A soil state analysis method and system based on complex loose body excavation

Abstract

The application discloses a soil state analysis method and system based on complex loose body tunneling, relates to the technical field of soil state monitoring, and comprises the following steps: collecting spiral conveying associated operation data in the tunneling process of a shield machine according to a unified time base, reading excavation section parameters, and generating spiral conveying time sequence data; screening an apparent stable soil dumping interval from the spiral conveying time sequence data according to soil dumping metering stability, and writing interval marks; calling load response data, soil dumping metering data and bin pressure response data in the apparent stable soil dumping interval, calculating soil dumping load and bin pressure decay deviation; combining advancing forming data, soil dumping metering data and load disturbance data to form implicit deviation characteristics; judging abnormal states of soil in a complex loose body according to the matching relationship between the implicit deviation characteristics and a state triggering rule set; and writing back the apparent stable soil dumping interval according to the judged state and confidence, and generating a corrected soil state analysis result.

Description

Technical Field

[0001] This invention relates to the field of soil condition monitoring technology, specifically to a method and system for soil condition analysis based on excavation of complex loose bodies. Background Technology

[0002] With the development of tunnel boring machine (TBM) construction monitoring and intelligent tunneling control technologies, TBM operation data has gradually expanded from single construction records to multi-source time-series data fusion analysis. Addressing the heterogeneous, compressible, and rearrangeable characteristics of complex loose materials within the soil chamber and spiral conveying channel, related technologies are beginning to focus on load changes, chamber pressure attenuation, volume balance, and micro-fluctuation characteristics during the soil removal process to aid in judging changes in soil condition during TBM tunneling.

[0003] Existing soil condition analysis techniques typically rely on single parameters such as dumping rate, silo pressure, or screw conveyor load for judgment, making it difficult to reveal the internal state changes of complex loose masses under seemingly stable dumping conditions. When the dumping rate and cumulative dumping volume remain stable, traditional methods tend to directly interpret this as a stable soil state, ignoring latent anomalies caused by soil compression, water precipitation, pore collapse, or particle rearrangement. Existing techniques lack sophisticated processing for time synchronization, missing data compensation, and noise retention of data from different sources, easily misinterpreting continuous high-amplitude load fluctuations as noise and failing to establish reference benchmarks for silo pressure attenuation based on dumping rate gradations, resulting in incomparable pressure responses under different dumping rhythms. Traditional analysis lacks multi-feature triggering rules and confidence level write-back mechanisms for soil compression, water precipitation, pore collapse, and particle rearrangement. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing methods for analyzing soil conditions during tunnel boring machine (TBM) excavation are prone to misjudging seemingly stable soil discharge as stable soil conditions, have difficulty identifying hidden anomalies inside complex loose bodies based on load, silo pressure, soil discharge measurement, and propulsion shaping data, lack multi-feature joint discrimination and confidence correction mechanisms for soil compression, water precipitation, pore collapse, and particle rearrangement, and lack a systematic analysis of the soil condition inside complex loose bodies during spiral conveying.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a soil condition analysis method based on tunneling of complex loose bodies, comprising: collecting spiral conveyor-related operational data during the tunneling process of a tunnel boring machine (TBM) according to a unified time reference, reading excavation section parameters, and generating spiral conveyor time-series data; filtering apparent stable soil discharge intervals from the spiral conveyor time-series data based on soil discharge metering stability, and writing interval markers; calling load response data, soil discharge metering data, and silo pressure response data within the apparent stable soil discharge intervals to calculate the deviations in soil discharge load and silo pressure attenuation; combining propulsion shaping data, soil discharge metering data, and load disturbance data to form implicit deviation features; determining abnormal soil conditions within the complex loose body based on the matching relationship between the implicit deviation features and the set of state triggering rules; and writing back the apparent stable soil discharge intervals based on the determined state and confidence level to generate corrected soil condition analysis results.

[0007] As a preferred embodiment of the soil condition analysis method for complex loose body tunneling described in this invention, the generation of spiral conveyor time-series data includes: collecting spiral conveyor-related operation data formed by the complex loose body in the soil chamber and spiral conveyor channel; the spiral conveyor-related operation data includes spiral conveyor operation data, soil discharge metering data, chamber pressure response data, and propulsion shaping data; the spiral conveyor operation data includes spiral conveyor speed, spiral conveyor torque, and spiral conveyor current; the soil discharge metering data includes soil discharge rate and cumulative soil discharge volume; the chamber pressure response data includes chamber pressure; the propulsion shaping data includes propulsion speed; the excavation cross-section parameters include the excavation cross-section area or equivalent excavation area used in the calculation of theoretical excavation volume; the cumulative soil discharge volume is calculated by time alignment, missing data filling, and sliding noise reduction of the soil discharge rate, along... The cumulative soil discharge volume is obtained by summing the data on the reference time axis. During the continuous tunneling process of the tunnel boring machine, the rotational speed, torque, and current are read from the screw conveyor control record; the soil discharge rate and cumulative soil discharge volume are read from the soil discharge metering record; the soil chamber pressure is read from the soil chamber pressure acquisition point; the propulsion speed is read from the propulsion control record, and the excavation section parameters are read from the construction parameter record. A reference time axis is established with a unified sampling period, and data from different sources are mapped to the reference time axis according to their temporal proximity. Missing data segments are marked as invalid and filled in according to the valid sampled values ​​before and after. Sliding noise reduction processing is performed on torque, current, soil discharge rate, and soil chamber pressure. The data are written into a data table according to the method that the same sampling time corresponds to a set of screw conveyor associated operation data and excavation section parameters, forming screw conveyor time sequence data corresponding to the screw conveyor process of complex loose materials.

[0008] As a preferred embodiment of the soil condition analysis method based on complex loose body tunneling described in this invention, the step of writing interval markers includes: setting a sliding time window in the spiral conveyor time series data; calling up soil discharge measurement data within each sliding time window, and determining the fluctuation state of the soil discharge rate of the complex loose body based on the average level and dispersion of the soil discharge rate; when the fluctuation ratio of the dispersion of the soil discharge rate relative to the average level is less than a preset soil discharge stability threshold, writing the corresponding sliding time window into a candidate stability marker; after writing the candidate stability marker, performing a monotonicity check on the cumulative soil discharge volume based on the change sequence of the cumulative soil discharge volume with the sampling time; when the cumulative soil discharge volume at the later sampling time is less than the previous sampling time... When calculating the cumulative soil discharge at a sampling time, the corresponding sampling interval is marked as a measurement anomaly interval. When the cumulative soil discharge meets the condition of monotonically non-decreasing, it is determined whether the increase in soil discharge between adjacent sampling times does not exceed the preset jump limit to determine the continuity of the cumulative soil discharge of complex loose bodies. When the candidate stability marker, the non-decreasing condition of cumulative soil discharge, and the continuity condition of cumulative soil discharge are all met, the corresponding sliding time window is confirmed as the apparent stable soil discharge interval of complex loose bodies. When the time interval between two adjacent apparent stable soil discharge intervals is lower than the preset merging interval, and the difference in the average level of soil discharge rate in the two intervals is lower than the preset mean difference threshold, the two intervals are merged and written into the same apparent stable soil discharge interval marker.

[0009] As a preferred embodiment of the soil condition analysis method based on complex loose soil excavation described in this invention, the calculation of the soil discharge load and pressure decay deviation includes: within each seemingly stable soil discharge interval, calling torque, current, and soil discharge rate, using the zero-value correction value of the soil discharge rate as a normalization benchmark, and calculating the torque load and current load corresponding to the unit volume of complex loose soil discharged; determining the changing trend of the torque load and current load corresponding to the unit volume of complex loose soil discharged according to the sampling time, and recording the changing patterns of continuous rise, staged abrupt changes, and intermittent peaks; calling the pressure response data within the same seemingly stable soil discharge interval, and forming the complex loose soil pressure according to the decrease in soil pressure per unit time. Attenuation trend; compare the silo pressure attenuation trend with the silo pressure attenuation reference benchmark under the preset rate class corresponding to the average level of the soil discharge rate; the silo pressure attenuation reference benchmark is the reference change relationship formed by the change of the silo pressure decrease over time in the time segment when the average level of the soil discharge rate is in the same preset rate class and is not written into the silo pressure attenuation deviation record; when the silo pressure attenuation trend meets the preset silo pressure attenuation deviation judgment condition, record the silo pressure attenuation deviation type and deviation duration, and write the corresponding result into the unit soil discharge load and silo pressure attenuation deviation; when the silo pressure attenuation trend does not meet the preset silo pressure attenuation deviation judgment condition, mark the corresponding time segment as the silo pressure attenuation non-deviation segment, and record the silo pressure attenuation deviation type as non-deviation.

[0010] As a preferred embodiment of the soil condition analysis method for excavation of complex loose bodies described in this invention, the process of forming the implicit deviation characteristics includes: calling the advance speed and excavation section parameters within the apparent stable soil dumping interval; multiplying the advance distance accumulated by the advance speed over time with the excavation section parameters to obtain the theoretical excavation volume of the complex loose body; calling the cumulative soil dumping volume at the start and end times of the same apparent stable soil dumping interval to obtain the actual soil dumping volume of the complex loose body; comparing the difference between the theoretical excavation volume and the actual soil dumping volume to form a volume balance deviation; performing moving average processing on torque and current respectively to obtain the low-frequency load change trend; and subtracting the corresponding low-frequency load change trend from the original torque and original current to obtain torque micro-fluctuation and current micro-fluctuation; recording the micro-fluctuation based on the degree of deviation of the micro-fluctuation amplitude from the corresponding window's conventional amplitude. The system analyzes the fluctuation intensity, peak occurrence frequency, and peak time distribution. It normalizes and encodes volume balance deviation, load change trends corresponding to the unit volume of soil discharged, silo pressure attenuation deviation type, torque micro-fluctuation characteristics, and current micro-fluctuation characteristics in a preset order to form implicit deviation features characterizing the internal changes of complex loose bodies. The torque and current micro-fluctuation characteristics respectively include the intensity, peak occurrence frequency, and peak time distribution of the corresponding micro-fluctuations. The normalization encoding uses the characteristic benchmark range corresponding to historical normal soil discharge intervals or stable segments of the current tunneling section as the comparison object. Features exceeding the upper limit of the characteristic benchmark range are encoded as positive deviations; features below the lower limit of the characteristic benchmark range are encoded as negative deviations; features within the characteristic benchmark range are encoded as no deviations, and the deviation level is determined according to the ratio of the deviation amount to the width of the characteristic benchmark range.

