Real-time updating method for discrete particle interfacial viscoelastic parameters based on genetic algorithm

Abstract

The application discloses a real-time updating method for discrete particle interface viscoelasticity parameters based on a genetic algorithm, and comprises the following steps: a material stress-strain model is established based on constitutive parameters in discrete element software, and a simulation stress-strain curve is obtained; an experimental stress-strain curve is obtained as benchmark data; the constitutive parameters are updated in real time based on fitting errors of the simulation stress-strain curve and the experimental stress-strain curve, until the fitting result meets preset conditions, and the current simulation stress-strain curve is output as a final simulation prediction curve. The application realizes automation and intelligentization of the parameter calibration process, and greatly improves the precision of the discrete element simulation.

Description

Technical Field

[0001] This invention relates to the field of computational materials mechanics and numerical simulation optimization, and in particular to a method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithms. Background Technology

[0002] The discrete element method (DEM) is a powerful numerical tool for analyzing the mechanical behavior of particle assemblies. It is widely used in simulating the stress and strain of viscoelastic materials. The accuracy of the DEM simulation fundamentally depends on whether the viscoelastic parameters in the constitutive model can truly reflect the intrinsic physical properties of the material.

[0003] Currently, the following process is commonly used when simulating viscoelastic behavior using the discrete element method (DEM): at the start of the simulation, a set of fixed viscoelastic parameters is set based on indoor experimental results, and the entire simulation process is run under these parameters to predict the material's mechanical response. This method is essentially a one-time, unidirectional prediction. However, the mechanical behavior of viscoelastic materials exhibits significant time-varying and path-dependent characteristics; its response within a loading cycle is not simply determined by the initial viscoelastic parameters. Therefore, relying on a fixed set of initial parameters for full-process prediction often fails to accurately capture the complex mechanical behavior exhibited by the material in the middle and later stages, leading to significant deviations between the simulation curve and the actual situation at critical stages. These deviations directly affect the reliability of engineering performance predictions based on simulation results.

[0004] In practical engineering, we can often obtain the true mechanical response data of materials at certain specific moments through experimental measurements. However, traditional fixed-parameter simulation methods cannot use these measured data to correct the simulation itself, which has certain inherent defects and leads to a reduction in simulation accuracy.

[0005] The existing invention patent CN120562223B, "A Method and System for Simulating and Predicting Inrush Water Applicable to Different Geological Media," while capable of extracting the medium structure and constructing an equivalent hydraulic network based on a discrete element model to calculate pressure and flow distribution, and then transmitting the hydraulic response back to the particle system through mapping to achieve water-solid coupling feedback updates, primarily focuses on the forward coupling prediction of "fluid-particle motion." This presents a barrier that makes it difficult to directly apply to real-time updates of material interface constitutive parameters. Specifically, there are barriers related to inconsistencies between the coupling quantity and the updating object, and inconsistencies in the closed-loop convergence target, creating a stability barrier. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention proposes a real-time update method for viscoelastic parameters of discrete particle interfaces based on genetic algorithms. This method drives parameter optimization through genetic algorithms and interacts with the discrete element simulation process in real time, achieving high-precision automatic parameter inversion.

[0007] The specific technical solution is as follows: A method for real-time updating of viscoelastic parameters at discrete particle interfaces based on a genetic algorithm includes: S1: Establish a material stress-strain model based on constitutive parameters in discrete element software and obtain simulated stress-strain curves; S2: Obtain the experimental stress-strain curve as reference data; S3: Based on the fitting error between the simulated stress-strain curve and the experimental stress-strain curve, determine whether the fitting result meets the preset conditions. If not, use a genetic algorithm to update the constitutive parameters in real time and repeat S1-S3; if yes, output the current simulated stress-strain curve as the final simulated prediction curve.

