Intelligent grouting control method and system

Abstract

The invention relates to the technical field of grouting control, in particular to an intelligent grouting control method and system.The intelligent grouting control method comprises the steps that multi-parameter time sequence data in the grouting process is collected in real time, and standardization processing is conducted to construct feature vectors; inputting the feature vectors into an integrated learning model, and synchronously executing grouting effect prediction based on time sequence analysis and abnormal working condition identification based on data distribution analysis; when an abnormal working condition is recognized, a control strategy corresponding to the abnormity is executed preferentially; when abnormity is not recognized, the control parameters are dynamically optimized according to the predicted grouting effect; issuing the control strategy or the optimization control parameter to grouting equipment for execution, and collecting feedback data after execution; on the basis of feedback data, online incremental learning is carried out on the integrated learning model, model parameters are adaptively optimized, prediction can be carried out in advance, collected data and final control form a closed loop, meanwhile, the method can also adapt to complex working conditions, and the intelligent grouting control efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of grouting control technology, and in particular to a smart grouting control method and system. Background Technology

[0002] In grouting construction in civil engineering, geological disaster management, and other fields, precise process control is crucial to ensuring project quality (such as the density and uniformity of the grout). Traditional grouting control methods mainly rely on the experience of construction personnel to set fixed parameters or on alarms based on simple thresholds. With technological advancements, solutions have emerged that use sensors to collect and monitor data.

[0003] However, these existing solutions often only provide post-event alarms and cannot perform predictive control before anomalies occur; they also cannot simultaneously address the conflicting goals of quality optimization and safety early warning, and make priority decisions; they are difficult to adapt to the complex and ever-changing geological conditions of different construction sites, resulting in poor universality and long-term attenuation of control effects. Summary of the Invention

[0004] This invention provides a smart grouting control method and system to address the shortcomings of existing grouting control systems that rely solely on a single sensor, making it impossible to predict in advance, adjust quality and early warning priorities, or adapt to complex working conditions.

[0005] On one hand, the present invention provides a smart grouting control method, comprising: Real-time acquisition of multi-parameter time-series data during the grouting process, followed by standardization processing to construct feature vectors; The feature vector is input into the ensemble learning model to simultaneously perform grouting effect prediction based on time series analysis and abnormal working condition identification based on data distribution analysis. When an abnormal operating condition is detected, the control strategy corresponding to the abnormality is executed first; when no abnormality is detected, the control parameters are dynamically optimized based on the predicted grouting effect. The control strategy or the optimized control parameters are sent to the grouting equipment for execution, and feedback data is collected after execution. Based on the feedback data, the ensemble learning model is subjected to online incremental learning to adaptively optimize the model parameters.

[0006] According to the intelligent grouting control method provided by the present invention, the step of inputting the feature vector into an ensemble learning model and simultaneously performing grouting effect prediction based on time series analysis and abnormal working condition identification based on data distribution analysis includes: The feature vectors are simultaneously input into a long short-term memory network and an isolated forest algorithm; The long short-term memory network is used to perform time-series analysis on the feature vector to output a predicted value of grouting effect. The isolated forest algorithm is used to analyze the data distribution of the feature vectors and output an abnormal operating condition identification signal.

[0007] According to the intelligent grouting control method provided by the present invention, the step of prioritizing the execution of the control strategy corresponding to the anomaly includes: The abnormal operating condition identification signal is analyzed to determine the abnormal type; Based on the preset mapping relationship between anomaly types and control strategies, the target control strategy corresponding to the anomaly type is invoked. The target control strategy includes at least an adjustment instruction for grouting pressure or flow rate. The current parameter optimization process is interrupted, and the target control strategy is sent to the grouting equipment for execution.

[0008] According to a smart grouting control method provided by the present invention, when no abnormality is detected, the control parameters are dynamically optimized based on the predicted grouting effect, including: The predicted grouting effect is compared with the preset target effect value, and the effect deviation is calculated; Based on the magnitude and direction of the effect deviation, and using a preset optimization rule base, the adjustment direction and magnitude of the grouting pressure or flow rate are determined. Based on the adjustment direction and amplitude, new control parameters are generated and sent to the grouting equipment for execution, so that the grouting effect approaches the target effect value.

