A method and system for detecting the construction quality of a TRD

Abstract

The application relates to the technical field of wisdom civil engineering, in particular to a TRD construction quality detection method and system. The method comprises the following steps: acquiring real-time construction data; performing soil property analysis and construction state analysis according to the real-time construction data to obtain soil property data and construction state data; performing construction state change analysis according to the soil property data and the construction state data to obtain construction state change data; and performing construction parameter optimization according to the construction state change data to obtain construction parameter optimization data, so as to perform TRD construction quality detection auxiliary operation. The application realizes deep analysis of the coupling relationship between soil and the construction state by analyzing the relationship between the soil mechanical properties (such as density, sound velocity, porosity, etc.) and the construction equipment operation state (such as cutter load, rotation speed, stirring depth, etc.), so that the construction efficiency and quality are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wisdom civil engineering, and particularly relates to a TRD construction quality detection method and system. BACKGROUND

[0002] In modern civil engineering and foundation construction, the TRD (Trench cutting Re-mixing Deep wall) method is widely used in underground continuous walls, soft soil foundation reinforcement and seepage prevention engineering. The TRD method realizes the full mixing of soil and curing agent through the rotation cutting and stirring of the cutter head, and the construction quality directly affects the stability and durability of the project. Under complex geological environment (such as hard interlayer, soft soil layer, etc.), the dynamic changes of soil properties and construction state often affect the construction quality, and the traditional method cannot realize real-time sensing and adjustment of these changes. SUMMARY

[0003] The present application provides a TRD construction quality detection method and system to solve at least one of the above technical problems.

[0004] The present application provides a TRD construction quality detection method, which comprises:

[0005] S1, acquiring real-time construction data;

[0006] S2, analyzing soil properties and construction state according to the real-time construction data to obtain soil property data and construction state data;

[0007] S3, analyzing the construction state changes according to the soil property data and the construction state data to obtain construction state change data;

[0008] S4, optimizing the construction parameters according to the construction state change data to obtain construction parameter optimization data for TRD construction quality detection auxiliary operation.

[0009] The present application realizes real-time perception of the whole construction process, ensuring the comprehensiveness and real-time nature of the data. Soil property analysis can identify the geological conditions (such as density, humidity, porosity, etc.) of the construction area, providing accurate basis for construction state change analysis. Construction state analysis can real-time master the equipment operation state (such as cutter speed, feeding speed, etc.), effectively monitor the construction progress and warn potential problems. Through dynamic analysis of soil property data and construction state data, the state change trend in the construction process is accurately predicted. Potential risks (such as cutter overload, uneven mixing, etc.) can be predicted, providing a scientific basis for optimizing construction parameters. The forward-looking nature of construction management is improved, and the quality risks caused by state change lag response are reduced. Based on the construction state change data, intelligent algorithms are used to dynamically adjust the construction parameters (such as mixing depth, rotation speed, etc.), ensuring that the construction quality meets the design requirements. Avoid wasting resources (such as excessive mixing agents or equipment wear) caused by improper parameter settings, and improve construction efficiency. A closed-loop control mechanism is formed to realize the complete process of "detection-analysis-optimization-feedback".

[0010] Optionally, S1 comprises:

[0011] S11, real-time acquisition of original construction data including soil resistance, cutter rotation speed and depth displacement through multi-modal sensors integrated on the TRD equipment;

[0012] S12, obtaining geographically labeled construction data by fusing geographic information of the original construction data;

[0013] S13, obtaining construction real-time data by real-time multi-source data integration of the geographically labeled construction data.

[0014] In the present application, multi-modal sensors can comprehensively cover soil properties and construction equipment states, ensuring multi-dimensionality and accuracy of data acquisition. Through accurate labeling of geographic information, the correspondence between construction data and actual location is ensured, supporting regional analysis of geological properties. Real-time multi-source data integration solves the problem of inconsistent data types of multi-modal sensors, ensuring that the integrated construction real-time data can be directly called.

[0015] Optionally, S2 comprises:

[0016] S21, soil property feature extraction and construction state feature extraction are performed on the construction real-time data, respectively obtaining soil property feature data and construction state feature data;

[0017] S22, feature multi-dimensional correlation is performed according to the soil property feature data and the construction state feature data, obtaining construction feature relationship data;

[0018] S23, soil model generation is performed on the construction feature relationship data, obtaining a soil mechanics model;

[0019] S24, soil property extraction is performed on the soil mechanics model to obtain soil property data;

[0020] S25, construction state modeling is performed on the construction feature relationship data to obtain a construction state model;

[0021] S26, construction state extraction is performed on the construction state model to obtain construction state data.

[0022] In the present application, the soil physical properties and the construction equipment state in the construction environment are accurately identified through multi-dimensional feature extraction. The dynamic interaction between soil properties and construction states is revealed, providing a basis for the cause analysis of construction problems. Combined with the construction feature relationship, a high-precision soil mechanics model that is more in line with the actual construction is generated. Based on the soil mechanics model, the key mechanical properties of the soil, such as shear strength and stiffness modulus, are extracted through stress-strain analysis and dynamic response analysis. The construction state model can predict key states (such as changes in equipment load) in real time, reducing the risk of construction accidents. The output of the construction state model is analyzed to extract key state features (such as equipment operating efficiency, mixing depth, and rotation speed) during the construction process, generating construction state data.

[0023] Optionally, the soil property feature extraction includes:

[0024] First soil property feature extraction is performed on the construction real-time data to obtain first soil property feature data;

[0025] Second soil property feature extraction is performed on the construction real-time data to obtain second soil property feature data, wherein the first soil property feature extraction and the second soil property feature extraction are different soil property feature extraction methods;

[0026] The first soil property feature data and the second soil property feature data are fused to obtain soil property feature data.

[0027] In the present application, the first soil property feature extraction and the second soil property feature extraction are used to analyze the construction real-time data in different ways, ensuring the multi-dimensionality and completeness of the soil property feature data. The diversity and depth of soil property analysis are enhanced, avoiding feature omission or distortion that may be caused by a single method. Combined with multiple soil property feature extraction methods, the respective technical advantages are utilized to generate more accurate and reliable property data. The feature fusion process can correct errors generated by different methods, improving data consistency. The limitations of a single property extraction method are eliminated in the fusion process, providing a comprehensive description of soil properties.

[0028] Optionally, the construction real-time data includes soil vibration wave data, the soil vibration wave data is collected by a seismic wave sensor or a high-frequency vibration sensor, and the first soil characteristic feature extraction includes:

[0029] Perform time domain characteristic analysis on the soil vibration wave data in the construction real-time data to obtain soil time domain characteristic data;

[0030] Perform frequency domain data variation on the soil time domain characteristic data to obtain soil frequency domain characteristic data;

[0031] Perform energy spectrum decomposition on the soil frequency domain characteristic data to obtain soil spectrum decomposition data;

[0032] Perform dynamic response parameter mapping on the soil spectrum decomposition data to obtain first soil characteristic feature data;

[0033] The energy spectrum decomposition includes:

[0034] Perform time-frequency joint transformation on the soil time domain characteristic data to obtain local time-frequency characteristic data;

[0035] Perform spectrum energy distribution extraction on the local time-frequency characteristic data to obtain spectrum energy distribution data;

[0036] Perform multi-scale spectrum decomposition on the spectrum energy distribution data to obtain multi-scale spectrum data;

[0037] Perform spectrum correlation analysis on the multi-scale spectrum data to obtain spectrum correlation matrix data;

[0038] Perform energy characteristic clustering processing on the spectrum correlation matrix data to obtain soil spectrum energy clustering characteristics;

[0039] Perform feature spectrum reconstruction on the soil spectrum energy clustering characteristics to obtain soil reconstructed spectrum data;

[0040] Generate soil spectrum decomposition data according to the soil reconstructed spectrum data to obtain soil spectrum decomposition data.