[0008] Furthermore, in step S3, a genetic algorithm is used to update the constitutive parameters in real time, specifically through the following sub-steps: (3.1) Calculate the fitting error based on the simulated stress-strain curve and the experimental stress-strain curve, and use the fitting error as the evaluation basis for the fitness function of the genetic algorithm; (3.2) Generate new constitutive parameter combinations through selection, crossover, and mutation operations; (3.3) Update the discrete element software with the new constitutive parameter combination and return to execute S1.

[0009] Furthermore, step (3.2) is specifically implemented through the following operations: (3.2.1) Selection: The roulette wheel selection method is adopted, and the selection operation is performed based on the fitting error to increase the probability that individuals with higher fitness will be selected as parents to enter the next generation of the population; (3.2.2) Crossover: Perform single-point crossover operation on the parent individuals according to the preset crossover probability to explore new constitutive parameter combinations; (3.2.3) Mutation: Randomly perturb some genes of offspring individuals to form mutated individuals and generate new constitutive parameter combinations; (3.2.4) Iteration: Repeat steps (3.3.1)-(3.3.3) to generate a new generation of population until the fitting error requirement is met.

[0010] Furthermore, the preset condition is that the fitting error is less than a preset threshold.

[0011] Furthermore, the constitutive model of the material stress-strain model is a viscoelastic model, and the constitutive parameters are viscoelastic parameters.

[0012] Furthermore, the viscoelastic model is the Burgers model, used to describe viscoelastic mechanical behavior.

[0013] Furthermore, the discrete element software establishes a connection with an external parameter optimization program through an application programming interface.

[0014] A real-time update system for the viscoelastic parameters of discrete particle interfaces based on genetic algorithms, used in the aforementioned real-time update method for the viscoelastic parameters of discrete particle interfaces based on genetic algorithms, includes: a discrete element model module, an interface module, a data acquisition module, an optimization module, and an update control module. The discrete element model module is used to establish a material stress-strain model in the discrete element software and set the viscoelastic constitutive parameters. The interface module is used to connect the discrete element model module and the optimization module through a programming interface; The data acquisition module is used to acquire experimental stress-strain curves; The optimization module adjusts the viscoelastic constitutive parameters in real time based on a genetic algorithm to minimize the fitting error between the simulation and experimental curves. The update control module is used to output the final simulated prediction curve when the fitting accuracy meets the requirements, and to call the optimization module to update the viscoelastic constitutive parameters when the requirements are not met.

[0015] Furthermore, the optimization module includes: a fitness calculation unit, a parameter generation unit, and an iteration control unit; The fitness calculation unit is used to calculate the fitting error; The parameter generation unit generates new parameters through a genetic algorithm. The iterative control unit is used to manage the iterative process of the simulated genetic algorithm.

[0016] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the aforementioned method for real-time updating of viscoelastic parameters of discrete particle interfaces based on a genetic algorithm.

[0017] The beneficial effects of this invention are: (1) By coupling the genetic algorithm with the discrete element simulation process online, the present invention realizes the real-time and automatic optimization of constitutive parameters, which greatly improves the efficiency of parameter calibration and the accuracy of simulation results, and provides a reliable method for predicting material properties.

[0018] (2) The method of this invention breaks the traditional one-way prediction mode and has the ability to update the constitutive parameters in the simulation model in real time at any time set by the user. By introducing the actual measured values ​​at a specific time as feedback signals into the model, the model is driven to dynamically adjust its internal parameters, so that the subsequent simulation curves can continuously approach the actual response of the material. This simulation-feedback-update closed-loop optimization mechanism can significantly improve the overall accuracy of the simulation curves throughout the process, making them more in line with the high-standard engineering prediction requirements. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithms in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the vertical permanent deformation simulation results of asphalt mixture under fixed parameters in an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram showing the result of parameter updating for simulating the vertical permanent deformation of asphalt mixture using the method of the present invention in an embodiment of the present invention.

