CN119989669B - A grouting diffusion prediction method and system with complex geological multi-attribute constraints - Google Patents
A grouting diffusion prediction method and system with complex geological multi-attribute constraintsInfo
- Publication number
- CN119989669B CN119989669B CN202510058184.5A CN202510058184A CN119989669B CN 119989669 B CN119989669 B CN 119989669B CN 202510058184 A CN202510058184 A CN 202510058184A CN 119989669 B CN119989669 B CN 119989669B
- Authority
- CN
- China
- Prior art keywords
- grouting
- diffusion
- model
- prediction
- grouting diffusion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Investigation Of Foundation Soil And Reinforcement Of Foundation Soil By Compacting Or Drainage (AREA)
- Consolidation Of Soil By Introduction Of Solidifying Substances Into Soil (AREA)
Abstract
本发明公开了一种复杂地质多属性约束的注浆扩散预测方法及系统，预测方法包括：获取复杂地质体的多源属性信息，所述多源属性信息包括裂隙特征、孔隙特征和涌水特性；将多源属性信息输入到训练好的注浆扩散动态预测模型中，输出复杂地质体的注浆扩散预测结果，所述注浆扩散结果包括渗透速率、压力分布和边界条件。本发明在对注浆扩散动态预测模型的训练过程中，结合深度学习模型与实时监测系统，在注浆过程中不断根据监测数据进行预测，并通过反馈机制调整预测结果，优化预测模型，提高预测结果的精度和可靠性。
The present invention discloses a method and system for predicting grouting diffusion for complex geological conditions with multi-attribute constraints. The prediction method comprises: obtaining multi-source attribute information of a complex geological body, the multi-source attribute information including fracture characteristics, pore characteristics, and water inflow characteristics; inputting the multi-source attribute information into a trained dynamic prediction model for grouting diffusion, and outputting a prediction result for the complex geological body, the grouting diffusion result including permeability, pressure distribution, and boundary conditions. During the training process of the dynamic prediction model for grouting diffusion, the present invention combines a deep learning model with a real-time monitoring system, continuously performs predictions based on monitoring data during the grouting process, and adjusts the prediction results through a feedback mechanism, optimizing the prediction model and improving the accuracy and reliability of the prediction results.
Description
Technical Field
The invention relates to the technical field of underground engineering grouting diffusion simulation, in particular to a grouting diffusion prediction method and system for multi-attribute constraint of complex geology.
Background
Grouting diffusion problems in complex geological environments have been a key technical challenge in engineering investigation and underground construction. Conventional grouting diffusion prediction methods usually focus on a single physical parameter (such as permeability or porosity), and often neglect complex multi-attribute coupling effects inside the geologic body. However, physical parameters such as cracks, pores, water burst and the like in the actual geologic body have obvious spatial and temporal changes, so that a single attribute model cannot fully and accurately simulate the diffusion behavior of slurry. Along with the increase of the complexity of the geologic body, the precision and the reliability of the traditional model in practical application are limited. In recent years, researchers try to remedy the defects through a mode of multi-source data fusion and dynamic coupling model, and particularly, certain progress is made in coupling analysis of multiple factors such as cracks, pores and water burst. The multi-attribute coupling model can reflect the diffusion process of slurry in complex geologic bodies more truly based on multi-source data such as resistivity, porosity and the like, but still faces technical problems such as real-time monitoring, dynamic feedback regulation and control and the like.
Most of the existing grouting diffusion simulation is based on a simplified physical model, and although the calculation efficiency is high, the requirements of accurate simulation and real-time performance in complex geologic bodies cannot be met. The rise of deep learning techniques provides a new solution for the processing and prediction of multi-source data. By means of a neural network and other deep learning methods, data from different sensors (such as CT imaging, laser scanning, seismic waves and the like) can be fused, dynamic calculation and prediction of physical parameters in complex geologic bodies are achieved, and the method has remarkable advantages in precise prediction of slurry diffusion paths. However, the current grouting diffusion model based on deep learning still faces the problems of difficult data acquisition, insufficient training samples, large real-time monitoring difficulty and the like, and is not widely applied to industrial practice. In the aspect of experimental platforms and monitoring systems, recent researches construct various geologic body models, and monitor grouting diffusion processes through real-time monitoring equipment (such as pressure sensors, resistivity imaging, flow meters and the like). The experimental devices provide more visual simulation environment for research, but have the limitations of insufficient multi-source data fusion, difficult comprehensive simulation of complex geologic bodies and the like. Therefore, despite a certain technical accumulation, the existing grouting diffusion prediction method and experimental platform still need to be further improved, and especially needs to be innovated in the aspects of data fusion, model optimization and real-time feedback regulation and control.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a grouting diffusion prediction method and a grouting diffusion prediction system for complex geological multi-attribute constraint, and the simulation result can reflect the slurry diffusion process in the complex geological body more accurately by a deep learning and multi-source data fusion method.
The technical scheme of the invention is as follows:
In a first aspect of the present invention, there is provided a method for grouting diffusion prediction for complex geological multi-attribute constraint, comprising:
acquiring multi-source attribute information of a complex geologic body, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics;
Inputting multi-source attribute information into a trained grouting diffusion dynamic prediction model, and outputting a grouting diffusion prediction result of a complex geologic body, wherein the grouting diffusion result comprises permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
in the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model.
In a second aspect of the invention, a complex geological multi-attribute constrained grouting diffusion prediction system is provided,
The multi-attribute information acquisition module is configured to acquire multi-source attribute information of the complex geologic body, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics;
the grouting diffusion prediction module is configured to input multi-source attribute information into a trained grouting diffusion dynamic prediction model and output grouting diffusion prediction results of the complex geologic body, wherein the grouting diffusion results comprise permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
in the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model.