[0041] The present application provides multi-dimensional analysis of soil dynamic characteristics and frequency domain energy characteristics, providing reliable data support for construction state adjustment and optimization. Through multi-scale spectrum decomposition and dynamic response parameter mapping of soil vibration wave data, the response characteristics of soil under complex construction conditions can be accurately identified. The adaptability of TRD construction under complex geological conditions is improved, and the construction error is reduced. The output soil characteristic data directly serves the intelligent construction model, providing a basis for whole-process monitoring and dynamic optimization of construction quality. Dynamic parameter mapping and feature data reconstruction reduce data noise and errors, making construction state prediction and parameter optimization more accurate. Through real-time monitoring and efficient data processing, construction risks are reduced, and overall construction efficiency is improved.

[0042] The time-domain data is converted into time-frequency distribution in the application, and the change rule of vibration wave characteristics with time and frequency is revealed. Local time-frequency characteristics that change dynamically are provided to support fine-grained analysis. The energy density at different frequencies is accurately calculated to quantify the response strength of the soil on the vibration frequency spectrum. Direct support is provided for identification and characteristic modeling of high-energy frequency bands. The energy characteristics of soil vibration waves are separated into high, medium, and low frequencies to provide hierarchical dynamic characteristic description. Data redundancy or frequency information loss in characteristic analysis is avoided. The correlation and coupling characteristics between frequency components are revealed to provide structured information for frequency domain modeling. Cross-band characteristic fusion is supported to improve the depth of data integration. The spectral energy characteristics are classified to identify the relationship between specific frequency bands and geological characteristics (such as faults and fractures). The clustering results directly serve the partition analysis and construction state adjustment of the construction area. The spectral atlas is reconstructed based on the clustering characteristic data to generate clearer energy distribution characteristics. The reconstructed atlas provides visual and intuitive support for construction state prediction.

[0043] Optionally, the construction real-time data includes soil acoustic wave data, the soil acoustic wave data is obtained by collecting through an acoustic wave sensor, and the second soil characteristic feature extraction includes:

[0044] Multi-band acoustic wave transmission analysis is performed on the soil acoustic wave data in the construction real-time data to obtain soil acoustic wave transmission data;

[0045] Sound velocity calculation is performed on the soil acoustic wave transmission data to obtain soil sound velocity data;

[0046] Acoustic wave reflection characteristic extraction is performed on the soil sound velocity data to obtain soil reflection characteristic parameter data;

[0047] Acoustic wave attenuation analysis is performed on the soil emission characteristic parameter data to obtain soil acoustic wave attenuation data;

[0048] Acoustic wave propagation simulation is performed on the soil acoustic wave attenuation data to obtain soil acoustic wave propagation data;

[0049] Soil characteristic parameter inversion is performed according to the soil acoustic wave propagation data to obtain second soil characteristic feature data.

[0050] The different frequency band acoustic wave data are analyzed in the application, and the transmission time, amplitude and phase change of each frequency band are extracted. According to the acoustic wave transmission path length and time difference, the acoustic wave propagation speed of the soil is calculated, and the acoustic velocity is calculated in real time by combining the environmental parameters, so as to accurately reflect the acoustic behavior of the soil under different depth, humidity and pressure conditions. By analyzing the reflection characteristics of the acoustic wave at different interfaces (such as cracks, soil layer junctions), the reflection coefficient and reflection amplitude are calculated. The energy attenuation coefficient of the acoustic wave during transmission in the soil medium is calculated, the absorption capacity of the soil to different frequency band acoustic waves is analyzed according to the frequency characteristics, the energy loss of the soil medium to the acoustic wave is evaluated in real time, and the absorption characteristics of the soil are quantified. The propagation path, waveform change and energy distribution of the acoustic wave in the soil are simulated, and the internal structure of the soil is intuitively reflected. Combined with the acoustic wave propagation data, the mechanical and physical characteristics of the soil are extracted through the inversion algorithm (such as full waveform inversion, Bayesian optimization), and high-precision soil density, elastic modulus, porosity and other key characteristics are generated through the inversion algorithm.

[0051] Optionally, wherein the soil reflection characteristic parameter data includes acoustic wave scattering characteristic data, and the extraction method of the acoustic wave scattering characteristic data includes:

[0052] According to the construction real-time data, a soil physical model is constructed to obtain discrete soil grid data;

[0053] According to the discrete soil grid data, acoustic wave source parameter labeling is performed to obtain acoustic wave source labeling data;

[0054] The acoustic wave source labeling data and the soil acoustic velocity data are subjected to acoustic wave scattering simulation to obtain scattering sound field distribution data;

[0055] The scattering intensity and directivity of the scattering sound field distribution data are analyzed to obtain the acoustic wave scattering characteristic data.

[0056] In the application, based on the construction real-time data (such as acoustic wave data, geological survey information), a discrete soil grid model is constructed, which contains physical parameters such as the density, porosity and humidity of the soil, supports the modeling of layered soil and heterogeneous medium, and adapts to the needs of different construction environments. The acoustic wave source parameters are labeled in the soil grid model, including the acoustic wave source position, frequency range, waveform type (such as sine wave, pulse wave) and the like. The propagation and scattering behavior of the acoustic wave in the soil is simulated to generate scattering sound field distribution data, and the reflection, scattering and diffraction phenomena caused by the acoustic wave encountering cracks, obstacles and the like in the soil are reproduced. The scattering intensity and directivity are calculated by quantitatively analyzing the scattering sound field distribution data to generate a scattering characteristic spectrum. The acoustic wave propagation path and energy loss are identified by using the spatial distribution and frequency response of the scattering sound field.

[0057] Optionally, S3 includes:

[0058] S31, construction state change processing is performed according to the soil characteristic data and the construction state data, and construction state change trend data is obtained;

[0059] S32, abnormal isolated points are extracted according to the construction state change trend data, and construction abnormal isolated point data is obtained;

[0060] S33, abnormal classification is performed according to the construction abnormal isolated point data, and construction abnormal classification data is obtained;

[0061] S34, change cause analysis is performed on the construction abnormal classification data, and construction change cause data is obtained;

[0062] S35, the construction state change trend data, the construction abnormal isolated point data and the construction change cause data are integrated, and construction state change data is obtained.

[0063] In the present application, the soil characteristic data and the construction state data are combined to generate the construction state change trend data, the dynamic change characteristics of the soil and the equipment state in the construction process are identified, and real-time adjustment is supported. The abnormal detection algorithm (such as the isolated forest algorithm and the DBSCAN clustering) is used to extract the construction abnormal isolated points from the trend data, the abnormal points deviating from the trend are efficiently identified, and potential problems are located. The construction abnormal isolated points are classified, different types of abnormalities are identified, the abnormalities are divided into equipment abnormalities, soil characteristic abnormalities and construction process abnormalities, etc., targeted support is provided for analysis and solution. The cause and effect relationship of the abnormal classification data is deduced, the deep reasons for the construction state change are identified, the cause and effect relationship behind the abnormality is revealed, and the decision maker is helped to accurately locate the abnormal cause. The unified presentation of the trend, the abnormality and the cause and effect analysis result is realized, and the data result of the construction state change is formed.