[0022] Figure 4 This is a schematic diagram of a real-time update system for the viscoelastic parameters of discrete particle interfaces based on a genetic algorithm, as described in an embodiment of the present invention. Detailed Implementation

[0023] The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. The objectives and effects of the present invention will become clearer as a result. The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0024] A method for real-time updating of viscoelastic parameters at discrete particle interfaces based on a genetic algorithm is presented. The core of this method lies in constructing a closed-loop system of "simulation-feedback-update," bridging discrete element simulation and experimental data through a genetic algorithm to achieve real-time and automatic updating of viscoelastic parameters. Figure 1 As shown, the method includes the following steps: S1. Model Establishment and Parameter Initialization: In the discrete element method software, a material stress-strain model is established based on constitutive parameters to obtain simulated stress-strain curves.

[0025] Preferably, the constitutive model of the material stress-strain model is a viscoelastic model, and the constitutive parameters are viscoelastic parameters. In this embodiment, the viscoelastic model is the Burgers model, used to describe viscoelastic mechanical behavior.

[0026] As an embodiment of the present invention, the vertical permanent deformation behavior of asphalt mixtures is taken as the research object. In discrete element method (DEM) software, a model characterizing the vertical permanent deformation behavior of asphalt mixtures is established based on viscoelastic parameters and a discrete particle system. Specifically, for the viscoelastic interaction between particles, the Burgers model is selected and set as the constitutive model. The Burgers model parameters are initialized, including: Maxwell model elastic modulus E1, Maxwell model viscosity η1, Kelvin model elastic modulus E2, and Kelvin model viscosity η2. These initial parameters are calibrated based on indoor tests.

[0027] S2. System Connection and Data Acquisition: Acquire experimental stress-strain curves as baseline data.

[0028] Furthermore, a connection is established with an external parameter optimization program (implemented in Python code, used to dynamically modify viscoelastic parameters and call the Discrete Element Software to perform simulations) through the Python application programming interface (API) provided by the Discrete Element Software. The obtained experimental stress-strain curves are input into the external optimization program for processing. This API allows the Python script to read the simulation results (deformation data) in real time and correct the constitutive parameters in the model (viscoelastic parameters in this embodiment). The method of this invention achieves distributed computing by interacting with the Discrete Element Software through an external server.

[0029] S3: Based on the fitting error between the simulated stress-strain curve and the experimental stress-strain curve, determine whether the fitting result (or fitting accuracy) meets the preset condition (fitting error is less than the preset threshold). If not, update the constitutive parameters in real time and repeat S1-S3; if yes, output the current simulated stress-strain curve as the final simulated prediction curve.

[0030] Furthermore, the constitutive parameters are updated using a genetic algorithm, specifically through the following sub-steps: (3.1) The fitting error is used as the evaluation criterion for the fitness function of the genetic algorithm. The smaller the fitting error, the better the fit and the higher the fitness. The expression for calculating the fitting error is as follows: In the formula, Error represents the fitting error, E e E represents the experimental reference value corresponding to the experimental stress-strain curve. s This represents the simulated value corresponding to the simulated stress-strain curve.

[0031] (3.2) New constitutive parameter combinations are generated through selection, crossover, and mutation operations. The specific operations are as follows: (3.2.1) Selection: The roulette wheel selection method is adopted, and the selection operation is performed based on the fitting error, so that individuals with higher fitness are selected as parents and enter the next generation of the population with a higher probability.

[0032] (3.2.2) Crossover: Perform a single-point crossover operation on the parent individuals according to the preset crossover probability to explore new parameter combinations.

[0033] (3.2.3) Mutation: With a small probability, some genes of offspring individuals are randomly perturbed to form mutated individuals, avoiding getting trapped in local optima and generating new constitutive parameter combinations.

[0034] (3.2.4) Iteration: Repeat steps (3.3.1)-(3.3.3) to generate a new generation of population until the fitting error requirement is met.

[0035] (3.3) Update the discrete element software with the new constitutive parameter combination and return to execute S1.

