SMW method for phased and step-by-step construction of piles
By employing the SMW method for phased and segmented pile construction, combined with real-time monitoring of the ratio of steel side resistance to sinking velocity and multi-level auxiliary measures, the problems of soil disturbance and steel control in deep and shallow foundation pit construction were solved, thereby achieving improved controllability and efficiency of construction.
Patent Information
- Application Number
- CN202611612775.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-19
- Publication Date
- 2026-11-17
AI Technical Summary
In the construction of deep and shallow foundation pits, existing technologies are unable to effectively avoid soil disturbance and deformation of the retaining structure during construction. Furthermore, the lack of proactive prediction and adaptive control methods during the insertion and removal of steel sections leads to increased construction complexity and uncertainty.
The SMW method of pile construction is adopted, with the construction time staggered from deep to shallow. By real-time monitoring of the ratio of the side resistance of the steel section to the sinking speed and its rate of change, the grouting parameters can be adjusted in a predictive manner. In addition, multi-level auxiliary measures are adopted during the steel section removal process to improve controllability.
It effectively avoids soil disturbance during deep and shallow foundation pit construction, improves the controllability of steel insertion and extraction, reduces construction complexity and cost, and ensures construction reliability and efficiency.
Smart Images

Figure CN123109152A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of building construction technology, specifically to a method for phased and segmented construction of piles using the SMW method. Background Technology
[0002] SMW (Soil Mixing Method) piles are a commonly used support method in deep foundation pit engineering. They form a composite wall that combines soil retention and water stopping functions by inserting H-beams into cement-soil mixing piles. When there are both deep and shallow foundation pits with varying elevations on the construction site, how to rationally arrange the construction sequence and support scheme for these pits is a technical problem that needs to be solved in foundation pit engineering.
[0003] For construction cases where deep and shallow foundation pits are arranged adjacent to each other, several existing construction methods are available for reference. For example, Chinese Patent No. CN107313431B discloses a method for simultaneous construction of adjacent foundation pits of unequal height, employing a coordinated construction approach that involves simultaneous excavation, support erection, and structural slab construction of both deep and shallow foundation pits. Chinese Patent No. CN110055973B discloses a retaining structure and water-stopping method for foundation pits under high-speed railway bridges when construction space is limited. This method involves segmented construction based on the distance between the foundation pit and the high-speed railway bridge, using sheet piles for water-stopping within the railway line safety protection zone, and triaxial cement-soil mixing piles and SMW method piles for water-stopping outside the railway line safety protection zone. Furthermore, Chinese Patent Application No. CN118600997A discloses a method for simultaneous construction of adjacent deep and shallow foundation pits under silty soil geological conditions. This method divides the shallow foundation pit into a first construction area and a second construction area, using the second construction area as a soil retention zone and forming a construction road, enabling simultaneous construction of the deep foundation pit and the shallow foundation pit in the first construction area.
[0004] While the aforementioned simultaneous construction method can shorten the nominal construction period, when deep and shallow foundation pits are excavated simultaneously, the lateral earth pressure of the deep foundation pit is transmitted to the shallow foundation pit, causing lateral compression on the retaining structure of the shallow foundation pit. When the stratum is multi-layered silt and the groundwater level is shallow, it can easily lead to increased horizontal displacement of the shallow foundation pit retaining structure, local leakage, and even slope slippage. Furthermore, with both deep and shallow foundation pits exposed to the atmosphere, the combined effects of rainfall and groundwater increase the risk of instability due to prolonged exposure over a large area, and monitored displacement may exceed warning values. Actual construction also requires frequent adjustments to the support axial force and reinforcement measures, further increasing organizational complexity and uncertainty.
[0005] In addition, some projects follow a deep-to-shallow construction sequence to avoid mutual soil disturbance, meaning that the shallow foundation pit is excavated only after the deep foundation pit is completed. In SMW (Sequencing Welded Mold) pile construction, the steel section must be inserted into the mixing pile before the cement-soil initially sets; inserting it too late will increase the insertion resistance due to cement-soil solidification. Conventional grouting parameters such as the water-cement ratio and grouting pressure are determined once before construction, making it difficult to quickly adjust when encountering sudden changes in resistance caused by local hard or weak layers. On-site judgment often relies on experience, lacking quantitative monitoring and control methods, leading to frequent occurrences of obstructed or incomplete steel insertion. For example, Chinese patent application CN122812241A discloses a closed-loop control wall-forming method and system for water-rich sandy soft soil strata using SMW, which dynamically adjusts grouting parameters based on geological zoning results and feedback from the H-beam insertion status. Chinese patent application CN102539024A discloses an automatic stress monitoring system for H-beams inserted using the SMW method, which collects strain data through strain gauges installed on the H-beams and uploads it for monitoring. While the aforementioned schemes incorporate insertion status feedback or stress monitoring, the control remains reactive: the machine only stops after the lateral resistance reaches the alarm threshold or the steel insertion is obstructed. Furthermore, they primarily monitor a single parameter, failing to normalize the lateral resistance and sinking velocity into a ratio reflecting the true resistance. They also cannot identify precursors based on the ratio change rate before the threshold is breached, quantitatively pre-adjust according to the deviation rate, and adaptively verify the results. Simultaneously, during the steel extraction stage, the extraction force fluctuates significantly along the depth due to the increase in cement-soil strength and changes in adhesion resistance. Existing projects often rely on experience for direct extraction, lacking graded auxiliary methods based on the deviation between the measured extraction force curve and the theoretical curve. Sudden increases in extraction force can easily cause steel deformation or damage to the mixing pile wall. Therefore, there is an urgent need for a steel insertion control method that can proactively predict, quantitatively pre-adjust, and adaptively verify before the threshold is breached, as well as a steel extraction auxiliary method that implements graded measures based on the deviation of the measured curve. Summary of the Invention
[0006] This invention provides a phased construction method for SMW (Sequential Slab Welding) piles. By constructing the deepest section first and then the shallowest, the construction of the deep and shallow foundation pits is staggered in time, avoiding mutual soil disturbance between the pits and eliminating the need for additional support axial force adjustment and reinforcement measures due to mutual disturbance. During the steel section insertion process, the ratio of lateral resistance to sinking speed and its first and second order rates of change are used to reflect the real-time resistance changes of the cement-soil to the steel section. Based on this, grouting parameters are preemptively adjusted before the ratio R(t) exceeds a preset threshold, improving the success rate of steel section insertion. During steel section removal, auxiliary measures are applied in stages based on the deviation between the measured extraction force and the theoretical curve, improving the controllability of steel section removal.
[0007] To achieve these objectives and other advantages according to the present invention, a method for phased construction of SMW piles is provided, comprising the following steps:
[0008] Step 1: Divide the construction site into a deep foundation pit area and a shallow foundation pit area, and install dewatering wells for pre-dewatering; Step 2: Construct SMW triaxial cement-soil mixing piles in the deep foundation pit area and insert H-beams; During the insertion process, collect the side resistance F(t) and sinking velocity v(t) in real time, calculate the ratio R(t) = F(t) / v(t), and calibrate the baseline ratio R0 = F0 / v0 of each soil layer by using the measured average side resistance F0 and the measured average sinking velocity v0 of the initial segment of each soil layer. When the second-order rate of change of R(t) with depth crosses zero and the data segments before and after it are compared to confirm the soil layer interface, recalibrate R0 according to the measured value of the initial segment of the new soil layer; Step 3: Calculate R(t) in real time. First-order and second-order rates of change are identified as precursors when multiple consecutive data points of R(t) move in the same direction and the change increases successively, or when the absolute value of the second-order rate of change continuously exceeds the threshold. Before R(t) exceeds the threshold, the grouting parameters are adjusted to make R(t) converge towards the baseline ratio of the soil layer: when R(t) is greater than R0, the water-cement ratio is increased or the grouting pressure is increased; when R(t) is less than R0, the water-cement ratio is decreased or the grouting pressure is decreased. The pre-adjustment amount is determined according to the interval to which the instantaneous deviation rate of R(t) relative to R0 belongs. After pre-adjustment, the first-order rate of change is monitored. If it approaches zero, it is locked; if it reverses, it is canceled and the original parameters are restored. Step 4: Construct SMW triaxial cement-soil mixing piles in the shallow foundation pit area and insert H-beams.
[0009] Preferably, in the SMW method for phased construction of piles, in step three, the change sequence of the first-order rate of change is monitored after pre-adjustment; when multiple consecutive data points in the sequence have the same sign and their absolute values decrease successively, it is determined that the first-order rate of change is in a state of monotonically decaying and approaching zero; the number of data points required for the first-order rate of change to reach zero is predicted based on the current decay rate; when the number of data points required to reach zero is less than a preset threshold, a locking command is issued in advance; if the predicted first-order rate of change does not decay to zero as expected, the locking is canceled and the original parameters are restored.
[0010] Preferably, in the SMW method for phased construction of piles, in step three, if the number of data points required for the predicted first-order rate of change to reach zero is less than a preset threshold, it is further determined whether the absolute value of the current first-order rate of change has dropped below a preset steady-state threshold; a locking command is issued only when the number of data points required to reach zero is less than the preset threshold and the absolute value of the current first-order rate of change has dropped below the preset steady-state threshold; when the absolute value of the first-order rate of change drops below the steady-state threshold, a locking countdown is started; if the first-order rate of change does not increase in the opposite direction during the countdown, the lock is activated when the countdown ends; if the first-order rate of change increases in the opposite direction during the countdown, the countdown is terminated and the original parameters are restored.
[0011] Preferably, the SMW method for phased construction of piles further includes: removing the H-beams in the shallow foundation pit area and filling the pile holes with grout before sealing the well; monitoring the pile extraction force in real time and plotting the pile extraction force-displacement curve during H-beam removal; calculating the deviation value and deviation change rate between the measured curve and the theoretical curve at each depth; when the deviation value exceeds the first preset value and the deviation change rate is positive, initiating the first-level auxiliary measure by injecting high-pressure water or bentonite slurry into the contact surface between the steel and the cement-soil; when the deviation change rate continues to increase beyond the preset rate, skipping the first-level measure and directly initiating the second-level auxiliary measure, i.e., applying high-frequency micro-vibration to assist in extraction; when applying high-frequency micro-vibration, retrieving the optimal frequency of the historical steel most similar to the current steel working condition from the extraction resistance database, and performing local frequency sweeping to both sides with the optimal frequency as the starting frequency; If no working condition meets the similarity requirement in the database, a hill climbing search is initiated simultaneously from multiple different starting frequencies. Each search thread independently records the minimum value of the pull-out force and its corresponding frequency. The pull-out forces of all threads are compared, and the frequency corresponding to the minimum value is selected as the initial search starting point. During the frequency sweep, the pull-out force signal and vibration acceleration signal are collected in real time. The cross power spectrum of the two signals is calculated, and the frequency corresponding to the peak value of the cross power spectrum is extracted as the effective dominant frequency. Frequency sweep optimization is performed with the effective dominant frequency as the center and a preset narrow band range as the sweep interval. The frequency with the minimum pull-out force is updated as the optimal frequency of the current steel section and stored in the pull-out resistance database. When the deviation value exceeds the second preset value and the deviation change rate is still positive, a level 3 auxiliary measure is initiated, namely, the pull-out is suspended and local segmented pull-out is performed.
[0012] Preferably, in the SMW method for phased construction of piles, when removing H-beams, the confidence level of the search results is calculated by retrieving historical steel sections most similar to the current steel section's working condition from the removal resistance database. When the confidence level is higher than a preset high threshold, the optimal frequency found is used as the starting frequency for local narrowband frequency sweeping. When the confidence level is lower than a preset low threshold, multiple starting points are simultaneously initiated for parallel hill climbing search, and the cross power spectrum of the pile extraction force signal and vibration acceleration signal is calculated to extract the effective dominant frequency. When the confidence level is between the preset high threshold and the preset low threshold, the optimal frequency found is used as one of the starting frequencies for parallel hill climbing and is searched in parallel with other preset starting frequencies.
