Controllable grouting method and grouting structure for cofferdam foundation of sandy gravel stratum

Abstract

The invention provides a sandy gravel stratum cofferdam foundation controlled grouting method and structure. The sandy gravel stratum cofferdam foundation controlled grouting method comprises the following steps that S1, construction hole sites are measured and lofted, and a plurality of hole sites are marked; the drilling machine is moved to the lofting hole position, casing-following drilling is carried out, a casing pipe follows up to a bed rock section, and one-time hole forming is carried out; s2, a grouting structure is installed in the casing pipe, the grouting structure comprises a stop-grouting plug and a grout outlet head, and the stop-grouting plug and the grout outlet head are installed at the position of the last section of the casing pipe in the hole; s3, a section of sleeve is pulled out through a tube drawing bench; s4, grouting is conducted, and grouting is stopped after grouting of the section is completed; s5, the steps S3-S4 are repeated till grouting in the hole is completed, grouting is stopped, and the grouting structure is moved to the next hole position; and S6, the steps S2-S5 are repeated till grouting of all the hole positions in the working area is completed. The slurry concentration is automatically controlled, the operation is simpler, the convenience during grouting is improved, and the strength of manual operation is reduced; and the device adapts to different grouting conditions, the adaptive capacity is better, and grouting is more convenient.

Description

Technical Field

[0001] The invention relates to the technical field of grouting, and in particular to a controlled grouting method and a grouting structure for a cofferdam foundation in a sand and gravel formation. Background Art

[0002] A cofferdam is a temporary retaining structure built for the construction of permanent water conservancy facilities in water conservancy project construction. Its function is to prevent water and soil from entering the construction site of the building so that drainage can be done in the cofferdam, foundation pits can be excavated, and buildings can be built. It is generally used in hydraulic structures. Except as part of a formal building, the cofferdam is generally dismantled after use. There are many types of cofferdams, including earth-rock cofferdams, concrete cofferdams, steel sheet pile cofferdams, etc. Their design and construction need to consider multiple factors such as water flow velocity, water depth, geological conditions, etc. to ensure construction safety and effectiveness.

[0003] Controlled grouting of cofferdams is a construction technology used to reinforce and seal earth-rock cofferdams. By adopting a reasonable grouting process, a certain amount of cementitious material is injected into the cofferdam foundation and filling materials, so that it flows, diffuses, fills and consolidates in the pores, cracks or cavities of the rock and soil, forming a stone body with certain strength and low permeability, thereby improving the strength and anti-seepage performance of the cofferdam.

[0004] The sand and gravel strata have the characteristics of strong discreteness and high permeability. The effect of grouting will directly affect the workload of regular drainage and even affect the safety of dam filling. However, there is no mature reference for the controlled grouting of cofferdams at present. Various situations are easily encountered during the grouting process. The existing technology completes the grouting in the hole at one time, which is not easy to adjust and control during the grouting process, resulting in poor grouting quality, which in turn affects the quality of the whole day. Summary of the invention

[0005] The main purpose of the present invention is to provide a controlled grouting method and a grouting structure for a cofferdam foundation in a sand and gravel formation, so as to solve the problems of inconvenient control of the grouting process and poor grouting quality.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A controlled grouting method for a cofferdam foundation in a sandy and gravel stratum comprises the following steps:

[0008] S1. Measure and lay out the construction holes and mark multiple holes;

[0009] Move the drilling rig to the designated hole location, carry out casing drilling, and follow the casing to the bedrock section to complete the hole drilling in one go;

[0010] S2. Install a grouting structure in the casing. The grouting structure includes a grout stopper and a grout outlet. The grout stopper and the grout outlet are installed to the last section of the casing in the hole.

[0011] S3. Use a pipe extractor to extract a section of casing;

[0012] S4. Grout, and stop grouting after this section of grouting is completed;

[0013] S5. Repeat S3 - S4 until all grouting in the hole is completed, stop grouting, and move the grouting structure to the next hole position;

[0014] S6. Repeat S2 - S5 until all hole positions in the working area are grouted.

[0015] In the preferred solution, before measuring and lofting the construction hole positions, install the water, electricity, and air supply to ensure stable water and electricity supply at the construction site; comprehensively level the construction site and build a grout transportation system;

[0016] When following the pipe during drilling, use a concentric three - wing pipe - following bit;

[0017] When extracting the pipe, according to the feedback of the drill pipe during the drilling process, if there is an air - filled layer, fill it with gravel, sand, and membrane bags.

[0018] In the preferred solution, S4 includes the following steps:

[0019] S41. Data collection and pre - processing;

[0020] S42. Build an LSTM model;

[0021] S43. Adjust the grouting concentration based on the collected data and the LSTM model.

[0022] In the preferred solution, in S41:

[0023] During the grouting process, use pressure sensors and flow sensors to collect the grouting pressure P and flow rate Q data in real - time, and record the corresponding time stamp t at the same time; collect a set of data (P i , Q i , t i ) at every fixed time interval Δt to build an initial data set:

[0024] D = {(P1, Q1, t1), (P2, Q2, t2), …, (P n , Q n , t n )};

[0025] Perform data cleaning on the data using an outlier detection algorithm based on statistical methods to process the noise or outliers in the data;

[0026] For the pressure data P and the flow rate data Q, calculate their mean and standard deviation σ P 、σ Q, if a certain data point satisfies or then it is determined that the data point is an outlier, which is removed, and linear interpolation is used to supplement the missing data, where k takes 2 - 3.

