Method, device and equipment for evaluating service life of viscous-elastic material and medium

Abstract

The invention discloses a life evaluation method and device for a viscous-elastic material, equipment and a medium, and the method comprises the steps: obtaining aging experiment data of a target viscous-elastic material under a preset aging test, training an initial sample generation model through the aging experiment data, and obtaining a preset sample generation model; inputting preset aging temperature data into the preset sample generation model to obtain target data; constructing an initial aging driving model through a Shakkov kinetic theory and an Andrus fracture mechanics theory; and solving parameters to be solved in the initial aging driving model based on the target data and a preset fitting rule to obtain a target aging driving model, and evaluating the service life of the target viscous-elastic material through the target aging driving model. Sample generation is performed through the preset sample generation model, so that test pieces required by the accelerated aging experiment are reduced, and the test cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of viscoelastic materials, and in particular to a life assessment method, device, equipment and medium for viscoelastic materials. Background Art

[0002] For the power source of viscoelastic materials, its main feature is "long-term storage, one-time use", that is, it is in a storage state for a long time. In practical applications, relevant personnel hope that viscoelastic materials have a long storage period and can accurately estimate this storage period to avoid waste or dangerous consequences caused by replacing too late. If the estimated life is longer than the actual life, the expired viscoelastic material will lead to unsatisfactory power output or even catastrophic explosion, endangering the safety of personnel; if the estimated time is too short, a large number of usable viscoelastic materials will be scrapped and destroyed in advance, causing huge economic losses. Therefore, the life assessment of viscoelastic materials is crucial.

[0003] In the prior art, in order to evaluate the storage life of viscoelastic materials, it is often necessary to establish a nonlinear aging-driven model of the viscoelastic material. When constructing the nonlinear aging-driven model of the viscoelastic material, experimental data (real data) is required to calculate the parameters in the nonlinear aging-driven model. When obtaining the experimental data (real data), due to relevant experimental regulations, for each additional aging temperature point, at least 72 test pieces need to be added, which increases the experimental cost. At the same time, due to the increase in test pieces, the complexity of the experiment also increases, resulting in the high cost and low efficiency of constructing the nonlinear aging-driven model of the viscoelastic material. Summary of the invention

[0004] Based on this, it is necessary to propose a life assessment method, device, equipment and medium for viscoelastic materials in response to the technical problems of low efficiency and complex process in the existing technology when constructing a nonlinear aging driving model of viscoelastic materials.

[0005] In a first aspect, a method for evaluating the life of a viscoelastic material is provided, the method comprising:

[0006] Acquire aging experimental data of a target viscoelastic material under a preset aging test, and train an initial sample generation model using the aging experimental data to obtain a preset sample generation model;

[0007] Inputting the preset aging temperature data into the preset sample generation model to obtain target data;

[0008] The initial aging driving model is constructed by using the Zhukov kinetic theory and Andrews fracture mechanics theory;

[0009] The parameters to be solved in the initial aging-driven model are solved based on the target data and preset fitting rules to obtain a target aging-driven model, and the life of the target viscoelastic material is evaluated by the target aging-driven model.

[0010] Optionally, before the step of training the initial sample generation model by using the aging experiment data to obtain the preset sample generation model, the step further includes:

[0011] Obtaining a nonlinear curve of an aging test temperature of the target viscoelastic material under a preset aging test;

[0012] Based on the nonlinear curve, a preset activation function corresponding to the nonlinear curve is selected from a preset activation function set as a target activation function;

[0013] Constructing an initial neural network including a preset number of hidden layers according to the target activation function;

[0014] A residual block is set between two adjacent hidden layers, an L1 regularization layer is set at the input end of the initial neural network, and an inverse regularization layer is set at the output end of the initial neural network to obtain an initial sample generation model, the output end of the L1 regularization layer is connected to the input end of the initial neural network, and the output end of the initial neural network is connected to the input end of the inverse regularization layer.

[0015] Optionally, the step of training the initial sample generation model through the aging experiment data to obtain a preset sample generation model includes:

[0016] Dividing the aging experiment data into a training set and a validation set, and inputting the training set into the initial sample generation model to obtain a first result corresponding to the training set;

[0017] Inputting the verification set into the initial sample generation model to obtain a second result corresponding to the verification set;

[0018] According to the first result and the second result, an L2 regularized loss function is used to update the parameters of the initial sample generation model through a back propagation algorithm and a quasi-Newton algorithm to obtain initial parameters;

[0019] The initial parameters are repeatedly optimized by a limited memory quasi-Newton algorithm until the L2 loss function converges to obtain target parameters;

[0020] The initial neural network in the initial sample generation model is updated according to the target parameters to obtain the target neural network, and then the preset sample generation model is obtained according to the target neural network.

[0021] Optionally, the step of inputting the preset aging temperature data into the preset sample generation model to obtain target data includes:

[0022] Inputting the preset aging temperature data into the L1 regularization layer in the preset sample generation model to obtain initial data;

[0023] Inputting the initial data into the target neural network in the preset sample generation model to obtain predicted experimental data under preset aging temperature data;

[0024] The predicted experimental data is input into the inverse normalization layer in the preset sample generation model to obtain the target data.

