Stamping die service life dynamic evaluation system and method based on multi-mode sensing

Abstract

The invention relates to a stamping die service life dynamic evaluation system and method based on multi-mode sensing. According to the system, a feature extraction module obtains stress distribution data and crack position information and inputs the data into a trained three-dimensional stress field model to obtain dynamic feature parameters of a stress concentration area under high-frequency cyclic loads; the crack prediction module judges the dynamic characteristic parameters based on a preset threshold value, and once the dynamic characteristic parameters exceed the preset threshold value, analysis is carried out according to the mapping relation between the stress level and the crack rate and the microstructure evolution rule at the same time, and a crack propagation prediction value is generated. And the service life evaluation module inputs the crack propagation prediction value into the trained health state evaluation model, and finally the service life evaluation value of the stamping die is obtained. The service life evaluation value of the stamping die under the existing working condition and the operation condition is accurately estimated, the production efficiency is effectively improved, the production cost is reduced, the potential crack risk of the die is perceived in advance, and accurate and dynamic die service life evaluation is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent detection technology, and in particular relates to a system and method for dynamically evaluating the life of a stamping die based on multimodal sensing. Background Art

[0002] With the development of intelligent detection technology, a dynamic assessment technology for the life of stamping dies based on multimodal sensing has emerged. In modern manufacturing, stamping technology is widely used as an important forming method. The performance and life of stamping dies are directly related to product quality, production efficiency and enterprise costs. Traditional stamping die life assessment methods mostly rely on a single data source or empirical judgment, which is difficult to fully and accurately reflect the actual condition of the die. As industrial production develops towards high precision, high efficiency and high reliability, dies operate under complex working conditions such as high-frequency cyclic loads, and their surface and internal stress changes, crack initiation and expansion, etc. are extremely complex. The rise of multimodal sensing technology makes it possible to obtain multi-dimensional status data of dies, but it is difficult to effectively integrate data and build an accurate and dynamic die life assessment method. Summary of the Invention

[0003] Based on this, it is necessary to provide a dynamic assessment system and method for stamping die life based on multimodal sensing, which can detect the potential crack risk of the die in advance and realize accurate and dynamic die life assessment, in order to address the above technical problems.

[0004] In a first aspect, the present application provides a dynamic assessment system for stamping die life based on multimodal sensing, comprising:

[0005] The feature extraction module is used to obtain stress distribution data on the mold surface and interior from the multimodal sensor network; and obtain crack location information from the acoustic emission signal; and is also used to input the stress distribution data and crack location information into the trained three-dimensional stress field model to obtain the dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading.

[0006] The crack prediction module is used to judge the dynamic characteristic parameters based on a preset threshold. If the dynamic characteristic parameters exceed the preset threshold, the mapping relationship between stress level and crack rate is combined with the microstructure evolution law to generate a crack extension prediction value.

[0007] The life assessment module is used to input the crack growth prediction value into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

[0008] In one embodiment, stress distribution data and crack location information are input into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading, including:

[0009] Obtain the three-dimensional grid node stress value corresponding to the stress distribution data and the expansion direction vector corresponding to the crack position information.

[0010] The node stress value and expansion direction vector are input into the trained three-dimensional stress field model, and the dynamic stress amplitude on the local grid node is calculated by combining the frequency and amplitude range in the high-frequency cyclic load parameters.

[0011] Based on the dynamic stress amplitude and the number of load cycles, a preset fatigue accumulation factor calculation formula is used to generate a stress fluctuation index associated with the crack propagation path.

[0012] Feature fusion is performed based on the stress fluctuation index to obtain the dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading; the dynamic characteristic parameters include the fatigue damage degree of the stress concentration area, the stress cycle change rate, and the fluctuation range of the stress concentration coefficient of the key parts.

[0013] In one embodiment, a stress fluctuation index associated with a crack propagation path is generated based on the dynamic stress amplitude and the number of load cycles using a preset fatigue accumulation factor calculation formula, including:

[0014] The stress fluctuation index is calculated using the following formula:

[0015]

[0016] Where Δσ represents the stress fluctuation index, K f represents the fatigue cumulative damage factor, σ i represents the stress amplitude of the i-th cycle, N i Indicates the corresponding number of load cycles, n indicates the total number of stress cycles; n i Indicates the actual number of cycles, N i represents fatigue life, γ i represents the fatigue damage weight coefficient.

[0017] In one embodiment, a crack growth prediction value is generated based on a mapping relationship between stress level and crack rate combined with analysis of microstructure evolution laws, including:

[0018] The dynamic feature parameters are obtained and compared with the preset threshold to obtain the feature judgment result.

[0019] The mapping relationship between stress level and crack rate is analyzed according to the characteristic judgment results to obtain the current stress state parameters.

[0020] The algorithm is used to extract stress state parameters and obtain the associated microstructure evolution sequence data.

[0021] A crack growth trend map is constructed based on the evolution sequence data and a path parameter set is generated using a time series prediction algorithm.

[0022] The crack propagation path coordinate matrix is constructed according to the path parameter set, and the crack propagation prediction value is calculated using the formula.

