Geological process label construction method and system based on shield operation parameters

Abstract

The embodiment of the application provides a geological process label construction method and system based on shield operation parameters, and belongs to the technical field of construction process information processing. The method comprises the following steps: acquiring shield multi-system operation parameters formed in a shield construction process, and constructing a multi-dimensional feature space corresponding to shield tunneling footage based on the shield multi-system operation parameters; calculating parameter variation range information between adjacent tunneling footage based on the multi-dimensional feature space, and identifying running state change boundary positions; segmenting the shield tunneling footage to obtain a plurality of running state segments; extracting segment representative features in each running state segment, and determining corresponding geological process labels based on the segment representative features; and outputting a geological process label sequence changing with the shield tunneling footage. The scheme constructs an inchage-aligned cooperative change structure and combines a segmentation and consistency checking mechanism, so that continuous, stable and structured expression of geological evolution characteristics in the shield tunneling process is realized.

Description

Technical Field

[0001] This invention relates to the field of construction process information processing technology, specifically to a method and system for constructing geological process tags based on shield tunneling operation parameters. Background Technology

[0002] During shield tunneling, numerous operational parameters are continuously generated by multiple subsystems, including the cutterhead, propulsion, screw conveyor, and grouting. These parameters, to a certain extent, reflect the interaction between the tunnel boring machine (TBM) and the surrounding geological strata. With the development of information-based construction and intelligent tunneling, analyzing geological conditions and the construction process using the multi-system operational parameters of the TBM has become one of the important research directions in the field of shield tunneling.

[0003] In practical engineering, geological conditions are usually described through survey reports or advanced geological forecasts, and given in the form of sections. However, tunnel boring machine (TBM) excavation is a continuous process advancing along the advance direction, and the operating parameters exhibit dynamic evolution characteristics in the advance dimension. Existing geological information is mostly represented in the form of discrete sections or cross-sections, making it difficult to establish a precise correspondence with the operating parameters in the advance dimension during continuous TBM excavation. This is especially true in areas with gradual geological changes or transition zones, where geological boundaries are often not abrupt but rather exhibit continuous changes within a certain range.

[0004] Existing analysis methods based on operational parameters primarily employ sample-level classification or clustering, treating each tunneling loop as an independent analysis unit and classifying geological formations or construction states based on parameter characteristics. However, there are significant coupling relationships among the operational parameters of multiple shield tunneling systems. Parameters from different subsystems typically exhibit coordinated changes during geological changes, rather than abrupt changes in a single parameter. If state classification is based solely on single-point characteristics or the assumption of sample independence, the collaborative evolutionary structure of operational parameters along the advance dimension is easily overlooked, leading to frequent jumps or unstable boundary phenomena in the classification results.

[0005] On the other hand, tunnel boring machines (TBMs) often experience short-term adjustments to operating conditions, construction control interventions, and equipment response delays, which can cause fluctuations or noise in operating parameters within local ranges. If a continuity constraint mechanism based on changing structures is lacking, and the condition is determined solely by the instantaneous amplitude of changes, short-term fluctuations can easily be misjudged as geological changes, thus affecting the stability and reliability of the segmentation results.

[0006] Therefore, without relying on real geological labels, how to identify the structural change boundary of the operating state based on the coordinated change characteristics of the operating parameters of the shield tunneling system in the tunneling advance direction, and on this basis construct continuous and stable geological process segment labels, while verifying the evolution consistency between adjacent segments, has become a technical problem that urgently needs to be solved in the field of shield tunneling construction information processing. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for constructing geological process tags based on shield tunneling operation parameters, so as to at least solve the problem in the prior art that it is difficult to identify structural changes in geological processes and construct a continuous and stable tag sequence based on operation parameters.

[0008] To achieve the above objectives, the first aspect of the present invention provides a method for constructing geological process labels based on shield tunneling operation parameters. The method includes: acquiring shield multi-system operation parameters formed during shield tunneling construction, and constructing a multi-dimensional feature space corresponding one-to-one with the shield tunneling advance based on the shield multi-system operation parameters; calculating parameter variation amplitude information between adjacent advances based on the multi-dimensional feature space, and identifying the boundary position of operation state change based on the parameter variation amplitude information; segmenting the shield tunneling advance based on the boundary position of operation state change to obtain multiple operation state segments; extracting representative features of each operation state segment, and determining the corresponding geological process label based on the representative features of the segment; performing consistency verification on the label continuity between adjacent operation state segments, and outputting a geological process label sequence that changes with the shield tunneling advance.

[0009] Optionally, the process involves acquiring the multi-system operating parameters of the shield tunneling system formed during shield tunneling construction, and constructing a multi-dimensional feature space that corresponds one-to-one with the shield tunneling advance based on these parameters. This includes: acquiring the multi-system operating parameters of the shield tunneling system formed along the tunneling advance direction during shield tunneling construction, and establishing a parameter-advance mapping relationship between the multi-system operating parameters according to the corresponding tunneling advance; performing advance resampling processing on the parameter-advance mapping relationship according to a preset standard advance interval, uniformly mapping the multi-system operating parameters of the shield tunneling system with different sampling frequencies to the same advance node, forming an advance alignment parameter sequence; performing advance offset correction processing on the operating parameters of each subsystem in the advance alignment parameter sequence based on a preset response delay coefficient, generating a delay correction parameter sequence; performing dimensionless processing on the delay correction parameter sequence, and uniformly transforming the change direction of each parameter, generating a scale-unified parameter sequence; constructing a parameter co-change matrix based on the scale-unified parameter sequence, and establishing a one-to-one correspondence between the parameter co-change matrix and the corresponding tunneling advance, thus constructing a multi-dimensional feature space for characterizing the changes in the operating state during shield tunneling.

[0010] Optionally, calculating the parameter variation amplitude information between adjacent tunneling advances based on the multidimensional feature space, and identifying the boundary position of the operating state change based on the parameter variation amplitude information, includes: performing a difference operation on the parameter cooperative variation matrix corresponding to adjacent tunneling advances in the multidimensional feature space to obtain a multidimensional variation vector sequence; performing sign consistency analysis on the variation of each dimension in the multidimensional variation vector sequence, and calculating the difference between the number of dimensions of change in the same direction and the number of dimensions of change in opposite directions to generate a cooperative variation direction index; performing a weighted summation operation on the variation of each dimension in the multidimensional variation vector sequence to obtain a cooperative variation intensity index; performing a normalization fusion operation on the cooperative variation direction index and the cooperative variation intensity index to generate a multidimensional cooperative variation index sequence; performing cumulative variation energy calculation on the multidimensional cooperative variation index sequence within a continuous preset number of tunneling advance intervals, and determining the corresponding tunneling advance as the boundary position of the operating state change when the cumulative variation energy exceeds a preset variation threshold.