[0064] Optionally, S4 comprises:

[0065] S41, construction optimization target data is obtained by generating an optimization target according to the construction state change data;

[0066] S42, construction optimization variable data is obtained by selecting an optimization variable according to the construction optimization target data;

[0067] S43, construction parameter optimization data is obtained by performing target optimization on the construction state change data according to the construction optimization variable data, so as to perform TRD construction quality detection auxiliary operation.

[0068] This invention utilizes in-depth analysis of construction status change data and optimization target generation to selectively adjust construction variables and optimize construction parameters, ensuring that construction quality meets design requirements. Combining real-time construction status change data with optimization target and variable selection enables rapid adaptation to complex or dynamic construction environments, enhancing the adaptability of construction equipment and processes. The precise optimization process reduces problems such as excessive equipment load and material waste, improving construction efficiency while ensuring high-quality construction results. By optimizing variable selection and parameter adjustment, unnecessary resource consumption and equipment wear are reduced. Scientific optimization lowers the risk of construction accidents or failures, ensuring construction safety.

[0069] Optionally, the present invention also provides a TRD construction quality inspection system for performing the TRD construction quality inspection method described above, the TRD construction quality inspection system comprising:

[0070] The construction real-time data acquisition module is used to acquire real-time construction data;

[0071] The real-time construction analysis module is used to perform soil property analysis and construction status analysis based on real-time construction data, and to obtain soil property data and construction status data.

[0072] The construction status change processing module is used to analyze construction status changes based on soil property data and construction status data to obtain construction status change data.

[0073] The construction parameter optimization module is used to optimize construction parameters based on construction status change data to obtain optimized construction parameter data for auxiliary TRD construction quality inspection.

[0074] The purpose of this invention is to ensure the comprehensiveness and accuracy of construction data by real-time acquisition of multi-dimensional data such as soil resistance, cutterhead rotation speed, and depth displacement through a multi-modal sensor network. It extracts parameters such as soil dynamics, density, and porosity. The invention dynamically monitors the status of construction equipment (e.g., cutterhead load, feed speed) to generate construction status data, reflecting the actual situation of equipment operation and construction technology in real time. It correlates soil characteristic data with construction status data in multiple dimensions to generate construction characteristic relationship data, providing input for status change analysis. The combination of soil mechanics model and construction status model improves the accuracy of construction characteristic modeling. It extracts and classifies anomalies from changing trends to accurately locate the types of construction anomalies (e.g., equipment failure, geological problems). Through causal analysis technology, it reveals the underlying causes of anomalies, providing a scientific basis for problem tracing and construction decision-making. Optimization algorithms are used to optimize the construction status data, dynamically adjusting construction parameters to ensure optimal construction quality and efficiency. Attached Figure Description

[0075] Other features, objects, and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments thereof, when read in connection with the following accompanying drawings:

[0076] Figure 1 A step flow chart of a TRD construction quality detection method of an embodiment is shown;

[0077] Figure 2 A step flow chart of a construction real-time data acquisition method of an embodiment is shown;

[0078] Figure 3 A step flow chart of a construction real-time analysis method of an embodiment is shown;

[0079] Figure 4 A step flow chart of a construction state change processing method of an embodiment is shown;

[0080] Figure 5 A step flow chart of a construction parameter optimization method of an embodiment is shown;

[0081] The implementation of the present application, functional features and advantages will be further described with reference to the accompanying drawings. DETAILED DESCRIPTION

[0082] The technical method of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0083] In addition, the accompanying drawings are only schematic illustrations of the present application, and are not necessarily drawn to scale. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0084] It should be understood that although the terms "first", "second", etc. can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the example embodiments, a first element can be called a second element, and similarly a second element can be called a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0085] Referring to Figures 1 to 5 The present application provides a TRD construction quality detection method, which comprises:

[0086] S1, acquiring construction real-time data;

[0087] In an embodiment, various sensors are installed on the construction equipment, including pressure sensors, acceleration sensors, acoustic wave sensors, and displacement sensors. Real-time acquisition of key data such as soil resistance, equipment vibration, acoustic wave propagation characteristics, and cutter depth. The collected data is transmitted to the central control system through the wireless transmission module, combined with the geographic information of the construction area, to realize the complete recording and spatial annotation of the construction real-time data.

[0088] In an embodiment, force sensors, acceleration sensors, displacement sensors, and pressure sensors are installed on the construction equipment (such as chain saws and cutter heads) to monitor the data in real time during equipment operation. In the soil of the construction area, buried conductivity detectors and moisture sensors are used to monitor the conductivity and humidity of the soil in real time. Through the industrial Internet of Things (IoT) gateway, the data collected by the sensors is transmitted to the field monitoring terminal in real time. The data includes time stamp, cutter position, soil conductivity, equipment applied pressure, and soil humidity. The collected data is denoised (such as using the moving average method), and the cleaned data set is obtained. The data is stored in groups and blocks according to time periods.

[0089] S2, soil property analysis and construction state analysis based on construction real-time data, to obtain soil property data and construction state data;

[0090] In an embodiment, in the control system, real-time data is processed for soil property analysis and construction state analysis: soil property analysis is performed through data correlation to evaluate soil density, humidity distribution, and resistance changes at different depths, and to generate a soil layered profile. Construction state analysis uses the vibration signal and applied pressure of the equipment to determine whether the operation of the construction equipment is normal, such as whether the equipment is overloaded or the soil resistance is abnormally increased.

[0091] In an embodiment, based on soil conductivity σ i and humidity H i , the soil density ρ is evaluated using the following formula: ρ = k1 × σ i + k2 × H i , where i is the time sequence term, k1 and k2 are empirical coefficients. According to the real-time soil property data, a soil profile is generated to represent the non-uniformity of the soil. Based on the mechanical feedback data F i and acceleration a i of the construction equipment, the force and motion state of the equipment are calculated: τ = F i / A, A is the contact area of the equipment cutter head. Combined with the equipment position data P i(x, y, z), to determine whether the construction state is normal, for example, whether overpressure or equipment jam occurs.

[0092] S3, construction state change analysis is performed according to the soil property data and the construction state data, to obtain construction state change data;

[0093] In an embodiment, the control system performs dynamic change analysis on the soil property data and the construction state data: the system identifies the change trend of the soil resistance and the equipment state in the construction process through a data trend analysis method, and draws a real-time state curve. At the same time, through an anomaly detection method, abnormal isolated points in the construction process are extracted, for example, sudden resistance increase or severe change of equipment vibration. The system performs causal analysis on the context data of the abnormal points, and marks the soil properties (such as local high compaction) or equipment parameters (such as pressure exceeding the normal range) that cause the abnormality.

[0094] In an embodiment, the construction state data (t i-1 , τ i-1 , a i-1 ) of the previous moment t i-1 is compared with the state (t i , τ i , a i ) of the current moment t i , the change rate is calculated: Δτ = (τ i - τ i-1 ) / (t i - t i-1 ), Δa = (a i - a i-1 ) / (t i - t i-1 ), a construction state change curve is generated, and it is detected whether an abnormality occurs in the construction process, such as sudden stress increase or acceleration fluctuation. The threshold range Δτ max - Δτ min and Δa max - Δa min are defined, when Δτ or Δa exceeds the threshold, the abnormality is marked, and the relevant time and position are recorded.

[0095] S4, construction parameter optimization is performed according to the construction state change data, to obtain construction parameter optimization data, for TRD construction quality detection auxiliary operation.