[0011] As a preferred embodiment of the soil condition analysis method based on complex loose body excavation described in this invention, the method for identifying abnormal soil conditions within the complex loose body includes: establishing a set of state triggering rules and confidence generation rules based on the latent deviation characteristics of the complex loose body extracted from historical normal soil discharge intervals, on-site verification abnormal intervals, and test calibration intervals; the set of state triggering rules includes soil compression triggering rules, water precipitation triggering rules, pore collapse triggering rules, and particle rearrangement triggering rules; when one of the torque load and current load corresponding to the unit volume of complex loose body soil discharged shows a continuous increase, the pressure decay trend is lower than the lower limit of the pressure decay reference benchmark and continues to exceed a preset time, and the theoretical excavation volume deviates continuously in a positive direction from the actual soil discharge volume, the soil compression triggering rule is matched to generate a soil compression candidate state; when one of the torque load and current load corresponding to the unit volume of complex loose body soil discharged shows a staged abrupt change, and the current micro-fluctuation is strong When the pressure rises and the deviation of the pressure decay changes direction between adjacent windows, the water precipitation trigger rule is matched to generate a water precipitation candidate state; when the soil pressure in the soil chamber drops suddenly within a short time window, and the torque micro-fluctuation or current micro-fluctuation shows a peak in the same time neighborhood, and the soil discharge rate does not show a corresponding continuous increase, the pore collapse trigger rule is matched to generate a pore collapse candidate state; when the torque micro-fluctuation and current micro-fluctuation show a peak sequence with adjacent peak time intervals within a preset interval range in multiple consecutive preset time windows, and the pressure decay deviation changes intermittently, the particle rearrangement trigger rule is matched to generate a particle rearrangement candidate state; for each candidate state, a comprehensive score is given according to the number of trigger features, deviation level, duration, and time consistency, and the comprehensive score is converted into a corresponding confidence level according to the confidence level generation rule; when the corresponding confidence level reaches the preset state threshold, the corresponding candidate state is confirmed as an abnormal state of soil inside a complex loose body.

[0012] As a preferred embodiment of the soil condition analysis method based on complex loose body tunneling described in this invention, the generation of the corrected soil condition analysis results includes: after each apparent stable dumping interval, reading the corresponding discrimination states and confidence levels of soil compression, water precipitation, pore collapse, and particle rearrangement in the complex loose body; when the confidence levels of all discrimination states are lower than a preset state threshold, retaining the dumping stability marker of the corresponding apparent stable dumping interval and writing an internal soil no-anomaly marker; when the confidence level of any discrimination state reaches the preset state threshold, correcting the corresponding apparent stable dumping interval from a single dumping stability marker. The system identifies anomalies within the complex loose body associated with apparent soil dumping stability. When the confidence levels of two or more discriminant states reach a preset state threshold, the primary state and accompanying state are determined according to their confidence levels, and a composite state label is written when the confidence level difference is below a preset distinction threshold. When the same primary state appears repeatedly in multiple consecutive apparent stable soil dumping intervals, a persistent state label is written; when the same primary state appears only in a single apparent stable soil dumping interval, an instantaneous state label is written. The output includes the interval start and end times, soil dumping stability labels, corrected soil conditions of the complex loose body, primary state, accompanying state, and confidence level.

[0013] As a preferred embodiment of the soil condition analysis system based on complex loose body tunneling described in this invention, the system includes: a data acquisition and calibration module, an interval screening module, a load attenuation module, a deviation construction module, a state discrimination module, and a result correction module. The data acquisition and calibration module collects the spiral conveyor-related operational data during the tunneling process of the tunnel boring machine (TBM) according to a unified time reference, reads the excavation section parameters, and generates spiral conveyor time-series data. The interval screening module filters seemingly stable soil discharge intervals from the spiral conveyor time-series data based on soil discharge metering stability and writes interval markers. The load attenuation module calls the load response data, soil discharge metering data, and silo pressure response data within the seemingly stable soil discharge intervals to calculate the deviations in soil discharge load and silo pressure attenuation. The deviation construction module combines the propulsion forming data, soil discharge metering data, and load disturbance data to form implicit deviation features. The state discrimination module determines the abnormal soil state inside the complex loose body based on the matching relationship between the implicit deviation features and the state triggering rule set. The result correction module rewrites the seemingly stable soil discharge intervals based on the determined state and confidence level to generate corrected soil condition analysis results.

[0014] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a method for analyzing the soil condition based on tunneling through complex loose bodies.

[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a soil condition analysis method based on tunneling through complex loose bodies.

[0016] The beneficial effects of this invention are as follows: The soil condition analysis method for tunneling in complex loose bodies provided by this invention forms a progressive analysis chain for the internal state of complex loose bodies through unified time-series data construction, screening of apparent stable dumping intervals, calculation of deviation between unit dumping load and chamber pressure attenuation, construction of volume balance and micro-fluctuation latent features, multi-state triggering rule discrimination, and confidence level write-back correction. This transforms internal soil changes, which are difficult to observe directly, into calculable, comparable, and traceable latent deviation features. By combining multiple features for triggering and confidence level correction, the risk of misjudgment based on a single parameter is reduced, enabling re-discrimination of the soil condition in apparent stable dumping intervals. This provides more accurate state data for adjusting shield tunneling parameters, controlling dumping rhythm, controlling chamber pressure, and improving slag. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Fig. 1 The present invention provides an overall flowchart of a soil condition analysis method for tunneling in complex loose bodies.

[0019] Fig. 2 Soil anomaly map provided by the present invention for a soil state analysis method based on complex loose body tunneling.

[0020] Fig. 3 A schematic diagram of a computer device provided by the present invention. Detailed Implementation

[0021] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0022] Reference Figs. 1-3 As an embodiment of the present invention, a method for soil condition analysis based on tunneling through complex loose bodies is provided, comprising:

[0023] S1: Collect the associated operation data of the screw conveyor during the tunneling process of the shield machine according to a unified time benchmark, and read the excavation section parameters to generate screw conveyor timing data.

[0024] Furthermore, the generation of spiral conveyor timing data includes collecting spiral conveyor-related operational data of complex loose materials formed within the soil chamber and spiral conveyor channel. This spiral conveyor-related operational data includes spiral conveyor operation data, soil discharge metering data, chamber pressure response data, and propulsion shaping data. The spiral conveyor operation data includes spiral conveyor speed, spiral conveyor torque, and spiral conveyor current. Soil discharge metering data includes soil discharge rate and cumulative soil discharge volume. Chamber pressure response data includes soil chamber pressure. Propulsion shaping data includes propulsion speed. Excavation cross-section parameters include the excavation cross-section area or equivalent excavation area used in the theoretical excavation volume calculation. The cumulative soil discharge volume is obtained by time alignment, missing data filling, and sliding noise reduction of the soil discharge rate, followed by accumulation along the reference time axis. (The last sentence appears to be incomplete and possibly refers to a separate process involving a tunnel boring machine.) During continuous tunneling, the rotational speed, torque, and current are read from the screw conveyor control record; the soil discharge rate and cumulative soil discharge volume are read from the soil discharge metering record; the soil chamber pressure is read from the soil chamber pressure acquisition point; the propulsion speed is read from the propulsion control record; and the excavation section parameters are read from the construction parameter record. A reference time axis is established with a unified sampling period, and data from different sources are mapped to the reference time axis according to their temporal proximity. Missing data segments are marked as invalid and filled in according to the valid sampling values ​​before and after. Sliding noise reduction processing is performed on torque, current, soil discharge rate, and soil chamber pressure. The screw conveying associated operation data and excavation section parameters corresponding to the same sampling time are written into a data table to form screw conveying time sequence data corresponding to the screw conveying process of complex loose materials.

[0025] It should be noted that one specific scheme for generating screw conveyor timing data includes, in this embodiment, the screw conveyor speed, screw conveyor torque, and screw conveyor current are read from the screw conveyor control record or drive controller; the soil discharge rate and cumulative soil discharge volume are read from the belt scale, slag metering device, or soil discharge metering record; the soil chamber pressure is read from the soil chamber pressure collection point; when there are multiple soil chamber pressure collection points, the pressure measurement point corresponding to the screw conveyor inlet area is used, or the average or weighted average of multiple pressure measurement points is used; the propulsion speed is read from the propulsion control record; and the excavation cross-sectional area or equivalent excavation area is read from the construction parameter record. To enable joint judgment of data from different sources at the same time, this embodiment establishes a reference time axis with a unified sampling period, represented as follows:

[0026]

[0027] in, Indicates the first At each reference sampling time, Indicates the start time of data analysis. Indicates a uniform sampling period. Indicates the maximum sampling sequence number within the analysis period. This indicates the sampling point index on the reference time axis. The uniform sampling period ranges from 1 second to 5 seconds. If the sampling period is too large, the peak load of the screw conveyor and the short-term sudden drop in soil pressure are easily smoothed out; if the sampling period is too small, the noise from the field sensors will increase the data processing burden. In this embodiment, considering the refresh frequency of the tunnel boring machine control record and the speed of change of the screw conveyor load, the uniform sampling period is preferably 2 seconds.

