A post-grouting construction method for improving low-strain integrity detection of square pile joint

Abstract

The application discloses a post-grouting construction method for improving low-strain integrity detection of square pile joints, and belongs to the technical field of pile foundation construction, and specifically comprises the following steps: performing low-strain detection on a target square pile joint, and extracting an initial reflection coefficient and a reference wave velocity; when the initial reflection coefficient exceeds a preset threshold value, performing hierarchical internal grouting on a joint part through a pre-buried pile grouting pipe; constructing a pressure-cumulative injection amount relationship curve, calculating unit pressure grouting amount and irreversible filling index; constructing an attribution mapping relationship between the unit pressure grouting amount and the irreversible filling index and a joint defect type, establishing a defect groutability discrimination model; and determining construction parameters according to model output. Through the injection capacity and filling response dual-dimensional characteristics in the hierarchical grouting process, the application realizes accurate discrimination on internal gap damage and joint cracking damage of the square pile, and improves the consistency of low-strain integrity detection results and grouting repair decisions.

Description

Technical Field

[0001] This invention relates to the field of pile foundation construction technology, and specifically to a post-grouting construction method for improving the low-strain integrity detection of square pile joints. Background Technology

[0002] Precast concrete square piles are widely used in various projects due to their high bearing capacity and fast construction speed. However, due to limitations in transportation and lifting capacity, multiple precast pile segments need to be connected to the design depth on-site through splicing. Joints are typically achieved through welding or mechanical connection. However, the repeated impact of the pile hammer during pile driving and the sudden changes in resistance when penetrating hard soil layers easily lead to stress concentration in the joint area. Simultaneously, residual welding stress, assembly gaps in connectors, and initial defects at the interface between the end plate and the concrete all make the joint a weak point in the overall pile bearing capacity. The low-strain reflected wave method, due to its ease of operation and low cost, is a primary means of detecting pile integrity. It generates stress waves through vibration at the pile top, which propagate along the pile body, producing reflected signals at the impedance change interface, which are then received and analyzed by sensors at the pile top.

[0003] In actual engineering, grouting pipes are pre-embedded inside square piles, with the lower outlet of the grouting pipe located at the designed depth of the joint and communicating with the joint. During pile driving, the compressive and tensile stress waves generated by the pile hammer impact are transmitted and reflected at the interface of the joint end plate. Local stress concentration can lead to voids inside the end plate or cracks at the welded position. Specifically, if the voids inside the joint end plate are not connected to the outside of the pile, grout can enter the voids during internal grouting through the grouting pipe and build pressure under the constraint of the surrounding wall, achieving effective filling. However, if the joint weld is incomplete or cracked, there is a communication channel with the outside of the pile. During internal grouting, grout flows out of the pile along the cracks, making it difficult to accumulate hydrostatic pressure within the defect space, thus limiting the effectiveness of internal grouting. However, existing technologies only judge the grouting effect based on the change in reflection coefficient before and after grouting, lacking a means to dynamically distinguish between the two types of defects during internal grouting. This makes it impossible to predict the connectivity of defects before or during grouting construction, leading to incorrect grouting path selection and affecting the quality of square pile joint integrity repair and the reliability of low-strain test results.

[0004] Therefore, how to provide a post-grouting construction method to improve the low-strain integrity detection of square pile joints, utilizing pre-embedded grout pipes within the square pile for graded internal grouting, and establishing an injectability identification mechanism for internal void damage and joint cracking damage of the square pile by coupling the dynamic response of pressure-cumulative injection volume during the grouting process with the wave velocity-reflection coefficient under stable and depressurized states, and adaptively determining construction parameters for continuing internal grouting or switching to external grouting based on the identification results, thereby achieving accurate classification and treatment of square pile joint defects and effectively improving the reliability of low-strain detection, is an urgent problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a post-grouting construction method for improving the low-strain integrity detection of square pile joints, and to solve the following technical problems: Existing grouting methods for square pile joints only evaluate the repair effect based on low-strain reflection coefficient anomalies and changes in reflection coefficient after grouting. They do not use pressure-injection volume and reflection coefficient response to distinguish between internal void damage and joint cracking damage in the square pile during the graded internal grouting process, and lack a closed-loop decision-making mechanism to dynamically switch between internal and external grouting based on the discrimination results.

[0006] The objective of this invention can be achieved through the following technical solutions: A post-grouting construction method for improving the low-strain integrity detection of square pile joints, wherein a grouting pipe is pre-embedded in the square pile and the lower end outlet of the grouting pipe is located at the designed depth of the joint and communicates with the joint location, the joint location being connected by welding; the method includes the following steps: S1, perform low-strain testing on the joint of the target square pile, extract the initial reflection coefficient and initial pile bottom reflection time at the design depth of the joint, and calculate the reference wave velocity; S2, when the initial reflection coefficient is greater than the preset reflection coefficient threshold, the joint is internally grouted in stages through the pile grouting pipe, and low strain detection is performed under the pressure stabilization state of each preset grouting pressure level to obtain the grouting pressure, cumulative injection volume, pressure stabilization reflection coefficient and pressure stabilization wave velocity. After depressurization, the recovered reflection coefficient and recovered wave velocity are obtained. S3, construct a pressure-cumulative injection volume relationship curve based on the grouting pressure and cumulative injection volume corresponding to each pressure level, calculate the unit pressure grouting volume based on the pressure-cumulative injection volume relationship curve, and calculate the irreversible filling index based on the initial reflection coefficient, the stable pressure reflection coefficient, and the recovery reflection coefficient; S4. Construct the attribution mapping relationship between unit pressure grouting volume and irreversible filling index and square pile joint defect type, and establish a square pile joint defect injectionability discrimination model. S5. Obtain the unit pressure grouting volume and irreversible filling index of the target square pile and input them into the defect injectability discrimination model. Determine the grouting construction parameters of the target square pile based on the output of the defect injectability discrimination model.

[0007] As a further aspect of the present invention: the specific process of performing low-strain testing on the joint in S1 is as follows: Velocity response signals under vertical excitation are collected by a velocity sensor placed on the pile top and averaged multiple times, wherein the vertical excitation is applied along the central axis of the pile top; the start time of the incident wave, the peak amplitude of the incident wave, and the start time of the reflected wave at the pile bottom are marked, and the difference between the start time of the reflected wave at the pile bottom and the start time of the incident wave is used as the initial pile bottom reflection time; the reference wave velocity is obtained by dividing twice the designed pile length by the initial pile bottom reflection time; the joint reflection time is calculated according to the joint design depth and the reference wave velocity, and the peak amplitude of the reflected wave is extracted within a preset time window with the joint reflection time as the center, and the peak amplitude of the reflected wave is divided by the peak amplitude of the incident wave to obtain the initial reflection coefficient.

[0008] As a further aspect of the present invention: in step S2, the specific process for obtaining the grouting pressure, cumulative injection volume, pressure-stabilizing reflection coefficient, and pressure-stabilizing wave velocity is as follows: During the graded grouting process, when the grouting pressure reaches any preset grouting pressure level and is maintained at a stable pressure, the current grouting pressure and cumulative injection volume are recorded, and the joint of the target square pile is subjected to low-strain testing again to obtain the stable pressure reflection coefficient and stable pressure wave velocity.

[0009] As a further aspect of the present invention: in step S2, the specific process of obtaining the restored reflection coefficient and restored wave velocity after depressurization is as follows: After closing the grouting valve and depressurizing to normal pressure and waiting for a preset time, low-strain testing is performed on the joint of the target square pile, and the obtained reflection coefficient and wave velocity are marked as the restored reflection coefficient and restored wave velocity, respectively.

[0010] As a further aspect of the present invention: the specific calculation process for the unit pressure grouting volume in S3 is as follows: From the pressure-cumulative injection volume relationship curve, extract the cumulative injection volume corresponding to each preset grouting pressure level; subtract the cumulative injection volume corresponding to two adjacent preset grouting pressure levels to obtain the cumulative injection volume difference; subtract the grouting pressure difference from the two adjacent preset grouting pressure levels to obtain the grouting pressure difference; divide the cumulative injection volume difference by the grouting pressure difference to obtain the unit pressure grouting volume for each pressure level segment; perform an arithmetic mean on the unit pressure grouting volumes for each pressure level segment to obtain the unit pressure grouting volume.

[0011] As a further aspect of the present invention: the specific calculation process of the irreversible filling index in S3 is as follows: For any preset grouting pressure level, the difference between the initial reflection coefficient and the recovered reflection coefficient is taken as the change in reflection coefficient caused by irreversible filling; the change in reflection coefficient caused by irreversible filling is divided by the initial reflection coefficient to obtain the irreversible filling index under that pressure level; the irreversible filling index under each pressure level is arithmetically averaged to obtain the irreversible filling index.

[0012] As a further aspect of the present invention: the specific construction process of the square pile joint defect injectability discrimination model in S4 is as follows: The unit pressure grouting volume and irreversible filling index of multiple square piles during the graded grouting process are obtained; the joint defect type of each square pile is determined by core drilling verification, wherein the joint defect type is internal void damage or joint cracking damage; a training sample set is constructed using the unit pressure grouting volume and the irreversible filling index as input features, and the internal void damage and joint cracking damage of the square pile as output labels, respectively; the training sample set is trained using a decision tree algorithm, and the feature space of the input features is recursively divided using the minimization of the Gini index as the splitting criterion, to generate a square pile joint defect injectability discrimination model with the unit pressure grouting volume and the irreversible filling index as input and the joint defect type as output.

