Method and system for predicting full mix proportion strength and modulus of sludge solidified soil

Abstract

The invention discloses a method and system for predicting the full mix proportion strength and modulus of sludge solidified soil, and relates to the field of building materials, and the method comprises the following steps: obtaining related input data of the sludge solidified soil; constructing a physical information neural network model, and embedding a physical equation describing the mechanical behavior of the sludge solidified soil into the neural network model as a constraint; training a physical information neural network model by using the input data to enable model parameters to converge step by step, and comprehensively considering experimental data errors and physical equation residual errors in the training process; and using the trained and converged physical information neural network model to predict the strength and modulus of the sludge solidified soil. The system comprises all modules for implementing all the steps. According to the method, a traditional neural network is combined with a physical equation for describing the sludge solidified soil, and the problems that a traditional prediction method excessively depends on an empirical formula, the data demand quantity is large, and the generalization performance is poor are effectively solved.

Description

Technical Field

[0001] The invention relates to the field of building materials, and more specifically to a method and system for predicting full mix strength and modulus of silt-solidified soil. Background Art

[0002] Silt-stabilized soil refers to soil material formed by solidifying dredged silt by adding cement and other curing agents. It is widely used in projects such as soft foundation reinforcement, roadbed filling and marine reclamation. Accurate prediction of the mechanical properties of silt-stabilized soil (such as compressive strength and deformation modulus) is crucial for engineering design and safety assessment. However, there are many shortcomings in existing prediction methods: traditional empirical methods usually fit empirical formulas based on limited indoor test data, which is difficult to cover all factors affecting strength and modulus, resulting in limited prediction accuracy; simple data-driven models (such as traditional BP neural networks or regression models) can use a large amount of test data for prediction, but when the amount of data is insufficient or exceeds the training range, the model often lacks physical constraints, and the prediction results may be unreasonable; purely relying on numerical simulation (such as finite element consolidation analysis) requires accurate acquisition of soil parameters, which is complex and time-consuming. In summary, the existing technology lacks a prediction method that can both integrate physical mechanisms and use measured data to improve the accuracy and generalization ability of strength and modulus prediction of silt-stabilized soil. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for predicting the full mix strength and modulus of silt-stabilized soil, so as to solve the problems mentioned in the background technology.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions: A method for predicting the full mix strength and modulus of silt-stabilized soil comprises the following steps: Obtain relevant input data for silt-stabilized soil; A physical information neural network model is constructed, and physical equations describing the mechanical behavior of silt-solidified soil are embedded into the neural network model as constraints; The physical information neural network model is trained using the input data to gradually converge the model parameters, and the experimental data error and the physical equation residual are comprehensively considered during the training process; The physical information neural network model after training convergence is used to predict the strength and modulus of silt-stabilized soil.

[0005] Optionally, the relevant input data include initial physical and chemical parameters and curing condition data of the silt-solidified soil; The initial physical and chemical parameters include moisture content, density, particle composition and curing agent dosage; The curing condition data include age, ambient temperature and humidity.

[0006] Optionally, the physical information neural network model is a multi-layer deep neural network, including an input layer, an output layer and at least two hidden layers; The input layer receives relevant input data of the silt-stabilized soil, and the output layer outputs predicted values ​​of strength and modulus of the silt-stabilized soil.

[0007] Optionally, physical mechanism constraint equations related to the strength and modulus of silt-solidified soil are introduced during the training process of the physical information neural network model, including control equations for soft soil consolidation and empirical equations for the strength growth of solidified soil.

[0008] Optionally, the control equation for soft soil consolidation includes the Terzaghi one-dimensional consolidation equation; The empirical equation for the growth of the strength of the stabilized soil includes: ; in, is the compressive strength of the solidified soil at the curing age t, is the strength stability limit, and k is the strength growth rate constant.

[0009] Optionally, the method further comprises constructing a loss function including physical constraints to train the physical information neural network model; The loss function is composed of a data fitting error term and a physical equation residual term; The data fitting error term is used to measure the deviation between the model prediction strength and modulus and the measured value; the physical equation residual term is used to measure the degree of violation of the model prediction result to the relevant physical equation; The neural network parameters are optimized by minimizing the loss function.