One or more of the technical schemes of the invention has the following beneficial effects:
(1) The invention performs coupling analysis based on multi-source information (fracture characteristics, pore characteristics and water burst characteristics) to dynamically simulate and predict permeability and diffusion state changes in the grouting diffusion process. And in the adopted grouting dynamic prediction model, a multi-attribute coupled permeability model and a dynamic equation are embedded to serve as physical constraint conditions, so that the output of the prediction model is ensured to accord with the fluid mechanics law. In addition, the permeability of the built experimental system can be monitored in real time along with the dynamic changes of time and space, and the permeability is updated to the input layer of the prediction model through a feedback mechanism, so that the space-time coupling characteristic is further optimized, and the diffusion path and the diffusion speed of slurry are accurately simulated. In this way, the multi-attribute coupled permeability model provides physical consistency constraint and real-time update support for dynamic prediction, and improves the accuracy and reliability of a prediction result. The prediction model can accurately predict permeability change and slurry diffusion behavior, can be widely applied to geological multi-attribute simulation and grouting process optimization in the fields of geological exploration, tunnel construction, scientific research experiments and the like, and has good expansibility, adaptability and stability.
(2) The multi-attribute coupled permeability model provided by the invention constructs a dynamic coupled permeability prediction model by introducing a plurality of physical parameters such as fracture characteristics, pore characteristics, water burst characteristics and the like, and is used as a key constraint condition and dynamic input characteristics in the prediction model in an integrated manner. The model establishes a quantitative relation between the multisource attribute and the permeability through dynamic coupling of fracture characteristics, pore characteristics and water burst characteristics, and reveals an influence mechanism of the quantitative relation on the permeability change. In the prediction model, output-permeability data of the permeability model is coupled as a physical constraint condition and is embedded into a physical constraint integration layer of the grouting diffusion dynamic prediction model, so that the model output is ensured to accord with a hydrodynamic rule.
(3) The dynamic grouting diffusion simulation model provided by the invention integrates a plurality of physical properties, and the simulation result can reflect the slurry diffusion process in the complex geologic body more accurately by a deep learning and multi-source data fusion method. Meanwhile, the model can be dynamically updated and optimized along with the input of real-time data, and the statics and the simplicity of the traditional model are overcome. The dynamic grouting diffusion simulation model can provide accurate slurry diffusion prediction, can adapt to the changes of different geological conditions and different physical parameters, and can adjust and optimize the diffusion process in real time. Under the condition of data missing or insufficient real-time monitoring data, the model can still fill in missing information through less sample learning and prediction, and the continuity and stability of the simulation process are ensured.
(4) According to the dynamic grouting diffusion simulation model, actual monitoring data (such as pressure, flow and concentration) and a prediction result are compared through a real-time feedback mechanism, input parameters or weights of a permeability model are reversely adjusted, accuracy of permeability prediction is optimized, and the permeability model and the diffusion prediction model are mutually supported by bidirectional coupling, so that dynamic simulation and prediction accuracy of slurry diffusion behaviors in complex geologic bodies are improved together.
Drawings
FIG. 1 is a flow chart of the construction, training and optimization of the dynamic prediction model of grouting diffusion in embodiment 1 of the present invention;
FIG. 2 is a schematic diagram of a simulation experiment system for grouting diffusion in example 2 of the present invention;
In the figure, 1 is a complex geological simulation unit, and 2 is a water injection grouting unit.
Detailed Description
The invention will be further described with reference to the drawings and examples.
Example 1
In an exemplary embodiment of the present invention, a method for predicting grouting diffusion with multi-attribute constraint in complex geology is provided, including:
Step one, multi-source attribute information of a complex geologic body is obtained, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics.
When the fracture characteristics are acquired, a three-dimensional laser scanner and a karst cave scanner are used for measuring the fracture characteristics of the rock mass, three-dimensional point cloud data of the fracture are generated, and the hand-held scanner is used for carrying out supplementary scanning on local details;
when the characteristics of the pores are acquired, acquiring porosity data by using a microcosmic CT (computed tomography) imager, and establishing a microcosmic model of pore distribution;
When the characteristic features of water burst are obtained, the distribution information of stratum moisture and water burst areas is measured by using resistivity equipment and ground penetrating radar, the permeability of different layers is measured by using an underground moisture layer test sampling system, the integral strength and crushing structure characteristics of the rock mass are obtained by combining an elastometer in the tunnel face and the model tunnel, and a flowmeter and a pressure sensor are arranged to obtain dynamic permeability data.
Inputting multi-source attribute information into a trained grouting diffusion dynamic prediction model, and outputting a grouting diffusion prediction result of the complex geologic body, wherein the grouting diffusion result comprises permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
in the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model.
Specifically, when a multi-attribute coupling permeability model is established, a statistical and machine learning method is used for modeling the coupling relation among a plurality of attribute information, and the influence mechanism of each physical attribute on permeability is revealed.
The multi-attribute coupling permeability model adopts key parameters such as fracture characteristics, pore characteristics, water burst characteristics and the like, and builds a constraint relation among the fracture, the pore and the permeability through dynamic coupling of multi-attribute information. The model integrates various geological information (such as fracture characteristics, pore characteristics and water burst characteristics) and reflects the compound influence of the geological information on the stratum permeability.
