Coastal tidal action hydrodynamic grouting simulation method and system based on deep learning

Abstract

The application discloses a deep learning-based simulation method and system for coastal tide action hydrodynamic grouting, relates to the technical field of hydrodynamic grouting simulation in a coastal environment, and inputs pretreated tidal data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction, so as to predict the dynamic change of grouting parameters under the action of tides; a grouting simulation model of a stratum under the action of tides is established based on a diffusion equation, the dynamic change of the grouting parameters is substituted into the grouting simulation model for grouting simulation calculation, the diffusion behavior of grout is obtained, multi-objective optimization of the grouting parameters is performed based on the diffusion behavior of the grout, and the optimal combination of the grouting parameters is obtained. The deep learning model and the physical constraint model are combined, the deep learning model is used to predict the dynamic change of the grouting parameters under the condition of tides, the dynamic change of the grouting parameters is used as input data of the physical constraint model, the grouting diffusion process under the dynamic environment of tides is accurately simulated, and the optimal combination of the grouting parameters is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dynamic water grouting simulation in coastal environment, and particularly relates to a dynamic water grouting simulation method and system based on deep learning under the action of coastal tides. BACKGROUND

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] The groundwater flow pattern in the coastal area is deeply affected by the periodic changes of tides, and especially large fluctuations may occur occasionally at the peak of tides, which brings challenges to the grouting treatment of tunnels and underground engineering. Due to the effects of multiple factors such as tidal erosion, fresh and salt water displacement and seawater erosion on rock mass for a long time, the engineering geology and hydrogeology conditions in the coastal area are extremely complex. As one of the main geological disasters in the development process of underground space in the coastal karst area, gushing water seriously threatens the engineering construction, ecological environment and people's life and property safety. Since the ocean is the discharge boundary of the aquifer, frequent and periodic tidal action will cause the water level of the aquifer to fluctuate, and thus the flow rate and even the flow direction of the groundwater near the boundary will change, thereby directly affecting the water quantity and water pressure of gushing water. Such fluctuations have a significant impact on engineering construction, especially underground engineering.

[0004] Grouting technology is an effective means to control gushing water, and can effectively seal and repair and reinforce the stratum. However, the traditional grouting method relies on empirical formula or numerical model, and lacks real-time response mechanism for parameter selection and adjustment, making it difficult to respond to the dynamic changes of the coastal dynamic water environment. Existing dynamic water grouting researches are mostly focused on constant flow rate conditions, without considering complex dynamic factors such as tidal periodic changes, tidal range and occasional large fluctuations. In addition, few methods can intelligently optimize the selection of grouting parameters to adapt to the tidal fluctuations over time and environmental changes. SUMMARY

[0005] To overcome the deficiencies of the prior art, the present application provides a dynamic water grouting simulation method and system based on deep learning under the action of coastal tides, which combines a deep learning model and a physical constraint model, predicts the dynamic changes of grouting parameters under tidal conditions through the deep learning model, and uses the predicted data as the input data of the physical constraint model, accurately simulates the grouting diffusion process under the dynamic tidal environment, and obtains the optimal combination of grouting parameters.

[0006] To achieve the above-mentioned purpose, one or more embodiments of the present application provide the following technical solutions:

[0007] In a first aspect, the present application provides a dynamic water grouting simulation method based on deep learning under the action of coastal tides, comprising:

[0008] Acquire and preprocess the tidal data, hydrogeological data and grouting parameter data of the coastal area;

[0009] Input the preprocessed tidal data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction, and predict the dynamic changes of the grouting parameters under the action of tides;

[0010] Based on the diffusion equation, a grouting simulation model of the stratum under the action of tides is established, the dynamic changes of the grouting parameters are substituted into the grouting simulation model for grouting simulation calculation, the diffusion behavior of the grout is obtained, and multi-objective optimization of the grouting parameters is performed based on the diffusion behavior of the grout to obtain the best combination of grouting parameters.

[0011] In a further technical solution, the deep learning model sequentially includes a convolutional neural network layer, an improved long short-term memory network layer and an improved self-attention mechanism.

[0012] In a further technical solution, the improved long short-term memory network layer enhances the modeling capability of tidal fluctuations by introducing tidal dynamic weights and periodic parameters, and its formula is expressed as:

[0013] f t =σ(W f ·[h t-1 ,x t ]+b f +γ f ·θ LSTM (t))

[0014] i t =σ(W i ·[h t-1 ,x t ]+b i +γ i ·θ LSTM (t))

[0015]

[0016]

[0017] o t =σ(W o ·[h t-1 ,x t ]+b o +γ o ·θ LSTM (t))

[0018] h t =o t ·tanh(C t )

[0019] where f t is the forget gate, i t is the input gate, is the candidate memory cell state, C t is the memory cell state, o t is the output gate, γ f , γ i , γ C , γ o are coefficients for adjusting the influence of tides on the forget, input, candidate memory, and output gates, σ is an activation function, W f , W i , W C , W o are weight matrices corresponding to the forget gate, input gate, candidate memory cell state, and output gate, respectively, h t is the hidden state at the current time step t, h t-1 is the hidden state at the previous time step before the current time step t, x t is the input feature at the current time step t, b f , b i , b C , b o are bias vectors corresponding to the forget gate, input gate, candidate memory cell state, and output gate, respectively, θ LSTM (t) is the dynamic weight of tides, p t is a periodic parameter.