[0013] Preferably, in the SMW method for segmented and phased pile construction, in step two or four, the second-order rate of change of the real-time ratio R(t) with depth is calculated; when the second-order rate of change changes from positive to negative or from negative to positive, the location is identified as a candidate zero-crossing point; after detecting a candidate zero-crossing point, the R(t) data in the first depth window before the zero-crossing point and the second depth window after the zero-crossing point are extracted, and the energy value of the data in the two windows is calculated; when the ratio of the energy value in the window after the zero-crossing point to the energy value in the window before the zero-crossing point exceeds a preset threshold, the candidate zero-crossing point location is confirmed as a soil layer interface, triggering the recalibration of the baseline ratio R0.
[0014] Preferably, in the SMW method for segmented and phased construction of piles, in step two or four, after detecting a candidate zero-crossing point, R(t) data before and after the zero-crossing point are extracted with multiple different window lengths centered on the candidate zero-crossing point. The energy value ratio of the window after the zero-crossing point to the window before the zero-crossing point is calculated for each window length. The maximum value is selected from the energy ratios corresponding to all window lengths. When the maximum value exceeds a preset threshold, the candidate zero-crossing point is confirmed as the soil layer interface, and the baseline ratio R0 is recalibrated.
[0015] Preferably, in the SMW method for segmented and phased construction of piles, in step two or four, the position of the zero-crossing point of the second-order rate of change is a candidate zero-crossing point; after detecting the candidate zero-crossing point, the R(t) waveform within the first depth window before the zero-crossing point is taken as the reference waveform, and the R(t) waveform within the second depth window after the zero-crossing point is taken as the comparison waveform; the shape similarity between the reference waveform and the comparison waveform is calculated; when the shape similarity is lower than a preset threshold, the candidate zero-crossing point position is confirmed as the soil layer interface and the baseline ratio R0 is recalibrated.
[0016] Preferably, in the SMW method for phased construction of piles, in step three, when determining the pre-adjustment amount of grouting parameters based on the instantaneous deviation rate, a deviation rate benchmark value k is set. ref The pre-adjustment amount is determined in stages according to the absolute value of the instantaneous deviation rate |k|: when |k| is less than 0.5k ref When |k| is at 0.5k, the pre-adjustment amount is 30% to 50% of the normal adjustment amount; ref Up to 1.5k ref When |k| is between 1.5k and 60% of the normal adjustment amount, the pre-adjustment amount should be between 80% and 1.5k. ref When adjusting, the pre-adjustment amount should be 100% to 120% of the normal adjustment amount.
[0017] Preferably, in the SMW method for phased construction of piles, in step three, before determining the pre-adjustment amount of grouting parameters, the response sensitivity of the current soil layer is calibrated: when the pre-adjustment conditions are met for the first time in the current soil layer, only the benchmark detection adjustment amount Δ0 is applied, where Δ0 is one-third of the lower limit of the preset adjustment amount; the measured ratio sequence of m consecutive depth points after application is collected, and the predicted ratio sequence is obtained by extrapolating the ratio sequence of the n depth points before application using a quadratic polynomial; the average value ΔR of the difference between the measured and predicted sequences at m depth points is used. p As the detection response quantity, ΔR p The normalized response sensitivity S of the soil layer is obtained by dividing the ratio of Δ0 to the baseline ratio R0; in subsequent pre-adjustment, the deviation ΔR of the current measured ratio from the baseline ratio R0 is used. cDivide the target adjustment amount by the sensitivity S, and take the smaller of the target adjustment amount and the preset upper limit of the adjustment amount as the actual pre-adjustment amount; when S is lower than the preset lower limit of sensitivity, the current soil layer is determined to be the grouting parameter adjustment failure layer, the current grouting parameters are kept unchanged, the single insertion stroke is shortened to the preset segment length, and the insertion is paused for a preset time after each segment is completed before continuing the insertion; the ratio of sensitivity S to the baseline of the soil layer R0 is stored in the insertion control database, and the subsequent pile construction of the same soil layer directly calls S, and no more benchmark detection adjustment amount Δ0 is applied.
[0018] The present invention has at least the following beneficial effects:
[0019] First, a phased construction method of deep excavation followed by shallow excavation is adopted, staggering the construction of deep and shallow foundation pits in time. The shallow foundation pit is excavated only after the deep foundation pit is backfilled. This avoids problems such as deformation of the retaining structure, heave of the base, and instability of the slope caused by the mutual disturbance of soil pressure due to the elevation difference when deep and shallow foundation pits are excavated at the same time. At the same time, it eliminates the need for additional support axial force adjustment and reinforcement measures due to mutual disturbance, and avoids work stoppage caused by excessive displacement. The overall construction period and cost are more controllable.
[0020] Second, during the steel section insertion process, by collecting the side resistance and sinking speed of the steel section in real time and calculating the ratio of the two parameters, the change of the resistance of cement and soil to the steel section can be reflected. Compared with single parameter monitoring, the insertion state of the steel section can be judged more accurately.
[0021] Third, by calculating the first and second rates of change of the ratio curve in real time, abnormal trends can be identified and grouting parameters can be adjusted in advance before the ratio exceeds the preset threshold. This transforms the traditional reactive adjustment into predictive control, which is beneficial to improving the passability of steel sections in complex strata.
[0022] Fourth, by setting a locking and undoing mechanism after pre-adjustment, the effectiveness of the pre-adjustment measures can be automatically determined based on the subsequent trend of the first-order rate of change, and the original parameters can be automatically restored when the pre-adjustment is excessive or the direction is incorrect, so that the adjustment of grouting parameters has adaptive capability.
[0023] Fifth, during the steel section removal process, different levels of auxiliary measures are automatically triggered based on the deviation rate between the measured pile removal force and the theoretical curve. When high-frequency micro-vibration is applied, the appropriate working frequency is determined by database retrieval or frequency sweep optimization, which helps to improve the steel section removal efficiency.
[0024] VI. During the steel section insertion process, by calculating the energy ratio or shape similarity of the ratio R(t) with depth in real time, it is possible to identify the soil layer interface and recalibrate the baseline ratio R0 of the soil layer before the threshold is broken, so that the insertion resistance of different soil layers is comparable, thereby improving the accuracy of steel section insertion process control in complex interactive strata.
[0025] 7. Before determining the pre-adjustment amount according to the deviation rate, the normalized response sensitivity of the current soil layer is first calibrated online using a single benchmark detection adjustment amount. Then, the target adjustment amount required to eliminate the deviation is calculated back based on the measured sensitivity, so that the pre-adjustment range matches the actual response characteristics of the soil layer, avoiding over-adjustment in sensitive soil layers and under-adjustment in sluggish soil layers. For soil layers with response sensitivity lower than the preset lower limit, the grouting parameter adjustment range is no longer increased. Instead, the thixotropic drag reduction characteristics of cement-soil are used to shorten the single insertion stroke and pause in segments to avoid ineffective grouting parameter adjustments.
[0026] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description
[0027] Figure 1 This is a flowchart of the SMW method for phased construction of piles in one of the technical solutions of the present invention;
[0028] Figure 2 This is a schematic diagram of the steel guide device structure in one of the technical solutions of the present invention;
[0029] Figure 3 This is a schematic diagram of the connection structure between the H-beam and the waler in one of the technical solutions of the present invention;
[0030] Figure 4 This is a schematic diagram of the ratio curve criterion for identifying soil layer interfaces in one of the technical solutions of the present invention;
[0031] Figure 5 This is a flowchart illustrating the triggering process of the three-stage auxiliary measures for steel section removal in one of the technical solutions of this invention.
[0032] Among them, 1: guide frame; 2: roller; 3: H-beam; 4: waler; 5: turnbuckle. Detailed Implementation
[0033] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.
[0034] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.
[0035] According to one embodiment of the present invention, such as Figure 1As shown, a phased construction method for SMW (Sequencing Welded Wrought Iron) pile foundations is provided. Specifically, before construction begins, the construction site is surveyed to determine the geological strata, groundwater level, surrounding buildings and structures, and underground pipelines. Based on the survey results, the construction site is divided into a deep foundation pit area and a shallow foundation pit area, and dewatering wells are installed in both areas. The deep foundation pit area has 11 dewatering wells (22m to 24m deep) and 4 observation wells (20m deep); the shallow foundation pit area has 12 dewatering wells (15m deep) and 3 observation wells (15m deep). The dewatering wells use sand-free pipes with an outer diameter of 400mm and a borehole diameter of not less than 600mm. The construction process for the dewatering wells is as follows: surveying and marking points, drilling with mud slurry, lowering filter pipes, filling the filter material with layers of 3mm to 7mm gravel and medium-coarse sand, and flushing the wells until the effluent is clear within 4 hours. After all the dewatering wells were completed, continuous pre-dewatering was carried out for 30 days in advance. Water level data was recorded daily through observation wells to ensure that the groundwater level in the work area was consistently 0.5m below the foundation pit.
[0036] After the dewatering is completed, construction of the SMW retaining system in the deep foundation pit area begins. First, surveying and layout are carried out to locate the positions of all three-axis cement-soil mixing piles in the deep foundation pit. Guide trenches with a depth of 1.0m to 1.5m are excavated, and steel guide devices are installed along the upper edge of the trenches. The three-axis cement-soil mixing piles use a diameter of 850mm and a spacing of 600mm. P.O42.5 grade ordinary Portland cement is used, with a cement admixture of not less than 20%, a water-cement ratio of 1.5, and a pumping pressure of 1.5MPa to 2.5MPa. After the three-axis pile driver is in place, the machine body is aligned, and the verticality of the column is controlled within 1 / 250. The mixing and sinking speed is controlled between 0.4m / min and 0.7m / min, and the lifting speed does not exceed 0.8m / min. H-beams should be inserted within 30 minutes after pile completion.
[0037] Before inserting the H-beams, a friction-reducing agent is applied to the surface of the steel section to facilitate later removal. Grouting pipes are pre-fixed to the inner side of the flanges of the steel section, arranged along its length, with grout outlets located at the insertion end and distributed along the side of the steel section. When applying the friction-reducing agent, the grout outlets are temporarily covered with adhesive tape to prevent blockage by cement and soil during insertion. The steel section used is 700mm × 300mm Q345B grade H-beams, arranged in a close-packed configuration. The steel section is inserted into the mixing pile via a guide device installed along the upper edge of the guide trench. The guide device is located within 1m above ground level and consists of two independently configured guide frames. Each guide frame has four adjustable rollers: two rollers correspond to the outer surfaces of the two flanges of the H-beam, and the other two rollers correspond to the two side surfaces of the web. Each roller is connected to the guide frame base via a precision screw; rotating the screw allows the roller to extend or retract horizontally, achieving independent fine-tuning in four directions. Before inserting the steel section, the design pile positions are laid out, and the extension of each roller is adjusted to ensure that the gap between the inner edge of the roller and the designed outer contour of the steel section is controlled within 5mm, forming a cross-shaped limiting space. During the steel section insertion process, a total station or theodolite is used to monitor the verticality and planar position of the steel section in real time. When a deviation is detected, the lead screw in the corresponding direction is rotated to push the roller to apply a corrective force to the steel section. The verticality deviation of the steel section insertion is controlled within 1 / 250, and the planar position deviation is controlled within 20mm.
[0038] During the insertion of the steel section, the lateral resistance and sinking velocity of the steel section are collected in real time. The lateral resistance of the cement-soil acting on the steel section during sinking can be collected by a strain gauge force measuring element attached to the surface of the steel section, or calculated from the difference in axial force at the upper and lower sections of the steel section; the sinking velocity can be collected by a distance measuring element set on the guide device. Let the lateral resistance collected at a certain moment be F(t) and the sinking velocity be v(t), then the real-time ratio R(t) is calculated, which is defined as R(t) = F(t) / v(t). When the steel section enters each soil layer, the average lateral resistance of the initial section of that soil layer is used as the basis for calculation. and average sinking speed Calculate the baseline ratio R0 of this soil layer, which is defined as follows: .
