Mold full life cycle management method, system, electronic equipment and storage medium

Abstract

The present application discloses a method, system, electronic device and storage medium for managing the entire life cycle of a mold, wherein the method includes: optimizing the mold design using a finite element simulation method during the mold design phase, and using finite element simulation results to assist in determining the selection and installation location of sensors based on the forming process conditions and range, so as to monitor the performance of the mold and the forming system during use online by installing sensors on the mold and its forming system; judging the online status of the mold and the forming system during use and the online health of the mold based on the performance data monitored online, and predicting the future health of the mold based on the online health of the mold over multiple cycles; and adjusting the mold maintenance strategy based on the offline status of the mold during the out-of-use period detected by the offline detection method. Through this application, active mold performance management is achieved, and the performance stability and efficiency of the mold during the entire life cycle of use are improved.

Description

Technical Field

[0001] The present application belongs to the technical field of production management, and more specifically, to a mold full life cycle management method, system, electronic device and storage medium. Background Art

[0002] The mold is one of the key tools in production. The quality and performance of the mold directly affect the quality and production cost of the product. There are many links and factors that affect the quality and performance of the mold, including the design, processing and mold trial in the mold factory, the use, maintenance and mold repair of the mold in the manufacturing factory. In the design stage, the designer will optimize the structure to ensure the design usage or life, design performance. The staff further ensures the performance of the factory mold through processing, mold trial and mold modification. Qualified molds will enter the molding production link. In this process, the operation and maintenance personnel will provide maintenance, maintenance and necessary repairs for the mold to ensure that the performance of the mold meets the production requirements.

[0003] The existing mold operation and maintenance methods usually provide regular maintenance and upkeep by the operation and maintenance personnel based on their experience and operation and maintenance procedures under normal circumstances. When the mold performance is abnormal, this is usually due to the identification of product quality problems caused by the abnormal mold performance. The operation and maintenance personnel first determine whether the product quality problem is related to the mold based on their experience. If it is related, the type and degree of the mold performance abnormality are further determined, and then corresponding maintenance measures are formulated based on experience or procedures. In extreme cases, the mold structure is damaged during use, which will also trigger the maintenance or repair of the mold. It can be seen that the existing method is untimely and passive. The handling of abnormalities needs to be based on the occurrence of product quality problems, or even batch product quality problems. More seriously, it can only be triggered by the occurrence of faults such as mold damage. This not only reduces production efficiency and increases production costs, but also increases the cost of mold maintenance and repair due to untimely handling. Summary of the invention

[0004] In view of the defects of the prior art, the purpose of this application is to provide a mold full life cycle management method, system, electronic device and storage medium, aiming to solve the problem of increased production costs and reduced production efficiency caused by the existing untimely and passive mold operation and maintenance methods.

[0005] To achieve the above objectives, in a first aspect, the present application provides a mold life cycle management method, comprising the following steps:

[0006] Step S101, using the finite element simulation method to optimize the mold design in the mold design stage, and using the finite element simulation results to assist in determining the selection and installation position of the sensor according to the forming process conditions and range, so as to install the sensor on the mold and its forming system, monitor the performance of the mold and the forming system during use online, and obtain the performance data monitored online;

[0007] Step S102: based on the performance data monitored online, determine the online status of the mold and the forming system during use and the online health of the mold, and predict the future health of the mold based on the online health of the mold over multiple cycles; based on the online status of the mold and the forming system, the online health and future health of the mold, and in combination with the offline status of the mold during the period of inactivity detected by the offline detection method, adjust the mold maintenance strategy.

[0008] In an optional example, step S102 specifically includes:

[0009] Based on the performance of the forming equipment in the performance data monitored online, determine whether the state of the forming equipment in the forming system is abnormal;

[0010] If the state of the forming equipment is abnormal, the adjustment and maintenance strategy of the forming equipment is triggered; if the state of the forming equipment is normal, the performance data monitored online is divided into performance data under each predefined working condition, and the characteristic similarity distance between it and the benchmark performance data under each predefined working condition is calculated; based on the forming environment component, mold posture component, mold structure component and forming performance component in the characteristic similarity distance, it is judged whether the forming environment, mold posture, mold structure and forming performance are abnormal, so as to trigger the forming environment adjustment strategy, mold posture adjustment strategy, mold structure maintenance strategy, mold performance compensation strategy and mold performance maintenance strategy;

[0011] Based on the feature similarity distance and the preset performance weight matrix, the comprehensive performance characterization vector is determined; the mold health function is applied to calculate the online health of the mold based on the comprehensive performance characterization vector; and it is determined whether the online health and future health of the mold are abnormal to trigger the mold maintenance strategy and the mold predictive maintenance strategy.

[0012] In an optional example, the mold health function is specifically calculated based on the following formula:

[0013]

[0014] Among them, h 0 is the initial health of the cycle, Y (t, m) is the comprehensive performance characterization vector at time t, m is the number of signals, and E (e j(t), j = 1, ..., n} is the equivalent number of events in the mold during the cycle, n is the event category, Δt is the effective duration of the mold during the cycle, w y That is {w y,i ,i=1,…,m} is the health impact coefficient of the performance characterization vector, w e That is {w e,j ,j=1,…,n} is the health impact coefficient of the event equivalent number, w t is the health impact coefficient of the effective use time of the mold, Δt is the effective use time of the mold within the cycle, θ is {θ y ,θ e ,θ t}Adjust parameters for the health function.

[0015] In an optional example, the method further includes:

[0016] When the difference between the online health of the mold and the offline evaluation result of the health exceeds a certain predefined deviation threshold, the mold health function is calibrated or trained; the offline evaluation result of the health is determined based on the offline state of the mold and the corresponding relationship between the preset offline state and the offline evaluation result;

[0017] The calibrated or retrained health function recalculates the online health of the mold;

[0018] The calibration is specifically performed by solving the following optimization problem to calibrate {θ y ,θ e ,θ t}:

[0019]

[0020] In the formula, d is the number of cycles, h(t i ) is the online health of the ith cycle, a(t i ) is the offline evaluation result of the health status in the i-th cycle;

[0021] The training is specifically carried out by solving the following optimization problem and training {w y ,w e ,w t} and {θ y ,θ e ,θ t}:

[0022]

[0023] In the formula, δw and δθ are the changes of w and θ respectively, λ w is the weight of w, λ θ is the weight of θ.

[0024] In an optional example, if the future health of the mold is abnormal, the future change trend of each component is predicted based on the prediction model corresponding to each component in the comprehensive performance characterization vector; based on the future change trend of each component, the component with the largest change is determined and used as the key component affecting the future health; based on the key component affecting the future health, the corresponding mold predictive maintenance strategy is determined;

[0025] The prediction models corresponding to the components are obtained by training based on the historical data of the components.

[0026] In a second aspect, the present application provides a mold life cycle management system, including:

[0027] The mold design and monitoring module is used to optimize the mold design by using the finite element simulation method in the mold design stage, and use the finite element simulation results to assist in determining the selection and installation location of sensors based on the forming process conditions and range, so as to install sensors on the mold and its forming system to monitor the performance of the mold and forming system during use online and obtain the performance data monitored online;

[0028] The mold performance evaluation and maintenance module is used to judge the online status of the mold and the forming system during use and the online health of the mold based on the performance data monitored online, and predict the future health of the mold based on the online health of the mold for multiple cycles; based on the online status of the mold and the forming system, the online health and future health of the mold, and combined with the offline status of the mold during the idle period detected by the offline detection method, adjust the mold maintenance strategy.

[0029] In a third aspect, the present application provides an electronic device comprising: at least one memory for storing programs; and at least one processor for executing the programs stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method described in the first aspect or any possible implementation of the first aspect.

[0030] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method described in the first aspect or any possible implementation of the first aspect.

[0031] In a fifth aspect, the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method described in the first aspect or any possible implementation of the first aspect.