[0027] In the preferred solution, in S42:

[0028] The pre - processed data is divided into a training set D train and a test set D test , where the training set accounts for 80% and the test set accounts for 20%; the data in the training set is normalized, and the pressure P and the flow rate Q are respectively mapped to the interval [0, 1]. The normalization formula is:

[0029]

[0030] In the formula, P min , P max are the minimum and maximum values of the pressure data in the training set, and Q min , Q max are the minimum and maximum values of the flow rate data in the training set;

[0031] An LSTM model is constructed, and the model structure is set to include an input layer, several LSTM hidden layers, and an output layer;

[0032] The input layer receives the normalized pressure and flow rate data, and the input at each time step is (P norm,i , Q norm,i );

[0033] The output layer outputs the predicted grouting concentration adjustment coefficient r;

[0034] The training set is used to train the LSTM model. The mean square error is used as the loss function, and the optimization algorithm selects Adam. The loss function formula is:

[0035]

[0036] In the formula, m is the number of training samples, r i pred is the grouting concentration adjustment coefficient predicted by the model, and r i true is the actual grouting concentration adjustment coefficient;

[0037] During the training process, the model parameters are continuously adjusted until the loss function converges, and a trained LSTM model is obtained.

[0038] In the preferred solution, in S43:

[0039] According to the grouting construction experience and the actual engineering requirements, the grouting process is divided into several time periods T1, T2, …, T m , and the length of each time period is T interval ;

[0040] At the beginning of each time period, the return slurry and pressure data collected in the previous ΔT within this time period are input into the trained LSTM model. The model outputs the grouting concentration adjustment coefficient r, and the current grouting concentration C is adjusted according to the adjustment coefficient r. The adjustment formula is:

[0041] C new = C×(1 + r);

[0042] In the formula, C new is the adjusted grouting concentration.

[0043] In the preferred solution, in S4:

[0044] S411. The pressure, flow rate, and density data during the grouting process are collected in real time through pressure sensors, flow sensors, and density sensors, and are subjected to cleaning, normalization, and standardization processing;

[0045] S412. Use the trained multi-layer perceptron neural network model, input the pressure, flow rate, and density data collected in real time, and predict the optimal grouting slurry concentration and grouting pressure;

[0046] S413. Set the initial grouting slurry concentration and grouting pressure according to the prediction results of the neural network model.

[0047] In the preferred solution, first grout with the initial grouting slurry concentration and grouting pressure. During the grouting process:

[0048] If the return slurry is normal or the pressure is within the normal range:

[0049] Grout the cement slurry with the normal initial grouting ratio;

[0050] Normal pressure rise and reach the end standard required by the design or reach the maximum injection volume specified by the design, and end the grouting of this section;

[0051] If the return slurry is small or the pressure is small:

[0052] Grout with a cement slurry one level thicker than the initial grouting ratio;

[0053] If normal pressure rise occurs during the grouting process, transform according to the water-cement ratio transformation principle; if there is no obvious effect, grout the paste-like slurry and grout according to the paste-like slurry transformation principle;

[0054] If there is no return slurry or no pressure:

[0055] Start grouting with the thickest cement slurry. During the grouting process, if the pressure rises normally, continue grouting according to the principle of water-cement ratio transformation; if there is still no pressure, use paste slurry for grouting; according to the effect during the grouting process, if there is obvious backflow or pressure, grout according to the principle of water-cement ratio transformation; if there is no obvious effect, use paste slurry with three levels of diffusivity for step-by-step grouting. If there is still no effect after grouting with the thickest paste slurry, end with a limited quantity.

[0056] The principle of water-cement ratio transformation is: according to the prediction result of the neural network model, dynamically adjust the grouting pressure:

[0057] P new = P old + α·ΔP;

[0058] Among them, P new is the adjusted grouting pressure, P old is the current grouting pressure, α is the adjustment coefficient, and ΔP is the pressure change amount predicted according to the neural network model;

[0059] According to the prediction result of the neural network model, dynamically adjust the slurry concentration:

[0060] C new = C old + β·ΔC;

[0061] Among them, C new is the adjusted slurry concentration, C old is the current slurry concentration, β is the adjustment coefficient, and ΔC is the concentration change amount predicted according to the neural network model;

[0062] The principle of paste slurry ratio change is: when the grouting volume of a certain ratio of paste slurry reaches 1m 3 , and the grouting pressure and injection rate do not change or change insignificantly, change the slurry step by step;

[0063] Use the K-means clustering algorithm to classify the grouting effect and evaluate the grouting effect under different grouting parameters;

[0064]

[0065] Among them, x and y are two data points, and n is the feature dimension;

[0066] According to the clustering result, evaluate the grouting effect under different grouting parameters and select the optimal combination of grouting parameters.

[0067] A grouting structure is used in a control grouting method for the foundation of a sand and gravel stratum cofferdam, and it also includes a casing. A grout stopper is provided inside the casing, and the grout stopper is provided with a grout inlet pipe and a sealing sleeve;

[0068] One end of the grout inlet pipe passes through the grout stopper and is connected with a grout outlet head. The other end of the grout inlet pipe is connected with a grouting device, a pressure gauge and a flow meter.

[0069] The grout stopper is provided with an air inlet pipe. One end of the air inlet pipe is communicated with the sealing sleeve. The sealing sleeve bulges due to air intake and abuts against the inner wall of the casing. The air inlet pipe is externally connected with an air intake device.