[0039] The first and second rates of change of R(t) are calculated in real time. The first rate of change refers to the change in R(t) between adjacent data points, i.e. This reflects the rate of change of R(t). The second-order rate of change refers to the change between adjacent first-order rates of change, i.e. This reflects the changing trend of the rate of change of R(t). When multiple consecutive data points of the R(t) curve move in the same direction and the change between adjacent points increases successively, it is judged as an acceleration approach precursor, indicating that R(t) is accelerating away from the baseline and is about to exceed the threshold. When the absolute value of the second-order rate of change continuously exceeds the preset threshold, it is judged as a curvature anomaly precursor, indicating that the curvature of the R(t) curve is changing at an accelerated pace, indicating that a significant deviation is about to occur. The number of consecutive data points in the above judgment can be 3 to 5; the preset threshold can be determined based on the statistical value of the absolute value of the second-order rate of change of the normal insertion section of similar soil layers in this work area, for example, taking 2 to 3 times its average value, or calibrated by field test when there is no historical measured data. When acceleration approach precursor or curvature anomaly precursor occurs, the grouting parameters should be adjusted in advance before R(t) exceeds the preset threshold. The preset threshold of R(t) can be determined based on the statistical upper limit of the measured data of R(t) of the normal insertion section of similar soil layers in this work area, or calibrated by field test when there is no historical measured data. The specific adjustment direction is as follows: R(t) should converge towards the baseline ratio of the soil layer. When R(t) is greater than the baseline ratio R0, meaning the steel section's settling resistance increases relatively, the water-cement ratio should be increased or the grouting pressure increased to enhance the fluidity of the cement-soil mixture and reduce its lateral resistance to the steel section. When R(t) is less than R0, the water-cement ratio should be decreased or the grouting pressure reduced to restore the strength of the cement-soil mixture while ensuring smooth insertion. These adjustments can be achieved by changing the water-cement ratio setting and the pumping pressure setting of the grouting pump. During steel section insertion, the grouting pressure and flow rate of the grout supplied to the contact surface between the steel section and the cement-soil mixture via the grouting pipe can also be changed, subject to limitations imposed by the design strength of the cement-soil mixture and the allowable range of the water-cement ratio. When adjusting the grouting parameters, calculate the instantaneous deviation rate of R(t) relative to R0, and determine the pre-adjustment amount of the grouting parameters according to the numerical range to which the rate belongs. The larger the rate, the larger the pre-adjustment amount. After pre-adjustment, continue to monitor the first-order rate of change of R(t). When the first-order rate of change approaches zero, lock the current grouting parameters. When the first-order rate of change reverses, cancel the pre-adjustment and restore the original parameters.
[0040] After all SMW method piles are constructed, the laitance at the pile top is removed, and a C30 grade reinforced concrete capping beam with a cross-section of 1.2m × 0.8m is poured. Core samples from cement-soil mixing piles must be cured for 28 days until their strength meets the standards before proceeding with the earthwork excavation process.
[0041] Deep foundation pit excavation adopts a layered excavation method, with support preceding excavation. The thickness of a single excavation layer does not exceed 2m, and the length of a single excavation does not exceed 30m. A 30cm layer is left at the base for manual cleaning. A dedicated excavation ramp with a 1:3 slope is constructed on-site, and the ramp surface is fully paved with steel plates. When the first layer of excavation reaches 0.5m below the capping beam, a 609mm diameter, 16mm wall thickness steel pipe is hoisted as the first support. A pre-applied axial force of 200kN is applied in stages using jacks and then stabilized and locked. Excavation can only continue downwards after the support has passed inspection. When the layered excavation reaches 0.5m above the corresponding elevations of the second and third supports, trenches are excavated sequentially, double-section 45C I-beam walers are installed, steel supports are erected, and prestressing is applied. The walers are fixed to the steel surface using pre-drilled holes on the H-beams with hooks, wire ropes, and turnbuckles. After the walers are installed, steel supports are erected, and prestressing is applied in stages.
[0042] After excavation to the foundation elevation in layers, the foundation pit was inspected, and construction of the main structure of the underground water tank, the external wall waterproofing, and the protective layer was carried out. After the underground structure concrete strength and waterproofing layer passed inspection, symmetrical layered backfilling with 2:8 lime-soil was used. When backfilling reached 0.5m below the third support, the prestressing was unloaded in stages, and the third steel support was removed. Backfilling was carried out layer by layer, and the second and first supports were removed sequentially. The boundary area between the deep foundation pit area and the shallow foundation pit area was backfilled to the foundation elevation of the shallow foundation pit. A hydraulic pile extractor was used to extract all the H-beams in the deep foundation pit in sections, and the gaps in the extracted steel were simultaneously sealed by pressure grouting.
[0043] After the deep foundation pit is completely backfilled and the steel sections are recycled, the overlapping triaxial cement-soil mixing piles at the junction are broken down to the bottom of the shallow foundation pit. Then, triaxial cement-soil mixing piles with a diameter of 850mm and a spacing of 600mm are constructed in the shallow foundation pit area. The H-beams are 700mm×300mm in size and are arranged in an alternating pattern. C30 grade capping beams of the same specifications are then poured.
[0044] During the construction of the shallow foundation pit, the dewatering system operated continuously, maintaining the water level 0.5m below the foundation. When excavating in layers to the anchor cable design elevation 2.5m below the capping beam, a jet grouting rig was used to drill holes at a 20° angle with a diameter of 150mm. Two steel strands with a diameter of 15.2mm and a strength grade of 1860MPa were used as anchor cables, and pressure grouting was performed with cement grout with a water-cement ratio of 0.5 to 0.55. After the anchor body strength reached 15MPa, a prestress of 90kN was applied to lock it in place, and additional tensioning was performed promptly when stress loss occurred. After the anchor cables passed inspection, the shallow foundation pit was excavated in layers to the foundation. After manual cleaning and inspection of the pit, the main structure of the shallow water tank was constructed.
[0045] After the shallow water tank structure and waterproofing are completed, symmetrical backfilling with 2:8 lime-soil is carried out, the H-beams in the shallow foundation pit are removed, and the gaps in the pile holes are filled with grout. After the soil covering construction of the top slab of all underground water tanks is completed, the sealing operation of the dewatering wells is carried out in a unified manner: the wells are filled with crushed stone filter layer, pressure grouting is carried out in sections, and micro-expansion concrete is poured to seal the wellhead, thus completing the complete set of support construction.
[0046] Throughout the construction period, monitoring points for horizontal displacement at the top of the slope, vertical displacement, deep inclinometers, axial force of steel supports, anchor cable tension, groundwater level, and settlement of surrounding buildings will be set up. During the first-stage deep foundation pit excavation phase, monitoring will be conducted 1 to 2 times daily. If the monitoring data reaches the warning value, construction will be suspended and reinforcement measures will be taken.
[0047] Compared to traditional methods of one-time enclosure and overall excavation of the entire foundation pit, and the problems of retaining deformation, base heave, and slope slippage caused by mutual disturbance of soil pressure due to elevation differences when constructing shallow pits first and then deep pits or simultaneously constructing shallow and deep pits, this method staggers the construction of deep and shallow foundation pits in time. The construction of the deep foundation pit is completed first, followed by the construction of the shallow foundation pit, thus avoiding mutual soil disturbance between the deep and shallow foundation pits. During the steel section insertion process, the grouting parameters are fed forward and controlled by real-time acquisition of the ratio of lateral resistance and sinking velocity. Based on this, the joint analysis of the first-order and second-order rates of change enables advanced prediction and early intervention of grouting parameter anomalies. Compared to traditional reactive adjustments, this method can respond more promptly to changes in soil conditions.
[0048] Furthermore, the aforementioned phased construction and real-time ratio control of steel section insertion work in tandem: shallow foundation pit support construction only begins after deep foundation pit backfilling and steel section recovery are completed. This ensures that steel section insertion in the shallow foundation pit area is not interfered with by the excavation and support erection of the deep foundation pit, allowing ratio monitoring and parameter pre-adjustment to be carried out under stable operating conditions. The improved steel section insertion efficiency also shortens the single-pile operation time, reducing the risk that the steel section may not be able to be inserted to the design elevation due to exceeding the initial setting time of the cement-soil mixture. Together, these two factors enhance the construction controllability of the deep and shallow foundation pit support system in complex, intersecting strata.
[0049] According to another embodiment of the present invention, a method for phased construction of SMW piles is provided, which refines the pre-adjustment of grouting parameters and the judgment of the pre-adjustment effect in step three above as follows: After the appearance of the precursor morphology, the instantaneous deviation rate of the current real-time ratio R(t) relative to the baseline ratio R0 is calculated. The deviation rate can be calculated using a sliding window linear fitting method, that is, taking a fixed number of data points forward from the current time as the endpoint, and performing linear fitting on the R(t) data within the window. The slope k of the fitted line is the instantaneous deviation rate at that time. The window size can be determined according to the data sampling frequency. For example, when the sampling frequency is one data point every 0.05m, the window can take 5 to 10 data points. A positive instantaneous deviation rate k indicates that R(t) has a positive deviation trend relative to R0, and a negative k indicates a negative deviation trend.
[0050] The pre-adjustment amount is adaptively determined based on the magnitude of the instantaneous deviation rate. The pre-adjustment amount can be divided into several levels: a smaller pre-adjustment is used when the absolute value of the deviation rate |k| is small, and a larger pre-adjustment is used when |k| is large. Specifically, a baseline value k for the deviation rate can be set. ref k ref The value is determined based on the statistical average deviation rate measured in the normal insertion section for similar soil layers in this work area. If no historical measured data is available, the value can be calculated by dividing the baseline ratio R0 by the depth interval of the data points from 0.5% to 2%. When |k| is less than 0.5k... ref When |k| is at 0.5k, the pre-adjustment amount is 30% to 50% of the normal adjustment amount; ref Up to 1.5k ref When |k| is between 1.5k and 60% of the normal adjustment amount, the pre-adjustment amount should be between 80% and 1.5k. ref At this time, the pre-adjustment amount is 100% to 120% of the normal adjustment amount. The normal adjustment amount refers to the adjustment range of the water-cement ratio or grouting pressure conventionally determined based on the degree of deviation between R(t) and R0. The specific numerical allocation of the pre-adjustment amount is determined in real time by the control system according to the above-mentioned classification rules based on the deviation rate calculated in real time.
[0051] After the pre-adjustment is implemented, the first-order rate of change of R(t) is continuously monitored to determine the effectiveness of the pre-adjustment measures and what subsequent actions should be taken. First-order rate of change This reflects the rate of change of R(t). The pre-adjusted first-order rate of change is continuously tracked to determine its trend.
[0052] When the first-order rate of change approaches zero, it indicates that R(t) has stabilized and no longer deviates from the baseline. At this point, the pre-adjustment measures have achieved the expected effect, and the current grouting parameters are locked in. A specific criterion for determining when the first-order rate of change approaches zero can be: within 5 to 10 consecutive data points, the absolute value of the first-order rate of change |ΔR| is less than the preset steady-state threshold ε.s Steady-state threshold ε s The baseline ratio R0 can be taken as 0.5% to 2%, and the specific value can be adjusted according to the variability of the soil layer on site. When the above conditions are met, the current grouting parameters will be used as fixed parameters for subsequent construction of this soil layer, and no further adjustments will be made.