[0009] According to a smart grouting control method provided by the present invention, the step of sending the control strategy or the optimized control parameters to the grouting equipment for execution includes: The control strategy or optimized control parameters are encapsulated into standard control commands that can be recognized by the underlying drive system of the grouting equipment. The standard control commands are sent to the corresponding grouting equipment actuators via industrial communication protocols. The status feedback signal from the actuator is received in real time to confirm that the control command has been successfully executed.

[0010] According to the intelligent grouting control method provided by the present invention, after real-time acquisition of multi-parameter time-series data during the grouting process, the method further includes: Analyze the noise characteristics of data from different sensors in a specific frequency band to identify noise sources; Dynamically select the corresponding digital filter type and parameters based on the noise source; The original sensor data is processed using the selected digital filter type and parameters, and the signal-to-noise ratio of the filtered data is verified.

[0011] According to the intelligent grouting control method provided by the present invention, the step of using the long short-term memory network to perform time-series analysis on the feature vector and outputting a predicted value of grouting effect further includes: The geological conditions of the current construction area are digitally coded. The encoded geological condition identifier vector is fused with the real-time acquired temporal feature vector to form an extended feature vector; The extended feature vector is input into the long short-term memory network to adaptively adjust the dependence weights on the geological background.

[0012] According to the intelligent grouting control method provided by the present invention, after collecting the feedback data after execution, it further includes: Collect the actual operating parameters of the grouting equipment actuator and compare them with the issued control commands to form a comparison data set; Calculate the continuous deviation between the actual operating parameters and the control commands; If the continuous deviation exceeds the preset allowable range, an equipment health status warning is generated, and a compensation value for the control command is calculated based on the continuous deviation to trigger system self-calibration.

[0013] According to the intelligent grouting control method provided by the present invention, the step of performing online incremental learning on the ensemble learning model based on the feedback data and adaptively optimizing the model parameters includes: The feedback data is correlated with the actual grouting effect evaluation indicators to generate new training samples with labels; Calculate the difference measure between the new training samples and the model's existing knowledge base, and select a subset of samples; The mini-batch gradient descent method is used to incrementally update the model parameters using the aforementioned sample subset. Verify the performance of the model after the incremental update, and adjust the learning rate parameter for subsequent incremental learning based on the verification results.

[0014] Secondly, the present invention provides a smart grouting control system, comprising: The acquisition module is used to acquire multi-parameter time-series data in real time during the grouting process and perform standardization processing to construct feature vectors; The execution module is used to input the feature vector into the ensemble learning model and simultaneously execute the grouting effect prediction based on time series analysis and the abnormal working condition identification based on data distribution analysis. When an abnormal working condition is identified, the control strategy corresponding to the abnormality is executed first. When no abnormality is identified, the control parameters are dynamically optimized according to the predicted grouting effect. The distribution module is used to distribute the control strategy or the optimized control parameters to the grouting equipment for execution, and to collect feedback data after execution; The feedback module is used to perform online incremental learning on the ensemble learning model based on the feedback data, and adaptively optimize the model parameters.

[0015] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the intelligent grouting control method as described above.

[0016] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent grouting control method as described above.

[0017] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the intelligent grouting control method as described above.

[0018] This invention provides a smart grouting control method and system that integrates data acquisition, synchronous prediction and identification, priority control, execution feedback, and online learning into a coherent automated process, forming a complete intelligent closed loop from perception to evolution. It utilizes an ensemble learning model to synchronously output prediction and identification results and makes decisions based on anomaly-priority rules, improving the reliability and efficiency of control. Through an online incremental learning mechanism, it can continuously evolve based on field feedback data, automatically adapting to changes in different geological conditions and equipment states. This solves the problem of poor universality of fixed models, ensuring long-term stable control effects and significantly reducing reliance on human experience. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the intelligent grouting control method provided in this embodiment; Figure 2 This is a schematic diagram of the intelligent grouting control system provided in this embodiment. Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Figure 1 This is a flowchart illustrating the intelligent grouting control method provided in this embodiment.