[0010] Optionally, the physical information neural network model adopts a multi-stage training optimization method: the training process is divided into multiple stages, including an early rapid intensity growth stage, a mid-term stable intensity growth stage, and a late intensity stabilization stage; the weight of the residual term of the physical equation in the loss function is dynamically adjusted at different training stages.

[0011] Optionally, the physical information neural network model is trained using an iterative optimization algorithm based on gradient descent. During the training process, a weight coefficient is set for the residual term of the physical equation to balance its contribution with that of the data error term. The training is terminated when the loss function converges to a preset threshold or reaches a maximum number of training iterations.

[0012] Optionally, the method utilizes multi-source data for training, including field sensor monitoring data and laboratory test data; The on-site sensor monitoring data include measured data of soft soil foundation settlement, pore water pressure or moisture content; The laboratory test data include the data obtained from the unconfined compressive strength and deformation modulus tests of silt-stabilized soil samples taken from the site; the data diversity and reliability of model training are improved through data preprocessing and fusion.

[0013] Optionally, the method is used in any of the following scenarios: Ground treatment engineering, predicting the strength and modulus of silt-solidified soil in soft soil foundation after reinforcement; Soft soil foundation reinforcement in road construction, predicting the strength and modulus of silt-solidified soil in roadbed; Dredger fill silt foundation treatment in marine engineering, predicting the strength and modulus of dredger fill silt soil after solidification treatment.

[0014] Therefore, the present invention also discloses a silt solidified soil full mix strength and modulus prediction system, including a data acquisition module, a physical information neural network calculation module and a result output module; The data acquisition module is used to collect the initial parameters and maintenance condition data of silt solidified soil. The physical information neural network calculation module is used to call the trained physical information neural network model according to the initial parameters and curing conditions to calculate the strength and modulus prediction results; The result output module is used to output the prediction results.

[0015] The advantage of the present invention over the prior art is that, by adopting a silt-solidified soil strength and modulus prediction method and system based on physical information neural network, the data-driven characteristics of the traditional neural network are combined with the consolidation equation and strength growth equation describing the physical process of silt-solidified soil, effectively overcoming the problems of the traditional prediction method being too dependent on empirical formulas, large data requirements and poor generalization performance. The present invention significantly improves the prediction accuracy and generalization ability of the model through physical constraints, and can still obtain prediction results that conform to the actual laws of the soil body even when there are relatively few experimental data or the prediction conditions exceed the experimental range. In addition, the present invention further adopts a multi-stage training optimization method, and dynamically optimizes the model training strategy stage by stage for the stage characteristics of silt-solidified soil in rapid growth in the early stage, stable growth in the mid-term, and gradually stabilization of strength in the late stage, so that the accuracy of the prediction model in the long-term prediction scenario is significantly improved, effectively overcoming the problem of poor prediction effect of the traditional data-driven model in the long term, and significantly enhancing the reliability and applicability in practical engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is the overall flow chart of the method of the present invention; Figure 2It is a comparison diagram of the prediction of compressive strength of silt-stabilized soil by the PINN model of the present invention and the traditional BP neural network; Figure 3 It is a relative error comparison diagram of the PINN model of the present invention and the traditional BP neural network in predicting the strength and deformation modulus of silt-solidified soil; Figure 4 It is a comparison of the extrapolated prediction graphs of the PINN model of the present invention and the traditional BP neural network when the curing age reaches 90 days. DETAILED DESCRIPTION

[0017] The specific implementation of the present invention will be described below in conjunction with the accompanying drawings.

[0018] The present invention provides a method and system for predicting the full mix strength and modulus of silt-stabilized soil based on a physical information neural network (PINN, Physics-Informed Neural Network), which overcomes the shortcomings of traditional methods by integrating physical constraints into the neural network model.

[0019] like Figure 1 As shown, the overall process of the method of the present invention is as follows: Obtain relevant input data for silt-stabilized soil; A physical information neural network model is constructed, and physical equations describing the mechanical behavior of silt-solidified soil are embedded into the neural network model as constraints; The physical information neural network model is trained using the input data to gradually converge the model parameters, and the experimental data error and the physical equation residual are comprehensively considered during the training process; The physical information neural network model after training convergence is used to predict the strength and modulus of silt-stabilized soil.

