Safety engineering integrated management system

Abstract

The invention relates to the technical field of risk analysis, in particular to a safety engineering comprehensive management system which comprises a geological parameter dynamic monitoring module, a risk threshold early warning module, a stability weight evaluation module and a risk comprehensive management and control module. According to the method, lithology data are extracted through a drilling profile to calculate the layer thickness sudden change rate and the strength attenuation rate, a lithology abnormal area is positioned through a variable coefficient method, lithology sudden change marks are analyzed through spatial clustering, a dynamic trend is captured through sliding window mean value comparison, and gradient amplitude and slope direction dispersion are calculated through a digital elevation model. Fusing the abnormal point density to generate a terrain instability index; establishing a multi-dimensional quantitative evaluation framework; carrying out fuzzy comprehensive evaluation and cross validation on the instability index and an engineering parameter trend; generating a dynamic priority sequence, carrying out stake number interval matching to construct a risk measure table, carrying out analytic hierarchy process dynamic resource allocation support drainage scheme, and carrying out real-time monitoring and prediction closed-loop iteration. And adaptively adjusting the control strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of risk analysis, and particularly to a comprehensive safety engineering management system. Background Art

[0002] The technical field of risk analysis involves the identification, assessment, and management of potential risk factors to reduce the probability and impact of accidents. The core content of this technical field includes risk identification, risk assessment, risk control, and monitoring, which systematically constructs a whole-process management system with the goal of safety assurance. Risk analysis covers multiple stages from preliminary risk screening to quantitative and qualitative assessment, relying on multidisciplinary theoretical bases such as statistical analysis, systems engineering, and human factors engineering methods, and is widely applied in industries such as engineering construction, manufacturing, transportation, and energy development. Its overall technical system is supported by data collection and processing, combined with refined means such as fault tree analysis, event tree analysis, and probabilistic risk assessment, to achieve comprehensive management of various systematic risks and accidental risks.

[0003] Among them, the comprehensive safety engineering management system refers to meeting the safety risk management requirements during the construction and operation of engineering projects, and using means such as risk identification, hierarchical classification management, and dynamic assessment to complete the full-cycle safety management process. The technical matters covered by this patent theme include formulating risk identification criteria based on the life cycle nodes of engineering projects, dividing risk levels according to different processes and environmental conditions, conducting safety assurance management by formulating risk response plans and resource allocation plans, continuously tracking and adjusting risk changes by adopting regular risk reviews and dynamic monitoring mechanisms, establishing a risk event database using historical data archiving and big data analysis technologies to support future risk prediction and decision-making.

[0004] The prior art relies on fixed-cycle screening and historical data attribution, making it difficult to capture the gradual anomalies of lithologic layer thickness and strength in real time, resulting in a lag in the identification of mutation regions. For example, when the layer thickness mutation rate exceeds the threshold, an alarm cannot be triggered in time. In the risk assessment process, a spatial correlation model between the layer thickness mutation rate and the strength attenuation rate is not established, and the anomalies of a single parameter are analyzed in isolation, making it impossible to quantify the chain effect of local anomalies on the stability of adjacent regions. The terrain stability assessment lacks a fusion calculation framework for slope change amplitude, aspect dispersion, and elevation anomaly density, and only uses a single terrain parameter to divide risk levels, easily ignoring the instability factors of complex terrain. When formulating risk control strategies, the geological feature differences in pile number intervals are not combined, and a general measure standard is adopted, resulting in insufficient matching between the support plan and the local terrain instability index. The dynamic monitoring mechanism lacks the function of predicting the change trend of parameters and relies on a fixed threshold to trigger a response, making it difficult to generate predictive measures in a timely manner when encountering slow-changing risks. The historical data and real-time monitoring data do not form a closed-loop verification, restricting the iterative optimization efficiency of the risk control plan during the construction process. Summary of the Invention

[0005] The object of the present invention is to solve the drawbacks existing in the prior art, and a comprehensive management system for safety engineering is proposed.

[0006] To achieve the above object, the present invention adopts the following technical solution: A comprehensive management system for safety engineering includes:

[0007] A risk threshold warning module, which is used to receive lithology mutation marks and perform spatial clustering analysis, compare the average value of the sliding window length dynamically adjusted based on the borehole spacing and the geological structure activity cycle for the layer thickness mutation rate and the strength attenuation rate, generate a segmented engineering warning signal when the threshold is exceeded, determine the window weight coefficient by using the moving average model regression analysis, and generate and transmit the warning signal distribution map and the parameter prediction trend to the risk comprehensive control module;

[0008] A stability weight evaluation module, which is used to calculate the grid slope amplitude change and the slope direction dispersion degree through a digital elevation model, generate a terrain instability index in combination with the elevation anomaly point density, normalize the slope amplitude change value by using the maximum and minimum method, output the prevention and control suggestion priority level and the instability index distribution, and transmit it to the risk comprehensive control module;

[0009] A risk comprehensive control module, which is used to receive the warning signal distribution map and match the stake number interval, call the parameter prediction trend to judge the risk evolution direction, generate a stake number - measure mapping relationship table in combination with the prevention and control suggestion priority level and the instability index distribution, construct the criterion layer weights of the support, drainage, and monitoring measures by using the analytic hierarchy process, and output the segmented engineering control plan bound to the stake number.

[0010] As a further scheme of the present invention, the segmented engineering warning signal includes a warning level, a warning section range, and a warning trigger type, the warning signal distribution map includes a warning stake number distribution, a spatial risk concentration area, and a warning signal density, the parameter prediction trend includes a layer thickness change trend, a strength evolution trend, and a comprehensive risk change trend, the terrain instability index includes a slope change index, a slope direction dispersion index, and an elevation anomaly index, the prevention and control suggestion priority level specifically is a support priority level, a drainage priority level, and a monitoring priority level, the instability index distribution includes a high - risk instability area, a medium - risk instability area, and a low - risk instability area, the stake number - measure mapping relationship table includes the support measures corresponding to the stake number, the drainage measures corresponding to the stake number, and the monitoring measures corresponding to the stake number, and the segmented engineering control plan includes a segmented support plan, a segmented drainage plan, and a segmented monitoring plan.

