Pile foundation construction quality monitoring method based on multi-data fusion

Abstract

The invention relates to the technical field of pile foundation construction, in particular to a pile foundation construction quality monitoring method based on multi-data fusion, which comprises the following steps of: establishing a local coordinate system by taking a pile foundation construction position as an original point, and collecting pile foundation position coordinates of each pile foundation during construction; according to the value conditions of the same pile foundation at multiple time points, construction deviation corresponding to the position coordinates of the pile foundation is calculated, and an abnormal area corresponding to the construction deviation is calculated by taking a construction abnormal point as a checking object; associating the abnormal region of the construction deviation with a strain value, performing spatial integration by using a strain gradient, and configuring a construction feature cluster under spatial evolution; performing multi-level condition judgment on each pile body by utilizing the construction characteristic cluster, and outputting a quality description label corresponding to each pile foundation; and on the basis of the quality description labels of all the pile foundations, the construction positions of the pile foundations are reset, and parameters of next-time construction are adjusted according to the hierarchical conditions related to the quality description labels. Accuracy and efficiency of pile foundation construction monitoring are achieved.

Description

Technical Field

[0001] This invention relates to the field of pile foundation construction technology, specifically a pile foundation construction quality monitoring method based on multi-data fusion. Background Technology

[0002] In building construction, pile foundations serve as crucial foundation structures, their bearing capacity and stability directly impacting the safety and durability of buildings. However, due to complex construction environments, diverse pile foundation types, and external influences, defects such as cracks, voids, and tilting may occur during pile foundation construction. Addressing these defects has evolved from traditional manual sampling and single-index testing to multi-parameter simultaneous monitoring.

[0003] For example, Chinese Patent Publication No. CN120739153A discloses a method and system for monitoring and controlling pile foundation construction, relating to the field of pile foundation construction. The method includes: acquiring operating parameters of construction equipment at multiple moments during the monitoring period; determining the safety operating coefficient of construction equipment based on the operating parameters; acquiring construction environment monitoring data of the construction area at multiple moments during the monitoring period; processing the construction environment monitoring data according to a trained construction environment risk prediction model to determine the construction environment risk coefficient; acquiring pile foundation performance monitoring data at multiple moments during the monitoring period; determining the pile foundation safety coefficient based on the pile foundation performance monitoring data; and generating a monitoring and control report based on the construction equipment safety operating coefficient, the construction environment risk coefficient, and the pile foundation safety coefficient.

[0004] For example, Chinese Patent Publication No. CN117966729A discloses a construction method for stiffened composite piles, including obtaining the depth of the soft soil foundation at the construction site and determining whether it is greater than a preset depth; if so, using graded surcharge preloading to treat the foundation; if not, using layered replacement treatment to treat the foundation; leveling the construction site and removing obstacles, performing positioning and layout, and embedding steel casings; performing high-pressure jet grouting cement mixing pile construction, collecting the construction parameters of the cement mixing piles in real time, and adaptively adjusting the operating parameters of the mixing pile machine in combination with a hardness evaluation model; performing prestressed pipe pile construction, and conducting bearing capacity tests after completion, including calculating the first bearing capacity of the stiffened composite pile, if it is less than a preset threshold, performing a pile foundation offset test, if offset occurs, using a genetic algorithm to reset the pile foundation, calculating the second bearing capacity, if it is still less than the preset threshold, adjusting the structure of the mixing pile machine to increase the diameter of the cement mixing pile.

[0005] Existing technologies acquire parameters from three aspects: equipment operation, environmental monitoring, and pile foundation performance. These parameters are then used to train risk prediction models for monitoring and control of pile foundation construction. Alternatively, soil hardness is assessed, and the bearing capacity at the corresponding hardness is used to interpret the radius of the mixing piles in the pile foundation construction, thereby adjusting the construction method. However, existing technologies tend to predict and handle pile foundation risks under a single process, easily neglecting the impact of corresponding deviations in pile foundation construction. This makes it difficult to correlate deviations during construction with the strain after the pile foundation is shaped, leading to misidentification of cumulative deviations during pile foundation construction and a fragmented, point-based approach to handling abnormal situations. Consequently, construction efficiency and the scope of construction monitoring are reduced. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a pile foundation construction quality monitoring method based on multi-data fusion, comprising: establishing a local coordinate system with the pile foundation construction location as the origin, collecting the pile foundation location coordinates during construction, and simultaneously recording the axial and horizontal strain values ​​of the pile foundation after construction, forming a time-series dataset of pile foundation construction.