[0011] Optionally, a normalization fusion operation is performed on the coordinated change direction index and the coordinated change intensity index to generate a multidimensional coordinated change index sequence, including: performing interval normalization processing on the coordinated change direction index sequence to obtain a direction-normalized index sequence, and limiting the direction-normalized index sequence to a preset direction index range; performing interval normalization processing on the coordinated change intensity index sequence to obtain an intensity-normalized index sequence, and limiting the intensity-normalized index sequence to a preset intensity index range; performing a linear fusion operation on the direction-normalized index sequence and the intensity-normalized index sequence based on preset fusion weights to obtain a fused coordinated index sequence; performing saturation pruning processing on the fused coordinated index sequence to obtain a pruned coordinated index sequence, and limiting the pruned coordinated index sequence to a preset coordinated index range; and performing sliding smoothing processing on the pruned coordinated index sequence to generate a multidimensional coordinated change index sequence.

[0012] Optionally, the tunnel boring machine's advance is segmented based on the boundary positions of the operational state changes to obtain multiple operational state segments. This includes: sorting the boundary positions of the operational state changes according to the tunneling advance sequence to generate a boundary position sequence; constructing candidate segments between adjacent boundary positions based on the boundary position sequence, and establishing a correspondence between each candidate segment and its corresponding tunneling advance range; calculating the segment length for each candidate segment, and performing adjacent segment merging processing on candidate segments whose segment length is less than a preset minimum segment length; and determining the merged segment set as multiple operational state segments.

[0013] Optionally, extracting representative features within each operational state segment and determining corresponding geological process labels based on these representative features includes: performing intra-segment feature aggregation operations on the corresponding multi-dimensional feature space within each operational state segment to generate a segment feature set; calculating the contribution of each parameter dimension to the changes within the operational state segment based on the segment feature set to generate a segment parameter contribution sequence; performing sorting processing on the segment parameter contribution sequence to determine the dominant parameter combination for the segment; extracting the main response feature vector for the segment based on the dominant parameter combination for the segment, using it as the representative feature of the operational state segment; and determining the corresponding geological process label based on the mapping rules between the representative features and preset geological process labels.

[0014] Optionally, the extraction rule for the segment main response feature vector is as follows: extract the parameter sequence corresponding to the segment's dominant parameter combination within the corresponding operating state segment; perform segment-level centering processing on the parameter sequence to generate a dominant parameter offset sequence; calculate the pairwise cooperative change coefficients between the parameters based on the dominant parameter offset sequence to generate a dominant parameter cooperative matrix; perform eigenvalue decomposition on the dominant parameter cooperative matrix to extract the main cooperative direction vector; and concatenate the main cooperative direction vector with the mean vector of the corresponding segment's dominant parameters to generate the segment main response feature vector.

[0015] Optionally, a consistency check is performed on the label continuity between adjacent operating state segments, and a geological process label sequence that changes with the tunnel boring machine's advance is output. This includes: sorting the operating state segments according to the tunnel boring machine's advance sequence to generate a segment label sequence; performing similarity calculation on the segment master response feature vectors corresponding to adjacent operating state segments to generate a segment similarity sequence; comparing the segment similarity sequence with a preset similarity threshold; when the segment similarity of adjacent operating state segments is greater than the preset similarity threshold, performing label merging processing to generate merged segment labels; when the segment similarity of adjacent operating state segments is less than the preset similarity threshold, keeping the segment labels unchanged; and performing advance expansion processing on the merged segment label sequence to generate a geological process label sequence that corresponds one-to-one with the tunnel boring machine's advance.

[0016] A second aspect of the present invention provides a geological process label construction system based on shield tunneling operation parameters. The system includes: a data acquisition unit for acquiring shield multi-system operation parameters generated during shield tunneling construction and constructing a multi-dimensional feature space corresponding one-to-one with the shield tunneling advance based on the shield multi-system operation parameters; a preprocessing unit for calculating parameter variation amplitude information between adjacent advances based on the multi-dimensional feature space and identifying the boundary position of operation state change based on the parameter variation amplitude information; a clustering unit for segmenting the shield tunneling advance based on the boundary position of operation state change to obtain multiple operation state segments; a verification unit for extracting representative features of each operation state segment and determining the corresponding geological process label based on the representative features of the segment; and an output unit for verifying the consistency of label continuity between adjacent operation state segments and outputting a geological process label sequence that changes with the shield tunneling advance.

[0017] On the other hand, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for constructing geological process tags based on tunnel boring machine operating parameters.

[0018] Through the above technical solution, this invention is based on the coordinated variation characteristics of the operating parameters of the shield tunneling system in the tunneling advance direction. By constructing a multi-dimensional feature space corresponding one-to-one with the tunneling advance, it identifies the structural change boundaries of the operating state and achieves stable segmentation of the shield tunneling process accordingly. Within each operating state segment, representative features are extracted and corresponding geological process labels are generated. Simultaneously, consistency verification is performed on labels of adjacent segments, effectively suppressing the impact of short-term fluctuations and local noise on the segmentation results, and improving the continuity and structural stability of the label sequence. Finally, a geological process label sequence that varies with the tunneling advance is formed, realizing the continuous expression of geological evolution characteristics during shield tunneling, and providing a stable and unified structured label foundation for subsequent construction state analysis and parameter modeling.

[0019] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the steps of a method for constructing geological process tags based on shield tunneling operation parameters provided by one embodiment of the present invention; Figure 2This is a detailed flowchart of step S10 of the geological process label construction method based on shield tunneling operation parameters provided in one embodiment of the present invention; Figure 3 This is a curve showing the variation of typical operating parameters along the tunneling ring number during shield tunneling, provided by one embodiment of the present invention. Figure 4 This is a system structure diagram of a geological process label construction system based on shield tunneling operation parameters provided by one embodiment of the present invention; Figure 5 This is an internal structural diagram of a computer device provided in one embodiment of the present invention. Detailed Implementation

[0021] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] like Figure 1 As shown, this invention provides a method for constructing geological process tags based on shield tunneling operation parameters, the method comprising: Step S10: Obtain the shield tunneling multi-system operation parameters formed during the shield tunneling construction process, and construct a multi-dimensional feature space that corresponds one-to-one with the shield tunneling advance based on the shield tunneling multi-system operation parameters.

[0023] Specifically, during tunnel boring machine (TBM) construction, multiple systems, including the cutterhead, propulsion, auger conveyor, and grouting, simultaneously generate numerous operating parameters. These parameters reflect the interaction between the TBM and the surrounding rock strata from different perspectives. This step establishes a parameter-footprint mapping relationship between the operating parameters of each subsystem and their corresponding tunneling advances. Based on a preset standard footprint interval, the parameter-footprint mapping relationship is resampled to map operating parameters with different sampling frequencies to the same footprint node, forming a footprint alignment parameter sequence. Further, a footprint offset correction process is performed on the operating parameters of each subsystem based on a preset response delay coefficient to eliminate the impact of response lag in different systems on the footprint alignment results. The parameter sequence after delay correction is dimensionless, and the direction of parameter change is unified. The parameters with unified scale are used to construct a parameter co-variation matrix according to parameter dimensions, and a one-to-one correspondence is established with the corresponding tunneling advances, forming a multi-dimensional feature space to characterize the evolution of the TBM's operating state, providing a basic data structure for subsequent boundary identification. Specifically, such as... Figure 2 Step S10 includes the following steps: Step S101: Obtain the shield multi-system operating parameters formed along the tunneling advance direction during shield construction, and establish a parameter-advance mapping relationship for the shield multi-system operating parameters according to the corresponding tunneling advance.