[0053] When the first rate of change reverses, i.e. ΔR(t) i ) and ΔR(t i-1 If the sign of R(t) is reversed, it indicates that the deviation direction of R(t) has reversed, and the pre-adjustment measures may have been excessive or incorrect in direction. In this case, the pre-adjustment should be cancelled and the original parameters restored. The specific criterion for determining if the first-order rate of change is reversed is: the sign of the first-order rate of change changes, and the first-order rate of change maintains the new sign for 2 to 3 consecutive data points after the change. When the above conditions are met, the grouting parameters should be restored to the original parameters, and further judgment should be made based on subsequent monitoring data.
[0054] If the first-order rate of change neither approaches zero nor reverses, but continues to maintain its original direction of change but at a slower rate, then the current preset parameters remain unchanged, and the subsequent changes in the first-order rate of change are monitored until it approaches zero or reverses, at which point the corresponding locking or undoing operation is performed.
[0055] Compared to existing technologies that adjust based on monitoring data exceeding limits and rely on lookup tables from a pre-set parameter database for adjustment, this implementation method adaptively determines the pre-adjustment amount through instantaneous deviation rate and judges the pre-adjustment effect by continuously monitoring the first-order rate of change after pre-adjustment: if it approaches zero, it is locked; if it reverses, it is canceled, forming an adaptive control closed loop that does not require a pre-set database and is entirely driven by real-time data.
[0056] According to another embodiment of the present invention, a method for phased construction of SMW piles is provided. After pre-adjustment, the change sequence of the first-order rate of change is monitored. The first-order rate of change ΔR(t) i )=R(t i )-R(t i-1 This reflects the rate of change of R(t). A sequence of changes refers to a sequence of first-order rate of change values arranged in chronological order, i.e., {ΔR(t)} i ),ΔR(t i-1 ),…,ΔR(t i-n+1 )}. New data points are continuously collected. For each new data point, the first-order rate of change for that point is calculated and added to the end of the change sequence. Simultaneously, the oldest data point at the beginning of the sequence is removed, maintaining a fixed sequence length. The sequence length n can be determined based on the data sampling frequency. For example, when the sampling frequency is one data point every 0.05 m, the sequence length can be 10 to 20 data points.
[0057] The sign and absolute value trends of the change sequence are determined. When multiple consecutive data points in the sequence have the same sign and their absolute values decrease successively, the first-order rate of change is determined to be in a monotonically decaying state approaching zero. Same sign means that all data points in the sequence are either all positive or all negative, i.e., sgn(ΔR(t) = ... j All are the same. The absolute value decreasing successively means that the absolute values of adjacent data points satisfy |ΔR(t)|. j )|>|ΔR(t j+1 The number of consecutive data points can be between 5 and 10. When the above conditions are met, it indicates that the rate of change of R(t) is continuously slowing down and there is no sign of reverse fluctuation. Based on this, it can be judged that the first rate of change is converging towards zero.
[0058] Predict the number of data points required for the first-order rate of change to reach zero based on the current decay rate. Decay rate k d The decay rate can be obtained by linearly fitting a series of recent data points in the change sequence; the slope of the fitted line is the decay rate. The number of data points used for fitting can be between 5 and 8. Based on the decay rate k... d And the value of the current first-order rate of change ΔR(t) current Extrapolate the number of data points *m* required for the first-order rate of change to continue decaying to zero at the current rate. The prediction can be calculated using linear extrapolation, i.e. This gives the number of data points required to predict that the value will reach zero.
[0059] When the number of data points m required for the prediction to reach zero is less than a preset threshold m th At that time, a locking command is issued in advance. Preset threshold m th The number of data points can be determined based on the data sampling frequency and the response speed of the control system; for example, 3 to 5 data points can be used. The predicted value m is less than m... th When this occurs, it indicates that the first-order rate of change will reach zero within a very short time. Locking the parameters in advance at this point can avoid control delays caused by waiting for confirmation. After the lock command is issued, the current grouting parameters are used as fixed parameters for subsequent construction of this soil layer, and no further adjustments are made.
[0060] If the predicted first-order rate of change does not decay to zero as expected, the lock is lifted and the original parameters are restored. Specific manifestations of this failure include: the sign of the first-order rate of change reverses, or the absolute value of the first-order rate of change stops decreasing or even increases within two to three consecutive data points. When these situations occur, it indicates that the decay trend upon which the previous prediction was based has changed, the condition for early locking no longer holds, the grouting parameters are restored to their original values, and the monitoring and judgment process is re-entered.
[0061] Compared to existing technologies that wait for parameters to stabilize before locking, and those that rely solely on whether the current value reaches a threshold to determine locking, this implementation method monitors the sequence of changes in the first-order rate of change, identifies its monotonically decaying trend toward zero, predicts the timing of reaching zero before it decays to zero, and issues a locking command in advance. It also sets up a mechanism to cancel prediction errors, thus forming a predictive locking method with fault tolerance.
[0062] According to another embodiment of the present invention, a method for phased construction of SMW piles is provided. After pre-adjustment, the decay trend of the first-order rate of change is predicted. When the number of data points m required to reach zero is less than a preset threshold m... th Instead of immediately locking, it further determines whether the absolute value of the current first-order rate of change, |ΔR|, has dropped to the preset steady-state threshold ε. s The following is the steady-state threshold ε. s It is a pre-set value used to determine whether the first-order rate of change has entered the steady-state region. Steady-state threshold ε s It can be determined based on the value of the baseline ratio R0, for example, it can be taken as 0.5% to 2% of R0. When |ΔR| < ε s When the change rate of R(t) is small enough, it indicates that the change rate of R(t) has entered an acceptable stable range.
[0063] A locking command is issued only when both of the following conditions are met: the number of data points m required to predict zero is less than a preset threshold m. th Furthermore, the absolute value of the current first-order rate of change, |ΔR|, has decreased to the preset steady-state threshold ε. s The simultaneous fulfillment of both conditions indicates that the first-order rate of change is approaching zero. At this point, its current value is within a stable range, and the reliability of the locking decision is correspondingly improved. If either of the two conditions is not met, monitoring continues, and locking is not executed.
[0064] When the absolute value of the first-order rate of change falls below the steady-state threshold, i.e., |ΔR| < ε s When the countdown begins, a lock countdown is initiated. The countdown duration can be determined based on the data sampling frequency, for example, corresponding to the acquisition time of 5 to 10 data points. During the countdown, the change in the first-order rate of change is continuously monitored. If the first-order rate of change does not increase in the opposite direction during the countdown, it indicates that R(t) remains stable within the stable range and there are no signs of re-deterioration. Locking is automatically executed when the countdown ends.
[0065] If the first-order rate of change increases in the opposite direction during the countdown, that is, the sign of the first-order rate of change changes or the absolute value |ΔR| increases again beyond the steady-state threshold ε. sThis indicates that the previous stable trend may have been just a temporary fluctuation, and R(t) may have re-entered a deviated state. At this point, the countdown is terminated and the original parameters are restored, all previous adjustments are undone, and monitoring and judgment are resumed from the initial state.
[0066] Compared to existing technologies that rely on a single condition to determine the locking timing, and those that rely on a fixed delay before forced locking, this implementation triggers locking by dual determination of predicted and steady-state conditions. A countdown verification window is set before locking, and the state of the first-order rate of change is continuously monitored within the window: if stable, locking occurs; if deteriorating, termination occurs, thus forming a locking decision mechanism with a verification step.
[0067] According to another embodiment of the present invention, a method for phased construction of SMW (Sectional-Mechanical Wrought Iron) piles is provided. After the steel section removal operation begins, the pile extraction force of the pile extractor is monitored in real time, and a pile extraction force-displacement curve is plotted based on the extraction depth. A pressure sensor can be integrated into the hydraulic system of the pile extractor, and the pile extraction force is calculated by collecting hydraulic oil pressure data. Displacement data can be obtained through a displacement sensor or depth encoder installed on the pile extractor. The measured pile extraction force-displacement curve is compared with the theoretical curve at each depth, and the deviation value and deviation change rate at each depth are calculated. The theoretical curve is a standard extraction resistance curve pre-calculated based on parameters such as the embedment depth of the steel section, cross-sectional dimensions, and design strength of the cement-soil. The triggering process for the three-level auxiliary measures is as follows: Figure 5 As shown.
[0068] The deviation value reflects the difference between the measured pile extraction force and the theoretical pile extraction force, while the deviation change rate reflects the rate at which the deviation changes with increasing depth. When the deviation value exceeds the first preset value δ1 and the deviation change rate is positive, the first-level auxiliary measures are activated. The first preset value δ1 can be determined as a percentage of the theoretical maximum pile extraction force, for example, it can be taken as 10% to 20% of the theoretical maximum pile extraction force. A positive deviation change rate indicates that the deviation of the measured pile extraction force from the theoretical value is intensifying with increasing depth, and the removal condition of the steel section is deteriorating. The first-level auxiliary measures involve injecting high-pressure water or bentonite slurry into the contact surface between the steel section and the cement-soil to reduce the extraction resistance by lubricating and softening the contact surface.
[0069] When the rate of change of deviation continues to increase beyond the preset rate r th When this occurs, the primary measure is skipped and the secondary auxiliary measure is initiated directly. Preset rate r th This can be determined based on engineering experience; for example, the deviation rate per meter of depth can be taken as exceeding the theoretical value by 5% to 10%. If the deviation rate continues to increase beyond r... th This indicates that the steel section's removal condition is rapidly deteriorating, and the resistance reduction effect of the primary measures may not be sufficient to curb the worsening trend. Therefore, stronger intervention measures are needed. The secondary auxiliary measure is to apply high-frequency micro-vibration to assist in the removal.
[0070] When applying high-frequency micro-vibration, the optimal frequency of the historical steel section most similar to the current steel section's working condition is first retrieved from the extraction resistance database. The extraction resistance database is gradually established during construction. After each steel section is extracted, its specifications, extraction depth, cement-soil age, extraction force at each frequency, and the finally determined optimal frequency are stored in the database. During retrieval, the current steel section's specifications, extraction depth, cement-soil age, and other parameters are matched with historical records in the database based on similarity. The optimal frequency corresponding to the record with the highest similarity is selected as the starting frequency f. start With f start Perform a local frequency sweep outwards from the center, with the sweep range being f. start ±2Hz to f start ±5Hz.
[0071] If no similar operating conditions are found in the database, a hill-climbing search is initiated simultaneously from multiple different starting frequencies. The number of starting frequencies can be 3 to 5, evenly distributed across the possible frequency range, such as 10Hz, 20Hz, and 30Hz. Each search thread runs independently, starting from its own starting frequency and fine-tuning the frequency upwards and downwards, measuring the pull-out force at each frequency, and fine-tuning the frequency in the direction of decreasing pull-out force until the minimum pull-out force for that thread is found and the corresponding frequency is recorded. After all threads have completed their tasks, the pull-out forces of each thread are compared, and the frequency corresponding to the minimum value is selected as the initial search starting point.
[0072] During the frequency sweep process, the pile extraction force signal and vibration acceleration signal are acquired synchronously in real time. The pile extraction force signal comes from the pressure sensor of the hydraulic system, and the vibration acceleration signal comes from the acceleration sensor mounted on the vibratory hammer or the surface of the steel profile. Both signals are acquired synchronously at the same sampling frequency to ensure that the data at the same moment are time-aligned. The cross-power spectrum S of the two signals is calculated. xy (f), which is defined as Where X(f) and T(f) are the Fourier transforms of the pile extraction force signal x(t) and the vibration acceleration signal y(t), respectively, Y*(f) is the conjugate of Y(f), and f is the frequency. The peak value of the cross-power spectrum reflects the intensity of the common frequency component in the two signals. The frequency corresponding to the peak value of the cross-power spectrum is extracted as the effective dominant frequency f. e This frequency is the dominant frequency that appears simultaneously in both the pile extraction force signal and the vibration acceleration signal, eliminating noise frequencies that only appear in a single signal. (The last part, "f," appears to be an unrelated fragment and is omitted from the translation.) e Centered on a preset narrowband range, a fine frequency sweep is performed. The frequency sweep range can be taken as f. e ±1Hz to f e ±3Hz. During the fine frequency sweep process, the frequency with the minimum pile extraction force is determined as the optimal frequency f for the current steel profile. opt and f optStore the data in the extraction resistance database for reference during subsequent steel section extraction.