[0023] like Figure 1 As shown in the figure, this embodiment provides a smart grouting control method, taking a civil engineering curtain grouting construction scenario (for a geological area with alternating layers of deep silty sand and gravel) as an example, and will describe it in detail. The method mainly includes the following steps: 101. Real-time acquisition of multi-parameter time-series data during the grouting process, and standardization processing to construct feature vectors.

[0024] Specifically, a sensor network is deployed at key nodes of the grouting system, including: pressure sensors at the grouting pump outlet and grouting borehole (collecting instantaneous and average pressure), electromagnetic flow sensors in the grout delivery pipeline (collecting instantaneous and cumulative flow), density sensors at the grout mixing tank and delivery end (collecting grout density), borehole depth sensors (collecting the current grouting section depth), and ambient temperature sensors (collecting the construction ambient temperature). All sensors collect data synchronously at a preset sampling frequency to form a multi-parameter time-series data stream.

[0025] Spectral analysis was used to decompose the data from each sensor into frequency bands to identify noise characteristics: if high-frequency spikes appeared in the mechanical vibration characteristic frequency band, the noise source was identified as the mechanical vibration of the grouting pump; if regular fluctuations appeared in the electromagnetic interference frequency band, it was identified as electromagnetic interference from the on-site power distribution box; if the data showed irregular random fluctuations covering the entire frequency band, it was identified as sensor noise. A dynamic matching filtering scheme was applied based on the noise source type: an adaptive Kalman filter was used for mechanical vibration noise, a Notch filter for electromagnetic interference, and a wavelet threshold filter for sensor noise, while adjusting parameters such as filter order and cutoff frequency. After filtering, the effect was verified by calculating the signal power to noise power ratio (SNR) to ensure that the SNR was improved above the preset threshold, thus avoiding noise data misleading subsequent decisions.

[0026] The denoised data is normalized to eliminate the influence of different units such as pressure (MPa), flow rate (L / min), and density (g / cm³). The time series segments of each parameter are extracted and spliced ​​according to a fixed time window to construct a time series feature vector with uniform dimensions.

[0027] 102. Input the feature vector into the ensemble learning model and simultaneously perform grouting effect prediction based on time series analysis and abnormal working condition identification based on data distribution analysis.

[0028] Specifically, the geological survey data of the current construction area is analyzed to extract core geological parameters: soil type (cohesive soil, sandy soil, gravelly soil), porosity (high / medium / low), permeability (strong / medium / weak), and groundwater level depth. One-heat coding is used to encode the classification parameters (soil type, permeability), and numerical mapping is used to encode the ordered parameters (porosity: high=3, medium=2, low=1), generating a geological condition identifier vector of dimension N. This vector is then concatenated column-wise with the temporal feature vector to form an extended feature vector, enabling the model to adapt to geological differences.

[0029] The ensemble learning model comprises two parallel modules: a Long Short-Term Memory (LSTM) network and an Isolation Forest algorithm. It expands the feature vector through multi-threaded synchronous processing. The LSTM network adopts an "input layer - 2 hidden layers - output layer" architecture. The input layer receives the expanded feature vector, and the hidden layer learns temporal correlations (such as the lag relationship between continuous pressure increase and density increase, and the mapping law between flow fluctuation period and grout diffusion range) through gating units (input gate, forget gate, and output gate). It also adaptively adjusts the weights according to the geological condition identifier vector (such as increasing the weight of permeability parameter in gravel layers and increasing the weight of porosity parameter in cohesive soil). Finally, it outputs predicted values ​​of grouting effect such as the density and uniformity of the reinforced body.

[0030] The Isolation Forest algorithm constructs multiple isolated trees by randomly selecting features and split points, recursively isolating the extended feature vectors, and calculating the anomaly score for each sample (the higher the score, the more significant the deviation from the normal distribution). A preset anomaly score threshold is used; when the score exceeds the threshold, a signal with an anomaly identifier is output, including the time point of the anomaly and related parameters (e.g., "t=10min, pressure surge, related parameters: pressure / flow rate"). If the score does not exceed the threshold, no anomaly signal is output.

[0031] The model coordination module aggregates LSTM predictions and isolated forest identification signals to ensure the timing alignment of the outputs from both tasks, providing synchronous data support for subsequent decision-making.