[0070] A pressure sensor is arranged inside the grout stopper. A slider is arranged at the bottom of the grout stopper. The slider is connected with the detection end of the pressure sensor.

[0071] The slider is slidably arranged inside the grout stopper, and a sealing ring is arranged at the sliding connection part.

[0072] In a preferred solution, the grout stopper includes a fixed seat and a connecting seat. The sealing sleeve is arranged on the outer periphery of the connecting seat.

[0073] A first air cavity is arranged inside the fixed seat. A second air cavity is arranged on the outer periphery of the connecting seat. The second air cavity is communicated with the inside of the sealing sleeve. A connecting hole is arranged inside the fixed seat and the connecting seat. The first air cavity and the second air cavity are communicated through the connecting hole. The air inlet pipe is communicated with the first air cavity.

[0074] A plurality of stabilizing mechanisms are arranged on the outer periphery of the fixed seat for increasing the stability of the grout stopper fixed inside the casing.

[0075] The stabilizing mechanism includes a piston cavity arranged inside the fixed seat. The piston cavity is communicated with the first air cavity. A piston is arranged inside the piston cavity. A plurality of connecting rods are arranged on the side of the piston away from the first air cavity.

[0076] A plurality of temporary storage grooves are arranged on the outer periphery of the fixed seat. A stabilizing plate is movably connected inside the temporary storage groove. The stabilizing plate is fixedly connected with the connecting rod.

[0077] A buffer pad is arranged on the side of the stabilizing plate away from the connecting rod.

[0078] The side of the stabilizing plate away from the connecting rod is arc-shaped.

[0079] The present invention provides a method for controlling grouting of a cofferdam foundation in a sandy pebble stratum and a grouting structure. By adopting the above solutions, the following beneficial effects are achieved:

[0080] 1. Automatically control the slurry concentration, with simpler operation, increased convenience during grouting, and reduced labor intensity of manual operation.

[0081] 2. Real-time control the grouting concentration according to the grouting situation, adapt to different grouting situations, have better adaptability, and be more convenient for grouting.

[0082] 3. Make real-time adjustments during the grouting process to facilitate ensuring the grouting effect and making the quality after grouting better.

[0083] 4. The grouting pipe and the grout stopper can be stably installed inside the casing to achieve a good grouting effect.

[0084] 5. The grout stopper is fixed and sealed by inflation, which is simpler and more efficient in operation during connection. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] The present invention will be further described below in conjunction with the drawings and embodiments:

[0086] Figure 1 is the construction flow chart of controlled grouting for sand and pebble strata cofferdam foundation

[0087] Figure 2 is the detailed flow chart of the selection and transformation of controlled grouting slurry for sand and pebble strata

[0088] Figure 3 is the experimental diagram of paste grouting slurry for sand and pebble control system

[0089] Figure 4 is the structural schematic diagram of the grouting structure of the present invention

[0090] Figure 5 is the enlarged structural schematic diagram of the grouting structure of the present invention

[0091] Figure 6 is the top view structural schematic diagram of the grouting structure of the present invention

[0092] Figure 7 is the front view structural schematic diagram of the grouting structure of the present invention

[0093] Figure 8 is the enlarged structural schematic diagram of the stabilizing mechanism of the present invention

[0094] In the figure:

[0095] casing 1, grout stopper 2, fixing seat 201, connecting seat 202, first air chamber 203, connecting hole 204, second air chamber 205, grout inlet pipe 3, grout outlet head 301, sealing sleeve 4, stabilizing mechanism 5, piston chamber 501, piston 502, connecting rod 503, temporary storage tank 504, stabilizing plate 505, buffer pad 506, air inlet pipe 6, pressure sensor 701, slider 702. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0096] Embodiment 1:

[0097] A method for controlled grouting of the cofferdam foundation in sand and pebble strata includes the following steps:

[0098] S1. Measuring and lofting the construction hole positions, marking multiple hole positions, and preferably using a total station for measuring and lofting the construction hole positions;

[0099] Move the drilling rig to the lofting hole position, carry out casing drilling, follow the casing into the bedrock section, and form a hole in one go;

[0100] When following the casing for drilling, use a concentric three-wing casing bit. The concentric three-wing casing bit has good drilling performance in sandy pebble formations;

[0101] Before measuring and lofting the construction hole position, complete the installation of water, wind, and electricity to ensure stable water and electricity supply at the construction site; comprehensively level the construction site and build the existing slurry conveying system.

[0102] S2. Install a grouting structure inside the casing. The grouting structure includes a grout stopper and a grout outlet head. The grout stopper and the grout outlet head are installed at the last section of the casing in the hole;

[0103] S3. Use a pipe extractor to pull out a section of the casing;

[0104] S4. Grout. Stop grouting after this section of grouting is completed;

[0105] S5. Repeat S3 - S4 until all grouting in the hole is completed, stop grouting, and move the grouting structure to the next hole position;

[0106] S6. Repeat S2 - S5 until all hole positions in the working area are grouted.

[0107] The grouting is completed through the above method, which is simple in operation, easy to control, and ensures the quality of grouting.

[0108] In the preferred solution, when pulling out the pipe, according to the feedback of the drill pipe during the drilling process, if there is an air layer, fill it with crushed stone, sand, and a membrane bag; during the drilling process, judge whether there is an air layer through feedback information such as the vibration and torque change of the drill pipe. For example, when the drill pipe suddenly vibrates significantly and the torque decreases, there may be an air layer. After discovering the air layer, fill it with crushed stone, sand, and a membrane bag. Fill it in the existing way, such as first mixing crushed stone and sand in a certain proportion, then loading them into a high-strength membrane bag, and sending the filling material into the air layer through existing special equipment to ensure dense filling and enhance the stability of the formation.