[0007] Based on the values ​​of the same pile foundation at multiple time points, scenario interaction analysis is performed on the pile foundation location coordinates to calculate the construction deviation corresponding to the pile foundation location coordinates. Using construction anomaly points as the verification objects, the abnormal areas corresponding to the construction deviations are deduced.

[0008] By associating abnormal areas of construction deviations with strain values ​​and spatially integrating them using strain gradients, construction feature clusters under spatial evolution are configured.

[0009] By utilizing construction feature clusters, multi-level condition judgments are performed on each pile body, and the quality description labels corresponding to each pile foundation are output based on the content that meets the conditions at each level.

[0010] Based on the quality description labels of each pile foundation, the construction position of the pile foundation is reset, and the parameters for the next construction are adjusted according to the hierarchical conditions involved in the quality description labels.

[0011] The beneficial effects of this invention are as follows: First, this invention establishes a local coordinate system with the pile foundation construction location as the origin, and simultaneously collects the pile foundation construction location coordinates and axial / horizontal strain values ​​after construction to form a pile foundation construction time-series dataset. By analyzing displacement changes under different numbers of monitoring points, the optimal number of monitoring points is selected to ensure the integrity of the time-series dataset, while avoiding monitoring point redundancy and improving the efficiency of current data acquisition.

[0012] Second, this invention performs scene interaction analysis on the position coordinates of the same pile foundation at multiple time points, calculates construction deviations, and uses construction anomaly points as verification objects to deduce the abnormal areas corresponding to the construction deviations. Then, the position coordinates are converted into attitude angles, and the construction deviation is calculated using the two-dimensional deviation of position coordinates and attitude angles. Combined with the average construction progress, similarity is obtained by calculating the construction deviations under the same or different time sequences, ensuring that the construction deviations fit the temporal characteristics of the construction scenario. Furthermore, by analyzing the trend and maximum construction deviation, abnormal areas are identified, achieving spatial connectivity judgment from abnormal points to abnormal areas. This improves the data's fit with the scenario and the accuracy of positioning under the current processing, providing a clear spatial range for subsequent quality traceability.

[0013] Third, this invention associates abnormal areas of construction deviations with strain values ​​and integrates them spatially based on strain gradients. First, it verifies the strain gradient to determine strain anomalies. Then, it combines construction deviations and strain gradients within the same strain gradient interval into spatial feature units, and verifies the axial / horizontal correlation in real time. This makes the correlation between strain anomalies and construction deviation anomalies more spatially distributed, improving the accuracy of anomaly analysis. Simultaneously, it uses cosine similarity to match and score spatial feature units of different anomaly types, ensuring the visibility and interpretability of the data format of each anomaly combination in global coordinates.

[0014] Fourth, this invention utilizes construction feature clusters to perform multi-level condition judgments on each pile body. Based on the compliance content under each level of condition, it outputs the corresponding quality description label for each pile foundation, making the relevant error problems during construction more interpretable and standardized, providing data guidance for subsequent parameter adjustments, and ultimately avoiding the one-sidedness and ambiguity of a single standard judgment, thus improving the accuracy and efficiency of construction quality analysis. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0016] Figure 1 This is a flowchart illustrating a pile foundation construction quality monitoring method based on multi-data fusion.

[0017] Figure 2 This is a flowchart illustrating step 2 of a pile foundation construction quality monitoring method based on multi-data fusion.

[0018] Figure 3 This is a flowchart illustrating step 3 of a pile foundation construction quality monitoring method based on multi-data fusion.

[0019] Figure 4 This is a flowchart illustrating step 4 of a pile foundation construction quality monitoring method based on multi-data fusion. Detailed Implementation

[0020] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0021] See Figure 1 A method for monitoring the construction quality of pile foundations based on multi-data fusion includes: S1, establishing a local coordinate system with the pile foundation construction location as the origin, collecting the pile foundation location coordinates during construction, and simultaneously recording the axial and horizontal strain values ​​of the pile foundations after construction to form a time-series dataset of pile foundation construction.

[0022] S2, based on the value of the same pile foundation at multiple time points, performs scene interaction analysis on the pile foundation position coordinates, calculates the construction deviation corresponding to the pile foundation position coordinates, and uses the construction anomaly points as the verification objects to infer the abnormal area corresponding to the construction deviation.

[0023] S3 associates abnormal areas of construction deviations with strain values, integrates them spatially using strain gradients, and configures construction feature clusters under spatial evolution.

[0024] S4 utilizes construction feature clusters to perform multi-level condition judgments on each pile body, and outputs the corresponding quality description labels for each pile foundation based on the content that meets the conditions at each level.

[0025] S5, based on the quality description labels of each pile foundation, resets the construction position of the pile foundation and adjusts the parameters for the next construction according to the hierarchical conditions involved in the quality description labels.