[0025] Optionally, the step of constructing an initial aging driving model by using the Rukov kinetics theory and the Andrews fracture mechanics theory includes:

[0026] The first formula is obtained by expressing the damage of stress to the main chain of the matrix of the target viscoelastic material through the Joukoff kinetic theory;

[0027] The energy parameter in the first formula is expressed by the energy required to break the chemical bonds between molecules in Andrews fracture mechanics theory, thereby obtaining the second formula;

[0028] Constructing a nonlinear aging driving model of the target viscoelastic material based on the second formula and a performance change rate formula of the target viscoelastic material;

[0029] When the actual working condition of the target viscoelastic material is a constant stress working condition, using the nonlinear aging driving model as the initial aging driving model;

[0030] When the actual working condition of the target viscoelastic material is a constant strain working condition, the nonlinear aging driving model is subjected to stress-strain conversion to obtain a converted nonlinear aging driving model, and the converted nonlinear aging driving model is used as the initial aging driving model.

[0031] Optionally, the step of performing stress-strain conversion on the nonlinear aging driving model to obtain a converted nonlinear aging driving model includes:

[0032] Acquire a first functional relationship, where the first functional relationship is used to characterize the functional relationship between the tensile strength of a target viscoelastic material, the crosslinking density of the target viscoelastic material, and the stretching ratio of the target viscoelastic material, and acquire a second functional relationship between the stretching ratio and a constant strain;

[0033] The stress-strain conversion is performed on the nonlinear aging driving model according to the first functional relationship and the second functional relationship to obtain a converted nonlinear aging driving model.

[0034] Optionally, the step of solving the unsolved parameters in the initial aging-driven model based on the target data and a preset fitting rule to obtain a target aging-driven model includes:

[0035] Obtaining an aging rate constant of the target viscoelastic material at different temperatures based on the target data and the preset fitting rule;

[0036] The aging rate constant and the temperature value corresponding to the aging rate constant form a data pair, and the data pair is substituted into the initial aging drive model to obtain an equation group;

[0037] The step of solving the parameters to be solved in the initial aging-driven model is completed by solving the set of equations to obtain parameter values ​​of the parameters to be solved;

[0038] Substituting the parameter values ​​into the initial aging-driven model, a target aging-driven model is obtained.

[0039] In a second aspect, a device for evaluating the life of a viscoelastic material is provided, the device comprising:

[0040] A data acquisition unit, used to acquire aging experimental data of a target viscoelastic material under a preset aging test, and train an initial sample generation model through the aging experimental data to obtain a preset sample generation model;

[0041] A data generating unit, used for inputting preset aging temperature data into the preset sample generating model to obtain target data;

[0042] Model building unit, used to build the initial aging drive model through Zhukov kinetic theory and Andrews fracture mechanics theory;

[0043] A model updating unit is used to solve the unsolved parameters in the initial aging-driven model based on the target data and a preset fitting rule to obtain a target aging-driven model, and to evaluate the life of a target viscoelastic material through the target aging-driven model.

[0044] In a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for evaluating the life of viscoelastic materials when executing the computer program.

[0045] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for evaluating the life of viscoelastic materials are implemented.

[0046] The present application obtains the aging test data of the target viscoelastic material under the preset aging test, and trains the initial sample generation model through the aging test data to obtain the preset sample generation model; inputs the preset aging temperature data into the preset sample generation model to obtain the target data; constructs the initial aging drive model through the Rukov kinetics theory and the Andrews fracture mechanics theory; solves the parameters to be solved in the initial aging drive model based on the target data and the preset fitting rules to obtain the target aging drive model, and evaluates the life of the target viscoelastic material through the target aging drive model. Sample generation is performed by the preset sample generation model, reducing the test pieces required for calculating the parameters to be solved in the initial aging drive model, reducing the test cost, and improving the efficiency of establishing the aging drive model. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more in line with non-brittle materials (viscoelastic materials). BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] in:

[0049] Figure 1 A diagram showing an application environment of a method for evaluating the life of a viscoelastic material in one embodiment;

[0050] Figure 2 A flowchart of a method for evaluating the life of a viscoelastic material in one embodiment;

[0051] Figure 3 is a structural block diagram of a device for evaluating the life of a viscoelastic material in one embodiment;

[0052] Figure 4 is a structural block diagram of a computer device in one embodiment;

[0053] Figure 5 It is a structural block diagram of a computer device in another embodiment. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] The life evaluation method of the viscoelastic material provided by the embodiment of the present invention can be applied in the following aspects: Figure 1 In the application environment, the experimental data measuring device (sensor) 110 communicates with the server 120, the server 120 obtains the aging experimental data of the target viscoelastic material under the preset aging test, and trains the initial sample generation model through the aging experimental data to obtain the preset sample generation model; the preset aging temperature data is input into the preset sample generation model to obtain the target data; the initial aging drive model is constructed through the Rukov kinetics theory and the Andrews fracture mechanics theory; the parameters to be solved in the initial aging drive model are solved based on the target data and the preset fitting rules to obtain the target aging drive model, and the life of the target viscoelastic material is evaluated through the target aging drive model. The sample generation is performed by the preset sample generation model, and the test pieces required for calculating the parameters to be solved in the initial aging drive model are reduced, the test cost is reduced, and the efficiency of establishing the aging drive model is improved. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more suitable for non-brittle materials (viscoelastic materials).