[0109] In this application, grouting is carried out in a single hole by means of step-by-step sectional grouting. Compared with the one-time grouting of the prior art, it is easier to control and adjust according to the actual grouting situation during grouting, ensuring the grouting effect and making the grouting quality better.

[0110] In the preferred solution, S4 includes the following steps:

[0111] S41. Data acquisition and preprocessing;

[0112] During the grouting process, the grouting pressure P and flow rate Q data are collected in real time using pressure sensors and flow sensors, and the corresponding time stamps t are recorded simultaneously; a set of data (P i , Q i , t i ) is collected every fixed time interval Δt, where Δt is 1 minute. After the start of grouting, the pressure and flow rate data are collected and recorded every minute to construct the initial data set:

[0113] D = {(P1, Q1, t1), (P2, Q2, t2), …, (P n , Q n , t n )};

[0114] Perform data cleaning using an outlier detection algorithm based on statistical methods to handle noise or outliers in the data;

[0115] For the pressure data P and the flow rate data Q, calculate their mean and standard deviation σ P , σ Q . If a certain data point satisfies or , then determine that the data point is an outlier, remove it, and use linear interpolation to supplement the missing data. Among them, k takes 2 - 3. For example, if the pressure data P i = 5 MPa and the standard deviation σ P = 0.5 MPa, and k takes 3, then the data point is an outlier and is removed. If there is missing data at a certain moment, such as the data missing at the 5th minute, linear interpolation can be performed based on the data at the 4th minute and the 6th minute.

[0116] S42. Construct an LSTM model;

[0117] Divide the preprocessed data into a training set D train and a test set D test in chronological order, where the training set accounts for 80% and the test set accounts for 20%; perform normalization processing on the training set data, map the pressure P and the flow rate Q to the interval [0, 1] respectively, and the normalization formula is:

[0118]

[0119] In the formula, P min , P max are the minimum and maximum values of the training set pressure data, and Q min , Q max are the minimum and maximum values of the training set flow rate data;

[0120] Build an LSTM model and set the model structure to include an input layer, several LSTM hidden layers, and an output layer;

[0121] The input layer receives the normalized pressure and flow data, and the input at each time step is (P norm,i ,Q norm,i );

[0122] The output layer outputs the predicted grouting concentration adjustment coefficient r;

[0123] Use the training set to train the LSTM model, adopt the mean square error as the loss function, and select Adam as the optimization algorithm. The loss function formula is:

[0124]

[0125] In the formula, m is the number of training samples, r i pred is the grouting concentration adjustment coefficient predicted by the model, and r i true is the actual grouting concentration adjustment coefficient;

[0126] During the training process, set the LSTM model to have 2 hidden layers, with 32 neurons in each layer. After multiple trainings, when the loss function value stabilizes within a small range, it is considered that the model converges, and the trained LSTM model is obtained.

[0127] S43. Adjust the grouting concentration based on the collected data and the LSTM model.

[0128] According to the grouting construction experience and the actual engineering requirements, divide the grouting process into several time periods T1, T2, …, T m , and the length of each time period is T interval ;

[0129] At the beginning of each time period, input the return slurry and pressure data collected in the previous ΔT within this time period into the trained LSTM model. The model outputs the grouting concentration adjustment coefficient r, and adjust the current grouting concentration C according to the adjustment coefficient r. The adjustment formula is:

[0130] C new = C × (1 + r);

[0131] In the formula, C new is the adjusted grouting concentration.

[0132] When adjusting the grouting concentration, combine the actual geological conditions and grouting effects, and perform appropriate manual intervention on the adjustment coefficient \(r\) to ensure that the adjusted concentration is reasonable. During each time period, continuously monitor the grouting parameters. If abnormal changes are found, re-enter the data in time for model prediction and concentration adjustment.

[0133] In the above - mentioned way, the grouting concentration can be controlled in real - time during the grouting process to adapt to different grouting situations. The operation is simple, the efficiency is higher, and the quality of grouting is guaranteed.

[0134] Example 2:

[0135] As Figure 1 、 2 and shown in Figure 3, a control grouting method for the cofferdam foundation in sandy cobble stratum includes the following steps:

[0136] S1. Measure and set out the construction hole positions, mark multiple hole positions. Preferably, a total station is used for measuring and setting out the construction hole positions.

[0137] Move the drill rig to the set - out hole position and conduct casing - following drilling. The casing is advanced into the bedrock section to form a hole in one go.

[0138] When conducting casing - following drilling, a concentric three - wing casing - following bit is used. The concentric three - wing casing - following bit has good drilling performance in sandy cobble stratum.

[0139] Before measuring and setting out the construction hole positions, complete the installation of water, wind, and electricity to ensure stable water and electricity supply at the construction site. Conduct a comprehensive leveling of the construction site and set up the existing slurry - conveying system.

[0140] S2. Install a grouting structure inside the casing. The grouting structure includes a grout - stopping plug and a grout - discharging head. The grout - stopping plug and the grout - discharging head are installed at the last section of the casing in the hole.

[0141] S3. Use a pipe - pulling machine to pull out a section of the casing.