[0096] In an embodiment, the system sets construction optimization goals based on the construction state change data, such as maximizing construction efficiency, minimizing energy consumption, reducing abnormal occurrences, etc. In combination with the capacity constraints of the construction equipment, the system optimizes and adjusts the rotation speed of the cutter head, the pressure exerted by the equipment, and the progress speed: after the optimization goal is determined, the system preferentially selects variables that have a significant impact on the goal, such as adjusting the rotation speed of the cutter head and the size of the thrust of the equipment. The system calculates new equipment operating parameters through an optimization algorithm and feeds them back to the construction equipment. The construction equipment operates according to the optimized parameters, and real-time monitoring of the construction data verifies the effectiveness of the optimization scheme, significantly improving the construction efficiency while reducing the occurrence of equipment overload and abnormal shutdown.

[0097] In an embodiment, a multi-dimensional optimization objective function is constructed based on soil property data p, construction state change data At, a, and construction parameters such as cutter head speed v, cutting angle q, and pressure F: where E(v, q) is an energy consumption model or energy consumption relationship mathematical expression, describing the energy consumption per unit time. S(F) is a construction stability model or construction stability mathematical expression, describing the relationship between abnormal rate and pressure. C1 is the energy consumption model weight coefficient, and C2 is the construction stability model weight coefficient. Based on historical data, the construction parameters are adjusted to gradually minimize the objective function. By feeding back the optimized parameters, the operating state of the equipment is updated. The optimized construction parameters (v * , q * , F * ) are obtained and fed back to the construction equipment to guide subsequent construction operations.

[0098] Optionally, S1 includes:

[0099] S11, real-time acquisition of original construction data including soil resistance, cutter head rotation speed, and depth displacement through multi-modal sensors integrated on the TRD equipment;

[0100] In an embodiment, the following sensors are installed on the TRD equipment: soil resistance sensor: real-time measurement of the resistance when the cutter head contacts the soil, recording information such as soil density. Rotation speed sensor: monitors the rotation speed of the cutter head for evaluating the efficiency of the equipment. Depth displacement sensor: measures the real-time position of the cutter head to accurately record the construction depth. During the construction process, these sensors collect data at fixed time intervals to form a stream of original construction data containing resistance values, rotation speeds, and depths.

[0101] S12, geographic information fusion of the original construction data to obtain geolabeled construction data;

[0102] In an embodiment, the construction equipment is equipped with a high-precision GPS positioning module, which is used to record the three-dimensional spatial position of the construction equipment in real time. The collected raw construction data is associated with the position information of the equipment, and geographical coordinates (such as latitude, longitude and depth) are added to each set of data. Through geographic information fusion, construction data with location labels is generated to accurately represent the construction status of different spatial points.

[0103] S13, real-time multi-source data integration is performed on the geolabeled construction data to obtain construction real-time data.

[0104] In an embodiment, the geolabeled data is integrated with multi-modal data collected by sensors under a unified time axis. Through timestamp alignment method, the data of soil resistance, rotation speed and depth displacement are synchronized, and real-time multi-source construction data is generated. The integrated construction real-time data can be further analyzed for soil property evaluation and construction state monitoring.

[0105] Optionally, S2 includes:

[0106] S21, soil property feature extraction and construction state feature extraction are performed on the construction real-time data to obtain soil property feature data and construction state feature data, respectively;

[0107] In an embodiment, the construction real-time data includes soil resistance, acoustic wave propagation characteristics, cutter depth displacement and equipment operating parameters. Using resistance data and acoustic wave characteristics, soil density, moisture and layering characteristics at different depths are extracted, and a soil layering profile is generated. Combined with the rotation speed and acceleration of the equipment, the dynamic parameters of the equipment operation are extracted, such as cutter rotation speed, equipment force, resistance change rate and torque change rate, to monitor whether the construction equipment is in normal working state.

[0108] S22, feature multi-dimensional correlation is performed according to the soil property feature data and the construction state feature data to obtain construction feature relationship data;

[0109] In an embodiment, the soil property feature data and the construction state feature data are jointly analyzed, and by matching the spatial position and time point of the data, the correlation between soil property changes and equipment operating state is found. For example, when the soil density is found to be higher in some areas, the rotation speed of the equipment decreases significantly, forming a set of construction feature relationship data with physical correlation.

[0110] In an embodiment, more importantly, the spatial alignment is performed according to the soil characteristic feature data and the construction state feature data to obtain construction feature alignment data; the physical feature screening is performed on the construction feature alignment data to obtain construction feature screening data; the maximum information coefficient calculation is performed on the construction feature screening data to obtain construction feature correlation data; the dynamic lag window analysis is performed on the construction feature correlation data to obtain construction feature association data; and the high-dimensional graph mapping is performed according to the construction feature association data to obtain construction feature relationship data.

[0111] According to the real-time position (such as the depth and horizontal position) of the construction equipment, the position matching is performed on the two groups of data. In the soil characteristic region with sparse data, the interpolation method is used to complete the missing data points, so that the construction state data can be aligned with the soil characteristic data at each spatial point. The features with significant physical association with the construction target (such as efficiency and abnormal rate) are preferentially retained. The features with large noise or redundant features are removed. The statistical analysis is performed on the distribution of each feature variable, and the features with abnormal data distribution or too small variation range (such as low correlation of soil sound velocity change) are removed. According to the dynamic characteristics of the construction equipment and the physical properties of the soil, the core features are selected: equipment parameters: cutter head rotation speed, advance speed, pressure. Soil parameters: resistance gradient, porosity, density. The maximum information coefficient (MIC) between each feature is calculated using the construction feature screening data: the MIC value is calculated for each pair of features (such as soil density and rotation speed, porosity and pressure). The MIC value ranges from 0 to 1, and the closer the value is to 1, the stronger the association between the two features. For example, it is found that the MIC value of the soil resistance gradient and the advance speed is high, indicating that there is a strong association between them. For each pair of features, a time window (such as 3 seconds, 5 seconds) is set to analyze the delay effect of the change of one feature on the other feature. The time window length is dynamically adjusted to find the lag time corresponding to the maximum correlation. For example, it is found that the change of the cutter head rotation speed has a 1-2 second lag on the advance speed, indicating that the construction equipment needs a certain reaction time for operation adjustment. According to the construction feature association data, the association relationship is mapped to a high-dimensional space to generate a feature relationship graph: an association graph is constructed, and each feature is taken as a node in the graph. According to the correlation strength between the features, a weighted edge is added, and the weight of the edge is determined by the MIC value and the lag analysis result. The complex relationship between the features is displayed through the high-dimensional graph: the edges between the nodes reflect the direct association of the features. The edges with larger weights reveal the key relationships in the construction process (such as the interaction between the resistance gradient and the cutter head rotation speed).

[0112] S23, soil model generation is performed on the construction feature relationship data to obtain a soil mechanics model;

[0113] In an embodiment, based on the construction feature relationship data, a soil mechanics model suitable for the construction area is generated through finite element analysis and finite element calculation, combined with soil resistance and density characteristics. The model can describe the bearing capacity and resistance variation characteristics of the soil at different depths.

[0114] S24, soil property extraction is performed on the soil mechanics model to obtain soil property data;

[0115] In an embodiment, using the soil mechanics model, the parameters in the model are further extracted to generate specific soil property data, including: soil density at different depths; resistance characteristics of each layer of soil; dynamic response characteristics of the soil.

[0116] S25, construction state modeling is performed on the construction feature relationship data to obtain a construction state model;

[0117] In an embodiment, based on the construction feature relationship data, a construction state model is established combined with the dynamic parameters of the equipment. The model can dynamically describe the relationship between the equipment operating parameters (such as rotation speed, pressure) and the construction efficiency, and at the same time reflect the potential triggering mechanism of construction anomalies. For example, a multiple linear regression model is used to describe a simple linear relationship. If the data presents nonlinear characteristics, a nonlinear regression model (such as a polynomial model) or a time series model is selected. Alternatively, in the case of considering the complex interaction between multiple variables, a tree-based model (such as random forest or gradient boosting tree) or a neural network model is used.