[0013] As a further aspect of the present invention: in step S5, the specific process for determining the grouting construction parameters of the target square pile is as follows: The unit pressure grouting volume and irreversible filling index of the target square pile are input into the square pile joint defect injectability discrimination model to obtain the defect type output by the model. When the defect type is internal void damage of the square pile, the grouting construction parameters are determined to continue internal grouting through the pile grouting pipe to the design final pressure, and grouting is stopped after the final pressure is stabilized for a preset time. When the defect type is joint cracking damage, the grouting construction parameters are determined to terminate internal grouting, drill a hole at the pile side position corresponding to the joint to install an external grouting pipe, and externally grout the defect space through the external grouting pipe to the design final pressure and stabilize the pressure.

[0014] As a further aspect of the present invention: it also includes, after the grout has solidified for a preset time after grouting is completed, a low-strain test is performed again to obtain the reflection coefficient after solidification; the difference between the initial reflection coefficient and the reflection coefficient after solidification is used as the reduction in reflection coefficient; if the reduction in reflection coefficient is greater than or equal to a preset reduction threshold, the grouting is deemed qualified; otherwise, core drilling verification is performed: if the defect type output by the defect injectability discrimination model is joint cracking damage, and the joint defect type determined by the core drilling is internal void damage of the square pile, then the unit pressure grouting volume and irreversible filling index of the target square pile are used as new features, and the internal void damage of the square pile is used as a label to form new training samples, and the square pile joint defect injectability discrimination model is updated; if the joint defect type determined by the core drilling is joint cracking damage, the grouting is deemed unqualified, and external grouting is re-performed on the target square pile.

[0015] The beneficial effects of this invention are: 1) This invention performs low-strain testing on the target square pile joint under both the stabilizing and depressurization recovery states at each preset pressure level during graded internal grouting. The stabilizing reflection coefficient and the recovery reflection coefficient are obtained, and the unit pressure grouting volume for each pressure level segment is obtained from the pressure-cumulative injection volume relationship curve. Simultaneously, the irreversible filling index is calculated using the difference between the initial reflection coefficient and the recovery reflection coefficient. It can be understood that when the defect space at the lower end of the pre-embedded grout pipe inside the square pile is not connected to the outside of the pile, the grout is constrained by the surrounding wall after entering the defect space. The cumulative injection volume increases significantly with increasing pressure, resulting in a high unit pressure grouting volume. After grout filling, the joint wave impedance tends to be continuous, and the reflection coefficient irreversibly decreases after depressurization, leading to a high irreversible filling index. When there is a connecting channel between the defect space and the outside of the pile, the grout flows out of the pile along a low-resistance path. The cumulative injection volume is difficult to convert into pressure accumulation within the defect space, resulting in a low or abnormal unit pressure grouting volume. After depressurization, the reflection coefficient basically recovers to the pre-grouting level, and the irreversible filling index is low. Therefore, the unit pressure grouting volume and irreversible filling index reflect the connectivity properties of the defect from the two dimensions of injection capacity and filling response, respectively, realizing an essential leap from relying on the single reflection coefficient change after grouting to judging the injectability of defects based on the dual characteristics of injection and filling during construction.

[0016] 2) This invention uses both unit pressure grouting volume and irreversible filling index as input features to construct an attribution mapping relationship with the type of defects in square pile joints. A decision tree algorithm is used to generate a defect injectability discrimination model by minimizing the Gini index, and grouting construction parameters are determined based on the defect type output by the model. It is understood that defects corresponding to internal void damage in square piles have closed defect spaces, allowing grout to enter through the pile grout pipe and effectively fill under pressure; therefore, both unit pressure grouting volume and irreversible filling index are high. Defect spaces corresponding to joint cracking damage are connected to the outside of the pile, resulting in continuous grout loss during internal grouting, making effective filling difficult; therefore, unit pressure grouting volume is low or abnormal, and the irreversible filling index is low. The decision tree, through recursive partitioning of the feature space, can automatically capture the essential differences in injection capacity and filling response between the two types of damage, thereby predicting defect injectability during construction. For joints identified as having internal void damage within the square pile, grouting continues through the pile grout pipe until the designed final pressure is reached and stabilized. For joints identified as having crack damage, internal grouting is terminated and external grouting is switched to external grouting, using the connecting channel as the grout inlet path for filling. This process changes the selection of the grouting path from uniform internal grouting to adaptive switching based on the defect connectivity attributes, avoiding repeated pressurization of defects that cannot be repaired by internal grouting, which would cause grout loss and disturbance of the surrounding soil.

[0017] 3) This invention performs low-strain testing on the target square pile joint again after the grouting construction is completed and the grout has solidified for a preset time. The reflection coefficient after solidification is compared with the initial reflection coefficient to obtain the reduction in reflection coefficient. When the reduction is lower than a preset threshold, core drilling verification is performed, and the defect injectability discrimination model is updated based on the core drilling results. It is understood that the reduction in reflection coefficient after solidification reflects the degree of improvement in the continuity of wave impedance in the joint area. Insufficient reduction may be due to incorrect selection of the internal grouting path or inadequate execution of the grouting process. Core drilling verification can directly determine the actual defect type. When the core drilling results show that the actual defect type is inconsistent with the model's discrimination results, it indicates that there is a deviation in the attribution mapping of this sample. Adding the unit pressure grouting volume, irreversible filling index, and actual defect type of this sample to the training set to retrain the decision tree can correct the split boundary of the model in this feature space region, making the discrimination results of subsequent similar samples closer to the actual defect type. If the core drilling results indicate joint cracking damage and the previous grouting path was not adjusted for this type, external grouting is performed again. This forms a closed-loop feedback mechanism encompassing internal grouting discrimination, grouting path switching, post-solidification testing, core drilling verification, and model updates. This allows the discrimination model to be gradually optimized as construction data accumulates, achieving a technological leap from one-time static discrimination to a dynamic closed-loop system of testing, grouting, verification, and updates. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a schematic diagram of a post-grouting construction method for improving the low-strain integrity detection of square pile joints according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, this invention is a post-grouting construction method for improving the low-strain integrity detection of square pile joints. A grouting pipe is pre-embedded within the square pile, and the lower outlet of the grouting pipe is located at the designed depth of the joint and communicates with the joint. The joint is connected by welding. The method includes the following steps: S1, perform low-strain testing on the joint of the target square pile, extract the initial reflection coefficient and initial pile bottom reflection time at the design depth of the joint, and calculate the reference wave velocity; A velocity sensor is installed on the top surface of the target square pile, positioned near the center of the pile top and tightly bonded to the pile surface using a coupling agent to minimize the contact impedance difference between the sensor and the pile top. A vibratory hammer is used to vertically excite the pile along its central axis. The hammer head material is selected based on the pile length and detection depth; for example, a hard nylon hammer head is used for shorter piles or shallower defects to generate higher-frequency stress waves, while an aluminum hammer head is used for longer piles to generate lower-frequency stress waves. The velocity response signal output from the velocity sensor is acquired using a signal acquisition device, with a sampling frequency of at least 20kHz and a sampling time sufficient to completely record the reflected wave from the pile bottom. Multiple velocity response signals acquired during excitation are superimposed and averaged at least 10 times. This ensures that the incident wave, joint reflection wave, and pile bottom reflection wave, all with consistent waveform characteristics, reinforce each other during averaging, while environmental vibrations and random electrical noise cancel each other out due to phase inconsistencies, resulting in an average velocity response signal with a high signal-to-noise ratio.

[0022] In the time-domain waveform of the average velocity response signal, the incident wave initiation time and the incident wave peak amplitude are marked. The incident wave initiation time is the moment when the leading edge of the first wave generated by excitation reaches the sensor, serving as the time zero point; the incident wave peak amplitude is the maximum absolute amplitude within the first wave packet, representing the energy strength of the incident wave. The initiation time of the pile bottom reflection wave is also marked, which is the moment when the stress wave propagates downward from the pile top to the pile bottom and reflects back to the pile top. The initial pile bottom reflection time is obtained by subtracting the pile bottom reflection time from the incident wave initiation time. Since the stress wave travels twice the pile length in one round trip within the pile, and the elastic wave velocity of the pile concrete is approximately constant under stress-free conditions, the reference wave velocity can be obtained by dividing twice the design pile length by the initial pile bottom reflection time using the known design pile length. The reference wave velocity reflects the average elastic wave propagation velocity of the pile concrete under natural conditions before grouting.

[0023] The joint reflection time is calculated based on the joint design depth and the reference wave velocity. The joint reflection time is the time required for the stress wave to propagate from the pile top to the joint design depth and then reflect back to the pile top; its value is twice the joint design depth divided by the reference wave velocity. The reason the joint reflection time can be calculated using the reference wave velocity is that the propagation speed of the stress wave in the pile is known, and the joint depth corresponds to a specific propagation time. This time allows us to determine the location of the reflected wave in the time-domain waveform. Using the joint reflection time as the center, the peak amplitude of the reflected wave is extracted within a preset time window. The length of the preset time window should completely cover the joint reflected wave packet to avoid overlap with adjacent reflected waves or noise. The peak amplitude of the reflected wave is divided by the peak amplitude of the incident wave to obtain the initial reflection coefficient. The initial reflection coefficient characterizes the degree of change in wave impedance at the joint location. When there are gaps on the inner side of the joint end plate, or defects such as incomplete welding or cracks at the weld location, the wave impedance at that location decreases, and the reflection coefficient increases. When the joint is intact, the wave impedance is continuous, and the reflection coefficient is very small. This yields the initial reflection coefficient and the reference wave velocity before grouting, providing a reference value for parameter calculations in subsequent grouting construction.