[0020] More specific embodiments are as follows: Firstly, the key parameters that affect the strength and modulus of silt-stabilized soil are collected, including the initial moisture content, density, particle size composition, type and dosage of curing agent, and curing and curing conditions such as age, ambient temperature and humidity.

[0021] In some embodiments, multi-source data can be used for training, including field sensor monitoring data and laboratory test data; the field sensor monitoring data includes measured data of soft soil foundation settlement, pore water pressure or moisture content; the laboratory test data includes data obtained from unconfined compressive strength and deformation modulus tests on silt-solidified soil samples taken from the site.

[0022] Then, a physical information neural network model is constructed, that is, the physical equations reflecting the mechanical behavior of silt-solidified soil are integrated as constraints in the deep neural network architecture.

[0023] Specifically, the neural network can adopt a fully connected feedforward network structure, for example, including several hidden layers to capture the nonlinear relationship between input parameters and output mechanical indicators. In order to ensure that the model prediction conforms to the engineering law, physical mechanism constraints such as soft soil consolidation and strength development are introduced in the network training process: on the one hand, the control equation of consolidation theory is used as a constraint condition to ensure that the soil mechanical response predicted by the model conforms to the trend of the consolidation process (for example, the law that the strength increases and the porosity decreases with the increase of consolidation degree); On the other hand, an empirical model of the growth of the strength of the consolidated soil with age (such as a curve relationship in which the strength gradually increases with time and tends to be stable) is introduced as a supplementary physical constraint. These physical constraints are added to the network training through the loss function, so that the network not only fits the existing data, but also makes the prediction results physically reasonable.

[0024] A commonly used physical formula is the Terzaghi one-dimensional consolidation equation, which describes the change of pore water pressure in the soil during the consolidation process. The expression is: ; in, is the excess pore water pressure, is the consolidation coefficient, t is the time, and z is the depth coordinate.

[0025] This equation can reflect the physical law of the gradual discharge of water and compression of soil during the consolidation process. When it is embedded in PINN, the partial derivatives can be calculated by automatic differentiation of the network output, for example and , and then construct the residual term This residual is added to the loss function as part of the physical constraint, forcing the network to predict results that not only fit the data but also try to satisfy the consolidation equation.

[0026] In addition to the consolidation equation, the empirical formula for the increase in strength over time is also a physical constraint worth introducing. For silt-stabilized soil, the compressive strength usually increases gradually with age, which can be described by the following expression: ; in, is the compressive strength of the solidified soil at the curing age t, is the strength stability limit, and k is the strength growth rate constant.

[0027] This formula captures the tendency of strength to stabilize during the curing process. The way to embed this formula in PINN is to compare the strength value predicted by the network with the value calculated by the formula, calculate the difference between the two as the residual, and include it in the loss function. In this way, the model will consider both experimental data and the physical laws of strength growth during training, thereby improving the rationality of the prediction. If the type of curing agent or environmental conditions are different, the parameters in the formula, such as k It can also be set as a learnable variable, which will be automatically optimized by the network based on the data.

[0028] Another physical formulation that can be introduced is based on the description of Biot's consolidation theory, which is more complex than the Terzaghi equation and is suitable for situations where nonlinear deformation or three-dimensional stress state of the soil is considered. Biot's theory involves coupled equations for displacement and pore water pressure. This theory is embedded in a similar way to the Terzaghi equation, by calculating the relevant derivatives through automatic differentiation and adding the equation residuals to the loss function. However, due to the high computational complexity, it is usually necessary to decide whether to adopt this more sophisticated model based on the specific needs of the problem.

[0029] The embedding of these physical formulas is usually achieved through the loss function mentioned later. In actual operation, the selection and embedding method of physical formulas need to be flexibly adjusted according to the specific characteristics of silt-stabilized soil. If the consolidation process exhibits obvious nonlinearity, or the chemical reaction of the curing agent significantly affects the strength growth, it may be necessary to introduce more complex formulas or modify the parameters of existing formulas. By setting some parameters to be learnable, PINN can automatically adapt to these changes during training. This flexibility, coupled with physical constraints, makes the model both accurate and reliable in predicting strength and modulus, while also being able to evaluate the degree of compliance with physical laws by analyzing the residual terms.