[0023] In one embodiment, a crack propagation path coordinate matrix is constructed based on the path parameter set, and a crack propagation prediction value is calculated using a formula including:

[0024] The crack length, crack width, and crack depth data in the path parameter set are mapped into the coordinate matrix using the preset coordinate set conversion rules to obtain the crack propagation path coordinate matrix.

[0025] The correlation between crack velocity, crack angle and crack force is calculated based on the crack propagation path coordinate matrix to obtain the crack propagation trend.

[0026] Among them, if the crack velocity is higher than the preset threshold, the linear regression algorithm is used to fit the crack length, crack width and crack depth in the crack propagation path coordinate matrix to obtain the crack propagation prediction value.

[0027] If the crack angle exceeds the preset range, the support vector machine algorithm is used to analyze the crack velocity and crack force in the crack growth path coordinate matrix to correct the crack growth prediction value.

[0028] If the crack force distribution is uneven, the decision tree algorithm is used to classify the crack angle and crack depth in the crack growth path coordinate matrix and adjust the crack growth prediction value.

[0029] In one embodiment, the correlation between the crack velocity, crack angle, and crack force is calculated based on the crack propagation path coordinate matrix to obtain the crack propagation trend, including:

[0030] The correlation between crack velocity, crack angle, and crack force is calculated using the following formula.

[0031]

[0032]

[0033] Where v represents the crack growth rate, K1 represents the first mode stress intensity factor, r represents the distance from the crack tip, θ represents the crack growth angle, and θ c represents the critical crack growth angle, K I and K II They represent the first and second mode stress intensity factors, F represents the driving force required for crack extension, γ represents the surface energy density, L represents the crack length, Represents the slope of the crack path.

[0034] In one embodiment, the crack growth prediction value is input into a trained health status assessment model to obtain a life assessment value of the stamping die, including:

[0035] The time-frequency features of the crack extension prediction value are extracted to obtain a multi-dimensional feature sequence.

[0036] The multidimensional feature sequence is input into the convolution kernel group to calculate the spatial correlation and obtain the crack evolution trend tensor.

[0037] The crack evolution trend tensor is transformed by attenuation function to obtain the probability density distribution of the remaining life.

[0038] The remaining life distribution value is obtained by assigning a dynamic weight coefficient according to the probability density distribution and performing calculations.

[0039] The remaining life distribution value is input into the trained health status assessment model and calculated using the formula to obtain the life assessment value of the stamping die.

[0040] The lifespan estimate is calculated using the following formula:

[0041]

[0042] Among them, E represents the life evaluation parameter, ω j Represents the weight coefficient of the jth health state, R j represents the remaining life value of the jth state, λ represents the decay coefficient, t represents the usage time, and n represents the number of healthy states.

[0043] In a second aspect, the present application also provides a method for dynamic assessment of stamping die life based on multimodal sensing, the method comprising:

[0044] The stress distribution data on the mold surface and interior are acquired from the multimodal sensor network; the crack location information is obtained from the acoustic emission signal; the stress distribution data and crack location information are input into the trained three-dimensional stress field model to obtain the dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading.

[0045] The dynamic characteristic parameters are judged based on the preset threshold. If the dynamic characteristic parameters exceed the preset threshold, the mapping relationship between stress level and crack rate is combined with the microstructure evolution law to generate a crack extension prediction value.

[0046] The crack growth prediction value is input into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

[0047] In a third aspect, the present application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program as described above in the system and method.

[0048] In a fourth aspect, the present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it is as described in the above system and method.

[0049] The above-mentioned multimodal sensing-based dynamic life assessment system and method for stamping dies includes three modules: feature extraction, crack prediction, and life assessment. The feature extraction module acquires stress distribution data on the die surface and interior, as well as crack location information from acoustic emission signals. This data is then input into a trained three-dimensional stress field model to obtain dynamic characteristic parameters for stress concentration areas under high-frequency cyclic loading. The crack prediction module determines these dynamic characteristic parameters based on a preset threshold. Once the dynamic characteristic parameters exceed the preset threshold, it analyzes the mapping relationship between stress level and crack rate, combined with microstructural evolution, to generate a crack propagation prediction value. The life assessment module inputs the crack propagation prediction value into a trained health status assessment model, ultimately obtaining a life assessment value for the stamping die. This value represents the estimated remaining useful life of the stamping die under current operating and working conditions. This system and method accurately estimates the life assessment value of the stamping die under existing operating and working conditions, effectively improving production efficiency and reducing production costs, enabling early detection of potential crack risks in the die and achieving accurate and dynamic die life assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 A structural block diagram of a stamping die life dynamic assessment system based on multimodal sensing provided by an embodiment of the present invention;

[0052] Figure 2 A flowchart of inputting stress distribution data and crack location information into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of stress concentration areas under high-frequency cyclic loading is provided in an embodiment of the present invention;

[0053] Figure 3 A flowchart for generating crack propagation prediction values based on the mapping relationship between stress level and crack rate combined with analysis of microstructural evolution laws provided in an embodiment of the present invention;

[0054] Figure 4 A flowchart of inputting a crack growth prediction value into a trained health status assessment model to obtain a life assessment value of a stamping die provided by an embodiment of the present invention;

[0055] Figure 5 A flowchart of a method for dynamically evaluating the life of a stamping die based on multimodal sensing is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0057] First, the implementation environment of the embodiment of the present application is described. Exemplarily, the implementation environment includes a sensor system, a data acquisition and transmission device, and a data processing and analysis unit.

