Construction quality dynamic monitoring method and system for heavy haul railway transition section inclined jet grouting pile reinforcement facing settlement control

Abstract

The present application relates to the technical field of construction monitoring, and discloses a method and system for dynamic monitoring of construction quality of oblique jet grouting piles in heavy haul railway transition section for settlement control, which comprises: real-time collection of construction parameters including real-time drilling speed, real-time grouting pressure and real-time grout flow rate and settlement records including real-time settlement and vibration interference values of subgrade surface and slope toe during the construction stage of oblique jet grouting piles; calculation of pile forming quality indicators based on the construction parameters and calculation of layered subgrade settlement rates based on the settlement records; construction of a three-dimensional correlation model with the construction parameters as the input layer, the pile forming quality indicators as the intermediate layer and the layered subgrade settlement rates as the output layer; when a settlement risk is determined to occur, reverse matching of the settlement risk traceability results is carried out based on the correlation model according to the same period parameters, construction parameter adjustment suggestions are determined, and a construction quality detection report is generated; and effective settlement control of the construction process of oblique jet grouting piles in heavy haul railway transition section is realized.

Description

Technical Field

[0001] This invention relates to the field of construction monitoring technology, and in particular to a method and system for dynamic monitoring of the construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railway for settlement control. Background Technology

[0002] In heavy-haul railway construction, the transition section, as a crucial link connecting different structures, is vital for the safe operation of the railway. Oblique jet grouting reinforcement technology is widely used in the construction of heavy-haul railway transition sections due to its ability to effectively enhance the bearing capacity of the foundation and control settlement. However, ensuring the construction quality of oblique jet grouting reinforcement and effectively controlling settlement have become key tasks in railway construction. Therefore, a dynamic monitoring method and system for the construction quality of oblique jet grouting reinforcement in heavy-haul railway transition sections, oriented towards settlement control, is of great significance. By monitoring various parameters and settlement conditions in real time during construction, potential quality problems and settlement risks can be identified in a timely manner, allowing for targeted adjustments and optimizations. This not only helps ensure the long-term stability of heavy-haul railway transition sections and reduce later maintenance costs, but also improves the safety and reliability of railway operation. With the continuous expansion of the scale of heavy-haul railway construction in my country, the requirements for construction quality and settlement control are becoming increasingly stringent. This technology has broad application prospects and is expected to promote the development of heavy-haul railway construction technology towards a more refined and intelligent direction.

[0003] However, existing construction quality monitoring technologies for inclined jet grouting pile reinforcement in heavy-haul railway transition sections lack comprehensive and timely understanding of actual construction conditions, making it difficult to accurately assess pile quality and settlement trends. The inability to effectively construct a correlation model between construction quality, pile quality, and settlement risk results in the inability to accurately trace the root cause of settlement risk and provide reasonable suggestions for adjusting construction parameters and construction quality inspection reports when abnormal subgrade settlement occurs. This hinders the provision of strong support for construction quality control, impacting the construction quality and long-term stability of heavy-haul railway transition sections.

[0004] Therefore, this invention proposes a method and system for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railway for settlement control. Summary of the Invention

[0005] This invention provides a method and system for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control. By collecting construction parameters and settlement records in real time, it comprehensively and promptly obtains key information during the construction process, providing an accurate data foundation for subsequent analysis. Based on the collected data, it calculates pile quality indicators and layered subgrade settlement rates, quantifying construction parameters and settlement conditions to facilitate in-depth analysis of their relationship. A correlation model is constructed between construction quality, pile quality, and settlement risk, systematically revealing the intrinsic links between each construction stage and settlement risk. When the subgrade settlement rate exceeds a threshold, the correlation model is used to back-match synchronous parameters to obtain settlement risk tracing results, accurately pinpointing the causes of settlement risk. Based on the tracing results, construction parameter adjustment suggestions are determined and a test report is generated, providing a scientific basis for construction adjustments, helping to optimize the construction process in a timely manner, effectively control the settlement of heavy-haul railway transition sections, and ensure railway construction quality and subsequent operational safety.

[0006] This invention provides a method for dynamic monitoring of the construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railway for settlement control, comprising:

[0007] Real-time acquisition of construction parameters and settlement records during the construction phase of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section.

[0008] The pile quality index was calculated based on the construction parameters, and the settlement rate of the layered subgrade was calculated based on the settlement records.

[0009] Using construction parameters as the input layer, pile quality indicators as the intermediate layer, and layered roadbed settlement rate as the output layer, a correlation model of construction quality, pile quality, and settlement risk is constructed using a multiple linear regression algorithm.

[0010] When it is determined that the settlement rate of the layered roadbed exceeds the corresponding preset settlement rate threshold, the settlement risk source tracing result is obtained by back matching the synchronous parameters based on the correlation model.

[0011] Based on the results of settlement risk tracing, recommendations for adjusting construction parameters are determined, and a construction quality inspection report is generated.

[0012] Preferably, real-time acquisition of construction parameters and settlement records during the construction phase of the inclined jet grouting pile includes:

[0013] The drilling rig integrates sensors to collect real-time drilling speed, real-time grouting pressure, and real-time grout flow rate as construction parameters during the construction phase of the inclined jet grouting pile.

[0014] Based on the settlement observation markers and vibration acceleration sensors installed on the roadbed surface and slope toe of the heavy-haul railway transition section, the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section are collected in real time as settlement records.

[0015] Preferably, the pile quality index is calculated based on construction parameters, and the settlement rate of the layered subgrade is calculated based on settlement records, including:

[0016] The pile diameter is calculated based on the real-time grouting pressure, real-time grout flow rate, and stratum correction coefficient in the construction parameters.

[0017] The uniformity of the pile body was calculated by the wavelet packet energy entropy of the real-time drilling speed.

[0018] Based on the vibration correction coefficient and the vibration disturbance value in the settlement record, the real-time settlement in the settlement record is corrected to obtain the real-time corrected settlement.

[0019] Based on the initial depth of the roadbed surface and the toe of the slope, the subgrade surface layer, the middle layer, and the bottom layer are divided.

[0020] The subgrade settlement rates for the surface layer, middle layer, and bottom layer of the subgrade were calculated based on the real-time corrected settlement amount and used as the layered subgrade settlement rates.

[0021] Among them, the pile quality indicators include pile diameter and pile uniformity.

[0022] Preferably, the training process of the association model includes:

[0023] The correlation coefficients between the intermediate layer and the input layer, and between the output layer and the intermediate layer, are used as model parameters. The parameters are optimized through cross-validation to ensure that the goodness of fit of the correlation model is not less than the preset goodness of fit threshold.

[0024] Preferably, the results of settlement risk tracing are obtained by back-matching contemporaneous parameters based on the correlation model, including:

[0025] A construction parameter vector is generated based on the current construction parameters, and the construction parameter vector is standardized to obtain a standardized construction parameter vector. The standardized construction parameter vector is then input into the association model for forward propagation calculation to obtain the theoretical value of the pile quality index corresponding to the intermediate layer.

[0026] The theoretical and actual pile quality indicators obtained continuously are fitted into a two-dimensional sequence data. The two-dimensional sequence data is dynamically time-normalized and the trend correlation is calculated. Based on the trend correlation, the theoretical and actual pile quality indicators obtained continuously are weighted and fused to obtain a comprehensive quality assessment vector.

[0027] The deviation between the theoretical value and the actual value of the pile quality index is projected onto the two-dimensional diagnostic plane to obtain the diagnostic projection point. Based on the historical database, the two-dimensional diagnostic plane is divided into polygon qualified domains, and the shortest Euclidean distance from the diagnostic projection point to the boundary of the polygon qualified domain is calculated.

[0028] When the shortest Euclidean distance is not less than the preset distance threshold, the stratum sensitivity coefficient matrix is ​​introduced to transform the coordinates of the comprehensive quality assessment vector, and the contribution of each pile quality index dimension to the comprehensive deviation index is calculated in the transformed space.

[0029] Based on the contribution of each pile quality indicator dimension to the comprehensive deviation index, the dominant deviation indicators are identified among the pile quality indicators. Based on the dominant deviation indicators, risk tracing is carried out in the risk reasoning knowledge graph to obtain the settlement risk tracing results.

[0030] Preferably, the polygonal qualified domain is divided in the two-dimensional diagnostic plane based on a historical database, including:

[0031] Extract the pile diameter and pile uniformity of all qualified construction stages from the historical database to form a set of qualified data points on the two-dimensional diagnostic plane, and extract the pile diameter and pile uniformity of all unqualified construction stages to form a set of unqualified data points on the two-dimensional diagnostic plane.

[0032] A density-based spatial clustering algorithm is used to cluster the qualified data point set and identify the set of qualified pile quality points that are density-connected.