[0024] S2, when the initial reflection coefficient is greater than the preset reflection coefficient threshold, the joint is internally grouted in stages through the pile grouting pipe, and low strain detection is performed under the pressure stabilization state of each preset grouting pressure level to obtain the grouting pressure, cumulative injection volume, pressure stabilization reflection coefficient and pressure stabilization wave velocity. After depressurization, the recovered reflection coefficient and recovered wave velocity are obtained. The preset reflection coefficient threshold is determined based on engineering experience and field tests. For example, for precast square pile joints, when the initial reflection coefficient is greater than 0.2, the joint is considered to have obvious defects and requires grouting repair. When the initial reflection coefficient exceeds this threshold, internal grouting is performed using the grouting pipe pre-embedded in the square pile. The lower outlet of the grouting pipe is located at the designed depth of the joint and is directly connected to the joint. Grout enters the joint defect space through the lower outlet of the grouting pipe. To ensure that the graded grouting data truly reflects the injectability of the joint defect itself, a low-pressure test injection can be applied through the grouting pipe before formal graded grouting to observe the changes in grouting pressure and injection volume response. If the pressure rises rapidly under low pressure and the injection volume is close to zero, it indicates that the lower outlet of the grouting pipe is blocked or the joint defect space does not meet the conditions for grouting, and it is necessary to clear the grouting pipe or adjust the grouting strategy first; if the pressure rises slowly under low pressure and is accompanied by an injection volume, it indicates that the grouting pipe is unobstructed and the joint defect space can be grouted, and graded grouting can proceed directly. During staged grouting, the preset grouting pressure level is determined based on the pile diameter, pile length, geological conditions, and grouting material. The difference between adjacent pressure levels is generally 0.2 MPa to 0.5 MPa, for example, set sequentially to 0.3 MPa, 0.6 MPa, 0.9 MPa, and 1.2 MPa. When each grouting pressure reaches the preset value and remains stable, the stabilization time is maintained for 3 to 5 minutes, allowing the grout to fully flow and penetrate within the defect space, while simultaneously ensuring a relatively stable stress state for the pile body and surrounding soil at that pressure level. During the stabilization state, the pressure value displayed by the grouting system and the cumulative injection volume are recorded. The cumulative injection volume is the total volume of grout injected through the pile grout pipe from the start of grouting to the current moment.

[0025] During staged grouting, the preset grouting pressure level is determined based on the pile diameter, pile length, geological conditions, and grouting material. The difference between adjacent pressure levels is generally 0.2 MPa to 0.5 MPa, for example, set sequentially as 0.3 MPa, 0.6 MPa, 0.9 MPa, and 1.2 MPa. When the grouting pressure reaches the preset value and remains stable at each level, the stabilization time is maintained for 3 to 5 minutes, allowing the grout to fully flow and permeate within the defect space, while simultaneously ensuring a relatively stable stress state for the pile body and surrounding soil at that pressure level. During the stabilization state, the pressure value displayed by the grouting system and the cumulative injection volume are recorded. The cumulative injection volume is the total volume of grout injected through the pile grout pipe from the start of grouting to the current moment. Subsequently, low-strain testing is performed on the target square pile, and the stabilization velocity response signal is collected, superimposed, and averaged to obtain the average stabilization velocity response signal. In the time-domain waveform of the average velocity response signal under pressure stabilization, the start time and peak amplitude of the incident wave are marked, as is the start time of the reflected wave at the pile bottom. The difference between the start time of the reflected wave at the pile bottom and the start time of the incident wave is taken as the reflection time at the bottom of the pressure-stabilized pile. The pressure-stabilized wave velocity is calculated based on the designed pile length and the reflection time at the bottom of the pressure-stabilized pile. The pressure-stabilized wave velocity reflects the average elastic wave propagation velocity of the pile concrete under the compression state of grouting pressure. The reflection time of the pressure-stabilized joint is calculated based on the designed joint depth and the pressure-stabilized wave velocity. The peak amplitude of the reflected wave is extracted within a preset time window centered on the reflection time of the pressure-stabilized joint. The peak amplitude of the reflected wave is divided by the peak amplitude of the incident wave to obtain the pressure-stabilized reflection coefficient.

[0026] After completing the pressure stabilization test, the grouting valve is closed to release pressure to atmospheric pressure, and a preset time is waited for, typically 5 to 10 minutes, to ensure sufficient recovery of elastic deformation of the pile body and surrounding soil, and to allow the grout flow within the defect space to stabilize. Then, low-strain testing is performed on the target square pile again, collecting and averaging the recovery velocity response signals to obtain the average recovery velocity response signal. In the time-domain waveform of the average recovery velocity response signal, the incident wave initiation time and peak amplitude are marked, as is the pile bottom reflection wave initiation time. The difference between the pile bottom reflection wave initiation time and the incident wave initiation time is taken as the recovery pile bottom reflection time. The recovery wave velocity is calculated based on the designed pile length and the recovery pile bottom reflection time. The recovery joint reflection time is calculated based on the joint design depth and the recovery wave velocity. The peak amplitude of the recovery reflection wave is extracted within a preset time window, centered on the recovery joint reflection time. The recovery reflection coefficient is obtained by dividing the peak amplitude of the recovery reflection wave by the peak amplitude of the incident wave.

[0027] Understandably, under stable pressure, the interface between the pile concrete and the joint defects is in a compressive elastic state, with increased wave velocity and decreased reflection coefficient. This change recovers as pressure is unloaded. However, grout filling of defects alters the joint medium impedance, a change that does not recover after pressure release. The recovered reflection coefficient and recovered wave velocity reflect the residual filling state of the joint defect space after pressure release. By collecting the reflection coefficient and wave velocity under stable pressure and pressure release recovery conditions respectively, it is possible to distinguish between the elastic changes caused by pressure and the irreversible changes caused by grout filling.

[0028] S3, construct a pressure-cumulative injection volume relationship curve based on the grouting pressure and cumulative injection volume corresponding to each pressure level, calculate the unit pressure grouting volume based on the pressure-cumulative injection volume relationship curve, and calculate the irreversible filling index based on the initial reflection coefficient, the stable pressure reflection coefficient, and the recovery reflection coefficient; Plotting grouting pressure on the x-axis and cumulative injection volume on the y-axis, the data points of grouting pressure and cumulative injection volume recorded at each preset grouting pressure level are plotted on a coordinate system and connected to obtain a pressure-cumulative injection volume relationship curve. This curve reflects the change in grout injection volume with increasing grouting pressure. The cumulative injection volume corresponding to each preset grouting pressure level is extracted from the curve. The difference between the cumulative injection volumes corresponding to two adjacent preset grouting pressure levels is obtained; the difference between two adjacent preset grouting pressure levels is obtained; and the difference between the cumulative injection volumes and the difference in grouting pressure is divided to obtain the unit pressure grouting volume for each pressure level segment. For example, when the grouting pressure increases from 0.3 MPa to 0.6 MPa, the cumulative injection volume increases from 5 L to 14 L, the difference in cumulative injection volume is 9 L, and the difference in grouting pressure is 0.3 MPa. Therefore, the unit pressure grouting volume for this pressure level segment is 30 L / MPa. The grouting volume per unit pressure is obtained by arithmetically averaging the grouting volume per unit pressure for each pressure level segment, thus yielding the grouting volume per unit pressure throughout the entire grading grouting process. The grouting volume per unit pressure reflects the average increase in grout injection volume under a unit increase in grouting pressure; the larger the value, the stronger the capacity of the defect space to accommodate the grout.

[0029] For any preset grouting pressure level, the change in reflection coefficient caused by irreversible filling is obtained by subtracting the recovered reflection coefficient from the initial reflection coefficient. This change represents the degree of permanent reduction in the joint reflection coefficient from before grouting to after pressure relief, i.e., the permanent change in joint wave impedance after the grout enters the defect space. Dividing the change in reflection coefficient caused by irreversible filling by the initial reflection coefficient yields the irreversible filling index at that pressure level. The arithmetic mean of the irreversible filling indices at each pressure level is used to obtain the irreversible filling index for the entire staged grouting process. The index generally ranges from 0 to 1. The closer it is to 1, the more significant the reduction in reflection coefficient after pressure relief, and the more fully the grout permanently fills the defect; the closer it is to 0, the more the reflection coefficient basically recovers to the pre-grouting level after pressure relief, and the grout fails to effectively enter the defect space and form irreversible filling. This index, by directly comparing the difference in reflection coefficients between the initial and recovered states, naturally eliminates the influence of stress elastic deformation under stable pressure, eliminating the need for additional stress correction.

[0030] S4. Construct the attribution mapping relationship between unit pressure grouting volume and irreversible filling index and square pile joint defect type, and establish a square pile joint defect injectionability discrimination model. Multiple square piles with known actual defect types and completed graded internal grouting were acquired as historical samples. For each historical square pile, the unit pressure grouting volume and irreversible filling index during the graded grouting process were calculated according to the aforementioned steps, yielding two characteristic values. Simultaneously, the actual defect type at the joint of the square pile was determined through core sampling or low-strain testing after grout solidification. During core sampling, a core sampling device was used to drill concrete core samples at the joint location to observe the connectivity of defects within the core samples. If the core sample shows that the defect is located inside the joint end plate, and the void is enclosed by the end plate and weld, with no connecting channel for grout to pass through to the outside of the pile, the grout can remain in the void and build up pressure after entering through the pile grout pipe. Therefore, the joint defect type of this sample is determined to be internal void damage of the square pile. If the core sample shows that there is a crack extending from the defect space to the outside of the pile at the joint, or that there is an open channel formed by incomplete filling or cracking at the weld position, the grout injected internally will flow out of the pile along the crack or channel, making it difficult to build up pressure and form effective filling in the defect space. Therefore, the joint defect type of this sample is determined to be joint cracking damage. The process of determining the defect type by low-strain testing after grouting solidification is as follows: After grouting is completed and the grout has solidified for a certain period of time, low-strain testing is performed again on the target square pile joint, and the reflection coefficient of the solidified joint is compared with the initial reflection coefficient before grouting. If the reflection coefficient after solidification is significantly lower than the initial reflection coefficient, it indicates that the grout has entered the defect space and changed the joint wave impedance, and is therefore determined to be internal void damage of the square pile; if the reflection coefficient after solidification is basically unchanged compared with the initial reflection coefficient, it indicates that the grout has failed to effectively enter the defect space, and is therefore determined to be joint cracking damage.