[0028] For data from different sources with inconsistent sampling times, the corresponding sampled values ​​are mapped to the reference time axis according to their temporal proximity. For any parameter, the most recent valid sampled value is selected before and after the reference sampling time. If the time difference between the most recent valid sampled value and the reference sampling time does not exceed half a unified sampling period, the corresponding sampled value is written to the reference sampling time; if it exceeds half a unified sampling period, the corresponding position is temporarily marked as missing. For data segments with consecutive missing durations not exceeding a preset missing duration, linear padding is performed using valid sampled values ​​before and after the missing data segment; for data segments with consecutive missing durations exceeding the preset missing duration, an invalid flag is written. The preset missing duration is preferably 4 unified sampling periods.

[0029] After time mapping, sliding noise reduction was performed on the screw conveyor torque, screw conveyor current, soil discharge rate, and soil chamber pressure. Sliding noise reduction used a sliding median approach with a window length of 3 to 7 sampling points. When a sample value undergoes a sudden change relative to the median level within an adjacent window, and the preceding and following sample values ​​do not support continuous change in the same direction, the corresponding sample value is discarded as an instantaneous jump value or replaced with the median value within the window. For screw conveyor torque and current, continuously occurring high-amplitude fluctuations are retained and not deleted as noise.

[0030] Data is written into a data table according to the method of corresponding a set of spiral conveyor operation data and excavation cross-section parameters at the same sampling time, forming spiral conveyor time sequence data for the spiral conveying process of complex loose materials. In the spiral conveyor time sequence data, each sampling time includes spiral conveyor speed, spiral conveyor torque, spiral conveyor current, soil discharge rate, cumulative soil discharge volume, soil chamber pressure, advance speed, excavation cross-section parameters, and validity markers.

[0031] It should also be noted that by uniformly mapping the associated operational data of the screw conveyor to the same reference time axis, and distinguishing between short-term missing, long-term missing, and instantaneous jump data during the mapping process, screw conveyor time-series data from different acquisition sources are formed. This allows a correspondence to be established between load response, soil discharge metering, silo pressure changes, and propulsion shaping at the same sampling point, avoiding misjudgments of the state due to sampling delays or data misalignment. This improves the temporal consistency of multi-source construction data and the reproducibility of state analysis, and reduces the risk of anomalies within complex loose bodies being masked by temporal mismatches or excessive noise reduction.

[0032] S2: Based on the stability of soil discharge metering, select the apparent stable soil discharge intervals from the screw conveyor timing data and write the interval markers.

[0033] Furthermore, the process of writing interval markers includes: setting a sliding time window in the spiral conveyor timing data; calling up the soil discharge measurement data within each sliding time window, and determining the fluctuation state of the soil discharge rate of complex loose materials based on the average level and dispersion of the soil discharge rate; when the fluctuation ratio of the dispersion of the soil discharge rate relative to the average level is less than a preset soil discharge stability threshold, writing the corresponding sliding time window into a candidate stability marker; after writing the candidate stability marker, performing a monotonicity check on the cumulative soil discharge volume based on the change sequence of the cumulative soil discharge volume with the sampling time; when the cumulative soil discharge volume at a later sampling time is less than the cumulative soil discharge volume at a previous sampling time, the corresponding sampling time window is marked as a candidate stability marker. The sample interval is marked as a measurement anomaly interval; when the cumulative soil discharge meets the monotonically non-decreasing condition, it is determined whether the increase in soil discharge between adjacent sampling times does not exceed the preset jump upper limit, thus determining the continuity of the cumulative soil discharge of complex loose bodies; when the candidate stability marker, the non-decreasing condition of cumulative soil discharge, and the continuous condition of cumulative soil discharge are all met simultaneously, the corresponding sliding time window is confirmed as the apparent stable soil discharge interval of complex loose bodies; when the time interval between two adjacent apparent stable soil discharge intervals is lower than the preset merging interval, and the difference in the average level of soil discharge rate in the two intervals is lower than the preset mean difference threshold, the two intervals are merged and written into the same apparent stable soil discharge interval marker.

[0034] It should be noted that one specific scheme for writing interval markers includes setting a sliding time window in the screw conveyor timing data. The length of the sliding time window covers multiple complete rotation cycles of the screw conveyor, while being less than the time required for adjusting the propulsion parameters to form a new stable response. Preferably, the sliding time window length is 30 to 120 seconds, and the window step size is 5 to 20 seconds. In this embodiment, a window length of 60 seconds and a window step size of 10 seconds are used. Let the first... Each sliding time window is The window contains The valid sampling point, the first one within the window The soil removal rate at each benchmark sampling point is The average level, dispersion, and fluctuation ratio of the soil dumping rate are expressed as follows:

[0035]

[0036]

[0037]

[0038] in, Indicates the first The average level of soil removal rate within a sliding time window Indicates the sequence number of the sliding time window. Indicates the first Number of valid sampling points within a sliding time window Indicates the first The degree of dispersion of the soil removal rate within each sliding time window This represents the soil removal rate parameter. Indicates the first The fluctuation ratio of soil removal rate within a sliding time window This represents the correction constant for the zero value of the soil dumping rate. Take 1% to 5% of the median value of the normal soil dumping rate on site, with 2% being preferred.

[0039] When the soil dumping rate fluctuates more Not higher than the preset soil dumping stability threshold At that time, the corresponding sliding time window is written into the candidate stability marker. Due to phenomena such as particle slippage, local bridging, and instantaneous collapse during the disposal of complex loose soil, this embodiment presets a soil disposal stability threshold. The range is 0.05 to 0.15, preferably 0.10. After writing the candidate stability marker, it is determined whether the cumulative soil discharge of the complex loose body is continuous. Specifically, within the same sliding time window, the difference in cumulative soil discharge at adjacent sampling times is recorded as... , represented as:

[0040]

[0041] in, Indicates the first The cumulative soil discharge at each benchmark sampling point This represents the cumulative amount of soil removed at the previous baseline sampling point. This represents the increment in soil discharge volume between adjacent sampling points. If the increment in soil discharge volume... Not less than the negative tolerance If the cumulative soil discharge volume may experience minor adverse fluctuations due to meter zeroing, belt scale vibration, or data rounding, it is considered to be within the allowable metering tolerance range, and the cumulative soil discharge volume is deemed to meet the non-decreasing condition. Indicates the soil discharge volume parameter; if Greater than the preset jump limit If this occurs, it is considered that there is an abnormal jump in the increase in soil discharge. Negative tolerance. To absorb minor reverse fluctuations caused by metering zeroing, belt scale vibration, or data rounding, the optimal value is 1% of the theoretical soil discharge volume within the window. Preset jump upper limit. Take 1.1 times the 95th percentile of the historical normal soil dumping increment.

[0042] When the corresponding sliding time window has been written with a candidate stability marker, and the incremental amount of soil discharge between adjacent benchmark sampling points within the window is not less than the negative tolerance and not greater than the preset jump limit, the corresponding sliding time window is confirmed as the apparent stable soil discharge interval of the complex loose body.

[0043] When the time interval between two adjacent seemingly stable spoil disposal intervals is less than a preset merging interval, and the difference in the average spoil disposal rate between the two intervals is less than a preset mean difference threshold, the two intervals are merged and written into the same seemingly stable spoil disposal interval label. The preset merging interval is preferably a sliding window step size, and the preset mean difference threshold is preferably 8% of the average spoil disposal rate.

[0044] It should also be noted that by calculating the average level, dispersion, and fluctuation ratio of the dumping rate through a sliding time window, and combining this with the non-decreasing nature of the cumulative dumping volume and the range of jumps in adjacent dumping volume increments, apparent stable dumping intervals for complex loose bodies are screened. Using measurement stability as a boundary condition, time intervals with low dumping rate fluctuations, continuous cumulative dumping volume, and no abnormal abrupt changes in dumping increments are extracted, focusing the internal state assessment on objects where "the dumping surface is stable, but the soil interior may have undergone changes." Adjacent apparent stable dumping intervals are then merged based on the time interval and the difference in the mean dumping rate, avoiding excessive segmentation of the same continuous dumping process and improving the engineering continuity of interval markings.

[0045] S3: Call the load response data, dumping metering data and silo pressure response data within the apparent stable dumping range to calculate the dumping load and silo pressure attenuation deviation.

[0046] Furthermore, the calculation of the deviation between the soil discharge load and the silo pressure decay includes: within each seemingly stable soil discharge interval, retrieving torque, current, and soil discharge rate; using the zero-corrected value of the soil discharge rate as a normalized benchmark; calculating the torque load and current load corresponding to the discharge of a unit volume of complex loose soil; determining the variation trend of the torque load and current load corresponding to the discharge of a unit volume of complex loose soil according to the sampling time, and recording the variation patterns of continuous rise, staged abrupt changes, and intermittent peaks; retrieving silo pressure response data within the same seemingly stable soil discharge interval, forming a silo pressure decay trend for complex loose soil based on the decrease in silo pressure per unit time; and comparing the silo pressure decay trend with the corresponding... The average level of soil discharge rate is compared with the reference benchmark for silo pressure attenuation under the preset rate range. The reference benchmark for silo pressure attenuation is the reference relationship formed by the change of silo pressure attenuation over time in the time segment where the average level of soil discharge rate is in the same preset rate range and is not recorded in the silo pressure attenuation deviation record. When the silo pressure attenuation trend meets the preset silo pressure attenuation deviation judgment condition, the silo pressure attenuation deviation type and deviation duration are recorded, and the corresponding results are written into the unit soil discharge load and silo pressure attenuation deviation. When the silo pressure attenuation trend does not meet the preset silo pressure attenuation deviation judgment condition, the corresponding time segment is marked as a silo pressure attenuation non-deviation segment, and the silo pressure attenuation deviation type is recorded as non-deviation.