[0033] An ensemble learning classification model is trained based on a set of qualified data points and a set of unqualified data points to obtain an ensemble qualification discrimination model.

[0034] Based on the integrated qualification discrimination model, the dense grid points of the two-dimensional diagnostic plane are predicted, the qualification probability of the dense grid points of the two-dimensional diagnostic plane is output, and the qualification probability distribution map is generated based on the qualification probability of the dense grid points of the two-dimensional diagnostic plane.

[0035] Based on the pass probability distribution map, multiple lines with similar pass probabilities are drawn in the two-dimensional diagnostic plane to divide the pass probability of each target.

[0036] Based on the set of qualified pile quality points and multiple lines of similar qualification probabilities for all targets, a polygonal qualified domain is divided in the two-dimensional diagnostic plane.

[0037] Preferably, based on the set of qualified pile quality points and multiple lines of similarity in the probability of all target qualified points, a polygonal qualified domain is divided in the two-dimensional diagnostic plane, including:

[0038] The set of qualified pile quality points is outlined to obtain the boundary of the qualified pile quality point set, and the physical center of the boundary of the qualified pile quality point set is marked.

[0039] The distance between the boundary points of the pile quality qualified point set on each straight line passing through the physical center and the points contained in each qualified probability similar line of each target qualified probability is regarded as the local distance between the boundary of the pile quality qualified point set and each qualified probability similar line of each target qualified probability in the corresponding straight line direction passing through the physical center.

[0040] Based on the local spacing between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target in all straight lines passing through the physical center, the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target are calculated.

[0041] Based on the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and all lines with similar qualified probabilities for each target, the matching degree between the boundary of the pile quality qualified point set and the qualified probability of each target is calculated.

[0042] Determine the matching weight of the pass probability of each target, and calculate the comprehensive matching degree of the pass probability of each target based on the matching weight of the pass probability of each target and the matching degree between the boundary of the pass point set of pile quality and the pass probability of each target.

[0043] Based on the boundary of all similar lines of the target qualification probability corresponding to the maximum comprehensive matching degree and the set of qualified points of pile quality, a polygonal qualification domain is divided in the two-dimensional diagnostic plane.

[0044] Preferably, based on the boundary of all similar lines of pass probability corresponding to the maximum comprehensive matching degree and the set of pass points of pile quality, a polygonal pass domain is divided in the two-dimensional diagnostic plane, including:

[0045] The verification region is defined as all simply connected regions formed by the intersection of all lines with similar pass probabilities corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality.

[0046] Calculate the pass / fail value for each verification region based on the pass / fail probability of all dense grid points within each verification region;

[0047] The region with the most frequent overlap between the lines of similar pass probabilities corresponding to the target pass probability corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality is taken as the core pass domain.

[0048] All verification areas with valid verification values ​​not less than the preset verification value threshold are merged with the core valid area and treated as polygonal valid areas in the two-dimensional diagnostic plane.

[0049] Preferably, based on the results of settlement risk tracing, recommendations for adjusting construction parameters are determined, and a construction quality inspection report is generated, including:

[0050] Based on the results of settlement risk tracing, recommendations for adjusting construction parameters were determined, and actual construction parameters after adjustment were collected. At the same time, new pile quality indicators were calculated, and new settlement records were obtained.

[0051] Calculate the compliance rate of construction parameters based on the current actual construction parameters;

[0052] Calculate the pile quality pass rate based on the new pile quality indicators;

[0053] Calculate the settlement risk control rate based on the new settlement records;

[0054] A construction quality inspection report is generated based on the compliance rate of construction parameters, the pass rate of pile quality, and the settlement risk control rate.

[0055] This invention provides a dynamic monitoring system for the construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railways, aimed at settlement control, comprising:

[0056] The construction record module is used to collect construction parameters and settlement records in real time during the construction of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section.

[0057] The noise reduction module is used to calculate pile quality indicators based on construction parameters and to calculate the settlement rate of layered subgrade based on settlement records.

[0058] The correlation analysis module is used to construct a correlation model of construction quality, pile quality, and settlement risk by taking construction parameters as the input layer, pile quality indicators as the intermediate layer, and roadbed settlement rate as the output layer, and using a multiple linear regression algorithm.

[0059] The risk tracing module is used to obtain the settlement risk tracing results by back-matching the synchronous parameters based on the correlation model when the settlement rate is determined to exceed the preset settlement rate threshold based on the initial settlement amount.

[0060] The quality monitoring module is used to determine construction parameter adjustment suggestions based on the settlement risk tracing results and generate construction quality inspection reports.

[0061] The beneficial effects of this invention compared to existing technologies are as follows: By collecting construction parameters and settlement records in real time, key information during the construction process can be obtained comprehensively and promptly, providing an accurate data foundation for subsequent analysis. Based on the collected data, pile quality indicators and layered subgrade settlement rates are calculated, quantifying the relationship between construction parameters and settlement, facilitating in-depth analysis of the relationship between the two. A correlation model of construction quality, pile quality, and settlement risk is constructed, systematically revealing the intrinsic connection between each stage of construction and settlement risk. When the subgrade settlement rate exceeds a threshold, the correlation model is used to back-match synchronous parameters to obtain settlement risk tracing results, accurately pinpointing the cause of the settlement risk. Based on the tracing results, construction parameter adjustment suggestions are determined and a test report is generated, providing a scientific basis for construction adjustments, helping to optimize the construction process in a timely manner, effectively control the settlement of heavy-haul railway transition sections, and ensure railway construction quality and subsequent operational safety.

[0062] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0063] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0064] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0065] Figure 1 This is a flowchart of a method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railway for settlement control, as described in this embodiment of the invention.

[0066] Figure 2 This is a core flowchart of risk tracing in an embodiment of the present invention;

[0067] Figure 3 This is a flowchart illustrating the dynamic delineation of the qualified polygonal region in an embodiment of the present invention. Detailed Implementation

[0068] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0069] like Figure 1 As shown, this invention provides an implementation method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control, comprising:

[0070] Real-time acquisition of construction parameters and settlement records during the construction phase of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section.

[0071] The pile quality index was calculated based on the construction parameters, and the settlement rate of the layered subgrade was calculated based on the settlement records.

[0072] Using construction parameters as the input layer, pile quality indicators as the intermediate layer, and layered roadbed settlement rate as the output layer, a correlation model of construction quality, pile quality, and settlement risk is constructed using a multiple linear regression algorithm.

[0073] When it is determined that the settlement rate of the layered roadbed exceeds the corresponding preset settlement rate threshold, the settlement risk source tracing result is obtained by back matching the synchronous parameters based on the correlation model.

[0074] Based on the results of settlement risk tracing, recommendations for adjusting construction parameters are determined, and a construction quality inspection report is generated.

[0075] In this embodiment of the dynamic monitoring method for the construction quality of inclined jet grouting pile reinforcement in the transition section of a heavy-haul railway aimed at settlement control:

[0076] The inclined jet grouting pile construction stage refers to the period during which inclined jet grouting piles are constructed in the transition section of heavy-haul railways. During this period, a series of operations from the start of construction to the formation of the pile body must be completed, which is also the stage for the collection of various data.

[0077] Construction parameters: These are important data collected in real time during the construction of inclined jet grouting piles, including real-time drilling speed, real-time grouting pressure, and real-time grout flow rate.

[0078] Settlement records are records of relevant data on the subgrade surface and slope toe of the transition section of heavy-haul railways, including real-time settlement and vibration disturbance values, used to monitor subgrade settlement changes and vibration effects during construction.

[0079] Real-time drilling speed: As one of the construction parameters, it is the drilling speed of the drill rod during the construction of inclined jet grouting piles. The data collection range is 5-30cm / min, the sampling frequency is ≥2Hz, and the data error is ≤±0.5%. It is obtained through the laser Doppler drilling speed sensor integrated into the drilling rig. This speed affects pile quality indicators such as pile uniformity.

[0080] Real-time grouting pressure: This is also a construction parameter, which is the pressure during the grouting process. The measurement range is 0-50MPa. It is collected by a fiber optic grating grouting pressure sensor with an accuracy of ±0.1MPa and an anti-electromagnetic interference level of IP68. Its magnitude will affect the pile diameter and other pile quality. The acquisition range is 20-30MPa.

[0081] Real-time grout flow rate: Also a construction parameter, it is collected by a Coriolis mass flow meter with a measurement range of 50-150 L / min and a mass flow rate error of ≤±0.5%. It is used together with grouting pressure to calculate the pile diameter, with a collection range of 80-120 L / min.

[0082] Transition sections of heavy-haul railways: These are the connecting parts of different structures in heavy-haul railways, such as the transition areas between road bridges and culverts. Due to structural differences, settlement problems are prone to occur, making them key areas for inclined jet grouting pile reinforcement and related monitoring.