[0142] S4. Grout. After this section of grouting is completed, stop grouting.

[0143] S5. Repeat S3 - S4 until all grouting in the hole is completed. Then stop grouting and move the grouting structure to the next hole position.

[0144] S6. Repeat S2 - S5 until all hole positions in the working area are grouted.

[0145] By the above - mentioned way, the grouting is completed. The operation is simple, easy to control, and the quality of grouting is guaranteed.

[0146] In a preferred solution, during pipe extraction, according to the feedback of the drill pipe during the drilling process, if there is an air layer, crushed stone, sand, and membrane bags are filled; during the drilling process, it is judged whether there is an air layer through feedback information such as the vibration and torque change of the drill pipe. For example, when the drill pipe suddenly vibrates significantly and the torque decreases, there may be an air layer. After discovering the air layer, crushed stone, sand, and membrane bags are used for filling. The existing method is used for filling, such as first mixing crushed stone and sand in a certain proportion, then loading them into a high-strength membrane bag, and sending the filling material into the air layer through existing special equipment to ensure dense filling and enhance the stability of the formation.

[0147] In S4:

[0148] S411. Real-time collect the pressure, flow rate, and density data during the grouting process through pressure sensors, flow sensors, and density sensors, and perform cleaning, normalization, and standardization processing;

[0149] The pressure sensor is installed at the outlet of the grouting pump, the flow sensor is installed in the grouting pipeline, and the density sensor is installed in the slurry storage tank. A set of data is collected at regular time intervals (such as 0.5 minutes) to construct a data set containing pressure P, flow rate Q, and density ρ. In the data cleaning stage, a variety of methods are combined, such as setting a reasonable data threshold range and using statistical methods to identify outliers.

[0150] For the normalization process, preferably map the pressure, flow rate, and density data to the [0,1] interval respectively, and the formulas are:

[0151]

[0152] The standardization process preferably uses the Z-score standardization method, and the formula is: Where X represents the pressure, flow rate, or density data, μ is the mean of the data, and σ is the standard deviation of the data. Assume that the mean μ of a set of pressure data is 1.3 MPa, the standard deviation σ is 0.2 MPa, and a certain pressure data P is 1.5 MPa. After standardization

[0153] S412. Use the trained multi-layer perceptron neural network model, input the real-time collected pressure, flow rate, and density data, and predict the optimal grouting slurry concentration and grouting pressure;

[0154] The multi-layer perceptron neural network model consists of an input layer, several hidden layers, and an output layer. The input layer receives the normalized and standardized data, the hidden layer performs feature extraction and transformation through a non-linear activation function (such as the ReLU function: f(x)=max(0,x)), and the output layer outputs the predicted optimal grouting slurry concentration and grouting pressure.

[0155] S413. Set the initial grouting slurry concentration and grouting pressure according to the prediction results of the neural network model.

[0156] In the preferred solution, first perform grouting with the initial grouting slurry concentration and grouting pressure. During the initial grouting process:

[0157] If the slurry return is normal or the pressure is within the normal range:

[0158] Use the normal starting grouting ratio to inject cement slurry, which is specifically determined according to the actual construction situation, such as a water-cement ratio of 1:1;

[0159] Continuously monitor the grouting pressure and flow rate. When the normal pressure starts and reaches the end standard required by the design (preferably, the pressure stabilization time exceeds 30 minutes and the injection rate is less than 5 L / min), or when the maximum injection volume specified by the design is reached, end the grouting of this section;

[0160] If the slurry return is small or the pressure is small:

[0161] Use a cement slurry one level thicker than the starting grouting ratio for injection. For example, when the starting grouting ratio is 1:1, adjust to inject cement slurry with a ratio of 0.9:1;

[0162] If normal pressure starts during the grouting process, use the water-cement ratio transformation principle for transformation, and dynamically adjust the grouting pressure and slurry concentration according to the prediction results of the neural network model; if there is no obvious effect, inject paste-like slurry and inject according to the paste-like slurry transformation principle;

[0163] If there is no slurry return or no pressure:

[0164] Start injecting with the thickest grade of cement slurry, such as cement slurry with a ratio of 0.6:1. If normal pressure starts during the injection process, continue injecting according to the water-cement ratio transformation principle; if there is still no pressure, inject paste-like slurry; according to the effect during the injection process, if there is obvious slurry return or pressure, inject according to the water-cement ratio transformation principle; if there is no obvious effect, inject using the three-stage diffusion degree of paste-like slurry step by step. If there is still no effect after injecting the thickest grade of paste-like slurry, then end with a limited amount;

[0165] The water-cement ratio transformation principle is: Dynamically adjust the grouting pressure according to the prediction results of the neural network model:

[0166] P new = P old + α·ΔP;

[0167] Wherein, P new is the adjusted grouting pressure, P old is the current grouting pressure, α is the adjustment coefficient, and ΔP is the pressure change amount predicted according to the neural network model;

[0168] Dynamically adjust the slurry concentration according to the prediction results of the neural network model:

[0169] C new = C old + β·ΔC;

[0170] Wherein, C new is the adjusted slurry concentration, C old is the current slurry concentration, β is the adjustment coefficient, and ΔC is the predicted concentration change amount according to the neural network model;

[0171] The principle of the change in the paste slurry ratio level is: when the perfusion volume of the paste slurry at a certain ratio level reaches 1 m 3 , and when both the grouting pressure and the injection rate remain unchanged or change insignificantly, the slurry is changed step by step;

[0172] As shown in Table 1 below, cement and bentonite paste slurries are prepared with a primary diffusion degree of 160 mm to 210 mm, a secondary diffusion degree of 120 mm to 160 mm, and a tertiary diffusion degree of 70 mm to 120 mm, and the mass mixing ratios of bentonite are 2%, 5%, and 20% in sequence.