[0020] In a further technical solution, the improved self-attention mechanism introduces a tidal feature time weight, which is represented by the formula:

[0021] θ attn (t) = α attn · (sin(ωt + φ) + γ attn · ΙΙ peak (t))

[0022] where α attn is a coefficient controlling the size of attention weight, ω is the angular frequency of tides, φ is the phase of tides, γ attn is a coefficient controlling the influence of the peak of tides on the attention mechanism, ΙΙ peak (t) is an indicator function.

[0023] In a further technical solution, the grouting simulation model adopts an improved diffusion equation, which is represented by:

[0024]

[0025] where C = C(x, y, z, t) is the concentration distribution of grout, x, y, z are spatial coordinates, t is time, and D(t) is the time-varying diffusion coefficient under the action of tidal fluctuations. S(t,C) is a source term.

[0026] Further technical solutions, the improved diffusion equation is discretized and solved by using finite difference method.

[0027] Further technical solutions, the multi-objective optimization of grouting parameters is realized by combining genetic algorithm and particle swarm optimization algorithm.

[0028] In a second aspect, the present application provides a deep learning-based simulation system for coastal tide action hydrodynamic grouting, comprising:

[0029] A data acquisition module configured to acquire and preprocess tide data, hydrogeological data and grouting parameter data in a coastal area;

[0030] A network model prediction module configured to input the preprocessed tide data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction, and predict the dynamic changes of grouting parameters under the action of tides;

[0031] A physical model simulation module configured to establish a grouting simulation model of strata under the action of tides based on a diffusion equation, substitute the dynamic changes of the grouting parameters into the grouting simulation model for grouting simulation calculation, obtain the diffusion behavior of grout, and perform multi-objective optimization of grouting parameters based on the diffusion behavior of grout to obtain the best combination of grouting parameters.

[0032] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the steps of the deep learning-based simulation method for coastal tide action hydrodynamic grouting according to the first aspect.

[0033] In a fourth aspect, the present application provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the deep learning-based simulation method for coastal tide action hydrodynamic grouting according to the first aspect.

[0034] The above one or more technical solutions have the following beneficial effects:

[0035] The present application combines a deep learning model with a physical constraint model to simulate the grouting process under tidal conditions. First, the deep learning model predicts the dynamic changes of grouting parameters under tidal conditions, and inputs the changes into the physical constraint model. Combined with the hydrogeological characteristics of the stratum, the diffusion equation is used for detailed grouting simulation calculation to obtain the diffusion and penetration of the grout under different tidal conditions, as well as the original hydrogeological characteristics of the stratum. The influence of dynamic water conditions on the grouting area is analyzed. By comprehensively considering various indicators of grouting effect (such as whether the grout diffusion range meets the expectation, whether the penetration depth meets the reinforcement requirements, etc.), the long-term stability of the grouting area under different tidal conditions is evaluated. The evaluation results are provided to the user in the form of a report, providing a scientific basis for the user's decision-making in coastal geotechnical engineering grouting construction.

[0036] The present application adopts a deep learning structure combining convolutional neural network, long short-term memory network and self-attention mechanism to construct a deep learning model, and improves the long short-term memory network and self-attention mechanism considering the tidal factor to predict the dynamic changes of grouting parameters under the action of tides, realizing the modeling of tidal periodic fluctuations and real-time response to occasional large fluctuations.

[0037] The present application proposes a grouting simulation model based on the diffusion equation to describe the diffusion behavior of grout under tidal fluctuation conditions. The model combines the time-varying diffusion coefficient of tidal action and the grouting source term, and adjusts the boundary conditions in real time to accurately simulate the grouting diffusion process under the dynamic environment of tides.

[0038] The present application realizes multi-objective optimization by combining genetic algorithm (GA) and particle swarm optimization (PSO) to optimize the parameter configuration of grouting, including grouting amount, grouting pressure, etc., to meet the parameter selection requirements under complex tidal conditions and obtain the best combination of grouting parameters under tidal conditions. BRIEF DESCRIPTION OF DRAWINGS

[0039] The drawings accompanying the specification of the present application form a part thereof, serve to provide further understanding of the present application, and together with the description, explain the present application. The present application is shown by way of illustration in the drawings and is not meant to be unduly limited thereto.