[0073] When the deviation exceeds the second preset value δ2 and the rate of change of the deviation is still positive, the third-level auxiliary measures are initiated. The second preset value δ2 is higher than δ1, for example, it can be taken as 30% to 40% of the theoretical maximum pile extraction force. If the deviation exceeds δ2 and the rate of change of the deviation is still positive, it indicates that the extraction resistance of the steel section has reached a very high level and is still increasing, and the first and second-level measures have failed to effectively improve the extraction status. At this time, the extraction is suspended, the steel section is locally cut, and the steel section is extracted segment by segment.
[0074] Compared to existing technologies that use fixed thresholds to trigger fixed measures in a tiered removal process, and those that rely on manual experience to determine when to switch removal methods, this implementation achieves dynamic, skip-step triggering of tiered measures through continuous monitoring of the deviation change rate: when the deterioration rate is rapid, it automatically skips lower-level measures and directly enters higher-level interventions, avoiding the time delays that may be caused by trying each level step by step. Simultaneously, an adaptive optimization of the vibration frequency is achieved through a combination of database retrieval, cross-power spectrum analysis, and multi-threaded hill-climbing search.
[0075] According to another embodiment of the present invention, a method for phased construction of SMW (Sectional Method for Welded Wrought Iron) piles is provided. When retrieving historical steel profiles most similar to the current steel profile condition from a pull-out resistance database, the confidence level C of the retrieval results is calculated simultaneously. The confidence level is a value between 0 and 1, used to measure the similarity between the retrieved historical records and the current condition. The calculation of the confidence level can comprehensively consider the matching degree of multiple condition parameters, including the matching degree w1 of steel profile specifications, the similarity of pull-out depth w2, and the similarity of cement-soil age w3, etc. Each parameter can be assigned different weights α1, α2, and α3. The weight values can be preset according to the importance of each parameter's influence on the optimal frequency, and α1 + α2 + α3 = 1. The total confidence level is obtained by weighted summation of the matching degrees of each parameter. .
[0076] The frequency sweep process is divided into three modes based on the confidence level C. When the confidence level is higher than a preset high threshold C... high At that time, enter fast mode. C high A value between 0.85 and 0.95 can be used. In fast mode, the historical steel sections retrieved are highly similar to the current steel sections in terms of operating conditions, and their optimal frequencies have high reference value. The retrieved optimal frequency f can be used directly. opt The starting frequency is used as the starting frequency, and a local narrowband frequency sweep is performed only in its vicinity. The sweep range can be taken as f. opt ±1Hz to f opt ±2Hz. Fast mode does not initiate multi-start parallel hill climbing search, nor does it calculate cross-power spectrum, and completes frequency locking in the shortest possible time.
[0077] When the confidence level is lower than the preset low threshold C low At that time, enter fine mode. C low A value between 0.4 and 0.6 can be used. In fine-grained mode, the historical steel sections retrieved have low similarity to the current steel sections in terms of working conditions, limiting the reference value of their optimal frequencies. Therefore, database search results are discarded, and a multi-starting-point parallel hill-climbing search is initiated simultaneously from multiple different starting frequencies. Each thread independently searches for its own minimum extraction force and corresponding frequency. Simultaneously, extraction force and vibration acceleration signals are acquired during the frequency sweep, and the cross-power spectrum of the two signals is calculated to extract the effective dominant frequency to guide fine-grained frequency sweeping.
[0078] When the confidence level C is between C low With C high In between, i.e., C low ≤C≤C high Then, it enters mixed mode. In mixed mode, the search results have some reference value but are not reliable enough. The optimal retrieved frequency f is used. opt As one of the starting frequencies for parallel hill climbing from multiple starting points, it is searched alongside other preset starting frequencies. The number of other preset starting frequencies can be 2 to 4, evenly distributed within the possible frequency range. After each thread searches in parallel, the minimum pull-out force and its corresponding frequency found are summarized and compared, and the optimal result is selected.
[0079] The switching between the three modes is completed automatically based on confidence level, requiring no manual intervention. The mode switching threshold C... low and C high Adjustments can be made based on actual construction conditions. For example, in sections with significant soil layer variations, C can be appropriately reduced. high Improve C low More refined or mixed modes are preferred to fully explore the frequency space; in sections with uniform soil layers, C can be appropriately increased. high More people are opting for the fast mode to improve efficiency.
[0080] Compared to existing technologies that employ a single frequency sweeping strategy and rely on manual judgment to select frequency sweeping parameters, this implementation method achieves automatic switching between fast mode, fine mode, and hybrid mode through confidence level classification: high confidence level directly utilizes historical experience to quickly lock the target, low confidence level conducts comprehensive exploration, and medium confidence level adopts a hybrid strategy, thereby improving frequency sweeping efficiency while ensuring optimization accuracy.
[0081] According to another embodiment of the present invention, a method for phased construction of SMW (Sectional-Mesh Welded) piles is provided. The soil interface identification method described in this embodiment can be implemented independently in both the deep and shallow foundation pit steel insertion stages. Side resistance and sinking velocity data are continuously collected during the steel insertion process, and the real-time ratio R(t) at each depth is calculated. As the steel continues to sink, the second-order rate of change of R(t) with depth is simultaneously calculated. The second-order rate of change is calculated by first calculating the first-order rate of change of R(t) between adjacent depth points. That is, the change of R(t) with depth; then calculate the difference between adjacent first-order rates of change. This yields the second rate of change. The second rate of change reflects the trend of the rate of change of R(t). When Δ 2 When R changes from a positive value to a negative value or from a negative value to a positive value, that is... and The sign of the curve is reversed, and this location is identified as a candidate zero-crossing point. A zero-crossing point indicates that the direction of curvature of the R(t) curve reverses at that point, marking a significant change in the curve's shape. The selection of candidate zero-crossing points relative to the first and second depth windows is as follows: Figure 4 As shown.
[0082] After detecting a candidate zero-crossing point, a segment of R(t) data within a fixed depth range L1 before the zero-crossing point is extracted as the front window data, and a segment of R(t) data within a fixed depth range L2 after the zero-crossing point is extracted as the back window data. The depth ranges L1 and L2 of the front and back windows can be the same or different. For example, L1 can be 0.3m to 0.5m, and L2 can be 0.15m to 0.3m. The selection of the window length needs to consider the soil layer thickness and the data sampling density; a shorter window can be used when the soil layer is thinner, and a longer window can be used when the soil layer is thicker.
[0083] Calculate the energy values E1 and E2 for the data in the front and back windows respectively. The energy value is calculated by summing the squares of the amplitudes of each data point within the window. Where n1 and n2 are the number of data points in the front and back windows, respectively. The energy value E1 of the front window reflects the overall strength of the R(t) signal before the zero-crossing point, and the energy value E2 of the back window reflects the overall strength of the R(t) signal after the zero-crossing point. When the energy ratio η = E2 / E1 exceeds the preset threshold η... th At that time, the candidate zero-crossing point is confirmed as the true soil layer interface. Energy ratio threshold η th It can be determined based on field tests or experience; for example, it can be taken as 1.5 to 3.0. η exceeds η th This indicates that the signal intensity of R(t) on both sides of the zero point has changed significantly. This change usually corresponds to the change in the mechanical properties of the soil layer, i.e., the soil layer interface.
[0084] After confirming the soil layer interface, the baseline ratio R0 is recalibrated, that is, the baseline ratio R0 of the soil layer is recalibrated according to the measured values of the initial segment of the new soil layer. The calibration method is as follows: taking the interface location as the starting point of the new soil layer, the average lateral resistance within a depth range of 0.3m to 0.5m of the initial segment is collected. and average sinking speed Based on this, the baseline ratio of the soil layer is calculated. This is used to determine the degree of deviation of R(t) within the soil layer.
[0085] Compared to existing technologies that rely solely on the abrupt change points or inflection points of single-parameter curves to determine soil layer interfaces, and those that rely on pre-construction well logging data for offline stratification, this implementation method identifies the morphological change position of the curve by using the zero-crossing point of the second-order rate of change, and then confirms it by using the energy ratio η=E2 / E1 of the windows before and after the zero-crossing point. This forms a two-stage determination process, which can complete the real-time identification of soil layer interfaces and the recalibration of the baseline during the steel section insertion construction process.
[0086] According to another embodiment of the present invention, a method for phased construction of SMW (Sequencing-Based Weft) piles is provided, in which a multi-window scanning strategy is adopted for the confirmation of candidate zero-crossing points. This method is also applicable to both the deep foundation pit steel insertion stage and the shallow foundation pit steel insertion stage; the identification processes of the two stages are carried out independently and do not affect each other. After a candidate zero-crossing point is detected, the energy ratio is not calculated using a single fixed window length. Instead, multiple different window lengths are used to extract R(t) data before and after the zero-crossing point, centered on the candidate zero-crossing point. The number of window lengths can be determined according to actual needs; for example, three to five different window lengths can be used. The range of window lengths can cover a certain interval from short to long; for example, a short window is 0.1m to 0.2m, a medium window is 0.3m to 0.4m, and a long window is 0.5m to 0.6m.
[0087] For each window length L k Let k range from 1 to N, where N is the total number of window lengths; calculate the ratio of the energy value of the window after the zero-crossing point to that of the window before the zero-crossing point for that window length. ,in and These represent the energy values of the front and back windows for the k-th window length, respectively. Different window lengths exhibit varying sensitivities to signal changes on both sides of the zero-crossing point: short windows are sensitive to local variations and can capture interface signals from thinner soil layers, but are also easily affected by local noise; long windows are sensitive to overall trends and have stronger noise resistance, but may smooth out signals from thin-layer interfaces. Therefore, a single window length cannot simultaneously balance sensitivity and noise resistance.
[0088] Select the maximum value η from the energy ratios corresponding to all window lengths.max The final energy ratio determination value at this zero-crossing point is, i.e. The logic behind selecting the maximum value is that the actual soil layer interface will cause a significant difference in signal intensity on both sides of the zero-crossing point. This difference will be reflected under multiple window lengths, and the energy ratio will reach its highest value when a certain window length happens to match the thickness of the soil layer and the fluctuation characteristics of the signal. By selecting the maximum value, the most suitable window length can be adaptively found under different soil layer thicknesses and different signal characteristics.
[0089] When η max >η th At that time, the candidate zero-crossing point is confirmed as the soil layer interface, and the baseline ratio R0 of the soil layer is recalibrated according to the measured value of the initial segment of the new soil layer. Energy ratio threshold η th You can choose between 1.5 and 3.0.
[0090] Compared to existing technologies that use a fixed window length to calculate the energy ratio and rely on manual selection of the window length, this embodiment uses multiple window lengths for scanning and the selection of the maximum energy ratio to enable the window length to adaptively match the interface signal characteristics of soil layers of different thicknesses, thus avoiding missed or incorrect judgments caused by improper selection of the window length.
[0091] According to another embodiment of the present invention, a method for phased construction of SMW (Sequencing Waveform Wave) piles is provided. The confirmation of candidate zero-crossing points uses a waveform shape similarity determination method, rather than energy value comparison. This embodiment can be implemented independently in both deep and shallow foundation pit steel insertion stages to identify the location where the steel crosses the soil layer interface. After detecting a candidate zero-crossing point, a depth range L preceding the zero-crossing point is extracted. ref Using the R(t) waveform within the range as a reference waveform, a depth range L is extracted after the zero-crossing point. comp The R(t) waveform within the range is used as the comparison waveform. ref and L comp The amplitudes can be the same, for example, all between 0.2m and 0.4m. Unlike the energy method, the shape similarity method does not focus on the magnitude of the signal amplitude, but only on the morphological characteristics of the waveform.