[0032] It can be understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0033] In general, the above technical solutions conceived by this application have the following beneficial effects compared with the prior art:

[0034] The present application provides a mold life cycle management method, system, electronic device and storage medium. Through real-time measurement of the mold status, a mold operation and maintenance process based on online calculation of mold performance and online analysis of mold performance anomalies is established to achieve active mold performance management, improve the performance stability and utilization efficiency of the mold during the entire life cycle of use, and reduce the increase in production costs and decrease in production efficiency caused by mold performance degradation and failure; different from the usual measurement perception of only a single device or a single state, the state of the entire forming system including the forming environment, the forming main machine and the forming auxiliary machine, and the mold is perceived and measured, so as to accurately and comprehensively reflect the working state of the forming system, and predict the health change trend of the mold, so as to provide a quantitative decision-making basis for more accurate mold performance management and maintenance; and the mold is tested offline during the out-of-use period to achieve more comprehensive inspection and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 One of the flow charts of the mold life cycle management method provided in the embodiment of the present application;

[0036] Figure 2 The second flowchart of the mold life cycle management method provided in the embodiment of the present application;

[0037] Figure 3 A schematic diagram of the structure of a mold life cycle management system provided in an embodiment of the present application;

[0038] Figure 4 A schematic diagram of the process flow of the mold management system provided in the embodiment of the present application;

[0039] Figure 5 A schematic diagram of the process flow of the mold life cycle management auxiliary system provided in an embodiment of the present application;

[0040] Figure 6 A schematic diagram of the process of mold performance, stress and load analysis and experiment and optimization adjustment provided in the embodiments of the present application;

[0041] Figure 7 A schematic diagram of the process of monitoring and detecting the status of molds and related equipment provided in the embodiment of the present application;

[0042] Figure 8A schematic diagram of online mold performance evaluation and health and trend prediction provided in an embodiment of the present application;

[0043] Fig. 9 A schematic diagram of a process flow of an online mold performance evaluation method provided in an embodiment of the present application;

[0044] Fig.10 A schematic diagram of the process flow of optimizing and outputting the mold operation and maintenance solution provided in the embodiment of the present application;

[0045] Fig.11 A schematic diagram of the relationship between state perception, performance evaluation and health calculation and operation and maintenance solution output provided in an embodiment of the present application;

[0046] Fig.12 The architecture diagram of the mold life cycle management system provided in the embodiment of the present application;

[0047] Fig.13 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0050] The present application embodiment provides a mold full life cycle management method, Figure 1 One of the flow charts of the mold life cycle management method provided in the embodiment of the present application is as follows: Figure 1 As shown, the method comprises the following steps:

[0051] Step S101, using the finite element simulation method to optimize the mold design in the mold design stage, and using the finite element simulation results to assist in determining the selection and installation position of the sensor according to the forming process conditions and range, so as to install the sensor on the mold and its forming system, monitor the performance of the mold and the forming system during use online, and obtain the performance data monitored online;

[0052] Step S102: based on the performance data monitored online, determine the online status of the mold and the forming system during use and the online health of the mold, and predict the future health of the mold based on the online health of the mold over multiple cycles; based on the online status of the mold and the forming system, the online health and future health of the mold, and in combination with the offline status of the mold during the period of inactivity detected by the offline detection method, adjust the mold maintenance strategy.

[0053] Specifically, the finite element simulation method is used to optimize the mold design, mold structure and mold material selection during the mold design stage. Further, based on the forming process conditions and range, the stress and strain distribution of the entire mold structure predicted by finite element simulation can help optimize the sensor type and parameter selection, as well as the installation method and location. For example, for the loads in the forming process including injection molding, extrusion and stamping, after obtaining the mold stress distribution and changes during the forming cycle, the stress measurement can be installed at the position where the change is most obvious. According to the physical quantities most related to the mold state change reflected in the simulation, such as temperature, pressure, deformation, displacement, amplitude, vibration frequency, etc. as the parameters to be measured, the required sensor type and the technical specifications of the sensor, such as accuracy, sensitivity, response time, operating temperature range, etc., are determined. At the same time, the location of the sensor is determined based on factors such as the geometric structure of the mold, the response sensitivity of each area, and the complexity of the engineering installation. Furthermore, the fatigue performance and life of the mold can be tested by simulating the cyclic loads encountered by the mold in actual use, and the experimental data will be further used for the optimization of sensor selection and installation location.

[0054] According to the selection and installation position of the sensors determined above, sensors are installed on the mold and its forming system to implement online monitoring of the performance of the mold and the forming system including the forming environment, the forming main machine and the forming auxiliary machine. The online state of the mold and the forming system during use, that is, the working state, is judged based on the performance data monitored online to determine whether it is abnormal, so as to trigger the corresponding maintenance strategy and effectively adjust the mold maintenance strategy. Sensors can include active sensors and passive sensors. Active sensors are installed on the forming main machine, the forming auxiliary machine and the mold. The working state of the forming main machine, the forming auxiliary machine and the mold can be tested by the signal excitation working mode. Passive sensors are installed on the mold to obtain the temperature distribution of the mold during use, the force, deformation, vibration and impact of the key position, etc., so as to accurately reflect the working state of the mold. By comprehensively sorting out the signals of different sensors, the performance data X(t,m,N) monitored online can be obtained, where t is the time, m is the number of signals, and N is the signal length.

[0055] The mold health function is applied to calculate the online health of the mold based on the performance data X(t,m,N), and the online health of the mold over multiple cycles is used to predict the health change trend of the mold; and the offline detection method is combined to detect the offline status of the mold during the idle period, evaluate whether the mold is fatigued or damaged, and trigger the corresponding maintenance strategy for comprehensive inspection and maintenance. Fatigue and damage of the mold are caused by long-term use, repeated cycle loads and other factors, which will shorten the life of the mold and affect production efficiency and product quality. Offline detection methods include but are not limited to appearance inspection. Metallographic detection, ultrasonic detection, magnetic sensors and vibration detection can be used to determine whether the mold has structural problems or other damage.

[0056] The time series of online mold health calculation results of known multiple consecutive cycles, {h(1), h(2), …, h(k)}, can be used to predict the health of the next p cycle points. from Start prediction, and iterate the prediction until the prediction of p future periods is completed. The specific prediction function, i.e., the prediction model, can adopt a periodic autoregressive moving average model (SARMA) based on time series, or a nonlinear autoregressive sliding average model (NARMA), or a machine learning method, such as using a tree structure model and a gradient boosting algorithm to directly construct a prediction function, which is not specifically limited in the present application embodiment.

[0057] Further use of the mold's online health h(t) and prediction of future health Determine whether the health or its prediction is normal, and determine the key reasons by analyzing the health function and trigger the corresponding maintenance method, i.e., maintenance strategy. This method will ensure that there are sufficient resources to maintain the mold when the health is abnormal, and avoid mold failure caused by health deterioration, which will lead to production interruption and product quality problems.

[0058] The method provided in the embodiment of the present application establishes a mold operation and maintenance process based on online calculation of mold performance and online analysis of mold performance anomalies through real-time measurement of mold status, thereby realizing active mold performance management, improving the performance stability and utilization efficiency of the mold during its entire life cycle, and reducing the increase in production costs and the decrease in production efficiency caused by mold performance degradation and failure. Different from the usual measurement and perception of only a single device or a single state, the method adopts the method of perceiving and measuring the state of the entire forming system including the forming environment, the forming main machine and the forming auxiliary machine, and the mold, so as to accurately and comprehensively reflect the working state of the forming system and predict the change trend of the health of the mold, thereby providing a quantitative decision-making basis for more accurate mold performance management and maintenance. The mold is also tested offline during the out-of-use period to realize more comprehensive inspection and maintenance.

[0059] Based on the above embodiment, step S102 specifically includes:

[0060] Based on the performance of the forming equipment in the performance data monitored online, determine whether the state of the forming equipment in the forming system is abnormal;

[0061] If the state of the forming equipment is abnormal, the adjustment and maintenance strategy of the forming equipment is triggered; if the state of the forming equipment is normal, the performance data monitored online is divided into performance data under each predefined working condition, and the characteristic similarity distance between it and the benchmark performance data under each predefined working condition is calculated; based on the forming environment component, mold posture component, mold structure component and forming performance component in the characteristic similarity distance, it is judged whether the forming environment, mold posture, mold structure and forming performance are abnormal, so as to trigger the forming environment adjustment strategy, mold posture adjustment strategy, mold structure maintenance strategy, mold performance compensation strategy and mold performance maintenance strategy;

[0062] Based on the feature similarity distance and the preset performance weight matrix, the comprehensive performance characterization vector is determined; the mold health function is applied to calculate the online health of the mold based on the comprehensive performance characterization vector; and it is determined whether the online health and future health of the mold are abnormal to trigger the mold maintenance strategy and the mold predictive maintenance strategy.

[0063] Specifically, the level value or range of the forming equipment performance in the performance data monitored online can be used to quickly determine whether the state of the forming main machine or the forming auxiliary machine is normal. For example, whether the barrel temperature of the injection molding machine is within a reasonable range, whether the vibration of the punching machine is within the allowable range, whether the negative pressure value of the exhaust gas outlet of the blister machine is normal, etc. Through the above evaluation and judgment, the abnormal state of the mold caused by the abnormal forming equipment can be quickly eliminated.