[0040] Figure 1 is a flowchart of the grouting simulation method of the present application. DETAILED DESCRIPTION

[0041] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0042] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0043] In the case of no conflict, the embodiments in the application and the features in the embodiments can be combined with each other.

[0044] Embodiment one

[0045] As Figure 1 shown, the embodiment discloses a deep learning-based coastal tidal action hydrodynamic grouting simulation method, which comprises the following steps:

[0046] S1: Obtain the tidal data, hydrogeological data and grouting parameter data of the coastal area and perform preprocessing;

[0047] In this embodiment, the tidal data collection: by setting multiple tide observation stations and tidal flow monitoring points in the coastal area, using sensors and other equipment to collect real-time data such as tide height, tidal flow speed, direction, period, etc. The collected data is periodically transmitted and summarized and stored in the data management module. In the data preprocessing stage, the collected tidal data is processed for missing values, such as using interpolation method to fill in the missing data points; identify and correct outliers to ensure the accuracy of the data.

[0048] Hydrogeological data collection: through geological exploration, drilling sampling and other means to obtain the hydrogeological parameters such as stratum lithology, permeability and porosity of the coastal area. After these data are sorted into a standard format, they are stored in the data management module. In the data preprocessing stage, the hydrogeological data is standardized to make its value range meet the requirements of deep learning model training.

[0049] Grouting parameter data collection: during the grouting construction process in the coastal area, record the grouting parameters such as grouting amount, grouting speed, grouting pressure and grouting viscosity. These data are transmitted to the data management module in real time. In the data preprocessing stage, the grouting parameter data is normalized to facilitate calculation and processing during model training.

[0050] In the data preprocessing stage, data standardization specifically refers to standardizing all continuous variable data to meet the requirements of model input, which is expressed by the formula:

[0051]

[0052] where x is the original value of the data, μ is the mean of the data, σ is the standard deviation of the data, x normalized is the standard value of the data.

[0053] The anomaly detection adopts the isolation forest to detect anomalies by constructing multiple trees through random partitioning of samples, and the specific process is as follows:

[0054] (1) Randomly sample the data set to generate multiple subsets.

[0055] (2) Each subset constructs an "isolation tree" (Isolation Tree), and in each tree, a feature and a split value are randomly selected each time to divide the samples.

[0056] (3) Calculate the average path length E(h(x)) of the sample, and the shorter the path length, the more abnormal the sample.

[0057] The anomaly score formula of the isolation forest is:

[0058]

[0059] where E(h(x)) is the average path length of sample x in the tree, n is the size of the data set, and c(n) is a constant for normalization. The closer s(x, n) is to 1, the more abnormal the sample, E() is the mathematical expectation, and the value calculated by averaging the path lengths of sample x in all isolation trees, h(x) represents the path length calculation function of sample x in a certain isolation tree.

[0060] S2: input the preprocessed tidal data, hydrogeological data and grouting parameter data into the pre-trained deep learning model for prediction, and predict the dynamic changes of the grouting parameters under the action of tides.

[0061] In this embodiment, the deep learning model adopts a deep learning structure combining convolutional neural network CNN, long short-term memory network LSTM and self-attention mechanism Attention to predict the changes of parameters in the grouting process, especially the dynamic changes of grouting parameters under the action of tides, and realizes the modeling of tidal periodic fluctuations and real-time response to occasional large fluctuations. Specifically, the deep learning model sequentially includes a convolutional neural network CNN layer, a long short-term memory network LSTM layer and a self-attention mechanism, the convolution kernel size, step length and other parameters of the CNN layer are set to effectively extract spatial features; the number of neurons, time step and other parameters of the LSTM layer are set to adapt to the processing of time series data; the weight parameters of the Attention mechanism are set to focus on important feature information. The number of neurons and connection mode of the input layer, hidden layer and output layer are determined to construct a complete deep learning model architecture.

[0062] S201: Convolutional Neural Network (CNN): used for preliminary processing of input data, especially to extract spatial features related to grouting environment (e.g., underground geological structure, soil layer properties, hydrological characteristics, and spatial patterns of tidal fluctuations, etc.). Convolutional operation can capture local patterns of geological features, reduce interference of irrelevant information, and provide more compact feature representation for subsequent layers. The output of the CNN layer is the extracted spatial feature map, which serves as the basis data for time series information input into the LSTM layer. The output of the CNN layer provides high-quality input features for LSTM, making subsequent time series modeling more effective. Convolutional operation:

[0063]

[0064] where i and j refer to the position index in the generated feature map (i represents the ith row and j represents the jth column), Filter mn is the convolution kernel, m represents the size of the convolution kernel in the horizontal direction, n represents the size of the convolution kernel in the vertical direction, σ is the activation function, b is the bias term, and Input is the original data input to the convolutional neural network.