[0056] In some scenarios, after receiving the aging experimental data sent by the experimental data measuring device (sensor) 110, the server 120 stores the aging experimental data in an internal fixed storage space. When it is necessary to build an aging-driven model, the server 120 obtains the aging experimental data of the target viscoelastic material under a preset aging test from the fixed storage space, and trains the initial sample generation model through the aging experimental data to obtain a preset sample generation model; inputs the preset aging temperature data into the preset sample generation model to obtain target data; constructs an initial aging-driven model through the Rukov dynamics theory and the Andrews fracture mechanics theory; solves the unsolved parameters in the initial aging-driven model based on the target data and preset fitting rules to obtain a target aging-driven model, and evaluates the life of the target viscoelastic material through the target aging-driven model.

[0057] The present invention is described in detail below through specific embodiments.

[0058] See also Figure 2 As shown, Figure 2 A schematic flow chart of a method for evaluating the life of a viscoelastic material provided by an embodiment of the present invention includes the following steps:

[0059] S101, obtaining aging experimental data of a target viscoelastic material under a preset aging test, and training an initial sample generation model using the aging experimental data to obtain a preset sample generation model;

[0060] Exemplarily, the preset aging test is an accelerated aging test under constant load or constant strain conditions, for example, under n constant load conditions, that is, n stress / strain levels are selected, and accelerated aging tests are performed at m temperature levels respectively, and p time points are selected to test the mechanical performance parameters of the target viscoelastic material. The initial sample generation model can be understood as a virtual sample generation model, which belongs to the virtual sample generation technology, that is, a prediction model of aging test data at other temperatures can be predicted by learning the characteristics of existing samples (aging test data). The aging test data is the degradation data of aging performance sensitive parameters under n constant load conditions and m temperature levels with aging time.

[0061] S102, inputting preset aging temperature data into the preset sample generation model to obtain target data;

[0062] Exemplarily, the preset aging temperature data is within the test temperature range (T min ,T max ), sampling temperature points at intervals of 1°C, and the obtained temperature points (T min+1 、T min+2 ,…,T max-1 ). The preset sample generation model is a virtual sample generation model that has learned the relationship between the existing samples (aging test data) and the aging temperature and can predict the aging test data at other temperatures. The target data is the degradation data of the aging performance sensitive parameters under the preset aging temperature data with the aging time.

[0063] S103, construct the initial aging driving model through Zhukov dynamics theory and Andrews fracture mechanics theory;

[0064] For example, since the Joukov kinetic theory does not consider the energy loss when the material is just broken, and is mainly applicable to brittle materials, it has great limitations. Therefore, based on the viscoelastic characteristics of polymers (target viscoelastic materials), the Andrews fracture mechanics theory involves the energy required for purely breaking the chemical bonds between molecules when creating a new surface. The energy required for purely breaking the chemical bonds between molecules when creating a new surface is substituted into the Joukov kinetic theory to construct an initial aging driven model that takes into account the energy loss when the material is just broken.

[0065] S104, solving the unsolved parameters in the initial aging-driven model based on the target data and a preset fitting rule to obtain a target aging-driven model, and evaluating the life of the target viscoelastic material through the target aging-driven model.

[0066] Exemplarily, the preset fitting rules are formulated according to the provisions of the "Accelerated Aging Test Method for Composite Viscoelastic Materials", and the parameters to be solved include but are not limited to hysteresis ratio, energy loss factor, chain break apparent activation energy and initial aging rate constant, etc.

[0067] By obtaining the aging test data of the target viscoelastic material under the preset aging test, and training the initial sample generation model through the aging test data, a preset sample generation model is obtained; the preset aging temperature data is input into the preset sample generation model to obtain the target data; the initial aging drive model is constructed through the Rukov kinetics theory and the Andrews fracture mechanics theory; the parameters to be solved in the initial aging drive model are solved based on the target data and the preset fitting rules to obtain the target aging drive model, and the life of the target viscoelastic material is evaluated through the target aging drive model. Sample generation is performed through the preset sample generation model, reducing the test pieces required for calculating the parameters to be solved in the initial aging drive model, reducing the test cost, and improving the efficiency of establishing the aging drive model. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more in line with non-brittle materials (viscoelastic materials).

[0068] In a possible implementation, optionally, before the step of training the initial sample generation model by using the aging experiment data to obtain the preset sample generation model, the method further includes:

[0069] Obtaining a nonlinear curve of an aging test temperature of the target viscoelastic material under a preset aging test;

[0070] Based on the nonlinear curve, a preset activation function corresponding to the nonlinear curve is selected from a preset activation function set as a target activation function;

[0071] Constructing an initial neural network including a preset number of hidden layers according to the target activation function;

[0072] A residual block is set between two adjacent hidden layers, an L1 regularization layer is set at the input end of the initial neural network, and an inverse regularization layer is set at the output end of the initial neural network to obtain an initial sample generation model, the output end of the L1 regularization layer is connected to the input end of the initial neural network, and the output end of the initial neural network is connected to the input end of the inverse regularization layer.