[0118] S26, construction state extraction is performed on the construction state model to obtain construction state data.

[0119] In an embodiment, using the construction state model, key state data in the construction process is extracted, such as: equipment operation stability; construction efficiency in different time periods; spatiotemporal distribution of potential anomalies (such as excessive equipment load or unstable operation).

[0120] Optionally, wherein the soil property feature extraction comprises:

[0121] First soil property feature extraction is performed on the construction real-time data to obtain first soil property feature data;

[0122] In an embodiment, the construction equipment collects soil resistance data of the cutterhead at different depths through the resistance sensor and the depth sensor. The system uses these data to perform first soil property feature extraction. In the construction area, the resistance values at different depths are analyzed to evaluate the change of soil density with depth. According to the gradient change of resistance, the stratification characteristics of the soil are identified, and stratified soil profile data are generated, so as to obtain first soil property feature data, which contains the density and stratification information of each layer of soil, and provides a basis for the mechanical property analysis of the soil.

[0123] Second soil property feature extraction is performed on the construction real-time data to obtain second soil property feature data, wherein the first soil property feature extraction and the second soil property feature extraction are different soil property feature extraction methods.

[0124] In an embodiment, the construction equipment also collects soil acoustic wave propagation data through the integrated acoustic wave sensor. The system uses the acoustic wave propagation characteristics to perform second soil property feature extraction. The acoustic wave propagation speed at different depths and positions is analyzed to evaluate the uniformity and porosity of the soil. Based on the reflection and attenuation characteristics of the acoustic wave, abnormal areas (such as high-density particles or voids) existing in the soil are further identified. The generated second soil property feature data reflects the acoustic characteristics and microstructure information of the soil, and supplements the results of the first method.

[0125] The first soil property feature data and the second soil property feature data are fused to obtain soil property feature data.

[0126] In an embodiment, the feature data obtained by the first and second extraction methods are fused, and the two groups of feature data are spatially aligned according to depth and position. The mechanical properties (resistance and density) and acoustic properties (sound speed and attenuation) of the soil are integrated to generate unified soil property feature data. The fused soil property feature data contains the mechanical properties, acoustic properties and stratification information of the soil, and provides more comprehensive basic data for construction parameter optimization and construction quality evaluation.

[0127] Optionally, wherein the construction real-time data includes soil vibration wave data, the soil vibration wave data is collected by a seismic wave sensor or a high-frequency vibration sensor, and the first soil property feature extraction includes:

[0128] Time domain characteristic analysis is performed on the soil vibration wave data in the construction real-time data to obtain soil time domain feature data.

[0129] In an embodiment, the soil vibration data is recorded in real-time by sensors, which contains the response signal of the soil under the action of the vibration source. The time variation of the vibration signal is statistically analyzed to extract the average value, peak value, and fluctuation amplitude of the signal. These time-domain characteristic data reflect the response strength of the soil to vibration and the overall characteristics of the signal waveform, such as the attenuation trend of the amplitude over time.

[0130] The time-domain characteristic data of the soil is subjected to frequency-domain data variation to obtain soil frequency-domain characteristic data;

[0131] In an embodiment, based on the time-domain characteristic data of the soil, the construction team performs frequency-domain analysis on the vibration signal, converts the time-domain signal to the frequency-domain signal, and extracts the amplitude and energy distribution at different frequencies. By observing the frequency spectrum, the main frequency components and energy concentration areas are identified, reflecting the frequency response characteristics of the soil.

[0132] The frequency-domain characteristic data of the soil is subjected to energy spectrum decomposition to obtain soil spectrum decomposition data;

[0133] In an embodiment, the frequency-domain data of the soil is subjected to energy spectrum decomposition, the energy distribution of the frequency-domain signal is divided into different frequency bands, and the energy contribution of each frequency band is calculated. Through the decomposition process, the system identifies that certain frequency bands have a larger contribution to the overall energy, and these frequency bands may correspond to special characteristics in the soil structure, such as dense regions or void regions.

[0134] The soil spectrum decomposition data is subjected to dynamic response parameter mapping to obtain first soil characteristic feature data;

[0135] In an embodiment, based on the energy spectrum decomposition data, the system further performs dynamic response analysis to extract the dynamic response parameters of the soil at specific frequencies, such as damping characteristics and vibration modulus. By mapping these parameters to the soil model of the construction area, the soil dynamic response distribution at different positions is generated.

[0136] The energy spectrum decomposition includes:

[0137] The time-domain characteristic data of the soil is subjected to time-frequency joint transformation to obtain local time-frequency characteristic data;

[0138] In an embodiment, the vibration signal of the soil during construction is collected, and the system performs time-frequency joint transformation on the time-domain signal to combine the time-domain characteristics and frequency-domain characteristics of the soil vibration signal, generating local features of the vibration signal on the time and frequency axes. The local time-frequency characteristic data represents the response of the soil to different frequency vibrations at different time periods, laying a foundation for energy distribution analysis.

[0139] The local time-frequency characteristic data is subjected to frequency spectrum energy distribution extraction to obtain frequency spectrum energy distribution data;

[0140] In an embodiment, based on the local time-frequency feature data, the system analyzes the spectral energy distribution of the soil vibration signal. The energy proportion of different frequency bands is extracted to determine the main energy concentration area of the vibration wave signal. Through this distribution extraction, the system can identify the absorption and transmission characteristics of the soil to different frequency vibrations.

[0141] Perform multi-scale spectral decomposition on the spectral energy distribution data to obtain multi-scale spectral data;

[0142] In an embodiment, the system further refines the spectral energy distribution data and performs multi-scale spectral decomposition. The spectrum is divided into multiple scale ranges (such as low frequency, medium frequency and high frequency), and the energy contribution of each scale is calculated respectively. Multi-scale decomposition reflects the response ability of soil to vibration of different frequency ranges, especially the difference in deep or shallow soil.

[0143] Perform spectral correlation analysis on the multi-scale spectral data to obtain spectral correlation matrix data;

[0144] In an embodiment, through the analysis of multi-scale spectral data, the system calculates the correlation between different frequency bands. By comparing the energy changes of each frequency band, the mutual influence of them in space or time is evaluated. Correlation analysis reveals the dynamic connection between different regions of soil, providing support for understanding the overall vibration behavior of soil.

[0145] Perform energy characteristic clustering processing on the spectral correlation matrix data to obtain soil spectral energy clustering features;

[0146] In an embodiment, the spectral correlation matrix data is subjected to clustering analysis. By grouping regions with similar energy characteristics in the spectrum into a class, representative soil spectral energy clustering features are generated. The clustering results reveal the regional characteristics in the soil, such as some regions showing high energy absorption or low energy propagation.

[0147] Reconstruct the feature map of the soil spectral energy clustering features to obtain soil reconstructed spectral data;

[0148] In an embodiment, the spectral feature map of the soil is reconstructed using the clustering results. The results of spectral energy clustering are mapped to the spatial position of the construction area to generate the reconstructed spectral map of the soil. The map clearly shows the distribution of the spectral characteristics of the soil, providing intuitive guidance for complex problems that occur during construction.

[0149] Generate soil spectral decomposition data based on the soil reconstructed spectral data to obtain soil spectral decomposition data.