[0083] Roadbed surface and slope toe: These are specific locations on the roadbed of heavy-haul railways. Monitoring equipment is evenly distributed on the roadbed surface at 2m intervals, with the intervals increasing to 0.5m at structural joints; at the slope toe, equipment is distributed at 1.5m intervals to house settlement observation markers and vibration acceleration sensors, acquiring real-time settlement and vibration disturbance values.

[0084] Real-time settlement: This is part of the settlement record. It refers to the real-time settlement value of the roadbed obtained by setting settlement observation points on the roadbed surface and slope toe during the construction process. The unit is mm. It can intuitively reflect the settlement of the roadbed.

[0085] Vibration interference value: This is part of the settlement record and is collected by a vibration acceleration sensor with a range of 0-20g and a frequency response of 0.1-1000Hz. It collects the instantaneous vibration peak value (unit: m / s²) and is used to correct the real-time settlement and reduce the interference of vibration on the settlement data.

[0086] Pile quality indicators: These are used to measure the quality of inclined jet grouting piles, including pile diameter and pile uniformity indicators. The pile diameter is calculated using real-time grouting pressure, real-time grout flow rate, and formation correction coefficients; the pile uniformity indicator is calculated using the wavelet packet energy entropy of the real-time drilling speed.

[0087] Layered subgrade settlement rate: Based on the initial depth of the subgrade surface and the toe of the slope, the subgrade surface layer (0-0.6m), the middle layer (0.6-1.2m), and the bottom layer (1.2-2.3m) are divided. The subgrade settlement rate of each layer is calculated based on the real-time corrected settlement amount, with the unit being mm / day, reflecting the rate of settlement change of the subgrade at different depths.

[0088] Using construction parameters as the input layer, pile quality indicators as the intermediate layer, and layered subgrade settlement rate as the output layer, a correlation model of construction quality, pile quality, and settlement risk is constructed using a multiple linear regression algorithm. This model takes real-time drilling speed, real-time grouting pressure, and real-time grout flow rate as inputs, pile diameter and pile uniformity indicators as intermediate links, and layered subgrade settlement rate as outputs, using a multiple linear regression algorithm to establish the mathematical relationship between the three. This model can analyze how construction parameters affect pile quality, and thus subgrade settlement risk. For example, a 1 MPa decrease in grouting pressure leads to a 5% reduction in pile diameter and an increase of 0.1 mm / day in the initial subgrade settlement rate. The model needs to be optimized through cross-validation to ensure that the goodness of fit is not less than a preset goodness of fit threshold.

[0089] Preset settlement rate threshold: This is a pre-set standard value for the settlement rate of the subgrade layers, such as 0.3 mm / day for the surface layer, 0.2 mm / day for the middle layer, and 0.15 mm / day for the bottom layer. It is used to determine whether the subgrade settlement is within a controllable range and serves as the basis for triggering the settlement risk tracing mechanism.

[0090] The settlement rate of the layered subgrade exceeds the corresponding preset settlement rate threshold: When the settlement rate of any layer of the subgrade, whether it is the surface layer, the middle layer, or the bottom layer, exceeds its respective preset settlement rate threshold, it indicates that there may be a risk of settlement in the layered subgrade, and the settlement risk tracing mechanism needs to be triggered to find the cause of the abnormal settlement.

[0091] Settlement risk tracing results: When the settlement rate of the layered subgrade exceeds the corresponding preset settlement rate threshold, the results are obtained by back-matching the synchronous parameters based on the correlation model. Through a series of calculations, such as analyzing the deviation between the theoretical and actual values ​​of pile quality indicators, the source of settlement risk is determined, such as "unstable grouting pressure leading to insufficient pile quality" or "excessive drilling speed leading to insufficient pile compaction".

[0092] Based on the results of settlement risk tracing, the following recommendations for adjusting construction parameters are provided: For example, if the risk source is fluctuation in grouting pressure, it is recommended to stabilize the grouting pump pressure at 24-26 MPa and control the grout flow rate at 100-110 L / min; if the problem is excessive drilling speed, it is recommended to reduce the drilling speed to 15-18 cm / min to ensure that the uniformity of the pile body meets the requirements, thereby controlling settlement risk and ensuring construction quality.

[0093] To accurately and in real-time collect construction parameters and settlement records during the construction phase of inclined jet grouting piles, and to provide reliable data for subsequent analysis, a method for real-time collection of construction parameters and settlement records during the construction phase of inclined jet grouting piles is proposed, including:

[0094] The drilling rig integrates sensors to collect real-time drilling speed, real-time grouting pressure, and real-time grout flow rate as construction parameters during the construction phase of the inclined jet grouting pile.

[0095] Based on the settlement observation markers and vibration acceleration sensors installed on the roadbed surface and slope toe of the heavy-haul railway transition section, the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section are collected in real time as settlement records.

[0096] In this embodiment:

[0097] Sensors integrated into the drilling rig: These are devices installed on the drilling rig to collect real-time construction parameters of inclined jet grouting piles, including the following types:

[0098] Laser Doppler drilling speed sensor: Installed at the drill rod feed drive motor, with a measurement range of 5-30 cm / min, a resolution of 0.05 cm / min, and a sampling frequency ≥2 Hz. It calculates the drilling speed by monitoring the motor speed and drill rod pitch, accurately measuring the drilling speed and providing data support for analyzing pile quality indicators such as pile uniformity. Furthermore, its vibration resistance level is suitable for the high-frequency vibration environment of heavy-haul railway construction.

[0099] Fiber Bragg grating grouting pressure sensor: Connected in series in the grouting pipeline, it measures pressure from 0 to 50 MPa with an accuracy of ±0.1 MPa and an IP68 electromagnetic interference resistance rating. This sensor uses a fiber Bragg grating sensing element to capture instantaneous pressure fluctuations, has a fast response time, and transmits data in real time via a specific bus. The measured grouting pressure plays a crucial role in calculating pile diameter and other pile quality indicators.

[0100] Coriolis slurry flow sensor: Designed based on the principle of electromagnetic induction, it is adapted to the high viscosity characteristics of cement slurry, with a measurement range of 50-150 L / min and a mass flow error of ≤±0.5%. It features automatic descaling function, with periodic cleaning by an internal ultrasonic cleaning unit and a filter screen at the inlet to prevent clogging, ensuring stable measurement accuracy. The collected slurry flow rate is used to calculate the pile diameter.

[0101] Settlement monitoring markers and vibration acceleration sensors are installed on the subgrade surface and toe of the slope in the transition section of the heavy-haul railway. These sensors are used to collect subgrade condition data in real time, as detailed below:

[0102] Settlement monitoring markers are evenly distributed at 2m intervals on the subgrade surface of the heavy-haul railway transition section, with a denser spacing of 0.5m at structural joints; and at 1.5m intervals at the toe of the slope. Fiber optic displacement sensors are used, with a measurement range of 0-50mm and an accuracy of ±0.05mm. They are primarily used to collect real-time settlement data of the subgrade surface and toe of the heavy-haul railway transition section, providing a data foundation for calculating the settlement rate of the layered subgrade and subsequently analyzing the subgrade settlement situation.

[0103] Vibration acceleration sensors: Installed at 1.5m intervals at the toe of the slope in the transition section of the heavy-haul railway, these are ICP-type piezoelectric sensors with a range of 0-20g and a frequency response of 0.1-1000Hz. Their function is to collect vibration interference values ​​in real time. These values ​​are used to correct the real-time settlement data collected by settlement monitoring points, eliminating interference from factors such as train vibration on the settlement data and ensuring that the settlement data more accurately reflects the subgrade settlement caused by construction.

[0104] To quantify the relationship between pile quality and subgrade settlement, a method is proposed to calculate pile quality indices based on construction parameters and to calculate the settlement rate of layered subgrades based on settlement records, including:

[0105] The pile diameter is calculated based on the real-time grouting pressure, real-time grout flow rate, and stratum correction coefficient in the construction parameters.

[0106] The uniformity of the pile body was calculated by the wavelet packet energy entropy of the real-time drilling speed.

[0107] Based on the vibration correction coefficient and the vibration disturbance value in the settlement record, the real-time settlement in the settlement record is corrected to obtain the real-time corrected settlement.

[0108] Based on the initial depth of the roadbed surface and the toe of the slope, the subgrade surface layer, the middle layer, and the bottom layer are divided.

[0109] The subgrade settlement rates for the surface layer, middle layer, and bottom layer of the subgrade were calculated based on the real-time corrected settlement amount and used as the layered subgrade settlement rates.

[0110] Among them, the pile quality indicators include pile diameter and pile uniformity.