[0024] Specifically, during tunnel boring machine (TBM) construction, multiple systems, including the cutterhead system, propulsion system, and screw conveyor system, continuously generate operational parameters that reflect the TBM's operating status at different tunneling locations. Because the timing and frequency of parameter acquisition vary across different systems, simply organizing the data chronologically is insufficient to accurately reflect the TBM's spatial operational changes.

[0025] This step obtains the timestamp information corresponding to the operating parameters of each subsystem and the tunneling footage records within the corresponding time period. Based on the correspondence between the timestamps and the tunneling footage records, the operating parameters of each subsystem are converted into data formats corresponding to the tunneling footage positions. For operating parameters sampled multiple times within the same tunneling footage interval, they are merged into the corresponding footage nodes according to preset mapping rules, and a parameter-footprint mapping table is generated.

[0026] By establishing a parameter-expansion mapping relationship, the operating parameters of the shield tunneling multi-system at different sampling frequencies and acquisition times are all expressed using the tunneling advance as a unified spatial index, providing a unified data benchmark for subsequent advance resampling processing and structural change analysis.

[0027] Step S102: Perform advance resampling processing on the parameter-advance mapping relationship according to the preset standard advance interval, and map the shield tunneling multi-system operating parameters with different sampling frequencies to the same advance node to form an advance alignment parameter sequence.

[0028] Specifically, in actual construction, the sampling frequencies of each subsystem often differ. For example, cutterhead torque may be collected at a higher frequency, while grouting pressure or propulsion displacement may be recorded at a relatively lower frequency. If the data is directly arranged along the time axis, the spatial correspondence of different parameters is prone to shift. This step uses tunneling footage as a unified spatial index, presets standard footage intervals, such as dividing spatial nodes according to fixed footage increments, and merges operating parameters falling within the same footage interval to the corresponding footage node.

[0029] For high-frequency sampled data, statistical aggregation can be performed within the standard advance interval; for low-frequency sampled data, interpolation or data preservation can be performed based on the records between adjacent advance nodes, thus ensuring that each standard advance node corresponds to a complete set of multi-system operating parameters. The value of the standard advance interval can be preset according to engineering requirements or adjusted according to the resolution of changes in tunneling advance. This step defines a unified expression of multi-system operating parameters in the spatial dimension, providing a consistent advance benchmark for subsequent identification of changing structures.

[0030] Step S103: Perform advance offset correction processing on the operating parameters of each subsystem in the advance alignment parameter sequence based on the preset response delay coefficient to generate a delay correction parameter sequence.

[0031] Specifically, during tunnel boring machine (TBM) excavation, different subsystems exhibit objective lags in their response to changes in the geological strata. For example, changes in propulsion force may precede changes in soil chamber pressure, while fluctuations in the auger conveyor torque may lag behind changes in cutterhead load. If the differences in response delays between systems are not considered, and analysis is performed directly based on the advance alignment parameter sequence, the same geological change may be expressed in a dispersed manner at different advance positions, weakening the accuracy of structural identification.

[0032] This step presets response delay coefficients for the operating parameters of each subsystem. These coefficients can be determined based on historical data statistics, system physical structure characteristics, or field calibration results. Based on these response delay coefficients, offset corrections are applied to the corresponding subsystem operating parameters along the advance dimension, ensuring that the responses of each subsystem to the same geological change are more consistent in the advance direction. The subsystem operating parameters, after advance offset correction, are then recombined to form a delay-corrected parameter sequence. This enhances the synergistic expressive ability of different system parameters in the spatial dimension, providing more consistent basic data for subsequent synergistic change analysis.

[0033] Step S104: Perform dimensionless processing on the delay correction parameter sequence and uniformly transform the change direction of each parameter to generate a scale-unified parameter sequence.

[0034] Specifically, the operating parameters of different subsystems differ significantly in physical dimensions and numerical ranges. For example, the order of magnitude of the cutterhead torque and the propulsion displacement are different, and the ranges of grouting pressure and screw conveyor speed are also inconsistent. If the delay correction parameter sequence is directly compared or calculated, parameters with larger numerical amplitudes tend to dominate in the change analysis, affecting the identification of the cooperative change structure.

[0035] This step performs dimensionless processing on each parameter in the delayed correction parameter sequence, ensuring that each parameter is expressed within a unified numerical range and eliminating the impact of dimensional differences on subsequent calculations. Simultaneously, considering that some parameters exhibit positive correlation trends during geological changes, while others show inverse trends, the direction of change for each parameter is uniformly transformed to ensure consistent directional significance when representing geological responses. After dimensionless processing and directional unification, a scale-uniform parameter sequence is formed, allowing each parameter to participate in subsequent collaborative change analysis within the same scale framework, improving the stability and comparability of identifying operational status changes.

[0036] Step S105: Construct a parameter co-variation matrix based on the scale-unified parameter sequence, and establish a one-to-one correspondence between the parameter co-variation matrix and the corresponding tunneling advance, thereby constructing a multi-dimensional feature space to characterize the changes in the operating state during shield tunneling.

[0037] Specifically, after obtaining the scale-unified parameter sequence, each parameter is within a unified numerical range and exhibits a consistent direction of change. This step no longer simply concatenates the parameters into a vector, but instead focuses on a structural expression of the collaborative changes between the parameters. For the same tunneling advance node, the corresponding multidimensional parameter values ​​are extracted, and the pairwise correlations between each parameter are calculated, forming a parameter collaborative change matrix. This matrix characterizes the coupling degree and collaborative direction of different subsystem parameters at that advance position, ensuring that the operational status is not only reflected by single-point values ​​but also by the structural relationships between multiple systems.

[0038] Each advance in tunneling corresponds to a parameter coordination variation matrix, and a one-to-one correspondence is established between this matrix and the corresponding advance. The parameter coordination variation matrices are arranged along the advance direction to form a multi-dimensional feature space, which enables the evolution structure of the tunnel boring machine's operating state during tunneling to be continuously expressed in the spatial dimension, providing structured input for subsequent boundary identification.

[0039] In one specific implementation, such as Figure 3 This paper describes the mechanical parameters of the TBM (Tunnel Boring Machine) collected during shield tunneling construction. In this embodiment, 16 TBM mechanical parameters from the cutterhead system, shield system, propulsion system, and screw conveyor system are selected as the analysis objects. For a dual-line shield tunneling project, data from ring numbers 50 to 1349 (1300 rings) are selected for the left line; data from ring numbers 50 to 1249 (1200 rings) are selected for the right line. Project data is stored in ring units, with each ring containing raw sampling data from multiple time points. To ensure the comparability of parameters between different ring segments, the data within each ring are statistically averaged to form a parameter sequence corresponding one-to-one with the ring number.

[0040] Among them, the penetration rate (PR) and advancement rate (AR) exhibit similar trends across multiple segments: higher overall values ​​in the early segments, a synchronous decrease in the middle segments, and relatively stable fluctuations in subsequent segments. The cutterhead angle (CA) shows minimal variation in most segments, with abrupt changes only at a few segment numbers; while the cutterhead rotation speed (RPM) exhibits clear segmented characteristics, forming relatively stable value ranges across different segments. This indicates that some parameters share similar trends and exhibit redundancy. Directly using the original parameters for subsequent machine learning analysis may affect the modeling results; therefore, preprocessing is necessary.