[0032] 103. When an abnormal working condition is identified, the control strategy corresponding to the abnormality shall be executed first; when no abnormality is identified, the control parameters shall be dynamically optimized based on the predicted grouting effect.

[0033] Specifically, if an anomaly detection signal is received, the anomaly handling process is immediately triggered: The signal analysis module extracts the associated parameters and change characteristics from the anomaly identifiers and matches them with a preset anomaly type library: sudden pressure increase (rate of change > threshold) and sudden flow decrease → pipeline blockage; sudden flow decrease and stable density → slurry supply interruption; sudden pressure decrease and sudden flow increase → formation slurry leakage; periodic fluctuations in pressure / flow → equipment mechanical failure.

[0034] The system queries the "Abnormality Type - Control Strategy" mapping table and invokes the target strategy: For pipeline blockage, it issues the command "Pressure drops to safety threshold + grouting pauses + backflushing for 10 seconds"; for slurry leakage in the formation, it issues the command "Reduce pressure by 20% + reduce flow rate by 30% + adjust slurry density". Simultaneously, an interrupt signal is triggered to pause the current parameter optimization process and lock the priority of the abnormality handling commands.

[0035] If no anomalies are identified, execute the quality optimization process: The predicted density and uniformity output by the LSTM are compared with the preset target values ​​(such as density ≥90% and uniformity coefficient of variation ≤15%), and the deviation value (deviation = predicted value - target value) and the deviation direction (positive deviation / negative deviation) are calculated.

[0036] Based on the preset optimization rule base, the decision adjustment scheme is as follows: When the absolute value of the negative deviation in density is >5%, increase the pressure (the adjustment range is positively correlated with the absolute value of the deviation) + fine-tune the flow rate (increase by 5%-10%); when the positive deviation in density is <3% and the uniformity meets the standard, slightly reduce the pressure (within 5%) to avoid over-grouting; when the uniformity has a negative deviation (coefficient of variation >15%), maintain stable pressure and optimize grout diffusion by pulse-type flow rate adjustment. Finally, optimized control parameters containing pressure, flow rate, and density adjustment values ​​are generated.

[0037] 104. Send the control strategy or optimized control parameters to the grouting equipment for execution, and collect feedback data after execution.

[0038] Specifically, the abnormal control strategy or optimization parameters are converted into a format recognizable by the underlying driver of the grouting equipment. The encapsulated content includes: instruction ID (unique identifier), execution parameters (such as pressure value, flow rate value), execution duration, and checksum (generated based on CRC32 algorithm). The encapsulation format is adapted according to the equipment communication protocol (Modbus / Profinet) to ensure instruction compatibility.

[0039] The encapsulation commands are sent to the corresponding actuators via industrial Ethernet (or RS485 bus, adapted for older equipment): pressure commands are sent to the grouting pump frequency converter, flow commands are sent to the electromagnetic flow regulating valve, and flushing commands are sent to the backflushing pump. After receiving the commands, the actuators drive the mechanical components to move (such as adjusting the pump speed and valve opening) through the hardware interface.

[0040] The actuator transmits status feedback signals in real time, including: command reception status (success / failure), execution progress (0%-100%), current actual operating parameters (real-time pressure, flow rate), and fault indicators (such as valve jamming, pump overload). The parsing module verifies the feedback signals: if reception fails, it triggers command retransmission (up to 3 times); if execution is abnormal, it generates a device fault warning; if the parameters match and execution is completed, it confirms successful command execution.

[0041] 105. Based on feedback data, perform online incremental learning on the ensemble learning model to adaptively optimize the model parameters.

[0042] Specifically, the actual operating parameters of the actuator are collected and compared with the issued command parameters to form a comparison data set, and the continuous deviation over 10 consecutive sampling periods is calculated (mean absolute deviation = Σ|actual value - command value| / 10).

[0043] If the continuous deviation exceeds the allowable range of equipment accuracy (e.g., pressure ±0.1MPa, flow rate ±1L / min), an equipment health warning is generated (marking possible causes such as "pressure sensor drift" or "valve response lag"); at the same time, a proportional-integral (PI) algorithm is used to calculate the compensation value (compensation value = proportional coefficient × deviation + integral coefficient × cumulative deviation), which is sent to the actuator to complete self-calibration and correct the control deviation.