[0073] Exemplarily, a nonlinear curve of the target viscoelastic material at an aging test temperature under a preset aging test is obtained, and the line shape of the nonlinear curve is identified, and the line shape of each preset activation function in the preset activation function set is obtained. The preset activation function may be, for example, a Sigmod (S-type growth curve) function, a Relu (nonlinear activation) function, a Tanh (hyperbolic tangent) function, etc. By comparing the line shape of the nonlinear curve with the line shapes of each preset activation function, a line shape having the highest similarity to the line shape of the nonlinear curve is selected as the target line shape, and then the preset activation function corresponding to the target line shape is selected as the target activation function. In this example, the Tanh (hyperbolic tangent) function is selected as the target activation function.

[0074] For example, in order to improve the performance of artificial neural networks (ANN), increasing the number of neurons in the hidden layer is an effective strategy. Therefore, by introducing multiple hidden layers in the neural network, the simple features learned by the network in the previous layers can be combined by the subsequent layers to learn more complex patterns and features. In this example, a deep neural network including 40 hidden layers is constructed as the initial neural network. 40 layers is a critical point. Adding neural layers after 40 layers will not significantly improve the generalization ability.

[0075] For example, when training the initial neural network, the problem of gradient vanishing or gradient exploding is faced. Therefore, by setting a residual block, the network can learn an identity mapping, which greatly alleviates the problem of gradient vanishing. The network no longer needs to learn a complex function mapping, but learns the difference from the input, thereby simplifying the learning task. The residual block is equivalent to adding the input from the previous hidden layer to the output of the next hidden layer. The residual block is set between two adjacent hidden layers, where a [l] is the output of the lth hidden layer, a [l+1] is the output of the l+1th hidden layer, a [l+2] is the output of the l+2th hidden layer.

[0076] Exemplarily, in order to enhance the generalization ability of the model and reduce the risk of overfitting, an L1 regularization layer is set at the input end of the initial neural network, and an inverse regularization layer is set at the output end of the initial neural network, so that L1 regularization and inverse regularization processing are performed on the data after input and before output, respectively, which can balance the relationship between model fitting and generalization. L1 regularization controls the complexity of the model by penalizing the absolute value of the model parameters, thereby sparseening the model parameters. L1 regularization can help achieve feature selection, making the model more concise, and is suitable for situations where features are sparse. The L1 regularization layer is used to execute the L1 regularization algorithm, and the inverse regularization layer is used to execute the inverse regularization algorithm.

[0077] In a possible implementation manner, the step of training the initial sample generation model through the aging experiment data to obtain the preset sample generation model includes:

[0078] Dividing the aging experiment data into a training set and a validation set, and inputting the training set into the initial sample generation model to obtain a first result corresponding to the training set;

[0079] Inputting the verification set into the initial sample generation model to obtain a second result corresponding to the verification set;

[0080] According to the first result and the second result, the parameters of the initial sample generation model are updated by using an L2 regularization loss function through a back propagation algorithm and a quasi-Newton algorithm to obtain initial parameters;

[0081] The initial parameters are repeatedly optimized by a limited memory quasi-Newton algorithm until the L2 loss function converges to obtain target parameters;

[0082] The initial neural network in the initial sample generation model is updated according to the target parameters to obtain the target neural network, and then the preset sample generation model is obtained according to the target neural network.

[0083] Exemplarily, when training the initial sample generation model, the back propagation algorithm (BP algorithm) and L2 regularization loss function are used to update and optimize the parameters. The optimization strategy uses the L-BFGS algorithm (quasi-Newton algorithm), which avoids the large amount of memory consumption required for large-scale optimization through limited memory approximation matrix. At the same time, the algorithm automatically adjusts the search step size to reduce the need to manually set the learning rate.

[0084] For example, the Adam algorithm is very popular when training deep learning models because it is relatively stable and fast. The Adam algorithm mainly adjusts the learning rate of each parameter by calculating the exponential moving average of the first moment (mean) and second moment (uncentered variance) of the gradient.

[0085] In a possible implementation manner, the step of inputting the preset aging temperature data into the preset sample generation model to obtain target data includes:

[0086] Inputting the preset aging temperature data into the L1 regularization layer in the preset sample generation model to obtain initial data;

[0087] Inputting the initial data into the target neural network in the preset sample generation model to obtain predicted experimental data under preset aging temperature data;

[0088] The predicted experimental data is input into the inverse normalization layer in the preset sample generation model to obtain the target data.

[0089] For example, the application of regularization in neural networks has many advantages. It can effectively avoid the problem of overfitting the model on the training data and improve the generalization ability of the model on unseen data. By constraining the complexity of the model, regularization can limit the parameter size of the model and reduce the variance of the model, thereby improving the stability and generalization ability of the model.

[0090] Exemplarily, the L1 regularization layer and the inverse regularization layer make the preset sample generation model more concise and suitable for the case where features are sparse.