[0041] Step S20: Calculate the parameter variation range information between adjacent tunneling advances based on the multi-dimensional feature space, and identify the boundary position of the operating status change based on the parameter variation range information.

[0042] Specifically, a difference operation is performed on the parameter coordination change matrix corresponding to adjacent tunneling advances in the multidimensional feature space to obtain a multidimensional change vector sequence; a sign consistency analysis is performed on the change quantities of each dimension in the multidimensional change vector sequence, and the difference between the number of dimensions of change in the same direction and the number of dimensions of change in opposite directions is calculated to generate a coordination change direction index; a weighted summation operation is performed on the change quantities of each dimension in the multidimensional change vector sequence to obtain a coordination change intensity index; a normalization fusion operation is performed on the coordination change direction index and the coordination change intensity index to generate a multidimensional coordination change index sequence; cumulative change energy is calculated on the multidimensional coordination change index sequence within a continuous preset number of tunneling advance intervals, and when the cumulative change energy exceeds a preset change threshold, the corresponding tunneling advance is determined as the boundary position of the operating state change.

[0043] In this embodiment of the invention, the multidimensional feature space corresponds to the parameter coordination change matrix at each tunneling advance node, and the position of the boundary of the operating state change is characterized by the difference in the coordination change matrices of adjacent advance nodes. A difference operation is performed on the parameter coordination change matrices corresponding to adjacent tunneling advance nodes to obtain a difference matrix, which is then converted into a multidimensional change vector sequence according to a preset flattening rule. The multidimensional change vector is used to describe the magnitude and direction of change of the multi-system parameter coupling structure from the previous advance node to the current advance node.

[0044] Sign consistency analysis uses the sign of the changes in each dimension of the multidimensional change vector as directional information. A sign value is calculated for each dimension of the change vector, and the number of dimensions with positive and negative signs is counted. The difference between these two values ​​is calculated as the cooperative change direction index, which characterizes the degree of directional consistency of the changes in the multi-system parameter coupling structure at the current advance node. The cooperative change intensity index is composed of the magnitudes of the changes in each dimension of the multidimensional change vector. A weighted summation is performed on the absolute values ​​of the changes in each dimension to obtain the intensity index. The weights are given by a preset weight vector to reflect the differences in the contribution of different parameter coupling terms in the change magnitude assessment.

[0045] The directional and intensity indices are then fused. Interval normalization is performed on both the directional and intensity index sequences to ensure they fall within a unified numerical range. Based on preset fusion weights, a linear fusion is performed on the normalized directional and intensity indices to obtain a multidimensional collaborative change index sequence. This multidimensional collaborative change index sequence is then saturated and pruned to suppress the impact of extreme values ​​on subsequent cumulative judgments, and a sliding smoothing process is applied to the pruned index sequence to reduce short-term fluctuations.

[0046] The boundary position of operational status change is determined by cumulative change energy. Cumulative energy is calculated on a smoothed multidimensional coordinated change index sequence within a predetermined number of consecutive tunneling footage intervals. The cumulative energy is the sum of squares of each coordinated change index within that interval or an equivalent cumulative amount. The cumulative energy is compared with a predetermined change threshold. When the cumulative energy exceeds the predetermined change threshold, the tunneling footage node is determined as the boundary position of operational status change; when the cumulative energy does not exceed the predetermined change threshold, the boundary position is not updated. The predetermined number of tunneling footage intervals is used to limit the minimum continuous change length for boundary identification, and the predetermined change threshold is used to limit the determination intensity of structural changes, thereby avoiding false boundaries triggered by single-point noise or local operating condition disturbances.

[0047] In one specific implementation, parameter variation information between adjacent tunneling advances is calculated based on the multidimensional feature space, and the location of the boundary of the change in operating status is identified based on the parameter variation information.

[0048] Let the sequence of tunneling advance nodes be as follows: The corresponding parameter cooperative change matrix is Perform differential operations on adjacent tunneling advance nodes: Difference matrix Transformed into a multidimensional transformation vector according to preset expansion rules. This vector is used to characterize the overall change state of the multi-system coupling structure at the current advance node relative to the previous node.

[0049] For the changing vector Consistency analysis of execution direction. Count the number of dimensions showing positive change. Number of negative change dimensions Construct a cooperative change direction index: in This is the dimension of the change vector. This index reflects whether the parameters of multiple systems exhibit a uniform change in the same direction at the current advance node.

[0050] By performing a weighted summation operation on the absolute values ​​of the changes in each dimension of the change vector, the coordinated change intensity index is obtained: in No. The intensity index of coordinated change at each advance node, These are preset weights used to reflect the degree of influence of coupling terms with different parameters. For the first At the first advance node, the first The amount of change in the dimensional parameter.

[0051] The directional and intensity indices were normalized to intervals, and then based on the fusion coefficient... Perform linear fusion: in, Indicates the first The multidimensional collaborative change index corresponding to each tunneling footage node Indicates the first Normalized cooperative change direction index of each advance node Indicates the first The normalized coordinated change intensity index of each advance node is used to obtain the multidimensional coordinated change index sequence. .

[0052] To avoid false triggering of boundary determination by single-point disturbances, in continuous Calculate the cumulative energy change within each tunneling footage interval: when Exceeding the preset change threshold At that time, the corresponding tunneling footage will be... The location was determined as the boundary of the change in operating state. Indicates the first The multidimensional cooperative change index is calculated at each advance node, where L is the preset window length. The direction index characterizes the consistency of the change structure, the intensity index characterizes the magnitude of the change, the fusion index comprehensively expresses the degree of structural abrupt change, and the cumulative change energy is used to suppress short-term noise fluctuations. These rules enable the identification of stable boundaries based on the cooperative change structure of multiple systems.

[0053] Specifically, the process of performing a normalization fusion operation on the coordinated change direction index and the coordinated change intensity index to generate a multidimensional coordinated change index sequence includes: performing interval normalization on the coordinated change direction index sequence to obtain a direction-normalized index sequence, and limiting the direction-normalized index sequence to a preset direction index range; performing interval normalization on the coordinated change intensity index sequence to obtain an intensity-normalized index sequence, and limiting the intensity-normalized index sequence to a preset intensity index range; performing a linear fusion operation on the direction-normalized index sequence and the intensity-normalized index sequence based on preset fusion weights to obtain a fused coordinated index sequence; performing saturation pruning on the fused coordinated index sequence to obtain a pruned coordinated index sequence, and limiting the pruned coordinated index sequence to a preset coordinated index range; and performing sliding smoothing on the pruned coordinated index sequence to generate a multidimensional coordinated change index sequence.

[0054] In this embodiment of the invention, the co-change direction index and the co-change intensity index typically differ in their numerical range and fluctuation scale. To ensure comparability between the two in the fusion calculation, interval normalization is performed on the direction index sequence, linearly mapping it to a preset direction index range, such as mapping it to a unified numerical range; similarly, interval normalization is performed on the intensity index sequence, limiting it to a preset intensity index range. The interval normalization process can employ a linear mapping method based on global extrema or a dynamic mapping method based on sliding window extrema. Its core purpose is to eliminate the differences in scale between different indices, giving them a unified dimensional basis.