The establishment process of the multi-attribute coupling permeability model comprises the following steps:
and collecting a plurality of attribute data (such as fracture characteristics, pore characteristics, water burst characteristics and the like) of the geologic body.
And (3) coupling analysis, namely modeling the coupling relation among a plurality of attributes by using a statistical and machine learning method (such as regression analysis, a support vector machine and the like) and revealing the influence mechanism of each physical attribute on permeability.
And constructing a permeability prediction model (multi-attribute coupling permeability model) based on time change, and dynamically adjusting by utilizing real-time data monitoring to solve the problem of permeability change of the geologic body under different working conditions.
The final model:
the permeability prediction model can provide permeability estimation under different conditions, and can adjust the permeability prediction value in real time, so that the diffusion path and the diffusion speed of slurry in a complex geologic body are more accurate. The permeability prediction model not only can be used for static permeability analysis, but also can be dynamically optimized and adjusted along with the input of real-time data.
The multi-attribute coupled permeability model is used integrally as a key constraint and a dynamic input feature in a grouting diffusion dynamic prediction model. The multi-attribute coupled permeability model establishes a quantitative relation between the multi-source attribute and the permeability through dynamic coupling of fracture characteristics, pore characteristics and water burst characteristics, and reveals an influence mechanism of the quantitative relation on the permeability change. In the grouting diffusion dynamic prediction model, permeability data, which is the output of the multi-attribute coupling permeability model, is used as a physical constraint condition and is embedded into a physical constraint integration layer of the grouting diffusion dynamic prediction model, so that the model output is ensured to accord with a hydrodynamic rule (such as Darcy's law and a diffusion equation).
In the embodiment, the structure of the grouting diffusion dynamic prediction model comprises a physical model and a deep learning model.
And establishing a physical equation describing the grouting diffusion process, and providing constraint conditions for the deep learning model by considering a plurality of physical properties (such as permeability, pressure distribution, porosity, fracture structure and the like).
And the deep learning model part processes multi-source data (such as pressure, temperature, resistivity and the like) from different sensors based on a neural network architecture (such as a convolutional neural network CNN, a long-short-term memory network LSTM and the like) and performs learning and prediction by combining physical constraints.
In the physical model, permeation rate, pressure distribution and diffusion boundaries are key physical properties that directly affect the slurry diffusion process. The dynamic constraint equation limits the output result range of the neural network through a physical model and ensures that the output result range accords with the principles of fluid mechanics and geology. The specific constraint equation includes:
darcy's law, the relationship between the rate of limiting slurry flow and pressure gradient.
And (3) a fluid dynamics equation, namely limiting the coupling relation between the pressure distribution and the permeability.
Diffusion equation describing the velocity of slurry diffusion versus boundary.
And the multi-physical field coupling is used for dynamically updating physical constraints by combining various attributes such as permeation rate, pressure distribution, boundary conditions and the like.
Coupling the physical properties with a dynamic constraint equation, coupling physical constraint conditions such as Darcy's law, a fluid dynamics equation, a diffusion equation and the like with a deep learning model, namely embedding the physical equation serving as a constraint condition of a neural network into a training process through a physical constraint loss function to ensure that a prediction result accords with a physical rule, and transmitting physical variables (such as permeation rate, pressure distribution, boundary conditions and the like) serving as input characteristics into the neural network to participate in learning and prediction together with other data.
Specifically, coupling the permeation rate with darcy's law, coupling the pressure distribution with the hydrodynamic equation, and coupling the diffusion boundary with the diffusion equation, includes:
1. Permeation rate and darcy law coupling
The expression of darcy's law is:
Wherein q is the permeation rate (flow rate), k is the permeation rate; pressure gradient, μ, fluid viscosity.
The specific coupling method comprises the following steps:
(1) Physical constraint layer
In the deep learning model, a ' physical constraint layer ' is introduced, and Darcy's law is embedded into a training loss function of the network. The error between the predicted permeation rate q pred of the neural network and the permeation rate q phys calculated according to darcy's law yields a first physical loss function:
the first loss function affects the network weight update, ensuring that the penetration rate of the network output complies with darcy's law.
(2) Input feature enhancement:
to determine the permeability k and the pressure gradient Directly used as input characteristics and added into an input layer of the deep learning model.
The network utilizes these input features to train with other geologic properties (fracture, porosity, temperature, etc.), enabling more accurate predictions.
2. Pressure distribution and hydrodynamic coupling
The hydrodynamic equation (pressure field equation) is as follows:
the specific coupling method comprises the following steps:
(1) Physical consistency loss:
the pressure field P pred output by the network needs to satisfy the above equation, and is implemented by the second physical constraint loss function:
wherein P pred represents the pressure field predicted by the neural network;
during the network training process, the second physical constraint loss function can enable the model to learn the pressure distribution conforming to the fluid dynamics rule.
(2) Input feature integration:
The permeability k and other parameters (such as boundary conditions, initial pressure distribution) in the fluid dynamics equation are taken as input features, and the historical spatiotemporal data are combined for learning.
(3) Numerical simulation correction:
solving a fluid dynamics equation by using CFD (computational fluid dynamics) and other methods, generating a high-quality training sample, and guiding the supervised learning of the network.
3. Diffusion boundary and spatial physical constraint coupling
The diffusion equation (Fick's law) is:
wherein, C is concentration field, u is flow velocity field, and D is diffusion coefficient.
The specific coupling method comprises the following steps:
(1) Concentration conservation constraint:
Introducing a third physical constraint loss during network training Ensuring that the concentration prediction results satisfy the diffusion equation:
where C pred represents the predicted concentration field of the neural network.