[0011] As a further scheme of the present invention, the dynamic adjustment rule of the sliding window length is that when the borehole spacing is less than 10 meters and the geological structure activity cycle is less than 1 year, the window length is set to 5 layers.

[0012] The normalization formula of the maximum-minimum value method is slope change amplitude = (original value - minimum value) / (maximum value - minimum value);

[0013] The weight of the criterion layer includes that the support measure weight is dynamically allocated according to the product result of the instability index distribution and the parameter prediction trend.

[0014] As a further solution of the present invention, the risk threshold warning module includes:

[0015] The spatial clustering sub-module calls the lithology mutation mark, extracts the coordinate data of the marked points, sets the clustering radius threshold based on the inverse proportional relationship between the layer thickness mutation rate and the marked point density, calculates the Euclidean distance between the marked points, merges the marked points with a distance less than the threshold into the same group, and generates lithology clustering partitions according to the spatial distribution of the groups;

[0016] The inverse proportional relationship sets the clustering radius threshold as clustering radius = 10 m / (layer thickness mutation rate × 0.1 + 1);

[0017] The mean comparison sub-module calls the lithology clustering partition, the layer thickness mutation rate, and the strength attenuation rate, intercepts continuous data segments within the partition according to the preset window length, calculates the arithmetic mean of the mutation rate and the attenuation rate within the window, and compares the mean with the engineering risk threshold set based on the regional geological stability classification to generate a super-threshold interval;

[0018] The setting rule of the regional geological stability classification is that the threshold of the first-level stable area = 0.3, the threshold of the second-level medium area = 0.5, and the threshold of the third-level high-risk area = 0.7;

[0019] The warning generation sub-module calls the super-threshold interval, performs a moving average calculation on the mutation rate and the attenuation rate within the interval according to the time series, sets the window weight coefficient to be 0.7 times that of the previous window determined by historical data regression analysis, and generates a warning signal distribution map and a parameter prediction trend.

[0020] As a further solution of the present invention, the stability weight evaluation module includes:

[0021] The slope feature quantification sub-module obtains digital elevation model data, calculates the slope angle difference between adjacent elevation points, extracts the maximum and minimum slope angle differences within the grid cell, generates the slope change amplitude, calculates the standard deviation after converting the slope direction angle to polar coordinates to eliminate the periodic influence, and outputs the slope direction dispersion;

[0022] The instability index calculation sub-module calls the slope change amplitude for normalization processing, converts the slope direction dispersion into a positive correlation coefficient through the inverse proportional function f(x) = 1 / (x + 1), superimposes the elevation anomaly point density data, and performs a weighted product operation according to the 4:3:3 weight based on the geological parameter sensitivity analysis to generate the terrain instability index of the center point of the grid cell;

[0023] The sensitivity analysis of the geological parameters determines the contribution rate weights of the slope amplitude variation, the aspect dispersion, and the density of elevation anomaly points through the principal component analysis method.

[0024] The prevention and control measure classification sub-module divides the terrain instability index into four intervals based on the fuzzy comprehensive evaluation method, defines the vertex parameters of the triangular membership function as the instability indices 0.2, 0.5, and 0.8, generates the prevention and control priority coefficients, and outputs the distribution of the global instability index.

[0025] As a further solution of the present invention, the risk comprehensive control module includes:

[0026] The stake number interval analysis sub-module obtains the pixel coordinates and the warning intensity values of the warning signal distribution map, converts the horizontal and vertical coordinates into stake number segment identifiers based on the stake number coding rule, calculates the absolute value of the warning intensity difference between adjacent stake numbers, compares with the warning fluctuation threshold set according to the regional historical disaster frequency, and generates a stake number interval table.

[0027] The rule for setting the regional historical disaster frequency is: when the disaster frequency ≥ 5 times / year, the fluctuation threshold = 0.5;

[0028] The measure mapping generation sub-module, based on the stake number interval table, calls the parameter prediction trend to judge the risk evolution direction, extracts the grade values of the prevention and control recommendation priorities and the grid cell values of the instability index distribution, divides the support measures into bolt support, pile foundation support, and retaining wall support types according to the priority level, and establishes a one-to-many mapping relationship between the stake number and the measure type and intensity, and generates a stake number-measure mapping table.

[0029] The measure priority ranking sub-module constructs the criterion layer and the scheme layer structure by using the analytic hierarchy process, sets the ratio of the support measure weight to the drainage measure weight as 3:2, calculates the weight value of each measure for the stake number risk control, generates the measure priority ranking in descending order of the weight value, and binds the stake number interval to output the segmented project control scheme.

[0030] As a further solution of the present invention, the system further includes:

[0031] The geological parameter dynamic monitoring module is used to extract the lithology distribution, the layer thickness sequence, and the shear strength sequence through the drilling profile, calculate the layer thickness mutation rate and the strength attenuation rate between adjacent segments, optimize the threshold generation logic by using the coefficient of variation method in combination with the dynamic coupling relationship between the layer thickness mutation rate and the strength attenuation rate, identify the lithology mutation abnormal area, generate a lithology mutation mark, and transmit it to the risk threshold warning module.

[0032] As a further solution of the present invention, the lithology mutation marker specifically includes mutation position identification, mutation region boundary, and mutation type classification. The layer thickness mutation rate includes layer thickness amplitude change, layer thickness variation degree, and local layer thickness difference. The strength attenuation rate includes shear strength amplitude change, strength decline rate, and abnormal strength gradient.

[0033] As a further solution of the present invention, the dynamic coupling relationship optimization threshold generation logic is to adjust the coefficient of variation threshold according to the correlation coefficient between the layer thickness mutation rate and the strength attenuation rate. When the correlation coefficient is greater than 0.5, the threshold is increased to 1.8 times the mean value of the coefficient of variation in the current dataset.

[0034] As a further solution of the present invention, the geological parameter dynamic monitoring module includes:

[0035] The lithology sequence extraction sub-module obtains the drilling profile data. Based on the mineral composition and particle size parameters in the lithology classification standard, the lithology horizons are divided by the interval segmentation method, the lithology codes of multiple horizons are extracted, and the layer thickness and shear strength data are statistically analyzed to generate a layer thickness sequence and a shear strength sequence.