[0031] For each historical pile, the unit pressure grouting volume and irreversible filling index of the pile are used as two input feature values, and the actual defect type of the pile determined by the above method is used as the output category, forming a training sample. Training samples from multiple historical piles together form a training sample set. A decision tree algorithm is used to train the training sample set, recursively partitioning the feature space of the input features using the minimization of the Gini index as the splitting criterion. Specifically, all training samples are placed at the root node, and the Gini index of the current node is calculated. Then, the unit pressure grouting volume and irreversible filling index are examined separately, and multiple possible splitting positions are tried for each feature, dividing the sample into two subsets. The weighted sum of the Gini indices of the two child nodes after the split is calculated. The feature and splitting position that minimizes the weighted sum are selected as the basis for this split, dividing the current node sample into two child nodes. The above process is repeated for each child node to continue searching for the optimal split until all samples within the node belong to the same defect type or further splitting is impossible. In the final generated decision tree, each internal node stores the feature and splitting position used in this split, and each leaf node corresponds to a defect type. This decision tree is a model for judging the injectability of defects in square pile joints. The model input is the unit pressure grouting volume and the irreversible filling index, and the output is the internal void damage or joint cracking damage of the square pile.

[0032] S5. Obtain the unit pressure grouting volume and irreversible filling index of the target square pile and input them into the defect injectability discrimination model. Determine the grouting construction parameters of the target square pile based on the output of the defect injectability discrimination model.

[0033] For the target square pile to be treated, after completing the graded internal grouting and corresponding testing, the unit pressure grouting volume and irreversible filling index of the target square pile are calculated according to the aforementioned steps. These two characteristic values ​​are input into the square pile joint defect injectability discrimination model. Starting from the root node of the decision tree, the model compares layer by layer. Based on the splitting features and splitting position saved at the current node, it determines whether to enter the left or right branch until it reaches the leaf node. The defect type marked by the leaf node is the defect type output by the model. When the defect type output by the model is internal void damage of the square pile, it indicates that there is a void inside the joint end plate that is wrapped by the end plate and the weld. The lower end outlet of the pile grout pipe is connected to this void, and there is no connecting channel between the void and the outside of the pile for grout to flow out. The grout can enter the defect space through the pile grout pipe and establish pressure under the constraint of the peripheral wall to achieve effective filling. At this time, the grouting construction parameters are determined to continue internal grouting through the pile grout pipe to the design final pressure, and stop grouting after the final pressure stabilization preset time. The final design pressure is predetermined based on the concrete strength of the target square pile and the geological conditions at the joint. The stabilization time is predetermined based on the flow properties of the grout and the geometric dimensions of the defect space. When the model outputs a defect type of joint cracking damage, it indicates the existence of a defect channel connecting the joint to the outside of the pile. Continuing internal grouting will not allow the grout to establish pressure within the defect space, only causing the grout to leak along the channel. In this case, the grouting parameters are set to terminate internal grouting. An external grouting pipe is drilled at the pile side location corresponding to the joint, and external grouting is performed on the defect space through the external grouting pipe until the final design pressure is reached and stabilized. During external grouting, the grout directly enters the connecting channel from the pile side and fills the defect space, using the channel as the grout entry path to achieve effective filling.

[0034] This invention utilizes multi-state low-strain detection during the graded internal grouting process to obtain the pressure-stabilized reflection coefficient and the restored reflection coefficient. The grouting volume per unit pressure is calculated from the pressure-cumulative injection volume relationship curve, and the irreversible filling index is calculated from the difference between the initial reflection coefficient and the restored reflection coefficient. These two factors are then used as input features to determine whether the defects in the square pile joint are internal void damage or joint cracking damage through a decision tree model. For internal void damage, internal grouting continues until the design final pressure; for joint cracking damage, internal grouting is terminated promptly and switched to external grouting. This avoids grout loss and soil disturbance caused by repeated pressurization of defects that cannot be repaired by internal grouting. This invention achieves on-site prediction of the injectability of square pile joint defects and adaptive switching of grouting paths, improving the consistency between low-strain integrity detection results and grouting repair decisions.

[0035] It is worth noting that diagnostic methods for internal defects in square pile joints already exist in this field, such as core sampling and sonic logging. Core sampling, which involves drilling concrete core samples at the joint to directly observe the defect morphology and connectivity, is a destructive testing method. The drilling process causes localized damage to the pile structure, and the core hole needs to be sealed and repaired. Furthermore, single-hole core sampling only reflects defect information near the hole location, offering limited representativeness of the spatial distribution of defects. Additionally, core sampling is costly and time-consuming, making it difficult to apply in real-time during grouting. Sonic logging requires pre-embedding sonic logging tubes during pile casting or splicing to determine joint integrity based on the sound wave propagation characteristics between the tubes. However, for precast square pile joints that have already been constructed, the on-site conditions for pre-embedding sonic logging tubes are often unavailable. Moreover, the logging tubes may become misaligned, blocked, or detached from the concrete during construction, leading to distorted sound signals and reduced reliability of the test results. None of the above methods can effectively determine the connectivity of defects in real-time during grouting, nor can they provide a basis for the dynamic selection of the grouting path. This invention utilizes pre-embedded grout pipes within square piles for graded internal grouting. During the grouting process, low-strain detection is performed at each preset pressure level under both stabilization and depressurization / recovery states. A pressure-cumulative injection volume relationship curve is constructed to calculate the grouting volume per unit pressure, and the irreversible filling index is calculated using the difference between the initial and recovered reflection coefficients. Based on this, a defect injectability discrimination model is built. This process does not rely on core sampling or pre-embedded sonic logging pipes and is applicable to precast square pile joints that have already been constructed. It can quickly determine whether joint defects are due to internal void damage or joint cracking damage based on real-time acquired unit pressure grouting volume and irreversible filling index during grouting construction. For joints identified as having internal void damage, grouting continues through the grout pipes until the designed final pressure is reached and stabilized, ensuring the grout fully fills and seals the defect space. For joints identified as having joint cracking damage, internal grouting is promptly terminated and switched to external grouting, using a connecting channel as the grout entry path for filling. This avoids repeated pressurization of defects inaccessible by internal grouting, preventing grout loss and disturbance of the surrounding soil. Therefore, in addition to existing core sampling and sonic logging methods, this invention provides a defect groutability assessment method applicable to completed square pile joints and which can be implemented simultaneously during grouting. This method overcomes the shortcomings of existing internal defect diagnosis methods, which are unable to dynamically guide the selection of grouting paths, and improves the coordination between low-strain integrity detection results and grouting repair decisions.

[0036] In a preferred embodiment of the present invention, the specific process of performing low-strain testing on the joint portion of the target square pile in step S1 is as follows: First, a velocity sensor is installed at a flat location on the pile top surface. Coupling agent is applied between the sensor and the pile top to ensure tight contact, reducing signal attenuation and multiple reflections at the interface. A vibratory hammer is used to vertically excite the pile along its central axis. The hammer strikes the pile top, generating a stress wave that propagates downwards. This stress wave is reflected when it encounters an interface with changing wave impedance, and the reflected wave propagates upwards back to the pile top where it is received by the velocity sensor. Since the signal generated by a single excitation contains environmental vibrations and random electrical noise, the velocity response signals acquired from multiple excitations are superimposed and averaged. Incident waves, joint reflections, and pile bottom reflections with consistent waveform characteristics from each signal are superimposed in phase to enhance them. Random noise cancels out during the averaging process due to phase inconsistencies, resulting in an average velocity response signal with a high signal-to-noise ratio. In the time-domain waveform of the average velocity response signal, the incident wave initiation time is first marked. This initiation time is the moment the leading edge of the first wave generated by excitation reaches the sensor, representing the zero point at which the stress wave begins to be received. Simultaneously, the peak amplitude of the incident wave is marked, i.e., the maximum absolute amplitude of the first wave packet, which reflects the energy of the incident wave. Then, the initiation time of the pile bottom reflection wave is marked. This initiation time is the moment when the stress wave propagates downwards from the pile top to the pile bottom and reflects back to the pile top. The difference between the pile bottom reflection wave initiation time and the incident wave initiation time yields the initial pile bottom reflection time. Since the stress wave travels twice the pile length in one round trip within the pile, and the wave velocity is approximately constant within the pile concrete, the reference wave velocity can be obtained by dividing twice the design pile length by the initial pile bottom reflection time using the known design pile length. This reference wave velocity represents the average elastic wave propagation velocity of the pile concrete before grouting. Based on the joint design depth and the reference wave velocity, the joint reflection time can be calculated. The joint reflection time is the time required for the stress wave to propagate downwards from the pile top to the joint design depth and then reflect back to the pile top; its value is twice the joint design depth divided by the reference wave velocity. This calculation is performed because the propagation speed of the stress wave within the pile is known, and the joint depth corresponds to a specific propagation time. This time allows us to determine the location of the joint reflection wave in the time-domain waveform. A preset time window is taken, centered on the joint reflection time, and the peak amplitude of the reflected wave is searched and extracted within this window. A time window is necessary because the actual joint reflection wave is not an isolated pulse but a wave packet containing multiple extreme points. Directly taking the amplitude at a single moment is easily affected by noise or interference from adjacent reflections. Taking a window of a certain width centered on the joint reflection time can completely encompass the joint reflection wave packet, and the maximum absolute amplitude within the window can be extracted as the peak amplitude of the reflected wave. Dividing the peak amplitude of the reflected wave by the peak amplitude of the incident wave yields the initial reflection coefficient. The reflection coefficient characterizes the degree of change in wave impedance at the joint location. When the joint has defects such as voids, mud inclusions, or cracks, the wave impedance at that location decreases and the reflection coefficient increases. When the joint is intact, the wave impedance is continuous and the reflection coefficient is very small.This yields the initial reflection coefficient and reference wave velocity before grouting, providing a reference value for stress correction and irreversible filling index calculation during subsequent grouting construction.