[0064] The performance data X(t,m,N) monitored online is segmented according to the predefined working conditions. The predefined working conditions divide the forming cycle into multiple stages based on the characteristics of the forming cycle during the use of the mold. For example, during injection molding, multiple stages such as closing, filling, holding pressure, cooling, mold opening, and demolding can be considered. These stages can be identified using the operating status of the injection molding machine. For stamping molding, stages such as feeding positioning, punching, and discharge can be considered. If the number of predefined working conditions is K, the signal length of each working condition is L. 1 ,L 2 ,…,L K , satisfying L 1 +L 2 +…+L k =N, after segmentation, X(t,m,N) will be converted into performance data X(t,m,N,K) under each predefined working condition.

[0065] The comparison between the benchmark performance data Z(m,N,K) and X(t,m,N,K) is achieved by calculating the feature similarity distance between the two, and the calculation process is defined as D(X,Z). After operating D(X,Z) between Z(m,N,K) and X(t,m,N,K), an m-row and K-column feature similarity distance matrix d(t,m,K) is obtained. According to the forming environment component, mold posture component, mold structure component and forming performance component in d(t,m,K), it can be judged whether the forming environment, mold posture, mold structure and forming performance are abnormal. The forming environment includes but is not limited to ambient temperature and humidity, incoming material status and foreign body interference.

[0066] Further, the comprehensive performance characterization vector Y(t,m) of the mold at a certain moment is calculated by weighting d(t,m,K) with the predefined performance weight matrix Q(m,K). The mold health function is applied to calculate the online health of the mold based on the comprehensive performance characterization vector; the online health of the mold over multiple cycles is then used to predict the future health of the mold; it is determined whether the online health and future health of the mold are abnormal. If abnormal, the key components that affect the health in Y(t,m) are identified through the analysis of the health function to trigger the corresponding mold maintenance strategy and mold predictive maintenance strategy.

[0067] It should be noted that by giving priority to judging whether the forming equipment and the forming environment in the forming system are abnormal, if there is an abnormality, the abnormal state of the mold caused by the abnormal forming equipment and the forming environment can be quickly eliminated, and the adjustment and maintenance strategy of the forming equipment and the forming environment adjustment strategy can be executed preferentially.

[0068] Based on any of the above embodiments, the mold health function is specifically calculated based on the following formula:

[0069]

[0070] Among them, h 0 is the initial health of the cycle, Y (t, m) is the comprehensive performance characterization vector at time t, m is the number of signals, and E (e j (t), j = 1, ..., n} is the equivalent number of events in the mold during the cycle, n is the event category, Δt is the effective duration of the mold during the cycle, w y That is {w y,i ,i=1,…,m} is the health impact coefficient of the performance characterization vector, w e That is {w e,j ,j=1,…,n} is the health impact coefficient of the event equivalent number, w t is the health impact coefficient of the effective use time of the mold, Δt is the effective use time of the mold within the cycle, θ is {θ y ,θ e,θ t}Adjust parameters for the health function.

[0071] It should be noted that the present application comprehensively considers the performance characterization vector Y(t,m) based on perception and detection at a certain time t, and the equivalent number of each event of the mold E: {e j (t), j = 1, ..., n} and the effective use time of the mold Δt, construct the mold health function, and realize the online and accurate calculation of the health of multiple dimensions, which is defined as h(t) = H(t,, h 0 ,Y,W,Δt,w y ,w e ,w u ,θ). At the same time, an adjustable parameter θ is introduced: {θ y ,θ e ,θ t}, the health function is calibrated using online health results and offline health measurement evaluation over multiple cycles.

[0072] Based on any of the above embodiments, if the future health of the mold is abnormal, the future change trend of each component is predicted based on the prediction model corresponding to each component in the comprehensive performance characterization vector; based on the future change trend of each component, the component with the largest change is determined as the key component affecting the future health; based on the key component affecting the future health, the corresponding mold predictive maintenance strategy is determined;

[0073] The prediction models corresponding to the components are obtained by training based on the historical data of the components.

[0074] In particular, the component with the largest change can be selected as the key component affecting future health, or multiple components ranked in descending order of change can be selected as key components, or the component whose change trend exceeds the allowable range can be selected as the key component. The embodiments of the present application do not make specific limitations on this.

[0075] The prediction model used in health prediction can be applied to each component, in which case the prediction model structures corresponding to each component are the same. Alternatively, a separate prediction model can be constructed for each component, in which case the prediction model structures corresponding to each component are different. This embodiment of the present application does not make any specific limitations on this.

[0076] Specifically, when the future health of the mold, i.e., the current h(t), is lower than the preset threshold, since the health h(t) is a function based on {Y, E, Δt}, the key components affecting the health will be identified through the analysis of the health function, and the corresponding mold maintenance plan, i.e., mold maintenance strategy, will be triggered immediately to ensure timely exception handling and health assurance. For example, the h(t) anomaly is related to the posture and position components in the performance vector Y(t, m), and the maintenance method triggered by this component is to adjust the mold installation method to restore the normal posture and position.

[0077] When the predicted health and its interval are lower than the preset threshold, the mold predictive maintenance plan will be triggered. The key components that affect the predicted health value will be identified by predicting each component in {Y, E, Δt}, and the corresponding predictive maintenance plan, i.e., mold predictive maintenance strategy, will be triggered. The time window, process, materials, tools, and professionals required for maintenance will be prepared according to the plan.

[0078] Based on any of the above embodiments, the method further includes:

[0079] When the difference between the online health of the mold and the offline evaluation result of the health exceeds a certain predefined deviation threshold, the mold health function is calibrated or trained; the offline evaluation result of the health is determined based on the offline state of the mold and the corresponding relationship between the preset offline state and the offline evaluation result;

[0080] The calibrated or retrained health function recalculates the online health of the mold;

[0081] The calibration is specifically performed by solving the following optimization problem to calibrate {θ y ,θ e ,θ t}:

[0082]

[0083] In the formula, d is the number of cycles, h(t i ) is the online health of the ith cycle, a(t i ) is the offline evaluation result of the health status in the i-th cycle;

[0084] The training is specifically carried out by solving the following optimization problem and training {w y ,w e ,w t} and {θ y ,θ e ,θ t}:

[0085]

[0086] In the formula, δw and δθ are the changes of w and θ respectively, λ w is the weight of w, λ θ is the weight of θ.

[0087] Specifically, the offline status of the mold may include the mold appearance and structural integrity. Based on the mold offline status detection result, the corresponding relationship between the offline status and the offline health evaluation result pre-set by the mold operation and maintenance staff is used to determine the offline health evaluation result {a(t 1 ),a(t 2 ),…,a(t d )}.

[0088] Establish a method to calibrate and train the health function using online health results and offline mold health assessments over multiple cycles to ensure the accuracy of the health function. The h(t) calculated by the calibrated or retrained health function will be used in the health prediction model to predict future health.

[0089] It is understandable that calibration only requires optimizing {θ y ,θ e ,θ t}, the amount of data required is relatively small, and the frequency of calibration can be relatively high in practical applications. Retraining also requires optimization of weight parameters, which requires a relatively large amount of data. In practical applications, the frequency of training can be relatively low, and conditional triggering is used after a certain amount of data has been accumulated.

[0090] Based on any of the above embodiments, in view of the deficiencies of the prior art, the present application provides a mold life cycle management method, the purpose of which is to timely identify mold performance abnormalities and categories by implementing mold status judgment, forming environment and forming performance judgment, mold performance evaluation, and mold health calculation and prediction during the use of the mold, thereby realizing active mold performance management, improving mold performance stability and use efficiency, and reducing production cost increases and production efficiency reductions caused by mold performance attenuation and failures. The present application is described in detail below.

[0091] Figure 2 The second flow chart of the mold life cycle management method provided in the embodiment of the present application is as follows: Figure 2 As shown, the mold life cycle management method provided by the embodiment of the present application may include:

[0092] S110 obtains mold information:

[0093] Specifically, mold information can be obtained from the mold management system. The mold information may include the mold number, the basic structure and size of the mold, the main materials used for the mold, substrate and parts, the specific size and precision requirements of the mold, the weight of the mold, the design service life of the mold including the expected usage time and number of uses, the repair and maintenance of the mold, the usage of the mold, the storage of the mold, the warranty period of the mold, the mold drawings, the mold inspection report, the manufacturer of the mold and contact information.