[0091] In a possible implementation, the step of constructing the initial aging driving model by using the Joachim Kinematics Theory and the Andrews Fracture Mechanics Theory includes:

[0092] The first formula is obtained by expressing the damage of stress to the main chain of the matrix of the target viscoelastic material through the Joukoff kinetic theory;

[0093] The energy parameter in the first formula is expressed by the energy required to break the chemical bonds between molecules in Andrews fracture mechanics theory, thereby obtaining the second formula;

[0094] Constructing a nonlinear aging driving model of the target viscoelastic material based on the second formula and a performance change rate formula of the target viscoelastic material;

[0095] When the actual working condition of the target viscoelastic material is a constant stress working condition, using the nonlinear aging driving model as the initial aging driving model;

[0096] When the actual working condition of the target viscoelastic material is a constant strain working condition, the nonlinear aging driving model is subjected to stress-strain conversion to obtain a converted nonlinear aging driving model, and the converted nonlinear aging driving model is used as the initial aging driving model.

[0097] For example, the viscoelastic material is a nonlinear viscoelastic material that will creep when subjected to stress. According to the fracture molecule theory, the damage of the stress to the main chain of the viscoelastic material adhesive matrix can be expressed by a modified first-order Zhurkov relationship:

[0098]

[0099] in, is the chain breaking rate, N tis the number of breakable chains at time t, A is the frequency factor, E is the apparent activation energy of chain breakage, R is the molar gas constant, T is the thermodynamic temperature, β is a constant, σ is the stress of the viscoelastic material, e is an exponential constant, and d is the operation prefix;

[0100] When it is assumed that N bonds must be broken before the remaining intact bonds can no longer bear the load, the rupture time formula (first formula) can be obtained from the above relationship:

[0101]

[0102] Among them, t 0 is the starting time of bond breaking, t f is the moment when the fth bond breaks, and ln() is a logarithmic function.

[0103] Exemplarily, the formula of Andrews fracture mechanics theory is expressed as:

[0104]

[0105] In the formula, E T The energy required to create a new surface; T represents the thermodynamic temperature; is the energy required for purely breaking the chemical bonds between molecules when creating a new surface; Φ is the energy loss factor; p is a factor related to the radius of curvature of the material stress-strain curve; g is the distribution function of the energy density W in the sample; δν is the normalized volume unit; γ is the hysteresis ratio, that is, the fraction of energy lost in a stress cycle; for elastic materials, γ = 0; for fully plastic materials, γ = 1; P is the sum of the entire stress field; PU is the sum near any point in the stress field, c is the system state for creating a new surface, k 1 is the performance change rate, ε 0 is the constant strain, ≡ is the identity symbol.

[0106] For example, [E-RTln(t f / t 0 )] (energy parameter) is the E in Andrews theory T0 (When creating a new surface, it is purely the energy required to break the chemical bonds between molecules), and the work done by the external stress, that is, E T (=βσ) not only breaks the chemical bonds, but also dissipates in irreversible energy loss. Then according to Andrews theory Then we have the second formula:

[0107]

[0108] Where T is the thermodynamic temperature and β is a constant

[0109] The performance change rate k1 =ΔP / t (performance change rate formula), substitute into the above formula, and we get:

[0110]

[0111] Where ΔP is the sum of the changes in the entire stress field, t is the time, and K 0 is the performance change rate at the initial time, K f is the rate of change of properties when the fth bond is broken, T is the thermodynamic temperature, and β is a constant;

[0112] By rearranging the above formula, the nonlinear aging driving model of the target viscoelastic material is obtained:

[0113]

[0114] Where T is the thermodynamic temperature and β is a constant.

[0115] In a possible implementation manner, the step of performing stress-strain conversion on the nonlinear aging driving model to obtain a converted nonlinear aging driving model includes:

[0116] Acquire a first functional relationship, where the first functional relationship is used to characterize the functional relationship between the tensile strength of a target viscoelastic material, the crosslinking density of the target viscoelastic material, and the stretching ratio of the target viscoelastic material, and acquire a second functional relationship between the stretching ratio and a constant strain;

[0117] The stress-strain conversion is performed on the nonlinear aging driving model according to the first functional relationship and the second functional relationship to obtain a converted nonlinear aging driving model.

[0118] For example, when the thermomechanical coupling under constant strain loading is considered, the nonlinear aging driven model can be subjected to stress-strain conversion to establish the relationship between constant strain, temperature and aging rate.

[0119] Specifically, a first functional relationship between the tensile strength (stress) σ of the composite viscoelastic material (the target viscoelastic material), the crosslinking density υ, and the stretching ratio λ is obtained:

[0120] σ=υRT(λ-λ -2 )

[0121] And obtain the second functional relationship between the stretching ratio and the constant strain: λ=1+ε 0 ;

[0122] Substitute the first functional relationship and the second functional relationship into the nonlinear aging driving model to obtain a converted nonlinear aging driving model:

[0123]

[0124] Where T is the thermodynamic temperature and β is a constant.