[0047] It should be noted that one method for calculating the deviation between the soil discharge load and the silo pressure attenuation specifically includes, within each seemingly stable soil discharge interval: calling the screw conveyor torque, screw conveyor current, and soil discharge rate to determine whether the screw conveyor load undergoes abnormal changes when discharging similar amounts of soil. Let the first... The torque of the screw conveyor at each reference sampling point is The current of the screw conveyor is The soil removal rate is The torque load and current load corresponding to the discharge of a unit volume of complex loose soil are expressed as follows:

[0048]

[0049] in, Indicates the first The torque load corresponding to a unit volume of complex loose soil is determined by sampling at each benchmark point. Indicates the first The current load corresponding to a unit volume of complex loose soil discharged from each benchmark sampling point. To determine the load variation trend, the apparent stable soil discharge interval was analyzed. and The time trend is calculated separately. This embodiment uses the least squares slope to avoid misjudgment caused by fluctuations at a single sampling point. For the first... The load sequence corresponding to each reference sampling point , It can be or The trend slope is expressed as:

[0050]

[0051] in, Represents the load sequence In the The slope of change within the apparent stable soil dumping interval Indicates the sequence number of the apparent stable soil dumping interval. Indicates the first A set of benchmark sampling point indices within an apparent stable spoil heap. Indicates the first The average value of sampling times within each apparently stable spoil disposal interval. Indicates the first The average value of the load sequence within each apparent stable dumping interval. When for hour, This represents the slope of the torque load change corresponding to the discharge of a unit volume of complex loose soil; when for hour, This represents the slope of the change in current load corresponding to the unit volume of complex loose soil discharged.

[0052] If the slope of the load change Greater than the corresponding continuously rising threshold If this state persists for at least three consecutive baseline sampling points, then the corresponding load sequence is recorded as a continuously rising pattern. When the difference between the average load values ​​of two adjacent sub-windows exceeds the stage mutation threshold... When the corresponding load sequence is recorded as a staged mutation pattern, and the local peak amplitude exceeds the peak threshold... When there is no continuous peak sequence between adjacent peaks, the corresponding load sequence is recorded as an intermittent peak pattern.

[0053] The sustained rise threshold, the stage mutation threshold, and the peak threshold are determined based on the slope distribution of similar load sequences in historical normal dumping intervals, the mean difference distribution of adjacent sub-windows, and the local peak amplitude distribution, respectively. For any load sequence , for or The continuously rising threshold is denoted as The stage mutation threshold is denoted as The peak threshold is denoted as , represented as:

[0054]

[0055]

[0056]

[0057] in, Represents the load sequence The continuously rising threshold and These represent the mean and standard deviation of the load change slope during the historical normal soil dumping interval, respectively. Represents the load sequence The stage mutation threshold, and These represent the mean and standard deviation of the difference in average load between adjacent sub-windows within the historical normal waste dumping interval, respectively. Represents the load sequence The peak threshold, and These represent the mean and standard deviation of the local peak amplitude within the historical normal soil dumping interval, respectively. In this embodiment, the three thresholds are preferably set as fixed judgment values ​​using the mean plus twice the standard deviation. When the on-site verification results indicate that there is an overall drift in the historical normal samples, only the historical sample set corresponding to the mean and standard deviation is updated, without changing the threshold coefficient of twice the standard deviation.

[0058] Within the same seemingly stable spoil disposal zone, silo pressure response data is used to generate a complex loose body silo pressure decay trend based on the rate of decrease in silo pressure per unit time. Let the starting time of the zone be... The end time is The corresponding earth pressures are as follows: and The trend of decreasing inventory pressure within the interval can be expressed as:

[0059]

[0060] in, Indicates the first The trend of decreasing silo pressure within the apparent stable soil dumping range. Indicates the pressure response parameters. and These represent the soil pressure at the beginning and end of the interval, respectively. When the soil pressure increases... It is a negative value; when the pressure in the earthwork chamber decreases, It is a positive value. To avoid the situation where the pressure decay of the soil chamber under different soil discharge rates cannot be directly compared, this embodiment first relies on the first... An apparent stable spoil disposal area The average level of soil removal rate within the area determines the corresponding rate grading. Specifically, let the first... An apparent stable spoil disposal area The average level of soil removal rate within the area is expressed as:

[0061]

[0062] in, Indicates the first A set of benchmark sampling point indices within an apparent stable spoil heap. Indicates the first Number of effective sampling points within an apparent stable spoil heap area Indicates the first The average disposal rate within a historically stable disposal range. The average disposal rate within a historically normal disposal range is then divided according to its quantile. Each preset rate level, preferred Let the first... The preset rate levels are: Then the first Rate classification of each apparent stable spoil heap zone Represented as:

[0063]

[0064] in, Indicates the first The preset rate grading number to which each apparently stable spoil disposal zone belongs. Indicates the number of the preset rate level. Indicates the first A preset rate level and These represent the lower and upper boundaries of the rate band, respectively. For the first... Each preset rate level is selected from historical normal spoil disposal intervals or stable segments of the current tunneling section, with the average spoil disposal rate falling into the range. Furthermore, the reference interval for recording the deviation of pressure decay was not included, thus forming a reference interval set. .for For each reference interval, the corresponding pressure decay trend is calculated, and a pressure decay reference benchmark is formed for that rate tier. Preferably, a set of reference intervals is used. The median value of the decreasing trend of the central warehouse pressure is used as the reference center value, and the allowable fluctuation range is formed by using the interquartile range, as shown below:

[0065]

[0066]

[0067] in, Indicates the first Reference center value for pressure attenuation under a preset rate range Represents the median function. Indicates the first The lower limit of the reference benchmark for warehouse pressure attenuation under each preset rate level. Indicates the lower limit of the reference benchmark. Indicates the first Upper limit of the reference benchmark for pressure decay under each preset rate level Indicates the upper limit of the reference benchmark. Represents the set of reference intervals The Middle The trend of decreasing pressure within a reference range and These represent the 25th percentile and the 75th percentile, respectively. Indicates the reference interval index. This represents the pressure decay parameter.

[0068] When the The rate classification of each apparent stable spoil heap is as follows: At that time, the trend of decreasing inventory pressure in this range will be observed. Compare with the reference benchmark under the corresponding rate bracket. If it meets the requirements... If this state persists for more than a preset duration, it will be recorded as the pressure decay being lower than the reference baseline. Indicates the first The decreasing trend of silo pressure in an apparent stable soil dumping zone. Indicates the first Rate classification of each apparent stable spoil heap zone The lower limit of the reference benchmark for warehouse pressure decay; if it meets the following conditions. If this state persists for more than a preset duration, it will be recorded as a pressure decay exceeding the reference baseline. Indicates the first Rate classification of each apparent stable spoil heap zone The upper limit of the pressure decay reference benchmark is set below. The preset duration is preferably three consecutive sliding sub-windows. For short-term sudden drops, this embodiment calculates the pressure decrease of the earth chamber within a preset short-term window. Let the starting point of the short-term window be the [number missing]. There are 1 reference sampling points, and the short-time window length is 1. For each sampling period, the short-term pressure drop is expressed as:

[0069]

[0070] in, Indicates from the first From the first reference sampling point to the second The decrease in soil pressure between the benchmark sampling points Indicates the first The soil pressure corresponding to each benchmark sampling point Indicates the first The soil pressure corresponding to each benchmark sampling point This indicates the number of sampling intervals contained within the short-time window. If... Greater than the preset drop threshold If this occurs, it is recorded as a short-term sudden drop in warehouse pressure. A preset sudden drop threshold is set. The optimal value is the 95th percentile of the historical normal short-term pressure drop range. This avoids misinterpreting normal, slow pressure release as a sudden drop, while also identifying short-term anomalies caused by pore collapse, local blockage, or rapid pressure release. For each type of silo pressure decay deviation, the corresponding deviation type, deviation start time, deviation end time, and deviation duration are recorded, and the corresponding results are entered into the unit discharge load and silo pressure decay deviation for complex loose materials.

[0071] It should also be noted that within the apparent stable discharge range, the discharge rate, after zero-value correction, is used as the normalized benchmark. The torque and current loads corresponding to the discharge unit volume of complex loose soil are calculated separately. Further analysis is conducted on load variations such as continuous increases, staged abrupt changes, or intermittent peaks. This transforms the original load of the screw conveyor from "instantaneous output" to "mechanical response required per unit discharge volume," reducing the direct impact of discharge rate differences on torque and current comparisons. This approach is more suitable for reflecting changes in transport conditions such as soil compaction, retention, blockage trends, or increased particle action. The silo pressure decay trend is categorized according to the discharge rate within the apparent stable discharge range and compared with a silo pressure decay reference benchmark under the same rate category, avoiding the inability to directly compare pressure release amplitudes under different discharge rhythms. By recording deviation types and durations such as silo pressure decay below or above the benchmark, or short-term sudden drops, physically meaningful load and pressure deviation inputs are provided for identifying internal anomalies, improving the sensitivity of identifying compaction retention, abnormal pressure release, and abrupt pore changes.

[0072] S4: Combining propulsion forming data, soil discharge measurement data, and load disturbance data, a hidden deviation characteristic is formed.