[0173]

[0174] Table 1

[0175] Use the K-means clustering algorithm to classify the grouting effect and evaluate the grouting effect under different grouting parameters;

[0176]

[0177] Wherein, x and y are two data points, and n is the feature dimension;

[0178] According to the clustering results, evaluate the grouting effect under different grouting parameters and select the optimal combination of grouting parameters. Select grouting pressure, flow rate, density, slurry injection volume, and slurry diffusion range, etc. as clustering features. Assume the data point x i = (p i , q i , d i , v i , s i ), where p i is the grouting pressure, q i is the flow rate, d i is the density, v i is the slurry injection volume, s iThis is the slurry diffusion range. The algorithm measures similarity by calculating the Euclidean distance between data points and continuously iteratively updates the clustering centers until the clustering results are stable. First, randomly select K initial clustering centers (the value of K is determined based on experience or multiple trials, such as K = 3). Then, calculate the distance from each data point to each clustering center and assign the data point to the nearest cluster. Next, recalculate the center of each cluster based on the assignment results. Repeat the above steps until the clustering centers no longer change or change very little, and the clustering process ends. After clustering is completed, evaluate the grouting effects of different clusters. For example, if the grouting pressure in a certain cluster is stable, the flow rate is moderate, the slurry injection volume reaches the expectation, the density meets the requirements, and the slurry diffusion range is reasonable, it indicates that the combination of grouting parameters corresponding to this cluster has a good effect. Select the parameter combination with the best grouting effect from different clusters as a reference for subsequent grouting construction to optimize the grouting process.

[0179] Embodiment 3:

[0180] As Figure 4 、 5 shown in Figures 6, 7, and 8, a grouting structure includes a casing 1, which is the casing during root canal drilling. A grout stopper 2 is provided inside the casing 1, and the grout stopper 2 is provided with a grout inlet pipe 3 and a sealing sleeve 4;

[0181] The sealing sleeve 4 is preferably made of polytetrafluoroethylene, soluble polytetrafluoroethylene, fluorinated ethylene propylene copolymer, polyether ether ketone, or liquid crystal polymer, etc., and has certain flexibility and corrosion resistance. During the working process, the sealing sleeve 4 is pressed against the inner wall of the casing 1 under the influence of air pressure to complete the sealing between the casing 1 and the grout stopper 2.

[0182] One end of the grout inlet pipe 3 passes through the grout stopper 2 and is connected with a grout outlet head 301. The other end of the grout inlet pipe 3 is connected to grouting equipment, a pressure gauge, and a flow meter; the slurry is directly fed through the grout inlet pipe 3 and the grout outlet head 301 in an existing manner, and the grouting pressure is detected by the pressure gauge, and the flow meter is used to detect the flow rate.

[0183] The grout stopper 2 is provided with an air inlet pipe 6. One end of the air inlet pipe 6 is communicated with the sealing sleeve 4, and the sealing sleeve 4 bulges due to the intake of air and presses against the inner wall of the casing 1. The air inlet pipe 6 is externally connected to an air intake device; it is used to intake air into the inside of the grout stopper 2 to make the sealing sleeve 4 bulge.

[0184] A pressure sensor 701 is provided inside the grout stopper 2, and a slider 702 is provided at the bottom of the grout stopper 2. The slider 702 is connected to the detection end of the pressure sensor 701;

[0185] During the grouting process, the reverse grout pressure of the cement slurry acts on the slider 702, and then acts on the pressure sensor 701 through the slider 702 to detect the grouting pressure.

[0186] The slider 702 is slidably arranged inside the grout plug 2, and a sealing ring is provided at the sliding connection to prevent the cement slurry from passing over the slider 702 and affecting the pressure sensor 701.

[0187] In a preferred solution, the grout plug 2 includes a fixed seat 201 and a connecting seat 202, and the sealing sleeve 4 is arranged on the outer periphery of the connecting seat 202;

[0188] A first air chamber 203 is arranged inside the fixed seat 201, a second air chamber 205 is arranged on the outer periphery of the connecting seat 202, the second air chamber 205 is communicated with the inside of the sealing sleeve 4, a connecting hole 204 is arranged in the fixed seat 201 and the connecting seat 202, the first air chamber 203 and the second air chamber 205 are communicated through the connecting hole 204, and the air inlet pipe 6 is communicated with the first air chamber 203;

[0189] The gas filled enters the first air chamber 203 through the air inlet pipe 6, then passes through the connecting hole 204 into the second air chamber 205 and finally inflates the sealing sleeve 4.

