CN-122022484-A - Tunnel engineering monitoring method and system based on multiple risk judgment

CN122022484ACN 122022484 ACN122022484 ACN 122022484ACN-122022484-A

Abstract

The invention discloses a tunnel engineering monitoring method and system based on multiple risk judgment, and relates to the technical field of safety monitoring, wherein the method comprises the steps of binding multisource monitoring data with working condition information, preprocessing, and predicting the future state of each monitoring index based on the preprocessed monitoring data to obtain abnormal risk level and uncertainty metric value; the method comprises the steps of carrying out extraction and weighted combination on a plurality of evolution features of each monitoring index based on preprocessed monitoring data to obtain evolution risk levels, carrying out logic arbitration through an engineering state machine based on the preprocessed monitoring data and corresponding working condition labels to obtain engineering risk levels, carrying out fusion decision on abnormal risk levels, evolution risk levels and engineering risk levels based on uncertainty metric values to generate final risk levels and interpretable cause chain information, and avoiding missed judgment or misjudgment caused by lack of working condition information or single view angle.

Inventors

LIAO KAI
DU YUNCHAO
ZHOU SHUAI
HE RONGGUO
FENG JIANJUN
ZHANG YUTAO

Assignees

中铁西南科学研究院有限公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1. The tunnel engineering monitoring method based on multiple risk judgment is characterized by comprising the following steps of: Acquiring multi-source monitoring data of tunnel engineering, binding each monitoring data with corresponding working condition information, and acquiring monitoring data with working condition labels; Preprocessing the monitoring data with the working condition label to obtain preprocessed monitoring data; predicting the future state of each monitoring index based on the preprocessed monitoring data to obtain an abnormal risk level and an uncertainty measurement value of each monitoring index; extracting and weighting and combining a plurality of evolution features of each monitoring index based on the preprocessed monitoring data to obtain an evolution risk level of each monitoring index; based on the preprocessed monitoring data and the corresponding working condition labels, carrying out logic arbitration through an engineering state machine to obtain the engineering risk level of each monitoring index; based on the uncertainty metric value of each monitoring index, fusion decision is carried out on the abnormal risk level, the evolution risk level and the engineering risk level of each monitoring index, and final risk level and interpretable reason chain information are generated.
2. The tunnel engineering monitoring method based on multiple risk judgment according to claim 1, wherein the preprocessing the monitoring data with the working condition label to obtain preprocessed monitoring data comprises: Performing time alignment and missing completion processing on the monitoring data with the working condition label to obtain a multi-dimensional monitoring data sequence with uniform and continuous time; judging whether the multidimensional monitoring data sequence meets the preset physical feasible region constraint or not; And carrying out projection correction on the abnormal data points which do not meet the physical feasible region constraint to obtain preprocessed monitoring data.
3. The tunnel engineering monitoring method based on multiple risk judgment according to claim 2, wherein the physical feasible region constraint comprises a value range constraint, a change rate constraint and a consistency constraint, the value range constraint is used for limiting a monitored value to be in a preset numerical range, the change rate constraint is used for limiting the change amplitude of the monitored value between adjacent time steps not to exceed a preset speed threshold, and the consistency constraint is used for limiting the monitored value difference of adjacent measuring points or a plurality of measuring points in the same section to meet a stress consistency condition of a preset geometric consistency condition.
4. The multiple risk judgment based tunnel engineering monitoring method according to claim 2, wherein the performing projection correction on the abnormal data that does not satisfy the physical feasible region constraint includes: determining outlier data points that do not satisfy the physical feasible region constraint; And replacing the original monitoring value of the abnormal data point with a value on the boundary of the corresponding physical feasible domain, or replacing the original monitoring value of the abnormal data with a feasible value which meets the constraint of the physical feasible domain and has the minimum correction quantity with the original monitoring value.
5. The tunnel engineering monitoring method based on multiple risk judgment according to claim 1, wherein predicting the future state of each monitoring index based on the preprocessed monitoring data to obtain the abnormal risk level and uncertainty metric value of each monitoring index comprises: based on the preprocessed monitoring data, predicting the future state of each monitoring index through a time sequence prediction model to obtain the prediction probability distribution of each monitoring index; calculating the overrun probability of each monitoring index exceeding the preset engineering threshold in the future based on the prediction probability distribution and the preset engineering threshold, wherein the preset engineering threshold is determined according to the working condition label carried by each monitoring index; Calculating the uncertainty measure based on the predictive probability distribution; And fusing the overrun probability and the uncertainty measurement value to obtain the abnormal risk level.