(2) Boundary conditions and initial condition constraints:
And taking the flow velocity field u and the diffusion coefficient D as network input characteristics, and dynamically simulating the relation between the diffusion process and the space constraint. During training, boundary conditions (such as concentration boundary values and flow velocity boundaries) and initial conditions (initial concentration distribution) are used as constraint conditions, so that the reasonable network output result is ensured.
By defining the physical loss functions (such as Darcy's law, fluid dynamics equation and diffusion equation), the physical law is embedded into the network training process, so that the model output is ensured to conform to the physical constraint. Key parameters (such as permeability, pressure gradient and concentration field) in the physical equation are directly used as network input and are learned together with other geological attribute data. And solving a physical equation by utilizing numerical simulation (CFD, FEM and the like), generating a training sample, and providing supervised learning data for deep learning.
In the grouting diffusion dynamic prediction model, dynamic coupling among permeation rate, pressure distribution and diffusion boundary is realized by the following mechanisms:
Physical constraints are embedded in the neural network-the input of the deep learning model includes a number of physical properties, such as permeation rate, pressure distribution, and diffusion boundaries, which are adjusted by physical constraint layers in the network. For example, during each prediction, the network not only depends on historical data, but also corrects according to the permeation rate, the pressure distribution and the diffusion boundary calculated by the physical model, so as to ensure that the prediction result is consistent with the physical rule.
Based on LSTM and other network structures, the model can process time-space sequence data, dynamically adjust pressure distribution and permeation rate on different time and space nodes, and predict slurry diffusion path and speed in real time.
And the feedback mechanism is that after real-time data (such as sensor data) is input, the model can adjust predicted values of the permeation rate, the pressure field and the diffusion boundary according to new data feedback, and grouting parameters are automatically adjusted through an optimization algorithm, so that the accuracy of the simulation process is ensured.
Through the coupling of the permeation rate, the pressure distribution and the diffusion boundary, the model realizes the accurate simulation and prediction of the slurry diffusion process in the complex geologic body. The dynamic coupling equation not only ensures physical consistency, but also realizes space-time dependence and real-time optimization on the basis of multi-source data through a deep learning algorithm, thereby improving the precision and the calculation efficiency of the model. The model can provide powerful technical support for underground construction, grouting optimization and the like.
In order to meet the above requirements on prediction and feedback of the grouting diffusion process in the complex geological environment, the grouting diffusion dynamic prediction model provided by the embodiment has a new deep learning model structure, namely a multi-mode space-time coupling self-adaptive neural network (MTSCNN), combines the multi-mode data fusion, space-time dynamic adjustment and feedback mechanism of deep learning, and can efficiently process the grouting diffusion simulation problem in the complex geological environment.
MTSCNN (Multimodal Temporal-Spatial Coupled Neural Network) is a multi-modal spatio-temporal coupled adaptive neural network architecture. The method combines multisource sensor data (such as pressure sensors, temperature sensors, resistivity imaging, hydrodynamics simulation data and the like) with spatial characteristics (such as cracks, porosity, permeability and the like) of the geologic body, and realizes real-time prediction and dynamic feedback of a complex underground grouting process through a time sequence modeling and feedback mechanism. Specifically, multiple influencing factors of slurry diffusion in complex geologic bodies are solved by combining physical characteristics such as permeation rate, pressure distribution, diffusion boundary and the like with a dynamic coupling equation and combining a deep learning method to predict and optimize. The core idea of the model is to combine the physical constraint condition with the prediction capability of deep learning, perform more accurate simulation and dynamically optimize key parameters in the grouting process.
The structure of the grouting diffusion dynamic prediction model comprises:
a. input layer Multi-modal data fusion
Multisource data input-input of a model includes data sources of multiple modalities, such as geological data-physical attributes including fracture, porosity, permeability, etc., as static inputs. Sensor data including real-time pressure field, temperature field, concentration field, flow field, etc. as dynamic input. And the numerical simulation data is hydrodynamic related data (such as a speed field and a pressure field) obtained through numerical simulation of a CFD model and the like, and is used as auxiliary information input.
Each type of data is subjected to separate preprocessing layers, such as normalization, convolution feature extraction of image data, and the like. The data of different modes are fused through the feature fusion layer so as to convert the data of different types into uniform feature representation.
B. Space-time modeling layer:
Time modeling-using LSTM (long short term memory network) to process time series information in input data. The model is responsible for capturing dynamic evolution rules in the grouting process, such as a slurry diffusion process, pressure change, temperature change and the like.
Spatial modeling, namely extracting spatial features by using a 3D convolutional neural network (3D CNN). The 3D CNN can effectively capture the spatial information in the geologic body, analyze the physical property changes of different positions such as cracks, pores and the like and the influence of the physical property changes on the slurry diffusion.
And a space-time coupling layer, wherein LSTM and 3DCNN are combined, and a space-time coupling layer is embedded in the model. This layer links the temporal and spatial features to better understand how the slurry diffusion process changes with time and spatial variations. Specifically, the space-time coupling layer combines dynamic evolution of the time sequence with distribution information of the spatial features to perform joint modeling.
C. Physical constraint integration layer:
Physical model embedding in order to ensure that the deep learning model conforms to the laws of fluid mechanics, in this layer, the model will embed the laws of physics into the neural network through "physical constraint loss functions" (e.g., darcy's law, continuous equations, etc.). The physical constraint loss function strengthens the adaptability of the model to the real physical process, so that the prediction result can follow the actual physical law.