[0073] Furthermore, the formation process of the latent deviation characteristics includes: within the apparent stable spoil disposal zone, the advance speed and excavation section parameters are called; the advance distance accumulated by the advance speed over time is multiplied by the excavation section parameters to obtain the theoretical excavation volume of the complex loose body; the cumulative spoil disposal volume at the start and end times of the same apparent stable spoil disposal zone is called to obtain the actual spoil disposal volume of the complex loose body, and the difference between the theoretical excavation volume and the actual spoil disposal volume is compared to form the volume balance deviation; moving average processing is performed on torque and current respectively to obtain the low-frequency load change trend, and the corresponding low-frequency load change trend is subtracted from the original torque and original current to obtain torque micro-fluctuation and current micro-fluctuation; based on the degree of deviation of the micro-fluctuation amplitude from the corresponding window's normal amplitude, the micro-fluctuation intensity, the number of peak occurrences, and the peak time are recorded. The distribution is as follows: the volume balance deviation, the load change trend corresponding to the unit volume of soil discharged, the type of silo pressure attenuation deviation, the torque micro-fluctuation characteristics, and the current micro-fluctuation characteristics are normalized and coded in a preset order to form implicit deviation characteristics that characterize the internal changes of complex loose bodies; the torque micro-fluctuation characteristics and the current micro-fluctuation characteristics include the intensity of the corresponding micro-fluctuation, the number of peak occurrences, and the peak time distribution, respectively; the normalization coding uses the characteristic benchmark range corresponding to the historical normal soil discharge interval or the stable segment of the current tunneling section as the comparison object; features above the upper limit of the characteristic benchmark range are coded as positive deviations; features below the lower limit of the characteristic benchmark range are coded as negative deviations; features within the characteristic benchmark range are coded as no deviations, and the deviation level is determined according to the ratio of the deviation amount to the width of the characteristic benchmark range.

[0074] It should be noted that a preferred scheme for forming implicit deviation characteristics representing internal soil changes specifically includes, within the apparent stable soil dumping zone, adjusting the advance rate and excavation cross-sectional parameters to calculate the theoretical excavation volume of the complex loose body. Let the excavation cross-sectional area or equivalent excavation area be... , No. The advance speed at each sampling moment is Then the first The theoretical excavation volume within an apparent stable spoil disposal zone is expressed as follows:

[0075]

[0076] in, Indicates the first Theoretical excavation volume of complex loose body within an apparent stable spoil heap area. This represents the excavated cross-sectional area or equivalent excavated area. By calling the cumulative discharge volume at the start and end times of the same apparent stable discharge interval, the actual discharge volume for complex loose bodies can be obtained.

[0077]

[0078] in, Indicates the first Actual amount of soil discharged from complex loose bodies within an apparent stable discharge zone. and These represent the cumulative excavation volume at the start and end of the interval, respectively. Further, the normalized difference between the theoretical excavation volume and the actual excavation volume is compared to form the volume balance deviation, expressed as:

[0079]

[0080] in, Indicates the first Deviation in volume balance within an apparent stable spoil disposal zone This represents the volume zero-value correction constant, preferably taken as 3% of the median value of the theoretical excavation volume within its normal range. When The value is consistently positive and deviates positively from the volume equilibrium threshold. When the theoretical excavation volume is greater than the actual excavation volume, complex loose materials may undergo compression, retention, or pore collapse within the soil chamber or screw conveyor channel; when The value remains negative and is less than the negative deviation threshold of volume balance. When the actual amount of soil removed exceeds the theoretical excavation volume, it indicates that there is an abnormality in the release or measurement of previously retained soil, and the volume balance deviates positively from the threshold. The preferred value is 0.1, which indicates a negative deviation from the threshold in volume balance. The preferred value is -0.1.

[0081] For load disturbance data, this embodiment performs moving average processing on the screw conveyor torque and screw conveyor current respectively to obtain the low-frequency load change trend. The corresponding low-frequency load change trend is then subtracted from the original torque and current to obtain torque micro-fluctuations and current micro-fluctuations. For the torque signal, the low-frequency trend and micro-fluctuations can be expressed as:

[0082]

[0083] in, Indicates the first The torque low-frequency load variation trend at each reference sampling point Indicates the first in the window The original torque at each reference sampling point Represents the moving average window The index of the reference sampling points within, Indicates the first A moving average window centered on each sampling time. This indicates the number of valid sampling points within the window. Indicates the first The torque fluctuates slightly at each sampling time. Indicates micro-fluctuation. Indicates the first in the window The original torque at each reference sampling point Indicates the first The torque low-frequency load variation trend at each reference sampling point. Micro-current fluctuations at each sampling time The original current is used in the same manner. The result is obtained by subtracting the low-frequency trend of the current. The moving average window preferably covers seven propeller rotation cycles. If the window is too short, the low-frequency trend will be mixed with particle impact signals; if the window is too long, propulsion control changes will be mistaken for micro-fluctuations. After obtaining the torque and current micro-fluctuations, the intensity, frequency of peak occurrences, and peak time distribution of the micro-fluctuations are calculated respectively. The intensity of the micro-fluctuations is preferably represented by the root mean square amplitude.

[0084]

[0085] in, Indicates the first Torque fluctuation intensity within the apparent stable soil dumping range Indicates the first The intensity of current fluctuations within the apparent stable soil discharge interval.

[0086] When determining the peak value, for the first... Any micro-fluctuation sequence within an apparent stable dumping interval Make a judgment. The peak value threshold is determined based on the statistical results of similar micro-fluctuation amplitudes in historical normal spoil disposal intervals or stable segments of the current tunneling section. In this embodiment, the mean plus twice the standard deviation is preferably used as a fixed threshold. Sampling points that meet the condition of exceeding the threshold and forming a local maximum relative to the preceding and following sampling points are determined as micro-fluctuation peak points, expressed as:

[0087]

[0088] in, The type is The threshold of micro-fluctuations and These represent the mean and standard deviation of similar micro-fluctuations in the historical normal spoil disposal interval or the current stable segment of the tunneling section, respectively. Indicates the first The type within the apparent stable spoil heap area is: The set of micro-fluctuation peak points.

[0089] Based on the set of peak points Statistically analyze the frequency and timing distribution of peak occurrences. The frequency of peak occurrences is a set. The number of elements in the middle, and the peak time distribution are ordered sets formed by arranging the sampling times corresponding to each peak point in chronological order. In order to determine whether the peak has periodicity related to the screw conveying process, this embodiment determines the blade passing cycle based on the average speed of the screw conveyor and the number of screw blades, and uses one to three blade passing cycles as the preset interval range of adjacent peaks, expressed as:

[0090]

[0091] in, Indicates the first The helical blades pass through a period of time within an apparent stable soil discharge zone. Indicates the first Average rotational speed of the screw conveyor within an apparent stable soil discharge zone; Indicates the number of helical blades; and This indicates the sampling time corresponding to two adjacent micro-fluctuation peaks.

[0092] When the same type of micro-fluctuation peak meets the adjacent peak interval condition within two consecutive preset time windows, the corresponding micro-fluctuation forms a peak sequence. For peak sequences formed by torque micro-fluctuations, torque micro-fluctuation features are written; for peak sequences formed by current micro-fluctuations, current micro-fluctuation features are written. The above peak occurrence frequency, peak time distribution, and peak sequence results are used for subsequent implicit deviation feature encoding and particle rearrangement trigger rule judgment.

[0093] The volume balance deviation, the load change trend corresponding to the unit volume of complex loose soil discharged, the type of silo pressure attenuation deviation, the torque micro-fluctuation characteristics, and the current micro-fluctuation characteristics are normalized and coded in a preset order to form implicit deviation characteristics characterizing the internal changes of complex loose soil. The normalization coding uses historical normal soil discharge intervals or stable segments of the current tunneling section as benchmarks. For any numerical characteristic... Normalization is used to obtain , represented as:

[0094]

[0095] in, Indicates the first Within the apparent stable spoil disposal zone, the first The normalization results of the features, Indicates the feature item number participating in the normalization coding. Indicates the first Within the apparent stable spoil disposal zone, the first Original eigenvalues, Indicates the first The reference center value of the feature in the reference interval. Indicates the first The scale parameter of the feature in the baseline interval, Indicates the first The correction constants corresponding to the feature terms. The normalization results of each term are combined in a preset order to obtain the first... The latent deviation feature vector corresponding to each apparent stable dumping interval .

[0096] Further record the direction and level of deviation for each feature. The direction of deviation is determined by... The sign of the deviation is determined, and the deviation level is divided into no deviation, slight deviation, moderate deviation, and severe deviation according to the magnitude of the absolute value. When When the deviation is less than the first-level threshold, it is recorded as no deviation or slight deviation. A deviation between the first-level threshold and the second-level threshold is considered moderate. Deviations exceeding the second-level threshold are considered severe. The first-level and second-level thresholds are preferably set to 1 and 2, respectively, corresponding to deviation levels of one and two scale units.

[0097] It should also be noted that the theoretical excavation volume is obtained by accumulating the advance speed and excavation section parameters, and then compared with the actual excavation volume obtained from the cumulative excavation volume within the same interval to form a volume balance deviation. Simultaneously, low-frequency trends in the torque and current of the screw conveyor are separated using moving averages, extracting torque micro-fluctuations, current micro-fluctuations, their peak frequency, and peak time distribution. The internal state of complex loose bodies is expanded from a single excavation volume or single load characteristic to a joint description of volume balance, mechanical load trends, pressure decay deviation, and micro-fluctuation structure. Normalization encoding is performed based on historical normal excavation intervals or current stable segments, allowing volume, pressure, load, and fluctuation characteristics of different dimensions to enter the same discrimination scale, while retaining the deviation direction and deviation level. The internal changes of complex loose bodies that cannot be directly observed are transformed into implicit deviation feature vectors, providing quantifiable, comparable, and traceable feature inputs for state classification.

[0098] S5: Based on the matching relationship between latent deviation characteristics and the set of state triggering rules, determine the abnormal state of soil inside complex loose bodies.