[0036] The mutation rate calculation sub-module calls the layer thickness sequence, calculates the absolute value of the difference between the current layer thickness and the previous layer thickness in the order of adjacent horizons. After unifying the layer thickness unit to meters, the difference calculation is performed, and the difference is divided by the previous layer thickness to obtain the layer thickness mutation rate of multiple horizons. It calls the shear strength sequence, calculates the absolute value of the difference between the current strength and the previous strength in the order of adjacent horizons. After unifying the shear strength unit to megapascals, the difference calculation is performed, and the difference is divided by the previous strength to obtain the strength attenuation rate of multiple horizons.

[0037] The anomaly identification sub-module calls the layer thickness mutation rate and the strength attenuation rate, calculates the mean and standard deviation of the two sequences, divides the standard deviation by the mean to obtain the coefficient of variation, and based on the lithology mutation threshold generation rule verified by the historical geological disaster case library, screens the horizon numbers with the coefficient of variation exceeding the threshold to generate lithology mutation markers.

[0038] The verification rule of the historical geological disaster case library is that when both the layer thickness mutation rate and the strength attenuation rate exceed the threshold, the coefficient of variation threshold is reduced to 1.2 times the mean value.

[0039] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0040] In the present invention, lithology distribution data is extracted from the drilling profile, and the layer thickness mutation rate and intensity attenuation rate are calculated. Combining with the coefficient of variation method, the lithology anomaly area is accurately located to reduce the identification deviation of a single static index. Spatial clustering analysis is used to aggregate the lithology mutation markers in the region, and the dynamic change trend of parameters is captured by comparing the moving window mean, so as to improve the sensitivity of the gradual risk warning. The digital elevation model is introduced to calculate the slope amplitude change and aspect dispersion degree, and the terrain instability index is generated by fusing the elevation anomaly point density, and a multi-dimensional quantitative evaluation framework for terrain stability is established. Based on the fuzzy comprehensive evaluation method, the terrain instability index and the predicted trend of engineering parameters are cross-validated to generate a dynamic priority ranking, so as to enhance the matching accuracy between the prevention and control measures and the risk evolution direction. The risk measure mapping relationship table is constructed by matching the pile number interval, and the dynamic resource allocation of the support and drainage schemes is carried out by combining the analytic hierarchy process, so as to realize the collaborative optimization of the measure combination, geological characteristics and construction stage. The real-time monitoring data and the prediction model form a closed-loop iteration, which promotes the adaptive adjustment of the control strategy with the progress of the project and reduces the response delay caused by the sudden change of environmental parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is the system flow chart of the present invention;

[0042] Figure 2 is the flow chart of the dynamic monitoring module of geological parameters of the present invention;

[0043] Figure 3 is the flow chart of the risk threshold warning module of the present invention;

[0044] Figure 4 is the flow chart of the stability weight evaluation module of the present invention;

[0045] Figure 5 is the flow chart of the risk comprehensive control module of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0046] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0048] Example 1, please refer to Figure 1 , a comprehensive safety engineering management system includes:

[0049] A geological parameter dynamic monitoring module, which is used to extract lithology distribution, layer thickness sequence, and shear strength sequence through drilling profiles, calculate the layer thickness mutation rate and strength attenuation rate of adjacent segments, optimize the threshold generation logic by combining the dynamic coupling relationship between the layer thickness mutation rate and strength attenuation rate using the coefficient of variation method, identify abnormal lithology mutation areas, generate lithology mutation marks, and transmit them to the risk threshold warning module;

[0050] A risk threshold warning module, which is used to receive lithology mutation marks and perform spatial clustering analysis, compare the mean values of the layer thickness mutation rate and strength attenuation rate with a sliding window length dynamically adjusted based on the borehole spacing and geological structure activity cycle, generate a segmented engineering warning signal when the threshold is exceeded, determine the window weight coefficient using a moving average model regression analysis, and generate and transmit the warning signal distribution map and parameter prediction trend to the risk comprehensive control module;

[0051] A stability weight evaluation module, which is used to calculate the grid slope amplitude change and slope direction dispersion through a digital elevation model, generate a terrain instability index by combining the elevation anomaly point density, normalize the slope amplitude change value using the maximum-minimum method, output the prevention and control recommendation priority and instability index distribution, and transmit them to the risk comprehensive control module;

[0052] A risk comprehensive control module, which is used to receive the warning signal distribution map and match the stake number interval, call the parameter prediction trend to judge the risk evolution direction, generate a stake number - measure mapping relationship table by combining the prevention and control recommendation priority and instability index distribution, construct the criterion layer weights of support, drainage, and monitoring measures using the analytic hierarchy process, and output a segmented engineering control plan for the bound stake number.

[0053] The lithology mutation markers specifically include mutation position identification, mutation area boundary, and mutation type classification. The layer thickness mutation rate includes the layer thickness variation range, the degree of layer thickness variation, and the local layer thickness difference. The strength attenuation rate includes the shear strength variation range, the strength decline rate, and the abnormal strength gradient. The segmented project warning signal includes the warning level, the warning section range, and the warning trigger type. The warning signal distribution map includes the warning stake number distribution, the spatial risk concentration area, and the warning signal density. The parameter prediction trend includes the layer thickness change trend, the strength evolution trend, and the comprehensive risk change trend. The terrain instability index includes the slope change index, the aspect divergence index, and the elevation anomaly index. The prevention and control recommendation priority specifically includes the support priority, the drainage priority, and the monitoring priority. The instability index distribution includes the high-risk instability area, the medium-risk instability area, and the low-risk instability area. The stake number - measure mapping relation table includes the support measures corresponding to the stake number, the drainage measures corresponding to the stake number, and the monitoring measures corresponding to the stake number. The segmented project control plan includes the segmented support plan, the segmented drainage plan, and the segmented monitoring plan.