D. feedback mechanism and adaptive adjustment:
and the feedback layer is used for comparing real-time monitoring data (such as sensor data of pressure, temperature, resistivity and the like) with a prediction result after each prediction, calculating an error and correcting the error through a back propagation algorithm. The purpose of the feedback mechanism is to use the actual field data for adaptive adjustment of the model parameters to optimize the subsequent prediction process.
And (3) self-adaptive learning rate adjustment, namely dynamically adjusting the learning rate of the model based on real-time feedback data so as to improve the response speed and the prediction accuracy of new data. The self-adaptive learning rate can help the model to quickly adjust and converge, avoid overfitting and ensure the real-time property of prediction.
E. output layer:
the prediction result is that the output layer generates a prediction result of grouting diffusion based on the predicted space-time characteristics, and the prediction result comprises the following steps:
Diffusion path and concentration distribution, which is to predict the diffusion path and concentration distribution of slurry in geologic body with time.
The pressure field and the speed field are used for predicting the change of pressure and flow rate in the grouting process.
Diffusion boundary-the boundary that predicts slurry diffusion, identifies the furthest location that the slurry may reach.
The training and optimizing process of the grouting diffusion dynamic prediction model is as follows:
a. Training data:
sources of training data include
The field experimental data is experimental data from different geological conditions, and comprises real-time monitoring data of various sensors.
And (3) numerical simulation data, namely generating training samples matched with the actual scene based on the existing hydrodynamic simulation result.
The historical grouting data comprises monitoring data and prediction results in the past grouting process and is used for model training and error correction.
B. And (3) loss function design:
The loss function is weighted in combination with physical constraint losses (e.g., darcy's law, fluid dynamics, etc.), predictive error losses (e.g., MSE or MAE), and feedback adjustment losses (e.g., error correction). By comprehensively optimizing the loss functions, the model is ensured to meet the requirement on precision, can also follow the physical rule and can be adaptively adjusted.
The technology is realized:
deep learning framework-model building and training using mainstream deep learning frameworks such as TensorFlow or PyTorch.
And the multi-mode data interface is used for fusing the data from different sensors through a unified data processing interface and ensuring the real-time transmission and processing of the data.
Physical model embedding, namely optimizing a neural network training process by using a loss function (such as hydrodynamic constraint, boundary condition and the like) based on physics, and ensuring that model output is consistent with a physical rule.
C. the optimization method comprises the following steps:
The model was optimized using Adam optimizer, adaptive learning rate, and dynamic update strategy. According to the real-time feedback information, the optimizer dynamically adjusts the learning rate to ensure the stability and the high efficiency of the model during real-time prediction.
In the training and optimizing process, the realization of a prediction and feedback mechanism is a key for ensuring that the model efficiently and accurately reflects the actual geologic body behaviors. In order to achieve the aim, the multi-source data fusion, the deep learning model, the space-time dynamic adjustment and the feedback mechanism work cooperatively, and the dynamic correction and optimization in the simulation process are ensured.
The feedback process ensures that the model can be continuously optimized in practical application, grouting parameters are dynamically adjusted, and simulation precision is improved. The feedback mechanism continuously adjusts the model prediction result based on the real-time sensor data and the change in the grouting process so as to realize more accurate simulation and control, and specifically comprises the following steps:
1. real-time data acquisition and input:
Sensor feedback, namely various sensors (such as a pressure sensor, a resistivity sensor, a temperature sensor and the like) installed on site can feed back data in the underground slurry diffusion process in real time, including the flow speed, the diffusion range, the pressure change and the like of the slurry.
And (3) inputting experimental data and field monitoring data in real time, wherein the data are continuously input into the deep learning model as new input characteristics so as to ensure that the model is continuously updated and optimized.
2. Dynamic adjustment and optimization:
physical constraints and model corrections:
By monitoring the data in real time, the model can dynamically adjust physical parameters (such as permeability, pressure distribution and the like) which are taken as constraint conditions to participate in the next prediction calculation.
Feedback correction of the deep learning model, namely, after receiving real-time data, the model carries out feedback correction on a prediction result of the model. For example, assuming that the actual slurry diffusion path deviates from the predicted path, the model adjusts the prediction at the next time by the newly input monitoring data, correcting the slurry diffusion path.
The feedback mechanism is realized:
and (3) back propagation and updating weight, namely when deviation exists between a model prediction result and actual data, the model adjusts the network weight through a back propagation algorithm (such as gradient descent) to gradually approach to a real diffusion process. Through continuous iterative training, the model can learn more accurate physical laws and update in real time.
And (3) real-time optimization, namely performing real-time optimization on the model according to feedback data, and adjusting a prediction result to match the real-time geologic body state. For example, the permeability is adjusted according to the pressure profile of the feedback, or the speed and path of slurry flow is adjusted according to the diffusion boundary of the feedback.
3. Feedback and optimization loop:
And (3) the feedback and prediction alternate process, namely, in the actual grouting process, real-time monitoring data are periodically input into the model, and each input leads to a new prediction result of the model. And the model adjusts and feeds back the prediction result to form a closed-loop optimization system. Whenever a certain variable in the grouting process changes (such as abnormal pressure or unexpected slurry diffusion), the model can be automatically corrected, and a new prediction path and a grouting strategy are generated after optimization.
The feedback optimization technology is realized:
Self-adaptive learning rate, in order to better perform real-time optimization, the feedback mechanism can be combined with the self-adaptive learning rate, so that the model can dynamically adjust the learning speed according to the change of real-time data in the training process, and the response efficiency is improved.