[0111] In this embodiment:

[0112] The pile diameter is calculated based on real-time grouting pressure, real-time grout flow rate, and soil correction factor. The calculation formula is as follows:

[0113]

[0114] in Indicates the diameter of the pile (unit: m). It is the real-time grouting pressure (unit: MPa). Real-time slurry flow rate (unit: L / min). It is the stratigraphic correction factor (0.8-0.9 for soft soil, 1.0-1.1 for sand, and 1.1-1.2 for clay). The time decay coefficient (0.02 min for soft soil) is given. -1 Sand layer 0.015min-1 clay 0.01min -1 ), This represents the construction time for a single pile (unit: min). This formula considers different geological characteristics and the effect of time on grout diffusion. For example, in soft soil strata with a grouting pressure of 25 MPa, a flow rate of 100 L / min, and a construction time of 30 min, substituting into the formula yields... The diameter of the pile is approximately 0.78m, which is used to accurately calculate the pile diameter and evaluate the quality of the pile.

[0115] The uniformity of the pile body was calculated using the wavelet packet energy entropy of the real-time drilling speed: wavelet packet decomposition was performed on the real-time drilling speed sequence (using the db4 wavelet basis, decomposition in 3 layers), which yielded energy values ​​for 8 frequency bands. Then through the formula The uniformity of the pile body is quantified by calculating the wavelet packet energy entropy. If the pile is deemed uniform, it means that the energy distribution is concentrated. This is determined to be non-uniform, meaning the energy distribution is dispersed. For example, if the drilling speed sequence has small fluctuations, the calculated... If the pile body is uniform, then it is determined that the pile body is uniform; if the fluctuation is large, If the pile is uneven, it is considered non-uniform. This method assesses the uniformity of the pile body by analyzing the energy distribution of the drilling speed, and is one of the important indicators of pile quality.

[0116] Based on the vibration correction factor and the vibration interference value in the settlement record, the real-time settlement in the settlement record is corrected to obtain the real-time corrected settlement: because vibration can interfere with the settlement data, correction is necessary. Based on the vibration correction factor... (0.05 for soft soil, 0.03 for sand, and 0.02 for clay) and vibration disturbance values ​​in the settlement record. (Unit: m / s²), through the formula To calculate the real-time corrected settlement. The corrected value. This is the real-time settlement measured by the settlement monitoring point (unit: mm), and the vibration duration is... This method eliminates the influence of vibration on settlement data, allowing the settlement data to more accurately reflect the roadbed settlement caused by construction.

[0117] Based on the initial depth of the subgrade surface and toe, the subgrade is divided into three layers: surface layer, middle layer, and bottom layer. According to the heavy-haul railway subgrade design specifications, different layers are defined based on the initial depth of the subgrade surface and toe. Specifically, the surface layer ranges from 0 to 0.6 m, the middle layer from 0.6 to 1.2 m, and the bottom layer from 1.2 to 2.3 m. This division facilitates more detailed settlement analysis of subgrades at different depths, as the stress and settlement characteristics of different subgrade layers may differ.

[0118] Based on the real-time corrected settlement, the subgrade settlement rates for the surface layer, middle layer, and bottom layer of the subgrade were calculated separately as the layered subgrade settlement rates: After obtaining the real-time corrected settlement, the settlement rates were calculated for the surface layer, middle layer, and bottom layer of the subgrade respectively. The calculation formula is as follows:

[0119]

[0120] in Corresponding to the top layer, middle layer, and bottom layer respectively. for The corrected settlement value of this layer at that time. for The corrected settlement value of this layer at that time. Use 24 hours (i.e., 1 day). By calculating the settlement rate of layered subgrades, we can more accurately understand the settlement changes of subgrades at different depths, providing a basis for judging subgrade settlement risk.

[0121] To improve the accuracy and reliability of the association model, the training process of the association model is proposed to include:

[0122] The correlation coefficients between the intermediate layer and the input layer, and between the output layer and the intermediate layer, are used as model parameters. The parameters are optimized through cross-validation to ensure that the goodness of fit of the correlation model is not less than the preset goodness of fit threshold.

[0123] In this embodiment, parameter optimization through cross-validation refers to dividing the data samples into multiple subsets. For example, using 5-fold cross-validation, the data is divided into 5 subsets. Each time, 4 subsets are selected as the training set to train the model, and the remaining subset is used as the validation set to test the model's performance. This process is repeated multiple times. In each iteration, the model learns parameters based on the training set and calculates evaluation metrics on the validation set. By continuously adjusting the parameters, the model performs well on each validation set, thereby finding the optimal parameter combination, improving the model's generalization ability, and avoiding overfitting.

[0124] In this embodiment, the correlation model fit goodness is used to measure the degree of model fit to the data, reflecting the closeness between the model's predicted values ​​and the actual observed values. The higher the value, the stronger the model's ability to explain the relationship between construction parameters, pile quality indicators and layered roadbed settlement rate.

[0125] In this embodiment, the preset goodness-of-fit threshold is a pre-set standard value. The goodness-of-fit of the constructed association model must reach or exceed this preset threshold to be considered qualified and able to meet the needs of actual application. For example, if it is set to 0.95, the goodness-of-fit of the model must reach 0.95 or above.

[0126] In this embodiment, the correlation coefficients between the intermediate layer and the input layer, and between the output layer and the intermediate layer are used as model parameters. The parameters are optimized through cross-validation to ensure that the goodness of fit of the correlation model is not less than a preset goodness of fit threshold. This means that when constructing the correlation model of construction quality, pile quality, and settlement risk, the construction parameters are used as the input layer, the pile quality index is used as the intermediate layer, and the settlement rate of the layered subgrade is used as the output layer. The mutual influence coefficients between the intermediate layer and the input layer, and between the output layer and the intermediate layer are used as model parameters. The parameters are adjusted using the cross-validation method to ensure that the goodness of fit of the model reaches or exceeds the preset standard, thus ensuring that the model accurately reflects the relationship between the three.

[0127] like Figure 2 As shown, in order to accurately pinpoint the causes of settlement risk, a method based on reverse matching of contemporaneous parameters using a correlation model is proposed to obtain settlement risk tracing results, including:

[0128] A construction parameter vector is generated based on the current construction parameters, and the construction parameter vector is standardized to obtain a standardized construction parameter vector. The standardized construction parameter vector is then input into the association model for forward propagation calculation to obtain the theoretical value of the pile quality index corresponding to the intermediate layer.

[0129] The theoretical and actual pile quality indicators obtained continuously are fitted into a two-dimensional sequence data. The two-dimensional sequence data is dynamically time-normalized and the trend correlation is calculated. Based on the trend correlation, the theoretical and actual pile quality indicators obtained continuously are weighted and fused to obtain a comprehensive quality assessment vector.

[0130] The deviation between the theoretical value and the actual value of the pile quality index is projected onto the two-dimensional diagnostic plane to obtain the diagnostic projection point. Based on the historical database, the two-dimensional diagnostic plane is divided into polygon qualified domains, and the shortest Euclidean distance from the diagnostic projection point to the boundary of the polygon qualified domain is calculated.

[0131] When the shortest Euclidean distance is not less than the preset distance threshold, the stratum sensitivity coefficient matrix is ​​introduced to transform the coordinates of the comprehensive quality assessment vector, and the contribution of each pile quality index dimension to the comprehensive deviation index is calculated in the transformed space.

[0132] Based on the contribution of each pile quality indicator dimension to the comprehensive deviation index, the dominant deviation indicators are identified among the pile quality indicators. Based on the dominant deviation indicators, risk tracing is carried out in the risk reasoning knowledge graph to obtain the settlement risk tracing results.

[0133] In this embodiment, a construction parameter vector is generated based on the current construction parameters. This involves integrating the currently collected real-time drilling speed, real-time grouting pressure, and real-time grout flow rate into a single vector for subsequent analysis. The construction parameter vector is then standardized using methods such as Z-score standardization to obtain a standardized construction parameter vector. According to the formula ( The sample mean. The standard deviation is converted to make parameters of different dimensions comparable, eliminating the influence of dimensions and obtaining a standardized construction parameter vector. This standardized construction parameter vector is then input into the correlation model for forward propagation calculation to obtain the theoretical values ​​of the pile quality indicators corresponding to the intermediate layer. This is based on the pre-constructed correlation model of construction parameters, pile quality, and settlement risk. The standardized construction parameter vector is input into the model, and calculations are performed according to the model's internal operational logic to obtain the theoretical values ​​of the pile quality indicators (pile diameter, pile uniformity index).