[0065] S202: Long Short-Term Memory Network (LSTM): LSTM focuses on time series analysis, capable of learning long-term and short-term dependencies in tidal data, particularly suitable for processing tidal periodic fluctuations and time patterns of grouting parameter changes. Through memory cells, it captures the dependencies between past and future in time series to cope with the dynamic changes of grouting parameters under the influence of tides. By relying on data dependencies in the time dimension, it predicts the trend of grouting parameters, especially in long-term cycles and short-term fluctuations, to accurately model tidal data. The output of the LSTM layer is a feature sequence containing time dependencies, which contains time series information of tidal fluctuations and changes in grouting effects, providing candidate information for the Attention mechanism.

[0066] Introducing tidal dynamic weight θ LSTM (t) represents the intensity and change of tidal fluctuations, reflecting the influence of tides on grout diffusion and flow. The intensity of tidal fluctuations may vary significantly at different time steps (such as high tide, low tide, or abnormal fluctuations). The intensity of tidal fluctuations changes with the tidal cycle and abnormal fluctuations. θ LSTM (t) can be modeled by the following formula:

[0067] θ LSTM (t) = α LSTM ·(sin(ωt+φ) + β LSTM ·ΙΙ abnormal (t))

[0068] where α LSTMis the adjustment coefficient to control the influence of tidal fluctuation intensity on memory; ω is the angular frequency of the tide, controlling the periodic change; φ is the initial phase of the tide, determining the starting position of the fluctuation; β LSTM is the influence coefficient of abnormal fluctuation, controlling the intensity of abnormal fluctuation on memory update; ΙΙ abnormal (t) is an indicator function, ΙΙ abnormal (t) = 1 when the time t is in the abnormal tidal fluctuation, otherwise 0.

[0069] Further, the periodicity parameter p t is introduced to represent the periodic change of the tidal fluctuation, reflecting the influence of the peak and trough of the tidal fluctuation on the grouting process, and adjusting the memory and update of the LSTM network combined with the tidal period and phase information. The periodicity parameter is modeled by the following formula:

[0070] p t = sin(ωt+φ)

[0071] where ω is the angular frequency of the tide, controlling the periodic change; t is the time, and φ is the initial phase of the tide, determining the starting position of the fluctuation.

[0072] Further, the two introduced parameters, i.e. the tidal dynamic weight and the periodicity parameter, are brought into the LSTM to enhance the modeling ability of the LSTM for the tidal fluctuation. These new variables can help the LSTM dynamically adjust its memory update and forgetting process during the grouting process, especially during periods of high tidal fluctuation intensity. The new LSTM formula is as follows:

[0073] f t = σ(W f · [h t-1 , x t ] + b f + γ f · θ LSTM (t))

[0074] i t = σ(W i · [h t-1 , x t ] + b i + γ i · θ LSTM (t))

[0075]

[0076]

[0077] o t = σ(W o · [h t-1 , x t ] + bo +γ o ·θ LSTM (t))

[0078] h t =o t ·tanh(C t )

[0079] where f t is the forget gate, i t is the input gate, is the candidate memory cell state, C t is the memory cell state, o t is the output gate, γ f , γ i , γ C , γ o are coefficients for adjusting the influence of tides on the forget, input, candidate memory, and output gates, σ is the activation function, W f , W i , W C , W o are the weight matrices corresponding to the forget gate, input gate, candidate memory cell state, and output gate, respectively, h t is the hidden state at the current time step t, h t-1 is the hidden state at the previous time step before the current time step t, x t is the input feature at the current time step t, b f , b i , b C , b o are the bias vectors corresponding to the forget gate, input gate, candidate memory cell state, and output gate, respectively, θ LSTM (t) is the tide dynamic weight, p t is the periodicity parameter.

[0080] By introducing the tide dynamic weight and periodicity parameter, the improved LSTM model can dynamically adjust the memory and update of the input features at different stages of tidal fluctuations. Especially at the peak, trough, and abnormal fluctuation periods of tides, the model can give more attention and adjust the grouting parameter prediction according to the tidal intensity. At the peak of tides, θ LSTM (t) will increase, and the input and state update of the model will be more concentrated to cope with possible strong tidal changes; during the low tide or abnormal fluctuation, the periodicity parameter p t also plays a regulating role, ensuring that the model can capture the influence of these key moments and avoid error accumulation.

[0081] S203: Attention mechanism: The attention mechanism is used to weight the focus on specific time steps or features in the time series to capture key data segments highly relevant to the grouting effect, especially the abnormally large fluctuations or key features in the tidal fluctuations. By assigning a weight to each time step of the LSTM output, the focus is on the time points that have a greater impact on the grouting effect (such as abnormal fluctuations or tidal peak points), thus improving the model's performance in large fluctuation scenarios. This weighted attention helps the model be more accurate when dealing with abnormal tidal fluctuations. After the attention mechanism is processed, the model generates a weighted feature representation that focuses more on the features of key time steps, making the model more targeted and accurate in predicting the grouting effect.