[0037] It is understandable that by performing multiple superposition and averaging processes, random disturbances caused by environmental noise and poor sensor contact are effectively suppressed, making the waveform characteristics of the incident wave, joint reflection wave, and pile bottom reflection wave clearer, thereby ensuring the accurate acquisition of the start-up time and peak amplitude. By accurately marking the start-up time of the incident wave and the start-up time of the pile bottom reflection wave, and calculating the reference wave velocity using the designed pile length, systematic errors caused by excitation delay or sensor response time can be eliminated, obtaining the true wave velocity of the pile body under stress-free conditions before grouting. Determining the joint reflection time based on the joint design depth and the reference wave velocity, and then extracting the peak amplitude of the reflected wave by truncating a window centered on this time, can separate the joint reflection from the pile bottom reflection and other secondary reflections, avoiding mistakenly treating the reflection waves from other interfaces as joint reflections, and ensuring that the initial reflection coefficient truly reflects the impedance change at the joint. The initial reflection coefficient, as a quantitative indicator of the severity of defects before grouting, is the benchmark for determining whether post-grouting construction is needed and for comparing the subsequent grouting effects; the reference wave velocity provides an initial reference for analyzing wave velocity changes during the grouting process. These two parameters together constitute the benchmark conditions for stress correction and irreversible filling index calculation in the subsequent grouting construction process, providing a reliable data basis for the final determination of the injectability of square pile joint defects and the determination of grouting construction parameters.

[0038] In another preferred embodiment of the present invention, the specific process of obtaining the grouting pressure, cumulative injection volume, pressure-stabilizing reflection coefficient, and pressure-stabilizing wave velocity in step S2 is as follows: During the graded grouting process, when the grouting pressure reaches any preset grouting pressure level and remains stable, the current grouting pressure value displayed by the pressure gauge installed on the grouting system is first read. Simultaneously, the cumulative grout volume injected from the start of grouting to the current moment is read from the cumulative flow recording device on the grouting equipment. This yields the grouting pressure and cumulative injection volume at that pressure level. At the same time, low-strain testing is performed on the target square pile. A velocity sensor is placed at the pile top and vertically vibrated along the central axis of the pile top to collect the velocity response signal under stable pressure. Because pumping vibration and construction noise exist during grouting, the signal obtained from a single vibration contains many interference components. Therefore, the stable pressure velocity response signals collected from multiple vibrations are superimposed and averaged. This ensures that the incident wave, joint reflected wave, and pile bottom reflected wave with consistent waveform characteristics reinforce each other during averaging, while random noise with unstable phases cancels each other out during superposition, thus obtaining the stable pressure average velocity response signal. In the time-domain waveform of the pressure-stabilized average velocity response signal, the incident wave initiation time is first marked. This time is the moment when the leading edge of the stress wave generated by excitation reaches the sensor, serving as the zero point of time. Simultaneously, the peak amplitude of the incident wave is marked, i.e., the maximum absolute amplitude within the first wave packet, representing the energy strength of the incident wave. Then, the initiation time of the pile bottom reflection wave is marked. This time is the moment when the stress wave propagates from the pile top to the pile bottom and reflects back to the pile top. The difference between the initiation time of the pile bottom reflection wave and the initiation time of the incident wave yields the pressure-stabilized pile bottom reflection time. Since the stress wave travels twice the pile length in one round trip within the pile, and the elastic wave velocity of the pile concrete under pressure-stabilized conditions can be considered constant within the pile length, the pressure-stabilized wave velocity can be obtained by dividing twice the design pile length by the pressure-stabilized pile bottom reflection time using the known design pile length. This pressure-stabilized wave velocity reflects the average elastic wave propagation speed of the pile concrete under compression under grouting pressure. Based on the joint design depth and the stabilizing wave velocity, the reflection time of the stabilizing joint is calculated. This time is the time required for the stress wave to propagate from the pile top to the joint design depth and back to the pile top, and its value is twice the joint design depth divided by the stabilizing wave velocity. Since the pile wave velocity changes under stabilizing conditions compared to before grouting, the position of the joint reflected wave in the time domain waveform will also shift accordingly. Therefore, recalculating the joint reflection time using the stabilizing wave velocity can accurately locate the position of the joint reflected wave under the current condition. Using the stabilizing joint reflection time as the center, the peak amplitude of the stabilizing reflected wave is extracted within a preset time window. The width of this window should completely cover the joint reflected wave packet to avoid overlap with adjacent reflected waves or noise. The peak amplitude of the stabilizing reflected wave is divided by the peak amplitude of the incident wave to obtain the stabilizing reflection coefficient. The stabilizing reflection coefficient characterizes the degree of change in wave impedance at the joint location under stabilizing conditions. Due to the elastic compression deformation of the interface between the pile concrete and the joint defect caused by the grouting pressure, this reflection coefficient includes both the elastic change component caused by pressure and the irreversible change component caused by the grout filling defect.

[0039] Understandably, low-strain testing under stable pressure conditions ensures that the pile concrete and joint defects are under stable pressure, preventing instability in wave velocity and reflection coefficient caused by pressure fluctuations. This guarantees that the stable pressure wave velocity and reflection coefficient have clear physical meaning. By recording the grouting pressure and cumulative injection volume, a correlation between pressure and grout injection volume can be established, providing raw data for subsequent calculations of grouting volume per unit pressure. Testing the stable pressure wave velocity and reflection coefficient reveals the actual response characteristics of increased pile wave velocity and decreased reflection coefficient under grouting pressure, providing data support for constructing wave velocity-pressure and reflection coefficient-pressure curves. Furthermore, the stable pressure wave velocity and reflection coefficient contain information on both stress elastic change and irreversible grout filling changes. By comparing these with the recovery value after pressure relief and the baseline value before grouting, and combining this with the stress sensitivity coefficient determined by the wave velocity-pressure relationship, the stress effect can be subtracted from the stable pressure test value, thus separating the irreversible impedance change caused by grout filling.

[0040] In another preferred embodiment of the present invention, the specific process of obtaining the restored reflection coefficient and restored wave velocity after depressurization in step S2 is as follows: First, after each stage of grouting and pressure stabilization testing is completed, the grouting valve on the grouting pipeline is closed to disconnect the grouting system from the internal defect space of the pile. Then, the pressure relief valve is opened or a pressure relief device is used to slowly release the pressure in the pile and pipeline to normal pressure. A preset time is allowed to allow the elastic deformation of the pile concrete and surrounding soil under grouting pressure to fully recover, while allowing any grout that may have entered the defect space to settle or partially solidify without pressure, thus avoiding transient disturbances at the moment of pressure relief from affecting the test results. Next, a velocity sensor is placed at the pile top, and vertical vibration is performed along the central axis of the pile top to collect the velocity response signal under the pressure relief recovery state. Since the stress state of the pile after pressure relief is different from that before grouting and under pressure stabilization, and there may be interference from grout flow or air bubble escape, the recovered velocity response signals collected from multiple vibrations are superimposed and averaged. This ensures that the incident wave, joint reflected wave, and pile bottom reflected wave with consistent waveform characteristics reinforce each other, while random noise cancels each other out, resulting in the recovered average velocity response signal. In the time-domain waveform of the recovered average velocity response signal, the incident wave initiation time and the incident wave peak amplitude are marked. The incident wave initiation time is the moment when the first excitation wave arrives at the sensor, and the incident wave peak amplitude is the maximum absolute amplitude of the first wave packet. Then, the initiation time of the pile bottom reflected wave is marked, and the difference between the pile bottom reflected wave initiation time and the incident wave initiation time is used to obtain the recovered pile bottom reflection time. Since the path of the stress wave from the pile top to the pile bottom and back to the pile top is twice the pile length, the wave velocity of the pile concrete in the recovered state is approximately constant within the pile length range. Therefore, using the known design pile length, twice the design pile length is divided by the recovered pile bottom reflection time to obtain the recovered wave velocity. Based on the joint design depth and the recovered wave velocity, the recovered joint reflection time is calculated, which is twice the joint design depth divided by the recovered wave velocity. This time corresponds to the propagation time of the stress wave from the pile top to the joint and back to the pile top in the recovered state. Centered on the recovered joint reflection time, the recovered reflected wave peak amplitude is extracted within a preset time window. This time window should completely cover the joint reflected wave packet to avoid interference from adjacent reflections. Dividing the peak amplitude of the recovered reflected wave by the peak amplitude of the incident wave yields the recovered reflection coefficient. The recovered reflection coefficient reflects the degree of change in wave impedance at the joint location after pressure relief. Since the elastic compression deformation caused by pressure relief has been largely recovered, if the grout has filled the defect space, the impedance of the defect medium will permanently change, and the recovered reflection coefficient will be lower than the initial reflection coefficient before grouting; if the grout has failed to fill the defect, the recovered reflection coefficient should be close to the initial reflection coefficient.

[0041] It is understandable that by comparing the wave velocity and initial reflection coefficient before grouting with the wave velocity and reflection coefficient under pressure stabilization, it can be determined whether the influence of grouting pressure on the wave velocity and reflection coefficient disappears after pressure relief. If the recovered wave velocity and reflection coefficient return to the level before grouting after pressure relief, it indicates that the changes observed during pressure stabilization mainly come from stress elastic compression; if the recovered wave velocity and reflection coefficient are significantly different from those before grouting, it indicates that the grout entered the defect space and produced irreversible impedance changes during pressure stabilization. Therefore, this step provides the necessary recovery state reference for subsequent stress correction using the wave velocity-pressure relationship and calculation of the irreversible filling index, enabling the final discrimination model to accurately distinguish between joint cracking damage and internal void damage of the square pile, and avoiding misjudging the stress effect as the grout filling effect.