[0044] The feedback data (execution parameters, equipment status, deviation values) are correlated with the actual measured performance indicators on site (such as the density of core drilling and the permeability of water pressure test) to label the samples with "performance labels" (such as "density 92%, no abnormality" and "density 85%, pipeline blockage").

[0045] Calculate the cosine similarity between the new sample and the historical samples in the model's knowledge base, and select samples with a similarity of <0.7 to form a subset (to ensure sample representativeness); use mini-batch gradient descent to input the subset into the model in batches, update the LSTM's gate weights and the tree structure parameters of the isolated forest, and calculate the loss function (such as mean squared error and anomaly detection accuracy) after each batch update until the loss converges.

[0046] Use a reserved test set to verify the performance of the updated model (prediction accuracy ≥ 90% and anomaly recall ≥ 85% are considered satisfactory). If the target is met, maintain the current learning rate. If the accuracy decreases, reduce the learning rate by 20%. If the recall is insufficient, increase the sample subset size to ensure that the model continues to adapt to changes in the field.

[0047] This embodiment integrates data acquisition, synchronous prediction and identification, priority control, execution feedback, and online learning into a coherent automated process, forming a complete intelligent closed loop from perception to evolution. It utilizes an ensemble learning model to synchronously output prediction and identification results and makes decisions based on anomaly-priority rules, improving the reliability and efficiency of control. Through an online incremental learning mechanism, it can continuously evolve based on field feedback data, automatically adapting to changes in different geological conditions and equipment states. This solves the problem of poor universality of fixed models, ensuring long-term stable control performance and significantly reducing reliance on human experience.

[0048] Based on the same general inventive concept, the present invention also provides an intelligent grouting control system.

[0049] Figure 2 This is a schematic diagram of the intelligent grouting control system provided in this embodiment.

[0050] like Figure 2 As shown in the figure, the intelligent grouting control system provided in this embodiment includes: The acquisition module 201 is used to acquire multi-parameter time-series data in real time during the grouting process and perform standardization processing to construct feature vectors; The execution module 202 is used to input the feature vector into the ensemble learning model and simultaneously execute the grouting effect prediction based on time series analysis and the abnormal working condition identification based on data distribution analysis. When an abnormal working condition is identified, the control strategy corresponding to the abnormality is executed first. When no abnormality is identified, the control parameters are dynamically optimized according to the predicted grouting effect. The distribution module 203 is used to distribute control strategies or optimized control parameters to the grouting equipment for execution and collect feedback data after execution. Feedback module 204 is used to perform online incremental learning on the ensemble learning model based on feedback data, and adaptively optimize the model parameters.

[0051] Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment.

[0052] like Figure 3 As shown, the electronic device may include: a processor 301, a communication interface 302, a memory 303, and a communication bus 304. The processor 301, communication interface 302, and memory 303 communicate with each other via the communication bus 304. The processor 301 can call logical instructions stored in the memory 303 to execute the intelligent grouting control method.

[0053] Furthermore, the logical instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0054] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the intelligent grouting control method provided by the above methods.

[0055] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the intelligent grouting control methods provided by the above methods.

[0056] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0057] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart grouting control method, characterized in that, include: Real-time acquisition of multi-parameter time-series data during the grouting process, followed by standardization processing to construct feature vectors; The feature vector is input into the ensemble learning model to simultaneously perform grouting effect prediction based on time series analysis and abnormal working condition identification based on data distribution analysis. When an abnormal operating condition is detected, the control strategy corresponding to the abnormality is executed first; when no abnormality is detected, the control parameters are dynamically optimized based on the predicted grouting effect. The control strategy or the optimized control parameters are sent to the grouting equipment for execution, and feedback data is collected after execution. Based on the feedback data, the ensemble learning model is subjected to online incremental learning to adaptively optimize the model parameters.

2. The intelligent grouting control method according to claim 1, characterized in that, The step of inputting the feature vector into the ensemble learning model and simultaneously performing grouting effect prediction based on time series analysis and abnormal working condition identification based on data distribution analysis includes: The feature vectors are simultaneously input into a long short-term memory network and an isolated forest algorithm; The long short-term memory network is used to perform time-series analysis on the feature vector to output a predicted value of grouting effect. The isolated forest algorithm is used to analyze the data distribution of the feature vectors and output an abnormal operating condition identification signal.