[0134] In this embodiment, the continuously obtained theoretical and actual pile quality indicators are fitted into a two-dimensional sequence of data. This involves arranging the theoretical and actual measured pile quality indicator values ​​obtained from multiple consecutive calculations sequentially to form two sequences, presented in a two-dimensional format for easier subsequent analysis of their relationship. Dynamic time warping is then applied to the two-dimensional sequence data, and the trend correlation is calculated. Dynamic time warping is an algorithm that finds the best match between two sequences when there are non-linear changes in the time series. This algorithm processes the two-dimensional sequence data to calculate the degree of correlation between the two sequences in terms of trend, measuring the similarity of the changing trends of the theoretical and actual values. Based on the trend correlation, the continuously obtained theoretical and actual pile quality indicators are weighted and fused to obtain a comprehensive quality assessment vector. That is, based on the calculated trend correlation, different weights are assigned to the theoretical and actual values, and they are fused together through weighted calculation to form a comprehensive quality assessment vector, which more comprehensively reflects the pile quality status. Assuming the theoretical pile diameter value sequence is... The actual value sequence is First, dynamic time warping is performed to calculate the trend correlation. The distances between corresponding elements of the theoretical and actual values ​​are calculated pairwise, such as the distance between 1.2 and 1.1, and these distances are combined into a distance matrix. Next, dynamic programming is used to find a path from the top left to the bottom right corner of this matrix. This process compares the cumulative distances of paths coming from above, left, and upper left, selecting the direction with the smallest cumulative distance. The cumulative distance of this optimal path is then normalized. For example, assuming the maximum possible distance is the maximum sum of the absolute differences between the elements of the two sequences, the ratio of the cumulative distance to the maximum possible distance is subtracted from 1, resulting in a trend correlation of approximately 0.477.

[0135] A weighted fusion based on trend correlation is used to obtain a comprehensive quality assessment vector. Based on the calculated trend correlation of 0.477, the theoretical value is weighted at 0.477, and the actual value weight is 1 minus 0.477, resulting in 0.523. For each time point, the theoretical and actual values ​​are multiplied according to their respective weights and then added together. For example, for the first time point, it's 0.477 multiplied by 1.2 plus 0.523 multiplied by 1.1, approximately equal to 1.148. This process is repeated to calculate the comprehensive value for each time point, ultimately forming the comprehensive quality assessment vector. This is used to more comprehensively reflect the quality of pile formation.

[0136] In this embodiment, the deviation between the theoretical and actual pile quality indicators is projected onto a two-dimensional diagnostic plane to obtain a diagnostic projection point. This involves projecting the deviation between the theoretical and actual values ​​of pile quality indicators such as pile diameter and pile uniformity onto a two-dimensional plane, using the pile diameter deviation as one coordinate axis and the pile uniformity deviation as another coordinate axis, to obtain a point representing the current pile quality deviation status, i.e., the diagnostic projection point.

[0137] In this embodiment, the historical database refers to data from the past three years on the construction of inclined jet grouting piles in the transition section of similar heavy-haul railways, showing whether the pile quality was qualified (pile diameter deviation ≤ ±10%, pile uniformity index meets requirements) or unqualified, and whether the stratified settlement rate met or did not meet the standards. This data is used for subsequent analysis and model training. The polygonal qualified domain defines the range of values ​​for the pile quality index within the qualified range.

[0138] In this embodiment, calculating the shortest Euclidean distance from the diagnostic projection point to the boundary of the qualified polygon domain is to use mathematical methods to calculate the straight-line distance from the diagnostic projection point to the nearest point on the boundary of the qualified polygon domain. This distance can be used as an indicator to assess the degree of deviation between the current pile quality and the qualified standard.

[0139] In this embodiment, the preset distance threshold is a pre-set distance standard. When the shortest Euclidean distance from the diagnostic projection point to the boundary of the qualified polygon domain is not less than this preset value, it indicates that the current pile quality deviates significantly from the qualified standard, which may pose a risk and requires further analysis.

[0140] In this embodiment, the stratum sensitivity coefficient matrix is ​​a matrix related to stratum characteristics. Different strata (soft soil, sand, clay, etc.) correspond to different matrix values, such as the soft soil stratum sensitivity coefficient matrix. Sandy stratigraphic sensitivity coefficient matrix This is used for subsequent coordinate transformation of the comprehensive quality assessment vector, taking into account the influence of geological factors on pile quality indicators.

[0141] In this embodiment, a soil sensitivity coefficient matrix is ​​introduced to transform the comprehensive quality assessment vector into coordinates. The contribution of each pile quality index dimension to the comprehensive deviation index is then calculated in the transformed space. This is achieved by multiplying the comprehensive quality assessment vector by the soil sensitivity coefficient matrix to perform the coordinate transformation. In the new space, the contribution ratio of each dimension, such as pile diameter deviation and pile uniformity deviation, to the comprehensive deviation index, which represents the overall degree of deviation, is analyzed. Assuming a heavy-haul railway transition section inclined jet grouting pile construction project has already undergone coordinate transformation and entered a new analysis space, we will now analyze the contribution ratio of pile diameter deviation and pile uniformity deviation to the comprehensive deviation index.

[0142] Suppose there are three piles. The diameter of the first pile is 0.1 meters smaller than the standard value. In terms of pile uniformity, it should have a uniform height, but in reality, there is a large difference in thickness from top to bottom. According to a certain evaluation standard, the uniformity deviation score is 8 points (out of 10, the lower the score, the greater the deviation). The diameter of the second pile is 0.05 meters larger than the standard value, and the pile uniformity deviation score is 6 points. The diameter of the third pile is 0.08 meters smaller than the standard value, and the pile uniformity deviation score is 7 points.

[0143] The overall deviation index is calculated to represent the overall degree of deviation. It's like assigning a comprehensive score to the deviations of these three stakes. After a series of analyses (without involving formulas), the overall deviation index is determined to be 7 points (assuming it's obtained through some comprehensive evaluation method).

[0144] Then, we analyze the contribution ratio of each dimension. For pile diameter deviation, we examine the impact of the diameter deviations of the three piles on the overall score of 7. For example, if the first pile has a large diameter deviation, its contribution to the comprehensive deviation index is relatively large, let's assume it accounts for 3 points out of 7; the second pile has a relatively small diameter deviation, let's assume it accounts for 1 point; and the third pile's diameter deviation is between the two, let's assume it accounts for 2 points. Thus, the total contribution of the three pile diameter deviations to the comprehensive deviation index is 6 out of 7, approximately 86%.

[0145] The same analysis applies to the uniformity deviation of the piles. The first pile has a uniformity deviation score of 8 points, which is assumed to account for 2 points of the comprehensive deviation index; the second pile has a uniformity deviation score of 6 points, which is assumed to account for 3 points; and the third pile has a uniformity deviation score of 7 points, which is assumed to account for 2 points. The uniformity deviation of the three piles accounts for 7 out of the total 7 points, so the contribution ratio of pile uniformity deviation to the comprehensive deviation index is 7 ÷ 7, which is 100%.

[0146] In this way, the contribution ratio of various dimensions, such as pile diameter deviation and pile uniformity deviation, to the comprehensive deviation index, which represents the overall degree of deviation, can be analyzed. Actual situations are more complex, involving more piles and more refined assessments, but the principle is similar.

[0147] In this embodiment, based on the contribution of each pile quality indicator dimension to the comprehensive deviation index, a dominant deviation indicator is identified among the pile quality indicators. This is achieved by comparing the contribution of each pile quality indicator dimension (such as pile diameter and pile uniformity) to the comprehensive deviation index. The pile quality indicator corresponding to the dimension with the larger contribution is the dominant deviation indicator, which plays a major role in the factors leading to settlement risk. Based on the dominant deviation indicator, risk tracing is performed in a risk reasoning knowledge graph to obtain settlement risk tracing results. The risk reasoning knowledge graph contains the causal relationships and edge weights between 12 types of construction anomalies, 8 types of quality defects, and 5 types of settlement risks. Starting with the dominant deviation indicator, a graph convolutional neural network (GCN) is used to reason about the risk propagation path in the knowledge graph, such as: grouting pressure fluctuation → pile diameter deviation as the dominant factor → excessive settlement of the subgrade surface. This yields the causes of settlement risk and related risk patterns, resulting in settlement risk tracing results. The method for determining the causal relationships and edge weights between nodes (i.e., 12 types of construction anomalies, 8 types of quality defects, and 5 types of settlement risks) in the risk reasoning knowledge graph includes: assuming that a large amount of historical data on the construction of inclined jet grouting piles in the transition section of heavy-haul railways has been collected, and that this data records various situations and corresponding results during the construction process.

[0148] The project includes 12 types of construction anomalies, such as unstable grouting pressure and excessive drilling speed; 8 types of quality defects, such as unqualified pile diameter and uneven pile body; and 5 types of settlement risks, such as excessive settlement of the subgrade surface and excessive settlement of the overall roadbed.