[0055] After normalization, a linear fusion operation is performed on the direction-normalized index sequence and the intensity-normalized index sequence according to preset fusion weights to obtain a fused synergy index sequence. The fusion weights are used to adjust the influence ratio of directional consistency and change intensity in the comprehensive judgment; their values ​​can be determined based on historical sample statistics, engineering experience, or model calibration results. The fused synergy index reflects both the degree of consistency of multidimensional parameter changes in direction and the overall magnitude of change.

[0056] To suppress the interference of extreme fluctuations on subsequent boundary determination, a saturation pruning process is performed on the fused synergistic index sequence, limiting values ​​exceeding the preset upper or lower limits of the synergistic index range to the corresponding range boundaries, resulting in a pruned synergistic index sequence. Subsequently, a sliding smoothing process is performed on the pruned synergistic index sequence, such as averaging or weighted smoothing based on a fixed-length sliding window, to reduce the risk of misjudgment caused by local peak fluctuations, ultimately generating a multidimensional synergistic change index sequence.

[0057] The above-mentioned normalization methods, fusion forms, pruning rules, and smoothing strategies all fall within the protection scope of this invention. Any technical solution that achieves unified scaling of the direction index and intensity index and forms a stable and coordinated change index sequence should fall within the protection scope of this invention.

[0058] In another possible implementation, the shield tunneling multi-system operating parameters are divided into several continuous advance segments based on the tunneling progress. Within each advance segment, integrity checks are performed on the shield tunneling multi-system operating parameters to identify missing or abnormal values ​​in local tunneling sections. For tunneling ring positions identified as abnormal within a certain advance segment, abnormal ring removal is only performed within the corresponding advance segment, thereby avoiding excessive impact of local anomalies on the global data structure.

[0059] After completing the segmented anomaly loop removal, correlation constraint processing is further introduced within each advance section. Based on the shield tunneling multi-system operating parameters after anomaly loop removal, the correlation relationships between different parameter dimensions are calculated, and parameter dimensions that satisfy the correlation constraint conditions are identified according to preset correlation constraint criteria. Subsequently, constraint processing is performed on the parameter dimensions that satisfy the correlation constraint conditions within each advance section, and the processing results are integrated in the advance direction to form a consistent multi-dimensional feature space.

[0060] Step S30: Based on the location of the boundary of the change in the operating state, the tunnel boring machine advance is segmented to obtain multiple operating state segments.

[0061] Specifically, the boundary positions of the operational status changes are sorted according to the tunneling footage sequence to generate a boundary position sequence; candidate segments are constructed between adjacent boundary positions based on the boundary position sequence, and a correspondence is established between each candidate segment and the corresponding tunneling footage range; the segment length of each candidate segment is calculated, and candidate segments with segment lengths less than the preset minimum segment length are merged into adjacent segments; the merged segment set is determined as multiple operational status segments.

[0062] In this embodiment of the invention, after identifying the boundary locations of operational state changes, each boundary location corresponds to several tunneling advance nodes. To ensure the spatial order of the segmentation results, the operational state change boundary locations are sorted according to the value of the tunneling advance, forming a monotonically increasing sequence of boundary locations. This sequence of boundary locations constitutes a set of structural change nodes during the shield tunneling process, used to define the spatial locations where operational states may change.

[0063] Based on this, the tunneling footage range between two adjacent boundary positions is used as candidate segments. The interval between the tunneling start position and the first boundary position, as well as the interval between the last boundary position and the current analysis endpoint, are included in the candidate segment set. Each candidate segment is clearly associated with its corresponding tunneling footage range, forming a segment-footprint mapping table for subsequent segment feature extraction and label generation.

[0064] Considering that short-term operating disturbances or local parameter fluctuations during actual construction may lead to overly dense boundary identification results in space, resulting in short candidate segments, this step calculates the segment length for each candidate segment. The segment length is determined by the difference between the segment's ending advance and its starting advance. The segment length is compared with a preset minimum segment length. When the segment length is less than the preset minimum segment length, the segment is considered not to have independent operational significance, and adjacent segments are merged. The merging rules can be forward merging, backward merging, or merging based on segment similarity, with the aim of eliminating fragmented segments caused by noise.

[0065] After merging, the remaining segments are reordered and numbered to obtain multiple operational segments. Each operational segment has continuity and a minimum length constraint in the tunneling advance dimension, thus ensuring the stability and rationality of the segment division structurally. This segment construction method, through boundary sorting, candidate segment generation, and length constraint merging mechanisms, achieves a spatial segmented representation of the structural changes during the tunnel boring process, providing a clear segment basis for subsequent segment representative feature extraction and geological process label determination.

[0066] In one specific implementation, to further enhance the robustness of identifying the boundary positions of operational state changes, the feature representations corresponding to each tunneling advance node in the multidimensional feature space can be used as observation samples to construct a Gaussian Mixture Model (GMM) for probabilistic modeling of the shield tunneling operation state. The GMM is used to characterize the distribution characteristics of multidimensional features under different potential operational states and outputs the posterior responsibility vector corresponding to each tunneling advance node. A state difference index is calculated based on the degree of change in the posterior responsibility vector between adjacent tunneling advance nodes. When the state difference index exceeds a preset change threshold within a consecutive preset number of tunneling advance intervals, the corresponding tunneling advance is determined as the boundary position of the operational state change. In this way, the GMM is not directly used to generate the final segment label, but rather serves as one of the means to detect structural changes in operational state, assisting in identifying the turning points of potential operational states during shield tunneling.

[0067] GMM estimates the parameters (mean μ) of a Gaussian distribution. k Covariance matrix Σ k The data is modeled using a Gaussian distribution and mixed weights π. k Composition. The Gaussian distribution, also known as the normal distribution, has the following probability density function: Where p represents the probability density function, x represents the input data points, μ represents the mean vector, Σ represents the covariance matrix, and D represents the dimension of the data. The mean is μ and the covariance is μ. The probability density value of the D-dimensional multivariate normal distribution at data point x.

[0068] A key step in Gaussian distribution model (GMM) is the iterative optimization of the Gaussian distribution parameters using the EM (Expectation-Maximization) algorithm. E represents the expectation step, which aims to calculate the posterior probability (response) of each data point belonging to each Gaussian distribution, as shown in the following equation: Where k represents the index of the target component being evaluated, j represents the traversal index in the normalization process, and j and k have the same value range (i.e., 1-K). The mixing weights for the j-th Gaussian component; Let be the mean vector corresponding to the j-th Gaussian component in the Gaussian mixture model; Let be the covariance matrix corresponding to the j-th Gaussian component in the Gaussian mixture model; M represents the maximization step, the purpose of which is to update the parameters of the Gaussian distribution and the mixture weights based on the response calculated in the E step, as shown in the following equation.

[0069] Among them, for the above formulas, γ(z) nk N(x) represents the posterior probability (response) that the nth data point (out of a total of N data points) belongs to the kth Gaussian distribution. n |μ k ,Σ k () represents the probability density function of the k-th Gaussian distribution at data point x. n The value at; μ k Indicates the updated mean; Σ k Represents the updated covariance matrix; π k This indicates the updated mixed weights.