[0094] S120 mold simulation (mold performance simulation and expected benchmark performance establishment):

[0095] Specifically, based on the forming process conditions and scope, the results of modeling and simulation are used to help optimize the sensor type and parameter selection, as well as the installation method and location. For example, for the loads in the forming process including injection molding, extrusion and stamping, after obtaining the mold stress distribution and changes during the forming cycle, the stress measurement can be installed at the position where the change is most obvious. Based on the physical quantities most relevant to the mold state change reflected in the simulation, such as temperature, pressure, deformation, displacement, amplitude, vibration frequency, etc., as the parameters to be measured, the required type of sensor and the technical specifications of the sensor, such as accuracy, sensitivity, response time, operating temperature range, etc., are determined. At the same time, the location of the sensor is determined based on factors such as the geometric structure of the mold, the response sensitivity of each area, and the complexity of the engineering installation.

[0096] The simulation results will be further used as expected performance data according to the forming process conditions and scope, so as to participate in online performance evaluation and health calculation as a benchmark.

[0097] S130 mold and related equipment status monitoring and detection, and offline detection:

[0098] Specifically, the state of the overall forming system including the forming environment, the forming main machine and the forming auxiliary machine, and the mold is sensed and measured, so as to accurately and comprehensively reflect the working state of the forming system. Specifically, it is necessary to measure or communicate the forming environment of the mold, including but not limited to the environmental conditions such as ambient temperature and humidity, the state of the incoming material, and the interference of foreign matter, so as to accurately reflect the state of the forming environment. It is necessary to measure or communicate the existing state measurement signals of the state of the related systems of the mold, such as the forming main machine and the forming auxiliary machine, so as to accurately reflect the working state of the forming main machine and the auxiliary machine. Active sensors can also be installed on the forming main machine and the auxiliary machine equipment, and the working state of the forming main machine and the auxiliary machine can be tested by using the signal excitation working mode. Passive sensing and measuring devices need to be installed on the mold to obtain the temperature distribution of the mold during use, the force, deformation, vibration and impact of the key positions, etc., so as to accurately reflect the working state of the mold. Active sensors can be installed on the mold. Unlike passive sensors, active sensors use the signal excitation working mode to test the working state of the forming mold. Offline detection methods are required to evaluate whether the mold is fatigued or damaged, and to evaluate the health score.

[0099] S140 mold performance assessment and health assessment, model calibration and training:

[0100] Specifically, the results of online monitoring and offline detection are used to judge the state of the mold, whether the forming environment, structure and position of the mold, and the forming performance are abnormal, and to predict the performance change trend of the mold. Among them, the key innovation is to use the online state perception and detection data during the forming process, and compare the similar feature distance with the benchmark performance data under each predefined working condition to establish d(t,m,K) and Y(t,m). And further judge and classify whether the performance is abnormal based on the comparison results.

[0101] Considering the performance representation vector Y(t,m) based on perception and detection at a certain time t, the equivalent number of events of the mold E: {e j (t), j = 1, …, n} and the effective use time of the mold Δt, construct the mold health function and implement the following online calculation of health including multiple dimensions.

[0102]

[0103] Using the time series of mold health calculation results of multiple consecutive cycles, {h(1), h(2), …, h(k)}, the health of the next p cycle points is predicted. The prediction method will use the following autoregressive single-step prediction model structure, iterating the prediction until the prediction of p future periods is completed. Further use the current health h(t) and the prediction of future health Determine whether the health or its prediction is normal, and determine the key reasons by analyzing the health function and trigger the corresponding maintenance method. This method will ensure that there are sufficient resources to maintain the mold when the health is abnormal, and avoid mold failure caused by health deterioration, which will lead to production interruption and product quality problems.

[0104] The health function is calibrated and trained by using online health results of multiple cycles and offline mold health evaluation to ensure the accuracy of the health function. The calibration process only adjusts the adjustable parameters {θ y ,θ e ,θ t ], because less valid sample data is required, it can occur at a higher frequency. Training is to optimize {w y ,w e ,w t} or optimize {w y ,w e ,w t} and {θ y ,θ e ,θ t}, because the number of parameters to be optimized is large, it can only be operated after accumulating enough valid data samples. Calibration and training can be triggered by conditional triggering, which is triggered only when the difference between the online health and offline evaluation results exceeds a certain predefined threshold.

[0105] S150 mold operation and maintenance solution optimization and output (mold operation and maintenance optimization module based on performance and health model):

[0106] Through real-time measurement of mold status, a mold operation and maintenance process based on online calculation of mold performance and online analysis of mold performance anomalies is established to improve the performance stability and efficiency of the mold throughout its life cycle, and reduce the increase in production costs and decrease in production efficiency caused by mold performance degradation and failures.

[0107] Figure 3 This is a schematic diagram of the structure of the mold life cycle management system provided in the embodiment of the present application. Figure 3 As shown, the mold life cycle management system S200 provided in the embodiment of the present application may include:

[0108] S210 mold management system:

[0109] Realize the entry, storage, query and management of basic information of molds, tracking of mold status, and process management of daily mold maintenance.

[0110] S220 mold life cycle management auxiliary system:

[0111] It realizes the main functions of the whole life cycle management of molds from design, status monitoring, operation and maintenance optimization to scrapping, including the main functional modules of mold design, simulation, experimental analysis and optimization adjustment of this application, status monitoring and detection of molds and related equipment, mold performance and health analysis, and mold operation and maintenance solution optimization. It realizes data interaction with the mold management system S210.

[0112] Figure 4 A schematic diagram of the process flow of the mold management system provided in the embodiment of the present application is as follows: Figure 4 As shown, module S210 may include:

[0113] S211 mold file management, establish a complete mold information management system to realize the collection, recording, classification, archiving and query of various mold information.

[0114] S212 mold usage plan, according to the company's production plan and order situation, to formulate a mold usage plan to ensure the rational use and maintenance of the mold, while ensuring the smooth implementation of the production plan and the stability of product quality.

[0115] S213 mold maintenance can formulate and track mold maintenance plans and execution, record mold maintenance conditions, and provide maintenance reminders and alarms in advance based on mold usage and maintenance plans. It has certain data analysis and statistical functions to provide decision support and management guidance, as well as historical maintenance record tracing and query.

[0116] S214 mold repair can achieve fast and accurate repair and reduce repair costs through a complete set of mold repair management processes. It mainly includes the formulation of repair plans, optimization of repair methods, steps, tools and materials, standardization and normalization of repair steps, methods, precautions and safety measures, and the entry, query and management of repair and repair records.

[0117] Figure 5 A schematic diagram of the process flow of the mold life cycle management auxiliary system provided in the embodiment of the present application is as follows: Figure 5 As shown, module S220 may include:

[0118] S221 Mold design, simulation, experimental analysis and optimization adjustment. Specifically, the finite element simulation method is used to optimize the mold design, mold structure and mold material selection in the mold design stage. Further, based on the forming process conditions and range, the results of modeling and simulation are used to help optimize the sensor type and parameter selection, as well as the installation method and location. For example, for the loads in the forming process including injection molding, extrusion and stamping, after obtaining the mold stress distribution and changes during the forming cycle, the stress measurement can be installed at the position where the change is most obvious. According to the physical quantities most related to the mold state change reflected in the simulation, such as temperature, pressure, deformation, displacement, amplitude, vibration frequency, etc., as the parameters to be measured, the required type of sensor and the technical specifications of the sensor, such as accuracy, sensitivity, response time, operating temperature range, etc., are determined. At the same time, the location of the sensor is determined based on factors such as the geometric structure of the mold, the response sensitivity of each area, and the complexity of the engineering installation. The results of the simulation are further used as the expected performance data, so as to participate in online performance evaluation and health calculation as a benchmark.

[0119] S222 mold and forming state perception, passive sensing measurement devices can be installed on the mold to obtain the temperature distribution, thermal stress, force at key positions, deformation, impact, etc. of the mold during use. To determine whether the mold is in normal working condition, or whether it needs maintenance or replacement.

[0120] Active sensors can be installed on the mold or its auxiliary system, and the working state of the forming mold can be tested by using a signal excitation working mode.

[0121] Offline testing methods can be used to assess whether the mold has fatigued or damaged.