[0058] In a dynamic stamping die life assessment system and method based on multimodal sensing, various sensors in the sensor system, such as force, displacement, temperature, and vibration, collect relevant physical quantity data during the stamping process in real time and output analog signals. The analog signals are then transmitted to a data acquisition and transmission device, where a data acquisition card converts the analog signals into digital signals. A signal amplifier amplifies the weak signals. If a wireless transmission module is used, the processed digital signals are then wirelessly transmitted to a data processing and analysis unit. If a wired connection is used, the data is directly transmitted via a cable to a data processing and analysis unit, such as an industrial computer or edge computing device with powerful computing capabilities, where in-depth data mining and die life assessment calculations are performed.

[0059] The sensor system integrates multiple types of sensors. The force sensor accurately measures the pressure exerted on the mold during the stamping process and can keenly capture the details of pressure changes; the displacement sensor is responsible for monitoring the displacement of the mold's moving parts to ensure that the accuracy and stability of the mold movement can be tracked in real time; the temperature sensor is arranged in the mold cavity, core and other heat-prone parts to dynamically monitor the temperature changes caused by friction during the stamping process; the vibration sensor is installed on the mold body or base to collect vibration signals when the mold is working. Abnormal vibration may indicate potential problems such as wear and loosening of the internal structure of the mold.

[0060] As a core component of data acquisition and transmission equipment, data acquisition cards (DACs) are capable of rapidly and accurately digitizing multiple analog signals from sensor systems and performing preliminary signal conditioning to ensure data quality and usability. Because sensor output signals are typically weak, signal amplifiers play an indispensable role. Their low noise, high gain, and excellent linearity boost signal strength to a level suitable for transmission. In the data transmission phase, wireless remote data transmission is implemented based on the actual application scenario and data transmission requirements.

[0061] Industrial computers in the data processing and analysis unit run complex data analysis algorithms and specialized mold life assessment software. Edge computing devices offer unique advantages in scenarios requiring extremely high real-time performance. They can quickly perform pre-processing operations such as screening, filtering, and feature extraction on sensor data close to the data source, effectively reducing data transmission volume and latency, and providing strong support for subsequent accurate analysis. Through in-depth mining and analysis of collected data, dynamic assessment of stamping die life is achieved.

[0062] In combination with the above implementation environment, the application scenarios of the embodiments of the present application are explained.

[0063] The embodiments of the present application provide a system and method for dynamically evaluating the life of a stamping die based on multimodal sensing. For example, the system and method for dynamically evaluating the life of a stamping die based on multimodal sensing provided in the embodiments of the present application can be applied to at least one of the following scenarios, including but not limited to.

[0064] First, this dynamic assessment system and method for stamping die life based on multimodal sensing is applied in automotive manufacturing scenarios. In automotive manufacturing, sensors are installed at the contact point between the stamping machine and the die, accurately capturing the significant pressure fluctuations during each stamping operation. Displacement sensors are located on active components such as the die slide and blank holder, monitoring their motion accuracy in real time to ensure dimensional consistency of vehicle body parts. Temperature sensors are placed within the die cavity to prevent degradation of the die material due to overheating. Vibration sensors are installed at the die base to promptly detect abnormal vibration caused by structural loosening or wear due to prolonged, high-load operation. Data acquisition and transmission equipment rapidly collects multimodal data, converts and conditions the signals using a high-speed data acquisition card, and transmits the data stably to an industrial computer on the workshop floor via wired transmission. The industrial computer runs life assessment software specifically customized for automotive stamping dies. Using complex algorithms, it deeply analyzes the data, assesses die life in real time, and provides early warning of die failures, ensuring efficient and high-quality automotive production.

[0065] Second, this multimodal sensing-based dynamic assessment system and method for stamping die lifespan is applied to electronic equipment production. Stamping dies in electronic equipment production are used to manufacture precision components such as mobile phone cases and computer radiators. The sensor system includes force sensors, which precisely measure minute stamping pressures to ensure the precision and quality of electronic component molding. These sensors are installed at subtle stress points in the mold. Displacement sensors closely monitor the displacement of minute moving parts in the mold to ensure that product dimensions meet stringent electronic equipment manufacturing standards. Temperature sensors are placed in key heat-generating areas of the mold. Although electronic equipment stamping dies experience relatively low pressures, high-frequency stamping can easily cause localized temperature rises, impacting mold lifespan. Vibration sensors detect subtle mold vibrations to prevent vibration-induced surface defects in components. Data acquisition and transmission equipment utilizes high-precision data acquisition cards and low-noise signal amplifiers to convert and amplify weak sensor signals and rapidly transmit the data to edge computing equipment in the data processing center. The edge computing equipment performs preliminary data screening and feature extraction, then transmits key data to backend servers for in-depth analysis. Using specialized algorithms, they dynamically assess mold lifespan, helping electronic equipment manufacturers efficiently plan mold maintenance and replacement.