[0036] Exemplary effect: such as Figure 2 As shown, in a simulation of the vertical permanent deformation behavior of an asphalt mixture, the simulated curve and the experimental curve under the initial parameters deviated significantly in the later stages, with the maximum fitting error reaching 0.15. This invention addresses this by setting a real-time update mechanism when the fitting error does not reach a preset threshold of 0.02, and the genetic algorithm automatically adjusts the viscoelastic parameters to more reasonable values; ultimately, as shown... Figure 3 As shown, the simulation curves for the entire process are in high agreement with the experimental curves, significantly improving the reliability of the predictions.

[0037] To implement the aforementioned method for real-time updating of viscoelastic parameters at discrete particle interfaces based on genetic algorithms, this embodiment also proposes a system for real-time updating of viscoelastic parameters at discrete particle interfaces based on genetic algorithms, such as... Figure 4 As shown, the system includes: a discrete element model module, an interface module, a data acquisition module, an optimization module, and an update control module.

[0038] The Discrete Element Model module is used to create geometric models, define particle contacts, set material stress-strain models, and set constitutive parameters in Discrete Element software.

[0039] The interface module connects the discrete element model module and the optimization module via a programming interface. It encapsulates all communication details with the discrete element software's Python API, automating parameter writing, simulation control, and data reading.

[0040] The data acquisition module is used to obtain experimental stress-strain curves from external files or databases. In this embodiment, the acquired experimental stress-strain curves are experimentally measured vertical permanent deformation data of asphalt mixtures.

[0041] The optimization module, based on a genetic algorithm, adjusts constitutive parameters in real time to minimize the fitting error between simulated and experimental stress-strain curves. The optimization module includes: a fitness calculation unit, a parameter generation unit, and an iteration control unit. The fitness calculation unit calculates the fitting error in real time; the parameter generation unit generates new combinations of constitutive parameters through genetic algorithm operations; and the iteration control unit manages the iteration process of the genetic algorithm itself.

[0042] The update control module, as the overall system controller, is used to output the final simulated prediction curve when the fitting accuracy meets the preset conditions, and to call the optimization module to update the constitutive parameters when the conditions are not met.

[0043] This invention also provides a computer-readable storage medium storing a program that, when executed by a processor (such as a server CPU), implements the method for real-time updating of viscoelastic parameters of discrete particle interfaces based on a genetic algorithm as described in the above embodiments. Specifically, the processor calls the computational kernel of the discrete element method software and coordinates the execution of the genetic algorithm. In a distributed computing environment, simulation tasks of different individuals can be distributed to multiple computing nodes for parallel processing, greatly improving optimization efficiency.

[0044] The computer-readable storage medium can be an internal storage unit of any data-processing device in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of any data-processing device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., mounted on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store computer programs and other programs and data required by any data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0045] This invention uses experimental stress-strain curves as benchmark data. It calculates the fitting error between the simulated and experimental stress-strain curves as the fitness evaluation criterion and employs a genetic algorithm to perform a global search for viscoelastic constitutive parameters, obtaining parameter combinations with smaller errors. Subsequently, the constitutive parameter combinations are automatically written back to the model via the discrete element method software's programming interface, triggering iterative simulations until the fitting error meets preset conditions. This mechanism completes the inversion channel of "observation bias - constitutive parameter update," thereby resolving the problem of inconsistency between the coupling quantity and the updating object.

[0046] Meanwhile, this invention uses fitting error as the core criterion for update triggering and termination, which further improves the closed-loop convergence target from coupled stability to the consistency between simulated response and experiment. While ensuring that coupled calculation is feasible, it realizes continuous calibration and real-time updating of constitutive parameters, thereby avoiding the problem of coupled stability but uncalibrated parameters and improving the reliability and accuracy of simulation prediction.

[0047] In summary, this invention provides a systematic solution to key technical barriers such as interface misalignment, insufficient real-time performance, and inconsistent convergence targets by establishing an error-driven genetic algorithm inversion mechanism, automating interface updates, and a closed-loop criterion system based on fitting error. This enables real-time updating and high-precision fitting of viscoelastic parameters of discrete particle interface materials.

[0048] It will be understood by those skilled in the art that the above descriptions are merely preferred examples of the invention and are not intended to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the invention should be included within the scope of protection of the invention.

