High polymer material performance prediction method and system based on deep learning

Abstract

The invention provides a high polymer material performance prediction method based on deep learning, and relates to the technical field of material performance prediction.The method comprises the specific steps that variable parameters influencing high polymer material performance are determined, multiple sets of samples are prepared and subjected to a tensile test and an accelerated aging test, and initial and final tensile performance parameters are obtained; and constructing and training an initial performance prediction model and an aging performance prediction model based on a deep learning network model, inputting related parameters into the models to obtain performance parameters, generating an environment severe coefficient through data processing, calculating a performance evaluation index in combination with a tensile performance parameter variation, and comparing the performance evaluation index with a preset threshold value to evaluate the material performance. According to the method, the performance change of the aged material can be accurately predicted, the performance evolution can be reflected, the prediction result is closer to reality, the material performance can be comprehensively evaluated, and an effective reference and a scientific basis are provided for long-term use, selection and application of the material.

Description

Technical Field

[0001] The present invention relates to the technical field of material property prediction, and specifically provides a method and system for predicting the properties of polymer materials based on deep learning. Background Art

[0002] Polymer materials have extremely wide applications in many fields such as modern industry and daily life. From key components in aerospace to plastic products for daily use, the quality and reliability of products are directly affected by the quality of their properties. Traditional research on the properties of polymer materials mainly relies on a large number of experimental tests and analysis methods based on experience and theory. In terms of experimental tests, it takes a lot of time, manpower, and material resources to prepare samples under different formulations and process conditions and conduct various tests on them, such as mechanical properties, thermal properties, chemical stability, etc. And the analysis methods based on experience and theory, although they can guide the design and property prediction of materials to a certain extent, their limitations become more and more obvious in the face of increasingly complex and diverse polymer material systems.

[0003] In the prior art, a method for predicting material properties based on deep learning disclosed in Publication No. CN111651916A includes the following steps: S1, establishing a finite element model; S2, using the finite element model to generate a material property relationship table; S3, establishing a deep learning model; S4, training the deep learning model to obtain a material property prediction model. This method can quickly and accurately complete the prediction of material properties.

[0004] However, there are still the following deficiencies. From the above statements, it can be seen that the prior art does not cover the content related to material aging. However, in actual applications, polymer materials will age due to environmental factors, resulting in property changes. Only predicting the initial properties cannot fully reflect the property evolution of materials during use and cannot provide effective property references for the long-term use of materials; in the prior art, a material property relationship table is generated through a finite element model. This method may be somewhat disconnected from the actual production and use scenarios. Although the finite element model can simulate some material properties, the actual material properties are affected by various complex factors, and the data generated only relying on the finite element model may not accurately reflect the real situation.

[0005] The above information disclosed in the background art section is only used to strengthen the understanding of the background of the present disclosure, so it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide a method and system for predicting the properties of polymer materials based on deep learning to solve the problems raised in the above background art.

[0007] To achieve the above purpose, the present invention provides the following technical solutions:

[0008] A method for predicting the properties of polymer materials based on deep learning, the specific steps include:

[0009] S1. Determine the variable parameters that affect the properties of polymer materials, construct multiple groups of different variable parameter combinations, prepare different polymer material sample groups according to different variable parameter combinations respectively, conduct a tensile test on one sample of the sample group to obtain the initial tensile property parameters of the sample. After an interval of T time periods, then conduct an accelerated aging test on each sample group with different accelerated aging condition data. After the test is completed, conduct a tensile test on another sample of the sample group to obtain the final tensile property parameters of the sample;

[0010] S2. Construct an initial property prediction model. The initial property prediction model is constructed based on a deep learning network model. Use the variable parameter combination of the sample as the input and the corresponding initial tensile property parameters as the labels to train the initial tensile property prediction model;

[0011] S3. Input the variable parameter combination of the polymer material to be predicted into the initial property prediction model to obtain the initial tensile property parameters of the polymer material to be predicted;

[0012] S4. Construct an aging property prediction model. The aging property prediction model is constructed based on a deep learning network model. Use the initial tensile property parameters and the accelerated aging condition data as the input and the corresponding final tensile property parameters as the labels to train the aging property prediction model;

[0013] S5. Input the initial tensile property parameters and the accelerated aging condition data of the polymer material to be predicted into the aging property prediction model to obtain the final tensile property parameters of the polymer material to be predicted;

[0014] S6. Conduct data processing and correlation analysis on the accelerated aging condition data to generate an environmental severity coefficient. Conduct numerical calculations on the change amount of the tensile property parameters of the polymer material to be predicted and the environmental severity coefficient to obtain the performance evaluation index of the polymer material to be predicted. Compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material.

[0015] Furthermore, the tensile property parameters include the breaking strength and elongation rate, the variable parameters include processing temperature, injection pressure, cooling rate, annealing temperature, annealing time data, and the accelerated aging condition data includes environmental temperature, environmental humidity, irradiation dose, salt spray concentration, and ozone concentration data.

[0016] Further, process the accelerated aging condition data, conduct a correlation analysis on the accelerated aging condition data and the severity of the accelerated aging conditions, and generate a first environmental severity coefficient. The formula is as follows:

[0017] HJxs 1 = β1WD + β2RH + β3I + β4SC + β5O

[0018] Where, HJxs 1 is the first environmental severity coefficient of the accelerated aging test. The first environmental severity coefficient is used to characterize the severity of the accelerated aging conditions. WD is temperature, RH is humidity, I is irradiation dose, SC is salt spray concentration, and O is ozone concentration;

[0019] In the formula, β1, β2, β3, β4, and β5 are the weight coefficients of temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration respectively in the first environmental severity coefficient formula. On the basis of β1 + β2 + β3 + β4 + β5 = 1, let 0 < β5 < β4 < β3 < β2 < β1 < 1.

[0020] Further, process different accelerated aging condition data, conduct a correlation analysis on the accelerated aging condition data and the severity of the accelerated aging conditions, and generate a second environmental severity coefficient. The formula is as follows:

[0021] HJxs 2 = γ1WD + γ2RH + γ3I + γ4SC + γ5O + γ6(WD × RH) + γ7(WD × O)

[0022] Where, HJxs 2 is the second environmental severity coefficient of the accelerated aging test. The second environmental severity coefficient is used to characterize the severity of the accelerated aging conditions. WD × RH is the interaction term of temperature and humidity, and WD × O is the interaction term of temperature and ozone concentration;

[0023] In the formula, γ1, γ2, γ3, γ4, γ5, γ6, and γ7 are the weight coefficients of temperature, humidity, irradiation dose, salt spray concentration, ozone concentration, the interaction term of temperature and humidity, and the interaction term of temperature and ozone concentration respectively in the first environmental severity coefficient formula. On the basis of γ1 + γ2 + γ3 + γ4 + γ5 + γ6 + γ7 = 1, let 0 < γ7 < γ6 < γ5 < γ4 < γ3 < γ2 < γ1 < 1.