[0066] Third, this multimodal sensing-based dynamic stamping die life assessment system and method is applied to aerospace component manufacturing. Aerospace component manufacturing places extremely high demands on stamping die precision and reliability. The sensor system comprehensively monitors die operation. Force sensors, installed in areas of the die subject to extreme loads, accurately measure forces during ultra-high-pressure stamping, assessing die performance under demanding conditions. Displacement sensors precisely track the displacement of complex moving parts in the die, ensuring high-precision molding of complex aerospace component shapes. Temperature sensors are located in areas of the die subject to significant thermal shock. Aerospace component stamping often involves specialized materials and processes, resulting in complex temperature fluctuations. Vibration sensors monitor die vibration in real time to prevent vibration-induced internal defects in the component. Data acquisition and transmission equipment utilizes a highly interference-resistant data acquisition card and a high-gain signal amplifier. Data is transmitted via high-speed, stable wired transmission methods, such as optical fiber, to an industrial-grade data processing and analysis unit with superior computing power. This unit runs a specialized aerospace die life assessment algorithm, combining multimodal data to accurately and dynamically assess die life.

[0067] In one embodiment, Figure 1 As shown, the present application provides a dynamic assessment system for stamping die life based on multimodal sensing, which may include:

[0068] Feature extraction module 101 is used to obtain stress distribution data on the mold surface and interior from a multimodal sensor network; and to obtain crack location information from acoustic emission signals; and is also used to input the stress distribution data and crack location information into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of stress concentration areas under high-frequency cyclic loading.

[0069] This module uses a multimodal sensor network and advanced sensing technology to comprehensively acquire stress distribution data on the mold surface and various key internal parts. The data reflects the real-time state and spatial distribution of the stress borne by the mold during the stamping operation. At the same time, through professional collection and in-depth analysis of acoustic emission signals, it can keenly capture the crack location information contained in the acoustic emission signal. Subsequently, the acquired stress distribution data and crack location information are input in an orderly and accurate manner into a three-dimensional stress field model that has been pre-trained, verified and optimized with a large amount of data. After receiving the input data, the dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loads are obtained in combination with the actual working condition parameters of the high-frequency cyclic load, such as the frequency range and amplitude of the load.

[0070] The crack prediction module 102 is used to judge the dynamic characteristic parameters based on a preset threshold. If the dynamic characteristic parameters exceed the preset threshold, the mapping relationship between stress level and crack rate is combined with the microstructure evolution law to generate a crack extension prediction value.

[0071] The module first compares the dynamic characteristic parameters with the thresholds pre-set in the system. Once the dynamic characteristic parameters exceed the preset threshold, it means that the operating status of the mold may deviate from the normal track, and there is a potential risk of crack expansion. At this time, the crack prediction module quickly initiates an in-depth analysis mechanism, which is based on the mapping relationship between stress level and crack rate mastered through long-term research and practice. This relationship reflects the law of change in the rate of crack initiation and expansion under different stress conditions. At the same time, combined with the evolution law of the material microstructure, because the microstructure of the mold material will change with factors such as stress state and temperature changes during the stamping process, the evolution of this microstructure is closely related to the generation and development of cracks. By comprehensively considering these two factors, using complex mathematical models and data analysis algorithms, the current state of the mold is deeply analyzed, and finally an accurate crack expansion prediction value is generated.

[0072] The life assessment module 103 is used to input the crack extension prediction value into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

[0073] This module will use the received crack propagation prediction value as key input data and import it into a health status assessment model that has been pre-trained with massive historical data, repeatedly debugged and optimized, and rigorously verified. After receiving the crack propagation prediction value, the model will comprehensively consider the actual working conditions of the current stamping die, such as stamping frequency, pressure intensity, working temperature environment and other operating conditions. At the same time, combined with the inherent properties of the die itself, such as material properties and structural design parameters, it uses a series of complex and precise algorithms for in-depth calculations and comprehensive analysis, and finally outputs the life assessment value of the stamping die. This life assessment value is of extremely high practical value. It intuitively and accurately presents the remaining effective time of the stamping die, which is scientifically estimated while maintaining the existing working conditions and operating conditions.

[0074] In one embodiment, Figure 2 As shown, inputting stress distribution data and crack location information into the trained three-dimensional stress field model to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading can include the following steps:

[0075] Step S201 : obtaining the three-dimensional grid node stress value corresponding to the stress distribution data and the expansion direction vector corresponding to the crack position information.

[0076] Step S202 : Input the node stress value and the expansion direction vector into the trained three-dimensional stress field model, and calculate the dynamic stress amplitude on the local grid node in combination with the frequency and amplitude range in the high-frequency cyclic load parameters.