[0054] The logic for generating the optimization threshold of the dynamic coupling relationship is to adjust the coefficient of variation threshold according to the correlation coefficient between the layer thickness mutation rate and the strength attenuation rate. When the correlation coefficient is greater than 0.5, the threshold is increased to 1.8 times the mean value of the coefficient of variation in the current dataset.

[0055] The dynamic adjustment rule for the sliding window length is that when the drilling hole spacing is less than 10 meters and the tectonic activity period is less than 1 year, the window length is set to 5 horizons.

[0056] The normalization formula of the maximum - minimum method is: slope variation amplitude = (original value - minimum value) / (maximum value - minimum value).

[0057] The weight of the criterion layer includes that the weight of the support measure is dynamically allocated according to the product result of the instability index distribution and the parameter prediction trend.

[0058] Please refer to Figure 2 , the geological parameter dynamic monitoring module includes:

[0059] The lithology sequence extraction sub - module obtains the drilling profile data, divides the lithology horizons by the interval segmentation method based on the mineral composition and particle size parameters in the lithology classification standard, extracts the lithology codes of multiple horizons, and statistically analyzes the layer thickness and shear strength data to generate the layer thickness sequence and the shear strength sequence.

[0060] First, obtain the drilling profile data of the boreholes (numbered ZK01, located at the stake number K1 + 200) along a specific engineering project (such as section K1 of G Expressway). This is an engineering geological log that details the information of the rock and soil layers exposed from the surface to the predetermined depth. According to the lithology classification standard based on mineral composition, particle size, and structure in the "Code for Geological Exploration and Survey of Building Engineering", the strata exposed by borehole ZK01 are divided.

[0061] In borehole ZK01, from the surface to a depth of 4.5 meters below the surface, the main mineral components of the geotechnical samples were identified as quartz and feldspar, and the content of fine particles with a particle size less than 0.075 mm exceeded 50%. This layer was determined to be silty clay and was assigned a lithology code . The thickness of this layer was directly measured to be 4.5 meters. Immediately following, from a depth of 4.5 meters to 11.2 meters, the main component changed to quartz grains, with the particle size concentrated in the range of 0.25 mm to 0.5 mm, conforming to the characteristics of medium sand, and the lithology code was determined , and its layer thickness was calculated by the depth difference, that is . Further down, from a depth of 11.2 meters to 20.0 meters, the retrieved core showed a granite structure, but the mineral composition had undergone significant alteration, and the joint fissures were very developed. It was comprehensively determined to be strongly weathered granite, with a lithology code , and the layer thickness . Through systematic analysis of the entire borehole profile, a multi-layer lithology code sequence of borehole ZK01 was extracted and denoted as .

[0062] At the same time, for each extracted lithology layer, the corresponding physical and mechanical property data were sorted out. These data were obtained from in-situ tests on-site (such as the standard penetration test SPT) or laboratory geotechnical tests (such as the direct shear test). The shear strength of the silty clay layer measured by the laboratory direct shear test kPa, the shear strength of the medium sand layer converted according to the standard penetration test results kPa, and the shear strength of the strongly weathered granite layer estimated by the rock point load test kPa. In this way, a shear strength sequence corresponding to the lithology code sequence and the layer thickness sequence was obtained , and its numerical unit was unified as kPa.

[0063] Finally, the processing result of this sub-module was to generate a layer thickness sequence (unit: meter) and a shear strength sequence (unit: kPa) reflecting the geological structure of borehole ZK01.

[0064] Table 1 Lithology Stratification and Parameter Table of Borehole ZK01

[0065]

[0066] As shown in Table 1, this table details the codes, depths, thicknesses, and measured or converted shear strength data corresponding to each lithology layer in borehole ZK01.

[0067] The mutation rate calculation sub-module calls the layer thickness sequence, calculates the absolute value of the difference between the current layer thickness and the previous layer thickness in the order of adjacent horizons, performs the difference calculation after unifying the layer thickness unit to meters, divides the difference by the previous layer thickness to obtain the layer thickness mutation rates of multiple horizons, calls the shear strength sequence, calculates the absolute value of the difference between the current strength and the previous strength in the order of adjacent horizons, performs the difference calculation after unifying the shear strength unit to megapascals, and divides the difference by the previous strength to obtain the strength attenuation rates of multiple horizons;

[0068] Call the layer thickness sequence generated by the previous sub-module (unit: meter) and the shear strength sequence (unit: kilopascal). The calculation process is carried out in the order of adjacent horizons, starting from the second horizon.

[0069] First, process the layer thickness sequence, whose unit is already meters and meets the calculation requirements. Calculate the absolute value of the layer thickness difference between the second horizon (medium sand) and its previous horizon (silty clay): meters. Calculate the absolute value of the layer thickness difference between the third horizon (strongly weathered granite) and its previous horizon (medium sand): meters. Then, divide these differences by the corresponding previous layer thicknesses respectively to obtain the layer thickness mutation rates. The layer thickness mutation rate at the interface of the second horizon . The layer thickness mutation rate at the interface of the third horizon . Thus, a layer thickness mutation rate sequence is generated.

[0070] Next, process the shear strength sequence , whose unit is kilopascal (kPa). In order to be consistent with other engineering parameters (usually in megapascal MPa) in subsequent calculations and conduct a unified risk assessment, it is necessary to convert the shear strength unit to megapascal (MPa). The conversion rule is based on the definition of SI prefixes. 1 MPa is equal to 1000 kPa, so dividing the kPa value by 1000 can obtain the MPa value. The converted shear strength sequence is MPa. Then, calculate the absolute value of the strength difference after converting the unit in the order of adjacent horizons. The absolute value of the strength difference between the second horizon and the first horizon: MPa. The absolute value of the strength difference between the third horizon and the second horizon: MPa. Divide these differences by the corresponding previous layer strengths (the values after converting the unit) respectively to obtain the strength change rates (here called strength attenuation rates, which can more accurately reflect the relative change amplitude of strength). The strength attenuation rate at the interface of the second horizon . The strength attenuation rate at the interface of the third horizon . Thus, a strength attenuation rate sequence is generated.

[0071] Finally, two sequences are output: the layer thickness mutation rate sequence and the intensity attenuation rate sequence .