Deep Reinforcement Learning (DRL) in more complex scenarios, reinforcement learning methods may be used to optimize the grouting strategy. The DRL can dynamically adjust grouting parameters according to environmental conditions (such as a pressure field, a slurry diffusion state and the like) and a reward function, so as to realize a long-term optimal strategy.
The MTSCNN model structure can provide an efficient and accurate grouting diffusion prediction and dynamic optimization scheme by fusing multisource data (such as sensor data, geological data and numerical simulation data), space-time modeling (using LSTM and 3D CNN), physical constraint embedding (hydrodynamic equations and the like) and feedback mechanisms (error feedback and self-adaptive adjustment). Through real-time feedback and self-adaptive learning, the model can continuously adjust and optimize the prediction result, solves the problem of insufficient simulation precision and real-time performance in the traditional method, and provides intelligent support for underground engineering in complex geological environment.
Example 2
In an exemplary embodiment of the present invention, a grouting diffusion prediction system for multi-attribute constraint of complex geology is provided, including:
The attribute information acquisition module is configured to acquire multi-source attribute information of the complex geologic body, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics;
the grouting diffusion prediction module is configured to input multi-source attribute information into a trained grouting diffusion dynamic prediction model and output grouting diffusion prediction results of the complex geologic body, wherein the grouting diffusion results comprise permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
in the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model.
Further, the grouting diffusion simulation experiment system, as shown in fig. 2, includes:
the complex geology simulation unit 1 is used for simulating complex geology, and comprises a model box, wherein various rock strata, sand layers and soil layers are filled in the model box, and different geological features of transverse and longitudinal cracks, sand layers, gravel layers, fault fracture zones and karst cave are embedded in the model box, so that a complex geological structure is simulated, for example:
the stratum structure is used for simulating various strata, sand layers and soil layers, including stratum coverage, crack setting, karst cave presetting and the like. The model material selects novel light cement and water-resistant resin, and water-resistant pigment is sprayed on the surface to simulate different rock stratum textures.
And (3) tunnel and structure penetration design, namely displaying rock stratum, water seepage condition, crack distribution and water burst paths of a tunnel penetration area in a transparent acrylic tunnel penetration model.
And (3) simulating geological layers, namely embedding different geological features such as transverse and longitudinal cracks, sand layers, gravel layers, fault fracture zones, karst cave and the like into the model so as to simulate various complex structures in natural geological environment.
The grouting unit 2 comprises a dynamic water circulation module and a grouting control module, wherein the dynamic water circulation module is used for simulating an underground water seepage process, providing constant-pressure water flow and simulating a natural underground water flow path in a model;
The data acquisition and monitoring unit comprises a pressure sensor, a flow sensor, a pore water pressure gauge, a resistivity imaging system and an ultrasonic sensor;
And the data processing unit is used for acquiring data and performing data processing. The desk computer and professional control software are equipped to support the functions of real-time data acquisition, processing and analysis. The software system has a data visualization function, and displays data such as a flowmeter, a pressure sensor, resistivity imaging and the like in real time so as to assist in real-time monitoring and regulating grouting processes. The remote control and monitoring are supported, parameters such as grouting pressure, flow and the like can be remotely regulated and controlled through a software interface, and meanwhile, data storage, historical data inquiry and experimental data derivation are carried out.
The operation steps are as follows:
1. device preparation and system initialization
And checking the integrity of each component of the geological model, and ensuring that the structure of the fissures, the pores, the karst cave and the sandy pebble stratum is preset correctly. And initializing a water injection grouting system to ensure the normal operation of water circulation and grouting equipment. And starting a data acquisition system, wherein the data acquisition system comprises a three-dimensional laser scanner, a karst cave scanner, a CT imager, resistivity equipment and the like, and checking the connection state to ensure that signal acquisition is smooth. And starting a computer and control software, and configuring test parameters including initial attributes of geologic models such as fracture distribution, pore structure, water burst area and the like.
2. Multi-source attribute measurement
(1) Crack characteristic measurement:
and measuring the crack characteristics (width, length and roughness) of the rock mass by using a three-dimensional laser scanner and a karst cave scanner to generate three-dimensional point cloud data of the crack. And carrying out supplementary scanning on the local details by using a handheld scanner so as to improve the resolution of the fracture model.
(2) Pore structure measurement:
And acquiring porosity data by using a microcosmic CT imager, and establishing a microcosmic model of pore distribution. And (3) combining a pore water pressure gauge, monitoring the pressure change of fluid in the pore in real time, and acquiring the associated data of the pore and the crack.
(3) Water burst and permeability characteristics measurement:
And measuring the distribution information of the stratum moisture and the water inrush area by using the resistivity equipment and the ground penetrating radar. And the permeability of different layers is measured through an underground water layering test sampling system, and the integral strength and the crushing structure characteristic of the rock mass are obtained by combining an elastometer. Inside the face and model tunnel, flow meters and pressure sensors are arranged to obtain dynamic permeability data.
3. Grouting diffusion process experiment
(1) Grouting system initialization:
Setting a water injection grouting experiment table at a specific position in the stratum model, and adjusting the initial flow and pressure of grouting according to experimental requirements. Grouting liquid (e.g., colored water or slurry) is configured to facilitate monitoring of its diffusion process.
(2) Monitoring the diffusion process in real time:
And starting the grouting system, gradually increasing grouting pressure, and monitoring the speed and direction of slurry diffusion.
Parameters such as flow, pressure, diffusion radius, etc. are recorded in real time, and the slurry diffusion path is visualized using control software. And monitoring permeability changes in the stratum through resistivity imaging, pore water pressure gauges and other equipment, and identifying the change trend of pores and cracks in the diffusion process.