[0077] Step S203 : generating a stress fluctuation index associated with the crack propagation path based on the dynamic stress amplitude and the number of load cycles using a preset fatigue accumulation factor calculation formula.

[0078] Step S204 , feature fusion is performed based on the stress fluctuation index to obtain dynamic feature parameters of the stress concentration area under high-frequency cyclic loading; the dynamic feature parameters include the fatigue damage degree of the stress concentration area, the stress cycle change rate, and the fluctuation range of the stress concentration coefficient of the key part.

[0079] Specifically, first obtain the three-dimensional grid node stress value corresponding to the stress distribution data, and the expansion direction vector corresponding to the crack location information. Then, input the node stress value and expansion direction vector into the trained three-dimensional stress field model, and combine the frequency and amplitude range in the high-frequency cyclic load parameters to accurately calculate the dynamic stress amplitude on the local grid node, so as to effectively capture the real-time stress state of the mold. Subsequently, based on the dynamic stress amplitude and the number of load cycles, the preset fatigue accumulation factor calculation formula is used to generate a stress fluctuation index closely related to the crack propagation path, thereby quantifying the characteristics of stress changes with time and load. Finally, feature fusion is performed based on the stress fluctuation index to obtain dynamic characteristic parameters covering the fatigue damage degree of the stress concentration area, the stress cycle change rate, and the fluctuation range of the stress concentration coefficient of key parts.

[0080] This embodiment can comprehensively and accurately analyze the complex mechanical state of the stress concentration area of the stamping die under high-frequency cyclic loads, providing crucial and highly reliable data basis for the subsequent accurate evaluation of the die life and prediction of crack propagation trends, effectively improving the accuracy and practicality of the entire die life dynamic evaluation system, helping enterprises to better control the equipment status during the production process and optimize production decisions.

[0081] In one embodiment, generating a stress fluctuation index associated with a crack propagation path based on the dynamic stress amplitude and the number of load cycles using a preset fatigue accumulation factor calculation formula may include the following steps:

[0082] The stress fluctuation index is calculated using the following formula:

[0083]

[0084] Where Δσ represents the stress fluctuation index, K f represents the fatigue cumulative damage factor, σ i represents the stress amplitude of the i-th cycle, N i Indicates the corresponding number of load cycles, n indicates the total number of stress cycles; n i Indicates the actual number of cycles, N i represents fatigue life, γ i represents the fatigue damage weight coefficient.

[0085] By applying a preset fatigue accumulation factor formula based on dynamic stress amplitude and the number of load cycles, a stress fluctuation index can be accurately calculated. This stress fluctuation index comprehensively and quantitatively reflects the stress fluctuation characteristics of the stamping die during operation as a function of load cycles and is closely related to the crack propagation path. In-depth analysis of the mold's fatigue damage mechanism provides key quantitative indicators, which helps to more accurately assess the mold's health under high-frequency cyclic loading. This in turn provides a solid data foundation for predicting mold life and formulating reasonable maintenance strategies. This improves the scientific nature and reliability of the entire dynamic assessment system for stamping die life, which is of great significance for ensuring the stability and efficiency of the production process.

[0086] In one embodiment, Figure 3 As shown in FIG, based on the mapping relationship between stress level and crack rate combined with the microstructure evolution law, the crack extension prediction value is generated, which can include the following steps:

[0087] Step S301: obtain dynamic feature parameters and compare them with preset thresholds to obtain feature judgment results.

[0088] Step S302: Analyze the mapping relationship between stress level and crack rate according to the feature judgment result to obtain the current stress state parameter.

[0089] Step S303: extract stress state parameters using an algorithm to obtain associated microstructure evolution sequence data.

[0090] Step S304: construct a crack growth trend map based on the evolution sequence data and generate a path parameter set using a time series prediction algorithm.

[0091] Step S305: construct a crack propagation path coordinate matrix according to the path parameter set, and calculate the crack propagation prediction value using a formula.

[0092] First, the dynamic characteristic parameters derived from the previous analysis are obtained and carefully compared with the thresholds preset in the system to obtain the characteristic judgment results. Then, based on the characteristic judgment results, the established mapping relationship between stress level and crack rate is deeply analyzed. The stress state parameters of the current mold are derived based on the mapping relationship, thereby accurately locating the stress condition the mold is currently experiencing. A specific algorithm is used to deeply mine and extract the obtained stress state parameters to obtain the associated microstructure evolution sequence data. Subsequently, based on the evolution sequence data, a map that intuitively presents the crack propagation trend is constructed, and with the help of a time series prediction algorithm, a path parameter set containing rich information is generated. Finally, a crack propagation path coordinate matrix is constructed based on the path parameter set, and then a special formula is used for precise calculation to finally obtain the crack propagation prediction value.

[0093] This embodiment analyzes and predicts the crack growth of stamping dies from multiple dimensions, fully considering key factors such as stress level, crack rate, and material microstructural evolution. This not only provides key data support for assessing the remaining life of stamping dies, helping companies predict when dies may fail in advance, but also helps companies rationally arrange mold maintenance plans and take timely measures before cracks propagate and affect production. This greatly improves the stability and continuity of the production process, avoids production stagnation caused by sudden mold failures, reduces production costs, enhances the company's competitiveness in the market, and effectively ensures the efficiency and safety of the entire production operation.

