, in particular to an evaluation method and system for mold life. Background In a modern manufacturing production system, a die is used as core equipment for realizing product molding, and the service life state of the die is directly related to production continuity, product qualification rate and comprehensive production cost, so that the die is a key control object for ensuring production efficiency of manufacturing enterprises. Along with the expansion of production scale and the improvement of product process complexity, the traditional die life management mode gradually shows various limitations, and the requirements on die management and control precision and efficiency in the intelligent manufacturing background can not be met. In the traditional management mode, the related data of the mould are stored in different carriers such as a production equipment control system, a manual standing book, an order management system and the like in a scattered manner, operation data such as the mould closing times, mould closing pressure and the like are recorded temporarily by a PLC (programmable logic controller), maintenance information is registered through a paper form or an Excel form, the production order requirement and the mould material parameter are respectively stored in an ERP system and a technical file, and the data format is not uniform and lacks a linkage mechanism. Meanwhile, the collected original data is not subjected to systematic pretreatment, invalid data generated by non-natural abrasion (such as die collapse and equipment abnormal impact) are not effectively screened, so that subsequent life assessment lacks standardized and high-quality structured data support, and accuracy of an assessment result is easily influenced by data deviation. In the process of early warning and life-prolonging calculation of the mold life, the existing mode relies on experience judgment of staff, the early warning proportion is set up to not combine the actual life consumption degree (such as life ratio and unit mold clamping abrasion amount) of the mold and the maintenance effect to establish quantitative calculation logic, the temporary life-prolonging value only refers to the requirement of a production order, the safety boundary of the rated life of the mold material is not fully considered, the problem that the mold fails in advance due to excessive life-prolonging or the production order cannot be delivered on schedule due to insufficient life-prolonging can occur, and a scientific double-constraint calculation system is lacked. In addition, the existing die life management strategies lack scene adaptation capability, do not differentiate different production order emergency degree, die operation load and other scenes, and all adopt a uniform evaluation period and early warning response mode, so that the requirements of fast response life adjustment cannot be met under the high emergency order scene, or the resource waste of excessive evaluation exists in low-load die management, the management precision and the production efficiency are difficult to balance, and the diversified manufacturing production requirements cannot be adapted. Disclosure of Invention According to the invention, through the full-period accurate management of the service life of the die, the accurate control of the service life of the die is realized, the production efficiency is improved, and the cost is reduced. The technical scheme provided by the invention is that the method for evaluating the service life of the die comprises the following steps: collecting multi-source data of a die, and generating a structured multi-source data set after light-weight pretreatment; based on the structured multi-source data set, combining the service life characteristics and the maintenance state of the die, calculating the early warning proportion through curve fitting and parameter fine tuning algorithm, calculating the service life increasing parameter through constraint matching formula, eliminating low-efficiency data, and generating high-precision evaluation parameter; constructing a double-time-domain prediction model based on the high-precision evaluation parameters, and outputting prediction results of short-term early warning risks and medium-term life consumption trends; performing abnormality judgment by adopting a dynamic threshold algorithm according to the prediction result, correcting core parameters through a data calibration mechanism according to the judgment result, generating an evaluation task adjustment instruction, and generating system optimization data according to an instruction execution result, model calibration information and mold operation data; and inputting the high-precision evaluation parameters, the system optimization data and the die use scene labels into a collaborati