[0082] Similarly, the periodicity and abnormality of tidal fluctuations are further introduced, and a tidal feature time weight θ attn (t) is added to adjust the attention allocation at a specific tidal period. The tidal feature time weight θ attn (t) is dynamically generated according to the intensity and periodicity of tidal fluctuations. The specific formula is as follows:

[0083] θ attn (t)=α attn ·(sin(ωt+φ)+γ attn ·ΙΙ peak (t))

[0084] where α attn is a coefficient that controls the size of the attention weight; ω is the angular frequency of the tide, which controls the periodicity of the tide; φ is the phase of the tide, which determines the starting time of the tidal fluctuations; ΙΙ peak (t) is an indicator function, when time t is at the peak of the tide, ΙΙ peak (t) = 1, otherwise 0; γ attn is a coefficient that controls the influence of the tidal wave peak on the attention mechanism.

[0085] Through this design, the model can adaptively adjust the attention weight of each time step according to the periodicity and abnormal fluctuation characteristics of the tide. When the tidal fluctuations reach the peak or occur abnormally, the weight will increase accordingly, strengthening the model's attention to these key periods.

[0086] Further, by introducing the tidal feature time weight, the improved attention mechanism formula is as follows:

[0087]

[0088] where e t is the attention score of the current time step t; h t-1 is the hidden state of the previous time step of the current time step t; xt It is the input feature of the current time step (such as tide level, current velocity, etc.); θ attn (t) represents the tidal characteristic time weight, reflecting the degree of influence of tidal fluctuations on the grouting process; T represents the vector transpose, w a These are the weights of the attention score; W h W x W θ It is the parameter matrix to be learned.

[0089] Here, by introducing θ attn The model can dynamically adjust the attention allocation at each time step according to the characteristics of tidal fluctuations, especially during the peak period of the tide or during abnormal fluctuations, and can strengthen the attention to important moments.

[0090] Furthermore, according to score e t Calculate the attention weight α t , is represented as:

[0091]

[0092] Among them, e t It is the attention score at the current time step t, α t This represents the attention weights of the model at time step t. l It is the attention score for all time steps l during computation, where l is an index representing the set of all time steps.

[0093] Due to the introduction of tidal characteristic time weight θ attn (t), the model can automatically allocate higher attention during the period when tidal fluctuations are most significant (such as the high tide period).

[0094] Furthermore, by weighting the output features, a weighted feature representation c is generated, which serves as the input for grouting process simulation:

[0095]

[0096] Among them, h t It is the output of the LSTM layer, representing the hidden state at time step t; α t It refers to the attention weights at time step t.

[0097] In this way, the model will pay more attention to the time steps when tidal fluctuations are most significant (such as tidal peaks, troughs, or sudden abnormal fluctuations), thereby improving the accuracy and stability of predictions.

[0098] The preprocessed historical data (including tidal data, hydrogeological data and grouting parameter data) is divided into different data sets according to a certain proportion (such as 80% training set, 10% validation set and 10% test set). The training set data is used to train the constructed deep learning model, and appropriate training parameters such as learning rate (such as 0.001), training rounds (such as 100 rounds) are set. The stochastic gradient descent algorithm is used as the training algorithm, and the weight parameters of the model are adjusted by minimizing the mean square error loss function. During the training process, the performance of the model is monitored every certain number of rounds (such as 10 rounds) using the validation set data. When the accuracy on the validation set no longer improves or signs of overfitting appear, adjust the learning rate or the number of hidden layer neurons and other hyperparameters for optimization. After multiple rounds of training and optimization, the test set data is used to evaluate the final optimized model to ensure that the model has good generalization ability.

[0099] When it is necessary to predict the dynamic changes of grouting parameters, the preprocessed historical data (including tidal data, hydrogeological data and grouting parameter data) and the set initial value of grouting parameters (determined according to the experience of similar coastal grouting projects in the past) are used as input data for the model. These data are input into the trained deep learning model, and the model predicts the changes of grouting parameters over time under the tidal conditions, including whether the grouting amount needs to be adjusted in time, how to change the grouting speed, etc., to adapt to the periodic changes of tides and possible abnormal situations (such as large tidal fluctuations), and provide optimization suggestions for grouting under abnormal tidal conditions (such as adjusting grouting rate, pressure, etc.).

[0100] S3: Establish a grouting simulation model of the stratum under the action of tides based on the diffusion equation, and substitute the dynamic changes of the grouting parameters into the grouting simulation model for grouting simulation calculation to obtain the diffusion behavior of the grout.