[0094] In one embodiment, constructing a crack propagation path coordinate matrix based on the path parameter set and calculating the crack propagation prediction value using a formula may include the following steps:

[0095] Step S401 : Mapping the crack length, crack width, and crack depth data in the path parameter set into a coordinate matrix using a preset coordinate set conversion rule to obtain a crack propagation path coordinate matrix.

[0096] Step S402 : calculating the correlation between the crack velocity, crack angle, and crack force based on the crack propagation path coordinate matrix to obtain the crack propagation trend.

[0097] Among them, if the crack velocity is higher than the preset threshold, the linear regression algorithm is used to fit the crack length, crack width and crack depth in the crack propagation path coordinate matrix to obtain the crack propagation prediction value.

[0098] If the crack angle exceeds the preset range, the support vector machine algorithm is used to analyze the crack velocity and crack force in the crack growth path coordinate matrix to correct the crack growth prediction value.

[0099] If the crack force distribution is uneven, the decision tree algorithm is used to classify the crack angle and crack depth in the crack growth path coordinate matrix and adjust the crack growth prediction value.

[0100] First, based on the preset coordinate set conversion rules, the crack length, crack width, and crack depth data, which reflect the geometric characteristics of the crack, are accurately mapped into the coordinate matrix, thereby constructing a coordinate matrix that comprehensively describes the crack propagation path. Then, with the help of the constructed crack propagation path coordinate matrix, the intrinsic correlation between the crack speed, crack angle, and crack force is deeply explored, and the crack propagation trend is clearly analyzed. In this process, differentiated algorithm strategies are adopted for different crack propagation conditions. When the crack speed is higher than the preset threshold, the linear regression algorithm is used to fit the crack length, crack width, and crack depth in the crack propagation path coordinate matrix to obtain a more accurate crack propagation prediction value. If the crack angle exceeds the preset range, the support vector machine algorithm is used to analyze the crack speed and crack force in the crack propagation path coordinate matrix, correct the crack propagation prediction value, and ensure the accuracy of the prediction result. When the crack force distribution is uneven, the decision tree algorithm is used to classify the crack angle and crack depth in the crack propagation path coordinate matrix, and the crack propagation prediction value is adjusted to make the prediction result more in line with the actual crack propagation situation.

[0101] Through precise coordinate mapping and multi-dimensional data analysis, comprehensive and in-depth information on crack growth is extracted, improving the accuracy and reliability of crack growth predictions. Based on these predictions, companies can proactively predict mold failure risks and rationally plan mold replacement and repairs. This effectively avoids production interruptions caused by sudden mold damage, reduces production costs, improves production efficiency, strengthens the company's competitive advantage, and ensures the stability and efficiency of the entire production operation.

[0102] In one embodiment, calculating the correlation between crack velocity, crack angle, and crack force based on the crack propagation path coordinate matrix to obtain the crack propagation trend may include the following steps:

[0103] The correlation between crack velocity, crack angle, and crack force is calculated using the following formula.

[0104]

[0105] Where v represents the crack growth rate, X1 represents the first mode stress intensity factor, r represents the distance from the crack tip, θ represents the crack growth angle, and θ c represents the critical crack growth angle, K I and K II They represent the first and second mode stress intensity factors, F represents the driving force required for crack extension, γ represents the surface energy density, L represents the crack length, Represents the slope of the crack path.

[0106] Calculations using a given formula incorporate multiple key factors in the crack growth process into a unified mathematical model. Leveraging the detailed information provided by the crack growth path coordinate matrix, this formula accurately quantifies the interactions and inherent connections between crack velocity, crack angle, and crack force. This quantitative analysis provides a deeper understanding of the physical mechanisms of crack growth and captures the dynamics of the crack growth process.

[0107] In the practical application of stamping dies, crack growth directly impacts the die's service life and performance. The crack growth trends derived through this method provide a key basis for die life assessment. By analyzing the key factors affecting crack growth, targeted improvements can be made to the die's structure and materials, enhancing its crack growth resistance, thereby improving production efficiency and product quality while reducing production costs.

[0108] In one embodiment, Figure 4 As shown, the crack growth prediction value is input into the trained health status assessment model to obtain the life assessment value of the stamping die, which may include the following steps:

[0109] Step S501 : extracting time-frequency features from crack extension prediction values to obtain a multi-dimensional feature sequence.

[0110] Step S502: Input the multidimensional feature sequence into the convolution kernel group to perform spatial correlation calculation to obtain the crack evolution trend tensor.

[0111] Step S503 , performing attenuation function transformation on the crack evolution trend tensor to obtain the probability density distribution of the remaining life.

[0112] Step S504 , assigning a dynamic weight coefficient according to the probability density distribution to perform calculations to obtain a remaining life distribution value.

[0113] Step S505 , inputting the remaining life distribution value into the trained health status assessment model and calculating using a formula to obtain the life assessment value of the stamping die.