[0125] In a possible implementation manner, the step of solving the unsolved parameters in the initial aging-driven model based on the target data and a preset fitting rule to obtain a target aging-driven model includes:

[0126] Obtaining an aging rate constant of the target viscoelastic material at different temperatures based on the target data and the preset fitting rule;

[0127] The aging rate constant and the temperature value corresponding to the aging rate constant form a data pair, and the data pair is substituted into the initial aging drive model to obtain an equation group;

[0128] The step of solving the parameters to be solved in the initial aging-driven model is completed by solving the set of equations to obtain parameter values ​​of the parameters to be solved;

[0129] Substituting the parameter values ​​into the initial aging-driven model, a target aging-driven model is obtained.

[0130] For example, in combination with the provisions of GJB 10021-2021 "Accelerated Aging Test Method for Composite Viscoelastic Materials" (the preset fitting rule), a data set of the aging rate constant K under different temperature conditions (K min+1 , K min+2 , …, K max-1 Finally, all aging rate constants K and their corresponding aging temperatures T are sorted out to obtain the data pair (K 1 ,T 1 ),(K 2 ,T 2 ),…,(K n ,T n ), all aging rate constants K and their corresponding aging temperatures T are obtained to obtain the data pair (K 1 ,T 1 ),(K 2 ,T 2 ),…,(K n ,T n ) is substituted into the initial aging-driven model to perform data fitting, that is, to obtain all the unknown coefficients to obtain the target aging-driven model.

[0131] In a second aspect, a device for evaluating the life of a viscoelastic material is provided, the device comprising:

[0132] The data acquisition unit 201 is used to acquire aging experimental data of the target viscoelastic material under a preset aging test, and train the initial sample generation model through the aging experimental data to obtain a preset sample generation model;

[0133] A data generating unit 202, configured to input preset aging temperature data into the preset sample generating model to obtain target data;

[0134] A model building unit 203 is used to build an initial aging driving model through the Zhukov kinetic theory and the Andrews fracture mechanics theory;

[0135] The model updating unit 204 is used to solve the unsolved parameters in the initial aging-driven model based on the target data and the preset fitting rules to obtain a target aging-driven model, and evaluate the life of the target viscoelastic material through the target aging-driven model.

[0136] In a possible implementation manner, the data acquisition unit 201 is used to:

[0137] Obtaining a nonlinear curve of an aging test temperature of the target viscoelastic material under a preset aging test;

[0138] Based on the nonlinear curve, a preset activation function corresponding to the nonlinear curve is selected from a preset activation function set as a target activation function;

[0139] Constructing an initial neural network including a preset number of hidden layers according to the target activation function;

[0140] A residual block is set between two adjacent hidden layers, an L1 regularization layer is set at the input end of the initial neural network, and an inverse regularization layer is set at the output end of the initial neural network to obtain an initial sample generation model, the output end of the L1 regularization layer is connected to the input end of the initial neural network, and the output end of the initial neural network is connected to the input end of the inverse regularization layer.

[0141] In a possible implementation manner, the data acquisition unit 201 is used to:

[0142] Dividing the aging experiment data into a training set and a validation set, and inputting the training set and the validation set into the initial sample generation model to obtain a first result corresponding to the training set;

[0143] Inputting the verification set into the initial sample generation model to obtain a second result corresponding to the verification set;

[0144] According to the first result and the second result, the parameters of the initial sample generation model are updated by using an L2 regularization loss function through a back propagation algorithm and a quasi-Newton algorithm to obtain initial parameters;

[0145] The initial parameters are repeatedly optimized by a limited memory quasi-Newton algorithm until the L2 loss function converges to obtain target parameters;

[0146] The initial neural network in the initial sample generation model is updated according to the target parameters to obtain the target neural network, and then the preset sample generation model is obtained according to the target neural network.

[0147] In a possible implementation manner, the data generating unit 202 is configured to:

[0148] Inputting the preset aging temperature data into the L1 regularization layer in the preset sample generation model to obtain initial data;

[0149] Inputting the initial data into the target neural network in the preset sample generation model to obtain predicted experimental data under preset aging temperature data;

[0150] The predicted experimental data is input into the inverse normalization layer in the preset sample generation model to obtain the target data.

[0151] In a possible implementation, the model building unit 203 is used to:

[0152] The first formula is obtained by expressing the damage of stress to the main chain of the matrix of the target viscoelastic material through the Joukoff kinetic theory;

[0153] The energy parameter in the first formula is expressed by the energy required to break the chemical bonds between molecules in Andrews fracture mechanics theory, thereby obtaining the second formula;

[0154] Constructing a nonlinear aging driving model of the target viscoelastic material based on the second formula and a performance change rate formula of the target viscoelastic material;

[0155] When the actual working condition of the target viscoelastic material is a constant stress working condition, using the nonlinear aging driving model as the initial aging driving model;

[0156] When the actual working condition of the target viscoelastic material is a constant strain working condition, the nonlinear aging driving model is subjected to stress-strain conversion to obtain a converted nonlinear aging driving model, and the converted nonlinear aging driving model is used as the initial aging driving model.

[0157] In a possible implementation, the model building unit 203 is used to:

[0158] Acquire a first functional relationship, where the first functional relationship is used to characterize the functional relationship between the tensile strength of a target viscoelastic material, the crosslinking density of the target viscoelastic material, and the stretching ratio of the target viscoelastic material, and acquire a second functional relationship between the stretching ratio and a constant strain;

[0159] The stress-strain conversion is performed on the nonlinear aging driving model according to the first functional relationship and the second functional relationship to obtain a converted nonlinear aging driving model.