[0072] The anomaly recognition sub-module calls the layer thickness mutation rate and the intensity attenuation rate, calculates the mean and standard deviation of the two sequences, divides the standard deviation by the mean to obtain the coefficient of variation, and generates a lithology mutation mark by screening the layer number with the coefficient of variation exceeding the threshold based on the lithology mutation threshold generation rule verified by the historical geological disaster case database;

[0073] The verification rule of the historical geological disaster case database is that when both the layer thickness mutation rate and the intensity attenuation rate exceed the threshold, the coefficient of variation threshold is reduced to 1.2 times the mean value.

[0074] The sub-module receives the layer thickness mutation rate sequence and the intensity attenuation rate sequence . First, calculate the statistical characteristics of these two sequences respectively:

[0075] The mean of the layer thickness mutation rate sequence ;

[0076] The standard deviation ;

[0077] The coefficient of variation of the layer thickness mutation rate ;

[0078] The mean of the intensity attenuation rate sequence ;

[0079] The standard deviation ;

[0080] The coefficient of variation of the intensity attenuation rate .

[0081] Next, set the lithology mutation threshold for anomaly identification. These thresholds are not set arbitrarily, but are based on a systematic analysis of the case database of historical geological disasters (such as landslides and collapses) in this area. By statistically analyzing the data distribution of the layer thickness mutation rate and the intensity attenuation rate at the locations where disasters occurred and did not occur in history, and using methods such as receiver operating characteristic curve (ROC) analysis, the optimal segmentation points that can better distinguish potential unstable layer interfaces are determined. In this project, the basic threshold of the layer thickness mutation rate , the basic threshold of the intensity attenuation rate are analyzed and determined. Similarly, by analyzing the distribution of the coefficient of variation in historical data, a basic coefficient of variation threshold is determined. When the coefficient of variation exceeds this value, it indicates that the dispersion degree of the parameter sequence has reached a statistically significant level.

[0082] Then, apply an adjustment rule verified based on the historical geological disaster case database: If the thickness mutation rate and the strength attenuation rate of a certain horizon interface both exceed their respective basic thresholds and , then for this horizon interface, the determination threshold of the coefficient of variation needs to be reduced and adjusted to multiplied by the mean of the two coefficients of variation ( and , here referring to the coefficient of variation of the entire sequence, used to set the determination criteria for this specific horizon).

[0083] Calculate this adjusted threshold: ; If the condition of both exceeding the threshold is not met, the coefficient of variation threshold of this horizon interface still adopts the basic value .

[0084] Now, screen two horizon interfaces of borehole ZK01: The second horizon interface (medium sand / silty clay interface): . Because , the thickness mutation rate exceeds the threshold. However, . Because , the strength attenuation rate does not exceed the threshold. Since both do not exceed the threshold simultaneously, the basic coefficient of variation threshold is adopted for this interface. Compare the coefficient of variation of the parameter sequence of this interface (here represented by the coefficient of variation of the entire sequence) with its threshold: and . Since both coefficients of variation do not exceed the threshold, this interface is not determined to be abnormal. The third horizon interface (strongly weathered granite / medium sand interface): . Because , the thickness mutation rate does not exceed the threshold. And . Because , the strength attenuation rate exceeds the threshold. Since both do not exceed the threshold simultaneously, the basic coefficient of variation threshold is also adopted for this interface. Compare the coefficient of variation: and . Since both coefficients of variation do not exceed the threshold, this interface is not determined to be abnormal either.

[0085] To demonstrate the identification of abnormal situations, introduce the data of another borehole ZK02 (located at the pile number K1 + 350). Assume that for a certain horizon interface k of ZK02, and are calculated. Compare the basic thresholds: and . Both conditions are met. Therefore, the coefficient of variation determination threshold for this interface is adjusted to . Assume that the coefficient of variation obtained by analyzing the ZK02 data is and . Compare the adjusted threshold: and . Both coefficients of variation exceed the adjusted threshold. Therefore, this stratigraphic interface of ZK02 (assumed to be at a depth of 15.3 meters) is determined to be an abnormal lithological mutation.

[0086] Filter out all stratigraphic interfaces determined to be abnormal (in this example, the 15.3-meter interface of ZK02), and generate lithological mutation markers for them. The marker contains information: location coordinates (ZK02, depth 15.3 meters), the corresponding layer thickness mutation rate and the intensity attenuation rate . These markers and their associated data will be passed to the next module.

[0087] Please refer to Figure 3 , the risk threshold warning module includes:

[0088] The spatial clustering sub-module calls the lithological mutation markers, extracts the coordinate data of the marker points, sets the clustering radius threshold based on the inverse proportional relationship between the layer thickness mutation rate and the marker point density, calculates the Euclidean distance between the marker points, merges the marker points with a distance less than the threshold into the same group, and generates lithological clustering partitions according to the spatial distribution of the groups;

[0089] Set the clustering radius threshold according to the inverse proportional relationship as clustering radius = 10 meters / (layer thickness mutation rate × 0.1 + 1);

[0090] The module processes the lithological mutation marker points screened by the anomaly identification sub-module from multiple boreholes (ZK01, ZK02, ZK03, etc.) within the region. Extract the three-dimensional spatial coordinates of each marker point and the corresponding layer thickness mutation rate . Assume the following marker point data has been obtained: Marker point A (from the 11.2-meter interface of borehole ZK01, not identified as abnormal, introduced here for demonstrating the complete clustering process, assumed to be marked): coordinates (unit: meters), layer thickness mutation rate . Marker point B (from the 15.3-meter interface of borehole ZK02, identified as abnormal): coordinates (unit: meters), layer thickness mutation rate . Marker point C (from the 9.5-meter interface of borehole ZK03, assumed to be identified as abnormal): coordinates (unit: meters), layer thickness mutation rate . Marker point D (from the 12.0-meter interface of borehole ZK04, assumed to be identified as abnormal): coordinates (Unit: meters), layer thickness mutation rate .