[0190] A plurality of stabilizing mechanisms 5 are arranged on the outer periphery of the fixed seat 201 to increase the stability of the grout plug 2 fixed in the casing 1;

[0191] The stabilizing mechanism 5 includes a piston chamber 501 arranged inside the fixed seat 201, the piston chamber 501 is communicated with the first air chamber 203, a piston 502 is arranged inside the piston chamber 501, and a plurality of connecting rods 503 are arranged on the side of the piston 502 away from the first air chamber 203;

[0192] A plurality of temporary storage grooves 504 are arranged on the outer periphery of the fixed seat 201, a stabilizing plate 505 is movably connected inside the temporary storage groove 504, and the stabilizing plate 505 is fixedly connected with the connecting rod 503;

[0193] A buffer pad 506 is arranged on the side of the stabilizing plate 505 away from the connecting rod 503 to increase the friction and stability after the stabilizing plate 505 contacts the inner wall of the casing 1;

[0194] The side of the stabilizing plate 505 away from the connecting rod 503 is arc-shaped to adapt to the inner wall of the casing 1;

[0195] During use, the gas filled enters the first air chamber 203 through the air inlet pipe 6, thereby pushing the piston 502 away from the first air chamber 203, then pushing the connecting rod 503 and the stabilizing plate 505, so that the stabilizing plate 505 tightly abuts against the inner wall of the casing 1, and the fixation of the grout plug 2 and the casing 1 can be completed.

[0196] This application is mainly applied to the grouting of the cofferdam foundation of the sandy cobble stratum, but is not limited to the grouting of the cofferdam foundation of the sandy cobble stratum, and is also applicable to other similar grouting sites.

[0197] A grouting structure applied is mainly used in a control grouting method for the cofferdam foundation of a sandy cobble stratum in this application, but is not limited to the control grouting method for the cofferdam foundation of a sandy cobble stratum in this application.

[0198] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations on the present invention. The protection scope of the present invention should be the technical solutions recorded in the claims, including the equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, the equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A controlled grouting method for a cofferdam foundation in a sandy and gravel formation, characterized by: The following steps are involved: S1. Measure and lay out the construction holes and mark multiple holes; Move the drilling rig to the designated hole location, carry out casing drilling, and follow the casing to the bedrock section to complete the hole drilling in one go; S2. Install a grouting structure in the casing. The grouting structure includes a grout stopper and a grout outlet. The grout stopper and the grout outlet are installed to the last section of the casing in the hole. S3, use the tube pulling machine to pull out a section of casing; S4, grouting, after this section of grouting is completed, grouting is stopped; S5, repeat S3-S4 until all grouting in the hole is completed, stop grouting, and move the grouting structure to the next hole position; S6. Repeat S2-S5 until all holes in the working area are grouted.

2. According to claim 1, a controlled grouting method for cofferdam foundation in sand and gravel formation, characterized in that: Before measuring and staking out the construction holes, complete the installation of wind, water and electricity to ensure a stable water and electricity supply at the construction site; fully level the construction site and build a slurry preparation and delivery system; When drilling with the pipe, a concentric three-wing pipe drill bit is used; When pulling out the pipe, according to the feedback of the drill rod during the drilling process, if there is an overhead layer, fill it with gravel, sand and film bags.

3. According to the method for controlled grouting of the cofferdam foundation in sand and gravel strata according to claim 1, it is characterized in that it comprises the following steps in S4: S41, data collection and preprocessing; S42, build LSTM model; S43. Adjust the grouting concentration based on the collected data and the LSTM model.

4. A controlled grouting method for cofferdam foundation in sand and gravel strata according to claim 3, characterized in that In S41: During the grouting process, pressure sensors and flow sensors are used to collect grouting pressure P and flow Q data in real time, and the corresponding timestamp t is recorded at the same time; a set of data (P i ,Q i ,t i ), build the initial data set: D={(P1,Q1,t1),(P2,Q2,t2),…,(P n ,Q n ,t n )}; An outlier detection algorithm based on statistical methods is used to clean data and process noise or outliers in the data; For pressure data P and flow data Q, calculate their mean and standard deviation σ P , σ Q , if a data point satisfies or The data point is determined to be an outlier, removed, and the missing data is supplemented using linear interpolation, where k is 2-3.

5. A controlled grouting method for cofferdam foundation in sand and gravel strata according to claim 4, characterized in that In S42: Divide the preprocessed data into training set D in chronological order train and the test set D test , of which the training set accounts for 80% and the test set accounts for 20%; the training set data is normalized, and the pressure P and flow Q are mapped to the [0,1] interval respectively. The normalization formula is: Where P min , P max is the minimum and maximum value of the pressure data in the training set, Q min , Q max is the minimum and maximum value of the flow data in the training set; Build an LSTM model, and set the model structure to include an input layer, several LSTM hidden layers, and an output layer; The input layer receives normalized pressure and flow data, and the input for each time step is (P norm,i ,Q norm,i ); The output layer outputs the predicted grouting concentration adjustment coefficient r; The LSTM model is trained using the training set, the mean square error is used as the loss function, and the optimization algorithm is Adam. The loss function formula is: In the formula, m is the number of training samples, r is i pred is the grouting concentration adjustment coefficient predicted by the model, and ritrue is the actual grouting concentration adjustment coefficient; During the training process, the model parameters are continuously adjusted until the loss function converges to obtain a trained LSTM model.

6. A controlled grouting method for cofferdam foundation in sand and gravel strata according to claim 5, characterized in that In S43: According to the grouting construction experience and actual engineering requirements, the grouting process is divided into several time periods T1, T2, …, T m , the length of each time period is Tinterval; At the beginning of each time period, the return grouting and pressure data collected before ΔT in the time period are input into the trained LSTM model, and the model outputs the grouting concentration adjustment coefficient r. The current grouting concentration C is adjusted according to the adjustment coefficient r. The adjustment formula is: C new =C×(1+r); In the formula, C new is the adjusted grouting concentration.