[0160] In a possible implementation, the model updating unit 204 is configured to:

[0161] Obtaining an aging rate constant of the target viscoelastic material at different temperatures based on the target data and the preset fitting rule;

[0162] The aging rate constant and the temperature value corresponding to the aging rate constant form a data pair, and the data pair is substituted into the initial aging drive model to obtain an equation group;

[0163] The step of solving the parameters to be solved in the initial aging-driven model is completed by solving the set of equations to obtain parameter values ​​of the parameters to be solved;

[0164] Substituting the parameter values ​​into the initial aging-driven model, a target aging-driven model is obtained.

[0165] By obtaining the aging test data of the target viscoelastic material under the preset aging test, and training the initial sample generation model through the aging test data, a preset sample generation model is obtained; the preset aging temperature data is input into the preset sample generation model to obtain the target data; the initial aging drive model is constructed through the Rukov kinetics theory and the Andrews fracture mechanics theory; the parameters to be solved in the initial aging drive model are solved based on the target data and the preset fitting rules to obtain the target aging drive model, and the life of the target viscoelastic material is evaluated through the target aging drive model. Sample generation is performed through the preset sample generation model, reducing the test pieces required for calculating the parameters to be solved in the initial aging drive model, reducing the test cost, and improving the efficiency of establishing the aging drive model. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more in line with non-brittle materials (viscoelastic materials).

[0166] In one embodiment, a computer device is provided. The computer device is a type of electronic device. The computer device may be a server. Its internal structure diagram may be as shown in FIG. Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, the functions or steps on the service side of a method for evaluating the life of a viscoelastic material are realized.

[0167] In one embodiment, a computer device is provided. The computer device is a type of electronic device. The computer device may be a client. Its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, the functions or steps on the client side of a method for evaluating the life of a viscoelastic material are realized.

[0168] In one embodiment, an electronic device is proposed, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: obtaining aging experimental data of a target viscoelastic material under a preset aging test, and training an initial sample generation model through the aging experimental data to obtain a preset sample generation model; inputting preset aging temperature data into the preset sample generation model to obtain target data; constructing an initial aging driven model through the Joukhoff dynamics theory and the Andrews fracture mechanics theory; solving the unsolved parameters in the initial aging driven model based on the target data and preset fitting rules to obtain a target aging driven model, and evaluating the life of the target viscoelastic material through the target aging driven model.

[0169] The present application obtains the aging test data of the target viscoelastic material under the preset aging test, and trains the initial sample generation model through the aging test data to obtain the preset sample generation model; inputs the preset aging temperature data into the preset sample generation model to obtain the target data; constructs the initial aging drive model through the Rukov kinetics theory and the Andrews fracture mechanics theory; solves the parameters to be solved in the initial aging drive model based on the target data and the preset fitting rules to obtain the target aging drive model, and evaluates the life of the target viscoelastic material through the target aging drive model. Sample generation is performed by the preset sample generation model, reducing the test pieces required for calculating the parameters to be solved in the initial aging drive model, reducing the test cost, and improving the efficiency of establishing the aging drive model. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more in line with non-brittle materials (viscoelastic materials).

[0170] In one embodiment, a computer-readable storage medium is proposed, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented: obtaining aging experimental data of a target viscoelastic material under a preset aging test, and training an initial sample generation model through the aging experimental data to obtain a preset sample generation model; inputting preset aging temperature data into the preset sample generation model to obtain target data; constructing an initial aging driven model through the Rokov dynamics theory and the Andrews fracture mechanics theory; solving the unsolved parameters in the initial aging driven model based on the target data and preset fitting rules to obtain a target aging driven model, and evaluating the life of the target viscoelastic material through the target aging driven model.

[0171] The present application obtains the aging test data of the target viscoelastic material under the preset aging test, and trains the initial sample generation model through the aging test data to obtain the preset sample generation model; inputs the preset aging temperature data into the preset sample generation model to obtain the target data; constructs the initial aging drive model through the Rukov kinetics theory and the Andrews fracture mechanics theory; solves the parameters to be solved in the initial aging drive model based on the target data and the preset fitting rules to obtain the target aging drive model, and evaluates the life of the target viscoelastic material through the target aging drive model. Sample generation is performed by the preset sample generation model, reducing the test pieces required for calculating the parameters to be solved in the initial aging drive model, reducing the test cost, and improving the efficiency of establishing the aging drive model. At the same time, the Rukov kinetics theory and the Andrews fracture mechanics theory are combined, and the problem that the Rukov kinetics theory does not consider the energy loss when the material is just broken is solved by the Andrews fracture mechanics theory, so that the Rukov kinetics theory can be more in line with non-brittle materials (viscoelastic materials).

[0172] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0173] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0174] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0175] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for evaluating the life of viscoelastic materials, It is characterized in that The method comprises: Acquire aging experimental data of a target viscoelastic material under a preset aging test, and train an initial sample generation model using the aging experimental data to obtain a preset sample generation model; Inputting the preset aging temperature data into the preset sample generation model to obtain target data; The initial aging driving model is constructed by using the Zhukov kinetic theory and Andrews fracture mechanics theory; The parameters to be solved in the initial aging-driven model are solved based on the target data and preset fitting rules to obtain a target aging-driven model, and the life of the target viscoelastic material is evaluated by the target aging-driven model.