[0026] In the current scenario, a high-precision total station / BeiDou positioning module can be used to calibrate the local coordinate system to ensure that each pile foundation can be mapped to the global coordinate system of the project. Then, with the design and construction center point of the pile foundation as the origin, the Z-axis along the design axis of the pile foundation, and the X and Y axes as horizontal planes, the position and attitude of the pile foundation construction are transformed into three-dimensional coordinates under the local coordinate system. By deploying fiber optic sensors or other sensors that can measure its coordinates at equal intervals along the Z-axis, each sensor can collect the strain values ​​of the pile body on the X, Y (horizontal) and Z (axial) axes.

[0027] If the type of pile foundation is taken into account, sensors can be installed on the pile body for precast piles, while sensors are tied to the reinforcing cage for cast-in-place piles to determine whether there is horizontal or axial strain after grouting.

[0028] The coordinates of the pile foundation collected at this time can reflect the verticality, position, diameter and length of the pile foundation during construction in real time; as well as the coordinate distribution under specific construction conditions such as penetration, driving speed and hole depth.

[0029] The strain value is used to reflect the stress, strain, displacement, etc. of the pile body. It can help determine the characteristics of the pile body integrity and stability, such as the stress of the borehole wall and the density of the concrete at each location after the pile foundation is constructed, so as to verify whether the pile foundation can be used normally after construction, and synchronize the measured strain value to the corresponding pile foundation construction location.

[0030] When forming the time-series dataset of pile foundation construction in step S1, the implementation method includes: S11, obtaining the pile foundation type at the current construction time. The pile foundation type can be divided into various types such as precast / cast-in-place. Using the pile foundation position coordinates corresponding to each pile foundation type, the displacement value of the pile foundation in each direction is calculated.

[0031] S12, using the displacement values ​​of the pile foundation in each direction as a guide, simultaneously verify the number of monitoring points when the time series dataset was acquired, and determine the optimal number of monitoring points based on the displacement change trends plotted under different numbers of monitoring points.

[0032] S13 uses the time series dataset corresponding to the optimal number of monitoring points as the output data.

[0033] At this point, displacement values ​​that change due to changes in the pile foundation's position coordinates during construction are recorded as displacement values ​​in the X, Y, and Z axes. This is used to explain whether the displacement values ​​in these three directions change after the pile foundation is driven into the soil or grouted. Then, based on the trend of these displacement values, the displacement values ​​are plotted in chronological order, with the timestamp as the horizontal axis and the displacement value as the vertical axis, forming a trend graph of the displacement values ​​in the three directions. The graph is then checked for discontinuities, abnormal fluctuations, and obvious curve overlaps. Displacement change trends obtained from different monitoring point deployments are used to identify the number of monitoring points corresponding to those without discontinuities, overlaps, or abnormal distortions. These numbers are considered the optimal number of monitoring points for the current scenario to monitor minor displacement changes occurring during construction for the corresponding pile foundation type.

[0034] When determining the optimal monitoring point, the number of monitoring points can be traversed. Based on the data input at each monitoring point, it can be determined whether there are data collection overlaps or data gaps at each monitoring point. If such gaps exist, the number of monitoring points can be adjusted at the corresponding locations until the acquired data shows complete data collection at the corresponding locations.

[0035] In one embodiment of the present invention, in step S2, the positional deviation between the pile foundation design and construction and the real-time construction of the pile foundation is recorded based on the real-time coordinates during the pile foundation construction process, and the corresponding construction deviation under real-time construction is determined to determine the processing part required for subsequent quality verification; for example, after the pile foundation construction is completed, its construction deviation and residual strain value are statistically analyzed to indicate whether its construction quality meets the requirements, and then the final output instruction is determined.

[0036] The implementation method of scene interaction analysis of pile foundation position coordinates in step S2 includes: converting the pile foundation position coordinates into attitude angles based on the construction scene corresponding to the current time series dataset. The attitude angles are used to check the verticality of the pile foundation during construction to ensure that there is no tilting construction during pile foundation construction. For example, when the bored pile is formed, its verticality deviation needs to be less than 1%. If the deviation is greater than 1%, an early warning will be triggered and the construction at the corresponding position will be adjusted.

[0037] The deviations corresponding to the pile foundation position coordinates and attitude angles are considered as the current construction deviations. The average construction progress corresponding to these deviations is then checked. Construction deviations are defined as a combination of pile horizontal deviation (deviation on the horizontal plane containing the XY axes) + verticality deviation rate + pile top elevation deviation (deviation on the Z-axis). The average construction progress is expressed as the ratio of the current depth of each pile to its design depth, used to verify the progress of pile driving. The deviations checked here are correlated with the construction progress by comparing the deviation of a specific pile with the average deviation of the group. Then, similarity calculations are performed based on the correlation of local construction in a single time series and the cumulative correlation across multiple time series to uncover the forms of correlation appearing in the current construction scenario.