[0122] S223 Mold performance and health trend (mold performance and health evaluation, prediction and calibration model), use the results of online monitoring and offline detection to judge the state of the mold, whether the forming environment, structure and position of the mold and the forming performance are abnormal, and predict the change trend of the mold health, so as to provide a decision-making basis for more accurate performance management and maintenance of the mold. Use the results of online monitoring and offline detection to judge the state of the mold, whether the forming environment, structure and position of the mold and the forming performance are abnormal, and predict the change trend of the mold performance. Among them, the online state perception and detection data during the forming process are used, and the similar feature distance is compared with the benchmark performance data under each predefined working condition. And further judge and classify whether the performance is abnormal based on the comparison results. Using the time series of the mold health calculation results of multiple consecutive cycles, {h(1), h(2),…, h(k)}, the health of the future p cycle points is predicted, A health function is calibrated and trained using online health results of multiple cycles and offline mold health assessment to ensure the accuracy of the health function.

[0123] S224 mold operation and maintenance solution optimization, based on the actual monitoring and detection status and change trend of the mold to optimize the mold maintenance strategy. Use the current health h(t) and the prediction of future health Determine whether the health or its prediction is normal, and determine the key reasons and trigger the corresponding maintenance methods by analyzing the health function. This method will ensure that there are sufficient resources to maintain the mold when the health is abnormal, and avoid mold failure caused by health deterioration, which will lead to production interruption and product quality problems. Specifically, when the current h(t) is lower than the preset threshold, since the health h(t) is a function based on {Y, E, Δt}, the key components affecting the health will be identified by analyzing the health function, and the corresponding mold maintenance plan will be triggered immediately, thereby ensuring timely exception handling and health assurance. When the predicted health and its interval are lower than the preset threshold, the mold predictive maintenance plan will be triggered. The key components affecting the predicted health value will be identified by predicting each component in {Y, E, Δt}, and the corresponding predictive maintenance plan will be triggered. The time window, process, materials, tools, and professionals required for maintenance will be prepared according to the plan. Among them, since the health h(t) is a function based on {Y, E, Δt}, its predicted value The health threshold below the preset threshold at k moments in the future (k = 1, ..., p) is related to the predicted value of a key component. If the components are assumed to be independent of each other, the key components can be identified by first applying the prediction model of h(t) to each component to calculate the predicted value of each component, such as the performance vector Y(t, m):

[0124]

[0125] …

[0126]

[0127] In addition, a prediction model may be established for each component. For example, for the performance vector Y(t,m), the following m prediction models may be established for m performance components.

[0128]

[0129] …

[0130]

[0131] By finding the components with the largest changes, key influencing factors can be identified to trigger maintenance plans.

[0132] Figure 6 A schematic diagram of the process of mold performance, stress and load analysis and experiment and optimization adjustment provided in the embodiment of the present application, such as Figure 6 As shown, module S221 may include:

[0133] S2211 mold stress analysis, specifically, considers the mold geometry, structural features, material properties, processing technology and process conditions, and predicts the stress and strain distribution of the entire mold structure through finite element analysis. This information will be used for sensor selection and optimization of installation location.

[0134] S2212 Temperature field analysis, based on the mold processing technology and process conditions and range, including the load and temperature changes and ranges during injection molding, extrusion, etc., obtains the temperature and thermal stress distribution inside the mold. This information, combined with simulation results such as pressure, will be used to establish a performance reference benchmark for the mold.

[0135] S2213 Cyclic Load Test, this optional step tests the fatigue performance and life of the mold by simulating the cyclic loads encountered by the mold in actual use. The experiment usually involves applying cyclic loads to the mold, then observing the mold's response, such as deformation, cracks, etc., and recording the fatigue life of the mold after a certain number of cyclic loads. The experimental data will be further used to optimize the sensor selection and installation location, as well as to establish a performance reference benchmark.

[0136] S2214 design optimization and adjustment, the service life and performance of the mold mainly depend on its structure and material. Therefore, the mold structure can be optimized to increase the strength and rigidity of the mold and reduce the generation and expansion of fatigue cracks. High-strength, high-toughness materials or high thermal conductivity materials can be selected to increase the service life of the mold. The wear resistance and corrosion resistance of the mold can be improved through surface treatment and coating. The mold cooling system can be designed to help the mold dissipate heat better, reduce the mold temperature, and reduce the generation of thermal stress and fatigue cracks, thereby extending the mold service life and improving the performance of the mold.

[0137] Figure 7 A schematic diagram of the process of monitoring and detecting the status of molds and related equipment provided in the embodiment of the present application, such as Figure 7 As shown, module S222 may include:

[0138] S2221 mold related host status, measure the state of the forming host or communicate the existing state measurement signal to accurately reflect the working state of the forming host. Active sensors can also be installed on the forming host to test the working state of the forming host using signal excitation. It further includes measuring or communicating the forming environment states such as, but not limited to, ambient temperature and humidity, incoming material state, and foreign body interference.

[0139] S2222 mold related auxiliary machine status, measure the state of the forming auxiliary machine or communicate the existing state measurement signal, accurately know the working state of the forming auxiliary machine. You can also install active sensors on the forming auxiliary machine equipment and use signal excitation to test the working state of the forming auxiliary machine.

[0140] S2223 online sensing of mold status, installing a passive sensor measurement device on the mold to obtain the temperature distribution of the mold during use, the force, deformation, vibration and impact of key positions, etc., to accurately reflect the working status of the mold.

[0141] S2224 online mold status detection, install active sensors on the mold, and use signal excitation to test the working status of the forming mold.

[0142] S2225 offline mold status detection uses offline detection methods to evaluate whether the mold is fatigued or damaged. Offline detection methods include but are not limited to appearance inspection. Metallographic testing, ultrasonic testing, magnetic sensors, and vibration testing can be used to determine whether the mold has structural problems or other damage.

[0143] Figure 8 A schematic diagram of online mold performance evaluation and health and trend prediction provided in an embodiment of the present application, such as Figure 8 As shown, module S223 may include:

[0144] S2231 Forming equipment status evaluation, obtain the value of the forming main machine or auxiliary machine status.

[0145] S2232 determines whether the molding equipment is in a normal state. It uses the level or range of the molding equipment to quickly determine whether the molding machine or auxiliary machine is in a normal state. For example, whether the barrel temperature of the injection molding machine is within a reasonable range, whether the vibration of the punching machine is within the allowable range, whether the negative pressure value of the exhaust gas outlet of the blister machine is normal, etc. Through the above evaluation and judgment, the abnormal state of the mold caused by the abnormal molding equipment can be quickly eliminated.

[0146] S2233 online mold performance evaluation uses online state perception and detection data during the forming process, and compares similar feature distances with benchmark performance data under each predefined working condition, and further determines and classifies whether the performance is abnormal based on the comparison results. Among them, the predefined working condition is to divide the forming cycle into multiple working condition stages based on the characteristics of the forming cycle. Because different stages in the forming process have different effects on the mold, the introduction of predefined working conditions can more accurately identify the type of abnormality and quantify the degree of performance change. The benchmark performance data can use the expected performance data based on simulation, or the reference performance data established based on the state perception and detection history data in the forming process. The benchmark can be recalibrated and adjusted with new simulation data or operation data.

[0147] The online calculation method of performance based on similar feature distance is defined in Figure 8 S2233, Fig. 9 A schematic diagram of the process flow of the online mold performance evaluation method provided in the embodiment of the present application is as follows: Fig. 9 As shown, at a certain time t, the module receives the online passive sensing and active measurement signal X(t,m,N 1 ,…,N m ), it is defined as:

[0148] x 1 (t,1),x 1 (t,2),…,x 1 (t,N 1 )

[0149] x 2 (t,1),x 2 (t,2),…,x 2 (t,N 2 ) ...

[0151] x m (t,1),x m (t,2),…,x m (t,N m )

[0152] Where m is the number of sensed and measured signals, N i (i=1,..,m) is each signal x i Length.