[0114] The lifespan estimate is calculated using the following formula:

[0115]

[0116] Among them, E represents the life evaluation parameter, ω j Represents the weight coefficient of the jth health state, R j represents the remaining life value of the jth state, θ represents the decay coefficient, t represents the usage time, and n represents the number of healthy states.

[0117] First, time-frequency features are extracted from the crack growth prediction value. Time-frequency analysis comprehensively analyzes crack growth characteristics from both the time and frequency dimensions, resulting in a multidimensional feature sequence containing information about crack growth at different time scales and frequency components. Next, the multidimensional feature sequence is input into a convolution kernel group. The convolution kernel group exploits the spatial correlations within the feature sequence through convolution operations, resulting in a crack evolution trend tensor. This tensor more clearly illustrates the evolution of cracks in both spatial and temporal dimensions, providing key information for accurately assessing die life. Subsequently, the crack evolution trend tensor is transformed using a decay function. Since die performance gradually degrades with age, the decay function simulates this degradation process, resulting in a probability density distribution of the remaining life. This distribution intuitively reflects the probability of the die falling within different remaining life intervals. Dynamic weighting coefficients are assigned to the probability density distribution for calculation. Different remaining life probabilities are assigned different weights. This dynamic weighting calculation yields a remaining life distribution value, making the assessment results more realistic. Finally, the remaining life distribution value is input into the trained health assessment model, and the life assessment value of the stamping die is calculated using a given formula.

[0118] This embodiment fully exploits the useful information in the crack extension prediction value through a series of operations such as time-frequency feature extraction, spatial correlation calculation, attenuation function transformation, and dynamic weighting, comprehensively and deeply considers the factors affecting the mold life, and improves the accuracy and reliability of the life assessment. From a practical application perspective, accurate life assessment values can provide strong support for the company's production management. Based on the assessment results, the company can reasonably arrange the replacement and maintenance plan of the mold, avoid production interruptions and increased costs caused by premature mold failure, and prevent the waste of resources caused by excessive maintenance, thereby improving production efficiency, reducing production costs, and enhancing the company's competitiveness in the market.

[0119] In one embodiment, Figure 5 As shown, the present application also provides a dynamic assessment method for stamping die life based on multimodal sensing, which may include the following steps:

[0120] Step S601: Obtain stress distribution data on the mold surface and interior from a multimodal sensor network; and obtain crack location information from acoustic emission signals; input the stress distribution data and crack location information into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading.

[0121] Step S602 : The dynamic characteristic parameter is judged based on a preset threshold. If the dynamic characteristic parameter exceeds the preset threshold, an analysis is performed based on the mapping relationship between stress level and crack rate combined with the microstructure evolution law to generate a crack extension prediction value.

[0122] In step S603, the crack extension prediction value is input into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

[0123] The above-mentioned dynamic assessment method for stamping die life based on multimodal sensing first obtains stress distribution data on the die surface and interior through a multimodal sensor network, simultaneously extracts crack location information from the acoustic emission signal, and inputs this data into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading. Subsequently, these dynamic characteristic parameters are compared with a preset threshold. Once the parameter exceeds the threshold, an in-depth analysis is performed based on the mapping relationship between stress level and crack rate combined with the microstructural evolution law to generate a crack propagation prediction value. Finally, the crack propagation prediction value is input into a trained health status assessment model, and the model calculation results in the stamping die life assessment value, which represents the estimated remaining effective time of the stamping die under the current working and operating conditions. The above method accurately estimates the life assessment value of the stamping die under the existing working and operating conditions, effectively improving production efficiency, reducing production costs, and detecting potential crack risks in the die in advance, thus achieving accurate and dynamic die life assessment.

[0124] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0125] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the system and method for dynamic evaluation of stamping die life based on multimodal sensing as described above are implemented.

[0126] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0127] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0128] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A dynamic assessment system for stamping die life based on multimodal sensing, characterized by: The system comprises: A feature extraction module is used to obtain stress distribution data on the mold surface and interior from a multimodal sensor network; and to obtain crack location information from acoustic emission signals; and is also used to input the stress distribution data and crack location information into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of stress concentration areas under high-frequency cyclic loading; a crack prediction module, configured to judge the dynamic characteristic parameters based on a preset threshold value, and if the dynamic characteristic parameters exceed the preset threshold value, generate a crack propagation prediction value by analyzing the mapping relationship between stress level and crack rate in combination with the microstructure evolution law; The life assessment module is used to input the crack extension prediction value into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

2. The system according to claim 1, wherein: The stress distribution data and crack location information are input into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic loading, including: Obtaining a three-dimensional grid node stress value corresponding to the stress distribution data and an expansion direction vector corresponding to the crack position information; The node stress value and the expansion direction vector are input into the trained three-dimensional stress field model, and the dynamic stress amplitude at the local grid node is calculated by combining the frequency and amplitude range in the high-frequency cyclic load parameters; Generate a stress fluctuation index associated with the crack propagation path based on the dynamic stress amplitude and the number of load cycles using a preset fatigue accumulation factor calculation formula; Feature fusion is performed based on the stress fluctuation index to obtain dynamic characteristic parameters of the stress concentration area under high-frequency cyclic load; the dynamic characteristic parameters include the fatigue damage degree of the stress concentration area, the stress cycle change rate and the fluctuation range of the stress concentration coefficient of the key part.