[0150] In an embodiment, from the soil reconstruction spectrum data obtained from the pre-analysis, the energy distribution, directionality and frequency response characteristics of each frequency band are extracted: spectral energy distribution: records the energy proportion of different frequency bands, reflecting the soil's absorption or reflection characteristics of sound waves in each frequency range. Directionality distribution: describes the distribution of sound energy in different propagation directions, revealing the non-uniformity inside the soil. Main frequency range: determines the resonance frequency and high-energy concentration frequency band of the soil.

[0151] According to the input soil reconstruction spectrum data, the system performs detailed decomposition of the spectral characteristics of the soil, such as frequency band division: dividing the frequency range into low, medium and high frequency regions, and selecting the frequency band range according to the physical properties of the soil (such as low frequency for coarse-grained soil and high frequency for fine-grained soil); segment energy calculation: calculating the total energy and energy proportion of each frequency band respectively to identify the absorption intensity of the soil to sound waves of each frequency band; local characteristic extraction: separately analyzing the spectral characteristics of a specific region (such as a region with strong sound wave reflection) to extract the special response of the local frequency band.

[0152] Further analysis of the dynamic behavior of the spectral characteristics with frequency, such as response curve fitting: using the reconstructed spectrum data to fit the dynamic response curve of the soil, showing the trend of the frequency response characteristics of the soil with frequency. Key characteristic parameters extraction, such as resonance frequency: the frequency at which the soil responds most strongly to sound waves, indicating the characteristic scale of the internal structure of the soil; energy attenuation rate: the attenuation rate of energy in different frequency bands with propagation distance, used to evaluate the sound absorption capacity of the soil; directionality parameters: changes in spectral characteristics in different propagation directions, used to identify non-uniform distribution inside the soil.

[0153] Correlate the decomposed spectral characteristics with the physical properties of the soil, such as the relationship between soil density and spectral energy: soil with higher density usually shows stronger low-frequency energy, while soil with lower density shows a higher proportion of high-frequency energy; the relationship between soil particle distribution and frequency band characteristics: coarse-grained soil tends to propagate at low frequencies, while fine-grained soil has better propagation characteristics at high frequencies; acoustic characteristics of abnormal regions: abnormal energy distribution is detected in certain regions, which may indicate cavities or changes in soil density.

[0154] Based on the above analysis results, soil spectrum decomposition data is generated, including segment energy distribution: energy proportion and total energy of each frequency band; dynamic response characteristics: dynamic response curve of the soil to sound waves, including resonance frequency, energy attenuation rate and directionality distribution; local characteristic description: spectral characteristic data of a specific region, used to identify abnormal regions and non-uniform soil distribution.

[0155] Optionally, the construction real-time data includes soil acoustic data, the soil acoustic data is obtained by an acoustic sensor, and the second soil characteristic feature extraction includes:

[0156] The soil sound wave data in the construction real-time data is subjected to multi-band sound wave transmission analysis to obtain soil sound wave transmission data.

[0157] In an embodiment, the sound wave sensor of the construction equipment emits sound waves to the soil at different depths and records the time of sound wave propagation in the soil and the received signal strength. The sound wave data is subjected to multi-band decomposition to analyze the propagation characteristics of sound waves in different frequency ranges. By comparing the propagation time and signal amplitude of each frequency band, the soil sound wave transmission data at different depths and frequency bands is obtained.

[0158] The soil sound velocity data is calculated based on the soil sound wave transmission data.

[0159] In an embodiment, the sound velocity distribution in the soil is calculated based on the sound wave transmission data. According to the time and distance of sound wave propagation, the sound velocity of the soil at different depths is estimated. The sound velocity data can reflect the distribution characteristics of soil density and porosity, thereby indirectly evaluating the compressibility and uniformity of the soil.

[0160] The soil reflection characteristic parameter data is obtained by extracting the sound wave reflection characteristics of the soil sound velocity data.

[0161] In an embodiment, the reflection characteristics of sound waves at the interface of soil layers are analyzed. According to the received reflected sound wave signals, the reflection intensity and time delay of different layer interfaces are extracted. By the change of reflection intensity, the system identifies the impedance difference of the soil layer interface, and determines the soil layering and local abnormal area.

[0162] The soil sound wave attenuation data is obtained by analyzing the sound wave attenuation of the soil emission characteristic parameter data.

[0163] In an embodiment, the propagation attenuation characteristics of sound waves in the soil are analyzed in depth. The signal strength attenuation of sound waves at different depths is calculated to evaluate the absorption ability of the soil to sound waves. The sound wave attenuation characteristics of the soil can reveal the size of the inter-particle gap and the internal microstructure characteristics.

[0164] In an embodiment, according to the propagation path of sound waves in the soil, the energy attenuation is calculated. The frequency band attenuation rate is calculated for each frequency band. The energy attenuation and the frequency band attenuation rate are integrated into soil sound wave attenuation data.

[0165] The soil sound wave propagation data is obtained by simulating the sound wave propagation of the soil sound wave attenuation data.

[0166] In an embodiment, based on the sound wave attenuation data, the soil sound wave propagation equation is established, the sound velocity and the path length are combined, the waveform superposition of the multi-path propagation is simulated, and the propagation of the sound wave in the soil is simulated. The soil sound wave propagation model is constructed to simulate the propagation path of the sound wave at different depths and directions. Through simulation, the energy distribution and propagation obstacles of the sound wave inside the soil are analyzed.

[0167] According to the soil sound wave propagation data, the soil characteristic parameter inversion is performed to obtain second soil characteristic feature data.

[0168] In an embodiment, the soil characteristic parameter is analyzed by inversion. Combined with the sound velocity and attenuation data, the key parameters such as the density, porosity and elastic modulus of the soil are obtained by inversion. The sound wave propagation in the soil conforms to the wave equation and is controlled by the density and elastic modulus of the soil through the soil sound wave equation. The sound velocity is determined by the density and elastic modulus of the soil, and the attenuation of the sound wave is determined by the porosity and inter-particle friction characteristics of the soil. Through the iterative inversion method, the sound wave propagation data is matched with the soil acoustic model. The error minimization target of the sound velocity and the attenuation is defined to ensure that the inversion parameters make the simulation results closest to the actual data. The density, porosity and elastic modulus are taken as the inversion parameters, and the initial values are set based on the geological investigation data of the construction area. The parameters are gradually adjusted according to the error value until the error meets the set accuracy requirement. Through the sound velocity data and the soil model, the density of each layer of soil is inverted. Higher sound velocity corresponds to higher density, and lower sound velocity corresponds to looser or higher water content soil layer. Based on the sound wave attenuation data, the porosity of the soil is estimated. High attenuation rate corresponds to soil with high porosity, because the sound wave loses more energy in the pores. The elastic modulus is calculated using the sound velocity and density inversion results. The elastic modulus reflects the ability of the soil to resist deformation, which is closely related to the pressure feedback of the construction equipment to the soil.

[0169] Optionally, wherein the soil reflection characteristic parameter data includes sound wave scattering characteristic data, the extraction method of the sound wave scattering characteristic data includes:

[0170] According to the construction real-time data, a soil physical model is constructed to obtain discrete soil grid data.

[0171] In an embodiment, based on the construction real-time data, including soil density, sound velocity and layer thickness information, a three-dimensional soil physical model is constructed. By layering the soil in the construction area, soil layers with different densities, sound velocities and impedances are divided. Using grid discretization technology, the soil model is divided into small discrete units, each unit containing specific physical parameters (such as density and sound velocity).

[0172] According to the discrete soil grid data, the sound wave source parameter labeling is performed to obtain sound wave source labeling data.

[0173] In one embodiment, based on discrete grid data, the system defines and labels parameters for the sound source. The location and depth of the sound source are determined and labeled as specific cells in the grid. The emission frequency, initial intensity, and propagation direction of the sound source are defined to ensure that the sound wave propagates in the model consistent with actual construction conditions.