[0024] Further, perform numerical calculations on the change in the rupture strength of the polymer material to be predicted and the first environmental severity coefficient to obtain the first performance evaluation index of the sample. The formula is as follows:

[0025]

[0026] Among them, XPxs 1 is the first performance evaluation index of the polymer material to be predicted. The first performance evaluation index is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change amount of the rupture strength and the first environmental severity coefficient;

[0027] In the formula, σ1 is the initial rupture strength of the polymer material to be predicted, and σ0 is the final rupture strength of the polymer material to be predicted.

[0028] Furthermore, numerical calculations are performed on the change amount of the elongation rate of the polymer material to be predicted and the corresponding second environmental severity coefficient to obtain the second performance evaluation index of the polymer material to be predicted. The formula is as follows:

[0029]

[0030] Among them, XPxs 2 is the second performance evaluation index of the polymer material to be predicted. The second performance evaluation index is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change amount of the elongation rate and the second environmental severity coefficient;

[0031] In the formula, ε1 is the initial elongation rate of the polymer material to be predicted, and ε0 is the final elongation rate of the polymer material to be predicted.

[0032] Furthermore, the performance evaluation index of the polymer material to be predicted is compared with the corresponding preset threshold value to comprehensively evaluate the performance quality of the polymer material. The specific process is as follows:

[0033] When any one of the performance evaluation indexes is not greater than the threshold value, that is, XPxs 1 ≤yz1 or XPxs 2 ≤yz2, it is considered that the performance of this polymer material is good;

[0034] When all the performance evaluation indexes are greater than the threshold value, that is, XPxs 1 >yz1 and XPxs 2 >yz2, it is considered that the performance of this polymer material is poor;

[0035] Among them, yz1 is the threshold value of the first performance evaluation index, and yz2 is the threshold value of the second performance evaluation index.

[0036] A polymer material performance prediction system based on deep learning. The system is used to execute any one of the above-mentioned polymer material performance prediction methods based on deep learning, including:

[0037] A dataset construction module, which is used to determine the variable parameters affecting the performance of polymer materials, construct multiple groups of different variable parameter combinations, prepare different groups of polymer material samples according to different variable parameter combinations respectively, conduct a tensile test on one sample of each sample group to obtain the initial tensile performance parameters of the sample, and after an interval of T time periods, conduct an accelerated aging test on each sample group respectively with different accelerated aging condition data. After the test ends, conduct a tensile test on another sample of each sample group to obtain the final tensile performance parameters of the sample;

[0038] An initial performance construction module, which is used to construct an initial performance prediction model. The initial performance prediction model is constructed based on a deep learning network model. The variable parameter combination of the sample is used as the input, and the corresponding initial tensile performance parameter is used as the label to train the initial tensile performance prediction model;

[0039] An initial parameter module, which is used to input the variable parameter combination of the polymer material to be predicted into the initial performance prediction model to obtain the initial tensile performance parameters of the polymer material to be predicted;

[0040] An aging performance construction module, which is used to construct an aging performance prediction model. The aging performance prediction model is constructed based on a deep learning network model. The initial tensile performance parameter and the accelerated aging condition data are used as the input, and the corresponding final tensile performance parameter is used as the label to train the aging performance prediction model;

[0041] An aging parameter module, which is used to input the initial tensile performance parameter and the accelerated aging condition data of the polymer material to be predicted into the aging performance prediction model to obtain the final tensile performance parameters of the polymer material to be predicted;

[0042] A data processing and prediction module, which is used to conduct data processing and correlation analysis on the accelerated aging condition data to generate an environmental severity coefficient, conduct numerical calculations on the change amount of the tensile performance parameter of the polymer material to be predicted and the environmental severity coefficient to obtain the performance evaluation index of the polymer material to be predicted, and compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] Aiming at the problem that the performance of polymer materials changes due to environmental factor aging in actual applications, the present invention specifically designs an aging performance prediction model. By conducting an accelerated aging test on samples, obtaining the initial and final tensile performance parameters and using them as training data, it can accurately predict the performance change of the material after aging, completely reflect the performance evolution of the material during use, and provide an effective performance reference for the long-term use of the material.

[0045] Rather than simply relying on a finite element model to generate a table of material property relationships, a group of polymer material samples with different combinations of variable parameters is prepared and directly tested on actual samples. These samples have undergone real accelerated aging conditions, taking into account the influence of various complex factors in actual production and use scenarios on material properties, making the prediction results closer to the real situation and avoiding the problem of being out of touch with reality.

[0046] The concepts of environmental severity coefficient and performance evaluation index are introduced. By processing the data of accelerated aging conditions and performing correlation analysis to generate the environmental severity coefficient, and then combining the change amount of the tensile property parameters of the polymer material to be predicted to calculate the performance evaluation index, and finally comparing it with a preset threshold, the performance advantages and disadvantages of the polymer material can be comprehensively and synthetically evaluated, providing a more scientific basis for the selection and application of materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0048] Figure 2 It is a block diagram of the module composition of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0050] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meaning understood by those of ordinary skill in the field to which the present invention belongs. The "first", "second", and similar terms used in the present invention do not indicate any order, quantity, or importance, but are only used to distinguish different components. The terms "comprising" or "including" and the like mean that the elements or objects appearing before the term cover the elements or objects listed after the term and their equivalents, without excluding other elements or objects. The terms "connected" or "coupled" and the like are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right", etc. are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0051] Example 1:

[0052] Please refer to Figure 1 , the present invention provides a technical solution:

[0053] A method for predicting the properties of polymer materials based on deep learning, the specific steps include:

[0054] S1. Determine the variable parameters that affect the properties of the polymer material, construct multiple groups of different variable parameter combinations, prepare different groups of polymer material samples according to different variable parameter combinations respectively, conduct a tensile test on one sample of each sample group to obtain the initial tensile property parameters of the sample. After an interval of T time periods, then conduct an accelerated aging test on each sample group respectively with different accelerated aging condition data. After the test is completed, conduct a tensile test on another sample of the sample group to obtain the final tensile property parameters;

[0055] Based on the above embodiments, the polymer material is at least one of polyethylene and polypropylene.