3. The system according to claim 2, characterized in that The method of generating a stress fluctuation index associated with a crack propagation path based on the dynamic stress amplitude and the number of load cycles using a preset fatigue accumulation factor calculation formula includes: The stress fluctuation index is calculated using the following formula: Where Δσ represents the stress fluctuation index, K f represents the fatigue cumulative damage factor, σ i represents the stress amplitude of the i-th cycle, N i Indicates the corresponding number of load cycles, n indicates the total number of stress cycles; n i Indicates the actual number of cycles, N i represents fatigue life, γ i represents the fatigue damage weight coefficient.

4. The system according to claim 1, wherein: The mapping relationship between stress level and crack rate is combined with the microstructure evolution law to analyze and generate the crack extension prediction value, including: Obtain dynamic feature parameters and compare them with preset thresholds to obtain feature judgment results; Analyzing the mapping relationship between stress level and crack rate according to the characteristic judgment result to obtain the current stress state parameter; Extracting the stress state parameters using an algorithm to obtain associated microstructure evolution sequence data; Constructing a crack growth trend map based on the evolution sequence data and generating a path parameter set using a time series prediction algorithm; A crack propagation path coordinate matrix is constructed according to the path parameter set, and a crack propagation prediction value is obtained by calculation using a formula.

5. The system according to claim 1, wherein: The method of constructing a crack propagation path coordinate matrix according to the path parameter set and calculating the crack propagation prediction value using a formula includes: Mapping the crack length, crack width, and crack depth data in the path parameter set into a coordinate matrix using a preset coordinate set conversion rule to obtain a crack propagation path coordinate matrix; Calculating the correlation between crack velocity, crack angle, and crack force based on the crack propagation path coordinate matrix to obtain the crack propagation trend; If the crack velocity is higher than a preset threshold, a linear regression algorithm is used to fit the crack length, crack width, and crack depth in the crack propagation path coordinate matrix to obtain a crack propagation prediction value; If the crack angle exceeds a preset range, a support vector machine algorithm is used to analyze the crack velocity and crack force in the crack propagation path coordinate matrix to correct the crack propagation prediction value; If the crack force distribution is uneven, a decision tree algorithm is used to classify the crack angle and crack depth in the crack propagation path coordinate matrix to adjust the crack propagation prediction value.

6. The system according to claim 5, characterized in that The calculation of the correlation between the crack velocity, crack angle, and crack force based on the crack propagation path coordinate matrix to obtain the crack propagation trend includes: The correlation between crack velocity, crack angle, and crack force is calculated using the following formula: Where v represents the crack growth rate, K1 represents the first mode stress intensity factor, r represents the distance from the crack tip, θ represents the crack growth angle, and θ c represents the critical crack growth angle, K I and K II They represent the first and second mode stress intensity factors, F represents the driving force required for crack extension, γ represents the surface energy density, L represents the crack length, Represents the slope of the crack path.

7. The system according to claim 1, wherein: Inputting the crack extension prediction value into a trained health status assessment model to obtain a life assessment value of the stamping die includes: Extracting time-frequency features of the crack extension prediction value to obtain a multi-dimensional feature sequence; Inputting the multidimensional feature sequence into a convolution kernel group to perform spatial correlation calculation to obtain a crack evolution trend tensor; Performing an attenuation function transformation on the crack evolution trend tensor to obtain a probability density distribution of the remaining life; Assigning a dynamic weight coefficient according to the probability density distribution to calculate and obtain the remaining life distribution value; Inputting the remaining life distribution value into the trained health status assessment model and calculating using a formula to obtain a life assessment value of the stamping die; The lifespan estimate is calculated using the following formula: Among them, E represents the life evaluation parameter, ω j Represents the weight coefficient of the jth health state, R j represents the remaining life value of the jth state, λ represents the decay coefficient, t represents the usage time, and n represents the number of healthy states.

8. A dynamic assessment method for stamping die life based on multimodal sensing, characterized in that: The method comprises: Acquire stress distribution data on the mold surface and interior from a multimodal sensor network; and obtain crack location information from acoustic emission signals; input the stress distribution data and crack location information into a trained three-dimensional stress field model to obtain dynamic characteristic parameters of stress concentration areas under high-frequency cyclic loading; The dynamic characteristic parameter is judged based on a preset threshold value. If the dynamic characteristic parameter exceeds the preset threshold value, a crack propagation prediction value is generated based on a mapping relationship between stress level and crack rate combined with a microstructure evolution law; The crack extension prediction value is input into the trained health status assessment model to obtain the life assessment value of the stamping die; the life assessment value represents the estimated remaining effective time of the stamping die under the existing working conditions and operating conditions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the system according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the system according to any one of claims 1 to 7 are implemented.

Images