[0091] For each marked point , according to its layer thickness mutation rate dynamically set its clustering radius threshold through the following inverse proportional relationship formula : , the meaning of this formula is: the larger the layer thickness mutation rate , the more drastic the lithology change, and the corresponding clustering radius is smaller, requiring points closer in space to be regarded as the same class; conversely, the gentler the change, the larger the radius. The constant 10 meters is the basic radius reference value, and 0.1 is the adjustment coefficient. These two values are calibrated based on the geological characteristics of this area and engineering experience, aiming to make the clustering scale adapt to the severity of geological changes.

[0092] Calculate the clustering radius of each marked point:

[0093] meters;

[0094] meters;

[0095] meters;

[0096] meters;

[0097] Next, calculate the Euclidean distance between all pairs of marked points :

[0098] meters;

[0099] meters;

[0100] meters;

[0101] meters;

[0102] meters;

[0103] meters;

[0104] Clustering rule: Compare the distance and between any two marked points with their respective clustering radii and . Only when and Only when both are satisfied will these two points be merged into the same group. A and B: , greater than and , not merged. A and C: , greater than and , not merged. A and D: . Because and , the condition is satisfied, and A and D are merged into the same group, denoted as G1. B and C: , greater than and , not merged. B and D: , greater than , not merged (even if less than is not okay, both sides need to be satisfied). C and D: , greater than and , not merged.

[0105] After cluster analysis, the marked points A and D form group G1, while the marked points B and C are independent. According to the spatial distribution ranges of these clustering results (group G1) and the isolated points (B, C), the lithology clustering partitions are defined. The area where G1 is located is defined as partition 1, the area near point B is defined as partition 2, and the area near point C is defined as partition 3. These partitions represent areas with similar lithology change characteristics or dense mutation points.

[0106] The mean comparison sub-module calls the lithology clustering partitions, the layer thickness mutation rate, and the strength attenuation rate, intercepts continuous data segments within the partition according to the preset window length, calculates the arithmetic mean of the mutation rate and the attenuation rate within the window, and compares the mean with the engineering risk threshold set based on the regional geological stability classification to generate an over-threshold interval;

[0107] The rules for setting the regional geological stability classification are as follows: the threshold for the first-level stable area = 0.3, the threshold for the second-level medium area = 0.5, and the threshold for the third-level high-risk area = 0.7;

[0108] The module uses the lithology clustering partitions (partition 1, partition 2, partition 3) divided in the previous step and the original layer thickness mutation rate and strength attenuation rate data sequences of all boreholes (such as ZK01, ZK02, ZK03, ZK04) within each partition. It is necessary to set the length of an analysis window. The setting rule is that considering the average borehole spacing in the area (20 meters) and the geological tectonic activity cycle (evaluated as 3 years), since the borehole spacing is greater than 10 meters and the activity cycle is greater than 1 year, a larger window length layer data points are adopted.

[0109] In partition 1 (the area containing marker points A and D, involving boreholes ZK01 and ZK04), along the borehole depth direction (or in time series if the data is time-sequential), select continuous data segments of several horizons. Select the interface data of 7 continuous horizons within the range of 25 meters to 48 meters in depth of borehole ZK04. The sequence of layer thickness mutation rates is , and the corresponding sequence of intensity attenuation rates is .

[0110] Calculate the arithmetic mean of the data within this window:

[0111] Mean of layer thickness mutation rate: ;

[0112] Mean of intensity attenuation rate: ;

[0113] Compare the two calculated mean values and with the engineering risk threshold. This threshold is set according to the regional geological stability classification. According to the latest regional geological investigation report, the area where this engineering project is located is rated as a "secondary medium-stable area". Based on the preset rules: "Threshold for primary stable area = 0.3, threshold for secondary medium area = 0.5, threshold for tertiary high-risk area = 0.7", so the engineering risk threshold is selected. Make the comparison: Within this window data segment, the mean values of both parameters do not exceed the risk threshold of 0.5. This indicates that the average change rate of geological parameters in this depth range (25 meters - 48 meters) is within the acceptable level of the "secondary medium-stable area" and does not reach the conditions for triggering an early warning.

[0114] Move the analysis window down one horizon along the depth, select 7 new data points, and repeat the above mean value calculation and comparison. When the window moves to the range of 35 meters to 55 meters in depth of borehole ZK04, the intercepted sequence of layer thickness mutation rates is , and the sequence of intensity attenuation rates is .

[0115] Calculate the mean values:

[0116] ;

[0117] ;

[0118] Compare with the threshold again: It still does not exceed the threshold. Continue to slide the window.

[0119] When the window moves to the range of 40 to 60 meters in depth of borehole ZK04, the sequence of layer thickness mutation rates intercepted is as follows: ;

[0120] The sequence of strength attenuation rates is as follows: ;

[0121] Calculate the mean value: ,

[0122] ; Compare the threshold : The average value of the strength attenuation rate exceeds the risk threshold . Therefore, record the depth interval (40 to 60 meters) covered by this window as the "super-threshold interval". All identified super-threshold intervals will be passed to the next module.

[0123] The early warning generation sub-module calls the super-threshold interval, performs a moving average calculation on the mutation rate and attenuation rate within the interval according to the time series, sets the window weight coefficient to be 0.7 times that of the previous window determined by historical data regression analysis, and generates an early warning signal distribution map and a parameter prediction trend.

[0124] Receive the super-threshold interval information identified by the previous module. For example, it is determined that the range of borehole ZK04 from 40 to 60 meters in depth is a super-threshold interval based on the strength attenuation rate. Extract the detailed strength attenuation rate data sequence within this interval: (Two subsequent points are added).

[0125] For this sequence Apply the moving average calculation to smooth the data and reveal the trend. The moving window size is set to 5 data points. It is necessary to set the weight coefficients in the calculation. These coefficients are determined by performing a time series regression analysis on the historical geological disaster evolution data in this area. The analysis results show that the data points in the recent period (shallower layer or more recent time) are more predictive of future trends. Based on this, determine that the influence weight of the previous window (representing the past state) on the prediction of the current window is , then the weight of the current data point is . Use the exponential smoothing method for calculation. The smoothed prediction value of the th point is given by the following formula: , where is the actual strength attenuation rate value of the th point in the sequence, is the smoothed prediction value calculated from the previous point. The initial value Set to the first value of the sequence 。

[0126] The calculation process is as follows:

[0127] ;

[0128] ;

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] ;

[0134] ;

[0135] The calculated smoothed sequence (rounded to three decimal places). Observing this sequence, its values show a continuous upward trend, especially in the second half, exceeding the threshold of 0.5 and continuing to increase. This indicates that the intensity attenuation rate not only exceeds the threshold but also has a tendency to deteriorate.