7. A controlled grouting method for cofferdam foundation in sand and gravel strata according to claim 1, The features in S4 are: S411, collecting pressure, flow and density data in the grouting process in real time through pressure sensors, flow sensors and density sensors, and performing cleaning, normalization and standardization processing; S412, using the trained multi-layer perceptron neural network model, inputting the real-time collected pressure, flow and density data, and predicting the optimal grouting slurry concentration and grouting pressure; S413. According to the prediction results of the neural network model, the initial grouting slurry concentration and grouting pressure are set.

8. A controlled grouting method for cofferdam foundation in sand and gravel strata according to claim 7, characterized in that: First, grouting is carried out with the initial grouting slurry concentration and grouting pressure. During the grouting process: If the slurry return is normal or the pressure is within the normal range: Use normal opening and pouring ratio to pour cement slurry; When the pressure is normally started and reaches the end standard required by the design or reaches the maximum injection volume specified in the design, the grouting of this section is ended; If the return slurry is small or the pressure is low: Use cement slurry one level thicker than the initial pouring ratio for pouring; If the pressure is normally generated during the grouting process, the water-cement ratio should be changed according to the principle of water-cement ratio change; if there is no obvious effect, the paste slurry should be poured according to the principle of paste slurry change; If there is no slurry return or no pressure: Use the most concentrated first-grade cement slurry to start pouring. If the pressure is normally generated during the pouring process, continue pouring according to the water-cement ratio conversion principle. If there is still no pressure, use paste slurry for pouring. According to the effect during the pouring process, if there is obvious slurry return or pressure, pour according to the water-cement ratio conversion principle. If there is no obvious effect, use paste slurry with three levels of diffusion to pour step by step. If there is still no effect after pouring to the most concentrated first-grade paste slurry, end the pouring with limited quantity. The principle of water-cement ratio transformation is: dynamically adjust the grouting pressure according to the prediction results of the neural network model: P new =P old +α·ΔP; Among them, P new is the adjusted grouting pressure, P old is the current grouting pressure, α is the adjustment coefficient, and ΔP is the pressure change predicted by the neural network model; According to the prediction results of the neural network model, the slurry concentration is dynamically adjusted: C new =C old +β·ΔC; Among them, C new is the adjusted slurry concentration, C old is the current slurry concentration, β is the adjustment coefficient, and ΔC is the concentration change predicted by the neural network model; The principle of changing the ratio of paste slurry is: when the injection volume of a certain ratio of paste slurry reaches 1m3, and the grouting pressure and injection rate have no change or the change is not significant, the slurry is changed step by step; The K-means clustering algorithm is used to classify the grouting effects and evaluate the grouting effects under different grouting parameters; Among them, x and y are two data points, and n is the feature dimension; According to the clustering results, the grouting effects under different grouting parameters are evaluated and the optimal grouting parameter combination is selected.

9. A grouting structure, characterized in that: A controlled grouting method for a cofferdam foundation in a sand and gravel formation used in any one of claims 1 to 8, further comprising a casing (1), a grout stopper (2) being provided in the casing (1), and the grout stopper (2) being provided with a grout inlet pipe (3) and a sealing sleeve (4); One end of the slurry inlet pipe (3) passes through the slurry stopper (2) and is connected to a slurry outlet head (301), and the other end of the slurry inlet pipe (3) is connected to a grouting device, a pressure gauge and a flow meter; The slurry stopper (2) is provided with an air inlet pipe (6), one end of which is in communication with the sealing sleeve (4), the sealing sleeve (4) is bulged up for air inlet and abuts against the inner wall of the sleeve (1), and the air inlet pipe (6) is externally connected to an air inlet device; A pressure sensor (701) is provided inside the slurry stopper (2), a slider (702) is provided at the bottom of the slurry stopper (2), and the slider (702) is connected to a detection end of the pressure sensor (701); The sliding block (702) is slidably arranged inside the slurry stopper (2), and a sealing ring is provided at the sliding connection.

10. The grouting structure according to claim 9, characterized in that: The slurry stopper (2) comprises a fixing seat (201) and a connecting seat (202), and the sealing sleeve (4) is arranged on the outer periphery of the connecting seat (202); A first air cavity (203) is provided in the fixing seat (201), a second air cavity (205) is provided on the outer periphery of the connecting seat (202), the second air cavity (205) is communicated with the interior of the sealing sleeve (4), a connecting hole (204) is provided in the fixing seat (201) and the connecting seat (202), the first air cavity (203) and the second air cavity (205) are communicated through the connecting hole (204), and the air inlet pipe (6) is communicated with the first air cavity (203); A plurality of stabilizing mechanisms (5) are provided on the outer periphery of the fixing seat (201) for increasing the stability of fixing the slurry stopper (2) in the casing (1); The stabilizing mechanism (5) comprises a piston chamber (501) disposed inside the fixing seat (201), the piston chamber (501) being in communication with the first air chamber (203), a piston (502) being disposed inside the piston chamber (501), and a plurality of connecting rods (503) being disposed on a side of the piston (502) away from the first air chamber (203); A plurality of temporary storage grooves (504) are provided on the outer periphery of the fixing seat (201), and a stabilizing plate (505) is movably connected in the temporary storage groove (504), and the stabilizing plate (505) is fixedly connected to the connecting rod (503); A buffer pad (506) is provided on one side of the stabilizing plate (505) away from the connecting rod (503); The side of the stabilizing plate (505) away from the connecting rod (503) is arc-shaped.

Images