6. The tunnel engineering monitoring method based on multiple risk judgment according to claim 1, wherein the multiple evolution features include trend features, drift features and acceleration features, the extracting and weighting combination are performed on the multiple evolution features of each monitoring index based on the preprocessed monitoring data to obtain the evolution risk level of each monitoring index, and the method comprises the following steps: Carrying out robust scale normalization processing on the preprocessed monitoring data to obtain normalized monitoring data and robust scale factors; based on the normalized monitoring data, extracting the long-term change slope of each monitoring index in a preset long-time window, and carrying out normalization processing based on the scale factors to obtain trend characteristics; Calculating the offset of the current monitoring value relative to the historical baseline value in a preset drift analysis window based on the normalized monitoring data, and carrying out normalization processing based on the scale factors to obtain drift characteristics; Based on the normalized monitoring data, extracting a short-term change slope of each monitoring index in a preset short-time window, calculating a difference value between the short-term change slope and the long-term change slope, and carrying out normalization processing on the difference value based on the scale factors to obtain an acceleration characteristic; And carrying out weighted combination on the trend characteristics, the drift characteristics and the acceleration characteristics of each monitoring index to obtain the evolution risk level of each monitoring index.
7. The tunnel engineering monitoring method based on multiple risk judgment according to claim 1, wherein the logic arbitration is performed by an engineering state machine based on the preprocessed monitoring data and the corresponding working condition label to obtain the engineering risk level of each monitoring index, comprising: configuring independent engineering state machine examples for each monitoring index, wherein each engineering state machine example comprises three states of an idle state, an active state and an upgrade state; Judging whether hit conditions are met or not based on the comparison result of the preprocessed monitoring data and a preset engineering threshold value in each monitoring period, if yes, judging that hit is performed once, and triggering hit counting to increase, if not, triggering hit counting to decrease, wherein the preset engineering threshold value is determined according to working condition labels carried by each monitoring index; When the duration from the first hit reaches a preset duration threshold and the hit count reaches a preset minimum number of continuous hits, switching the engineering state machine instance from an idle state to an active state, and determining the engineering risk level according to the comparison result; when the duration of the engineering state machine instance in the activated state reaches a preset upgrading time threshold, switching the engineering state machine instance from the activated state to the upgrading state, and improving the engineering risk level; and when the monitoring period is finished, outputting the engineering risk level by the engineering state machine instance.
8. The multiple risk-discrimination based tunnel engineering monitoring method of claim 7, further comprising, after switching the engineering state machine instance from an idle state to an active state or switching the engineering state machine instance from an active state to an upgraded state: When the engineering state machine instance is in an activated state, the engineering risk level is lower than a preset level, and the duration of continuous miss reaches a preset hysteresis time threshold, switching the engineering state machine instance from the activated state to an idle state; and when the engineering state machine instance is in an upgrading state and the duration time of the continuous miss reaches a preset hysteresis time threshold, switching the engineering state machine instance from the upgrading state to an idle state.
9. The tunnel engineering monitoring method based on multiple risk judgment according to claim 1, wherein the performing a fusion decision on the abnormal risk level, the evolution risk level and the engineering risk level of each monitoring index based on the uncertainty metric value of each monitoring index to generate a final risk level and interpretable cause chain information comprises: Judging whether the engineering risk level of each monitoring index reaches a preset risk level; if the engineering risk level of any monitoring index reaches a preset risk level, taking the engineering risk level of the monitoring index as the final risk level; If the engineering risk level of the monitoring index does not reach the preset risk level, judging whether the uncertainty measurement value of the monitoring index is smaller than or equal to a preset uncertainty threshold or not and whether the evolution risk level of the monitoring index reaches the preset risk level or not; if the uncertainty metric value of the monitoring index is smaller than or equal to a preset gating threshold value and the evolution risk level of the monitoring index reaches a preset risk level, taking the evolution risk level of the monitoring index as the final risk level; If the uncertainty metric value of the monitoring index is larger than a preset gating threshold value or the evolution risk level of the monitoring index does not reach a preset risk level, taking the maximum value of the abnormal risk level and the engineering risk level of the monitoring index as the final risk level; While generating the final risk level, the interpretable cause chain information is synchronously recorded.
10. The utility model provides a tunnel engineering monitoring system based on multiple risk judgement, its characterized in that includes: the data acquisition module is used for acquiring multi-source monitoring data of tunnel engineering, binding each piece of monitoring data with corresponding working condition information, and acquiring monitoring data with a working condition label; The preprocessing module is used for preprocessing the monitoring data with the working condition label to obtain preprocessed monitoring data; The abnormal risk calculation module is used for predicting the future state of each monitoring index based on the preprocessed monitoring data to obtain the abnormal risk level and uncertainty metric value of each monitoring index; The evolution risk calculation module is used for extracting and weighting and combining a plurality of evolution features of each monitoring index based on the preprocessed monitoring data to obtain an evolution risk level of each monitoring index; The engineering state machine module is used for carrying out logic arbitration through the engineering state machine based on the preprocessed monitoring data and the corresponding working condition labels to obtain the engineering risk level of each monitoring index; And the fusion decision module is used for carrying out fusion decision on the abnormal risk level, the evolution risk level and the engineering risk level of each monitoring index based on the uncertainty metric value of each monitoring index, and generating final risk level and interpretable reason chain information.