[0092] Calculate the shape similarity ρ between the reference waveform and the comparison waveform. The shape similarity can be calculated using a normalized cross-correlation method. Specifically, the reference waveform R... ref (z) and comparison waveform R comp (z) Subtract their respective mean μ ref and μ comp Then divide by their respective standard deviations σ ref and σ compThis normalizes the amplitude of the two waveforms to the same scale. The normalized reference waveform is... The normalized comparison waveform is Then, the cross-correlation coefficient ρ between the two normalized waveforms is calculated, which is defined as follows: Where n is the number of data points within the window. The cross-correlation coefficient ρ ranges between -1 and 1. The closer to 1, the more similar the shapes of the two waveforms; the closer to -1, the opposite the shapes; and the closer to 0, the uncorrelated shapes. The cross-correlation coefficient ρ is the numerical value of shape similarity.
[0093] When the shape similarity ρ is lower than the preset threshold ρ th When, i.e., ρ < ρ th This confirms that the candidate zero-crossing point is the soil layer interface. Shape similarity threshold ρ th A value between 0.3 and 0.6 is acceptable. A value below this threshold indicates a significant change in the waveform morphology within the two windows before and after the zero-crossing point. This morphological change reflects a shift in the pattern of R(t) variation with depth, corresponding to changes in the mechanical properties of the soil layer. When ρ ≥ ρ th If the waveform shape is basically the same before and after the zero point, the zero point may be caused by local noise or small fluctuations and is not considered as a soil layer interface.
[0094] After confirming the interface, the baseline ratio R0 of the soil layer is recalibrated according to the measured value of the starting section of the new soil layer, and the calibration method is as described above.
[0095] Compared to existing technologies that rely on signal amplitude differences to identify interfaces using energy methods and inflection point methods that only focus on single-point numerical abrupt changes, this implementation method determines interfaces by comparing the shape similarity ρ of waveforms before and after crossing the zero point. This method is completely independent of the absolute amplitude of the signal, thus effectively identifying interfaces even in weak interface scenarios where the signal strength difference between the two sides is small but the waveform shape changes. Furthermore, the shape similarity calculation is based on a normalized waveform, eliminating the influence of amplitude changes on the determination result and ensuring that the determination criteria remain consistent across different depths and soil layers.
[0096] In one embodiment of the construction method of the present invention, the insertion of the steel profile and the fixing of the waler adopt the following steel profile guiding construction system, such as... Figure 2 and Figure 3 As shown, the steel section guiding construction system consists of two parts: a steel section guiding device and a waler fixing device. The steel section guiding device is installed above the pile position of the cement-soil mixing pile to be inserted into the H-section steel 3, and is used to guide and correct the H-section steel 3 during the insertion of the H-section steel 3 into the cement-soil mixing pile.
[0097] The steel section guiding device comprises two independently arranged guide frames 1, one upper and one lower. Both guide frames 1 are horizontally arranged, with the upper guide frame 1 located 0.5m to 0.6m above the ground and the lower guide frame 1 located 0.5m to 0.6m below the upper guide frame 1, i.e., near or slightly above the ground. The two guide frames 1 are connected by vertical columns, the bottom of which are fixed to the ground on both sides of the guide trench or mounted on pre-embedded anchor bolts, ensuring the stability of the entire guiding device during steel section insertion. Each guide frame 1 is equipped with four adjustable rollers 2, two of which correspond to the outer surfaces of the two flanges of the H-beam 3, and the other two correspond to the two side surfaces of the web. The four rollers 2 are arranged around the cross-section of the H-beam 3 from four directions, forming a cross-shaped limiting space.
[0098] Each roller 2 is connected to the base of the guide frame 1 via a precision lead screw. The precision lead screw is arranged horizontally, passing through a through hole in the base of the guide frame 1 and mounted on the base of the guide frame 1 via a bearing seat. It can only rotate and cannot move axially. A handwheel or a drive motor can be installed on the extended end of the precision lead screw. A lead screw nut that mates with the precision lead screw is fixed on the mounting bracket of the roller 2. When the precision lead screw is rotated, the lead screw nut moves axially along the precision lead screw, causing the roller 2 to extend and retract horizontally. Since each roller 2 is driven by an independent lead screw, the four rollers 2 can achieve independent fine adjustment in four directions. The rollers 2 can be made of polyurethane or high-polymer wear-resistant nylon material, with a smooth surface to reduce friction with the surface of the H-beam 3 and avoid scratching the anti-friction coating on the surface of the H-beam. A scraper can be installed on the side of the roller 2. The blade of the scraper is in close contact with the surface of the roller 2. During the rotation of the roller 2, the cement slurry or mud adhering to the surface of the roller 2 is scraped off, preventing the accumulation of debris from affecting the normal rotation and guiding accuracy of the roller 2.
[0099] The working principle of the steel section guiding device is as follows: Before inserting the H-beam 3, according to the design pile position layout results, the operator rotates each precision screw to adjust the extension of the four rollers 2, so that the gap between the inner edge of the roller 2 and the designed outer contour of the H-beam 3 is controlled within 5mm, and the four rollers 2 together form a cross-shaped limiting space. After the H-beam 3 is lifted by the hoisting equipment, it passes through the cross-shaped limiting space, and a small gap is maintained between the inner edge of the roller 2 and the surface of the H-beam 3. During the steel section insertion process, the verticality and planar position of the H-beam 3 are monitored in real time from the ground using a total station or theodolite. Once the H-beam 3 is detected to have deviated in a certain direction, the operator immediately rotates the precision screw in the corresponding direction, pushing the roller 2 in that direction toward the surface of the H-beam 3. After the roller 2 contacts the surface of the H-beam 3, it applies a corrective force to the steel section, pushing the steel section back to the designed position. Because two layers of guide frames 1 are set up, the steel section is constrained at both the upper and lower height positions, which can effectively control the overall tilt of the steel section and avoid the possible rotation of the steel section when there is only one layer of guide frame 1.
[0100] The waler fixing device is installed on the inserted H-beam 3 to fix the waler 4 to the surface of the H-beam 3. The waler 4 is a large-section steel beam horizontally arranged on the side wall of the foundation pit, usually located in the middle or lower middle part of the H-beam 3, corresponding to the height of the foundation pit side wall that needs support. The waler fixing device includes pre-drilled holes, hooks, wire ropes, and turnbuckles 5. The pre-drilled holes are opened at the junction of the web and flange of the H-beam 3. For different specifications of H-beam 3, the specific location of the junction of the web and flange is determined according to the cross-sectional dimensions of the steel section. This junction is the corner area where the outer surface of the web and the inner surface of the flange meet. For a 700mm×300mm H-beam 3, this area is located on both sides of the web, extending approximately 20mm to 30mm downwards from the inner surface of the flange along the height direction of the web; for other specifications of steel sections, the location of this area can be determined in the same way according to the geometric dimensions of the web and flange in the cross-sectional diagram of the steel section. The reserved holes are located in the middle of this area, with the center of the hole situated on the mid-surface of the web thickness direction, ensuring a smooth transition between the hole wall and the steel profile surface. The reserved holes are drilled along the length of the H-beam 3 at the design elevation of the walers 4; that is, circular holes are drilled at the junction of the web and flange on the H-beam 3 corresponding to the installation height of each waler 4. The number of reserved holes is determined based on stress calculations; two to three holes can be reserved at the corresponding positions of each waler 4.
[0101] The hook is an L-shaped component, with one end bent to form a hook for inserting into a pre-drilled hole and hooking onto the hole wall; the other end is a straight rod section for connecting the wire rope. When installing the waler 4, the hook is inserted into the pre-drilled hole from one side of the H-beam 3. After passing through the hole, the hook hangs down naturally, hooking onto the edge of the hole wall, thus securing the hook to the H-beam 3. One end of the wire rope is connected to the straight rod section at the lower end of the hook, and the other end passes through the turnbuckle 5 and connects to a pre-set suspension point on the waler 4. The turnbuckle is a connector with a threaded adjusting rod; rotating the middle sleeve of the turnbuckle changes the distance between the two connection points, thereby tightening or loosening the wire rope. During installation, the wire rope is initially tightened, and then the turnbuckle 5 is rotated to further tighten the wire rope. Under the tension of the wire rope, the waler 4 is pressed against the surface of the H-beam 3 until the flange of the waler 4 is tightly fitted against the flange surface of the H-beam 3. The wire rope is set at an angle, so while the waler 4 is horizontally pressed against the surface of the steel section, the weight of the waler 4 is borne by the vertical component of the wire rope, thus eliminating the need to weld supporting brackets onto the steel section. Because the wire rope and turnbuckle 5 are adjustable, even if there is a slight positional deviation in the H-beam 3 during insertion, the waler 4 can be tightly fitted against the surface of the steel section by adjusting the length of the wire rope, without the need to add shims between the steel section and the waler 4 to adjust the gap. The entire installation process does not require any welding work in the foundation pit. In other embodiments, the waler 4 can also be fixed to the H-beam 3 by conventional fixing methods such as welding steel brackets to the flange of the steel section or fixing the waler 4 to the top of the steel section.
[0102] Compared to existing technologies that weld steel brackets to the flanges of the H-beam to support the waler 4, or use a sleeve method to attach the waler 4 to the top of the H-beam, this embodiment uses pre-drilled holes at the junction of the web and flange of the H-beam 3, along with hooks, wire ropes, and turnbuckles 5, to fix the waler 4 to the surface of the H-beam. The waler 4 and the H-beam are flexibly connected, eliminating the need for welding during installation and avoiding the risks of welding fumes, harmful gases, and open flames in the confined space of the foundation pit. Furthermore, the hook-and-hole connection is simple and reliable, and installation and disassembly are convenient, facilitating the recycling and reuse of the H-beam. The steel guide device and the waler fixing device share the same set of H-beams 3: the guide device ensures that the H-beams 3 are inserted precisely in the design position, and that the reserved hole is accurately positioned at the design elevation of the waler 4. This provides a positioning basis for the installation of the hooks of the waler fixing device. The two devices work together to form a complete construction system from steel insertion to waler 4 fixing.
[0103] According to another embodiment of the present invention, a method for phased construction of SMW piles is provided. Based on determining the pre-adjustment amount according to the grading of instantaneous deviation rates, it further introduces online calibration of the response sensitivity of the current soil layer to solve the problem of mismatch between the aforementioned grading scale and the actual response characteristics of the soil layer. During the insertion of the steel section, the change in R(t) caused by the same grouting parameter adjustment applied to different soil layers is not the same: in sensitive soil layers such as silt, a small parameter adjustment will cause a significant change in R(t); in insensitive soil layers such as silty clay, the same adjustment amount causes a very small change in R(t). If the relative adjustment amount is only obtained by looking up the table according to the grading of deviation rates, even the smallest setting in sensitive soil layers will result in over-adjustment, triggering cancellation and repeated adjustments, consuming the time window for steel section insertion; in insensitive soil layers, even the largest setting is insufficient to eliminate deviation, and the steel section still faces the risk of obstruction.
[0104] To eliminate the aforementioned uncertainties, this implementation method does not directly apply adjustment amounts according to the grade when pre-adjustment is initially required. Instead, a baseline detection adjustment amount Δ0 is applied for calibration. Δ0 is taken as one-third of the lower limit of the adjustment amount corresponding to the grade to which the soil layer belongs. This ensures that the adjustment amount itself does not cause significant deviation, and that the resulting change in R(t) can be distinguished by the sensor. After applying Δ0, ratio data of m consecutive depth points are collected. Since R(t) itself has a natural variation trend with depth, the measured sequence contains both the response caused by the adjustment and the natural drift of the strata. Therefore, it is necessary to separate the two: extrapolate the ratio sequence of n depth points before applying Δ0 using a quadratic polynomial to obtain the values of R(t) at the m depth points that should exist without adjustment, as the predicted sequence. Then, subtract the measured sequence from the predicted sequence point by point and take the average to obtain ΔR. pThis refers to the change in ratio caused solely by this single detection adjustment. Both m and n represent the number of data points in the depth direction, which can be determined based on the sampling interval and soil thickness; for example, m and n can be 5 to 10 respectively.