3. Data processing and coupling analysis
(1) Multi-source data acquisition and fusion:
Data of each sensor and measuring equipment is imported, including a fracture three-dimensional model, porosity distribution, permeability, pressure change and the like. Preprocessing the collected fracture, pore and permeability data, unifying coordinates and data formats, and ensuring the consistency of data fusion.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (9)
1. A grouting diffusion prediction method of complex geological multi-attribute constraint is characterized by comprising the following steps:
acquiring multi-source attribute information of a complex geologic body, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics;
Inputting multi-source attribute information into a trained grouting diffusion dynamic prediction model, and outputting a grouting diffusion prediction result of a complex geologic body, wherein the grouting diffusion result comprises permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
In the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model;
The grouting diffusion dynamic prediction model comprises an input layer, a space-time modeling layer, a physical constraint integration layer, a feedback mechanism, a self-adaptive adjustment layer and an output layer.
2. The complex geologic multi-attribute constrained grouting diffusion prediction method of claim 1, wherein the dynamic equations comprise darcy's law, a hydrodynamic equation and a diffusion equation, and wherein in training the grouting diffusion dynamic prediction model, the permeation rate is coupled with darcy's law, the pressure distribution is coupled with the hydrodynamic equation, and the diffusion boundary is coupled with the diffusion equation.
3. The complex geologic multi-attribute constrained grouting diffusion prediction method of claim 2, wherein the permeation rate is coupled to darcy's law to obtain a first physical loss function:
;
In the formula,Representing the predicted penetration rate of the neural network; Representing the permeation rate obtained by numerical simulation;
Coupling the pressure distribution and the hydrodynamic equation to obtain a second physical loss function :
;
In the formula,Representing the pressure field predicted by the neural network, wherein k represents permeability;
Coupling the diffusion boundary with the diffusion equation to obtain a third physical loss function :
;
In the formula,Representing the predicted concentration field of the neural network, u representing the flow velocity field, and D representing the diffusion coefficient.
4. The complex geologic multi-attribute constrained grouting diffusion prediction method of claim 1, wherein the optimization of the grouting diffusion dynamic prediction model comprises:
The model adjusts the network weight through a back propagation algorithm when deviation exists between the model prediction result and actual experimental data, and gradually approaches to the actual diffusion process;
and (3) real-time optimization, namely performing real-time optimization on the model according to feedback data, and adjusting a prediction result to match the real-time geologic body state.
5. The method for predicting grouting diffusion of complex geological multi-attribute constraint according to claim 1, wherein when fracture characteristics are acquired, a three-dimensional laser scanner and a karst cave scanner are used for measuring the fracture characteristics of a rock mass, three-dimensional point cloud data of the fracture are generated, and a handheld scanner is used for carrying out supplementary scanning on local details;
when the characteristics of the pores are acquired, acquiring porosity data by using a microcosmic CT (computed tomography) imager, and establishing a microcosmic model of pore distribution;
When the characteristic features of water burst are obtained, the distribution information of stratum moisture and water burst areas is measured by using resistivity equipment and ground penetrating radar, the permeability of different layers is measured by using an underground moisture layer test sampling system, the integral strength and crushing structure characteristics of the rock mass are obtained by combining an elastometer in the tunnel face and the model tunnel, and a flowmeter and a pressure sensor are arranged to obtain dynamic permeability data.
6. The complex geologic multi-attribute constrained grouting diffusion prediction method of claim 1, wherein when the multi-attribute coupled permeability model is established, the coupling relation among a plurality of attribute information is modeled by using a statistical and machine learning method, and the influence mechanism of each physical attribute on the permeability is revealed.
7. A complex geological multi-attribute constrained grouting diffusion prediction system, comprising:
The multi-attribute information acquisition module is configured to acquire multi-source attribute information of the complex geologic body, wherein the multi-source attribute information comprises fracture characteristics, pore characteristics and water burst characteristics;
the grouting diffusion prediction module is configured to input multi-source attribute information into a trained grouting diffusion dynamic prediction model and output grouting diffusion prediction results of the complex geologic body, wherein the grouting diffusion results comprise permeation rate, pressure distribution and boundary conditions;
The construction and training process of the grouting diffusion dynamic prediction model specifically comprises the following steps:
building a grouting diffusion simulation experiment system, and performing experimental simulation on a grouting diffusion process of complex geology;
Acquiring multi-source attribute information and permeability of a simulated stratum of an experimental system, establishing a multi-attribute coupling permeability model, and embedding the multi-attribute coupling permeability model into the grouting diffusion dynamic prediction model;
Performing numerical simulation on an experimental system, obtaining the permeation rate, pressure distribution and diffusion boundary in the grouting experimental process, and training a grouting diffusion dynamic prediction model;
After training, the pressure, the resistivity and the concentration of the grouting experiment at the current moment are obtained in real time, and are input into a grouting diffusion dynamic prediction model to predict the grouting diffusion process at the next moment;
In the prediction process, the grouting diffusion dynamic prediction model continuously adjusts the prediction result through a feedback mechanism according to the grouting result of the grouting experiment so as to optimize the grouting diffusion dynamic prediction model;
The grouting diffusion dynamic prediction model comprises an input layer, a space-time modeling layer, a physical constraint integration layer, a feedback mechanism, a self-adaptive adjustment layer and an output layer.