[0070] The purpose of running GMM in this application is to determine the optimal classification tree k and output the geological label corresponding to each ring of data. Since GMM is based on a probabilistic model, both BIC (Bayesian Information Criterion) and AIC (Akaike Information Criterion) can be used to measure the effectiveness of data interpretation and penalize model complexity, as shown in the following equation: in, This represents the maximum likelihood value of the model, m represents the total number of model parameters, and n represents the total number of sample points.

[0071] After iterating through k, the corresponding information criterion values ​​for k are obtained, which can help determine a reasonable value for k. Machine learning requires evaluation index (EM) values ​​to further evaluate the algorithm results. Since reliable geological label values ​​are not yet available, three types of internal evaluation indices were selected to evaluate the clustering results (profile coefficient SC, Calinski-Harabasz index CHI, and Davies-Bouldin index DBI), as shown in the following formula: Where i represents the cluster index of the sample point, j is the cluster index variable used to traverse other clusters besides the cluster to which sample point i belongs, j ≠ i, a represents the average distance (cohesion) between sample points in the same cluster, b represents the average distance (separation) from a sample point to sample points in the nearest cluster, k represents the number of clusters, and B k W represents the inter-class scatter matrix. k T represents the within-class scatter matrix. r Let σ represent the sum of the diagonal elements of the matrix, σ represent the standard deviation of the distances from sample points to their respective cluster centers, and d represent the distance between two clusters. For BIC and AIC, smaller values ​​indicate more effective clustering. For SC, CHI, and DBI, values ​​closer to 1, larger values, and smaller values ​​indicate more reliable clustering results, respectively.

[0072] In another possible implementation, when using a Gaussian mixture model to perform soft clustering analysis on the shield tunneling multi-system operating parameters after two-stage preprocessing, a segmented modeling approach based on the continuity of tunneling progress is introduced. Specifically, the shield tunneling multi-system operating parameters are first divided into several continuous progress segments according to the tunneling progress direction. Within each progress segment, the corresponding shield tunneling multi-system operating parameters are input into the Gaussian mixture model for soft clustering analysis. This method allows the Gaussian mixture model to more fully characterize the distribution characteristics of the shield tunneling operation state within local progress segments, avoiding interference with the overall modeling results due to significant differences in operating characteristics before and after the tunneling process.

[0073] After completing the soft clustering analysis within each advance segment, the soft clustering analysis results for each segment are spliced ​​and integrated according to the tunneling advance sequence to form a sequence of soft clustering analysis results covering the entire shield tunneling process. By combining segmented modeling with overall integration, the continuity of soft clustering analysis can be maintained while reflecting the changing characteristics of the shield's operating state at different tunneling stages in greater detail, making the soft clustering analysis results more closely match the actual evolution of the shield's operating state during tunneling.

[0074] It should be noted that the above-described method for identifying the boundary location of operational state changes based on Gaussian mixture models is only one specific implementation of the present invention and does not constitute a limitation on the implementation form of the present invention. Without departing from the technical concept of the present invention, other probability distribution models or state partitioning models can also be used to detect operational state changes in the advance dimension. For example, a model segmentation method based on Bayesian information criteria can be used to determine the boundary by the optimal switching position of model parameters in the advance direction; a state transition probability change detection method based on hidden Markov models can also be used to identify operational state turning points by abrupt changes in the state transition matrix; a segmentation method based on spectral clustering, density peak clustering, or change point detection algorithms can also be used to identify structural abrupt changes in the multidimensional feature space. Any technical solution that determines the boundary of operational state changes based on the distribution structure changes of multidimensional features in the tunneling advance dimension and performs segmentation accordingly should be considered to fall within the protection scope of the present invention.

[0075] Step S40: Extract representative features of each of the aforementioned operating state segments, and determine the corresponding geological process labels based on the representative features of the segments.

[0076] Specifically, feature aggregation operations are performed on the multi-dimensional feature space corresponding to each operational state segment to generate a segment feature set; the contribution of each parameter dimension to the changes within the operational state segment is calculated based on the segment feature set to generate a segment parameter contribution sequence; the segment parameter contribution sequence is sorted to determine the dominant parameter combination of the segment; the main response feature vector of the segment is extracted based on the dominant parameter combination of the segment as the segment representative feature of the operational state segment; and the corresponding geological process label is determined based on the segment representative feature and the preset geological process label mapping rule.

[0077] Furthermore, the extraction rule for the segment main response feature vector is as follows: extract the parameter sequence corresponding to the segment's dominant parameter combination within the corresponding operating state segment; perform segment-level centering processing on the parameter sequence to generate a dominant parameter offset sequence; calculate the pairwise cooperative change coefficients between the parameters based on the dominant parameter offset sequence to generate a dominant parameter cooperative matrix; perform eigenvalue decomposition on the dominant parameter cooperative matrix to extract the main cooperative direction vector; and concatenate the main cooperative direction vector with the mean vector of the corresponding segment's dominant parameters to generate the segment main response feature vector.

[0078] In this embodiment of the invention, each operating state segment corresponds to a continuous tunneling footage range. Within this range, the multidimensional feature space is composed of parameter co-variation matrices at several tunneling footage nodes. Intra-segment feature aggregation operations are performed on the multidimensional feature space within the operating state segment. Statistical aggregation or structural aggregation methods can be used to integrate the feature representations corresponding to each footage node within the segment, forming a segment feature set. This segment feature set is used to characterize the overall structural representation of the operating state within the segment, rather than the instantaneous state at a single point.

[0079] After obtaining the segment feature set, the contribution of each parameter dimension involved in constructing the multidimensional feature space to the overall structural change within the operating state segment is calculated. The contribution can be determined based on the parameter's fluctuation amplitude, frequency of change, or weight influence in the collaborative change matrix, thereby generating a segment parameter contribution sequence. This sequence is then sorted, and parameter dimensions with contributions within a preset range are selected to form the segment's dominant parameter combination. This method filters out parameters with a significant impact on the operating state structure from all parameters, making subsequent feature representation more focused.

[0080] The main response feature vector of the segment is extracted based on the dominant parameter combination of the segment. Within the corresponding operating state segment, the parameter sequence corresponding to the dominant parameter combination is extracted, and this parameter sequence is then centered within the segment. Specifically, the average value of the parameter within the segment is used as a benchmark, and the values ​​at each advance node are offset to obtain the dominant parameter offset sequence. This processing is used to eliminate the influence of the overall amplitude level of the segment and highlight the relative change structure of the parameter within the segment.

[0081] Based on this, pairwise cooperative variation coefficients between the dominant parameters are calculated to construct a dominant parameter cooperative matrix. This cooperative matrix reflects the coupling relationship and common variation trend of each dominant parameter within a segment. Eigenvalue decomposition is performed on the dominant parameter cooperative matrix to extract the eigenvector corresponding to the largest eigenvalue, which serves as the segment's main cooperative direction vector. This main cooperative direction vector is used to characterize the dominant structural direction of multi-parameter cooperative variation within the segment.