[0030] During the model training phase, the experimental data collected above are used for supervised learning to construct a composite loss function that includes physical constraints.

[0031] For example, the loss function can consist of two parts: The first is the data error loss, which is used to measure the difference between the strength and modulus output by the network and the actual experimental value; The second is the physical equation loss, which is used to measure the degree of violation of the network output to the relevant physical equations (such as consolidation equations and strength-age models). By assigning a certain weight coefficient to the physical equation loss, the role of physical constraints and data fitting in the training process can be balanced.

[0032] The formula is as follows: ; L is the total loss function; L dis the data loss term, is the mean square error between the predicted value and the actual measured value; L p It is the residual term of the physical equation, which is used to measure the degree of violation of the model prediction to the above physical equation; is the weight coefficient used to balance the two losses. The value range of is usually between 0 and 1. When the data is sufficient and reliable, A smaller value (such as 0.01 to 0.1) can be used to give priority to fitting the data; when the data is sparse or the noise is large, A larger value (such as 0.5 to 1) should be used to strengthen the effect of physical constraints.

[0033] In some embodiments, considering that the strength and modulus development of silt-stabilized soil has obvious stage rules, it is particularly manifested as follows: In the early stage (0-14 days), the strength increases rapidly; In the middle period (14-28 days), the rate of strength growth slows down and gradually stabilizes; In the late stage (after 28 days), the strength increases slowly and gradually approaches a stable limit value.

[0034] The traditional single-stage training method of neural networks or data-driven models in the existing technology cannot fully utilize and reflect the characteristics of this phased change, resulting in poor model generalization, especially a significant decrease in the accuracy of long-term predictions.

[0035] The present invention proposes a multi-stage PINN training optimization method for the special property of silt solidified soil, which specifically includes: In the first stage (initial model rapid training stage), the PINN model is trained first using the test data of the early age (such as 0-14 days) to establish a preliminary model so that the model can accurately capture the law of rapid strength growth. Since the data is sufficient and accurate, we hope that the model can fit the actual data characteristics of early rapid strength growth as soon as possible. Therefore, the weight of the physical constraint term in this stage is relatively small, and more emphasis is placed on the data fitting error term, allowing the model to learn features from the data more freely. In the second stage (intensive training in the stable stage), when the model converges to a certain degree, the data of the medium-term age (such as 14 to 28 days) are added, and the weight of the physical constraint term in the loss function is adjusted to emphasize the gradual law of strength growth and gradually converge to the trend of the stable stage; because the model needs to reflect the physical law that the strength of silt-solidified soil gradually increases and tends to be stable, the weight of the physical constraint term is appropriately increased at this time, so that the model more strictly abides by the trend of stable strength growth, rather than relying entirely on limited data fitting, so that the model is more in line with physical reality; In the third stage (long-term generalization training), a small amount of late-stage (28-90 days or even longer) test data is gradually added in the subsequent stages, and the constraints on the residual terms of the physical equation are further strengthened, so that the generalization ability of the PINN model is significantly improved; because late-stage test data are often more sparse, the model directly relies on data fitting and there will be a greater generalization risk. At this time, the weight of the physical constraint term needs to be further increased, even significantly higher than the first two stages. This strengthened physical constraint forces the model to more closely follow the theoretical trend of the long-term strength of the solidified soil tending to be stable. Even if it exceeds the original training data range, the model can give reasonable predictions and significantly improve the generalization performance.

[0036] During the training process, the gradient descent optimization algorithm (such as Adam optimizer combined with L-BFGS, etc.) is used to iteratively adjust the network parameters so that the total loss is gradually reduced. In order to improve the training effect, hyperparameters such as learning rate, batch size, and training rounds can be adjusted, and an early stopping strategy can be used to avoid overfitting. When the loss function converges to a predetermined threshold, the training ends and a physically calibrated neural network model is obtained.

[0037] After the PINN model obtained by the method of the present invention is trained, the strength and modulus of new input conditions can be predicted.

[0038] The user only needs to input the initial state parameters of the silt-solidified soil and the predetermined curing conditions (for example, if you want to predict the strength at 28 days), the model can output the corresponding strength and deformation modulus prediction values. Because the model integrates physical laws, even if the input parameters exceed the original data range, the model can still give results that conform to engineering common sense, which improves the credibility of the prediction.