[0149] First, we statistically analyze the frequency of simultaneous occurrences of construction anomalies and quality defects in historical data. For example, in 100 construction cases, unstable grouting pressure was found 20 times, and in 15 of these instances, it was also accompanied by substandard pile diameter. Therefore, the frequency of this relationship—unstable grouting pressure leading to substandard pile diameter—is 15 times. This statistical analysis is performed for each combination of construction anomalies and quality defects.

[0150] Next, calculate the probability of each causal relationship. Divide the frequency calculated above by the total number of times the construction anomaly occurs to obtain the probability of this causal relationship. For example, the probability that unstable grouting pressure leads to an unqualified pile diameter is 15 ÷ 20 = 0.75. This probability can be used as a preliminary weight for the edges between them.

[0151] The causal relationship between quality defects and settlement risk, and the calculation of edge weights, follow a similar method. The frequency of simultaneous occurrences of quality defects and settlement risk is statistically analyzed, and then divided by the total number of occurrences of the quality defect to obtain the corresponding probability, which serves as the edge weight. For example, if the pile diameter is substandard 30 times, and 20 of these occurrences are accompanied by excessive settlement of the subgrade surface, then the probability that substandard pile diameter leads to excessive subgrade surface settlement is 20 ÷ 30 ≈ 0.67. This is the edge weight between them.

[0152] By calculating the frequency and probability of various events occurring simultaneously in historical data, the causal relationships and edge weights among 12 types of construction anomalies, 8 types of quality defects, and 5 types of settlement risks in the risk reasoning knowledge graph can be determined. This information can then be used to analyze how various factors during construction can cause quality problems and settlement risks.

[0153] like Figure 3 As shown, in order to provide a standard area for judging the quality of pile foundation, a polygonal qualified region is proposed to be divided in the two-dimensional diagnostic plane based on a historical database, including:

[0154] Extract the pile diameter and pile uniformity of all qualified construction stages from the historical database to form a set of qualified data points on the two-dimensional diagnostic plane, and extract the pile diameter and pile uniformity of all unqualified construction stages to form a set of unqualified data points on the two-dimensional diagnostic plane.

[0155] A density-based spatial clustering algorithm is used to cluster the qualified data point set and identify the set of qualified pile quality points that are density-connected.

[0156] An ensemble learning classification model is trained based on a set of qualified data points and a set of unqualified data points to obtain an ensemble qualification discrimination model.

[0157] Based on the integrated qualification discrimination model, the dense grid points of the two-dimensional diagnostic plane are predicted, the qualification probability of the dense grid points of the two-dimensional diagnostic plane is output, and the qualification probability distribution map is generated based on the qualification probability of the dense grid points of the two-dimensional diagnostic plane.

[0158] Based on the pass probability distribution map, multiple lines with similar pass probabilities are drawn in the two-dimensional diagnostic plane to divide the pass probability of each target.

[0159] Based on the set of qualified pile quality points and multiple lines of similar qualification probabilities for all targets, a polygonal qualified domain is divided in the two-dimensional diagnostic plane.

[0160] In this embodiment, the historical database is a collection of stored data, including various data on the construction of inclined jet grouting piles in the transition section of heavy-haul railways in the past, such as construction parameters at different stages, pile quality indicators, and subgrade settlement data, which provide data support for subsequent analysis and model training.

[0161] In this embodiment, the qualified construction stage refers to the construction period during the construction of inclined jet grouting piles, where all construction parameters, pile quality indicators, and subgrade settlement meet the pre-set qualified standards. For example, the pile diameter deviation is within the specified range, the pile uniformity meets the standards, and the settlement rate of the layered subgrade is normal. The qualified data point set is the collection of pile quality indicator values ​​gathered during the qualified construction stage.

[0162] In this embodiment, the unqualified construction phase, as opposed to the qualified construction phase, refers to the period during the construction of inclined jet grouting piles where at least one condition fails to meet the pre-set qualified standard, such as excessive deviation in pile diameter, substandard pile uniformity, or abnormal roadbed settlement rate. The unqualified data point set is a collection of pile quality index values ​​formed from data collected during the unqualified construction phase.

[0163] In this embodiment, a density-based spatial clustering algorithm is used to cluster the qualified data point set and identify the set of qualified pile quality points that are densely connected. That is, using an algorithm such as DBSCAN, based on the density relationship between data points, data points that are densely connected (the number of data points within a certain radius reaches a set threshold) in the qualified data point set are divided into a set. These sets represent different categories with similar pile quality characteristics and are all qualified, which facilitates the classification and analysis of qualified pile quality.

[0164] In this embodiment, an ensemble learning classification model is trained based on a set of qualified data points and a set of unqualified data points to obtain an ensemble qualification discrimination model. This involves using multiple basic classifiers (such as decision trees, Naive Bayes, etc.) to form an ensemble learning model, using the set of qualified data points and the set of unqualified data points as training data, allowing the model to learn the feature differences between qualified and unqualified data, and training it through methods such as Bagging and Boosting, ultimately obtaining an ensemble qualification discrimination model that can determine whether data is qualified or not.

[0165] In this embodiment, a dense grid point prediction is performed on the two-dimensional diagnostic plane based on the integrated qualification discrimination model, and the qualification probability of the dense grid points on the two-dimensional diagnostic plane is output. A qualification probability distribution map is generated based on the qualification probability of the dense grid points on the two-dimensional diagnostic plane. The two-dimensional diagnostic plane is divided into dense grid points, and the feature information of each grid point (such as construction parameters, pile quality indicators and other information) is input into the integrated qualification discrimination model. The model predicts the qualification probability of the pile quality represented by each grid point. These probability values ​​are presented on the plane to form a qualification probability distribution map, which intuitively shows the qualification probability distribution of pile quality at different locations.

[0166] In this embodiment, multiple lines with similar pass probabilities for each target pass probability are drawn in the two-dimensional diagnostic plane based on the pass probability distribution map. This is done by finding data points with similar probability values ​​(such as 0.61 and 0.608) in the map for each set target pass probability (e.g., 0.6). These points are then connected to form lines. These lines represent the approximate boundary of the region where the pile quality reaches the target pass probability on the two-dimensional diagnostic plane, which helps to analyze the distribution of the pass probability of pile quality in more detail.

[0167] like Figure 3 As shown, to optimize the rationality of polygonal qualified domain division, a method is proposed to divide the polygonal qualified domain in the two-dimensional diagnostic plane based on the set of qualified points of pile quality and multiple lines of similarity of qualified probabilities of all targets, including:

[0168] The set of qualified pile quality points is outlined to obtain the boundary of the qualified pile quality point set, and the physical center of the boundary of the qualified pile quality point set is marked.

[0169] The distance between the boundary points of the pile quality qualified point set on each straight line passing through the physical center and the points contained in each qualified probability similar line of each target qualified probability is regarded as the local distance between the boundary of the pile quality qualified point set and each qualified probability similar line of each target qualified probability in the corresponding straight line direction passing through the physical center.

[0170] Based on the local spacing between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target in all straight lines passing through the physical center, the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target are calculated.

[0171] Based on the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and all lines with similar qualified probabilities for each target, the matching degree between the boundary of the pile quality qualified point set and the qualified probability of each target is calculated.

[0172] Determine the matching weight of the pass probability of each target, and calculate the comprehensive matching degree of the pass probability of each target based on the matching weight of the pass probability of each target and the matching degree between the boundary of the pass point set of pile quality and the pass probability of each target.

[0173] Based on the boundary of all similar lines of the target qualification probability corresponding to the maximum comprehensive matching degree and the set of qualified points of pile quality, a polygonal qualification domain is divided in the two-dimensional diagnostic plane.

[0174] In this embodiment, outlining the set of qualified pile quality points to obtain the boundary of the qualified pile quality point set and marking the physical center of the boundary of the qualified pile quality point set refers to drawing the outer contour of the previously identified qualified pile quality point set to obtain the boundary of this point set. Then, by calculating the geometric center, such as by using the mean of the coordinates of the point set on a two-dimensional plane, the physical center of the shape enclosed by this boundary is determined. This center position is of great reference significance for subsequent analysis.

[0175] In this embodiment, based on the local spacing between the boundary of the pile quality qualified point set and the line of approximation of each target qualified probability along all straight lines passing through the physical center, the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and the line of approximation of each target qualified probability are calculated. This means that the distance between the boundary of the pile quality qualified point set and the line of approximation of each target qualified probability is measured along all straight lines passing through the previously marked physical center; these distances are the local spacings. Mathematical operations such as summation and averaging are performed on all these local spacings to obtain the comprehensive spacing, which reflects the overall distance between the two. Simultaneously, by analyzing the dispersion of these local spacings, such as by calculating the standard deviation, the spacing uniformity is obtained, thus measuring whether the distribution of these local spacings is uniform.