[0042] In another preferred embodiment of the present invention, the specific calculation process for the unit pressure grouting volume in step S3 is as follows: First, the cumulative injection volume corresponding to each preset grouting pressure level is extracted from the established pressure-cumulative injection volume relationship curve. This curve is plotted during the staged grouting process, with each grouting pressure level as the abscissa and the cumulative grout volume injected at that pressure level as the ordinate. During extraction, for example, if the preset grouting pressure levels are Level 1, Level 2, and Level 3, three cumulative injection volume values ​​can be read from the corresponding curves. Subtracting the cumulative injection volumes corresponding to two adjacent preset grouting pressure levels (e.g., subtracting Level 1 from Level 2) yields the actual increase in grout injection volume within the pressure range from Level 1 to Level 2, i.e., the cumulative injection volume difference. Simultaneously, subtracting two adjacent preset grouting pressure levels (e.g., subtracting Level 1 from Level 2) yields the grouting pressure difference for that pressure range. Dividing the cumulative injection volume difference by the grouting pressure difference yields the unit pressure grouting volume for that pressure level segment, which physically represents the increase in grout injection volume corresponding to each additional unit pressure within that pressure range. The reason this result can be obtained through division is that the relationship between the cumulative injection volume and grouting pressure during the grouting process can be approximated as linear between adjacent pressure levels. The ratio of the increase in grout injection volume to the pressure increase reflects the average capacity of grout to enter the defect space within that pressure range. The above operation is repeated for each adjacent pressure level to obtain the unit pressure grouting volume for each pressure level. Then, the unit pressure grouting volume for each pressure level is arithmetically averaged to obtain the unit pressure grouting volume for the entire staged grouting process. The resulting unit pressure grouting volume represents the average increase in grout injection volume per unit pressure for the target square pile joint throughout the entire grouting pressure variation range, and is a parameter comprehensively reflecting the defect space's capacity to accommodate grout.

[0043] It is understandable that by arithmetically averaging the unit pressure grouting volume for each pressure level segment, deviations caused by local disturbances or instantaneous flow fluctuations in single-segment data can be smoothed out, resulting in a stable and reliable characteristic parameter. This unit pressure grouting volume directly participates in the construction of the subsequent defect injectability discrimination model. When the joint defect is a joint crack, the grout can continuously enter the defect space along the connecting channel under pressure, and the cumulative injection volume increases significantly with increasing pressure, resulting in a high unit pressure grouting volume for each pressure segment. When the joint defect is an internally closed type of damage, the grout cannot enter the defect space, and the cumulative injection volume mainly comes from a small amount of expansion of the grouting pipeline and sealing device, resulting in a low unit pressure grouting volume for each pressure segment. Therefore, the unit pressure grouting volume can effectively reflect the differences in injectability of the defect space, providing input features for the decision tree model, enabling the model to distinguish between joint crack damage and internally closed type damage, and thus determine reasonable grouting construction parameters. This step is a key link connecting the on-site grouting construction data with the defect type discrimination model, directly affecting the accuracy of subsequent grouting decisions.

[0044] In another preferred embodiment of the present invention, the specific calculation process of the irreversible filling index in step S3 is as follows: For any preset grouting pressure level, obtain the recovered reflection coefficient after pressure stabilization and pressure relief recovery tests at that pressure level, and simultaneously obtain the initial reflection coefficient obtained from low-strain testing before grouting. Subtract the recovered reflection coefficient from the initial reflection coefficient to obtain the change in reflection coefficient caused by irreversible filling. The physical meaning of this change is: the permanent reduction in reflection coefficient at the designed joint depth from before grouting to after pressure relief. The reason why the change in reflection coefficient caused by irreversible filling can be obtained from the difference between the initial reflection coefficient and the recovered reflection coefficient is that the reduction in reflection coefficient under pressure stabilization includes contributions from both the elastic compression of grouting pressure and grout filling. After pressure relief, the elastic compression deformation caused by grouting pressure basically disappears, and the recovered reflection coefficient only retains the permanent change in medium impedance caused by grout filling. Therefore, subtracting the initial reflection coefficient from the recovered reflection coefficient can naturally deduct the interference of stress elastic change and directly obtain the change in irreversible reflection coefficient caused by grout filling. Dividing the change in reflection coefficient caused by irreversible filling by the initial reflection coefficient yields the irreversible filling index at that pressure level. For example, if the initial reflection coefficient is 0.45 and the recovered reflection coefficient is 0.30 at a certain preset grouting pressure level, then the change in reflection coefficient caused by irreversible filling is 0.15. Dividing this by the initial reflection coefficient of 0.45 yields an irreversible filling index of 0.33 at that pressure level. This index represents the proportion of the permanent reduction in the joint reflection coefficient relative to before grouting, and its value ranges from 0 to 1. The above calculation is performed for each preset grouting pressure level to obtain the irreversible filling index at each pressure level. Then, the arithmetic mean of the irreversible filling indices at each pressure level is taken to obtain the irreversible filling index for the entire staged grouting process. The purpose of taking the arithmetic mean for each pressure level is that the degree to which grout enters the defect space varies at different pressure levels during staged grouting. The irreversible filling index of a single pressure level may be affected by local transient factors; averaging can smooth out fluctuations and obtain a stable characteristic parameter representing the overall filling effect.

[0045] It is understandable that the irreversible filling index directly reflects the degree to which the grout permanently fills the defect space. For internal void damage in square piles, the sealing layer at the lower end of the grout pipe is broken and the defect space is not connected to the outside of the pile. After the grout enters, it is constrained by the surrounding wall. After depressurization, the reflection coefficient decreases significantly and irreversibly, resulting in a high irreversible filling index. For joint crack damage, the defect space is connected to the outside of the pile. During internal grouting, the grout flows out along the channel. After depressurization, the reflection coefficient basically returns to the level before grouting, resulting in a low irreversible filling index. Therefore, this index can effectively distinguish the connectivity attributes of the two types of defects from the perspective of filling response. Using the irreversible filling index and the grouting volume per unit pressure as input features of the defect injectability discrimination model allows the model to utilize information from both injection capacity and filling response, improving the accuracy of distinguishing between internal void damage and joint crack damage in square piles. This provides a reliable basis for determining whether to continue internal grouting or switch to external grouting, avoiding repeated pressurization of defects that cannot be repaired by internal grouting, which could lead to grout loss and disturbance of the surrounding soil.

[0046] In another preferred embodiment of the present invention, the specific construction process of the square pile joint defect injectability discrimination model in step S4 is as follows: First, several square piles with known actual defect types after completing graded internal grouting were collected as historical samples. For each historical square pile, the unit pressure grouting volume and irreversible filling index obtained during the graded grouting process were extracted according to the aforementioned steps. These two values ​​serve as two features describing the grouting response characteristics of the square pile. The actual defect type was obtained by core drilling at the joint of each historical square pile after grouting was completed and the grout had solidified. The connectivity of the defect space in the core sample was directly observed. If the core sample showed that the defect was located inside the joint end plate, the void was enclosed by the end plate and weld, and there was no connecting channel for grout to pass through to the outside of the pile body. After the grout entered through the pile grout pipe, it could stay in the void and build up pressure. Then, the joint defect type of the square pile was marked as internal void damage. If the core sample showed that there was a crack extending from the defect space to the outside of the pile body at the joint, or that there was an open channel formed by incomplete filling or cracking at the weld position, and the grout would flow out of the pile body along the crack or open channel during internal grouting, then the joint defect type of the square pile was marked as joint cracking damage. For each historical pile, the obtained unit pressure grouting volume and irreversible filling index are used as the two input features of the sample, and the joint defect type determined by core drilling is used as the output label of the sample, forming a training sample. For example, a historical pile with a high unit pressure grouting volume and a high irreversible filling index corresponds to the output label of internal void damage in the pile; another historical pile with a low unit pressure grouting volume and a low irreversible filling index corresponds to the output label of joint cracking damage. The training samples of multiple historical piles are aggregated to form a training sample set. Next, a discriminative model is generated using this training sample set. The generation method is as follows: all training samples are placed in the same initial set, and an optimal partitioning method is found within this set. A feature is used as the partitioning criterion, and a value of that feature is used as the partitioning point to divide the initial set into two subsets, such that the defect types of the samples in each subset are as consistent as possible. The specific approach involves examining two features: unit pressure grouting volume and irreversible filling index. Multiple possible partitioning locations are explored for each feature. After each partition, a weighted sum of the degree of mixture of defect types in the samples of the two subsets is calculated. The feature and partitioning location that minimizes this weighted sum are selected as the first partition. For example, if partitioning at a certain location based on the irreversible filling index results in the majority of samples in the left subset having internal void damage in square piles, while the majority of samples in the right subset have joint crack damage, then this partitioning location is optimal. This process is repeated for each resulting subset to continue searching for the optimal partition, until all samples in the subset belong to the same defect type or further partitioning is not possible. This layer-by-layer partitioning process divides the feature space into multiple non-overlapping regions, each corresponding to a specific defect type.For a new square pile, its unit pressure grouting volume and irreversible filling index are simply input into the partitioning structure. The two feature values ​​are compared step by step along the path from the initial set to the subsets, and finally it falls into a certain region. The defect type corresponding to that region is the discrimination result. This process automatically learns the correspondence between unit pressure grouting volume and irreversible filling index and joint defect type through training samples, generating a square pile joint defect injectability discrimination model.