2. The method according to claim 1, It is characterized in that Before the step of training the initial sample generation model through the aging experiment data to obtain the preset sample generation model, the method further includes: Obtaining a nonlinear curve of an aging test temperature of the target viscoelastic material under a preset aging test; Based on the nonlinear curve, a preset activation function corresponding to the nonlinear curve is selected from a preset activation function set as a target activation function; Constructing an initial neural network including a preset number of hidden layers according to the target activation function; A residual block is set between two adjacent hidden layers, an L1 regularization layer is set at the input end of the initial neural network, and an inverse regularization layer is set at the output end of the initial neural network to obtain an initial sample generation model, the output end of the L1 regularization layer is connected to the input end of the initial neural network, and the output end of the initial neural network is connected to the input end of the inverse regularization layer.

3. The method according to claim 2, It is characterized in that The step of training the initial sample generation model through the aging experiment data to obtain the preset sample generation model includes: Dividing the aging experiment data into a training set and a validation set, and inputting the training set into the initial sample generation model to obtain a first result corresponding to the training set; Inputting the verification set into the initial sample generation model to obtain a second result corresponding to the verification set; According to the first result and the second result, the parameters of the initial sample generation model are updated by using the L2 loss function through the back propagation algorithm and the quasi-Newton algorithm to obtain initial parameters; The initial parameters are repeatedly optimized by a quasi-Newton algorithm with limited memory until the L2 loss function converges to obtain the target parameters; The initial neural network in the initial sample generation model is updated according to the target parameters to obtain the target neural network, and then the preset sample generation model is obtained according to the target neural network.

4. The method according to claim 3, It is characterized in that The step of inputting the preset aging temperature data into the preset sample generation model to obtain target data includes: Inputting the preset aging temperature data into the L1 regularization layer in the preset sample generation model to obtain initial data; Inputting the initial data into the target neural network in the preset sample generation model to obtain predicted experimental data under preset aging temperature data; The predicted experimental data is input into the inverse normalization layer in the preset sample generation model to obtain the target data.

5. The method according to claim 1, It is characterized in that The steps of constructing the initial aging driving model by using the Zhukov kinetics theory and the Andrews fracture mechanics theory include: The first formula is obtained by expressing the damage of stress to the main chain of the matrix of the target viscoelastic material through the Joukoff kinetic theory; The energy parameter in the first formula is expressed by the energy required to break the chemical bonds between molecules in Andrews fracture mechanics theory, thereby obtaining the second formula; Constructing a nonlinear aging driving model of the target viscoelastic material based on the second formula and a performance change rate formula of the target viscoelastic material; When the actual working condition of the target viscoelastic material is a constant stress working condition, using the nonlinear aging driving model as the initial aging driving model; When the actual working condition of the target viscoelastic material is a constant strain working condition, the nonlinear aging driving model is subjected to stress-strain conversion to obtain a converted nonlinear aging driving model, and the converted nonlinear aging driving model is used as the initial aging driving model.

6. The method according to claim 5, It is characterized in that The step of performing stress-strain conversion on the nonlinear aging driving model to obtain a converted nonlinear aging driving model comprises: Acquire a first functional relationship, where the first functional relationship is used to characterize the functional relationship between the tensile strength of a target viscoelastic material, the crosslinking density of the target viscoelastic material, and the stretching ratio of the target viscoelastic material, and acquire a second functional relationship between the stretching ratio and a constant strain; The stress-strain conversion is performed on the nonlinear aging driving model according to the first functional relationship and the second functional relationship to obtain a converted nonlinear aging driving model.

7. The method according to claim 1, It is characterized in that The step of solving the unsolved parameters in the initial aging-driven model based on the target data and the preset fitting rule to obtain the target aging-driven model includes: Obtaining an aging rate constant of the target viscoelastic material at different temperatures based on the target data and the preset fitting rule; The aging rate constant and the temperature value corresponding to the aging rate constant form a data pair, and the data pair is substituted into the initial aging drive model to obtain an equation group; The step of solving the parameters to be solved in the initial aging-driven model is completed by solving the set of equations to obtain parameter values ​​of the parameters to be solved; Substituting the parameter values ​​into the initial aging-driven model, a target aging-driven model is obtained.

8. A life evaluation device for viscoelastic materials, It is characterized in that The device comprises: A data acquisition unit, used to acquire aging experimental data of a target viscoelastic material under a preset aging test, and train an initial sample generation model through the aging experimental data to obtain a preset sample generation model; A data generating unit, used for inputting preset aging temperature data into the preset sample generating model to obtain target data; Model building unit, used to build the initial aging drive model through Zhukov kinetic theory and Andrews fracture mechanics theory; A model updating unit is used to solve the unsolved parameters in the initial aging-driven model based on the target data and a preset fitting rule to obtain a target aging-driven model, and to evaluate the life of a target viscoelastic material through the target aging-driven model.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program. It is characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images