[0056] Based on the above embodiments, the tensile property parameters include the breaking strength and the elongation at break, and the variable parameters include the processing temperature, injection pressure, cooling rate, annealing temperature, and annealing time data. The accelerated aging condition data includes the environmental temperature, environmental humidity, irradiation dose, salt spray concentration, and ozone concentration data.

[0057] Based on the above embodiments, the numerical values of the processing temperature, injection pressure, cooling rate, annealing temperature, and annealing time are set as follows:

[0058] Processing temperature: 180 °C, 200 °C, 220 °C;

[0059] Injection pressure: 50 MPa, 80 MPa, 100 MPa;

[0060] Cooling rate: 1 °C / min - 10 °C / min, 10 °C / min - 50 °C / min;

[0061] Annealing temperature: 80 °C, 100 °C;

[0062] Annealing time: 1 h, 2 h.

[0063] Based on the above embodiments, when conducting a tensile test on the sample, it is preferred to use a dumbbell-shaped sample with a gauge length of 25 mm and a thickness of 2 mm, and at the same time keep the test environment at a laboratory temperature of (23 ± 2) °C and a humidity of (50 ± 5) %RH, and the test standard is a tensile test (constant rate of 50 mm / min).

[0064] Based on the above embodiments, the breaking strength and the elongation at break are two main indicators of the tensile properties, which respectively measure the tensile strength and the deformation ability of the material. Together, they provide the ability of the material to withstand the maximum load and the maximum deformation during the tensile process.

[0065] Among them, the breaking strength reflects the anti-fracture ability of the material under stress, and the elongation at break reflects the plasticity of the material, which can withstand large deformations without breaking.

[0066] In the practical applications of polymer materials, not only the strength requirements need to be met, but also the material should be able to deform appropriately under different environments or working conditions. Therefore, the combination of the breaking strength and elongation can comprehensively reflect the adaptability and stability of the material, helping to predict the performance of the material in a specific environment.

[0067] Based on the above embodiments, a tensile test is performed on one sample of the sample group to obtain the initial breaking strength and elongation of the sample. The specific process is as follows:

[0068] The sample is symmetrically clamped in a pneumatic fixture, and a video extensometer is installed to monitor the deformation of the sample in real time. A tensile load is applied at a constant rate (50 mm / min) until the sample breaks. The stress-strain curve, load-displacement data, and real-time images are recorded synchronously. The stress is obtained by dividing the load (force) by the initial cross-sectional area of the sample, and the strain is the ratio of the elongation of the sample to the original length;

[0069] Through the load-displacement curve, the maximum load at the time of sample fracture (i.e., the maximum force applied during the test) is found;

[0070] The initial cross-sectional area of the sample is calculated by measuring the width and thickness of the sample;

[0071] Initial breaking strength = maximum load / initial cross-sectional area of the sample;

[0072] Through the video extensometer, the elongation of the sample at the time of fracture is obtained. Initial elongation = (sample length at fracture - initial length) / initial length × 100%.

[0073] Based on the above embodiments, the process of obtaining the breaking strength and elongation under accelerated aging conditions is the same as the process and calculation formula for obtaining the breaking strength and elongation before accelerated aging.

[0074] Based on the above embodiments, after collecting the breaking strength and elongation, these parameters are respectively subjected to maximum-minimum normalization processing, and then the normalized data is used for the subsequent analysis processing, so that in the subsequent analysis processing, various data can be analyzed under the same dimension, avoiding the problem that some data is ignored due to different dimensions.

[0075] Before and after the accelerated aging test, the measured breaking strength and elongation are collected multiple times (such as 3 groups). The same type of data is averaged, and the finally obtained average value is used as the corresponding data for the breaking strength and elongation. The breaking strength and elongation calculated later are both the data after averaging.

[0076] S2. Construct an initial performance prediction model, which is constructed based on a deep learning network model. Use the variable parameter combinations of the samples as input and the corresponding initial tensile performance parameters as labels to train the initial tensile performance prediction model;

[0077] S3. Input the variable parameter combinations of the polymer material to be predicted into the initial performance prediction model to obtain the initial tensile performance parameters of the polymer material to be predicted;

[0078] S4. Construct an aging performance prediction model, which is constructed based on a deep learning network model. Use the initial tensile performance parameters and accelerated aging condition data as input and the corresponding final tensile performance parameters as labels to train the aging performance prediction model;

[0079] S5. Input the initial tensile performance parameters and accelerated aging condition data of the polymer material to be predicted into the aging performance prediction model to obtain the final tensile performance parameters of the polymer material to be predicted;

[0080] Based on the above embodiments, both the initial performance prediction model and the aging performance prediction model are composed of deep learning networks based on multi-layer perceptrons. The deep neural network of the multi-layer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer all have at least two neurons and all use ReLU (Rectified Linear Unit), that is, the rectified linear unit, as the activation function;

[0081] In the initial performance prediction model, the input features of the deep learning network of the multi-layer perceptron include: processing temperature, injection pressure, cooling rate, annealing temperature, and annealing time data, a total of 5 features.

[0082] The structure of the deep learning network of the multi-layer perceptron is as follows:

[0083] Input layer: Receive the input of 5 features;

[0084] First hidden layer: Has 128 neurons and uses ReLU as the activation function;

[0085] Second hidden layer: Has 64 neurons and also uses the ReLU activation function;

[0086] Third hidden layer: Has 32 neurons and uses the ReLU activation function;

[0087] Output layer: Has 1 neuron, which is the initial tensile performance parameter of the sample.

[0088] In the aging performance prediction model, the input features of the deep learning network of the multi-layer perceptron include: initial fracture strength, initial elongation rate, processing temperature, injection pressure, cooling rate, annealing temperature, and annealing time data, a total of 7 features.

[0089] The structure of the deep learning network of the multi-layer perceptron is as follows:

[0090] Input layer: Receives the input of 7 features;

[0091] First hidden layer: Has 128 neurons and uses ReLU as the activation function;

[0092] Second hidden layer: Has 64 neurons and also uses the ReLU activation function;

[0093] Third hidden layer: Has 32 neurons and uses the ReLU activation function;

[0094] Output layer: Has 1 neuron, which is the final tensile performance parameter of the sample.