[0039] The method and system of the present invention can be implemented on a computer in the form of software, or integrated into a field monitoring system to predict the performance of foundation soil in real time. It has a wide range of applications and is suitable for various foundation projects such as soft soil foundation treatment, road subgrade reinforcement, and marine fill reinforcement, providing reliable data support for engineering design.

[0040] In order to verify the effectiveness of the method of the present invention, an experimental example was carried out.

[0041] The experiment selected a silt soil sample from a coastal engineering project, with basic physical properties of about 60% water content and 1.5g / cm³ natural density. Cement was added to the silt as a curing agent, with the mass ratio of 5%, 10% and 15% respectively, and standard samples were prepared and cured indoors for different ages (7 days, 14 days, 28 days).

[0042] Unconfined compressive strength tests were conducted on the solidified soil samples of various ages to obtain compressive strength data, and the corresponding compression deformation modulus was calculated through the stress-strain curve. Several groups of sample data were obtained (for example, strength range 0.5-2.0MPa, modulus range 50-200MPa).

[0043] 80% of the data is used for training, and 20% of the data is used for validation testing. A PINN model is constructed for the above data prediction: the network structure is input layer-3 hidden layers-output layer, the input includes parameters such as curing agent dosage, initial moisture content, age, etc., and the output is the corresponding compressive strength and deformation modulus.

[0044] The hidden layer uses the ReLU activation function, and the number of neurons in each layer is 32, 16, and 8 respectively.

[0045] In terms of physical constraints, the relationship between the evolution of porosity with time in one-dimensional consolidation theory (such as Terzaghi's one-dimensional consolidation equation) and the empirical formula for strength growth are introduced as constraints.

[0046] In the loss function, the data error term uses the mean square error between the predicted and measured values ​​of strength and modulus, and the physical constraint term is defined based on whether the strength output by the network changes with age in accordance with the selected empirical model. The weight coefficient of the physical constraint term is set to 0.1, and the Adam optimization algorithm is used for training 10,000 iterations with an initial learning rate of 0.001. During the training process, the learning rate is appropriately adjusted according to the performance of the validation set to accelerate convergence.

[0047] After training is complete, the model is applied to the test set data for prediction.

[0048] The results show that the predicted values ​​of the PINN model established by the method of the present invention are in good agreement with the measured values: the average relative error of the predicted values ​​of compressive strength is about 5%, and the average relative error of the predicted values ​​of deformation modulus is about 8%. In contrast, when the traditional BP neural network model (without introducing physical constraints) with the same training data is used for prediction, the average relative errors of strength and modulus reach 12% and 15%, respectively. It can be seen that the neural network that introduces physical information significantly improves the prediction accuracy. More importantly, when making extrapolated predictions for situations beyond the range of training data (for example, predicting the strength after 90 days of curing), the model obtained by the method of the present invention still gives a reasonable strength growth trend and value; while the prediction results of the traditional neural network model fluctuate greatly when there is a lack of corresponding data support, and the physical rationality is poor. More specifically: Figure 2 shows the comparison between the PINN model and the traditional BP neural network (Back Propagation Neural Network) in predicting the compressive strength of silt-stabilized soil. It can be clearly seen from the figure that the prediction results of the PINN model are closer to the actual experimental data, indicating that the accuracy of the PINN model in predicting strength is significantly higher than that of the traditional BP neural network.

[0049] Figure 3 compares the relative errors of the PINN model and the traditional BP neural network in predicting the strength and deformation modulus of silt-stabilized soil. It can be clearly seen from the figure that the strength prediction error of the PINN model is about 5%, and the modulus prediction error is about 8%, which are significantly lower than the prediction error of the BP model, reflecting the advantage of the PINN model in prediction accuracy.

[0050] Figure 4 is an extrapolated prediction chart for a maintenance age of 90 days, showing the prediction performance of the PINN model and the traditional BP neural network model beyond the training data range. The PINN model gives a more reasonable and robust trend extrapolation prediction, while the BP model shows obvious unreasonable prediction fluctuations at a long age (such as 90 days), which reflects the stronger generalization performance of the PINN model due to the introduction of physical constraints.