[0105] With ΔR p The ratio of Δ0 to the baseline ratio R0 of the soil layer is then divided to obtain the normalized response sensitivity S. The physical meaning of S is the relative change in the ratio caused by a unit relative adjustment in the soil layer. A large S value indicates that the soil layer is sensitive to grouting parameters, while a small S value indicates insensitivity. In subsequent pre-adjustments of this soil layer, the deviation ΔR from the current measured ratio relative to the baseline ratio R0 is used as the basis for the sensitivity. c Dividing by S, the target adjustment amount needed to eliminate the deviation can be calculated. This target adjustment amount is then reduced by the upper limit of the adjustment amount corresponding to the soil layer's classification, and constrained by the allowable ranges of water-cement ratio and grouting pressure, yielding the actual pre-adjustment amount. Thus, the pre-adjustment amount is changed from being obtained from a table to being calculated based on measured sensitivity, automatically being smaller in sensitive soil layers and larger in insensitive soil layers. A single adjustment can bring R(t) back to near the target range, avoiding repeated adjustments.
[0106] For soil layers with a response sensitivity S lower than the preset lower limit, it indicates that the soil layer itself is not sensitive to the adjustment of grouting parameters, meaning that further increasing the adjustment range will not achieve the corresponding drag reduction effect. In this case, if the adjustment is still increased in stages, it will only waste grouting material and cause the cement-soil strength to deviate from the design value. Therefore, this embodiment uses a mechanical method to reduce drag for this type of soil layer: keeping the current grouting parameters unchanged, shortening the single insertion stroke to a preset segment length, pausing for a preset time after each segment insertion, and then continuing insertion. The structural strength of the cement-soil will temporarily decrease after being agitated. The aforementioned segmented pauses allow the cement-soil near the steel section end to undergo local reshaping during each pause, thereby reducing resistance during the next segment insertion. The segment length and pause time can be determined based on the sinking speed and the sensitivity of the soil layer; for example, the segment length can be 0.3m to 0.5m, and the pause time can be 5s to 15s.
[0107] Each calibration yields a normalized response sensitivity S and a baseline ratio R0 for that soil layer, both stored in the insertion control database. Since the mechanical properties of the same soil layer are essentially the same at different pile locations within the same work area, when subsequent piles are constructed to that soil layer, S can be directly retrieved from the database and the pre-adjustment amount can be calculated in the aforementioned manner, eliminating the need for repeated detection and adjustment, thus reducing the number of detections and shortening the single-pile control time. Records in the database are associated with the depth range of the soil layer interface and the baseline ratio R0; if the difference between the soil layer interface depth or the baseline ratio R0 at a certain pile location and existing records exceeds a preset range, the pile location will be recalibrated.
[0108] Compared to methods that determine grouting parameter adjustment amounts solely based on deviation rate grading and methods that perform single-factor coarse adjustments based only on the magnitude of the deviation rate, this implementation method calibrates the normalized response sensitivity of the soil layer online through a single benchmark detection adjustment amount. This sensitivity is then used to calculate the target adjustment amount needed to precisely eliminate the deviation, ensuring the adjustment range matches the actual response characteristics of the soil layer. This solves the problem of over-adjustment of sensitive layers and under-adjustment of insensitive layers caused by the mismatch between the grading scale and the formation response characteristics. Simultaneously, for soil layers with sensitivity below the preset lower limit, a segmented, paused mechanical drag reduction method is used, avoiding ineffective increases in adjustment amounts in soil layers where grouting parameter adjustment has failed.
[0109] Application examples
[0110] The expansion project of a reclaimed water plant is located in an alluvial-diluvial landform unit in the North China Plain. The shallow part of the site contains a 4m to 6m thick layer of fill, followed by layers of silty clay, silt, and silty sand. The groundwater level is about 2.5m deep, and the permeability coefficient is about 0.2m / d. The project includes a temporary storage tank, a regulating tank, and an AO biological treatment tank. The temporary storage tank is 13.2m deep and constitutes a primary foundation pit, while the regulating tank (7.6m deep) and the AO biological treatment tank (6.1m deep) constitute secondary foundation pits. The maximum height difference between the deep and shallow foundation pits is 7.1m. A phased construction scheme, starting with the deep pits and then moving to the shallow ones, is adopted.
[0111] Before construction, the site was leveled and surveyed, dividing the site into a deep foundation pit area and a shallow foundation pit area. In the deep foundation pit area, 11 dewatering wells with a depth of 22m to 24m and 4 observation wells were installed. In the shallow foundation pit area, 12 dewatering wells with a depth of 15m and 3 observation wells were installed. The dewatering wells used 400mm outer diameter sand-free concrete filter pipes with a borehole diameter of 650mm. The filter material was 3mm to 7mm round gravel and medium-coarse sand. The boreholes were sealed with clay balls 2m below the ground. The wells were flushed within 4 hours after completion until the water was clear. Pre-dewatering was carried out 30 days in advance to lower the water level to 0.5m below the foundation.
[0112] After the water level is reached, the SMW retaining system for the deep foundation pit is constructed: a 1.2m deep guide trench is excavated along the edge of the foundation pit, and two independent guide frames with a layer spacing of 0.6m are installed on the guide trench. Each layer is equipped with four adjustable rollers, two of which correspond to the outer surfaces of the two flanges of the steel section, and the other two correspond to the two side surfaces of the web plate. Four-way independent micro-adjustment is achieved through precision screw rods.
[0113] The three-axis cement-soil mixing piles are constructed with a diameter of 850mm, a spacing of 600mm, and a single-hole connection to form a water-stop curtain; the cement content is 20%, the water-cement ratio is 1.5, and the pumping pressure is 1.8MPa to 2.2MPa; the mixing and sinking speed is 0.5m / min to 0.6m / min, the lifting speed is no more than 0.8m / min, the verticality of the column is within 1 / 250, and H-beams are inserted within 30 minutes for each pile.
[0114] The H-beams are made of 700mm×300mm Q345B grade steel, arranged closely together. Before insertion, two coats of thick-film asphalt-based friction reducer are applied, and the grout outlet of the grouting pipe is temporarily covered with tape. A cross-shaped limiting space is formed by the guide device rollers. During insertion, the verticality and plane position are monitored bidirectionally with a theodolite. If there is a deviation, the corresponding precision screw is rotated to correct the deviation. The verticality deviation is controlled within 1 / 250 and the plane position deviation does not exceed 20mm.
[0115] During the insertion process, the side resistance F(t) is collected in real time by the strain gauge force measuring element on the steel surface, and the sinking velocity v(t) is collected by the laser rangefinder of the guide device. The real-time ratio R(t) = F(t) / v(t) is calculated. When the steel enters each soil layer, the baseline ratio of the soil layer is calculated based on the average side resistance and the average sinking velocity within the initial 0.5m range.
[0116] The first and second order rates of change of R(t) are calculated in real time according to the aforementioned step three to identify the precursors. Before R(t) exceeds the threshold, the grouting parameters are adjusted. The pre-adjustment amount is determined according to the instantaneous deviation rate. After pre-adjustment, the first order rate of change is monitored to lock or cancel it. The threshold and pre-adjustment amount range are the same as described above.
[0117] After all the mixing piles are completed, the laitance on the pile top is removed to the design elevation. A 1.2m×0.8m C30 cap beam is poured to form a whole. After the core sample strength meets the standard, excavation begins. Excavation is carried out in layers, with support first and then excavation. Each layer is no more than 2m and each excavation is no more than 30m. Each layer is supported by 609mm diameter and 16mm wall thickness steel pipes, with a pre-applied axial force of 200kN. Double 45C I-beam walers are also erected. The walers are fixed and tightened to fit the flanges by using pre-drilled holes, hooks, wire ropes and turnbuckles. No welding is required in the pit throughout the process.
[0118] After excavating to a depth of -13.2m from the foundation for inspection, the main structure of the temporary storage pool was constructed. Subsequently, symmetrical layered backfilling with 2:8 lime-soil was carried out, and each support was removed layer by layer. The steel sections were extracted in sections using a hydraulic pile extractor, and the gaps were sealed by pressure grouting with a water-cement ratio of 0.5 and a pressure of 0.5MPa to 1.0MPa.
[0119] After backfilling the deep foundation pit and recycling the steel sections, the mixing piles at the junction were broken down to -7.6m from the bottom of the shallow foundation pit. Subsequently, three-axis cement-soil mixing piles with a diameter of 850mm and a spacing of 600mm were constructed in the shallow foundation pit. H-beams were arranged in an alternating pattern and C30 cap beams of the same specification were poured.
[0120] The shallow foundation pit dewatering operation continued. When the excavation reached the design elevation of the anchor cable, the holes were drilled at a 20° inclination angle with a diameter of 150mm. The anchor cable consisted of two 15.2mm diameter, 1860MPa steel strands with a free section of 6m and an anchoring section of 10m. The water-cement ratio was 0.5 to 0.55, and the pressure was 0.5MPa to 1.0MPa. After the strength reached 15MPa, the grouting was staged and locked at 90kN. After acceptance, the excavation was carried out to the regulating pool -7.6m and the AO biochemical pool -6.1m, and the main structure of the shallow water pool was constructed.
[0121] After the shallow pool structure and waterproofing are completed, symmetrical backfilling with 2:8 lime-soil is carried out, the shallow foundation pit steel is removed and the pile holes are filled with grout. After the top slab is covered with soil, the well is sealed in a unified manner: a crushed stone filter layer is filled to 2m below the wellhead, segmented pressure grouting is carried out, and a micro-expansion concrete of not less than 500mm is poured at the wellhead.
[0122] During the construction period, approximately 80 monitoring points were set up to monitor the horizontal displacement of the slope top, vertical displacement, deep inclination, axial force of steel supports, anchor cable tension, groundwater level, and settlement of surrounding buildings. During the excavation of the primary foundation pit, monitoring was conducted twice daily. When the warning value was reached at 80%, the frequency was increased to four times. Work was suspended and reinforcement was carried out when the warning value was reached. Throughout the entire process, the maximum horizontal displacement of the foundation pit did not exceed 25 mm, and the settlement of surrounding buildings did not exceed 5 mm, all of which met the design requirements.
[0123] In this invention, the aforementioned steel guide construction system is used. The steel is inserted and straightened by the guide device. The waler is fixed by pre-drilled holes, hooks, wire ropes and turnbuckles, without the need for welding inside the pit.
[0124] To verify the technical effectiveness of this invention compared to traditional construction methods, a set of comparative test data based on the expansion project of a certain reclaimed water plant is provided. The test section is a multi-layered stratum with a shallow groundwater level, consistent with the geological conditions of the aforementioned application example. This set of comparisons focuses on the steel section insertion process. The steel section insertion obstruction rate described below is calculated as the ratio of the number of obstruction events to the total number of inserted sections.
[0125] The comparative project adopted the traditional manual experience-based control method. Before insertion, the grouting parameters such as water-cement ratio and grouting pressure were determined once according to the mix design. During insertion, the construction personnel judged abnormalities by touch and resistance gauge readings. The machine was only stopped when the side resistance reached the alarm threshold or the steel section was blocked. Adjustments were made based on experience or the cement-soil was allowed to be temporarily remolded. This comparative project completed the insertion of 400 steel sections in the first section of the same project.