8. The complex geologic multi-attribute constrained grouting diffusion prediction system of claim 7, wherein the grouting diffusion simulation experiment system comprises:
the complex geological simulation unit is used for simulating complex geology;
The grouting unit comprises a dynamic water circulation module and a grouting control module, wherein the dynamic water circulation module is used for simulating the groundwater seepage process, providing constant-pressure water flow and simulating a natural groundwater flow path in a model; the grouting control module is used for realizing accurate control of grouting speed, pressure and flow, and monitoring pressure and flow data in the grouting process through a multi-source sensor;
The data acquisition and monitoring unit comprises a pressure sensor, a flow sensor, a pore water pressure gauge, a resistivity imaging system and an ultrasonic sensor;
And the data processing unit is used for acquiring data and performing data processing.
9. The complex geologic multi-attribute constrained grouting diffusion predicting system of claim 7, wherein the complex geologic simulation unit comprises a model box filled with a plurality of rock formations, sand layers and soil layers and embedded with different geologic features of transverse and longitudinal fissures, sand layers, gravel layers, fault zones and karst cave to simulate complex geologic structures.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510058184.5A CN119989669B (en) | 2025-01-14 | 2025-01-14 | A grouting diffusion prediction method and system with complex geological multi-attribute constraints |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510058184.5A CN119989669B (en) | 2025-01-14 | 2025-01-14 | A grouting diffusion prediction method and system with complex geological multi-attribute constraints |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119989669A CN119989669A (en) | 2025-05-13 |
| CN119989669B true CN119989669B (en) | 2025-09-05 |
Family
ID=95636363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510058184.5A Active CN119989669B (en) | 2025-01-14 | 2025-01-14 | A grouting diffusion prediction method and system with complex geological multi-attribute constraints |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119989669B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120334019B (en) * | 2025-06-16 | 2025-09-16 | 天津市博川建设工程有限公司 | Indoor shear test method and system for the interface between mud mixed pile and paste slurry |
| CN120508175B (en) * | 2025-07-18 | 2025-09-30 | 山东科技大学 | Self-adaptive grouting regulation and plugging method based on real-time monitoring of mining fracture |
| CN120575912B (en) * | 2025-08-01 | 2025-10-03 | 中能化江苏矿山生态研究院有限公司 | Grouting water plugging method for coal mining area |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117852416A (en) * | 2024-03-07 | 2024-04-09 | 山东大学 | Multimode grouting precontrolled analysis method and system based on digital geological model |
| CN119290676A (en) * | 2024-09-27 | 2025-01-10 | 中国矿业大学 | A morphology prediction system and method for tunnel grouting diffusion |
-
2025
- 2025-01-14 CN CN202510058184.5A patent/CN119989669B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117852416A (en) * | 2024-03-07 | 2024-04-09 | 山东大学 | Multimode grouting precontrolled analysis method and system based on digital geological model |
| CN119290676A (en) * | 2024-09-27 | 2025-01-10 | 中国矿业大学 | A morphology prediction system and method for tunnel grouting diffusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119989669A (en) | 2025-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN119989669B (en) | A grouting diffusion prediction method and system with complex geological multi-attribute constraints | |
| RU2321064C2 (en) | Method for building a reversible three-dimensional hydrodynamic model of earth, calibrated in real time during drilling | |
| CN115062368B (en) | Digital twin body and fine modeling system and method for tunnel long pipe shed | |
| US11846175B2 (en) | Estimating reservoir production rates using machine learning models for wellbore operation control | |
| CN116911115A (en) | Digital twin tunnel and intelligent construction method and system | |
| US20160209546A1 (en) | Static earth model calibration methods and systems using tortuosity evaluations | |
| CN113742827B (en) | A Construction Method of Highway Slope Monitoring Network System Based on Finite Difference Analysis | |
| CN117787102A (en) | Disaster digital twin simulation method based on refined dynamic geological model combined with GEOAI | |
| CN119962262B (en) | A method and system for evaluating carbon dioxide storage capacity, and a computer-readable medium | |
| NO20180710A1 (en) | Methods of selecting an earth model from a plurality of earth models | |
| CN119992947B (en) | Virtual-real fusion collaborative test method and system for disaster grouting prevention and control in sandy and gravel strata | |
| AU2013397497A1 (en) | Static earth model calibration methods and systems using permeability testing | |
| CN116817777A (en) | Deformation prediction method of tunnel surrounding rock based on high-precision sensors and Transformer | |
| CN120410223B (en) | Underground construction decision-making method based on three-dimensional geological modeling and risk hot zone identification | |
| CN115308803A (en) | Coal seam thickness prediction method, device, equipment and medium | |
| CN119227498A (en) | Analysis method of time-dependent deformation mechanism of large-scale excavation high slope under ultra-deep unloading conditions | |
| CN120068619A (en) | Targeting positioning grouting high-efficiency plugging and reinforcing method and system based on deep learning | |
| CN118686142A (en) | A digital monitoring method for caisson construction | |
| CN116227335B (en) | Three-dimensional geostress prediction method based on parallel neural network with physical constraints | |
| CN119989668B (en) | Permeability prediction method and system based on multi-source attribute coupling | |
| CN119754828B (en) | A method and system for automatically aligning upper anchor rod equipment based on a lifting robot | |
| CN119375969B (en) | A method for testing formation rock and soil parameters based on a controlled neutron source, a ground acquisition computer, a system and a computer-readable storage medium | |
| CN120541927A (en) | A method, system, equipment and medium for intelligent construction of tunnels based on digital twins | |
| CN120781296A (en) | Logging-while-drilling data lag compensation method based on deep learning | |
| AU2025201663A1 (en) | Method and system for multi-scale attribute characterization and permeability hazard assessment of fractured rock mass in tunnel |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |