Anchoring rod self-adapting grouting control system

Abstract

The application discloses an anchor rod self-adaptive grouting control system, which comprises an environmental parameter collection module, an initial parameter prediction module and a grouting data collection module.The environmental parameter collection module is used for collecting construction environmental parameters and consolidation strength data.The initial parameter prediction module is used for inputting the construction environmental parameters and the consolidation strength data into a pre-constructed initial parameter prediction model and outputting initial parameters.The grouting data collection module is used for numbering N anchor rods, obtaining anchor rod numbers, and grouting the N anchor rods according to the initial parameters.The intelligent sensing collection network is established to collect monitoring data, and the monitoring data is summarized to a central adaptive control module.The application improves the efficiency and safety of grouting construction through integrated environmental parameter monitoring, intelligent data processing and adaptive control algorithm, thereby improving the overall quality and durability of the project.

Description

Technical Field

[0001] This application relates to the field of automation control technology, and in particular to an adaptive grouting control system for anchor bolts. Background Technology

[0002] The construction process of anchor bolt grouting generally includes the following: Based on the grouting process requirements, configure the appropriate grouting equipment and grouting materials. Inject the grouting material into the anchor bolt hole using a grouting pump. During the grouting process, it is necessary to control the grouting speed, grouting volume, and grouting pressure to ensure that the grouting material can uniformly fill the entire hole. After grouting, the consolidation time of the grouting material needs to be controlled to ensure that the grouting material can completely solidify and form a strong bond. After the grouting process is completed, the grouting effect is tested and evaluated.

[0003] However, currently, the grouting parameters of anchor bolt grouting technology, such as grouting pressure and grouting volume, still rely on manual experience for calculation. If intermittent grouting is to be carried out, the interval time also needs to be estimated based on manual experience. However, when dealing with multiple anchor bolts or grouting operations in batches, the flexibility and accuracy of manual experience often lead to problems with the efficiency and safety of grouting construction, resulting in huge trial and error costs for the execution of grouting design. Summary of the Invention

[0004] To address the aforementioned problems, embodiments of the present invention provide an adaptive grouting control system for anchor bolts, the system comprising:

[0005] An environmental parameter collection module is used to collect construction environmental parameters and consolidation strength data.

[0006] An initial parameter prediction module is used to input construction environment parameters and consolidation strength data into a pre-constructed initial parameter prediction model and output initial parameters.

[0007] The grouting data collection module is used to number N anchor bolts and obtain anchor bolt numbers; grouting is performed on the N anchor bolts according to initial parameters; an intelligent sensor collection network is established to collect monitoring data, and the monitoring data is aggregated to the central adaptive control module;

[0008] A central adaptive control module, which uses monitoring data to obtain grouting adjustment parameters;

[0009] The grouting control module is used to transmit the grouting adjustment parameters back to the anchor control unit of the corresponding anchor number through an intelligent sensor collection network. The anchor control unit adjusts the grouting parameters according to the grouting adjustment parameters.

[0010] The human-computer interaction module adopts a graphical user interface for manual operation.

[0011] Furthermore, the training method for the initial parameter prediction model includes:

[0012] All construction environment parameters and consolidation strength data are used as inputs to the initial parameter prediction model. The initial parameter prediction model outputs the predicted initial parameters for each set of construction environment parameters and consolidation strength data, and uses the actual initial parameters corresponding to each set of construction environment parameters and consolidation strength data as the prediction target. The training objective is to minimize the sum of the first prediction accuracies of all predicted initial parameters. The initial parameter prediction model is trained until the sum of the first prediction accuracies converges, at which point training stops. The initial parameter prediction model is a convolutional neural network model.

[0013] Furthermore, the intelligent sensing and data collection network includes N edge computing nodes and a central processing server. The central processing server receives monitoring data from the N edge computing nodes and filters the monitoring data using a Kalman filter joint model.

[0014] Furthermore, the modeling method for the Kalman filter joint model includes:

[0015] Construct the state equations, which include:

[0016] ;

[0017] In the formula, For the first Each anchor rod at time The state vector; This is the state transition matrix; To control the input matrix; For control input; The noise is the process noise, with a mean of 0 and a covariance of . Gaussian noise; and Number the anchor bolts, and 0 < ≤N, 0< ≤N;

[0018] Construct observation equations, which include:

[0019] ;

[0020] In the formula, For the first Observed values ​​of each anchor bolt; The observation matrix; To observe the noise, a Gaussian white noise model with a mean of 0 and a covariance of is used. .

[0021] Furthermore, the filtering method of the Kalman filter joint model includes:

[0022] Using the state prediction equation, we obtain the first... Each anchor rod at time The predicted state value;

[0023] Using the error covariance prediction equation, we obtain the first... Each anchor rod at time The prediction error covariance matrix;

[0024] By combining the covariance of prediction error, observation matrix, observation equation, and observation noise,

[0025] Obtain the Kalman gain;

[0026] Calculate the residuals, and then update the predicted values ​​using the state update equation to obtain the updated predicted state values. At the same time, update the error covariance using the error covariance update equation based on the Kalman gain.

[0027] Furthermore, the method for obtaining grouting adjustment parameters includes: dividing the monitoring data into N data zones based on the anchor bolt number; using an optimization algorithm to determine the target grouting volume for each anchor bolt; merging the target grouting volume with the data zone to which the anchor bolt number belongs to form a dedicated data zone; inputting the dedicated data zone into a pre-built adaptive control model; and outputting the grouting adjustment parameters.

[0028] Furthermore, the method for obtaining the target grouting volume includes:

[0029] Initialize the particle swarm, where each particle represents a target grouting volume combination and has a length of N;

[0030] Calculate the fitness value using the fitness function;

[0031] The particle position and velocity are updated using position and velocity update equations, and the calculation is performed iteratively. In each iteration, the fitness value of each particle is recalculated. The iteration terminates when a termination condition is met. The termination condition includes:

[0032] The maximum number of iterations is reached; the change in the global optimal fitness is less than a preset threshold.

[0033] Furthermore, the training method for the adaptive control model includes:

[0034] All data zones are used as input to the adaptive control model. The adaptive control model outputs the grouting adjustment parameters predicted for each data zone and uses the actual grouting adjustment parameters corresponding to each data zone as the prediction target. The training objective is to minimize the sum of the second prediction accuracies of all predicted grouting adjustment parameters. The adaptive control model is trained until the sum of the second prediction accuracies converges, at which point training stops. The adaptive control model is a support vector machine.

[0035] The technical effects and advantages of the adaptive grouting control system for anchor bolts provided by this invention are as follows:

[0036] This invention improves the efficiency and safety of grouting construction by integrating environmental parameter monitoring, intelligent data processing, and adaptive control algorithms, thereby enhancing the overall quality and durability of the project. Real-time monitoring of construction environmental parameters (such as temperature, humidity, air pressure, and soil type) ensures that the grouting process adapts to actual construction conditions. Different environmental factors significantly affect the performance of grouting materials; accurate data collection helps adjust construction parameters, thus improving grouting quality. By inputting construction environmental parameters and consolidation strength data into the initial parameter prediction model, the required grouting volume, pressure, and mix ratio for anchor bolts before construction begins can be efficiently estimated, ensuring compliance with project requirements from the initial stage. This approach lays a solid foundation for subsequent construction. Continuous monitoring of each anchor bolt's status via an intelligent sensor network enables rapid acquisition of real-time data. This data is then analyzed by a central adaptive control module to provide precise grouting adjustment parameters. The data-driven decision-making mechanism effectively addresses potential changes and uncertainties during construction. Training and updating the system using machine learning and optimization algorithms continuously improves its predictive accuracy and adaptability. Introducing a Kalman filter joint model to handle data interactions between multiple anchor bolts allows for a more comprehensive consideration of their mutual influences, providing more accurate state estimation and decision support when dealing with noise and uncertainties. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the connection of an adaptive grouting control system for anchor bolts in Embodiment 1;

[0038] Figure 2 This is a schematic diagram of the intelligent sensor collection network connection in Example 2; Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] Example 1:

[0041] Please see Figure 1 As shown in this embodiment, an adaptive grouting control system for anchor bolts includes:

[0042] The environmental parameter collection module collects construction environmental parameters and consolidation strength data.

[0043] The initial parameter prediction module inputs construction environment parameters and consolidation strength data into the pre-built initial parameter prediction model and outputs the initial parameters.

[0044] The grouting data collection module numbers the N anchor bolts to obtain their numbers; the N anchor bolts are grouted according to the initial parameters; an intelligent sensor collection network is established to collect monitoring data, and the monitoring data is aggregated to the central adaptive control module.

[0045] The central adaptive control module uses monitoring data to obtain grouting adjustment parameters;

[0046] The grouting control module transmits the grouting adjustment parameters back to the anchor control unit corresponding to the anchor bolt number through an intelligent sensor collection network. The anchor control unit then adjusts the grouting parameters according to the grouting adjustment parameters.

[0047] The human-computer interaction module uses a graphical user interface for manual operation.

[0048] The graphical user interface (GUI) displays real-time monitoring data, system status, and grouting parameters, helping users quickly understand and operate the system. The human-computer interaction module serves as an input window for preset values, allowing users to manually input or adjust certain parameters (such as target grouting volume and construction environment parameters) and feed this information back to the system, prompting model updates and optimizations.

[0049] Construction environment parameters include temperature, humidity, air pressure, and preset soil type. Temperature and humidity are measured using sensors. The temperature of the environment and soil affects the rheological properties and curing time of the grouting material. Air humidity and soil moisture content affect the performance of the grouting material. Air pressure is mainly used to determine the air pressure range. The impact of low altitude and high altitude on construction conditions is different. Preset soil types include clay, sand, and gravel, etc. Different soil types also have different effects on the permeability of the grouting material. Consolidation strength data includes consolidation strength, soil void ratio, and pore water pressure. Consolidation strength data is directly related to the bearing capacity and stability of the anchor. Consolidation strength is the shear strength of the soil under certain conditions, used to characterize the stability of the soil. Soil void ratio is the ratio of the volume of soil pores to the volume of solid particles, affecting the soil consolidation process and strength. Pore water pressure is the pressure of water inside the soil, which is crucial to the consolidation and bearing capacity. It can be monitored using a pore water pressure gauge (such as a piezoelectric pressure gauge).

[0050] Training methods for initial parameter prediction models include:

[0051] All construction environment parameters and consolidation strength data are used as inputs to the initial parameter prediction model. The initial parameter prediction model outputs the predicted initial parameters for each set of construction environment parameters and consolidation strength data, and uses the actual initial parameters corresponding to each set of construction environment parameters and consolidation strength data as the prediction target. The training objective is to minimize the sum of the first prediction accuracies of all predicted initial parameters.

[0052] The formula for calculating the first prediction accuracy is as follows: ,in, Each set of construction environment parameters and consolidation strength data is assigned a number. For the highest prediction accuracy, For the first The initial parameters for prediction corresponding to construction environmental parameters and consolidation strength data. For the first The actual initial parameters corresponding to the construction environment parameters and consolidation strength data are used to train the initial parameter prediction model until the sum of the first prediction accuracies reaches convergence. The initial parameter prediction model is a convolutional neural network model.

[0053] The initial parameters include grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time. The grout mix ratio is a fixed ratio, which can be the predicted value output by the initial parameter prediction model or a preset value. If there is no preset value with a fixed ratio, the initial parameters are the predicted values ​​output by the initial parameter prediction model. If there is a preset value with a fixed ratio, the initial parameters are the preset values ​​with a fixed ratio. The interval time is because grouting needs to be intentionally interrupted when performing intermittent grouting.

[0054] It should be noted that the initial parameters are range values, which serve as a safe range before the start of anchor grouting construction. Correct initial parameters can ensure grouting quality, enhance the anchoring effect of anchors, improve construction safety and project durability. Only after grouting will a unified adjustment be made for each anchor.

[0055] The method for obtaining grouting adjustment parameters includes: dividing the monitoring data into N data zones based on the anchor bolt number; using an optimization algorithm to determine the target grouting volume for each anchor bolt; merging the target grouting volume with the data zone to which the anchor bolt number belongs into a dedicated data zone; inputting the dedicated data zone into a pre-built adaptive control model; and outputting the grouting adjustment parameters.

[0056] The data area includes grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time;

[0057] By dividing the monitoring data into multiple data zones, it is possible to better analyze and process the characteristics of different anchor bolts. Each anchor bolt may exhibit different behaviors due to its specific location, construction conditions, or environment, so this division can help optimize the grouting volume and adjust parameters.

[0058] The data area includes grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time;

[0059] The data section includes target grouting volume, grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time.

[0060] The construction of the data zone makes the data input to the adaptive control model more comprehensive and richer, thus allowing the model to better understand the relationship between the target state and the actual state of each anchor bolt; without the integration of target grouting volume, the model may not be able to fully evaluate the anchor bolt performance and adjust the parameters required.

[0061] The data section includes target grouting volume, grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time.

[0062] The monitoring data includes grouting volume, grouting speed, grouting pressure, grout mix ratio, and interval time.

[0063] It should be noted that the initial parameters are range values, while the monitoring data are precise values. The monitoring data is collected after grouting and is measured during actual operation, hence it is a precise value.

[0064] N is the preset number of anchor bolts, representing the number of anchor bolts in the working state.

[0065] Methods for obtaining the target grouting volume include:

[0066] Initialize the particle swarm, where each particle represents a target grouting volume combination and has a length of N;

[0067] Calculate fitness values ​​using the fitness function. ;

[0068] Fitness functions include:

[0069] ;

[0070] In the formula, Number the anchor bolts. Number the particles, and 0 < ≤N, 0< ≤N; For the first The weight of each anchor bolt is preset according to construction requirements or importance.

[0071] The target state refers to the ideal anchor bolt performance indicators set based on engineering requirements or design standards. It is typically the desired state to be achieved, including:

[0072] Displacement: The maximum allowable displacement of the anchor bolt after grouting under preset conditions.

[0073] Stress: The maximum stress that the anchor bolt should withstand under preset conditions.

[0074] Grouting pressure: Under preset conditions, the optimal grouting pressure is achieved, thereby improving the consolidation strength and load-bearing capacity of the anchor bolt.

[0075] Grouting volume: Under preset conditions, the grouting volume reaches the target grouting volume, which is the preset value.

[0076] Stability index: The requirements for the interaction force between the anchor bolt and the surrounding soil.

[0077] Actual condition refers to the performance indicators of the anchor bolt under actual on-site operating conditions obtained through monitoring equipment, reflecting the true performance of the anchor bolt during actual construction or use.

[0078] The purpose of the optimization algorithm is to adjust the grouting amount of each anchor rod to bring the actual state close to or reach the target state.

[0079] The particle position and velocity are updated using position and velocity update equations, and the calculation is performed iteratively. In each iteration, the fitness value of each particle is recalculated. The iteration terminates when a termination condition is met. The termination condition includes:

[0080] The maximum number of iterations is reached; the change in the global optimal fitness is less than the preset threshold (i.e., convergence).

[0081] It should be noted that the termination condition can be terminated if one of the termination conditions is met; it is not mandatory that both conditions be met.

[0082] The location update equations include:

[0083] ;

[0084] In the formula, For the first The particle in the first Position in the next iteration; This represents the number of iterations. For the first The particle in the first Speed ​​in the next iteration.

[0085] The velocity update equations include:

[0086] ;

[0087] In the formula, Inertial weights control the degree to which the particle's historical velocity affects its current velocity; and The learning factor, usually a positive value, represents the particle's path to its optimal position. and global best position Learning inclination; and is a random number in the range [0, 1], used to increase the randomness and diversity of the search.

[0088] The reason for confirming the grouting volume of a single section first is that determining the grouting volume in advance can reduce the trial and error costs of adjustments during construction, minimize the impact of subsequent adjustments, and analyze the relationship between grouting volume and effect data, which helps to make more accurate subsequent plans and also provides the best human-computer interaction effect.

[0089] Training methods for adaptive control models include:

[0090] All data zones are used as input to the adaptive control model. The adaptive control model outputs the grouting adjustment parameters predicted for each data zone, uses the actual grouting adjustment parameters corresponding to each data zone as the prediction target, and minimizes the sum of the second prediction accuracies of all predicted grouting adjustment parameters as the training objective.

[0091] The formula for calculating the second prediction accuracy is as follows: ,in, Each data section is assigned a unique number. For the second prediction accuracy, For the first The data zone corresponds to the predicted grouting adjustment parameters. For the first The actual grouting adjustment parameters corresponding to the data set area; the adaptive control model is trained until the sum of the second prediction accuracies reaches convergence and training stops; the adaptive control model is a support vector machine.

[0092] Adaptive control models can adjust in real time based on new input data, enhancing the system's sensitivity to environmental changes. For example, if the actual grouting effect monitored does not match expectations, the model can quickly adjust prediction parameters to reflect the new construction status.

[0093] An anchor bolt control unit is a device or module specifically designed to control and adjust the anchor bolt grouting process. By receiving grouting adjustment parameters from a central adaptive control module, it controls crucial parameters related to the anchor bolt, such as grouting volume, pressure, and flow rate, ensuring that the anchor bolt meets predetermined performance standards during construction. Upon receiving instructions from the central adaptive control module, the anchor bolt control unit can execute specific operations, such as starting or stopping the grouting process, and adjusting the grouting pressure or flow rate. It translates system decisions into actual operations, making the grouting process more automated and efficient.

[0094] Example 2:

[0095] like Figure 2 As shown, the intelligent sensing and data collection network includes N edge computing nodes and a central processing server. Various sensor devices (such as pressure sensors, flow meters, and other high-precision instruments) are attached to the anchor bolts to acquire multi-dimensional data about the anchor bolts and their surrounding environment. Each anchor bolt serves as an edge computing node. The edge computing nodes and the central processing server transmit data using low-power wide-area network technology or 5G technology. The central processing server is connected to a human-machine interface module. The central processing server receives monitoring data from the N edge computing nodes and filters the data. Given that there are N edge computing nodes, there may be interference between them, and a single Kalman filter may not be sufficient to effectively handle the correlation between the states of each anchor bolt. Therefore, a joint Kalman filter model is used for filtering.

[0096] Modeling methods for the Kalman filter joint model include:

[0097] Constructing the state equations:

[0098] For the There are several anchor rods, and the state vector of each anchor rod is assumed to be... This includes quantities such as pressure, flow rate, and displacement. Assuming that the states of the N anchors influence each other, this influence is defined by the state transition matrix, and the state equations include:

[0099] ;

[0100] In the formula, For the first Each anchor rod at time The state vector; Let be the state transition matrix, describing the state transition matrix. The anchor bolt is paired with the first The impact of individual anchor bolts; The control input matrix describes the control input pair for the th... The influence of the state of individual anchor bolts; To control inputs (such as external stress); The noise is assumed to be process noise, with a mean of 0 and a covariance of . Gaussian noise; and Number the anchor bolts, and 0 < ≤N, 0< ≤N;

[0101] Construct the observation equation:

[0102] The readings of each anchor are also affected by the readings of other anchors, and the observation equations include:

[0103] ;

[0104] In the formula, For the first Observed values ​​of each anchor bolt; Let be the observation matrix, describing the first... The state of the anchor bolt for the first... The impact of individual anchor bolt observations; To observe the noise, a Gaussian white noise model is used, assuming a mean of 0 and a covariance of . .

[0105] The filtering methods of the Kalman filter joint model include:

[0106] Using the state prediction equation, we obtain the first... Each anchor rod at time Predicted state values The state prediction equations include:

[0107] ;

[0108] In the formula, For the first Each anchor rod at time The latest state estimate (updated state) is used to calculate the predicted state.

[0109] Using the error covariance prediction equation, we obtain the first... Each anchor rod at time Prediction error covariance matrix The error covariance prediction equation includes:

[0110] ;

[0111] In the formula, For the first Each anchor rod at time The error covariance matrix; This is the transpose symbol.

[0112] The Kalman gain is obtained by combining the covariance of the prediction error, the observation matrix, the observation equation, and the covariance of the observation noise. The calculation methods include:

[0113] ;

[0114] Calculate the residuals (i.e., the difference between the observed and predicted values). After obtaining the residuals, update the predicted values ​​using the state update equation to obtain the updated predicted state values. Simultaneously, the error covariance is updated using the error covariance update equation based on the Kalman gain. .

[0115] The state update equations include:

[0116] ;

[0117] The error covariance update equation includes:

[0118] ;

[0119] In the formula, It is an identity matrix.

[0120] The Kalman filter joint model can estimate the state of multiple interdependent anchors, comprehensively considering the interference and correlation between the anchors. It uses linear state equations and observation equations, and optimizes the overall system estimate through prediction and update mechanisms, especially performing well in the presence of noise and uncertainty.

[0121] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0122] The above description is merely a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present application, based on the technical solution and concept of the present application, should be covered within the scope of protection of the present application.

Claims

1. An adaptive grouting control system for anchor bolts, characterized in that, include: An environmental parameter collection module is used to collect construction environmental parameters and consolidation strength data. An initial parameter prediction module is used to input construction environment parameters and consolidation strength data into a pre-constructed initial parameter prediction model and output initial parameters. Grouting data collection module, the grouting data collection module is used to number N anchor rods and obtain anchor rod numbers; N anchor bolts were grouted according to the initial parameters; Establish an intelligent sensor collection network to collect monitoring data and aggregate the monitoring data to the central adaptive control module; A central adaptive control module, which uses monitoring data to obtain grouting adjustment parameters; The method for obtaining grouting adjustment parameters includes: dividing the monitoring data into N data zones based on the anchor bolt number; using an optimization algorithm to determine the target grouting volume for each anchor bolt; merging the target grouting volume with the data zone to which the anchor bolt number belongs to form a dedicated data zone; inputting the dedicated data zone into a pre-built adaptive control model; and outputting grouting adjustment parameters. The grouting control module is used to transmit the grouting adjustment parameters back to the anchor control unit of the corresponding anchor number through an intelligent sensor collection network. The anchor control unit adjusts the grouting parameters according to the grouting adjustment parameters. The human-computer interaction module adopts a graphical user interface for manual operation.

2. The adaptive grouting control system for anchor bolts as described in claim 1, characterized in that, The training method for the initial parameter prediction model includes: All construction environment parameters and consolidation strength data are used as inputs to the initial parameter prediction model. The initial parameter prediction model outputs the predicted initial parameters for each set of construction environment parameters and consolidation strength data, and uses the actual initial parameters corresponding to each set of construction environment parameters and consolidation strength data as the prediction target. The training objective is to minimize the sum of the first prediction accuracies of all predicted initial parameters. The initial parameter prediction model is trained until the sum of the first prediction accuracies converges, at which point training stops. The initial parameter prediction model is a convolutional neural network model.

3. The adaptive grouting control system for anchor bolts as described in claim 1, characterized in that, The intelligent sensing and data collection network includes N edge computing nodes and a central processing server. The central processing server receives monitoring data from the N edge computing nodes and filters the monitoring data using a Kalman filter joint model.

4. The adaptive grouting control system for anchor bolts as described in claim 3, characterized in that, The modeling method for the Kalman filter joint model includes: Construct the state equations, which include: ； In the formula, For the first Each anchor rod at time The state vector; This is the state transition matrix; To control the input matrix; For control input; The noise is the process noise, with a mean of 0 and a covariance of . Gaussian noise; and Number the anchor bolts, and 0 < ≤N, 0< ≤N; Construct observation equations, which include: ； In the formula, For the first Observations of each anchor bolt; The observation matrix; To observe the noise, a Gaussian white noise model with a mean of 0 and a covariance of is used. .

5. The adaptive grouting control system for anchor bolts as described in claim 4, characterized in that, The filtering method of the Kalman filter joint model includes: Using the state prediction equation, we obtain the first... Each anchor rod at time The predicted state value; Using the error covariance prediction equation, we obtain the first... Each anchor rod at time The prediction error covariance matrix; By combining the covariance of prediction error, observation matrix, observation equation, and observation noise, Obtain the Kalman gain; Calculate the residuals, and then update the predicted values ​​using the state update equation to obtain the updated predicted state values. At the same time, update the error covariance using the error covariance update equation based on the Kalman gain.

6. The adaptive grouting control system for anchor bolts as described in claim 1, characterized in that, The method for obtaining the target grouting volume includes: Initialize the particle swarm, where each particle represents a target grouting volume combination and has a length of N; Calculate the fitness value using the fitness function; The particle position and velocity are updated using position and velocity update equations, and the calculation is performed iteratively. In each iteration, the fitness value of each particle is recalculated. The iteration terminates when a termination condition is met. The termination condition includes: The maximum number of iterations is reached; the change in the global optimal fitness is less than a preset threshold.

7. The adaptive grouting control system for anchor bolts as described in claim 1, characterized in that, The training method for the adaptive control model includes: All data zones are used as input to the adaptive control model. The adaptive control model outputs the grouting adjustment parameters predicted for each data zone and uses the actual grouting adjustment parameters corresponding to each data zone as the prediction target. The training objective is to minimize the sum of the second prediction accuracies of all predicted grouting adjustment parameters. The adaptive control model is trained until the sum of the second prediction accuracies converges, at which point training stops. The adaptive control model is a support vector machine.

Images