[0095] The process of training the initial performance prediction model is as follows:

[0096] Using the variable parameter combinations of the samples as the input quantity and the corresponding initial tensile performance parameters as the labels for training, and using the mean squared error as the loss function. When the mean squared error is within the range of [0, 0.01], the training of the initial performance prediction model is completed.

[0097] The process of training the aging performance prediction model is as follows:

[0098] Using the initial tensile performance parameters of the samples and the accelerated aging condition data as the input quantity and the corresponding final tensile performance parameters as the labels for training, and using the mean squared error as the loss function. When the mean squared error is within the range of [0, 0.01], the training of the aging performance prediction model is completed.

[0099] S6. Process and analyze the correlation of different combinations of accelerated aging conditions to generate an environmental severity coefficient. Perform numerical calculations on the change amount of the tensile performance parameters of the polymer material to be predicted and the environmental severity coefficient to obtain the performance evaluation index of the polymer material to be predicted. Compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material.

[0100] Based on the above embodiments, generally, with the changes in temperature, humidity, irradiation amount, salt fog concentration, and ozone concentration, that is, deviating from the standard range in the natural aging state, which means that the temperature, humidity, irradiation amount, salt fog concentration, and ozone concentration are all higher than the values in the natural aging state, and within the ranges of temperature, humidity, irradiation amount, salt fog concentration, and ozone concentration, as their values increase, the rupture strength of the polymer material will decrease.

[0101] For example, the actual operating temperature of the polymer material is generally between -20°C and 60°C. To accelerate the aging process, the environmental temperature is increased to 80°C or even higher in the accelerated aging test. This can speed up the physical and chemical changes inside the material, such as the breakage and crosslinking of molecular chains, so as to more quickly observe the changes in material properties (such as rupture strength);

[0102] The normal environmental humidity is around 30% - 80%. In the accelerated aging, the humidity is increased to more than 90% or even close to the saturated humidity. The high-humidity environment will accelerate the aging processes such as hydrolysis and swelling of the polymer material, especially having a more obvious impact on the rupture strength of some hydrophilic polymer materials;

[0103] Under natural aging conditions, the irradiation energy such as sunlight received by the polymer material is relatively limited. In the accelerated aging test, high-intensity ultraviolet lamps and other equipment are used to significantly increase the irradiation amount, accelerating the photo-oxidative aging process of the polymer material, and causing the rupture strength of the material to change more quickly;

[0104] In non-coastal and other special areas, the salt fog concentration in the actual environment is extremely low or even negligible. However, in the accelerated aging test, a certain concentration of salt fog environment is configured to simulate the impact of coastal or high-salt environments on the rupture strength of the polymer material;

[0105] The ozone concentration in the atmosphere is usually very low, generally around 0.02 - 0.05 ppm. In the accelerated aging test, the ozone concentration is increased to several hundred ppm. Because ozone has strong oxidizing properties, it can quickly react with the polymer material, accelerating the decrease in the rupture strength of the polymer material.

[0106] Therefore, the temperature, humidity, irradiation amount, salt fog concentration, and ozone concentration are all negatively correlated with the rupture strength.

[0107] Based on the above embodiments, since the first environmental severity coefficient is used to characterize the severity of the accelerated aging conditions, when the first environmental severity coefficient is larger, it indicates that the accelerated aging conditions are more severe, which in turn leads to a decrease in the rupture strength. Therefore, the temperature, humidity, irradiation amount, salt fog concentration, and ozone concentration are positively correlated with the first environmental severity coefficient.

[0108] Based on the above embodiments, the accelerated aging condition data is processed, and a correlation analysis is performed on the accelerated aging condition data and the severity of the accelerated aging conditions to generate a first environmental severity coefficient. The formula is as follows:

[0109] HJxs 1 = β1WD + β2RH + β3I + β4SC + β5O

[0110] Where, HJxs 1 is the first environmental severity coefficient of the accelerated aging test. The first environmental severity coefficient is used to characterize the severity of the accelerated aging conditions. WD is temperature, RH is humidity, I is irradiation dose, SC is salt spray concentration, and O is ozone concentration;

[0111] In the formula, β1, β2, β3, β4, and β5 are the weight coefficients of temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration respectively in the first environmental severity coefficient formula;

[0112] The reasons for using the above functional form to express the functional relationship between temperature, humidity, irradiation dose, salt spray concentration, ozone concentration, and the first environmental severity coefficient are as follows:

[0113] First, the formula uses a linear superposition method. After multiplying temperature (WD), humidity (RH), irradiation dose (I), salt spray concentration (SC), ozone concentration (O) by their corresponding weight coefficients and then adding them up, the first environmental severity coefficient is obtained. This is because in the actual environment, the effects of these factors on the environmental severity do not exist in isolation, but interact and jointly affect. Through superposition, the contributions of various factors to the environmental severity can be comprehensively considered, and the severity of the accelerated aging conditions can be fully reflected.

[0114] Second, different environmental factors have different degrees of influence on material aging or environmental severity. For example, during the aging process of some materials, changes in temperature may have a greater impact on their performance decline, while in other cases, the effects of humidity or irradiation dose may be more critical. The weight coefficients in the formula can be adjusted and determined according to the actual situation and a large amount of experimental data, so as to accurately reflect the relative importance of each factor to the environmental severity. In this way, the roles of various factors in the comprehensive environmental impact can be more accurately quantified.

[0115] Third, from the perspective of data processing and correlation analysis, the linear function form is convenient for statistical analysis and establishing a mathematical model. Methods such as multiple linear regression can be used to determine the weight coefficients based on a large amount of experimental data, so as to establish a quantitative relationship between environmental factors and environmental severity.

[0116] Temperature is often a key factor affecting material aging and environmental changes. In many chemical reactions, an increase in temperature affects the reaction rate. For example, in the aging of polymer materials, high temperature accelerates the thermal motion of molecular chains, leading to increased aging reactions such as molecular chain breakage and cross-linking, and thus causing a decline in material properties. Compared with factors such as relative humidity and irradiation dose, the influence of temperature changes on the microstructure and properties of materials is more direct and intense, so a relatively large weight coefficient is assigned to it.