[0051] It can be seen that the PINN-based silt-stabilized soil strength and modulus prediction method of the present invention has higher accuracy and robustness than traditional methods, and will be more reliable in engineering applications. This experimental example verifies the effectiveness of the present invention and shows that incorporating physical constraints into neural networks can make full use of existing test data and follow the laws of soil mechanics, thereby achieving accurate prediction of the mechanical properties of silt-stabilized soil.

[0052] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting the full mix strength and modulus of silt-stabilized soil, characterized in that: The following steps are involved: Obtain relevant input data for silt-stabilized soil; A physical information neural network model is constructed, and physical equations describing the mechanical behavior of silt-solidified soil are embedded into the neural network model as constraints; The physical information neural network model is trained using the input data to gradually converge the model parameters, and the experimental data error and the physical equation residual are comprehensively considered during the training process; The physical information neural network model after training convergence is used to predict the strength and modulus of silt-stabilized soil.

2. The method according to claim 1, characterized in that: The relevant input data include initial physical and chemical parameters of the silt-solidified soil and curing condition data; The initial physical and chemical parameters include moisture content, density, particle composition and curing agent dosage; The curing condition data include age, ambient temperature and humidity.

3. The method according to claim 1, characterized in that: The physical information neural network model is a multi-layer deep neural network, including an input layer, an output layer and at least two hidden layers; The input layer receives relevant input data of the silt-stabilized soil, and the output layer outputs predicted values ​​of strength and modulus of the silt-stabilized soil.

4. The method according to claim 1, characterized in that: Introducing physical mechanism constraint equations related to the strength and modulus of silt-solidified soil during the training of the physical information neural network model, including a control equation for soft soil consolidation and an empirical equation for the growth of solidified soil strength; The control equation of soft soil consolidation includes Terzaghi one-dimensional consolidation equation; The empirical equation for the growth of the strength of the stabilized soil includes the following: ； in, is the compressive strength of the solidified soil at the curing age t, is the strength stability limit, and k is the strength growth rate constant.

5. The method according to claim 1, characterized in that: The method further includes constructing a loss function including physical constraints to train the physical information neural network model; The loss function is composed of a data fitting error term and a physical equation residual term; The data fitting error term is used to measure the deviation between the model prediction strength and modulus and the measured value; the physical equation residual term is used to measure the degree of violation of the model prediction result to the relevant physical equation; The neural network parameters are optimized by minimizing the loss function.

6. The method according to claim 5, characterized in that The physical information neural network model adopts a multi-stage training optimization method: the training process is divided into multiple stages, including an early stage of rapid intensity growth, a mid-term stage of stable intensity growth, and a late stage of intensity stabilization; the weight of the residual term of the physical equation in the loss function is dynamically adjusted at different training stages.

7. The method according to claim 5, characterized in that: The physical information neural network model is trained using an iterative optimization algorithm based on gradient descent. During the training process, a weight coefficient is set for the residual term of the physical equation to balance its contribution with the data error term. The training is terminated when the loss function converges to a preset threshold or reaches a maximum number of training iterations.

8. The method according to claim 1, characterized in that: The method utilizes multi-source data for training, including field sensor monitoring data and laboratory test data; The on-site sensor monitoring data include measured data of soft soil foundation settlement, pore water pressure or moisture content; The laboratory test data include the data obtained from the unconfined compressive strength and deformation modulus tests on silt-stabilized soil samples taken from the field.

9. The method according to claim 1, characterized in that: The method is used in any of the following scenarios: Ground treatment engineering, predicting the strength and modulus of silt-solidified soil in soft soil foundation after reinforcement; Soft soil foundation reinforcement in road construction, predicting the strength and modulus of silt-solidified soil in roadbed; Dredger fill silt foundation treatment in marine engineering, predict the strength and modulus of dredger fill silt soil after solidification treatment.

10. A system for predicting the full mix strength and modulus of silt-stabilized soil, characterized in that: It includes a data acquisition module, a physical information neural network calculation module and a result output module; The data acquisition module is used to collect the initial parameters and maintenance condition data of silt solidified soil. The physical information neural network calculation module is used to call the trained physical information neural network model according to the initial parameters and curing conditions to calculate the strength and modulus prediction results; The result output module is used to output the prediction results.

Images