[0038] The construction process of each pile foundation is checked using the average construction progress. The similarity of construction deviations of adjacent pile foundations in the same time series and the similarity of cumulative construction deviations in different time series are calculated respectively. The similarity in the same time series is obtained by using cosine similarity to determine the similarity between adjacent pile foundations. The similarity of adjacent pile foundations in different time series can be calculated by dynamic time warping (DTW). By finding the Euclidean distance of all point pairs in the time series of construction deviations of two pile foundations and minimizing the sum of the distances of all point pairs, the cumulative amount is calculated. The smaller the distance, the higher the similarity. This is used to obtain the interaction content of two adjacent pile foundations in the corresponding scenario.

[0039] When the similarity value is maximized in different time series and in the same time series, the corresponding data is regarded as the result of the interaction analysis of the current scene.

[0040] As for the maximum similarity value, it essentially involves determining which of the adjacent pile foundations has a higher similarity value in the same time series and different time series, in order to explain the deviation pattern between adjacent piles. For example, when the similarity value is the highest in the same time series, it will point to local common factors, such as geological anomalies (local weak layers), synchronous errors in construction operations (incorrect setting of pile driver parameters), soil squeezing effect (that is, when driving piles, the center pile is not driven first to reduce soil squeezing, or the amount of soil squeezing is too large after driving piles at the corresponding location). These problems mostly reflect similarity at specific points in time, and it is necessary to determine the construction deviation in the current scenario based on this part of the similarity data.

[0041] When the similarity is greatest at different time series, it can point to problems in process reproducibility, such as periodic equipment failures like drill bit wear, operator habitual deviations, and repeated entry into soil layers with the same geological anomalies. These similar data can represent the main content of the current scenario analysis, and the changing trend of construction deviations under the corresponding scenario can be inferred based on these data.

[0042] When estimating abnormal areas of construction deviations, the data from the current scenario interactive analysis will be used to statistically analyze the range of values ​​for changes under different time series, as well as the corresponding slope values. This data will then be projected from local coordinates to global coordinates to establish a correlation between local and global changes.

[0043] like Figure 2 As shown, when calculating the abnormal area corresponding to the construction deviation in step S2, the implementation method includes: S21, according to the construction deviation output by each pile foundation during scene interaction analysis, calculate the construction deviation change curve at each pile foundation; the construction deviation change curve will focus on the construction deviation generated at each depth during the pile driving process, that is, with the actual depth of the pile foundation into the soil as the horizontal axis and the construction deviation as the vertical axis, record the deviation generated during the current construction; such as how many meters of construction deviation are generated at each depth of descent, and how many millimeters of construction deviation are generated, to explain the slope value of the curve.

[0044] S22, based on the trend of each construction deviation change curve, synchronize the local construction deviation of each pile foundation to the global coordinate system of the construction scene; here, the trend can be defined in five ways: linear increase, linear decrease, small fluctuation, significant change and basically stable, to define the current construction deviation.

[0045] The criteria for determining each trend direction will be set based on its slope value and coefficient of variation (the ratio of standard deviation to mean). For example, "basically stable" means that the absolute value of the slope value is ≤0.5 and the coefficient of variation is ≤0.2, indicating that the deviation does not change significantly with depth, and the curve is close to a horizontal straight line; "small fluctuations" means that the absolute value of the slope value is located in (0.5, 2] and the coefficient of variation is located in [0.2, 0.5], indicating that there is no significant increase or decrease overall, but there are small fluctuations in some areas; "linearly increasing" means that the slope value is >2 and the coefficient of variation is ≤0.3, indicating that the deviation increases uniformly with depth, the curve is an upward curve, and the dispersion is low; "linearly decreasing" means that the slope value is <-2 and the coefficient of variation is ≤0.3. For values ​​≤0.3, the deviation decreases uniformly with increasing depth, forming a downward-sloping curve with low dispersion. Significant abrupt changes will be addressed by processing sudden increases and decreases, such as a sudden increase or decrease of 50% of the allowable deviation threshold, an absolute difference in slope between adjacent points greater than 5, or a coefficient of variation greater than 0.5 and a deviation value greater than 80% of the allowable deviation threshold. In these cases, data meeting any one of these three conditions will be considered as significant abrupt changes, providing a data basis for current construction deviation identification and processing. The remaining unspecified value ranges represent transitional areas of linear fluctuation during construction and will be directly input into subsequent regional anomaly detection processing.