[0082] The main cooperative direction vector is concatenated with the mean vector of the corresponding segment's dominant parameters to form the segment's main response feature vector. This segment's main response feature vector simultaneously contains structural relationship information and amplitude level information of the dominant parameters within the segment, enabling it to characterize the features of the operating state segment at both structural and intensity levels.

[0083] The corresponding geological process labels are determined by matching the main response feature vector of the section with a preset geological process label mapping rule. The mapping rule can be determined based on the distribution position of the main response feature vector of the section in the feature space, its similarity to historical section features, or its correspondence with predefined geological response patterns. Through the above steps, a structured mapping from operational sections to geological process labels is achieved, ensuring that the generated labels not only reflect numerical changes but also possess multi-system collaborative structural characteristics, thereby enhancing the stability and interpretability of the labels during continuous tunneling.

[0084] In one specific implementation, for a certain operating state segment Its tunneling advance range is Select within this section One advance node And read the scale uniformity parameter sequence vector corresponding to each advance node. Where p is the number of parameter dimensions. Intra-segment feature aggregation uses a method of convergence based on the set of advance nodes. The segment feature set is defined as... .

[0085] Calculation of the contribution of parameters in section 1 The j-th parameter is used to calculate the energy contribution within the segment: in Let j be the mean of the j-th dimension parameter within the segment. and These are the mean values ​​of the j-th and q-th dimensions of the parameter within the segment, respectively. This yields the segment parameter contribution sequence. .right Sort in descending order, and take the cumulative sum of contributions that reaches the preset cumulative percentage threshold. The first r parameters constitute the dominant parameter combination of the segment. .

[0086] The calculation of the dominant response feature vector of segment k involves extracting the parameter sequence corresponding to the dominant parameter combination within segment k. Centralized processing is performed on it: in, To The parameter vector after centering In the The parameter vector corresponding to the combination of dominant parameters extracted at each advance node. for The mean value represents the amplitude level information of the segment, based on Construct the dominant parameter coordination matrix: right Perform eigenvalue decomposition and select the largest eigenvalue. corresponding unit eigenvector As the main collaborative direction vector, the segment's main response feature vector is constructed according to the concatenation rules: in Provides segment amplitude level information. It provides information on the direction of the segment's collaborative structure, and together the two constitute the representative features of the segment.

[0087] Step S50: Perform consistency verification on the continuity of labels between adjacent operating state sections, and output the geological process label sequence that changes with the tunnel boring machine's advance.

[0088] Specifically, the operating state segments are sorted according to the tunneling footage sequence to generate a segment label sequence; similarity calculation is performed on the main response feature vectors of adjacent operating state segments to generate a segment similarity sequence; the segment similarity sequence is compared with a preset similarity threshold, and when the segment similarity of adjacent operating state segments is greater than the preset similarity threshold, label merging is performed to generate merged segment labels; when the segment similarity of adjacent operating state segments is less than the preset similarity threshold, the segment labels remain unchanged; the segment label sequence after merging is processed by footage expansion to generate a geological process label sequence that corresponds one-to-one with the tunneling footage.

[0089] In this embodiment of the invention, after determining the geological process labels for each operational segment, the operational segments are first sorted according to the tunneling footage, forming a segment label sequence arranged in spatial order. ,in This represents the total number of segments. Each operational segment corresponds to a segment's main response feature vector. This is used to characterize the multi-system collaborative structure features of this section.

[0090] To assess the evolutionary continuity between adjacent segments, adjacent segments are... and The similarity is calculated using the main response feature vectors of the segments. Similarity can be measured using metrics such as cosine similarity or Euclidean distance, for example: Obtain the segment similarity sequence The segment similarity is compared with a preset similarity threshold. Comparison. When If two segments are considered to have high consistency in structural response, then label merging can be performed, i.e., the segments... The tag is updated to a segment. The labels, or the two can be unified into a new merged section label; when At that time, keep the original labels unchanged.

[0091] After the merging process is completed, the updated segment label sequence is subjected to footage unfolding processing. Each segment label is assigned to all footage nodes within its corresponding tunneling footage range, forming a geological process label sequence that corresponds one-to-one with the tunneling footage. Through the above consistency verification and merging mechanism, the frequent switching of labels caused by local fluctuations can be suppressed, ensuring the continuity and stability of the label sequence in the spatial dimension, so that the final output geological process labels are more in line with the actual evolution law of geological response during shield tunneling.

[0092] In another possible implementation, when outputting the geological process label sequence that varies with the tunnel boring machine's (TBM) advance, a label smoothing and reorganization method based on advance segments is introduced to further enhance the continuity and stability of the geological process label sequence. Specifically, in the soft clustering analysis results that have completed credibility verification, the TBM tunneling process is first divided into segments according to a preset advance segment length, so that each segment covers multiple consecutive advance positions.

[0093] Within each tunneling advance segment, the distribution of geological process labels corresponding to each tunneling advance position within that segment is statistically analyzed, and representative geological process labels for that segment are determined based on the statistical results. These representative geological process labels are used to characterize the dominant geological process features of the shield tunneling operation within that advance segment, thereby reducing the impact of label fluctuations at individual tunneling advance positions on the overall sequence continuity.

[0094] Subsequently, the representative geological process labels corresponding to each advance segment are arranged in ascending order of the tunnel boring machine's advance, and associated with the corresponding advance segment, forming a segmented and continuously distributed sequence of geological process labels along the tunnel boring direction. This method preserves the trend of geological process changes while avoiding frequent switching of geological process labels due to fluctuations in local operating parameters, making the output geological process label sequence more consistent with the actual gradual evolution of geological conditions during tunnel boring.

[0095] like Figure 4As shown, this invention provides a geological process tagging system based on shield tunneling operation parameters. The system includes: a data acquisition unit for acquiring shield multi-system operation parameters generated during shield tunneling construction and constructing a multi-dimensional feature space corresponding one-to-one with the shield tunneling advance based on the shield multi-system operation parameters; a preprocessing unit for calculating parameter variation amplitude information between adjacent tunneling advances based on the multi-dimensional feature space and identifying the boundary position of operation state change based on the parameter variation amplitude information; a clustering unit for segmenting the shield tunneling advance based on the boundary position of operation state change to obtain multiple operation state segments; a verification unit for extracting representative features of each operation state segment and determining the corresponding geological process tag based on the representative features of the segment; and an output unit for verifying the consistency of tag continuity between adjacent operation state segments and outputting a geological process tag sequence that changes with the shield tunneling advance.

[0096] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for constructing geological process tags based on tunnel boring machine operating parameters.

[0097] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A06. The network interface A02 is used for communication with external terminals via a network connection. When the processor A01 executes the computer program B02, it implements a method for constructing geological process tags based on tunnel boring machine (TBM) operating parameters.

[0098] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0099] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0100] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for constructing geological process tags based on shield tunneling operation parameters, characterized in that, The method includes: The shield tunneling multi-system operation parameters generated during the shield tunneling process are obtained, and a multi-dimensional feature space corresponding one-to-one with the shield tunneling advance is constructed based on the shield tunneling multi-system operation parameters; The parameter variation range information between adjacent tunneling advances is calculated based on the multidimensional feature space, and the boundary position of the operating status change is identified based on the parameter variation range information. Based on the boundary position of the change in the operating state, the tunnel boring machine advance is segmented to obtain multiple operating state segments; Extract representative features of each of the aforementioned operating state segments, and determine the corresponding geological process labels based on the representative features of the segments; The consistency of labels between adjacent operating sections is checked, and the geological process label sequence that changes with the tunnel boring machine's advance is output.

2. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 1, characterized in that, Obtain the shield tunneling multi-system operating parameters generated during shield tunneling construction, and construct a multi-dimensional feature space corresponding one-to-one with the shield tunneling advance based on the shield tunneling multi-system operating parameters, including: Obtain the shield multi-system operating parameters formed along the tunneling advance direction during shield construction, and establish a parameter-advance mapping relationship for the shield multi-system operating parameters according to the corresponding tunneling advance; According to the preset standard advance interval, the parameter-advance mapping relationship is subjected to advance resampling processing, and the shield tunneling multi-system operating parameters with different sampling frequencies are uniformly mapped to the same advance node to form an advance alignment parameter sequence. Based on a preset response delay coefficient, the operating parameters of each subsystem in the advance alignment parameter sequence are subjected to advance offset correction processing to generate a delay correction parameter sequence. The delay correction parameter sequence is subjected to dimensionless processing, and the change direction of each parameter is uniformly transformed to generate a scale-unified parameter sequence. Based on the scale-unified parameter sequence, a parameter co-variation matrix is ​​constructed, and a one-to-one correspondence is established between the parameter co-variation matrix and the corresponding tunneling advance, thereby constructing a multi-dimensional feature space for characterizing the changes in the operating state during shield tunneling.

3. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 1, characterized in that, Based on the multidimensional feature space, the parameter variation range information between adjacent tunneling advances is calculated, and based on the parameter variation range information, the location of the operational status change boundary is identified, including: Perform a difference operation on the parameter co-change matrix corresponding to adjacent tunneling advances in the multidimensional feature space to obtain a multidimensional change vector sequence; A sign consistency analysis is performed on the change in each dimension of the multidimensional change vector sequence, and the difference between the number of dimensions of change in the same direction and the number of dimensions of change in opposite directions is calculated to generate a cooperative change direction index. A weighted summation operation is performed on the changes in each dimension of the multidimensional change vector sequence to obtain the collaborative change intensity index; A normalization fusion operation is performed on the coordinated change direction index and the coordinated change intensity index to generate a multidimensional coordinated change index sequence. The cumulative change energy is calculated for the multidimensional collaborative change index sequence within a continuous preset number of tunneling advance intervals, and when the cumulative change energy exceeds a preset change threshold, the corresponding tunneling advance is determined as the boundary position of the operating state change.

4. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 3, characterized in that, Perform a normalization fusion operation on the coordinated change direction index and the coordinated change intensity index to generate a multidimensional coordinated change index sequence, including: The coordinated change direction index sequence is subjected to interval normalization processing to obtain a direction normalization index sequence, and the direction normalization index sequence is limited to a preset direction index range; The coordinated change intensity index sequence is subjected to interval normalization processing to obtain an intensity normalized index sequence, and the intensity normalized index sequence is limited to a preset intensity index range; A linear fusion operation is performed on the direction-normalized index sequence and the intensity-normalized index sequence based on a preset fusion weight to obtain a fusion synergy index sequence; The fusion synergy index sequence is subjected to saturation pruning to obtain a pruned synergy index sequence, and the pruned synergy index sequence is limited to a preset synergy index range; The trimmed synergistic index sequence is subjected to sliding smoothing to generate a multidimensional synergistic change index sequence.

5. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 1, characterized in that, Based on the boundary positions of the aforementioned operational state changes, the tunnel boring machine advance is segmented to obtain multiple operational state segments, including: The boundary positions of the changes in the operating state are sorted according to the tunneling advance sequence to generate a boundary position sequence; Based on the boundary position sequence, candidate segments are constructed between adjacent boundary positions, and a correspondence is established between each candidate segment and the corresponding tunneling footage range; Calculate the segment length for each candidate segment, and merge adjacent segments for candidate segments whose segment length is less than the preset minimum segment length; The merged set of segments is determined as multiple running status segments.

6. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 1, characterized in that, Extracting representative features from each of the aforementioned operational state segments, and determining corresponding geological process labels based on these representative features, including: Perform intra-segment feature aggregation operation on the multi-dimensional feature space corresponding to each operating state segment to generate segment feature set; Based on the segment feature set, the contribution of each parameter dimension to the change in the operating state segment is calculated, and a segment parameter contribution sequence is generated. The contribution sequence of the segment parameters is sorted to determine the combination of dominant parameters for the segment; Based on the combination of dominant parameters of the segment, the main response feature vector of the segment is extracted as the segment representative feature of the operating state segment; Based on the representative features of the section and the preset geological process label mapping rules, the corresponding geological process label is determined.

7. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 6, characterized in that, The extraction rules for the main response feature vector of a section are as follows: Extract the parameter sequence corresponding to the dominant parameter combination of the corresponding operating state segment; Perform intra-segment centering on the parameter sequence to generate a dominant parameter offset sequence; Based on the dominant parameter offset sequence, calculate the cooperative change coefficients between each pair of parameters to generate the dominant parameter cooperative matrix; Perform eigenvalue decomposition on the dominant parameter cooperative matrix to extract the main cooperative direction vector; The main collaborative direction vector is concatenated with the mean vector of the dominant parameters of the corresponding segment to generate the segment main response feature vector.

8. The method for constructing geological process tags based on shield tunneling operation parameters according to claim 6, characterized in that, The consistency of labels between adjacent operational sections is checked, and the geological process label sequence that changes with the tunnel boring machine's advance is output, including: The operating status sections are sorted according to the tunneling footage sequence to generate a section label sequence; Perform similarity calculation on the main response feature vectors of adjacent operating state segments to generate a segment similarity sequence; The segment similarity sequence is compared with a preset similarity threshold. When the segment similarity of adjacent running state segments is greater than the preset similarity threshold, a label merging process is performed to generate merged segment labels. When the segment similarity between adjacent operating state segments is less than the preset similarity threshold, the segment label remains unchanged; After the merging process is completed, the segment label sequence is subjected to advance expansion processing to generate a geological process label sequence that corresponds one-to-one with the tunneling advance.

9. A geological process tagging system based on shield tunneling operation parameters, characterized in that, The system includes: The acquisition unit is used to acquire the shield tunneling multi-system operating parameters generated during the shield tunneling process, and to construct a multi-dimensional feature space that corresponds one-to-one with the shield tunneling advance based on the shield tunneling multi-system operating parameters; The preprocessing unit is used to calculate the parameter variation range information between adjacent tunneling advances based on the multidimensional feature space, and to identify the boundary position of the operating status change based on the parameter variation range information; Clustering units are used to segment the tunnel boring machine advance based on the location of the boundary of the change in the operating state, so as to obtain multiple operating state segments; The verification unit is used to extract representative features of each of the said operating state segments and determine the corresponding geological process label based on the representative features of the segments; The output unit is used to verify the consistency of the label continuity between adjacent operating state sections and output the geological process label sequence that changes with the tunnel boring machine's advance.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the geological process tagging method based on shield tunneling operation parameters as described in any one of claims 1-8.

Images