[0101] In this embodiment, under the action of tides, the grouting process is affected by various dynamic factors, including tidal level changes, flow rate fluctuations and grout diffusion characteristics. Therefore, the present application innovatively proposes a grouting simulation model based on the diffusion equation, which is used to describe the diffusion behavior of grout under tidal fluctuations, specifically: grout diffusion range, grout penetration depth, grout flow path, grout flow rate change, etc. The model combines the time-varying diffusion coefficient of tidal action and the grouting source term, and accurately simulates the grouting diffusion process in the dynamic tidal environment by adjusting the boundary conditions in real time.

[0102] S301: Physical constraint model based on diffusion equation (grouting simulation model)

[0103] In the traditional diffusion model, the diffusion process of slurry usually relies on the diffusion coefficient of the material. However, under the influence of tides, the grouting process is affected by factors such as tidal level, flow rate changes, etc., resulting in dynamic changes in diffusion coefficient and flow behavior. In order to more accurately simulate this process, the invention uses the following improved diffusion equation:

[0104]

[0105] where C = C(x, y, z, t) is the concentration distribution of slurry, x, y, z are spatial coordinates, and t is time; D(t) is the time-varying diffusion coefficient under the action of tidal fluctuations, representing the speed of tidal-induced flow and slurry diffusion; is the Laplacian of concentration, representing the spatial variation of concentration; S(t, C) is the source term, representing the grouting source or other disturbance sources induced by tides.

[0106] Further, the tidal fluctuations dynamically change the diffusion coefficient, and the mathematical description of the time-varying diffusion coefficient D(t) is as follows:

[0107] D(t) = D0·(1 + α·sin(ωt + φ))

[0108] where D0 is the baseline diffusion coefficient, the diffusion value when there is no tidal fluctuation; α is the tidal fluctuation intensity coefficient; ω is the tidal angular frequency, related to the tidal period; φ is the phase of the tidal fluctuation, used to reflect the initial state of the tide.

[0109] This formula reflects the periodic modulation of tidal fluctuations on the diffusion speed of slurry, making the diffusion model flexible to adapt to the dynamic environment of tides.

[0110] Further, the source term S(t, C) is used to describe the characteristics of the slurry injection process, including grouting rate, decay behavior, etc. Its expression is:

[0111] S(t, C) = γ·C in (t) - β·C

[0112] where γ is the grouting rate coefficient; C in (t) is the concentration of the grouting source, representing the concentration of the slurry at the injection point; β is the decay coefficient of the slurry, describing the concentration reduction of the slurry due to mixing, sedimentation, etc. during the diffusion process; C = C(x, y, z, t) is the concentration distribution of the slurry, x, y, z are spatial coordinates, and t is time.

[0113] The source term model, combined with the conditions of tidal fluctuations, can accurately reflect the kinetic characteristics of slurry injection and diffusion.

[0114] Further, in the tidal fluctuation environment, the diffusion equation needs to be combined with boundary conditions and initial conditions. The boundary condition: the tidal prediction result generated by the deep learning model provides the dynamic change of the grouting parameters, denoted as: wherein represents the predicted grouting parameter change. The initial condition: the concentration distribution at the initial time of grouting is set by the C=C(x, y, z, 0) grouting scheme.

[0115] S302: Since the improved diffusion equation is a partial differential equation (PDE), it needs to be solved by a numerical method. In the present application, the finite difference method (FDM) is used to discretize and solve the diffusion equation.

[0116] (1) Spatial and temporal discretization

[0117] Let the spatial grid points be x i , y j , z k , and the time step be Δt, then the discretized form of the diffusion equation is:

[0118]

[0119] wherein, is the concentration at the jth time step and the ith spatial point; is the discrete form of the Laplacian operator, which is expressed according to the finite difference method as:

[0120] (2) Time stepping

[0121] Use explicit or implicit methods for time stepping.

[0122] Explicit method:

[0123] Implicit method (more suitable for larger time steps):

[0124]

[0125] (3) Real-time adjustment

[0126] After each time step calculation, the diffusion coefficient D(t) and the source term S(t, C) are dynamically updated in combination with the dynamic change of the grouting parameters generated by the deep learning model, ensuring that the simulation result can reflect the dynamic change of the tide.

[0127] S303: Combine the deep learning model with the physical constraint model for mutual optimization, including the deep learning model outputting the boundary conditions and diffusion coefficients of the physical constraint model, and the physical constraint model feeding back the training data of the deep learning model. Specifically, the deep learning model, i.e., the CNN+LSTM+Attention model, generates the dynamic changes of grouting parameters, which are used to update the boundary conditions and diffusion coefficients of the diffusion equation in real time; the solution of the physical constraint model, i.e., the diffusion equation, can be used for training the deep learning model to further optimize the prediction of the influence of tidal fluctuations on the grouting process.

[0128] S304: Realize multi-objective optimization by combining genetic algorithm (GA) and particle swarm optimization (PSO) to optimize the parameter configuration of grouting, including grouting amount, grouting pressure, etc., to meet the parameter selection requirements under complex tidal conditions. GA is used to generate the initial population, and the diversity of solutions is expanded through selection, crossover and mutation operations, while PSO performs local search and accelerates convergence on the initial solutions provided by GA to find the best combination of grouting parameters.