Claims

1. A method for real-time updating of viscoelastic parameters at discrete particle interfaces based on a genetic algorithm, characterized in that, include: S1: Establish a material stress-strain model based on constitutive parameters in discrete element software and obtain simulated stress-strain curves; S2: Obtain the experimental stress-strain curve as reference data; S3: Based on the fitting error between the simulated stress-strain curve and the experimental stress-strain curve, determine whether the fitting result meets the preset conditions. If not, use a genetic algorithm to update the constitutive parameters in real time and repeat S1-S3; if yes, output the current simulated stress-strain curve as the final simulated prediction curve.

2. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithm according to claim 1, characterized in that, In step S3, a genetic algorithm is used to update the constitutive parameters in real time, which is implemented through the following sub-steps: (3.1) Calculate the fitting error based on the simulated stress-strain curve and the experimental stress-strain curve, and use the fitting error as the evaluation basis for the fitness function of the genetic algorithm; (3.2) Generate new constitutive parameter combinations through selection, crossover, and mutation operations; (3.3) Update the discrete element software with the new constitutive parameter combination and return to execute S1.

3. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithms according to claim 2, characterized in that, Step (3.2) is specifically implemented through the following operations: (3.2.1) Selection: The roulette wheel selection method is adopted, and the selection operation is performed based on the fitting error to increase the probability that individuals with higher fitness will be selected as parents to enter the next generation of the population; (3.2.2) Crossover: Perform single-point crossover operation on the parent individuals according to the preset crossover probability to explore new constitutive parameter combinations; (3.2.3) Mutation: Randomly perturb some genes of offspring individuals to form mutated individuals and generate new constitutive parameter combinations; (3.2.4) Iteration: Repeat steps (3.3.1)-(3.3.3) to generate a new generation of population until the fitting error requirement is met.

4. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithm according to claim 1, characterized in that, The preset condition is that the fitting error is less than a preset threshold.

5. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithm according to claim 1, characterized in that, The constitutive model of the material stress-strain model is a viscoelastic model, and the constitutive parameters are viscoelastic parameters.

6. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithms according to claim 5, characterized in that, The viscoelastic model is the Burgers model, which is used to describe viscoelastic mechanical behavior.

7. The method for real-time updating of viscoelastic parameters of discrete particle interfaces based on genetic algorithm according to claim 1, characterized in that, The discrete element software establishes a connection with an external parameter optimization program through an application programming interface.

8. A real-time update system for viscoelastic parameters of discrete particle interfaces based on genetic algorithms, used to implement the real-time update method for viscoelastic parameters of discrete particle interfaces based on genetic algorithms as described in any one of claims 1-7, characterized in that, include: Discrete element model module, interface module, data acquisition module, optimization module, and update control module; The discrete element model module is used to establish a material stress-strain model in the discrete element software and set the viscoelastic constitutive parameters. The interface module is used to connect the discrete element model module and the optimization module through a programming interface; The data acquisition module is used to acquire experimental stress-strain curves; The optimization module adjusts the viscoelastic constitutive parameters in real time based on a genetic algorithm to minimize the fitting error between the simulation and experimental curves. The update control module is used to output the final simulated prediction curve when the fitting accuracy meets the requirements, and to call the optimization module to update the viscoelastic constitutive parameters when the requirements are not met.

9. The real-time update system for viscoelastic parameters of discrete particle interfaces based on genetic algorithms according to claim 8, characterized in that, The optimization module includes: a fitness calculation unit, a parameter generation unit, and an iteration control unit; The fitness calculation unit is used to calculate the fitting error; The parameter generation unit generates new parameters through a genetic algorithm. The iterative control unit is used to manage the iterative process of the simulated genetic algorithm.

10. A computer-readable storage medium, characterized in that, It stores a program that, when executed by a processor, implements the method for real-time updating of viscoelastic parameters of discrete particle interfaces based on a genetic algorithm, as described in any one of claims 1-7.

Images