[0136] According to the degree of exceeding the threshold (the average value is higher than the threshold of 0.5 by ), and the predicted trend (the smoothed value continues to rise), determine the warning level for this interval. Set the warning level rules: When the degree of exceeding the threshold is 0 - 10% and the trend is stable, it is a first-level (low-risk) warning; when the threshold is exceeded by 0 - 10% but the trend is upward, or the threshold is exceeded by 10% - 30%, it is a second-level (medium-risk) warning; when the threshold is exceeded by more than 30%, it is a third-level (high-risk) warning. In this example, the threshold is exceeded by 3.8% and the trend is upward, so it is determined as a second-level (medium-risk) warning.

[0137] Associate this warning information with the location. The ZK04 borehole is located near the highway stake number K1 + 800, and the depth of the interval exceeding the threshold is 40 - 60 meters. Finally, generate a warning signal distribution map, mark the area from stake number K1 + 780 to K1 + 820 (covering the affected range) and the depth of 40 - 60 meters as the second-level warning area on the engineering drawing, and indicate that the main triggering parameter is the intensity attenuation rate. At the same time, output a parameter prediction trend curve graph, showing the smoothed value of the intensity attenuation rate and the short-term prediction based on the last point (such as predicting that the intensity attenuation rate may reach 0.65 within the next 5 meters of depth).

[0138] Please refer to Figure 4 The stability weight evaluation module includes:

[0139] The slope feature quantization sub-module obtains digital elevation model data, calculates the difference in slope angles between adjacent elevation points, extracts the maximum and minimum slope angle differences within the grid cell, generates the slope variation amplitude, converts the slope direction angle to polar coordinates, calculates the standard deviation to eliminate the periodic influence, and outputs the slope direction dispersion;

[0140] This module processes digital elevation model (DEM) data covering the project area. A DEM raster data with a spatial resolution of 5 meters is selected 5-meter DEM raster data. For each grid cell (located in the row, column) to calculate the terrain features.

[0141] First, calculate the slope variation amplitude. For the grid , consider itself and the eight adjacent grids around it (forming a 3 ×3 window). Calculate the slope value of each grid within this 3 ×3 window (the slope is the angle of the maximum change rate of elevation on the horizontal plane, which can be calculated from the elevation difference and the horizontal distance). Find the maximum value and the minimum value among these 9 slope values. The slope variation amplitude is defined as the difference between the two: . For a specific grid cell , the maximum slope calculated within its 3 ×3 neighborhood is , the minimum slope is , then the slope variation amplitude of this grid .

[0142] Secondly, calculate the slope direction dispersion. First, calculate the slope direction of each grid within the 3 ×3 window (the projection direction of the slope vector on the horizontal plane, represented by 0 - 360 degrees, with due north being 0 degrees). To avoid the 0 / 360 degree boundary problem and effectively calculate the dispersion degree, convert each slope direction angle to a two-dimensional unit vector . For a grid with a slope direction of 45 degrees (northeast direction), its vector is . Calculate the average vector of the 9 such unit vectors within the 3 ×3 window, where and . Calculate the average slope direction according to the average vector. This function can correctly handle angles in all quadrants. Then, calculate the slope direction of each grid within the window and the average slope direction The angular difference (note that the case of crossing 0 / 360 degrees needs to be handled, and the acute angle difference is taken). Finally, calculate the standard deviation of these 9 angular differences to obtain the slope direction dispersion of this central grid of the grid . For the grid , the calculated slope direction dispersion is .

[0143] This module outputs two quantified terrain feature values for each grid cell in the study area : the amplitude variation of slope (such as ) and the slope direction dispersion (such as ).

[0144] The instability index calculation sub-module calls the amplitude variation of slope for normalization, converts the slope direction dispersion into a positive correlation coefficient through the inverse proportional function f(x)=1 / (x + 1), superimposes the elevation anomaly point density data, and performs a weighted product operation according to the 4:3:3 weight based on the sensitivity analysis of geological parameters to generate the terrain instability index at the center point of the grid cell;

[0145] The sensitivity analysis of geological parameters determines the contribution rate weights of the amplitude variation of slope, the slope direction dispersion, and the elevation anomaly point density through the principal component analysis method;

[0146] Receive the amplitude variation of slope calculated for each grid in the previous step and the slope direction dispersion . It is also necessary to introduce a third terrain factor: the elevation anomaly point density . This data is obtained by analyzing the DEM, identifying the points (anomaly points) where the elevation has a significant deviation from the average elevation of its local neighborhood, and then counting the density of these anomaly points within the neighborhood of each grid . The elevation anomaly point density at grid is points per 100 square meters.

[0147] In order to comprehensively evaluate the terrain instability risk, it is necessary to combine these three heterogeneous factors ( , , ). First, perform data standardization processing to eliminate the differences in units and numerical ranges. Normalize the amplitude variation of slope . It is necessary to first count the amplitude variation of slope of all grids in the entire study area to obtain the minimum value and the maximum value . After the whole area statistics, it is determined that and . Use the min-max scaling method for normalization: , for the slope direction dispersion , the original text provides a conversion function such that the greater the dispersion, the smaller the factor value. The calculated . This converted factor is dimensionless. For the density of elevation anomaly points points per 100 square meters, normalization is also performed. The statistical density value range for the entire region is [0, 0.5] points per 100 square meters. Normalization:

[0148] ;

[0149] Next, determine the weights of these three standardized / converted factors ( , , ) when calculating the terrain instability index. The weights are determined based on a sensitivity analysis of the relationship between the historical data of geological disasters in the area and these terrain factors. Using the principal component analysis method (PCA), identify the principal component most strongly associated with geological instability, and then analyze the loadings (contribution rates) of these three factors on this principal component to determine the weights. The analysis results show that the contribution rate of the normalized slope range is 40%, the contribution rate of the (converted) aspect dispersion factor is 30%, and the contribution rate of the normalized elevation anomaly point density is 30%. Therefore, the weights are set as , , .