[0047] It is understandable that the unit pressure grouting volume and irreversible filling index describe the characteristics of joint defects from two dimensions: injection capacity and filling response, respectively. For square piles with internal void damage, the defect space is closed, allowing grout to continuously enter and establish effective filling under pressure; therefore, both the unit pressure grouting volume and irreversible filling index are relatively high. Conversely, for joint crack damage, the defect space is connected to the outside of the pile, causing grout to leak along the channel during internal grouting, making effective filling difficult; thus, both the unit pressure grouting volume and irreversible filling index are relatively low. The training sample set gathers two real-world defect types and their corresponding characteristics from multiple projects. Through layer-by-layer segmentation, the optimal feature combination and segmentation position for distinguishing the two types of defects are automatically found, without the need for manually pre-setting thresholds. After this discrimination model is generated, in new square pile grouting construction, the unit pressure grouting volume and irreversible filling index calculated in real-time during construction can be quickly used to output the defect type discrimination result, providing a direct basis for determining whether to continue internal grouting or switch to external grouting. If the model identifies the damage as internal voids in the square pile, grouting continues through the pile grouting pipe until the design final pressure is reached and stabilized, ensuring that the grout fully fills and seals the defect space. If the model identifies the damage as joint cracking, internal grouting is terminated, and an external grouting pipe is installed in a hole drilled on the pile side for external grouting, utilizing the connecting channel as the grout entry path for effective repair. Thus, this step transforms historical construction experience into a reusable discrimination model, allowing the determination of defect injectability to be completed during construction. This avoids repeated pressurization of defects that cannot be repaired by internal grouting, which could lead to grout loss and disturbance of the surrounding soil. It also improves the coordination and accuracy between the low-strain integrity detection results of the square pile joint and the grouting repair decision.

[0048] In another preferred embodiment of the present invention, the specific process for determining the grouting construction parameters of the target square pile in step S5 is as follows: First, the two feature values ​​of unit pressure grouting volume and irreversible filling index, calculated during the graded internal grouting process of the target square pile, are obtained. These two values ​​are then input into the pre-constructed square pile joint defect injectability discrimination model. This discrimination model internally stores the splitting rules formed during the training phase by progressively dividing the feature space. After inputting the two feature values, starting from the first layer of partitioning nodes, the feature value used by the current node is compared with the partitioning position stored in the node. If the current node is partitioned based on unit pressure grouting volume, the unit pressure grouting volume of the target square pile is compared with this partitioning position. If it is less than the partitioning position, it moves to one side of the sub-node; if it is greater than or equal to the partitioning position, it moves to the other side of the sub-node. If the current node is partitioned based on the irreversible filling index, the comparison is performed in the same way, and the corresponding sub-node is entered. This comparison is continued layer by layer until the final node, where further partitioning is no longer possible. The defect type marked on the final node is the defect type output by the model. The reason this output process can determine the defect type based on two features is that the discrimination model has already found the optimal division position to distinguish between internal void damage and joint cracking damage of the square pile using multiple known samples during the training phase. Therefore, when the two feature values ​​of the target square pile fall into a certain feature region, the defect type corresponding to that region is the discrimination result of the target square pile. When the defect type output by the model is internal void damage of the square pile, it indicates that there is a void inside the joint end plate that is wrapped by the end plate and the weld, the lower end outlet of the pile grout pipe is connected to the void, and there is no connecting channel between the void and the outside of the pile for grout to flow out. At this point, the grouting construction parameters are determined to continue internal grouting through the pile grouting pipe until the designed final pressure. Specifically, the grouting pump is started, and grout is continuously injected into the joint voids through the pile grouting pipe. Simultaneously, the pressure gauge reading on the grouting system is observed. When the grouting pressure gradually rises to the designed final pressure, the pressure increase is stopped, and the grouting pressure is maintained at this designed final pressure level for a preset stabilization time. This allows the grout to fully penetrate into the various micro-areas within the voids under pressure. After the stabilization time is completed, the grouting valve is closed to stop grouting. The principle behind this is that the internal voids of the joint are enclosed by the end plate and weld, and the grout, once inside, is constrained by the surrounding walls and cannot escape. As the injection volume increases, the hydraulic pressure within the voids gradually rises, allowing the grout to enter the tiny gaps within the voids under pressure, achieving effective filling. The designed final pressure ensures that the grout has sufficient driving force to enter the micro-areas of the voids, and the stabilization time ensures that the grout has sufficient time to flow, penetrate, and expel moisture within the voids. When the model outputs a defect type of joint cracking damage, it indicates that there is an open channel formed by incomplete grouting or cracking at the weld position of the joint. During internal grouting, the grout will flow directly to the outside of the pile along the open channel and cannot stay and accumulate in the defect space. The grouting pressure is mainly consumed in pushing the grout to flow along the external channel and is difficult to be converted into hydrostatic pressure in the defect space.At this point, the grouting construction parameters are set to terminate internal grouting. Specifically, the grouting pump is immediately stopped, the grouting valve on the pile grouting pipe is closed, and grout injection through the pile grouting pipe is ceased. The completed grouting pressure and cumulative injection volume for the target square pile are recorded. Subsequently, a hole is drilled at the pile side location corresponding to the joint, reaching the depth of the joint defect space. An external grouting pipe is installed within the hole, with one end connected to the joint defect space and the other end connected to the grouting equipment. External grouting is then performed through the external grouting pipe to the design final pressure and stabilized. The principle behind this is that although joint cracking damage limits the effectiveness of internal grouting, the connecting cracks or open channels themselves constitute a natural channel for grout to enter the defect space from the pile side. By drilling a hole at the corresponding location on the pile side and installing an external grouting pipe, grout can directly enter the connecting channel from the pile side and fill the defect space, using the channel as a grout entry path for effective repair.

[0049] It is understandable that by using models to determine the unit pressure grouting volume and irreversible filling index, it is possible to quickly identify whether the defect space is closed or connected to the outside of the pile during construction. This avoids the loss of grout caused by uniformly using internal grouting for all defects, or the increased construction complexity caused by using external grouting for closed defects. For internal void damage in square piles, continuing to grout through the pile grout pipe to the design final pressure and stabilizing the pressure can take advantage of the closed defect space and the pressure-building properties of the grout, ensuring that the grout fully fills the defect, making the joint impedance tend to be continuous, and providing a physical basis for the reduction of the reflection coefficient in low-strain testing after grouting. For joint cracking damage, timely termination of internal grouting and switching to external grouting can prevent the grout from continuously flowing away along the connecting channel under the internal grouting path, reducing grout waste and disturbance to the surrounding soil. At the same time, using the connecting channel as the grout entry path for external grouting allows the grout to directly enter the defect space from the pile side to achieve filling. Therefore, this step directly transforms the defect injectability judgment results into specific grouting construction operations, realizing a closed-loop connection from detection and identification to construction decision-making, improving the pertinence and effectiveness of square pile joint repair construction, and ultimately enhancing the consistency between low-strain integrity detection results and grouting repair decisions.

[0050] In another preferred embodiment of the invention, the process further includes first waiting for a preset time for the grout to solidify. This preset time is determined based on the grout mix ratio and ambient temperature. The purpose is to allow the injected grout to transform from a liquid state into a solid stone body, forming a stable filling structure with the defect surface. After the solidification time is reached, low-strain testing is performed again on the joint of the target square pile. The specific operation is basically the same as the pre-grouting test: a velocity sensor is arranged at the top of the pile, and vertical vibration is applied along the central axis of the pile top. The velocity response signals are collected multiple times and averaged to obtain the average velocity response signal after solidification. The incident wave initiation time and the peak amplitude of the incident wave are marked in the time-domain waveform. The initiation time of the reflected wave at the pile bottom is also marked. The wave velocity after solidification is calculated from the pile bottom reflection time. The joint reflection time is then calculated based on the joint design depth. The peak amplitude of the reflected wave at the joint after solidification is extracted within a preset window centered on this time and divided by the peak amplitude of the incident wave to obtain the reflection coefficient after solidification. Since the wave impedance of the grout after solidification is close to that of concrete, if the grout effectively fills the joint defect space, the wave impedance at the joint changes from a low-impedance defect medium to a high-impedance continuous medium, and the reflection coefficient will decrease accordingly. If the grout fails to fill effectively, the change in the joint reflection coefficient is minimal. Subtracting the solidified reflection coefficient from the initial reflection coefficient obtained before grouting yields the reduction in reflection coefficient, which reflects the degree of reduction in the difference in wave impedance before and after grouting. If the reduction in reflection coefficient is greater than or equal to the preset reduction threshold, it indicates that the reflection coefficient after solidification has significantly decreased compared to before grouting, the grout has effectively filled the defect space, the joint integrity has improved, and the grouting is deemed qualified. If the reduction in reflection coefficient is less than the preset reduction threshold, it indicates that the reflection coefficient after solidification has not changed much compared to before grouting. The grout may not have entered the defect space, or the characteristics of the defect itself may not match the model's judgment results, or the grouting process may not have been executed properly. In this case, it cannot be directly determined that the defect is unsuitable for grouting; core drilling verification of the target square pile is required. During core drilling verification, a core sample is taken from the joint area to directly observe the connectivity of the defects in the concrete core sample. If the core sample shows that the defect is located inside the joint end plate, and the void is enclosed by the end plate and weld, with no connecting channel for grout to pass through to the outside of the pile, it is determined to be internal void damage of the square pile. If the core sample shows that there is a crack extending from the defect space to the outside of the pile at the joint, or that there is an open channel formed by incomplete grouting or cracking at the weld position, it is determined to be joint crack damage. When the defect type previously output by the defect injectability discrimination model is joint crack damage, but the joint defect type determined by the core drilling is internal void damage of the square pile, it indicates that the model has made a directional deviation in the discrimination of the target square pile, misjudging a closed void that should be repaired by internal grouting as a connected defect that requires external grouting.At this point, the unit pressure grouting volume and irreversible filling index of the target square pile are used as new features, and the internal void damage of the square pile is used as the output label to form a new training sample. This sample is added to the original training sample set to retrain the square pile joint defect injection capability discrimination model. This corrects the split boundary of the model in the corresponding feature area, preventing subsequent similar samples from being misjudged in the same direction. When the joint defect type determined by core drilling is joint cracking damage, it indicates that the defect itself has the conditions for grout entry, but the previous grouting construction failed to achieve effective filling. This may be due to process reasons such as the misalignment of the external grouting pipe, improper grout ratio, or insufficient pressure. The grouting is judged as unqualified, and external grouting is re-performed for the target square pile. External grouting and testing are carried out again according to the preset pressure level until the grouting qualification standard is met.