[0129] (1) Objective function

[0130] F(X) = η1·grouting effect - η2·cost + η3·fluctuation adaptability

[0131] Where X is the parameter to be optimized, including grouting amount, pressure, flow rate, etc.; grouting effect represents the optimization target of grouting effect, such as penetration depth or flow rate; cost represents the cost control target, including grouting material and time cost; fluctuation adaptability measures the robustness of grouting parameters under different tidal fluctuation conditions. η1, η2, η3 represent the weight coefficients for balancing the relationship between grouting effect, cost and fluctuation adaptability, respectively.

[0132] (2) Optimization algorithm

[0133] Genetic algorithm (GA): generate multiple initial candidate solutions (different combinations of grouting parameters), calculate the fitness based on the simulation results of physical equations. Through selection, crossover and mutation operations, the search range is expanded to find better solutions, and the next generation of candidate solutions is generated. The final population is passed to PSO for further optimization.

[0134] Offspring = α·Parent1 + (1-α)·Parent2

[0135] Where α is the crossover coefficient, usually set to 0.5.

[0136] Particle swarm optimization (PSO): particle swarm optimization (PSO) simulates the cooperative search process of a group, and each solution is regarded as a "particle" and updated according to the individual's historical best position and global best position. PSO is used to perform local search on the population provided by GA to accelerate convergence to the optimal solution. The formula is represented as:

[0137]

[0138] where ω is the inertia weight, controlling the dependence of the particle on the previous velocity; c1, c2 are learning factors, adjusting the influence of individual cognition and group cognition; r1, r2 are uniformly distributed random numbers, used to increase search diversity; denotes the velocity of the i-th particle at the t+1 time step, i is the index of the particle, used to distinguish different particle individuals, t represents the time step, p i denotes the individual historical best position of particle i, denotes the velocity of the i-th particle at the t time step, x denotes the particle, g denotes the group global best position, that is, the optimal solution position found by the entire particle swarm in the search process.

[0139] S305: Real-time monitoring of grouting process parameters (such as grouting pressure, grout flow rate, etc.) through sensors, and adjusting the grouting strategy in combination with the output of numerical simulation, that is, the best grouting parameter combination obtained after numerical simulation optimization.

[0140] The main simulation process is as follows:

[0141] (1) Real-time data acquisition: Collect real-time data from sensors and input into the deep learning model.

[0142] (2) Feedback adjustment: Compare the optimal parameters predicted by the deep learning model with the real-time data feedback from the sensors, and if there is a significant deviation, adjust the parameters.

[0143] (3) Update and display: Apply the adjusted parameters to the grouting equipment, and display the simulation results in the form of charts in real time.

[0144] The present application uses a deep learning model to predict the dynamic changes of grouting parameters, which is based on historical data and the rules learned by the model, and estimates the possible changes of grouting parameters over time under the action of tides. It mainly reflects how grouting parameters such as grouting volume, grouting speed, grouting pressure change over time during the tidal fluctuation process, helping to understand the dynamic characteristics of the grouting process, but this is not necessarily the optimal parameter combination. In S304, the best grouting parameter combination obtained by using genetic algorithm (GA) and particle swarm optimization (PSO) for multi-objective optimization is obtained after considering the grouting effect (such as whether the grout diffusion range meets the expectation, whether the penetration depth meets the reinforcement requirements, etc.), cost (grouting material and time cost), and fluctuation adaptability (robustness of grouting parameters under different tidal fluctuation conditions), etc. This stage is based on the prediction of the deep learning model, and further adjusts and selects the parameters through optimization algorithms.

[0145] Embodiment Two

[0146] The embodiment discloses a deep learning-based simulation system for hydrodynamic grouting under coastal tidal action, comprising:

[0147] A data acquisition module configured to acquire and preprocess tidal data, hydrogeological data and grouting parameter data in a coastal area;

[0148] A network model prediction module configured to input the preprocessed tidal data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction to predict dynamic changes of the grouting parameters under tidal action;

[0149] A physical model simulation module configured to establish a grouting simulation model of a stratum under tidal action based on a diffusion equation, substitute the dynamic changes of the grouting parameters into the grouting simulation model for grouting simulation calculation to obtain diffusion behavior of grout, and perform multi-objective optimization of the grouting parameters based on the diffusion behavior of the grout to obtain an optimal combination of the grouting parameters.

[0150] Embodiment Three

[0151] The purpose of the embodiment is to provide a computing device, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method of Embodiment One.

[0152] Embodiment Four

[0153] The purpose of the embodiment is to provide a computer-readable storage medium, a computer-readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to execute the steps of the method of Embodiment One.