[0099] Furthermore, refer to Fig. 2The method for identifying abnormal soil states within complex loose soil masses includes establishing a set of state triggering rules and confidence generation rules based on latent deviation characteristics of complex loose soil masses extracted from historical normal soil discharge intervals, on-site verification abnormal intervals, and test calibration intervals. The state triggering rule set includes soil compression triggering rules, water content precipitation triggering rules, pore collapse triggering rules, and particle rearrangement triggering rules. When either the torque load or current load corresponding to a unit volume of complex loose soil discharged shows a continuous increase, the silo pressure decay trend is lower than the lower limit of the silo pressure decay reference benchmark and continues for more than a preset time, and the theoretical excavation volume deviates continuously in a positive direction from the actual soil discharge volume, the soil compression triggering rule is matched to generate a soil compression candidate state. When either the torque load or current load corresponding to a unit volume of complex loose soil discharged experiences a staged abrupt change, the intensity of current micro-fluctuations increases, and the silo pressure decay deviation occurs between adjacent windows... When the direction of soil growth changes, a water precipitation trigger rule is matched to generate a water precipitation candidate state. When the soil pressure in the soil chamber suddenly drops within a short time window, and the torque or current micro-fluctuation spikes in the same time neighborhood, and the soil discharge rate does not show a corresponding continuous increase, a pore collapse trigger rule is matched to generate a pore collapse candidate state. When the torque and current micro-fluctuations show a peak sequence with adjacent peak time intervals within a preset interval range in multiple consecutive preset time windows, and the pressure decay deviation changes intermittently, a particle rearrangement trigger rule is matched to generate a particle rearrangement candidate state. For each candidate state, a comprehensive score is given according to the number of trigger features, deviation level, duration, and time consistency, and the comprehensive score is converted into a corresponding confidence level according to the confidence level generation rule. When the corresponding confidence level reaches a preset state threshold, the corresponding candidate state is confirmed as an abnormal state of soil inside a complex loose body.

[0100] It should be noted that one method for identifying abnormal soil conditions within complex loose masses specifically includes, after obtaining the first... The latent deviation feature vector corresponding to each apparent stable dumping interval Subsequently, based on the hidden deviation characteristics of complex loose bodies extracted from historical normal spoilage intervals, on-site verification anomaly intervals, or test calibration intervals, a set of state triggering rules is established. Historical normal spoilage intervals are historical time intervals where spoilage metering is stable, silo pressure decay does not exceed the allowable range of the reference benchmark, and no abnormal spiral spoilage occurs in the construction records. On-site verification anomaly intervals are time intervals where anomalies are confirmed through on-site spoilage observation, construction logs, manual sampling, or equipment alarms. Test calibration intervals are time intervals where soil compression, water precipitation, pore collapse, or particle rearrangement is confirmed through indoor or on-site spiral conveying tests.

[0101] The set of state triggering rules includes soil compression triggering rules, water precipitation triggering rules, pore collapse triggering rules, and particle rearrangement triggering rules. Soil compression, water precipitation, pore collapse, and particle rearrangement are abnormal soil states within complex loose materials that occur during compression, decompression, transportation, or disturbance within the soil chamber and spiral conveyor channel. To avoid misjudgment due to a single parameter, this embodiment employs a multi-feature combination triggering method. For soil compression candidate states, this embodiment sets a soil compression triggering flag. When the first A candidate soil compression state is generated when an apparent stable soil dumping interval simultaneously satisfies the following conditions: a continuous increase in the load of the discharged unit complex loose soil, a decrease in silo pressure below the reference baseline, and a continuous positive deviation from the volume balance. This state is represented as follows:

[0102]

[0103] in, Indicates the first Whether an apparent stable soil dumping interval matches the soil compression triggering rule, with a value of 1 indicating a match and a value of 0 indicating a mismatch; Indicates the first The slope of the change in torque load corresponding to the unit volume of complex loose soil discharged within an apparent stable soil discharge interval. This indicates the threshold for a sustained increase in torque load. Indicates the first The slope of the current load change corresponding to the unit volume of complex loose soil discharged within a seemingly stable soil discharge interval. This indicates the threshold for a sustained increase in current load. and The mean plus twice the standard deviation was used as the fixed decision value. or This indicates an increase in the load required to discharge a unit volume of complex loose soil. This indicates that the pressure reduction trend is below the lower limit of the reference benchmark under the corresponding rate tier; This indicates that the theoretical excavation volume is greater than the actual soil discharge volume and the deviation reaches the preset level; when both conditions are met, it means that the complex loose body may be compacted, retained, or insufficiently discharged in the soil chamber or spiral conveyor channel.

[0104] For the candidate state of water precipitation, this embodiment sets a water precipitation trigger flag. , Indicates the first Does the apparent stable soil discharge interval match the water content precipitation triggering rule? Water content precipitation typically manifests as phased load changes, increased micro-current fluctuations, and a change in the direction of pressure decay between adjacent intervals. To express the change in the direction of pressure decay, the deviation of the pressure decay center is defined as:

[0105]

[0106] in, Indicates the first The deviation of the silo pressure decay trend of an apparent stable spoil disposal zone from the reference center value of its corresponding rate bracket. Indicates the first The pressure decay trend of each apparent stable dumping zone is relative to the reference center value of pressure decay under the corresponding rate bracket. When Compared with the previous apparent stable spoil disposal zone When the signs are opposite and the deviation exceeds the preset pressure direction change threshold, it is determined that the pressure decay deviation has changed direction. Indicates the first The deviation of the silo pressure decay trend in an apparent stable dumping zone from the reference center value of its corresponding rate bracket. The water precipitation triggering rule is expressed as follows:

[0107]

[0108] in, Indicates the first Markers of stage-specific abrupt changes in torque load within an apparent stable soil discharge zone. Indicates the first Markers of stage-specific abrupt changes in current load within an apparent stable soil discharge zone; The threshold value for representing the intensity of micro-current fluctuations is the mean plus twice the standard deviation, which is used as a fixed judgment value. This indicates a change in the direction of pressure decay deviation between adjacent seemingly stable dumping zones. When both conditions are met... or , as well as This indicates that the water content of complex loose materials may change, leading to phased changes in the relationship between adhesion, flowability, and pressure release.

[0109] For pore collapse candidate states, this embodiment sets a pore collapse trigger flag. , Indicates the first Does the apparent stable soil discharge interval match the pore collapse triggering rule? Pore collapse is not simply manifested as an increase in soil discharge volume, but often as a short-term sudden drop in soil chamber pressure, a spike in load fluctuations, and a non-synchronous and continuous increase in the soil discharge rate. Let the first... The maximum short-term soil pressure decrease within the apparent stable soil dumping interval is expressed as:

[0110]

[0111] in, Indicates the first The maximum decrease in soil pressure within a preset short-term window within an apparent stable soil dumping interval. This represents the maximum value function. The pore collapse triggering rule is expressed as:

[0112]

[0113] in, Indicates the adjacent first The average change in the dumping rate between apparent stable dumping intervals; The threshold for torque micro-fluctuation intensity is represented by the mean plus twice the standard deviation as a fixed judgment value. The threshold for a sustained increase in the soil discharge rate is preferably 5% of the average soil discharge rate in the previous apparent stable soil discharge interval. When the soil chamber pressure experiences a short-term sudden drop and the load fluctuations increase synchronously, but the soil discharge rate does not show a corresponding sustained increase, the pressure release is more likely to originate from the collapse of the complex loose body pore structure or the destruction of local channels, rather than the reduction in chamber pressure caused by normal continuous soil discharge.

[0114] For particle rearrangement candidate states, this embodiment sets a particle rearrangement trigger flag. Particle rearrangement typically exhibits quasi-periodic load peaks and intermittent pressure deviations. Let... Indicates the first The torque peak sequence markers for each apparent stable soil dumping interval Indicates the first The current peak sequence is marked for each apparent stable soil discharge interval; when a torque peak sequence or current peak sequence exists, the corresponding mark is set to 1. Further settings include a silo pressure intermittent deviation mark. When the first When there are discontinuous deviation segments in silo pressure decay within multiple sub-windows of an apparent stable spoil disposal interval, ,otherwise The particle rearrangement triggering rule is expressed as follows:

[0115]

[0116] in, Indicates the first Does the apparent stable soil displacement zone match the particle rearrangement triggering rule? Indicates the candidate state of particle rearrangement; Indicates the first The deviation of the silo pressure decay in the apparent stable soil dumping zone shows intermittent changes. When the first Each apparent stable dumping interval satisfies the torque peak sequence marking. or And the warehouse pressure deviates intermittently from the mark. This indicates that complex loose particles may repeatedly engage, disengage, and re-aggregate within the spiral groove, resulting in a quasi-periodic distribution of the load peaks. Therefore, the first... The set of candidate states corresponding to each apparent stable spoil disposal interval can be represented as:

[0117]

[0118] in, Indicates the first A set of candidate states generated from an apparent stable spoil heap. If If it is an empty set, it means that no candidate states of internal soil anomalies have been generated in the apparently stable soil dumping interval; if If a condition contains one or more states, the corresponding candidate states and trigger markers are input into the subsequent confidence level generation rules. For each candidate state, a comprehensive score is calculated based on the number of trigger features, deviation level, duration, and time consistency. The comprehensive score is then converted into the corresponding confidence level according to the confidence level generation rules. Let the anomaly state type of the soil within a complex loose mass be... , It can represent soil compression, water precipitation, pore collapse, or particle rearrangement; the corresponding state triggering rules include If there are triggering features, the comprehensive score of the candidate state can be expressed as:

[0119]

[0120] in, Indicates the first State type in an apparent stable spoil disposal zone Overall score Indicates the first State type in an apparent stable spoil disposal zone The corresponding number of triggering features, This indicates the trigger feature number in the state triggering rule. Indicates the first Each triggering feature in the state type The weights in express In the apparent stable spoil disposal zone, the first Whether a trigger feature is triggered, the value is 1 if triggered, and 0 if not triggered; express In the apparent stable spoil disposal zone, the first The deviation level normalized value of each trigger feature; express In the apparent stable spoil disposal zone, the first The duration coefficient of each triggering feature; express In the apparent stable spoil disposal zone, the first The temporal consistency coefficient between each triggering feature and other triggering features. The corresponding confidence level is obtained using a normalized comprehensive score, expressed as:

[0121]

[0122] in, Indicates state type confidence level This represents the minimum value function. When the corresponding confidence level reaches a preset state threshold, the corresponding candidate state is confirmed as an abnormal state of the soil inside the complex loose body. The preset state threshold is preferably 0.7.