[0174] Sound wave scattering simulation was performed on the sound source annotation data and soil sound velocity data to obtain the scattered sound field distribution data;

[0175] In one embodiment, sound wave propagation and scattering in soil were simulated by combining source-labeled data and soil sound velocity data. The simulation simulated the propagation path of sound waves from the source point in different regions of the soil. When the sound waves encountered interfaces between different soil layers, scattering phenomena, including the direction and intensity of reflection, refraction, and multiple scattering, were recorded. The simulation yielded data on the distribution of the scattered sound field in the soil, demonstrating the energy distribution and scattering characteristics of the sound waves.

[0176] The scattering intensity and directionality of the scattered sound field distribution data are analyzed to obtain the sound wave scattering characteristic data.

[0177] In one embodiment, the scattered sound field distribution data were analyzed in detail. The sound wave scattering intensity in different regions of the soil was calculated, and high-scattering and low-scattering regions were identified. The propagation direction of the scattered sound waves was analyzed to generate a directional distribution map of sound wave scattering, revealing the influence of the soil's internal structure on sound wave propagation.

[0178] Optionally, S3 includes:

[0179] S31. Based on soil property data and construction status data, process the construction status change data to obtain construction status change trend data.

[0180] In one embodiment, the changing trends of soil property data and construction status data over time are analyzed, such as using linear trends or recurrent convolutional neural network algorithms, to generate construction status change trend data. For example, as equipment pressure increases with construction depth, soil resistance also shows a gradual increasing trend, but abnormal fluctuations may occur at certain locations.

[0181] S32. Extract outlier points based on the construction status change trend data to obtain construction outlier point data;

[0182] In one embodiment, based on construction status change trend data, the system further identifies anomalies during construction. Using an anomaly template library or an isolated decision tree algorithm, analysis of the trend data reveals that pressure and resistance data suddenly deviate from normal ranges within certain time periods, such as an abnormal increase in equipment pressure at a certain depth. These deviations are marked as isolated construction anomalies, and the time, location, and related parameters of the anomaly are recorded.

[0183] In an embodiment, according to the construction state change trend data, the statistical method is used to detect the isolated points. The threshold is set k is an empirical coefficient, and δ is a data standard deviation. When a certain dimension data exceeds this range, it is marked as an abnormal point. The density analysis method (such as the local outlier factor method) is used to further identify the isolated points in the high-dimensional data. The local outlier factor LOF of each point is calculated: Where p is the target point (the point that needs to be judged whether it is an outlier point), q is the other point in the neighborhood of p (that is, the point set within the near neighbor range of p), is the k-neighbor set of point p, which contains the k nearest points to p, and O(q) is the local reachable density of point q. When LOF> the preset isolated point abnormal threshold, it is marked as an abnormal point. The LOF value describes the comparison between the local density of point p and the density of its neighborhood points: when LOF(p)≈1: it indicates that the local density of point p is similar to the local density of its neighborhood points, and p is not an outlier. When LOF(p)>1: it indicates that the density of point p is lower than that of the neighborhood points, and it is an outlier. When LOF(p)<1: it indicates that the density of point p is higher than that of the neighborhood points, and it is not an outlier.

[0184] S33, according to the construction abnormal isolated point data, the abnormal classification is carried out, and the construction abnormal classification data is obtained;

[0185] In an embodiment, the extracted abnormal isolated point data is classified. According to the characteristics of the abnormal occurrence (such as sudden increase of resistance, decrease of rotation speed or intensification of equipment vibration), the abnormal points are divided into different types by using the support vector machine algorithm or the abnormal event parameter library. For example, some abnormalities are classified as device overload induced abnormalities, and others are caused by sudden change of soil density.

[0186] In an embodiment, the context features of the abnormal points are extracted, including the trend changes before and after the occurrence of the abnormality. The unsupervised classification algorithm (such as the clustering method) is used to divide the abnormal points into several categories. The center point of each category is calculated and marked as the construction abnormal classification label data to obtain the construction abnormal classification data.

[0187] S34, change causal analysis is carried out on the construction abnormal classification data, and construction change causal data is obtained;

[0188] In an embodiment, the classified abnormal data is used for change causal analysis. Combined with the soil property data, the relationship between the abnormality and the soil attribute (such as the density and the porosity) is analyzed. For example, the abnormal points in a certain area are attributed to the significant increase of the soil layer resistance at the position, which causes the equipment operation load to exceed the normal range.

[0189] In one embodiment, time series analysis is used to determine the causal impact of soil properties or construction status data on anomalies based on causality tests. X→Y (causal) if Var(Y|X) < Var(Y), where X→Y indicates that variable X has a causal impact on variable Y, Var(Y|X) is the remaining fluctuation range of Y given X, Var(Y) is the overall change range of Y, that is, the degree of fluctuation of YYY when X is not considered, and Var(Y|X) < Var(Y) means that the conditional variance of Y given X is less than the total variance, indicating that X can significantly reduce the uncertainty of Y, demonstrating that X has a causal impact on Y. Regression analysis is performed on the time series before and after the anomaly point to extract the causal path. The causal model is linear regression: Y = β0 + β1X t-1 + μ t , where Y is the dependent variable, the target variable affected by the causal variable, β0 is the intercept, the base level when X is constant, β1 is the regression coefficient, representing the t-1 causal effect intensity of X on Y, X t-1 is the lagged causal variable, the variable affecting the current Y, and μ t is the error term, representing unexplained random factors or other influences. When β1 ≠ 0 and the significance test passes, the causal relationship is recorded. Construction change causal data = {(X→Y, β1}.

[0190] S35. Integrate the construction status change trend data, construction anomaly outlier data, and construction change causal data to obtain the construction status change data.

[0191] In one embodiment, the construction status change trend data, outlier data, and change causal data are integrated. A comprehensive construction status change report is constructed, covering construction dynamic trends, anomaly characteristics, and their causes.

[0192] Optionally, S4 includes:

[0193] S41. Generate optimization objectives based on the construction status change data to obtain construction optimization objective data;

[0194] In one embodiment, according to the construction status change data, including the dynamic change trends of pressure, rotational speed, and soil resistance, construction optimization objectives are set. Objective 1: Maximize construction efficiency - increase the construction depth per unit time. Objective 2: Minimize equipment energy consumption - reduce the energy consumption of the equipment in high-resistance areas. Objective 3: Minimize the anomaly occurrence rate - reduce the frequency of anomaly outliers, for example, avoid equipment shutdowns caused by excessive equipment pressure.

[0195] S42. Select optimization variables based on the construction optimization objective data to obtain construction optimization variable data;

[0196] In an embodiment, controllable construction variables are identified from the construction state change data: Cutterhead rotation speed: directly affects construction efficiency and soil resistance balance. Equipment applied pressure: controls the cutting ability of the equipment on the soil, affecting the abnormality occurrence rate. Advancing speed: adjusts the advancing speed of the construction equipment to adapt to different soil conditions.

[0197] S43, target optimization of the construction state change data according to the construction optimization variable data, to obtain construction parameter optimization data for TRD construction quality detection auxiliary work.

[0198] In an embodiment, the selected optimization variables are adjusted to achieve the optimization target: Cutterhead rotation speed optimization: in high resistance areas, reduce the rotation speed to avoid equipment overload; in low resistance areas, increase the rotation speed to increase construction efficiency. Equipment pressure optimization: near abnormal isolated points, appropriately reduce the equipment pressure to reduce the risk of abnormality occurrence; in uniform soil areas, maintain a higher pressure to improve cutting ability. Advancing speed optimization: slow down the advancing speed in dense soil areas to ensure smooth operation of the equipment; appropriately accelerate the advancing speed in loose soil areas to improve construction efficiency.