[0046] The above explanations based on trend trends are illustrative. Depending on the soil conditions and location of the current construction site, as well as the deviations in the X, Y, and Z axes, different values ​​will be used to determine the current deviation of the pile foundation after construction to different depths.

[0047] S23 synchronously correlates the depth and trend of each pile foundation in the global coordinate system. Using the maximum construction deviation at the same depth and the maximum construction deviation under different trends as the verification content, it performs regional anomaly judgment on each pile foundation and obtains the output anomaly area. Anomalies occurring at the same depth or trend will be regarded as individual anomaly points. By using the distribution location of individual anomaly points in the global coordinate system, they will be grouped into a contiguous or adjacent spatial region, and then the anomaly region will be marked.

[0048] In step S23, when determining regional anomalies for each pile foundation, the implementation method includes: based on the construction deviation at any depth, using the maximum construction deviation as a benchmark, sorting all pile foundations from largest to smallest to generate a sorting list; starting from the first position of the sorting list, verifying whether the construction deviation of each pile foundation exceeds the allowable deviation threshold, and configuring the construction anomaly point corresponding to the construction deviation; when the current maximum construction deviation is greater than the allowable deviation threshold, the corresponding position is regarded as a construction anomaly point, and the allowable deviation threshold can be obtained according to the construction specifications of the current construction scenario.

[0049] Based on construction deviations under any trend, construction anomalies corresponding to construction deviations are configured according to the range of values ​​of construction deviations and average construction deviations under the same trend. When analyzing the trend, data with significant abrupt changes are mainly extracted. At the same time, it is also necessary to consider construction anomalies that may exist in other trends. At this time, the average construction deviation will be introduced, and the part exceeding the average construction deviation will also be regarded as construction anomalies here.

[0050] Based on the planar location, depth, and station number of each construction anomaly point, the corresponding anomaly area is defined. Each anomaly point represents an anomalous pile, and the anomaly area can be categorized according to the value of the anomalous pile, such as the anomaly area corresponding to tilted piles, the anomaly area for excessive settlement, etc. Furthermore, based on the concentrated distribution of construction anomalies, the coordinate range of the concentrated pile foundation anomalies, as well as the coordinate range of the scattered or strip-shaped distribution, is described to define the current anomaly area. The output anomaly area is also further categorized by its location, such as being at the construction edge or construction center, to indicate the current anomaly area.

[0051] In one embodiment of the present invention, in step S3, based on the depth value during construction, the abnormal area marked by the construction deviation is synchronized to each pile foundation, and according to the construction deviation and strain gradient that appear at the corresponding position of the pile foundation, it is checked whether the abnormality of the construction deviation leads to the abnormal synchronization of the strain gradient, thereby explaining the impact of the slight deviation during construction on the subsequent strain in the scenario where construction deviation occurs.

[0052] like Figure 3 As shown, when associating the abnormal area of ​​construction deviation with the strain value in step S3, the implementation method includes: S31, based on the location of each pile foundation, simultaneously check the strain gradient of each pile foundation, and determine the strain abnormal point corresponding to each pile foundation.

[0053] When the current input data is time series data, the strain gradient can be explained according to the rate of change of strain over time at the same measuring point, or the strain difference between adjacent measuring points can be calculated along the axial or circumferential direction of the pile foundation. The relative abrupt change in the strain gradient can be explained in a spatial or temporal correlation manner, thereby explaining the co-occurrence of construction deviations and strain gradient anomalies.

[0054] It should be noted that the strain anomaly mentioned below is an abbreviation for the strain gradient anomaly of the current scene.

[0055] When determining the strain anomaly points corresponding to each pile foundation, the following methods are used: taking the anomaly area corresponding to the construction deviation as the verification object, and using the rate of change of strain value over time to determine whether there is an anomaly in the strain gradient at the current location.

[0056] When there is an anomaly in the strain gradient at the current location, the strain gradient in the corresponding direction is checked simultaneously with the anomaly in the axial and horizontal directions corresponding to the construction deviation point, and the correlation between the strain anomaly point and the axial and horizontal directions is marked.

[0057] If there is no anomaly in the strain gradient at the current location, check the strain gradient of the remaining area excluding the abnormal area. If an anomaly is confirmed, mark the strain anomaly point.