[0117] The magnitude relationship of the weighting coefficients is set as follows:

[0118] Temperature often plays a dominant role among the many factors affecting the fracture strength of materials. An increase in temperature significantly changes the internal structure and intermolecular forces of materials. For example, for polymer materials, high temperature intensifies the thermal motion of molecular chains, resulting in weakened entanglement between molecular chains, reduced toughness of the material, and thus making it more prone to fracture. In metal materials, high temperature triggers changes in crystal structure, generates thermal stress, accelerates material fatigue and crack propagation, and causes a decrease in fracture strength. Compared with other factors such as humidity and irradiation dose, the influence of temperature changes on the microstructure and mechanical properties of materials is more direct, so the largest weight coefficient β1 is assigned to it in the first environmental severity coefficient formula.

[0119] The influence of humidity on the fracture strength of materials is also relatively crucial, but to a lesser extent than temperature. Humidity mainly affects material properties through moisture absorption. For hydrophilic materials, after absorbing water, it causes volume expansion, generates internal stress, and at the same time, water may also participate in chemical reactions, accelerating material degradation. However, the influence of humidity usually requires a certain time accumulation, and in many cases, the influence of humidity on the fracture strength of materials needs to act synergistically with other factors to be more obvious. For example, in a high-temperature and high-humidity environment, the aging and fracture risk of materials increase, and the influence of a single humidity change on the fracture strength of materials is relatively limited, so β2 is less than β1.

[0120] Irradiation (such as ultraviolet irradiation) causes photochemical reactions in materials and destroys the molecular structure. However, in a general environment, the energy and intensity of irradiation are relatively limited, and materials have a certain resistance to irradiation. Only under long-term and high-intensity irradiation will the fracture strength of materials be significantly affected. Compared with temperature and humidity, the influence of irradiation on the fracture strength of materials is relatively weak in terms of both universality and intensity, so its weight coefficient β3 is less than β2 in the formula.

[0121] Salt spray mainly affects metal materials and acts on metals through electrochemical corrosion. For most non-metallic materials such as plastics and rubbers, the change in salt spray concentration has a minimal impact on their properties, and the affected range is relatively narrow. Irradiation, on the other hand, generally affects various materials. Irradiation such as ultraviolet light can trigger photochemical reactions in most materials, damaging the molecular structure of the materials. For example, plastics exposed to sunlight for a long time will become brittle and fade due to irradiation. Therefore, when reflecting the severity of the environment, the irradiation dose has a higher weight, and β4 is less than β3.

[0122] Salt spray mainly affects the corrosion of metal materials. For most non-metallic materials, the affected range and degree of salt spray concentration are relatively limited. Metals will undergo electrochemical corrosion in a salt spray environment, forming corrosion products, weakening the load-bearing capacity of the materials, and resulting in a decrease in the rupture strength. However, this effect is mainly limited to metal materials. Although ozone has strong oxidizing properties, in the natural environment, its concentration is usually low. Only in specific industrial environments or for materials sensitive to ozone, ozone will have a relatively obvious impact on the rupture strength of the materials. Therefore, the influence of salt spray concentration and ozone concentration on the rupture strength of materials is relatively small, and in some application scenarios of metal materials, the influence of salt spray on the rupture strength is relatively more prominent than that of ozone. So, β5 is less than β4.

[0123] Therefore, on the basis of β1 + β2 + β3 + β4 + β5 = 1, let 0 < β5 < β4 < β3 < β2 < β1 < 1.

[0124] As an implementation manner, the value range of β1 is 0.2 - 0.25, the value range of β2 is 0.15 - 0.2, the value range of β3 is 0.25 - 0.3, the value range of β4 is 0.1 - 0.15, and the value range of β5 is 0.05 - 0.1. The specific values are set by technicians according to the actual situation and are not limited here.

[0125] On the basis of the above embodiments, generally, as the temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration change, that is, deviate from the standard range in the natural aging state, namely the temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration are all higher than the values in the natural aging state, and within the ranges of temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration, as their values increase, the elongation of the polymer material will decrease.

[0126] For example, when the temperature in the natural environment is about 25°C, its elongation is 200%. When the temperature is increased to 80°C for an accelerated aging test for a period of time, the thermal degradation reaction of the polymer material intensifies, the internal structure of the material changes, and it becomes more brittle. When testing its elongation again, it drops to 100%.

[0127] In an environment with a relative humidity of 50%, its elongation is approximately 60%. When the environmental humidity rises to 90% and after a period of aging, nylon molecules undergo hydrolysis, the molecular chains become shorter, the flexibility of the material deteriorates, and the elongation decreases to 30%.

[0128] In a normal indoor environment, under natural light, its elongation can be maintained at approximately 600%. If it is placed in an artificial accelerated aging test chamber with a high irradiation dose to simulate strong ultraviolet irradiation, after a certain period of time, due to photooxidation, the PE molecular chains crosslink, the film becomes hard, and the elongation drops significantly to 200%.

[0129] In an environment with a salt spray concentration of 0.1%, its elongation can be maintained at around 300%. When the salt spray concentration rises to 5% for an accelerated aging test, chloride ions erode the polymer material, causing degradation of the molecular chains, a decrease in the toughness of the material, and the elongation reduces to 150%.

[0130] In a normal atmospheric environment with an ozone concentration of 0.02 ppm, its elongation can reach 700%. If it is placed in an accelerated aging environment with an ozone concentration of 100 ppm, ozone rapidly reacts with the double bonds in the polymer material, the molecular chains break, the rubber becomes hard and brittle, and the elongation decreases to 300%.

[0131] Therefore, temperature, humidity, irradiation dose, salt spray concentration, and ozone concentration are all negatively correlated with elongation.

[0132] Based on the above embodiments, since the second environmental severity coefficient is used to characterize the severity of the accelerated aging conditions, the greater the second environmental severity coefficient, the more severe the accelerated aging conditions, which in turn leads to a decrease in elongation. Therefore, there is a positive correlation between temperature, humidity, irradiation dose, salt spray concentration, ozone concentration, and the second environmental severity coefficient.

[0133] Based on the above embodiments, different combinations of accelerated aging conditions are processed for data and correlation analysis to generate the second environmental severity coefficient, according to the following formula:

[0134] HJxs 2 =γ1WD + γ2RH + γ3I + γ4SC + γ5O + γ6(WD×RH) + γ7(WD×O)

[0135] Where, HJxs 2 is the second environmental severity coefficient, which is used to characterize the severity of the accelerated aging conditions. WD×RH is the interaction term of temperature and humidity, and WD×O is the interaction term of temperature and ozone concentration.

[0136] where γ1, γ2, γ3, γ4, γ5, γ6, and γ7 are the weight coefficients of temperature, humidity, irradiation dose, salt fog concentration, ozone concentration, the interaction term of temperature and humidity, and the interaction term of temperature and ozone concentration in the first environmental severity coefficient formula, respectively;

[0137] The reasons for using the above functional form to express the functional relationship between temperature, humidity, irradiation dose, salt fog concentration, ozone concentration, and the second environmental severity coefficient are as follows:

[0138] First, in the process of material accelerated aging, the synergistic effect of temperature and humidity cannot be underestimated. When the temperature rises, the molecular activity inside the material increases. At this time, if the humidity also increases, water is more likely to penetrate into the material, accelerating hydrolysis, swelling and other reactions of the material, and then having a greater negative impact on the elongation of the material. For example, in a high-temperature and high-humidity environment, the molecular chains of some polymer materials will expand due to water absorption, and the intermolecular force weakens, resulting in a decrease in elongation. Introducing the interaction term (WD×RH) into the formula can effectively quantify the impact of the combined action of temperature and humidity on the environmental severity, making the second environmental severity coefficient more accurately reflect the actual situation.

[0139] Second, ozone has strong oxidizing properties. In a high-temperature environment, its oxidation activity will be further enhanced. High temperature promotes the molecular movement of the material to intensify, and ozone is more likely to chemically react with the material molecules, destroying the molecular structure of the material, thereby reducing the elongation of the material. For example, in a high-temperature environment, the aging effect of ozone on rubber materials is more obvious, accelerating the hardening and embrittlement of rubber, resulting in a decrease in its elongation. Through the interaction term (WD×O), the contribution of the interaction between temperature and ozone concentration to the environmental severity can be fully considered, making the coefficient calculation more comprehensive.

[0140] Third, from the perspective of data processing, the performance change data of materials under different accelerated aging conditions are complex and diverse. By constructing a multiple linear function that includes various environmental factors and their interaction terms, these experimental data can be better fitted. When performing multiple linear regression analysis, the weight coefficients from γ1 to γ7 are determined according to a large amount of experimental data, and an accurate mathematical relationship between environmental factors and the second environmental severity coefficient can be established, improving the accuracy of predicting the degree of material aging.

[0141] Fourth, this functional form can not only be used to calculate the second environmental severity coefficient, but also provides convenience for in-depth study of the synergistic mechanism between environmental factors. By analyzing the weight coefficients, the relative importance of different environmental factors and their interactions on material aging can be understood, and the action laws of each factor in the process of accelerating material aging can be clarified, providing a more valuable theoretical basis for material aging research and the formulation of protection measures.

[0142] The magnitude relationship of the weighting coefficients is set as follows:

[0143] The influence of temperature on material aging is universal and fundamental. In most chemical reactions, an increase in temperature affects the reaction rate and directly impacts the molecular structure and properties of materials. For example, high temperature can accelerate the thermal degradation of polymer materials, causing molecular chains to break, thereby deteriorating the material properties. Compared with factors such as humidity and irradiation dose, the influence of temperature change is more direct and intense, so γ1 is the largest.

[0144] Humidity also plays an important role in material aging, especially in chemical reactions involving water. In a high-humidity environment, moisture may cause problems such as hydrolysis and corrosion of materials. However, the influence of humidity usually needs to cooperate with other factors, and its effect is not as significant as that of temperature when acting alone, so γ2 is less than γ1.

[0145] Irradiation can trigger the photochemical reaction of materials and damage the molecular structure. However, in the natural environment or general accelerated aging tests, the irradiation intensity and time are limited, and some materials have a certain tolerance to irradiation, so its influence degree is lower than that of temperature and humidity, thus γ3 is less than γ2.

[0146] Salt spray mainly affects metal materials and reduces their performance through electrochemical corrosion. However, for most non-metal materials, the influence of salt spray concentration is relatively small and the influence range is relatively narrow, so γ4 is less than γ3.

[0147] In a salt spray environment, metal materials will corrode rapidly, forming corrosion products on the surface, reducing the strength and toughness of the materials, and seriously damaging the material structure when severe. For example, metal buildings by the sea are eroded by salt spray for a long time, and the strength of the metal structure decreases, affecting safety. The oxidation effect of ozone on materials is usually relatively slow. In the ordinary environment, the influence degree on material properties is not as strong as the corrosion effect of salt spray on metals, so γ5 is less than γ4.

[0148] The influence of ozone concentration on material aging has a certain universality. As long as ozone exists in the environment, it may react with materials oxidatively, damage the material structure, and reduce the material properties. The interaction of temperature and humidity, although significantly affecting material aging under specific conditions, may not be in extreme states simultaneously in some environments, resulting in an unclear interaction and a relatively narrow influence range, so γ6 is less than γ5.

[0149] Temperature and humidity are extremely common environmental factors in the process of material aging and are prevalent in various natural and artificial environments. Their interaction has a more extensive influence on material aging, covering almost all types of materials. The interaction of irradiation dose and salt spray concentration is relatively limited. Salt spray mainly affects metal materials, and irradiation requires specific light conditions, which are not available in all scenarios. Therefore, the influence range of the interaction between temperature and humidity is wider and the weight coefficient is higher, so γ7 is less than γ6.

[0150] Therefore, on the basis of γ1 + γ2 + γ3 + γ4 + γ5 + γ6 + γ7 = 1, let 0 < γ7 < γ6 < γ5 < γ4 < γ3 < γ2 < γ1 < 1.

[0151] As an implementation, the value range of γ1 is 0.25 - 0.3, the value range of γ2 is 0.2 - 0.25, the value range of γ3 is 0.15 - 0.2, the value range of γ4 is 0.1 - 0.15, the value range of γ5 is 0.05 - 0.1, the value range of γ6 is 0.05 - 0.08, and the value range of γ7 is 0.02 - 0.05. The specific values are set by technicians according to the actual situation and are not limited here.

[0152] On the basis of the above embodiments, the change amount of the rupture strength of the polymer material to be predicted and the first environmental severity coefficient are numerically calculated to obtain the first performance evaluation index of the polymer material to be predicted. The formula is as follows:

[0153]

[0154] Among them, XPxs 1 is the first performance evaluation index of the polymer material to be predicted. The first performance evaluation index is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change amount of the rupture strength and the first environmental severity coefficient. When the change amount of the rupture strength of the polymer material to be predicted is smaller and the first environmental severity coefficient is larger, the first performance evaluation index is smaller, and the performance of the polymer material is better.

[0155] In the formula, σ1 is the initial rupture strength of the polymer material to be predicted, and σ0 is the final rupture strength of the polymer material to be predicted.

[0156] In the formula, there is an inverse relationship between the first performance evaluation index and the first environmental severity coefficient, indicating that under relatively harsh environmental conditions, if the change in rupture strength is small, it means that the performance of the polymer material is good; otherwise, it means that the performance of the polymer material is poor.

[0157] On the basis of the above embodiments, the change amount of the elongation rate of the polymer material to be predicted and the corresponding second environmental severity coefficient are numerically calculated to obtain the second performance evaluation index of the polymer material to be predicted. The formula is as follows:

[0158]

[0159] Among them, XPxs 2The second performance evaluation index of the polymer material to be predicted, which is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change in elongation and the second environmental severity coefficient. When the change in elongation of the polymer material to be predicted is smaller and the second environmental severity coefficient is larger, the first performance evaluation index is smaller, and the weather resistance of the polymer material is better;

[0160] In the formula, ε1 is the initial elongation of the polymer material to be predicted, and ε0 is the final elongation of the polymer material to be predicted;

[0161] Similarly, in the formula, there is an inverse relationship between the second performance evaluation index and the second environmental severity coefficient. This means that under harsh environmental conditions, if the change in elongation is small, it indicates that the performance of the polymer material is good, and vice versa, it indicates that the performance of the polymer material is poor.

[0162] Based on the above embodiments, the performance evaluation index of the polymer material to be predicted is compared with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material. The specific process is as follows:

[0163] When any one of the performance evaluation indexes is not greater than the threshold, that is, XPxs 1 ≤yz1 or XPxs 2 ≤yz2, it is considered that the performance of the polymer material is good;

[0164] When all the performance evaluation indexes are greater than the threshold, that is, XPxs 1 >yz1 and XPxs 2 >yz2, it is considered that the performance of the polymer material is poor;

[0165] Among them, yz1 is the threshold of the first performance evaluation index, yz2 is the threshold of the second performance evaluation index. yz1 and yz2 can be obtained by analyzing past experimental data, finding the performance evaluation indexes of the material under different environmental conditions, calculating and analyzing the performance evaluation indexes, calculating the quantiles of each performance evaluation index, and based on the data distribution, selecting a reasonable threshold. It can be considered to use the median as the threshold.

[0166] Please refer to Figure 2 , the present invention also provides a technical solution:

[0167] A polymer material performance prediction system based on deep learning, which is used to execute any one of the above-mentioned polymer material performance prediction methods based on deep learning, including:

[0168] A data set construction module, which is used to determine variable parameters that affect the performance of polymer materials, construct multiple groups of different variable parameter combinations, prepare different groups of polymer material samples according to different variable parameter combinations respectively, conduct a tensile test on one sample of each sample group to obtain the initial tensile performance parameters of the sample, and after an interval of T time periods, conduct an accelerated aging test on each sample group with different accelerated aging condition data respectively. After the test is completed, conduct a tensile test on another sample of each sample group to obtain the final tensile performance parameters of the sample;

[0169] An initial performance construction module, which is used to construct an initial performance prediction model. The initial performance prediction model is constructed based on a deep learning network model. The variable parameter combination of the sample is used as the input, and the corresponding initial tensile performance parameter is used as the label to train the initial tensile performance prediction model;

[0170] An initial parameter module, which is used to input the variable parameter combination of the polymer material to be predicted into the initial performance prediction model to obtain the initial tensile performance parameters of the polymer material to be predicted;

[0171] An aging performance construction module, which is used to construct an aging performance prediction model. The aging performance prediction model is constructed based on a deep learning network model. The initial tensile performance parameter and the accelerated aging condition data are used as the input, and the corresponding final tensile performance parameter is used as the label to train the aging performance prediction model;

[0172] An aging parameter module, which is used to input the initial tensile performance parameter and the accelerated aging condition data of the polymer material to be predicted into the aging performance prediction model to obtain the final tensile performance parameters of the polymer material to be predicted;

[0173] A data processing and prediction module, which is used to conduct data processing and correlation analysis on the accelerated aging condition data to generate an environmental severity coefficient, conduct numerical calculations on the change amount of the tensile performance parameter of the polymer material to be predicted and the environmental severity coefficient to obtain a performance evaluation index of the polymer material to be predicted, and compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material.

[0174] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to get a formula that is closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0175] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0176] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] As described above, the above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all of them should be covered by the protection scope of the present application.

Claims

1. A method for predicting the properties of polymer materials based on deep learning, characterized in that: The specific steps are as follows: S1. Determine the variable parameters that affect the performance of the polymer material, construct multiple groups of different variable parameter combinations, prepare different groups of polymer material samples according to different variable parameter combinations respectively, conduct a tensile test on one sample of the sample group to obtain the initial tensile performance parameters of the sample. After an interval of T time periods, then conduct an accelerated aging test on each sample group with different accelerated aging condition data respectively. After the test is completed, conduct a tensile test on another sample of the sample group to obtain the final tensile performance parameters of the sample; S2. Construct an initial performance prediction model. The initial performance prediction model is constructed based on a deep learning network model. Use the variable parameter combination of the sample as the input and the corresponding initial tensile performance parameter as the label to train the initial tensile performance prediction model; S3. Input the variable parameter combination of the polymer material to be predicted into the initial performance prediction model to obtain the initial tensile performance parameters of the polymer material to be predicted; S4. Construct an aging performance prediction model. The aging performance prediction model is constructed based on a deep learning network model. Use the initial tensile performance parameter and the accelerated aging condition data as the input and the corresponding final tensile performance parameter as the label to train the aging performance prediction model; S5. Input the initial tensile performance parameter and the accelerated aging condition data of the polymer material to be predicted into the aging performance prediction model to obtain the final tensile performance parameters of the polymer material to be predicted; S6. Process the accelerated aging condition data and conduct a correlation analysis to generate an environmental severity coefficient. Perform a numerical calculation on the change amount of the tensile performance parameter of the polymer material to be predicted and the environmental severity coefficient to obtain the performance evaluation index of the polymer material to be predicted. Compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance quality of the polymer material.