[0123] It should also be noted that, based on the implicit deviation feature vector, state triggering rules for soil compression, water precipitation, pore collapse, and particle rearrangement are constructed separately, and combined judgments are made on changes in unit discharge load, deviation of silo pressure attenuation, deviation of volume balance, intensity of micro-fluctuations, peak sequence, and changes in discharge rate. This avoids the crude approach of directly concluding soil state based on the increase or decrease of a single parameter, ensuring that different abnormal states correspond to different feature combinations. Candidate states are comprehensively scored according to the number of triggering features, deviation level, duration, and temporal consistency, and converted into confidence scores. This makes state confirmation no longer a simple rule hit, but reflects the degree of support of multiple features for the same state. It can reduce the probability of single-factor misjudgment when there are sensor fluctuations, short-term pressure disturbances, or local load spikes, improve the distinguishability between soil compression, water precipitation, pore collapse, and particle rearrangement, and generate interpretable state confidence results.

[0124] S6: Based on the discrimination state and confidence level, write back the apparent stable soil dumping interval to generate the corrected soil condition analysis results.

[0125] Furthermore, the generated corrected soil condition analysis results include: after each seemingly stable dumping interval, reading the corresponding discrimination states and confidence levels for soil compression, water precipitation, pore collapse, and particle rearrangement in the complex loose body; when the confidence levels of all discrimination states are lower than the preset state threshold, retaining the dumping stability marker for the corresponding seemingly stable dumping interval and writing an internal soil anomaly marker; when the confidence level of any discrimination state reaches the preset state threshold, correcting the corresponding seemingly stable dumping interval from a single dumping stability marker to one where apparent dumping stability is accompanied by internal anomalies in the complex loose body. Constant labeling; when the confidence of two or more discrimination states reaches the preset state threshold, the main state and the accompanying state are determined according to the confidence level, and a composite state label is written when the confidence difference is lower than the preset discrimination threshold; when the same main state appears repeatedly in multiple seemingly stable dumping intervals, a persistent state label is written; when the same main state appears only in a single seemingly stable dumping interval, an instantaneous state label is written; the output includes the interval start and end time, dumping stability label, corrected complex loose soil state, main state, accompanying state, and confidence level of the soil state analysis results.

[0126] It should be noted that one specific scheme for generating the corrected soil condition analysis results includes, after each seemingly stable dumping interval, reading the corresponding discrimination states and confidence levels for soil compression, water precipitation, pore collapse, and particle rearrangement in the complex loose body. If the confidence levels of all discrimination states are lower than a preset state threshold, the dumping stability marker for the corresponding seemingly stable dumping interval is retained, and an internal soil anomaly marker is added. The output results indicate that the complex loose body not only remains stable in dumping measurement within the corresponding time interval, but also does not exhibit any internal anomalies supported by load, silo pressure, volume balance, and micro-fluctuations.

[0127] If the confidence level of any discrimination state reaches the preset state threshold, the corresponding apparent stable dumping interval is corrected from a single dumping stability marker to an apparent dumping stability marker accompanied by anomalies within complex loose bodies. Although the dumping rate and cumulative dumping volume appear stable, the internal soil state is not directly determined to be stable, but rather written back as an internal anomaly state based on the identification results; the preset state threshold ranges from 0.6 to 0.8, preferably 0.7, and is continuously corrected based on the on-site verification results.

[0128] When the confidence levels of two or more discrimination states reach a preset state threshold, the primary state and accompanying states are determined according to their confidence levels. The maximum confidence level among the four state categories corresponds to the primary state, and other states reaching the preset state threshold are considered accompanying states. When the difference between the confidence level of the primary state and the confidence level of the second-highest state is less than a preset discrimination threshold, a composite state label is written. The preset discrimination threshold ranges from 0.05 to 0.15, preferably 0.1.

[0129] When the same primary state occurs repeatedly within multiple seemingly stable spoil disposal intervals, a persistent state marker is written; when the same primary state occurs only within a single seemingly stable spoil disposal interval, an instantaneous state marker is written. Multiple seemingly stable spoil disposal intervals are preferably two adjacent intervals. Based on this, the soil condition analysis results are output. The soil condition analysis results include the interval start and end times, spoil disposal stability markers, corrected complex loose soil condition, primary state, accompanying state, confidence level, and persistent or instantaneous state markers. The output results can serve as the data basis for adjusting shield tunneling parameters, spoil disposal rhythm, soil chamber pressure control, or spoil improvement strategies.

[0130] It should also be noted that the discrimination state and confidence level are written back to the corresponding apparent stable spoil heap interval, and the original spoil heap stability markers are corrected according to the preset state thresholds. When the confidence levels of all states are below the threshold, the internal soil quality is not abnormally marked; when the confidence level of any state reaches the threshold, the single spoil heap stability marker is corrected to an apparent spoil heap stability accompanied by an anomaly marker inside the complex loose body, so that spoil heap measurement stability is no longer directly equivalent to soil quality stability. For intervals where multiple anomalous states reach the threshold simultaneously, the main state and accompanying state are determined according to the confidence level, and a composite state marker is written when the confidence level difference is small, which can express the superposition of anomalies inside the complex loose body; when the same main state appears repeatedly in multiple consecutive apparent stable spoil heap intervals, a persistent state marker is written, otherwise an instantaneous state marker is written, thereby distinguishing between continuous evolution anomalies and occasional disturbances.

[0131] One embodiment of the present invention provides a soil condition analysis system based on complex loose body tunneling, including a data acquisition and calibration module, an interval screening module, a load attenuation module, a deviation construction module, a condition discrimination module, and a result correction module.

[0132] The system comprises the following modules: Acquisition and Calibration Module: Acquires spiral conveyor-related operational data during the tunnel boring machine's excavation process according to a unified time benchmark, reads excavation section parameters, and generates spiral conveyor time-series data; Interval Filtering Module: Filters seemingly stable soil discharge intervals from the spiral conveyor time-series data based on soil discharge metering stability and writes interval markers; Load Attenuation Module: Calls load response data, soil discharge metering data, and silo pressure response data within the seemingly stable soil discharge intervals to calculate the deviations in soil discharge load and silo pressure attenuation; Deviation Construction Module: Combines propulsion shaping data, soil discharge metering data, and load disturbance data to form implicit deviation features characterizing internal soil changes; State Discrimination Module: Based on the matching relationship between implicit deviation features and the set of state triggering rules, identifies abnormal soil states within complex loose bodies; Result Correction Module: Writes back the seemingly stable soil discharge intervals based on the identified state and confidence level to generate corrected soil state analysis results.

[0133] Reference Fig. 3This embodiment also provides a computer device applicable to the soil condition analysis method based on complex loose body tunneling, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the soil condition analysis method based on complex loose body tunneling as proposed in the above embodiment.

[0134] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0135] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the soil condition analysis method for complex loose body tunneling as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

Claims

1. A method for analyzing soil condition during tunneling in complex loose soil, characterized in that, include: Collect the associated operation data of the screw conveyor during the tunnel boring machine excavation process according to a unified time benchmark, and read the excavation section parameters to generate screw conveyor time sequence data; Based on the stability of the soil dumping measurement, the apparent stable soil dumping intervals are selected from the screw conveyor time series data and marked as intervals. Call the load response data, dumping metering data and silo pressure response data within the apparent stable dumping range to calculate the dumping load and silo pressure attenuation deviation; By combining propulsion shaping data, soil discharge measurement data, and load disturbance data, implicit deviation characteristics are formed; Based on the matching relationship between latent deviation characteristics and the set of state triggering rules, the abnormal state of soil inside complex loose bodies is determined; Based on the discrimination state and confidence level, the apparent stable soil dumping interval is rewritten to generate the corrected soil condition analysis results.

2. The soil condition analysis method based on complex loose mass tunneling as described in claim 1, characterized in that: The generated spiral conveyor timing data includes... Collect spiral conveying-related operational data of complex loose materials formed in the soil bin and spiral conveying channel. The spiral conveying-related operational data includes spiral conveying operation data, soil discharge metering data, bin pressure response data, and propulsion shaping data. The screw conveyor operation data includes screw conveyor speed, screw conveyor torque, and screw conveyor current; the soil discharge metering data includes soil discharge rate and cumulative soil discharge volume; the silo pressure response data includes silo pressure; the propulsion shaping data includes propulsion speed; and the excavation cross-section parameters include the excavation cross-section area or equivalent excavation area used in the calculation of theoretical excavation volume. The cumulative soil discharge volume is obtained by adding up the soil discharge rate along the reference time axis after time alignment, missing data filling and sliding noise reduction. During the continuous tunneling process of the tunnel boring machine, the rotational speed, torque and current are read from the screw conveyor control record; the soil discharge rate and cumulative soil discharge volume are read from the soil discharge metering record; the soil chamber pressure is read from the soil chamber pressure acquisition point; the propulsion speed is read from the propulsion control record; and the excavation section parameters are read from the construction parameter record. A baseline time axis is established with a uniform sampling period, and data from different sources are mapped to the baseline time axis according to their temporal proximity. Invalid flags are written for missing data segments and the data is filled in with valid sampled values ​​before and after. Sliding noise reduction is performed on torque, current, soil discharge rate and soil chamber pressure. The data is written into a data table according to the method of corresponding a set of spiral conveying associated operation data and excavation section parameters at the same sampling time, forming spiral conveying time sequence data corresponding to the spiral conveying process of complex loose materials.

3. The soil condition analysis method based on complex loose mass tunneling as described in claim 2, characterized in that: The write interval markers include, Set a sliding time window in the screw conveyor timing data; Within each sliding time window, the soil dumping measurement data is retrieved, and the fluctuation state of the soil dumping rate of complex loose bodies is determined based on the average level and dispersion of the soil dumping rate. When the dispersion of the soil dumping rate fluctuates less than the average level, the corresponding sliding time window is written into the candidate stability flag. After writing the candidate stability marker, the monotonicity of the cumulative soil discharge is checked based on the order of change of the cumulative soil discharge with the sampling time; when the cumulative soil discharge at the later sampling time is less than the cumulative soil discharge at the previous sampling time, the corresponding sampling interval is marked as a measurement anomaly interval. When the cumulative soil discharge meets the monotonically non-decreasing condition, determine whether the increase in soil discharge between adjacent sampling times does not exceed the preset jump limit to determine the continuity of the cumulative soil discharge of complex loose bodies. When the candidate stability marker, the non-decreasing condition of cumulative soil discharge, and the continuous condition of cumulative soil discharge are all met simultaneously, the corresponding sliding time window is identified as the apparent stable soil discharge interval of the complex loose body. When the time interval between two adjacent apparent stable dumping intervals is less than the preset merging interval, and the difference in the average dumping rate between the two intervals is less than the preset mean difference threshold, the two intervals are merged and written into the same apparent stable dumping interval marker.

4. The soil condition analysis method based on complex loose mass tunneling as described in claim 3, characterized in that: The deviation of the calculation of soil dumping load and silo pressure attenuation include, Within each apparent stable soil discharge interval, torque, current, and soil discharge rate are used. The value of the soil discharge rate after zero correction is used as the normalization benchmark. The torque load and current load corresponding to the unit volume of complex loose soil discharged are calculated respectively. The changing trends of torque load and current load corresponding to the unit volume of complex loose soil discharged are determined according to the sampling time, and the changing patterns of continuous rise, staged sudden change and intermittent peak are recorded. Within the same apparent stable soil dumping range, the silo pressure response data is retrieved, and the complex loose body silo pressure decay trend is formed according to the rate of decrease in silo pressure per unit time. Compare the silo pressure decay trend with the silo pressure decay reference benchmark under the preset rate class corresponding to the average level of the corresponding soil discharge rate. The reference benchmark for silo pressure attenuation is the reference change relationship formed by the change of silo pressure attenuation over time in a time segment when the average level of soil discharge rate is in the same preset rate range and is not recorded in the silo pressure attenuation deviation record. When the silo pressure decay trend meets the preset silo pressure decay deviation judgment condition, record the silo pressure decay deviation type and deviation duration, and write the corresponding results into the unit soil discharge load and silo pressure decay deviation. When the pressure decay trend does not meet the preset pressure decay deviation judgment condition, the corresponding time segment is marked as the pressure decay non-deviation segment, and the pressure decay deviation type is recorded as non-deviation.

5. The soil condition analysis method based on complex loose mass tunneling as described in claim 4, characterized in that: The formation process of the latent deviation feature includes, Within the apparent stable spoil disposal zone, the advance speed and excavation section parameters are used. The theoretical excavation volume of the complex loose body is obtained by multiplying the advance distance formed by the advance speed over time with the excavation section parameters. The cumulative excavation volume at the start and end times of the same apparent stable excavation interval is used to obtain the actual excavation volume of the complex loose body. The difference between the theoretical excavation volume and the actual excavation volume is compared to form the volume balance deviation. The torque and current are processed by moving average to obtain the low-frequency load change trend. The corresponding low-frequency load change trend is then subtracted from the original torque and original current to obtain the torque micro-fluctuation and current micro-fluctuation. Based on the degree of deviation of the micro-fluctuation amplitude from the corresponding window's normal amplitude, record the micro-fluctuation intensity, the number of peak occurrences, and the peak time distribution; The volume balance deviation, the load change trend corresponding to the unit volume of soil discharged, the deviation type of silo pressure attenuation, the torque micro-fluctuation characteristics, and the current micro-fluctuation characteristics are normalized and encoded in a preset order to form implicit deviation characteristics that characterize the internal changes of complex loose bodies. The torque micro-fluctuation characteristics and current micro-fluctuation characteristics respectively include the intensity of the corresponding micro-fluctuation, the number of times the peak occurs, and the peak time distribution; The normalized coding uses the characteristic benchmark range corresponding to the historical normal spoil disposal interval or the current stable segment of the tunneling section as the comparison object; features above the upper limit of the characteristic benchmark range are coded as positive deviations; features below the lower limit of the characteristic benchmark range are coded as negative deviations; features within the characteristic benchmark range are coded as no deviations, and the deviation level is determined according to the ratio of the deviation amount to the width of the characteristic benchmark range.

6. The soil condition analysis method based on complex loose mass tunneling as described in claim 5, characterized in that: The method for identifying abnormal soil conditions within complex loose bodies includes... Based on the hidden deviation features of complex loose bodies extracted from historical normal spoilage intervals, on-site verification abnormal intervals, and test calibration intervals, a set of state triggering rules and confidence generation rules are established. The set of state triggering rules includes soil compression triggering rules, water precipitation triggering rules, pore collapse triggering rules, and particle rearrangement triggering rules; When either the torque load or the current load corresponding to the unit volume of complex loose soil discharged increases continuously, the silo pressure decay trend is lower than the lower limit of the silo pressure decay reference benchmark and continues to exceed the preset time, and the theoretical excavation volume deviates continuously in a positive direction from the actual soil discharge volume, the soil compression triggering rule is matched to generate a soil compression candidate state. When one of the torque load and current load corresponding to the unit volume of complex loose soil discharged undergoes a sudden change, the intensity of the current micro-fluctuation increases, and the silo pressure decay deviates from the direction of change between adjacent windows, the water precipitation triggering rule is matched to generate a water precipitation candidate state. When the soil pressure in the soil chamber suddenly drops within a short time window, and the torque or current fluctuations spike in the same time neighborhood, and the soil discharge rate does not show a corresponding continuous increase, the pore collapse triggering rule is matched to generate a pore collapse candidate state. When torque micro-fluctuations and current micro-fluctuations exhibit a peak sequence with adjacent peak time intervals within a preset interval range within multiple consecutive preset time windows, and the pressure decay deviation shows intermittent changes, the particle rearrangement trigger rule is matched to generate a particle rearrangement candidate state. For each candidate state, a comprehensive score is calculated based on the number of triggering features, deviation level, duration, and time consistency. The comprehensive score is then converted into a corresponding confidence level according to the confidence level generation rules. When the corresponding confidence level reaches the preset state threshold, the corresponding candidate state is confirmed as an abnormal state of soil inside a complex loose body.

7. The soil condition analysis method based on complex loose mass tunneling as described in claim 6, characterized in that: The generated corrected soil condition analysis results include: After each apparent stable soil dumping interval is completed, the corresponding discrimination states and confidence levels of complex loose soil compression, water precipitation, pore collapse and particle rearrangement are read; When the confidence level of all the discrimination states is lower than the preset state threshold, retain the soil dumping stability mark of the corresponding apparent stable soil dumping interval and write the internal soil quality non-abnormal mark. When the confidence level of any discrimination state reaches the preset state threshold, the corresponding apparent stable dumping interval is modified from a single dumping stability marker to an apparent dumping stability accompanied by anomaly markers inside complex loose bodies. When the confidence of two or more discrimination states reaches the preset state threshold, the main state and the accompanying state are determined according to the confidence level, and a composite state label is written when the confidence difference is lower than the preset discrimination threshold. When the same primary state occurs repeatedly in multiple consecutive apparent stable dumping intervals, a persistent state flag is written; when the same primary state occurs only in a single apparent stable dumping interval, a transient state flag is written. The output includes the start and end times of the interval, soil stabilization markers, corrected soil condition of complex loose soil, main state, associated state, and confidence level.

8. A soil condition analysis system based on complex loose body tunneling, employing the soil condition analysis method based on complex loose body tunneling as described in any one of claims 1 to 7, characterized in that: It includes a data acquisition and calibration module, an interval filtering module, a load attenuation module, a deviation construction module, a state discrimination module, and a result correction module; The acquisition and calibration module is used to acquire the spiral conveyor-related operation data during the tunneling process of the shield machine according to a unified time reference, and read the excavation section parameters to generate spiral conveyor time sequence data. The interval screening module is used to screen the apparent stable soil discharge intervals from the screw conveyor time series data based on the soil discharge metering stability, and write the interval markers. The load attenuation module is used to call the load response data, soil discharge metering data and silo pressure response data within the apparent stable soil discharge range, and calculate the deviation of soil discharge load and silo pressure attenuation. The deviation construction module is used to combine propulsion forming data, soil discharge measurement data, and load disturbance data to form implicit deviation features; The state discrimination module is used to determine the abnormal state of soil inside complex loose bodies based on the matching relationship between latent deviation features and the set of state triggering rules. The result correction module is used to write back the apparent stable soil discharge interval based on the discrimination state and confidence level, and generate the corrected soil condition analysis results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the soil condition analysis method based on complex loose body tunneling as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the soil condition analysis method based on complex loose body tunneling as described in any one of claims 1 to 7.

Images