[0058] At this point, the abnormal gradient within the abnormal area is first checked. Based on areas without construction deviations, the strain change rate over time at the same measuring point is extracted. This rate is then labeled as the normal strain gradient range in the form of average ± three standard deviations. If the strain gradient at the current location exceeds this range, it is considered an anomaly at that location. For locations with anomalies, the axial and horizontal deviations corresponding to the occurrence of the construction deviation are simultaneously checked. Within the time period of the anomaly, the strain difference between adjacent measuring points is calculated along the pile foundation axis or circumferentially. It should be noted that the strain gradient measured circumferentially is consistent with the horizontal plane direction it represents. To explain the strain patterns observed on the X and Y axes: if strain anomalies exist in both the circumferential and axial directions without directional differences, they are marked as global strain anomalies; if there is only one directional anomaly, it is marked as a circumferential strain anomaly or an axial strain anomaly; if there are anomalies in both directions, and one direction is significantly higher than the other (e.g., if the value in one direction exceeds 50% relative to the overall value, it can be considered the main anomaly direction), and this direction is consistent with the main anomaly direction of the construction deviation, it is marked as a directional strain anomaly; and if the main direction of the strain anomaly is not consistent with the main anomaly direction of the construction deviation, it is marked as a strain anomaly with inconsistent direction.

[0059] Subsequently, in the abnormal areas of construction deviation, some without strain gradient anomalies are not processed. They are only marked by the relevant description of construction deviation. The remaining areas, excluding the abnormal areas, are checked only by the rate of change of strain value over time. When they exceed the normal range, they are decomposed into strain anomalies in the corresponding directions. At this time, only the contents in the axial and horizontal directions are recorded. It is not necessary to mark strain anomalies in the same direction.

[0060] The data recorded in these records will demonstrate the causal relationship between construction deviations and strain anomalies, serving as the data basis for verifying feedback instructions.

[0061] S32, based on the location of strain anomalies and construction anomalies, combines the construction deviations and strain gradients within the same strain gradient range one by one to form spatial feature units.

[0062] Spatial feature units can be categorized into three anomalous combinations: construction deviation anomaly + strain anomaly, strain anomaly + no construction deviation anomaly, and no strain anomaly + construction deviation anomaly. This is used to quantify the coupling effect between geometric deviation and structural response, and through matching processing of different spatial units, the output construction feature cluster is obtained.

[0063] For example, if the strain is abnormal but the construction deviation is small, the materials or maintenance can be checked. If the construction deviation is large and the strain is abnormal, geological re-exploration and process review can be carried out to determine the feedback that needs to be done in the current scenario.

[0064] S33, verify the positional matching between each spatial feature unit, and combine the verified spatial feature units with the pile foundation type, construction scenario and global coordinate system to configure the construction feature cluster under spatial evolution.

[0065] When verifying the positional matching between spatial feature units, the implementation method includes: for any two spatial feature units, converting the data corresponding to strain anomaly points and construction anomaly points into feature vectors, and recording the matching score between each spatial feature unit in the form of cosine similarity.

[0066] For spatial feature elements that simultaneously exhibit strain gradient anomalies and construction deviation anomalies, the matching score between each spatial feature element is calculated and recorded by combining the spatial overlap rate and direction of its location.

[0067] For spatial feature elements with only strain gradient anomalies and only construction deviation anomalies, the matching score between each spatial feature element is calculated and recorded by matching the same anomaly form.

[0068] Select any two spatial feature units corresponding to the maximum matching score as the output data to complete the location matching processing.

[0069] In pile foundations where two types of anomalies coexist, the spatial overlap rate of any two spatial feature units is calculated based on the regions corresponding to strain anomalies and construction deviation anomalies. This represents the concentrated occurrence of anomalies within a specific depth region, emphasizing the distribution of gradient anomalies and construction anomalies occurring simultaneously within similar depth intervals, thereby completing the position matching of spatial feature units. Direction is determined by converting the descriptions within the aforementioned strain anomaly points, such as directional strain anomalies, strain anomalies with inconsistent directions, and global strain anomalies, into feature values ​​to fill the dimension of the current feature vector, thus completing the processing of related feature clusters.

[0070] As for the data under a single anomaly, the strain anomaly and construction deviation anomaly are counted separately, and similar descriptive content is obtained by calculating the value of the feature vector at the corresponding location.

[0071] It should be noted that the calculated cosine similarity is greater than or equal to 0.6; otherwise, the corresponding data is not considered a group of related spatial feature units. The process iterates through the data one by one until all data is matched to obtain spatial feature units with specific feature requirements. These matched spatial feature units will be grouped and described according to pile foundation type, construction scenario, and global coordinate system. Pile foundation type and construction scenario can be understood as the same pile foundation type (e.g., precast pile feature cluster / cast-in-place pile feature cluster) and the same construction scenario (e.g., precast pile hammer-driven cluster, cast-in-place pile concrete pouring cluster). The global coordinate system is described according to the partitions divided during construction, such as the center, edge, and other specific construction locations.