[0154] The steps involved in the devices of Embodiments Three and Four above correspond to the method of Embodiment One, and the specific embodiments can be referred to the relevant description part of Embodiment One. The term "computer-readable storage medium" should be understood to include a single medium or multiple media of one or more instruction sets; it should also be understood to include any medium capable of storing, encoding or carrying instruction sets for execution by a processor and causing the processor to perform any of the methods in the present application.

[0155] Those skilled in the art should understand that each module or step of the present application described above can be realized by a general computer device, alternatively, they can be realized by program codes executable by a computing device, so that they can be stored in a storage device for execution by a computing device, or they can be made into individual integrated circuit modules, or a plurality of modules or steps among them can be made into a single integrated circuit module to realize. The present application is not limited to any specific combination of hardware and software.

[0156] The above merely provides the preferred embodiments of the present application, but not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0157] The above merely provides the preferred embodiments of the present application, but not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A deep learning-based simulation method for coastal tidal hydrodynamic grouting, characterized in that, The method comprises the following steps: acquiring and preprocessing tidal data, hydrogeological data and grouting parameter data of a coastal area; inputting the preprocessed tidal data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction to predict dynamic changes of grouting parameters under the action of tides; establishing a grouting simulation model of a stratum under the action of tides based on a diffusion equation, substituting the dynamic changes of the grouting parameters into the grouting simulation model for grouting simulation calculation to obtain diffusion behavior of grout, and performing multi-objective optimization of the grouting parameters based on the diffusion behavior of the grout to obtain an optimal combination of grouting parameters; the grouting simulation model adopts an improved diffusion equation, which is expressed as: wherein, is the slurry concentration distribution, , , are spatial coordinates, is time, is the time varying diffusion coefficient under the action of tidal fluctuations, is the Laplacian of concentration, is the source term.

2. The deep learning-based simulation method of coastal tidal hydrodynamic grouting according to claim 1, wherein, the deep learning model comprises a convolutional neural network layer, an improved long short-term memory network layer and an improved self-attention mechanism in sequence. 3.The deep learning based coastal tidal hydro-mud simulation method according to claim 2, wherein, the improved long short-term memory network layer enhances the modeling capability of tidal fluctuations by introducing tidal dynamic weights and periodic parameters, and its formula is expressed as: in, For the Gate of Oblivion For input gate, Candidate memory cell state, For the state of the memory unit, For output gate, , , , It is a coefficient used to adjust the effect of tides on forgetting, input, candidate memory, and output gate. It is an activation function. , , , These are the weight matrices corresponding to the forget gate, input gate, candidate memory cell states, and output gate, respectively. It is the current time step The hidden state, It is the current time step The hidden state of the previous time step, It is the current time step Input features, , , , These are the bias vectors corresponding to the forget gate, input gate, candidate memory cell state, and output gate, respectively. It is a tidal dynamic weight. It is a periodic parameter. 4.The deep learning based simulation method of coastal tidal hydrodynamic grouting according to claim 2, wherein, the improved self-attention mechanism introduces tidal feature time weights, and its formula is expressed as: wherein, is a coefficient controlling the size of the attention weight, is the angular frequency of the tide, is the phase of the tide, is a coefficient controlling the influence of the tide wave crest on the attention mechanism, is an indicator function. 5.The deep learning based simulation method of coastal tidal hydrodynamic grouting according to claim 1, wherein, the improved diffusion equation is discretized and solved by using the finite difference method. 6.The deep learning based simulation method of coastal tidal hydrodynamic grouting according to claim 1, wherein, The multi-objective optimization of the grouting parameters is realized by combining a genetic algorithm and a particle swarm optimization algorithm.

7. A deep learning-based simulation system for coastal tidal hydrodynamic grouting, characterized in that, The method comprises the following steps: a data acquisition module configured to acquire and preprocess tidal data, hydrogeological data and grouting parameter data of a coastal area; a network model prediction module configured to input the preprocessed tidal data, hydrogeological data and grouting parameter data into a pre-trained deep learning model for prediction to predict dynamic changes of grouting parameters under the action of tides; a physical model simulation module configured to establish a grouting simulation model of a stratum under the action of tides based on a diffusion equation, substitute the dynamic changes of the grouting parameters into the grouting simulation model for grouting simulation calculation to obtain diffusion behavior of grout, and perform multi-objective optimization of the grouting parameters based on the diffusion behavior of the grout to obtain an optimal combination of grouting parameters; the grouting simulation model adopts an improved diffusion equation, which is expressed as: wherein is the slurry concentration distribution, , , are spatial coordinates, is time, is the time varying diffusion coefficient under the action of tidal fluctuations, is the Laplacian of concentration, is the source term.

8. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the steps in the deep learning-based simulation method for dynamic water grouting under the action of coastal tides according to any one of claims 1-6.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the steps in the deep learning-based simulation method for dynamic water grouting under the action of coastal tides according to any one of claims 1-6.

Images