[0051] It is understandable that low-strain testing should be performed after grout solidification to reflect the final state of the bond between the grout aggregate and the defect surface. This avoids unstable reflection coefficient changes caused by testing before the grout solidifies, thus providing a more accurate evaluation of the actual grouting effect. The reduction in reflection coefficient, as a quantitative indicator of the improvement in joint impedance before and after grouting, can objectively determine whether the grouting has achieved its intended purpose, providing a clear basis for whether further verification is needed. When the reduction in reflection coefficient is insufficient, core drilling can directly obtain the actual defect type, distinguishing whether the lack of improvement in the reflection signal is due to a mismatch between the defect's inherent characteristics and the model's judgment, or inadequate execution of the grouting process. This avoids misclassifying repairable defects as unrepairable and abandoning repair, and also avoids the waste caused by repeatedly grouting defects that cannot be repaired through the current grouting path. When the core drilling results are inconsistent with the model's judgment results, the sample is fed back into the training set to retrain the model, allowing the model to gradually correct itself in the judgment of subsequent similar samples, improving the accuracy of defect injectability judgment. For cases where core drilling confirms joint cracking damage but grouting is substandard, repeating external grouting can remedy the process defects and ensure that groutable defects are effectively filled. In summary, this step combines grouting effect evaluation, defect authenticity verification, and dynamic model updating, transforming post-grouting construction of square pile joints from a one-time decision into a feedback-enabled, correctable closed-loop process. Ultimately, this improves the reliability of low-strain integrity testing results and the rationality of grouting repair construction.

[0052] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A post-grouting construction method for improving the low-strain integrity detection of square pile joints, characterized in that, A grouting pipe is pre-embedded inside the square pile, and the lower outlet of the grouting pipe is located at the designed depth of the joint and communicates with the joint. The joint is connected by welding. The method includes the following steps: S1, perform low-strain testing on the joint of the target square pile, extract the initial reflection coefficient and initial pile bottom reflection time at the design depth of the joint, and calculate the reference wave velocity; S2, when the initial reflection coefficient is greater than the preset reflection coefficient threshold, the joint is internally grouted in stages through the pile grouting pipe, and low strain detection is performed under the pressure stabilization state of each preset grouting pressure level to obtain the grouting pressure, cumulative injection volume, pressure stabilization reflection coefficient and pressure stabilization wave velocity. After depressurization, the recovered reflection coefficient and recovered wave velocity are obtained. S3, construct a pressure-cumulative injection volume relationship curve based on the grouting pressure and cumulative injection volume corresponding to each pressure level, calculate the unit pressure grouting volume based on the pressure-cumulative injection volume relationship curve, and calculate the irreversible filling index based on the initial reflection coefficient, the stable pressure reflection coefficient, and the recovery reflection coefficient; S4. Construct the attribution mapping relationship between unit pressure grouting volume and irreversible filling index and square pile joint defect type, and establish a square pile joint defect injectionability discrimination model. S5. Obtain the unit pressure grouting volume and irreversible filling index of the target square pile and input them into the defect injectability discrimination model. Determine the grouting construction parameters of the target square pile based on the output of the defect injectability discrimination model.

2. The post-grouting construction method for improving the low-strain integrity detection of square pile joints according to claim 1, characterized in that, In step S1, the specific process of performing low-strain testing on the joint is as follows: Velocity response signals under vertical excitation are collected by a velocity sensor placed on the pile top and averaged multiple times, wherein the vertical excitation is applied along the central axis of the pile top; the start time of the incident wave, the peak amplitude of the incident wave, and the start time of the reflected wave at the pile bottom are marked, and the difference between the start time of the reflected wave at the pile bottom and the start time of the incident wave is used as the initial pile bottom reflection time; the reference wave velocity is obtained by dividing twice the designed pile length by the initial pile bottom reflection time; the joint reflection time is calculated according to the joint design depth and the reference wave velocity, and the peak amplitude of the reflected wave is extracted within a preset time window with the joint reflection time as the center, and the peak amplitude of the reflected wave is divided by the peak amplitude of the incident wave to obtain the initial reflection coefficient.

3. The post-grouting construction method for improving the low-strain integrity detection of square pile joints according to claim 2, characterized in that, In step S2, the specific process for obtaining the grouting pressure, cumulative injection volume, pressure-stabilizing reflection coefficient, and pressure-stabilizing wave velocity is as follows: During the graded grouting process, when the grouting pressure reaches any preset grouting pressure level and is maintained at a stable pressure, the current grouting pressure and cumulative injection volume are recorded, and the joint of the target square pile is subjected to low strain testing again. The obtained reflection coefficient and wave velocity are marked as the stable pressure reflection coefficient and stable pressure wave velocity, respectively.

4. The post-grouting construction method for improving the low-strain integrity detection of square pile joints according to claim 2, characterized in that, In step S2, the specific process of obtaining the restored reflection coefficient and restored wave velocity after depressurization is as follows: After closing the grouting valve and depressurizing to normal pressure and waiting for a preset time, low-strain testing is performed on the joint of the target square pile, and the obtained reflection coefficient and wave velocity are marked as the restored reflection coefficient and restored wave velocity, respectively.

5. The post-grouting construction method for improving low-strain integrity detection of square pile joints according to claim 1, characterized in that, In S3, the specific calculation process for the unit pressure grouting volume is as follows: From the pressure-cumulative injection volume relationship curve, extract the cumulative injection volume corresponding to each preset grouting pressure level; subtract the cumulative injection volume corresponding to two adjacent preset grouting pressure levels to obtain the cumulative injection volume difference; subtract the grouting pressure difference from the two adjacent preset grouting pressure levels to obtain the grouting pressure difference. Divide the cumulative injection volume difference by the grouting pressure difference to obtain the unit pressure grouting volume for each pressure level range; The grouting volume per unit pressure is obtained by arithmetically averaging the grouting volume per unit pressure for each pressure level range.

6. The post-grouting construction method for improving the low-strain integrity detection of square pile joints according to claim 1, characterized in that, In S3, the specific calculation process of the irreversible filling index is as follows: For any preset grouting pressure level, the difference between the initial reflection coefficient and the recovered reflection coefficient is taken as the change in reflection coefficient caused by irreversible filling; the change in reflection coefficient caused by irreversible filling is divided by the initial reflection coefficient to obtain the irreversible filling index at that pressure level. The irreversible filling index is obtained by arithmetically averaging the irreversible filling indices at each pressure level.

7. The post-grouting construction method for improving the low-strain integrity detection of square pile joints according to claim 1, characterized in that, In S4, the specific construction process of the square pile joint defect injectability discrimination model is as follows: The unit pressure grouting volume and irreversible filling index of multiple square piles during the graded grouting process are obtained; the joint defect type of each square pile is determined by core drilling verification, wherein the joint defect type is internal void damage or joint cracking damage; a training sample set is constructed using the unit pressure grouting volume and the irreversible filling index as input features, and the internal void damage and joint cracking damage of the square pile as output labels, respectively; the training sample set is trained using a decision tree algorithm, and the feature space of the input features is recursively divided using the minimization of the Gini index as the splitting criterion, to generate a square pile joint defect injectability discrimination model with the unit pressure grouting volume and the irreversible filling index as input and the joint defect type as output.

8. A post-grouting construction method for improving low-strain integrity detection of square pile joints according to claim 1, characterized in that, In step S5, the specific process for determining the grouting construction parameters of the target square pile is as follows: The unit pressure grouting volume and irreversible filling index of the target square pile are input into the square pile joint defect injectability discrimination model to obtain the defect type output by the model. When the defect type is internal void damage of the square pile, the grouting construction parameters are determined to continue internal grouting through the pile grouting pipe to the design final pressure, and grouting is stopped after the final pressure is stabilized for a preset time. When the defect type is joint cracking damage, the grouting construction parameters are determined to terminate internal grouting, drill a hole at the pile side position corresponding to the joint to install an external grouting pipe, and externally grout the defect space through the external grouting pipe to the design final pressure and stabilize the pressure.

9. A post-grouting construction method for improving low-strain integrity detection of square pile joints according to claim 8, characterized in that, This also includes a low-strain test to measure the reflection coefficient after solidification after the grout has solidified for a preset time after grouting is completed; the difference between the initial reflection coefficient and the solidified reflection coefficient is used as the reduction in reflection coefficient; if the reduction in reflection coefficient is greater than or equal to a preset reduction threshold, the grouting is deemed qualified; otherwise, core drilling verification is performed: if the defect type output by the defect injectability discrimination model is joint cracking damage, and the joint defect type determined by the core drilling is internal void damage of the square pile, then new training samples are formed using the unit pressure grouting volume and irreversible filling index of the target square pile as new features, and the internal void damage of the square pile as a label, and the square pile joint defect injectability discrimination model is updated; if the joint defect type determined by the core drilling is joint cracking damage, the grouting is deemed unqualified, and external grouting is re-performed for the target square pile.

Images