Description

Tunnel engineering monitoring method and system based on multiple risk judgment Technical Field The invention relates to the technical field of safety monitoring, in particular to a tunnel engineering monitoring method and system based on multiple risk judgment. Background In the construction stage and the operation stage of tunnel engineering, the problems of complex surrounding rock conditions, strong construction disturbance, uncertain stress state of a supporting structure and the like are generally faced. In order to ensure engineering safety, sensing systems such as displacement, settlement, convergence, support internal force, vibration, environmental parameters and the like are generally arranged in engineering practice, and the state of a tunnel structure is continuously monitored. The existing tunnel engineering monitoring and early warning method mostly adopts a fixed threshold judgment mechanism, namely, when a certain monitored quantity exceeds a preset control value, an alarm is triggered. However, the fixed threshold is difficult to adapt to the changes of different construction stages, different surrounding rock conditions and working conditions, false alarm is easy to generate, and the binary judgment of yes or no is performed only based on the instantaneous state of the current monitoring value, so that the risk cognition visual angle is single. This results in a system that is not able to predict the probability of future overruns, nor is it difficult to identify adverse evolution trends (e.g., sustained acceleration, trend drift) contained in the data. Therefore, for the potential risks of slow accumulation or accelerated development, the system usually sends out an alarm when the risks actually burst (the numerical value exceeds the limit), valuable early warning windows are missed, and serious missed report risks exist. In summary, the prior art has the problems of high false alarm rate and high missing report risk due to lack of working condition information and single risk cognition visual angle. Disclosure of Invention The invention aims to solve the technical problems of high false alarm rate and high missing report risk of the existing fixed threshold judgment method, and provides a tunnel engineering monitoring method and system based on multiple risk judgment. The invention is realized by the following technical scheme: In a first aspect, the present invention provides a tunnel engineering monitoring method based on multiple risk judgment, including: Acquiring multi-source monitoring data of tunnel engineering, binding each monitoring data with corresponding working condition information, and acquiring monitoring data with working condition labels; Preprocessing the monitoring data with the working condition label to obtain preprocessed monitoring data; predicting the future state of each monitoring index based on the preprocessed monitoring data to obtain an abnormal risk level and an uncertainty measurement value of each monitoring index; extracting and weighting and combining a plurality of evolution features of each monitoring index based on the preprocessed monitoring data to obtain an evolution risk level of each monitoring index; based on the preprocessed monitoring data and the corresponding working condition labels, carrying out logic arbitration through an engineering state machine to obtain the engineering risk level of each monitoring index; based on the uncertainty metric value of each monitoring index, fusion decision is carried out on the abnormal risk level, the evolution risk level and the engineering risk level of each monitoring index, and final risk level and interpretable reason chain information are generated. Optionally, the preprocessing the monitoring data with the working condition label to obtain preprocessed monitoring data includes: Performing time alignment and missing completion processing on the monitoring data with the working condition label to obtain a multi-dimensional monitoring data sequence with uniform and continuous time; judging whether the multidimensional monitoring data sequence meets the preset physical feasible region constraint or not; And carrying out projection correction on the abnormal data points which do not meet the physical feasible region constraint to obtain preprocessed monitoring data. Optionally, the physical feasible region constraint comprises a value range constraint, a change rate constraint and a consistency constraint, wherein the value range constraint is used for limiting the monitoring value to be in a preset numerical range, the change rate constraint is used for limiting the change amplitude of the monitoring value between adjacent time steps not to exceed a preset speed threshold, and the consistency constraint is used for limiting the monitoring value difference of the adjacent measuring points or a plurality of measuring points in the same section to meet a stress consistency condition of a preset geometr