[0072] In one embodiment of the present invention, in step S4, the feature information presented by the construction feature cluster is used as a basis for matching search to determine the content that conforms to the conditions at each level, and then a description label corresponding to the pile foundation is generated.

[0073] like Figure 4 As shown, when performing multi-level condition judgment on each pile body in step S4, the implementation method includes: S41, using the spatial feature similarity of construction feature clusters, construction scene rules and coordinate association as hierarchical conditions, searching for construction feature clusters, and determining the description label of the current construction feature cluster under the corresponding hierarchical conditions.

[0074] S42 connects the description labels under each level of condition according to their level, and uses them as the output quality description labels.

[0075] Spatial feature similarity can be used as the highest level, construction scene rules as the second level, and coordinate association as the third level to associate the current data.

[0076] Spatial feature similarity refers to the cosine similarity of specific spatial feature units and the description content under pile foundation type; construction scenario rules describe the current construction scenario, geological parameters, and construction methods, determining the data content related to the current anomaly. Construction scenario rules are based on a rule base predefined by historical experience or specifications. For example, if the pile foundation is located in a silty soil layer and is constructed using the hammer-driving method, when the construction deviation exceeds 5cm, the rule determines it as a soft soil squeezing effect. In practical applications, a rule base can be established based on engineering specifications or expert experience to match construction feature clusters. As for coordinate association, it verifies the partition location (center / edge), adjacent structures, etc., determines the labels under spatial coordinate association, and uses the label combination description as the quality description label for the current output.

[0077] In one embodiment of the present invention, in step S5, the construction parameters are reset and adjusted according to the quality description label corresponding to the current scene, and the parameters for the next construction are adjusted according to the relevant hierarchical conditions.

[0078] When adjusting the parameters for the next construction in step S5, the implementation method includes: for each quality description label associated with the hierarchical conditions, determining the parameters to be adjusted at different locations based on the construction locations involved in the hierarchical conditions, and taking the part with the largest coverage of the parameters to be adjusted as the parameters currently output.

[0079] By resetting the actual construction location to the design coordinates based on the parameters corresponding to each level of conditions, adjustments are made according to the part with the largest parameter coverage. For example, in abnormal area A, 10 piles exhibit slight X-axis deviation anomalies and a single pile has an abnormal penetration. The parameters to be adjusted for this level of conditions are positioning calibration frequency, pile driving speed, hammer force, and penetration control value. Among these four parameters, positioning calibration frequency and pile driving speed are for the scenario of handling 10 piles, while hammer force and penetration control value are for the scenario of handling a single pile with abnormal penetration. It can be seen that positioning calibration frequency and pile driving speed cover a larger number of piles. At this time, the parameter coverage represents the number of piles corresponding to the parameters to be adjusted. Priority is given to adjusting the parameters with a large range of data as output parameters, and these are reset and adjusted sequentially to determine the parameters for the next construction.

[0080] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A method for monitoring the construction quality of pile foundations based on multi-data fusion, characterized in that, include: A local coordinate system was established with the pile foundation construction location as the origin. The pile foundation location coordinates during construction were collected, and the axial and horizontal strain values ​​of the pile foundation after construction were recorded to form a time series dataset of pile foundation construction. Based on the value of the same pile foundation at multiple time points, the pile foundation location coordinates are analyzed by scene interaction, the construction deviation corresponding to the pile foundation location coordinates is calculated, and the abnormal area corresponding to the construction deviation is deduced by taking the construction abnormal point as the verification object. By associating abnormal areas of construction deviations with strain values ​​and spatially integrating them using strain gradients, construction feature clusters under spatial evolution are configured. By utilizing construction feature clusters, multi-level condition judgments are performed on each pile body, and the quality description labels corresponding to each pile foundation are output based on the content that meets the conditions at each level. Based on the quality description labels of each pile foundation, the construction position of the pile foundation is reset, and the parameters for the next construction are adjusted according to the hierarchical conditions involved in the quality description labels.

2. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 1, characterized in that, When generating a time-series dataset for pile foundation construction, the implementation methods include: Obtain the pile foundation type during the current construction, which can be either precast or cast-in-place. Calculate the displacement value of the pile foundation in each direction using the pile foundation location coordinates corresponding to each pile foundation type. Using the displacement values ​​of the pile foundation in various directions as a guide, the number of monitoring points when the time series dataset was acquired was checked simultaneously, and the optimal number of monitoring points was determined according to the displacement change trend plotted under different numbers of monitoring points. The time series dataset corresponding to the optimal number of monitoring points is used as the output data.

3. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 1, characterized in that, The methods for implementing scene interactive analysis of pile foundation location coordinates include: Based on the construction scenario corresponding to the current time series dataset, the pile foundation location coordinates are converted into attitude angles; The deviations corresponding to the pile foundation position coordinates and attitude angles are regarded as the current construction deviations, and the average construction progress corresponding to the construction deviations is checked. The construction process of each pile foundation is checked by the average construction progress. The similarity of construction deviations of adjacent pile foundations under the same time sequence and the similarity of cumulative construction deviations under different time sequences are calculated respectively. When the similarity value is maximized in different time series and in the same time series, the corresponding data is regarded as the result of the interaction analysis of the current scene.

4. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 1, characterized in that, When calculating the abnormal areas corresponding to construction deviations, the methods include: Calculate the construction deviation change curve for each pile foundation based on the construction deviation output during scene interaction analysis. Based on the trend of each construction deviation change curve, the local construction deviations of each pile foundation are synchronized to the global coordinate system of the construction scene. The depth and trend of each pile foundation in the global coordinate system are synchronously correlated. The maximum construction deviation at the same depth and the maximum construction deviation under different trends are used as the verification content to determine the regional anomalies of each pile foundation and obtain the output anomaly area.

5. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 4, characterized in that, When determining regional anomalies for each pile foundation, the implementation methods include: Based on the construction deviation at any depth, and taking the maximum construction deviation as the benchmark, all pile foundations are sorted from largest to smallest to generate a sorting list. Starting from the first position of the sorting list, the construction deviation of each pile foundation is verified one by one to see if it exceeds the allowable deviation threshold, and the construction anomaly points corresponding to the construction deviation are configured. Based on the construction deviation under any trend, the construction anomaly points corresponding to the construction deviation are configured according to the range of values ​​of the construction deviation and the average construction deviation under the same trend. Based on the plane location, depth and station number of each construction anomaly point, the anomaly area corresponding to each construction anomaly point is defined.

6. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 1, characterized in that, When associating abnormal areas of construction deviations with strain values, the methods include: Based on the location of each pile foundation, the strain gradient of each pile foundation is checked simultaneously to determine the strain anomaly points corresponding to each pile foundation; Based on the location of strain anomalies and construction anomalies, construction deviations and strain gradients within the same strain gradient range are combined one by one to form spatial feature units. Verify the positional matching between each spatial feature unit, and then combine the verified spatial feature units with the pile foundation type, construction scenario, and global coordinate system to configure a construction feature cluster under spatial evolution.

7. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 6, characterized in that, The methods for determining the strain anomaly points corresponding to each pile foundation include: Using the abnormal areas corresponding to construction deviations as the verification objects, and using the rate of change of strain values ​​over time, determine whether there are any abnormalities in the strain gradient at the current location; When there is an anomaly in the strain gradient at the current location, the strain gradient in the corresponding direction is checked simultaneously with the anomaly in the axial and horizontal directions corresponding to the construction deviation point, and the correlation between the strain anomaly point and the axial and horizontal directions is marked. If there is no anomaly in the strain gradient at the current location, check the strain gradient of the remaining area excluding the abnormal area. If an anomaly is confirmed, mark the strain anomaly point.

8. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 6, characterized in that, When verifying the positional matching between various spatial feature units, the implementation methods include: For any two spatial feature units, the data corresponding to strain anomalies and construction anomalies are transformed into feature vectors, and the matching scores between each spatial feature unit are recorded in the form of cosine similarity. For spatial feature elements that simultaneously exhibit strain gradient anomalies and construction deviation anomalies, the matching score between each spatial feature element is calculated and recorded by combining the spatial overlap rate and direction of its location. For spatial feature elements with only strain gradient anomalies and only construction deviation anomalies, the matching score between each spatial feature element is calculated and recorded in the manner of matching the same anomaly form. Select any two spatial feature units corresponding to the maximum matching score as the output data to complete the location matching processing.

9. The pile foundation construction quality monitoring method based on multi-data fusion according to claim 1, characterized in that, When performing multi-level condition judgments on each pile, the implementation methods include: Using spatial feature similarity, construction scene rules, and coordinate association as hierarchical conditions, the construction feature clusters are searched to determine the description label of the current construction feature cluster under the corresponding hierarchical conditions. The description labels under each level of condition are connected one by one according to their level to form the output quality description labels.

10. A method for monitoring the construction quality of pile foundations based on multi-data fusion according to claim 1, characterized in that, When adjusting the parameters for the next construction phase, the methods include: For each quality description label associated with a hierarchical condition, the parameters to be adjusted at different locations are determined based on the construction locations involved in the hierarchical condition, and the part with the largest coverage of the parameters to be adjusted is regarded as the current output parameter.

Images