[0153] The X(t,m,N 1 ,…,N m ) is used as the data source for mold performance evaluation. First, it is converted into X(t,m,N) through filtering, oversampling and signal synchronization, which is defined as:

[0154] x 1 (t,1),x 1 (t,2),…,x 1 (t,N)

[0155] x 2 (t,1),x 2 (t,2),…,x 2 (t,N)

[0156] …

[0157] x m (t,1),x m (t,2),…,x m (t,N)

[0158] Where N = max(N 1 ,N 2 ,…,N m ). The module then divides X(t,m,N) according to predefined working conditions. The predefined working conditions divide the forming cycle into multiple stages based on the characteristics of the forming cycle during the use of the mold. For example, during injection molding, multiple stages such as closing, filling, holding pressure, cooling, mold opening, and demolding can be considered. These stages can be identified using the operating status of the injection molding machine. For stamping molding, stages such as feeding positioning, blanking, and discharge can be considered. If the number of predefined working conditions is K, the signal length of each working condition is L 1 ,L 2 ,…,L K , satisfying L 1 +L 2 +…+L k =N, after segmentation, X(t,m,N) will be converted to X(t,m,N,K), which is defined as:

[0159] X 1,1 ,X 1,2 ,…,X 1,K

[0160] X 2,1 ,X 2,2 ,…,X 2,K

[0161] …

[0162] X m,1 ,X m,2 ,…,X m,K

[0163] Where X i,j (i=1,…,m,j=1,…,K). For example, where X 1,1 , defined as:

[0164] x 1 (t,1),x 1 (t,2),…,x 1 (t,L 1 )

[0165] X m,1 for:

[0166] x m (t,1),x m (t,2),…,x m (t,L 1 )

[0167] X 1,2 for:

[0168] x 1 (t,L 1 +1),x 1 (t,L 1 +2),…,x 1 (t,L 1 +L 2 -1)

[0169] X m,2 for:

[0170] x m (t,L 1 +1),x m (t,L 1 +2),…,x m (t,L 1 +L 2 -1)

[0171] The performance evaluation at a certain moment will compare the online measurement X(t,m,N,K) based on the working condition segmentation with the preset benchmark data Z(m,N,K). Z(m,N,K) has the same structure as X(t,m,N,K) and is defined as:

[0172] Z 1,1 ,Z 1,2 ,…,Z 1,K

[0173] Z 2,1 ,Z 2,2 ,…,Z 2,K

[0174] …

[0175] Z m,1 ,Z m,2 ,…,Z m,K

[0176] Z(m,N,K) can be derived from the simulation results of S221, or can be obtained or updated through experiments. The comparison between Z(m,N,K) and X(t,m,N,K) is achieved by calculating the characteristic similarity distance between the two. The calculation process is defined as D(X,Z), which is specifically achieved by calculating the characteristic similarity distance between the online measurement data of each signal under each working condition and the benchmark data. 1,1 :

[0177] x 1 (t,1),x 1 (t,2),…,x 1 (t,L 1 )

[0178] and Z in Z(m,N,K)1,1 :z 1 (1),z 1 (2),…,z 1 (L 1 ) as an example, the feature similarity distance between the two is d(t,1,1). The feature similarity distance can be calculated using the following methods, including but not limited to the following.

[0179] 1) Distance between vectors: This includes, for example, Euclidean distance: Manhattan distance: And the generalized Minkowski distance

[0180] 2) Vector similarity: This includes, for example, the cosine similarity calculation used to measure the cosine of the angle between two vectors, and the Pearson correlation coefficient used to measure the linear correlation between two vectors.

[0181] 3) Feature quantity difference: For example, X 1,1 With Z 1,1 The maximum difference between max(X 1,1 )-max(Z 1,1 ), mean difference, median difference, standard deviation difference, rate of change difference, etc.

[0182] In this solution, a predefined feature similarity distance function is used for each signal and each segment of the data. For example, the injection segment considers the rate of change or slope of the signal, and the pressure extreme value is considered in the pressure holding segment. Finally, after operating D(X,Z) between Z(m,N,K) and X(t,m,N,K), a feature similarity distance matrix d(t,m,K) with m rows and K columns is obtained.

[0183] Fig. 9 In the specific implementation of S2233, the comprehensive performance characterization vector Y(t,m) of the mold is further calculated by using the predefined performance weight matrix Q(m,K) to perform performance weighting on d(t,m,K), which is defined as:

[0184] Y(t,m)=[y 1 (t) … y m (t)]

[0185] y 1 (t)=q(1,1)×d(t,1,1)+…+q(1,k)×d(t,1,k)

[0186] y 2 (t)=q(2,1)×d(t,2,1)+…+q(2,k)×d(t,2,k)

[0187] …

[0188] y m (t)=q(m,1)×d(t,m,1)+…+q(m,k)×d(t,m,k)

[0189] The weight setting of Q(m,K) reflects the correlation and importance of each signal and each segment and the overall performance. For example, the vibration or deformation signal of the stamping die has a low correlation with the forming performance in the feeding section, but reflects the forming performance in the blanking section. This can be achieved by lowering the weight of the feeding section and raising the weight of the blanking section. The comprehensive performance characterization vector Y(t,m) will be used as the output of S2233 and enter S2237.

[0190] S2234 determines whether the forming environment is normal, using the relevant forming machine, forming auxiliary machine and incoming material status components in d(t,m,K) from S2233, such as the servo response performance of the injection molding machine, the balance of the column, the hydraulic oil level, the cooling return water temperature, the incoming material temperature and humidity, etc. Only when all judgments are normal will it enter the next link. The abnormal judgment output of S2234 will enter S224 and enter S2243. The normal judgment output of S2234 will enter S2235.

[0191] S2235 determines the normal state of mold posture / structure, using the components of the relevant mold posture and structure in d(t,m,K) from S2233 and passed through S2234. For example, the position of the mold relative to the forming host in d(t,m,k), the angle of the ejector plate of the injection mold, the horizontality and verticality of the stamping mold, etc. are used to determine the position and posture. The stress, image, ejection force, gate pressure, etc. in d(t,m,k) are used to determine the structure of parts such as inserts and valve needles, mold deformation, and mold surface finish. The judgment result will only enter the next link when all are normal. The output of S2235 abnormal judgment will enter S224 and enter S2244 and S2245 for processing. The output of S2235 normal judgment will enter S2236.

[0192] S2236 determines the normal forming performance, using the components of the relevant forming performance in d(t, m, K) from S2233 and passing through S2234 and S2235. For example, for injection molding, the injection molding performance is determined by the injection cavity pressure, injection cavity temperature, structural characterization of injection molded parts, and appearance characterization of injection molded parts in d(t, m, k). For stamping forming, the acceleration of the stamping die forming process, the deformation of the stamping die forming process, the structural characterization of the stamping parts, and the appearance characterization of the stamping parts in d(t, m, k) are used to determine the stamping forming performance. The judgment results will only enter the next link when all are normal. The abnormal judgment output of S2236 will enter S224 and enter S2246 and S2247 for processing.

[0193] S2237 mold health and trend prediction, the mold health function at a certain time t is established based on the performance characterization vector Y(t,m) of perception and detection, the equivalent number of mold events E: {e j (t), j = 1, ..., n} and the effective use time of the mold Δt, so as to realize the online calculation of health in multiple dimensions, which is defined as h(t) = H(t, h 0 ,Y,E,Δt,w y ,w e ,w u ,θ). We further introduce an adjustable parameter θ:{θ y ,θ e ,θ t}, the health function is calibrated using online health results and offline health measurement evaluation over multiple cycles.

[0194] The specific implementation of this health is as follows:

[0195]

[0196] The variables are defined as follows:

[0197] h 0 : Initial health of the cycle

[0198] Y:Y(t,m), performance representation vector at time t

[0199] E:{e j (t), j = 1, ..., n}, the equivalent number of events in the mold during the cycle, n is the event category, and each event includes but is not limited to maintenance, parts replacement, mold repair, etc.

[0200] w y :{w y,i ,i=1,…,m}, the health impact coefficient of the performance characterization vector

[0201] w e :{w e,j ,j=1,…,n}, health impact coefficient of event equivalent number

[0202] w t :Health influence coefficient of effective mold use time

[0203] Δt: The effective duration of the mold in a cycle. The effective duration here can be the accumulation of the actual use time of the mold, or it can be replaced by the number of times the mold is used.

[0204] θ:{θ y ,θ e ,θ t}, health function adjustment parameters

[0205] The health prediction is based on the time series of mold health calculation results of multiple known consecutive cycles, {h(1), h(2), …, h(k)}, so as to predict the health of p future cycle points.

[0206] The forecasting method will use the following autoregressive one-step forecasting model structure, iterating the forecast until the forecast of p future periods is completed:

[0207]

[0208]

[0209] …

[0210]

[0211] The prediction function f can use the periodic autoregressive moving average model (SARMA) based on time series. When nonlinear transformation is introduced, the NARMA model can be used, f(g(h(k),δ),g(h(k-1),δ),…,g(h(kq),δ)), where the function g is the introduced nonlinear transformation and δ is the transformation parameter. Machine learning methods can also be used, such as using tree structure models and gradient boosting algorithms to directly construct function f.

[0212] S2238 mold health assessment, based on the results of mold offline status detection, mold operation and maintenance personnel evaluate and record the mold health, {a(t 1 ),a(t 2 ),…,a(t d )}. The evaluation result will enter S2239 to calibrate or train the online health model.

[0213] S2239 mold health trend calibration uses multiple cycles of online health results and offline mold health assessment to calibrate and train the health function to ensure the accuracy of the health function. The calibration process only adjusts the adjustable parameters {θ y ,θ e ,θ t}, because less valid sample data is required, it can occur at a higher frequency. Training is to optimize {w y ,w e ,w t} or optimize {w y ,w e ,w t} and {θ y ,θ e ,θ t}, because the number of parameters to be optimized is large, it can only be operated after sufficient valid data samples have been accumulated. It should be noted that the triggering of calibration and training can be conditionally triggered, and is triggered only when the difference between the online health of S2237 and the offline evaluation result of S2238 exceeds a certain predefined threshold.

[0214] The online health of multiple cycles is defined as {h(t 1 ),h(t 2 ),…,h(t d )}, where d is the number of cycles. Offline mold health assessment is Figure 8 S2238, this assessment will be used by relevant mold operation and maintenance personnel Figure 7 The S2225 mold status is evaluated by offline detection, and the evaluation result is {a(t 1 ),a(t 2 ),…,a(t d )}.

[0215] For {θ y ,θ e ,θ t The calibration process will be implemented by solving the following optimization problem:

[0216]

[0217] Specifically, since the health h(t) is related to the adjustable parameter {θ y ,θ e ,θ t} is a linear relationship. The above optimization problem can be solved by the least square method when d is greater than 3. In practice, it is recommended to select d much greater than 3 to avoid overfitting of adjustable parameters.

[0218] When d is large enough, we can solve the following optimization problem to train {w y ,w e ,w t} and {θ y ,θ e ,θ t ],

[0219]

[0220] In order to avoid large fluctuations in parameters, the method design can add a penalty term to the parameter changes to solve the following problems:

[0221]

[0222] Where δw and δθ are the changes of w and θ.

[0223] The h(t) calculated by the calibrated or retrained health function will be used in the health prediction model to predict future health.

[0224] S22310 Determine whether the mold is healthy or trending normally, using the current health h(t) and the prediction of future health Check whether the health or its prediction is lower than the preset threshold. If it is higher than the preset threshold, it is considered normal, otherwise it is considered abnormal. The judgment result needs to be further analyzed by the health function to determine the key cause and trigger the corresponding maintenance method.

[0225] Fig.10 A schematic diagram of the process flow of optimizing and outputting the mold operation and maintenance solution provided in the embodiment of the present application, such as Fig.10 As shown, module S224 may include:

[0226] S2241 Forming equipment adjustment, by adjusting the process parameters of the forming main machine or forming auxiliary machine, the abnormal status problem can be compensated.

[0227] S2242 Forming equipment maintenance, handles abnormal status problems by debugging, maintaining, repairing or replacing the forming main machine or forming auxiliary machine.

[0228] S2243 Forming environment adjustment, first solves abnormal problems including but not limited to environmental temperature and humidity, incoming material status and foreign body interference. Further solves the abnormal status problems of the forming main machine and forming auxiliary machine that are not solved in S2241 and S2242. Such problems are usually due to the lack of direct perception and measurement of the relevant status of the forming main machine and forming auxiliary machine for such abnormal phenomena, and cannot be effectively identified. Usually, they need to be inferred through analysis of the forming process or forming result data.

[0229] S2244 posture adjustment, restore the mold posture by adjusting the mold installation method, position, angle, horizontality and verticality.

[0230] S2245 structural maintenance, through re-polishing, mold repair and replacement, mold, replacement of spare parts and other methods to restore the structural integrity of the mold to meet the forming requirements.

[0231] S2246 performance compensation restores the forming performance by adjusting the forming process, the operating mode of the forming main machine, the operating mode of the forming auxiliary machine, etc.

[0232] S2247 performance maintenance, restore forming performance by maintaining, repairing or replacing the mold, forming main machine and forming auxiliary machine.

[0233] S2248 mold maintenance, this application refers to a method of triggering mold maintenance using the current health h(t) and analysis based on the health model. Specifically, when the current h(t) is lower than the preset threshold, since the health h(t) is a function based on {Y, E, Δt}, the key components affecting the health will be identified through the analysis of the health function, and the corresponding mold maintenance plan will be triggered immediately to ensure timely exception handling and health assurance. For example, the h(t) anomaly is related to the posture and position components in the performance vector Y(t, m), and the maintenance method triggered by this component is to adjust the mold installation method to restore the normal posture and position.

[0234] S2249 predictive maintenance, in this application, refers to a mold predictive maintenance plan that will be triggered when the predicted health and its interval are lower than a preset threshold. The method will identify the key components that affect the predicted health value by predicting each component in {Y, E, Δt}, and trigger the corresponding predictive maintenance plan.

[0235] Fig.11 A schematic diagram of the relationship between state perception, performance evaluation and health calculation and operation and maintenance solution output provided in the embodiment of the present application, specifically, the relationship between S2234, S2235 and S2236 of S223 and S2243, S2244, S2245, S2246, S2247, S2248 and S2249 of S224 is shown in FIG. Fig.11 Hint.

[0236] Different from the existing system that only records the asset information and data of the mold, this application needs to record the data and information related to the mold itself, the forming environment and the forming equipment in all aspects of the design, simulation, use and maintenance of the mold. Specifically, in the design and simulation stage, the stress concentration areas of the mold and the locations of the structures prone to fatigue defects are recorded. This information serves as an important reference for the management of the entire life cycle and can guide the installation location of the sensor and the key areas of focus for offline detection. In the operation stage, the perception and measurement information of the mold status, the forming environment and the forming equipment status is recorded. In the operation and maintenance stage, the offline detection and evaluation results of the mold, the number of various anomalies and failures, and the maintenance and repair information are recorded. The recorded information and data will be used for the above model-based feature distance comparison, comprehensive performance evaluation, health calculation, prediction and calibration, performance judgment and performance repair method determination.

[0237] Finally, it should be noted that the present application establishes or improves a mold performance management system to realize the entry, storage, query and management of relevant information and data of molds, forming environments and forming equipment, as well as the establishment, call, adjustment and push of model results of relevant models. The state of the overall forming system including the forming environment, the forming main machine and the forming auxiliary machine, and the mold is perceived and measured, so as to accurately and comprehensively reflect the working state of the forming system. A mold performance, health evaluation, prediction and calibration model is established, and the results of online monitoring and offline detection are used to judge the state of the mold, determine whether the forming environment and forming performance are abnormal, and predict the change trend of the mold health, so as to provide a quantitative decision-making basis for more accurate mold performance management and maintenance. And further based on the mold performance and health model, a mold abnormal performance analysis method is established to provide a basis for the selection of mold operation and maintenance methods. Furthermore, in order to ensure the reliability of the online health model, the present application uses the data and judgment results of the appearance state and structural integrity detection of the mold during the deactivation period to calibrate and train the online mold health model. This application realizes key status monitoring and active performance management of the entire life cycle of the mold, improves the performance stability and utilization efficiency of the mold, and reduces the increase in production costs and decrease in production efficiency caused by mold performance degradation and failure.

[0238] This application improves the efficiency and stability of the mold, reduces production costs, and improves production efficiency by realizing the management of the entire life cycle of the mold. On the one hand, by integrating or developing a basic mold information management module, the basic information of the mold, such as mold model, specifications, manufacturer, design service life, etc., can be entered, stored, queried and managed. On the second hand, a mathematical model of the mold is established through mold design software and finite element analysis of mold performance, and the performance of the mold during use is simulated and analyzed so as to optimize and adjust it in the design stage. On the third hand, various sensors and detection devices are installed on the mold or mold-related auxiliary equipment to monitor and detect the state of the mold during use in real time, and to discover and handle abnormal situations. At the same time, during the period when the mold is out of service for maintenance, the mold is tested offline for comprehensive inspection and maintenance. On the fourth hand, a mold performance evaluation and life prediction model is established by analyzing the historical state data and operation data of the mold, so as to realize the precise management and maintenance of the mold. On the fifth hand, the mold operation and maintenance process is optimized by combining the real-time online status of the mold and the offline detection results, so as to improve the service life and stability of the mold and reduce maintenance costs and downtime. The sixth aspect is to analyze the data collected by the mold management system and generate reports to support decision-making in the production process. At the same time, a friendly operation interface is provided for different users, and corresponding permissions are assigned according to user roles to achieve system security and availability.