[0176] In this embodiment, the matching degree between the boundary of the pile quality qualified point set and the qualified probability of each target is calculated based on the comprehensive spacing and spacing uniformity of the boundary of the pile quality qualified point set and all lines with similar qualified probabilities. This is achieved by combining the comprehensive spacing and spacing uniformity obtained earlier with certain weights (e.g., the weight of the comprehensive spacing is 0.6 and the weight of the spacing uniformity is 0.4) to obtain a value that represents the matching degree between the boundary of the pile quality qualified point set and the qualified probability of each target. This matching degree reflects the similarity or conformity between the two.

[0177] In this embodiment, determining the matching weight of the pass probability of each target is to use the ratio of the pass probability of each target to the sum of the pass probabilities of all targets as the matching weight of the pass probability of that target.

[0178] In this embodiment, based on the matching weight of each target qualification probability and the matching degree between the boundary of the pile quality qualification point set and each target qualification probability, the comprehensive matching degree of each target qualification probability is calculated. This is achieved by multiplying the matching weight corresponding to each target qualification probability by the previously calculated matching degree, and finally obtaining the comprehensive matching degree of each target qualification probability. This comprehensive matching degree can more comprehensively and accurately reflect the relationship between the boundary of the pile quality qualification point set and different target qualification probabilities, providing a stronger basis for subsequent evaluation and decision-making.

[0179] like Figure 3 As shown, to further refine the division of the polygonal qualified domain, a method is proposed to divide the polygonal qualified domain in the two-dimensional diagnostic plane based on the boundary of all qualified probability similar lines corresponding to the target qualified probability with the maximum comprehensive matching degree and the boundary of the set of qualified points for pile quality. This includes:

[0180] The verification region is defined as all simply connected regions formed by the intersection of all lines with similar pass probabilities corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality.

[0181] Calculate the pass / fail value for each verification region based on the pass / fail probability of all dense grid points within each verification region;

[0182] The region with the most frequent overlap between the lines of similar pass probabilities corresponding to the target pass probability corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality is taken as the core pass domain.

[0183] All verification areas with valid verification values ​​not less than the preset verification value threshold are merged with the core valid area and treated as polygonal valid areas in the two-dimensional diagnostic plane.

[0184] In this embodiment, based on the pass probability of all dense grid points in each verification area, the pass verification value of each verification area is calculated. This is done by collecting the pass probabilities predicted by the integrated passability discrimination model of all dense grid points in each verification area, and averaging them using methods such as averaging and weighted averaging (averaging after assigning different weights according to the importance of the grid points) to obtain a value representing the overall passability of the verification area, namely the pass verification value, which is used to assess the probability of the pile quality of the verification area being qualified.

[0185] In this embodiment, the region with the most frequent overlap between the lines of similar pass probabilities corresponding to the target pass probability with the maximum comprehensive matching degree and the boundary of the pile quality pass point set refers to the target pass probability corresponding to the maximum comprehensive matching degree after calculating the comprehensive matching degree of each target pass probability. Then, all lines of similar pass probabilities for this target pass probability are determined, and these lines are compared with the boundary of the pile quality pass point set to find the region with the most overlaps. This region plays an important indicative role in judging the fit between the pile quality and the target pass probability.

[0186] In this embodiment, the preset verification value threshold is a pre-set standard value. The qualified verification value of each verification area is compared with it. If the qualified verification value reaches or exceeds the threshold, it indicates that the pile quality of the corresponding verification area is likely to be qualified. If it is lower than the threshold, it indicates that there may be a problem with the pile quality. This serves as a reference standard for judging whether the pile quality of the verification area meets the standard.

[0187] To provide detailed evidence for construction quality assessment, recommendations for adjusting construction parameters based on settlement risk tracing results are proposed, and a construction quality inspection report is generated, including:

[0188] Based on the results of settlement risk tracing, recommendations for adjusting construction parameters were determined, and actual construction parameters after adjustment were collected. At the same time, new pile quality indicators were calculated, and new settlement records were obtained.

[0189] Calculate the compliance rate of construction parameters based on the current actual construction parameters;

[0190] Calculate the pile quality pass rate based on the new pile quality indicators;

[0191] Calculate the settlement risk control rate based on the new settlement records;

[0192] A construction quality inspection report is generated based on the compliance rate of construction parameters, the pass rate of pile quality, and the settlement risk control rate.

[0193] In this embodiment, collecting the actual construction parameters after adjustment, calculating new pile quality indicators, and obtaining new settlement records means that after adjusting the construction parameters, real-time drilling speed, real-time grouting pressure, and real-time grout flow rate are collected again during actual construction. Based on these new parameters, new pile quality indicators such as pile diameter and pile uniformity are calculated according to predetermined formulas and methods. Simultaneously, new real-time settlement and vibration disturbance values ​​are collected using equipment installed on the roadbed surface and slope toe, forming new settlement records for subsequent comprehensive evaluation of the construction effect.

[0194] In this embodiment, calculating the compliance rate of construction parameters based on current actual construction parameters involves comparing the currently collected actual construction parameters, such as real-time drilling speed, real-time grouting pressure, and real-time grout flow rate, with pre-set standard parameter ranges. The number of actual parameters falling within the standard range is counted, and then divided by the total number of parameters. The resulting percentage is the compliance rate of construction parameters, used to measure the degree to which the actual construction parameters meet the standards. For example, if there are 3 parameters, and 2 of them are within the standard range, the compliance rate is approximately 2 ÷ 3 ≈ 66.7%.

[0195] In this embodiment, the pile quality pass rate is calculated based on new pile quality indicators. This involves comparing the newly calculated indicators, such as pile diameter and pile uniformity, with the corresponding pass standards. The number of pile quality indicators that meet the pass standards divided by the total number of indicators gives the pile quality pass rate, which is used to determine the proportion of piles that meet the quality requirements. Assuming there are two pile quality indicators, both of which meet the standards, the pile quality pass rate is 100%.

[0196] In this embodiment, the settlement risk control rate is calculated based on new settlement records. This involves analyzing the roadbed settlement based on newly acquired settlement records. By comparing these records with pre-set settlement risk control standards, the proportion of actual settlement within a controllable risk range is calculated; this is the settlement risk control rate, used to evaluate the effectiveness of settlement risk control. For example, a range of allowable settlement over a certain period is set, the number of times this range is recorded in the new settlement records is counted, and this count is divided by the total number of records to obtain the settlement risk control rate.

[0197] In this embodiment, a construction quality inspection report is generated based on the compliance rate of construction parameters, the pass rate of pile formation quality, and the settlement risk control rate. This report integrates the calculated compliance rate of construction parameters, the pass rate of pile formation quality, and the settlement risk control rate, and elaborates on whether the construction parameters meet the standards, the pass rate of pile formation quality, and the settlement risk control situation in detail according to a certain format and requirements. This forms a comprehensive inspection report that reflects the construction quality and provides a clear basis for construction quality assessment and subsequent decision-making.

[0198] This invention proposes an implementation method for a dynamic monitoring system of construction quality for inclined jet grouting pile reinforcement in the transition section of heavy-haul railways, oriented towards settlement control, including:

[0199] The construction record module is used to collect construction parameters and settlement records in real time during the construction of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section.

[0200] The noise reduction module is used to calculate pile quality indicators based on construction parameters and to calculate the settlement rate of layered subgrade based on settlement records.

[0201] The correlation analysis module is used to construct a correlation model of construction quality, pile quality, and settlement risk by taking construction parameters as the input layer, pile quality indicators as the intermediate layer, and roadbed settlement rate as the output layer, and using a multiple linear regression algorithm.

[0202] The risk tracing module is used to obtain the settlement risk tracing results by back-matching the synchronous parameters based on the correlation model when the settlement rate is determined to exceed the preset settlement rate threshold based on the initial settlement amount.

[0203] The quality monitoring module is used to determine construction parameter adjustment suggestions based on the settlement risk tracing results and generate construction quality inspection reports.