[0126] This invention employs the phased construction method and real-time ratio closed-loop control method of this invention, namely, real-time calculation of the ratio R(t)=F(t) / v(t) and its first and second order rates of change, identification of precursors before threshold exceedance and determination of pre-adjustment amount according to instantaneous deviation rate, and monitoring of first order rate of change after pre-adjustment to lock or cancel; 500 steel sections are inserted into the project under the condition that the other conditions are consistent with the comparative example in the adjacent second section of the same project.
[0127] The comparison results show that the incidence of steel section insertion obstruction decreased from 12.5% in the comparative example to 1.8% in the present invention, a decrease of 85.6%. Predictive control before threshold breakthrough can reduce steel section insertion obstruction. The first insertion success rate increased from 82% to 98%, an increase of 16 percentage points.
[0128] The average insertion time per grouting tube was reduced from 45 minutes to 32 minutes, a reduction of approximately 28.9%. Predictive pre-adjustment made the grouting parameters more closely match the formation, reducing repeated stirring and downtime caused by insertion obstruction, and the cement admixture was stably controlled above the design requirements.
[0129] The walers are fixed with pre-drilled holes, hooks, wire ropes, and turnbuckles. The positioning deviation is controlled from ±8mm to within ±2mm, and no welding inside the pit is required throughout the process.
[0130] The above effects are achieved through the combined efforts of normalized ratio monitoring, rate of change precursor identification, and closed-loop quantitative pre-adjustment, which solves the problems of poor controllability of steel section insertion, low success rate, low efficiency, and material waste.
[0131] In addition, the above-mentioned project adopted a phased construction approach, starting with the deep pits and then moving to the shallow pits. The excavation of the deep and shallow foundation pits was staggered. The construction of the shallow foundation pit retaining structure only began after the deep foundation pits were backfilled and the steel sections were recovered. No sudden displacement occurred during the entire construction process due to soil disturbance between the deep and shallow foundation pits. The deformation of the retaining structure and the settlement of the surrounding buildings were all controlled within the design requirements, which verified the isolation effect of phased construction on the mutual disturbance of adjacent foundation pits.
[0132] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.
Claims
1. A method for the construction of a SMW pile in zones in steps, characterized in that, Includes the following steps: Step 1: Divide the construction site into deep foundation pit area and shallow foundation pit area, and set up dewatering wells for pre-dewatering; Step 2: Construct SMW triaxial cement-soil mixing piles in the deep foundation pit area and insert H-beams; during the insertion process, collect the side resistance F(t) and sinking velocity v(t) in real time, calculate the ratio R(t) = F(t) / v(t), and calibrate the soil layer baseline ratio R0 = F0 / v0 by comparing the measured average side resistance F0 and the measured average sinking velocity v0 of the initial segment of each soil layer. When the second-order rate of change of R(t) with depth crosses zero and is confirmed as the soil layer interface by comparing the data segments before and after it, recalibrate R0 according to the measured value of the initial segment of the new soil layer. Step 3: Calculate the first and second order rates of change of R(t) in real time. When multiple consecutive data points of R(t) move in the same direction and the change increases successively, or when the absolute value of the second order rate of change continuously exceeds the threshold, it is judged as a precursor. Before R(t) exceeds the threshold, adjust the grouting parameters to make R(t) converge towards the soil baseline ratio: when R(t) is greater than R0, increase the water-cement ratio or increase the grouting pressure; when R(t) is less than R0, decrease the water-cement ratio or decrease the grouting pressure. Determine the pre-adjustment amount according to the interval to which the instantaneous deviation rate of R(t) relative to R0 belongs. After pre-adjustment, monitor the first order rate of change. If it approaches zero, lock it; if it reverses, cancel and restore the original parameters. Step 4: Construct SMW three-axis cement-soil mixing piles in the shallow foundation pit area and insert H-beams.
2. The method of claim 1, wherein, In step three, the change sequence of the first-order rate of change is monitored after pre-adjustment; when multiple consecutive data points in the sequence have the same sign and their absolute values decrease successively, it is determined that the first-order rate of change is in a state of monotonically decaying and approaching zero; the number of data points required for the first-order rate of change to reach zero is predicted based on the current decay rate; when the number of data points required to reach zero is less than the preset threshold, a locking command is issued in advance; if the predicted first-order rate of change does not decay to zero as expected, the locking is canceled and the original parameters are restored.
3. The method of claim 2, wherein, In step three, if the number of data points required to predict the first-order rate of change to reach zero is less than a preset threshold, it is further determined whether the absolute value of the current first-order rate of change has dropped below a preset steady-state threshold. A locking command is issued only when the number of data points required to predict zero is less than the preset threshold and the absolute value of the current first-order rate of change has dropped below the preset steady-state threshold. When the absolute value of the first-order rate of change drops below the steady-state threshold, a locking countdown is started. If the first-order rate of change does not increase in the opposite direction during the countdown, the lock is activated when the countdown ends. If the first-order rate of change increases in the opposite direction during the countdown, the countdown is terminated and the original parameters are restored.
4. The method of claim 1, wherein, Also includes: After removing the H-beams in the shallow foundation pit area and filling the pile holes with grout, the well is sealed; during the removal of the H-beams, the pile extraction force is monitored in real time and the pile extraction force-displacement curve is plotted. Calculate the deviation value and the rate of change of the measured curve and the theoretical curve at each depth; when the deviation value exceeds the first preset value and the rate of change of the deviation is positive, activate the first-level auxiliary measures to inject high-pressure water or bentonite slurry into the contact surface between the steel section and the cement-soil. When the deviation rate continues to increase beyond the preset rate, the first-level measures are skipped and the second-level auxiliary measures are directly initiated, namely, high-frequency micro-vibration is applied to assist in the removal. When applying high-frequency micro-vibration, the optimal frequency of the historical steel section most similar to the current steel section working condition is retrieved from the removal resistance database, and a local frequency sweep is performed to both sides starting from the optimal frequency. If there is no working condition that meets the similarity requirement in the database, a hill climbing search is initiated simultaneously from multiple different starting frequencies. Each search thread independently records the minimum value of the pile removal force and the corresponding frequency it finds. The pile removal forces of all threads are compared and the frequency corresponding to the minimum value is selected as the initial search starting point. During the frequency sweeping process, the pile extraction force signal and vibration acceleration signal are collected in real time and synchronously. The cross power spectrum of the two signals is calculated, and the frequency corresponding to the peak value of the cross power spectrum is extracted as the effective main frequency. The frequency sweeping optimization is performed with the effective main frequency as the center and the preset narrow band range as the frequency sweeping interval. The frequency with the minimum pile extraction force is updated as the optimal frequency of the current steel section and stored in the extraction resistance database. When the deviation value exceeds the second preset value and the deviation change rate is still positive, the third-level auxiliary measures are activated, namely, suspending the removal and performing local segmented removal.
5. The method of claim 4, wherein, When removing H-beams, the system retrieves historical steel sections from the removal resistance database that are most similar to the current steel section's working condition and calculates the confidence level of the retrieval results. When the confidence level is higher than a preset high threshold, the system directly uses the retrieved optimal frequency as the starting frequency for local narrowband frequency sweeping. When the confidence level is lower than a preset low threshold, the system simultaneously initiates multi-starting-point parallel hill climbing search from multiple different starting frequencies and synchronously calculates the cross-power spectrum of the pile extraction force signal and the vibration acceleration signal to extract the effective dominant frequency. When the confidence level is between the preset high threshold and the preset low threshold, the retrieved optimal frequency is used as one of the starting frequencies for multi-starting-point parallel hill climbing and is searched in parallel with other preset starting frequencies.
6. The method of claim 1, wherein, In step two or four, the second-order rate of change of the real-time ratio R(t) with depth is calculated; when the second-order rate of change changes from positive to negative or from negative to positive, the position is identified as a candidate zero crossing point; after the candidate zero crossing point is detected, the R(t) data in the first depth window before the zero crossing point and the second depth window after the zero crossing point are extracted, and the energy value of the data in the two windows is calculated. When the ratio of the energy value of the window after the zero-crossing to the energy value of the window before the zero-crossing exceeds a preset threshold, the candidate zero-crossing position is confirmed as the soil layer interface, triggering the recalibration of the baseline ratio R0.
7. The method of claim 6, wherein, In step two or four, after a candidate zero-crossing point is detected, R(t) data before and after the zero-crossing point are extracted using multiple different window lengths, with the candidate zero-crossing point as the center. The energy value ratio of the window after the zero-crossing point to the window before the zero-crossing point is calculated for each window length. The maximum value is selected from the energy ratios corresponding to all window lengths. When the maximum value exceeds a preset threshold, the candidate zero-crossing point is confirmed as the soil layer interface, and the baseline ratio R0 is recalibrated.
8. The SMW method for phased construction of piles as described in claim 1, characterized in that, In step two or four, the position of the zero-crossing point of the second-order rate of change is the candidate zero-crossing point; after the candidate zero-crossing point is detected, the R(t) waveform in the first depth window before the zero-crossing point is taken as the reference waveform, and the R(t) waveform in the second depth window after the zero-crossing point is taken as the comparison waveform. Calculate the shape similarity between the reference waveform and the comparison waveform; when the shape similarity is lower than the preset threshold, confirm the candidate zero-crossing point as the soil layer interface and trigger the recalibration of the baseline ratio R0.
9. The SMW method for phased construction of piles as described in claim 1, characterized in that, In step three, when determining the pre-adjustment amount of grouting parameters according to the instantaneous deviation rate, a deviation rate reference value k is set ref , and the pre-adjustment amount is determined according to the absolute value |k| of the instantaneous deviation rate: when |k| is less than 0.5k ref , the pre-adjustment amount is 30% to 50% of the normal adjustment amount; when |k| is between 0.5k ref and 1.5k ref , the pre-adjustment amount is 60% to 80% of the normal adjustment amount; and when |k| is greater than 1.5k ref , the pre-adjustment amount is 100% to 120% of the normal adjustment amount.
10. The SMW method for segmented and phased pile construction as described in claim 1, characterized in that, In step three, before determining the pre-adjustment amount of grouting parameters, the current soil layer is first calibrated for response sensitivity: when the pre-adjustment condition is first met in the current soil layer, only the baseline detection adjustment amount Δ0 is applied, and Δ0 takes one-third of the lower limit of the preset adjustment amount; the measured ratio sequence of the continuous m depth points after application is collected, and the predicted ratio sequence is obtained by extrapolating the ratio sequence of the n depth points before application by a quadratic polynomial; the average value ΔR of the difference between the measured and predicted sequences at the m depth points is calculated p As a detection response, ΔR p is divided by the baseline ratio R0 to obtain the normalized response sensitivity S of the soil layer; In the subsequent pre-adjustment, the deviation AR of the current measured ratio relative to the baseline ratio R0 c Divide by the sensitivity S to obtain the target adjustment amount, and take the smaller one of the target adjustment amount and the preset adjustment amount upper limit value as the actual pre-adjustment amount; When S is lower than the preset sensitivity lower limit, the current soil layer is determined to be a layer where the grouting parameter adjustment has failed. The current grouting parameters are kept unchanged, and the single insertion stroke is shortened to the preset segment length. After each segment insertion is completed, the insertion is paused for a preset time before continuing. The ratio of sensitivity S to the baseline of the soil layer R0 is stored in the insertion control database. For subsequent pile construction in the same soil layer, S is directly called and no more benchmark detection adjustment amount Δ0 is applied.
Citation Information
Patent Citations
System for automatically monitoring stress of interpolation H-shaped steel by soil mixing wall (SMW) construction method
CN102539024A
Synchronous construction method for adjacent foundation pits of unequal height
CN107313431B
Support Structure and Water-stopping Methods for Foundation Pit under High-Speed Railway Bridges When Construction Space is Limited
CN110055973B
Synchronous construction method for deep and shallow adjacent foundation pits under silt soil geologic condition
CN118600997A
A method and system for closed-loop control of SMW wall construction in water-rich sandy soft soil layer
CN122812241A