[0239] Based on any of the above embodiments, the present application embodiment provides a mold life cycle management system. Fig.12 The architecture diagram of the mold life cycle management system provided in the embodiment of the present application is as follows: Fig.12 As shown, the system includes:

[0240] The mold design and monitoring module 1210 is used to optimize the mold design by using the finite element simulation method in the mold design stage, and use the finite element simulation results to assist in determining the selection and installation position of the sensor according to the forming process conditions and range, so as to install sensors on the mold and its forming system, monitor the performance of the mold and the forming system during use online, and obtain the performance data monitored online;

[0241] The mold performance evaluation and maintenance module 1220 is used to determine the online status of the mold and the forming system during use and the online health of the mold based on the performance data monitored online, and to predict the future health of the mold based on the online health of the mold over multiple cycles; based on the online status of the mold and the forming system, the online health and future health of the mold, and in combination with the offline status of the mold during the period of inactivity detected by the offline detection method, adjust the mold maintenance strategy.

[0242] It is understandable that the detailed functional implementation of each of the above modules can be found in the introduction of the aforementioned method embodiment, and will not be repeated here.

[0243] Based on the method in the above embodiment, an embodiment of the present application provides an electronic device. Fig.13 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Fig.13 As shown, the electronic device may include: a processor 1310, a communication interface 1320, a memory 1330 and a communication bus 1340, wherein the processor 1310, the communication interface 1320 and the memory 1330 communicate with each other through the communication bus 1340. The processor 1310 may call the logic instructions in the memory 1330 to execute the method in the above embodiment.

[0244] In addition, the logic instructions in the above-mentioned memory 1330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.

[0245] Based on the method in the above embodiment, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method in the above embodiment.

[0246] Based on the method in the above embodiment, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method in the above embodiment.

[0247] It is understandable that the processor in the embodiment of the present application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor.

[0248] It should be understood that the various numerical numbers involved in the embodiments of the present application are only used for the convenience of description and are not used to limit the scope of the embodiments of the present application.

[0249] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A mold life cycle management method, characterized in that: The following steps are involved: Step S101, using the finite element simulation method to optimize the mold design in the mold design stage, and using the finite element simulation results to assist in determining the selection and installation position of the sensor according to the forming process conditions and range, so as to install the sensor on the mold and its forming system, monitor the performance of the mold and the forming system during use online, and obtain the performance data monitored online; Step S102, judging the online status of the mold and the forming system during use and the online health of the mold based on the performance data monitored online, and predicting the future health of the mold based on the online health of the mold over multiple cycles; Adjust mold maintenance strategy based on the online status of the mold and forming system, the online health and future health of the mold, and the offline status of the mold during the idle period detected by offline detection methods; Step S102 specifically includes: Based on the performance of the forming equipment in the performance data monitored online, determine whether the state of the forming equipment in the forming system is abnormal; If the state of the forming equipment is abnormal, the adjustment and maintenance strategy of the forming equipment is triggered; if the state of the forming equipment is normal, the performance data monitored online is divided into performance data under each predefined working condition, and the characteristic similarity distance between it and the benchmark performance data under each predefined working condition is calculated; based on the forming environment component, mold posture component, mold structure component and forming performance component in the characteristic similarity distance, it is judged whether the forming environment, mold posture, mold structure and forming performance are abnormal, so as to trigger the forming environment adjustment strategy, mold posture adjustment strategy, mold structure maintenance strategy, mold performance compensation strategy and mold performance maintenance strategy; Based on the feature similarity distance and the preset performance weight matrix, the comprehensive performance characterization vector is determined; the mold health function is applied to calculate the online health of the mold based on the comprehensive performance characterization vector; whether the online health and future health of the mold are abnormal is determined to trigger the mold maintenance strategy and mold predictive maintenance strategy; Based on the offline status, the mold is evaluated to see if it is fatigued or damaged, and the corresponding maintenance strategy is triggered to perform comprehensive inspection and maintenance; the offline detection method includes but is not limited to appearance inspection, and metallographic detection, ultrasonic detection, magnetic sensor and vibration detection are used to determine whether the mold has structural problems or other damage.

2. The method according to claim 1, characterized in that The mold health function is calculated based on the following formula: Among them, h0 is the initial health of the cycle, Y (t, m) is the comprehensive performance characterization vector at time t, m is the number of signals, and E (e j (t), j = 1, ..., n} is the equivalent number of events in the mold within the cycle, n is the event category, Δt is the effective duration of the mold use within the cycle, w y That is, {w y，i , i = 1, ..., m} is the health impact coefficient of the performance characterization vector, w e That is, {w e，j , j = 1, ..., n} is the health impact coefficient of the event equivalent number, w t is the health impact coefficient of the effective use time of the mold, θ, that is, {θ y ,θ e ,θ t }Adjust parameters for the health function.

3. The method according to claim 2, characterized in that Also includes: When the difference between the online health of the mold and the offline evaluation result of the health exceeds a certain predefined deviation threshold, the mold health function is calibrated or trained; the offline evaluation result of the health is determined based on the offline state of the mold and the corresponding relationship between the preset offline state and the offline evaluation result; The calibrated or retrained health function recalculates the online health of the mold; The calibration is specifically performed by solving the following optimization problem to calibrate {θ y ,θ e ,θ t }: In the formula, d is the number of cycles, h(t i ) is the online health of the ith cycle, a(t i ) is the offline evaluation result of the health status in the i-th cycle; The training is specifically carried out by solving the following optimization problem and training {w y , w e , w t } and {θ y ,θ e ,θ t }: In the formula, δw and δθ are the changes of w and θ respectively, λ w is the weight of w, λ θ is the weight of θ.

4. The method according to claim 1, characterized in that: If the future health of the mold is abnormal, the future change trend of each component is predicted based on the prediction model corresponding to each component in the comprehensive performance characterization vector; based on the future change trend of each component, the component with the largest change is determined and used as the key component affecting the future health; Determine the corresponding mold predictive maintenance strategy based on the key components that affect future health; The prediction models corresponding to the components are obtained by training based on the historical data of the components.

5. A mold full life cycle management system, characterized in that: include: The mold design and monitoring module is used to optimize the mold design by using the finite element simulation method in the mold design stage, and use the finite element simulation results to assist in determining the selection and installation location of sensors based on the forming process conditions and range, so as to install sensors on the mold and its forming system to monitor the performance of the mold and forming system during use online and obtain the performance data monitored online; The mold performance evaluation and maintenance module is used to judge the online status of the mold and the forming system during use and the online health of the mold based on the performance data monitored online, and to predict the future health of the mold based on the online health of the mold over multiple cycles; Adjust mold maintenance strategy based on the online status of the mold and forming system, the online health and future health of the mold, and the offline status of the mold during the idle period detected by offline detection methods; Mold performance evaluation and maintenance module, specifically for: Based on the performance of the forming equipment in the performance data monitored online, determine whether the state of the forming equipment in the forming system is abnormal; If the state of the forming equipment is abnormal, the adjustment and maintenance strategy of the forming equipment is triggered; If the state of the forming equipment is normal, the performance data monitored online is divided into performance data under each predefined working condition, and the characteristic similarity distance between the performance data and the benchmark performance data under each predefined working condition is calculated; Based on the forming environment component, mold posture component, mold structure component and forming performance component in the feature similarity distance, it is judged whether the forming environment, mold posture, mold structure and forming performance are abnormal, so as to trigger the forming environment adjustment strategy, mold posture adjustment strategy, mold structure maintenance strategy, mold performance compensation strategy and mold performance maintenance strategy; Based on the feature similarity distance and the preset performance weight matrix, a comprehensive performance representation vector is determined; Apply the mold health function to calculate the online health of the mold based on the comprehensive performance characterization vector; Determine whether the online health and future health of the mold are abnormal to trigger mold maintenance strategies and mold predictive maintenance strategies; Based on the offline status, the mold is evaluated to see if it is fatigued or damaged, and the corresponding maintenance strategy is triggered to perform comprehensive inspection and maintenance; the offline detection method includes but is not limited to appearance inspection, and metallographic detection, ultrasonic detection, magnetic sensor and vibration detection are used to determine whether the mold has structural problems or other damage.

6. An electronic device, characterized in that: include: at least one memory for storing a computer program; At least one processor is used to execute the program stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program runs on a processor, the processor is caused to execute the method according to any one of claims 1 to 4.

8. A computer program product, characterized in that When the computer program product runs on a processor, the processor is enabled to execute the method according to any one of claims 1 to 4.

Images