3. The intelligent grouting control method according to claim 1, characterized in that, The priority execution of the control strategy corresponding to the anomaly includes: The abnormal operating condition identification signal is analyzed to determine the abnormal type; Based on the preset mapping relationship between anomaly types and control strategies, the target control strategy corresponding to the anomaly type is invoked. The target control strategy includes at least an adjustment instruction for grouting pressure or flow rate. The current parameter optimization process is interrupted, and the target control strategy is sent to the grouting equipment for execution.

4. The intelligent grouting control method according to claim 1, characterized in that, When no anomalies are detected, the control parameters are dynamically optimized based on the predicted grouting effect, including: The predicted grouting effect is compared with the preset target effect value, and the effect deviation is calculated; Based on the magnitude and direction of the effect deviation, and using a preset optimization rule base, the adjustment direction and magnitude of the grouting pressure or flow rate are determined. Based on the adjustment direction and amplitude, new control parameters are generated and sent to the grouting equipment for execution, so that the grouting effect approaches the target effect value.

5. The intelligent grouting control method according to claim 1, characterized in that, The step of sending the control strategy or the optimized control parameters to the grouting equipment for execution includes: The control strategy or optimized control parameters are encapsulated into standard control commands that can be recognized by the underlying drive system of the grouting equipment. The standard control commands are sent to the corresponding grouting equipment actuators via industrial communication protocols. The status feedback signal from the actuator is received in real time to confirm that the control command has been successfully executed.

6. The intelligent grouting control method according to claim 1, characterized in that, After acquiring multi-parameter time-series data during the real-time grouting process, the method further includes: Analyze the noise characteristics of data from different sensors in a specific frequency band to identify noise sources; Dynamically select the corresponding digital filter type and parameters based on the noise source; The original sensor data is processed using the selected digital filter type and parameters, and the signal-to-noise ratio of the filtered data is verified.

7. The intelligent grouting control method according to claim 2, characterized in that, The step of performing time-series analysis on the feature vector using the long short-term memory network to output a predicted value for grouting effect further includes: The geological conditions of the current construction area are digitally coded. The encoded geological condition identifier vector is fused with the real-time acquired temporal feature vector to form an extended feature vector; The extended feature vector is input into the long short-term memory network to adaptively adjust the dependence weights on the geological background.

8. The intelligent grouting control method according to claim 1, characterized in that, After collecting the feedback data after execution, the process also includes: Collect the actual operating parameters of the grouting equipment actuator and compare them with the issued control commands to form a comparison data set; Calculate the continuous deviation between the actual operating parameters and the control commands; If the continuous deviation exceeds the preset allowable range, an equipment health status warning is generated, and a compensation value for the control command is calculated based on the continuous deviation to trigger system self-calibration.

9. The intelligent grouting control method according to any one of claims 1-8, characterized in that, The step of performing online incremental learning on the ensemble learning model based on the feedback data and adaptively optimizing the model parameters includes: The feedback data is correlated with the actual grouting effect evaluation indicators to generate new training samples with labels; Calculate the difference measure between the new training samples and the model's existing knowledge base, and select a subset of samples; The mini-batch gradient descent method is used to incrementally update the model parameters using the aforementioned sample subset. Verify the performance of the model after the incremental update, and adjust the learning rate parameter for subsequent incremental learning based on the verification results.

10. A smart grouting control system, characterized in that, include: The acquisition module is used to acquire multi-parameter time-series data in real time during the grouting process and perform standardization processing to construct feature vectors; The execution module is used to input the feature vector into the ensemble learning model and simultaneously execute the grouting effect prediction based on time series analysis and the abnormal working condition identification based on data distribution analysis; when an abnormal working condition is identified, the control strategy corresponding to the abnormality is executed first. If no anomalies are detected, the control parameters are dynamically optimized based on the predicted grouting effect. The distribution module is used to distribute the control strategy or the optimized control parameters to the grouting equipment for execution, and to collect feedback data after execution; The feedback module is used to perform online incremental learning on the ensemble learning model based on the feedback data, and adaptively optimize the model parameters.

Images