[0204] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in transition sections of heavy-haul railways for settlement control, characterized in that, include: Real-time acquisition of construction parameters and settlement records during the construction phase of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section. The pile quality index was calculated based on the construction parameters, and the settlement rate of the layered subgrade was calculated based on the settlement records. Using construction parameters as the input layer, pile quality indicators as the intermediate layer, and layered roadbed settlement rate as the output layer, a correlation model of construction quality, pile quality, and settlement risk is constructed using a multiple linear regression algorithm. When it is determined that the settlement rate of the layered roadbed exceeds the corresponding preset settlement rate threshold, the settlement risk source tracing result is obtained by back matching the synchronous parameters based on the correlation model. Based on the results of settlement risk tracing, recommendations for adjusting construction parameters are determined, and a construction quality inspection report is generated; The training process of the association model includes: The correlation coefficients between the intermediate layer and the input layer, and the correlation coefficients between the output layer and the intermediate layer are used as model parameters. The parameters are optimized through cross-validation so that the goodness of fit of the correlation model is not less than the preset goodness of fit threshold. Among them, the results of settlement risk tracing are obtained by back-matching contemporaneous parameters based on the correlation model, including: A construction parameter vector is generated based on the current construction parameters, and the construction parameter vector is standardized to obtain a standardized construction parameter vector. The standardized construction parameter vector is then input into the association model for forward propagation calculation to obtain the theoretical value of the pile quality index corresponding to the intermediate layer. The theoretical and actual pile quality indicators obtained continuously are fitted into a two-dimensional sequence data. The two-dimensional sequence data is dynamically time-normalized and the trend correlation is calculated. Based on the trend correlation, the theoretical and actual pile quality indicators obtained continuously are weighted and fused to obtain a comprehensive quality assessment vector. The deviation between the theoretical value and the actual value of the pile quality index is projected onto the two-dimensional diagnostic plane to obtain the diagnostic projection point. Based on the historical database, the two-dimensional diagnostic plane is divided into polygon qualified domains, and the shortest Euclidean distance from the diagnostic projection point to the boundary of the polygon qualified domain is calculated. When the shortest Euclidean distance is not less than the preset distance threshold, the stratum sensitivity coefficient matrix is ​​introduced to transform the coordinates of the comprehensive quality assessment vector, and the contribution of each pile quality index dimension to the comprehensive deviation index is calculated in the transformed space. Based on the contribution of each pile quality indicator dimension to the comprehensive deviation index, the dominant deviation indicators are identified among the pile quality indicators. Based on the dominant deviation indicators, risk tracing is carried out in the risk reasoning knowledge graph to obtain the settlement risk tracing results.

2. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control as described in claim 1, characterized in that, Real-time acquisition of construction parameters and settlement records during the construction phase of the inclined jet grouting piles, including: The drilling rig integrates sensors to collect real-time drilling speed, real-time grouting pressure, and real-time grout flow rate as construction parameters during the construction phase of the inclined jet grouting pile. Based on the settlement observation markers and vibration acceleration sensors installed on the roadbed surface and slope toe of the heavy-haul railway transition section, the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section are collected in real time as settlement records.

3. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control as described in claim 1, characterized in that, The pile quality index is calculated based on construction parameters, and the settlement rate of the layered subgrade is calculated based on settlement records, including: The pile diameter is calculated based on the real-time grouting pressure, real-time grout flow rate, and stratum correction coefficient in the construction parameters. The uniformity of the pile body was calculated by the wavelet packet energy entropy of the real-time drilling speed. Based on the vibration correction coefficient and the vibration disturbance value in the settlement record, the real-time settlement in the settlement record is corrected to obtain the real-time corrected settlement. Based on the initial depth of the roadbed surface and the toe of the slope, the subgrade surface layer, the middle layer, and the bottom layer are divided. The subgrade settlement rates for the surface layer, middle layer, and bottom layer of the subgrade were calculated based on the real-time corrected settlement amount and used as the layered subgrade settlement rates. Among them, the pile quality indicators include pile diameter and pile uniformity.

4. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control as described in claim 1, characterized in that, Based on historical databases, polygonal qualified regions are divided within the two-dimensional diagnostic plane, including: Extract the pile diameter and pile uniformity of all qualified construction stages from the historical database to form a set of qualified data points on the two-dimensional diagnostic plane, and extract the pile diameter and pile uniformity of all unqualified construction stages to form a set of unqualified data points on the two-dimensional diagnostic plane. A density-based spatial clustering algorithm is used to cluster the qualified data point set and identify the set of qualified pile quality points that are density-connected. An ensemble learning classification model is trained based on a set of qualified data points and a set of unqualified data points to obtain an ensemble qualification discrimination model. Based on the integrated qualification discrimination model, the dense grid points of the two-dimensional diagnostic plane are predicted, the qualification probability of the dense grid points of the two-dimensional diagnostic plane is output, and the qualification probability distribution map is generated based on the qualification probability of the dense grid points of the two-dimensional diagnostic plane. Based on the pass probability distribution map, multiple lines with similar pass probabilities are drawn in the two-dimensional diagnostic plane to divide the pass probability of each target. Based on the set of qualified pile quality points and multiple lines of similar qualification probabilities for all targets, a polygonal qualified domain is divided in the two-dimensional diagnostic plane.

5. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control, as described in claim 4, is characterized in that... Based on the set of qualified pile quality points and multiple lines of similarity in the probability of all target qualifications, a polygonal qualified domain is divided in the two-dimensional diagnostic plane, including: The set of qualified pile quality points is outlined to obtain the boundary of the qualified pile quality point set, and the physical center of the boundary of the qualified pile quality point set is marked. The distance between the boundary points of the pile quality qualified point set on each straight line passing through the physical center and the points contained in each qualified probability similar line of each target qualified probability is regarded as the local distance between the boundary of the pile quality qualified point set and each qualified probability similar line of each target qualified probability in the corresponding straight line direction passing through the physical center. Based on the local spacing between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target in all straight lines passing through the physical center, the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and the line of similarity to each qualified probability of each target are calculated. Based on the comprehensive spacing and spacing uniformity between the boundary of the pile quality qualified point set and all lines with similar qualified probabilities for each target, the matching degree between the boundary of the pile quality qualified point set and the qualified probability of each target is calculated. Determine the matching weight of the pass probability of each target, and calculate the comprehensive matching degree of the pass probability of each target based on the matching weight of the pass probability of each target and the matching degree between the boundary of the pass point set of pile quality and the pass probability of each target. Based on the boundary of all similar lines of the target qualification probability corresponding to the maximum comprehensive matching degree and the set of qualified points of pile quality, a polygonal qualification domain is divided in the two-dimensional diagnostic plane.

6. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control as described in claim 5, is characterized in that, Based on the boundary of all similar lines of pass probabilities corresponding to the target pass probability with the maximum comprehensive matching degree and the set of pass points of pile quality, a polygonal pass domain is divided in the two-dimensional diagnostic plane, including: The verification region is defined as all simply connected regions formed by the intersection of all lines with similar pass probabilities corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality. Calculate the pass / fail value for each verification region based on the pass / fail probability of all dense grid points within each verification region; The region with the most frequent overlap between the lines of similar pass probabilities corresponding to the target pass probability corresponding to the maximum comprehensive matching degree and the boundary of the set of pass points for pile quality is taken as the core pass domain. All verification areas with valid verification values ​​not less than the preset verification value threshold are merged with the core valid area and treated as polygonal valid areas in the two-dimensional diagnostic plane.

7. The method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement in heavy-haul railway transition sections for settlement control as described in claim 1, characterized in that, Based on the results of settlement risk tracing, recommendations for adjusting construction parameters are determined, and a construction quality inspection report is generated, including: Based on the results of settlement risk tracing, recommendations for adjusting construction parameters were determined, and actual construction parameters after adjustment were collected. At the same time, new pile quality indicators were calculated, and new settlement records were obtained. Calculate the compliance rate of construction parameters based on the current actual construction parameters; Calculate the pile quality pass rate based on the new pile quality indicators; Calculate the settlement risk control rate based on the new settlement records; A construction quality inspection report is generated based on the compliance rate of construction parameters, the pass rate of pile quality, and the settlement risk control rate.

8. A dynamic monitoring system for the construction quality of inclined jet grouting pile reinforcement in the transition section of heavy-haul railway for settlement control, characterized in that, A method for dynamic monitoring of construction quality of inclined jet grouting pile reinforcement for heavy-haul railway transition sections, oriented towards settlement control, as described in any one of claims 1 to 7, comprising: The construction record module is used to collect construction parameters and settlement records in real time during the construction of inclined jet grouting piles. The construction parameters include real-time drilling speed, real-time grouting pressure and real-time grout flow rate. The settlement records include the real-time settlement and vibration disturbance values ​​of the roadbed surface and slope toe of the heavy-haul railway transition section. The noise reduction module is used to calculate pile quality indicators based on construction parameters and to calculate the settlement rate of layered subgrade based on settlement records. The correlation analysis module is used to construct a correlation model of construction quality, pile quality and settlement risk by taking construction parameters as the input layer, pile quality index as the intermediate layer and layered roadbed settlement rate as the output layer, and using a multiple linear regression algorithm. The risk tracing module is used to obtain the settlement risk tracing results by back-matching the synchronous parameters based on the correlation model when the settlement rate is determined to exceed the preset settlement rate threshold based on the initial settlement amount. The quality monitoring module is used to determine construction parameter adjustment suggestions based on the settlement risk tracing results and generate construction quality inspection reports.

Images