2. The method for predicting the properties of polymer materials based on deep learning according to claim 1, wherein: The tensile performance parameters include the breaking strength and elongation rate. The variable parameters include processing temperature, injection pressure, cooling rate, annealing temperature, and annealing time data. The accelerated aging condition data includes environmental temperature, environmental humidity, radiation dose, salt spray concentration, and ozone concentration data.

3. The method for predicting the properties of polymer materials based on deep learning according to claim 2, wherein: Process the accelerated aging condition data and conduct a correlation analysis on the accelerated aging condition data and the severity of the accelerated aging condition to generate a first environmental severity coefficient. The formula is as follows: HJxs 1 = β1WD + β2RH + β3I + β4SC + β5O Among them, HJxs 1 is the first environmental severity coefficient of the accelerated aging test. The first environmental severity coefficient is used to characterize the severity of the accelerated aging conditions. WD is the temperature, RH is the humidity, I is the irradiation dose, SC is the salt spray concentration, and O is the ozone concentration; In the formula, β1, β2, β3, β4, and β5 are the weight coefficients of temperature, humidity, radiation dose, salt spray concentration, and ozone concentration in the first environmental severity coefficient formula respectively. On the basis of β1 + β2 + β3 + β4 + β5 = 1, let 0 < β5 < β4 < β3 < β2 < β1 < 1.

4. The method for predicting the properties of polymer materials based on deep learning according to claim 3, wherein: Process different accelerated aging condition data and conduct a correlation analysis on the accelerated aging condition data and the severity of the accelerated aging condition to generate a second environmental severity coefficient. The formula is as follows: HJxs 2 = γ1WD + γ2RH + γ3I + γ4SC + γ5O + γ6(WD × RH) + γ7(WD × O) Among them, HJxs 2 is the second environmental severity factor for the accelerated aging test. The second environmental severity factor is used to characterize the severity of the accelerated aging conditions. WD×RH is the interaction term of temperature and humidity, and WD×O is the interaction term of temperature and ozone concentration; In the formula, γ1, γ2, γ3, γ4, γ5, γ6, and γ7 are the weight coefficients of temperature, humidity, irradiation dose, salt fog concentration, ozone concentration, the interaction term of temperature and humidity, and the interaction term of temperature and ozone concentration in the first environmental severity coefficient formula. On the basis of γ1 + γ2 + γ3 + γ4 + γ5 + γ6 + γ7 = 1, it is stipulated that 0 < γ7 < γ6 < γ5 < γ4 < γ3 < γ2 < γ1 < 1.

5. The method for predicting the properties of polymer materials based on deep learning according to claim 4, characterized in that: Perform numerical calculations on the change in the rupture strength of the polymer material to be predicted and the first environmental severity coefficient to obtain the first performance evaluation index of the polymer material to be predicted. The formula is as follows: Among them, XPxs 1 is the first performance evaluation index of the polymer material to be predicted. The first performance evaluation index is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change amount of the rupture strength and the first environmental severity coefficient; In the formula, σ1 is the initial rupture strength of the polymer material to be predicted, and σ0 is the final rupture strength of the polymer material to be predicted.

6. The method for predicting the properties of polymer materials based on deep learning according to claim 5, characterized in that: Perform numerical calculations on the change in the elongation of the polymer material to be predicted and the corresponding second environmental severity coefficient to obtain the second performance evaluation index of the polymer material to be predicted. The formula is as follows: Among them, XPxs 2 is the second performance evaluation index of the polymer material to be predicted. The second performance evaluation index is used to comprehensively evaluate the performance of the polymer material to be predicted by combining the change amount of elongation and the second environmental severity coefficient; In the formula, ε1 is the initial elongation of the polymer material to be predicted, and ε0 is the final elongation of the polymer material to be predicted.

7. The method for predicting the properties of polymer materials based on deep learning according to claim 6, wherein: Compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold to comprehensively evaluate the performance of the polymer material. The specific process is as follows: When any one of the performance evaluation indices is not greater than the threshold value, i.e., XPxs 1 ≤yz1 or XPxs 2 ≤yz2, it is considered that the performance of the polymer material is good; When all performance evaluation indices are greater than the threshold, i.e., XPxs 1 > yz1 and XPxs 2 > yz2, it is considered that the performance of the polymer material is poor; Among them, yz1 is the threshold of the first performance evaluation index, and yz2 is the threshold of the second performance evaluation index.

8. A polymer material property prediction system based on deep learning, the system is used to execute a polymer material property prediction method based on deep learning according to any one of claims 1-7, characterized in that: Including: A data set construction module, which is used to determine the variable parameters affecting the performance of the polymer material, construct multiple groups of different variable parameter combinations, prepare different polymer material sample groups according to different variable parameter combinations, conduct a tensile test on one sample of the sample group to obtain the initial tensile performance parameters of the sample. After an interval of T time periods, then use different accelerated aging condition data to conduct accelerated aging tests on each sample group respectively. After the test is completed, conduct a tensile test on another sample of the sample group to obtain the final tensile performance parameters of the sample; An initial performance construction module, which is used to construct an initial performance prediction model. The initial performance prediction model is constructed based on a deep learning network model. Use the variable parameter combination of the sample as the input and the corresponding initial tensile performance parameter as the label to train the initial tensile performance prediction model; An initial parameter module, which is used to input the variable parameter combination of the polymer material to be predicted into the initial performance prediction model to obtain the initial tensile performance parameters of the polymer material to be predicted; An aging performance construction module, which is used to construct an aging performance prediction model. The aging performance prediction model is constructed based on a deep learning network model. Use the initial tensile performance parameter and the accelerated aging condition data as the input and the corresponding final tensile performance parameter as the label to train the aging performance prediction model; An aging parameter module, which is used to input the initial tensile performance parameter and the accelerated aging condition data of the polymer material to be predicted into the aging performance prediction model to obtain the final tensile performance parameters of the polymer material to be predicted; A data processing and prediction module, which is used to process the accelerated aging condition data and conduct correlation analysis to generate an environmental severity coefficient, perform numerical calculations on the change amount of the tensile property parameters of the polymer material to be predicted and the environmental severity coefficient to obtain the performance evaluation index of the polymer material to be predicted, compare the performance evaluation index of the polymer material to be predicted with the corresponding preset threshold, and comprehensively evaluate the performance quality of the polymer material.

Images