[0199] Optionally, the present application also provides a TRD construction quality detection system for executing the TRD construction quality detection method as described above, the TRD construction quality detection system comprising:

[0200] A construction real-time data acquisition module for acquiring construction real-time data;

[0201] A construction real-time analysis module for soil property analysis and construction state analysis according to the construction real-time data, to obtain soil property data and construction state data;

[0202] A construction state change processing module for construction state change analysis according to the soil property data and the construction state data, to obtain construction state change data;

[0203] A construction parameter optimization module for construction parameter optimization according to the construction state change data, to obtain construction parameter optimization data for TRD construction quality detection auxiliary work.

[0204] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the attached application file rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the application file are intended to be included in the present application.

[0205] The foregoing is considered as illustrative only of the principles of the application. Numerous modifications and changes will readily occur to those skilled in the art, and it is intended to embrace all such modifications and changes that fall within the scope of the application. Accordingly, the application is not to be restricted in scope to the specific embodiments disclosed herein but is to be accorded the full scope that the principles and novel features request appropriately granted.

Claims

1. A TRD construction quality inspection method, characterized in that, The method includes: S1. Obtain real-time construction data; S2. Extract soil characteristic features and construction status features from real-time construction data to obtain soil characteristic feature data and construction status feature data, respectively. Perform multi-dimensional feature correlation based on the soil characteristic feature data and construction status feature data to obtain construction feature relationship data. This construction feature relationship data includes a feature relationship graph with nodes representing cutterhead rotation speed, feed speed, and pressure from the construction status feature data, and resistance gradient, porosity, and density from the soil characteristic feature data. The edge weights are determined by the MIC values ​​between features and the hysteresis analysis results. Generate a soil model from the construction feature relationship data to obtain a soil mechanics model. This model is generated based on the construction feature relationship data, combined with soil resistance and density features, through finite element analysis and calculation. Extract soil characteristics from the soil mechanics model to obtain soil characteristic data. Model the construction status based on the construction feature relationship data to obtain a construction status model. This construction status model includes a model based on the construction feature relationship data and combined with the equipment's dynamic parameters, used to dynamically describe the relationship between equipment operating parameters and construction efficiency. Extract construction status from the construction status model to obtain construction status data. S3. Analyze the changes in construction status based on soil property data and construction status data to obtain construction status change data; S4. Optimize construction parameters based on construction status change data to obtain optimized construction parameter data for auxiliary TRD construction quality inspection.

2. The method according to claim 1, characterized in that, S1 includes: By integrating multimodal sensors into the TRD equipment, raw construction data, including soil resistance, cutterhead rotation speed, and depth displacement, are collected in real time. Geographic information fusion is performed on the original construction data to obtain geographically labeled construction data; Real-time multi-source data integration is performed on geographically labeled construction data to obtain real-time construction data.

3. The method according to claim 1, characterized in that, Soil characteristic feature extraction includes: First soil characteristic features are extracted from real-time construction data to obtain first soil characteristic feature data; The second soil characteristic feature is extracted from the real-time construction data to obtain the second soil characteristic feature data. The first soil characteristic feature extraction and the second soil characteristic feature extraction are different soil characteristic feature extraction methods. The first soil characteristic data and the second soil characteristic data are fused to obtain soil characteristic data.

4. The method according to claim 3, characterized in that, The real-time construction data includes soil vibration data, which is acquired through seismic wave sensors or high-frequency vibration sensors. The first soil characteristic feature extraction includes: Time-domain characteristic analysis was performed on soil vibration data in real-time construction data to obtain soil time-domain characteristic data. By performing frequency domain transformations on the soil time-domain characteristic data, soil frequency domain characteristic data can be obtained. Energy spectrum decomposition was performed on the soil frequency domain characteristic data to obtain soil spectrum decomposition data; Dynamic response parameters were mapped from the soil spectral decomposition data to obtain the first soil characteristic data; The energy spectrum decomposition includes: Time-frequency joint transformation is performed on soil time-domain characteristic data to obtain local time-frequency characteristic data; Spectral energy distribution data is obtained by extracting the spectral energy distribution from the local time-frequency feature data. Multi-scale spectral decomposition is performed on the spectral energy distribution data to obtain multi-scale spectral data; Spectral correlation analysis was performed on multi-scale spectral data to obtain spectral correlation matrix data; Energy characteristic clustering was performed on the spectral correlation matrix data to obtain the soil spectral energy clustering characteristics; The soil spectral energy clustering features were reconstructed using feature maps to obtain reconstructed soil spectral data; Soil spectrum decomposition data is generated by reconstructing soil spectrum data.

5. The method according to claim 3, characterized in that, The real-time construction data includes soil acoustic wave data, which is collected by acoustic wave sensors. The second soil characteristic feature extraction includes: Multi-band acoustic wave transmission analysis was performed on the soil acoustic wave data in the real-time construction data to obtain soil acoustic wave transmission data. The sound velocity is calculated from the soil sound wave transmission data to obtain soil sound velocity data; The sound wave reflection characteristics of the soil sound velocity data were extracted to obtain soil reflection characteristic parameter data; Sound wave attenuation analysis was performed on the soil emission characteristic parameter data to obtain soil sound wave attenuation data; Sound wave propagation data of soil sound wave attenuation data is obtained by simulating sound wave propagation in soil. Soil property parameters are inverted based on soil acoustic wave propagation data to obtain second soil property characteristic data.

6. The method according to claim 5, characterized in that, The soil reflectance parameter data includes acoustic scattering characteristic data, and the methods for extracting acoustic scattering characteristic data include: A soil physical model was constructed based on real-time construction data to obtain discrete soil grid data. The acoustic source parameters are labeled based on the discrete soil grid data to obtain the acoustic source labeled data; Sound wave scattering simulation was performed on the sound source annotation data and soil sound velocity data to obtain the scattered sound field distribution data; The scattering intensity and directionality of the scattered sound field distribution data are analyzed to obtain the sound wave scattering characteristic data.

7. The method according to claim 1, characterized in that, S3 include: Based on soil property data and construction status data, the construction status change processing is performed to obtain construction status change trend data; Anomaly isolated points are extracted based on the construction status change trend data to obtain construction anomaly isolated point data; Anomaly classification is performed based on isolated construction anomaly data to obtain construction anomaly classification data; A causal analysis of changes in construction anomaly classification data was performed to obtain causal data of construction changes. By integrating the data on construction status change trends, isolated construction anomalies, and causal data on construction changes, we obtain the construction status change data.

8. The method according to claim 1, characterized in that, S4 include: Optimization targets are generated based on construction status change data to obtain construction optimization target data; Based on the construction optimization target data, optimization variables are selected to obtain construction optimization variable data; Based on the construction optimization variable data, the construction status change data is optimized to obtain construction parameter optimization data, which is used to assist in TRD construction quality detection.

9. A TRD construction quality inspection system, characterized in that, For performing the TRD construction quality inspection method as described in claim 1, the TRD construction quality inspection system comprises: The construction real-time data acquisition module is used to acquire real-time construction data; The real-time construction analysis module is used to perform soil property analysis and construction status analysis based on real-time construction data, and to obtain soil property data and construction status data. The construction status change processing module is used to analyze construction status changes based on soil property data and construction status data to obtain construction status change data. The construction parameter optimization module is used to optimize construction parameters based on construction status change data to obtain optimized construction parameter data for auxiliary TRD construction quality inspection.

Images