[0150] Calculate the terrain instability index of the grid using the weighted sum method :

[0151] ;

[0152] ;

[0153] ;

[0154] , and the calculated terrain instability index of the grid is 0.287. Repeat this calculation for all grids in the study area to generate a terrain instability index distribution map.

[0155] The prevention and control measure classification sub-module divides the terrain instability index into four intervals based on the fuzzy comprehensive evaluation method, defines the vertex parameters of the triangular membership function as instability indices 0.2, 0.5, and 0.8, generates the prevention and control priority coefficients, and outputs the overall instability index distribution.

[0156] This module uses the overall terrain instability index generated by the previous module Distribution map. The fuzzy comprehensive evaluation method is used to divide the risk levels of the terrain instability index for each grid. Four risk levels are set: low risk (L), medium - low risk (ML), medium - high risk (MH), and high risk (H). Membership functions are defined for each level to describe the degree to which a given index value belongs to each level.

[0157] The triangular membership function is used, and its shape is determined by the vertex parameters. The vertex parameters are set to 0.2, 0.5, and 0.8. These vertices define the positions where the membership degrees of each risk level reach the peak or start / end. The specific definitions are as follows: Low risk (L): The membership function is 1 at and drops to 0 at and is 0 for . (This is a right - descending semi - trapezoid or simplified to linearly descend in the interval [0, 0.2]) - A more reasonable definition might be that the peak is at 0 and the support interval is [0, 0.2]. Assume the latter: . Medium - low risk (ML): The membership function is a triangle with the vertex at and the support interval (non - zero range) is . Calculation formula: If , ; if , . Medium - high risk (MH): The membership function is a triangle with the vertex at and the support interval is . Calculation formula: If , ; if , . High risk (H): The membership function is a triangle with the vertex at and the support interval is (assuming the upper limit is 1.0). Calculation formula: If , ; if , .

[0158] For the terrain instability index of the grid calculated in the previous module, calculate its membership degrees to these four risk levels: : Since , the second formula is applicable, : Since , the first formula is applicable, (since );

[0159] Compare the four membership values: . The maximum value is . According to the principle of maximum membership, it is determined that the risk level of this grid is medium - low risk (ML).

[0160] Set a prevention and control priority coefficient for each risk level to guide the formulation of subsequent measures. Setting rules: the coefficient for low risk (L) is 1, the coefficient for medium - low risk (ML) is 2, the coefficient for medium - high risk (MH) is 3, and the coefficient for high risk (H) is 4. Therefore, the prevention and control priority coefficient of grid is .

[0161] Apply this fuzzy evaluation and priority coefficient allocation process to all grid cells in the terrain instability index distribution map, and finally generate a risk level distribution map and a prevention and control priority coefficient distribution map covering the entire region.

[0162] Please refer to Figure 5 , the risk comprehensive management and control module includes:

[0163] The stake number interval analysis sub - module obtains the pixel coordinates and warning intensity values of the warning signal distribution map, converts the horizontal and vertical coordinates into stake number section identifiers based on the stake number coding rule, calculates the absolute value of the difference in warning intensity between adjacent stake numbers, compares it with the warning fluctuation threshold set according to the historical disaster frequency of the region, and generates a stake number interval table;

[0164] The rule for setting the historical disaster frequency of the region is: when the disaster frequency ≥ 5 times / year, the fluctuation threshold = 0.5;

[0165] This module processes the warning signal distribution map generated by the risk threshold warning module, and this map shows the warning intensity values along the engineering line (such as roads, railways) . At the same time, the stake number coding rule of the project is required. The starting stake number of the project is K0 + 000, and a large stake number (K0 + 100, K0 + 200…K1 + 000…) is set every 100 meters along the line direction, and a small stake number mark is set every 10 meters.

[0166] First, convert the position information on the warning signal map (such as the distance along the line ) into a stake number identifier. At 1535 meters from the starting point, its stake number is K1 + 535. Extract the warning intensity values at every 10 - meter interval (a small stake number interval) along the line to obtain a sequence. The warning intensity value sequence of a section of the line is as follows:

[0167] ;

[0168] ;

[0169] ;

[0170] ;

[0171] ;

[0172] ;

[0173] ;

[0174] ;

[0175] ;

[0176] Calculate the absolute value of the difference in warning intensity between adjacent station numbers (at 10-meter intervals):

[0177] K1+510 vs K1+500: ;

[0178] K1+520 vs K1+510: ;

[0179] K1+530 vs K1+520: ;

[0180] K1+540 vs K1+530: ;

[0181] K1+550 vs K1+540: ;

[0182] K2+110 vs K2+100: ;

[0183] K2+120 vs K2+110: ;

[0184] Set a warning fluctuation threshold to determine whether the change in warning intensity is drastic. The setting of this threshold is based on the historical frequency of geological disasters in the study area. Query the records to obtain the average number of geological disasters occurring in this area per year in the past ten years times. According to the setting rule: "If times / year, then ; if times / year, then ". Since , select .

[0185] Compare the absolute value of the strength difference between adjacent station numbers calculated with the threshold Comparison: For the section from K1+500 to K1+550, all differences (0.05, 0.10, 0.05, 0.02, 0.02) are less than 0.5. For the section from K2+100 to K2+110, the difference . For the section from K2+110 to K2+120, the difference .

[0186] Meanwhile, define the risk level of the early warning intensity itself: the intensity value is high risk, is medium risk, is low risk. Volatility assessment: the absolute value of the difference is high volatility, is low volatility.

[0187] Based on the risk level of the early warning intensity and the volatility assessment results, merge consecutive stake number sections with similar states to divide different stake number risk intervals. Stake number K1+500 to K1+510: intensity 0.60, 0.65 (medium risk), volatility 0.05 (low). Stake number K1+510 to K1+520: intensity 0.65, 0.75 (medium -> high risk), volatility 0.10 (low). Stake number K1+520 to K1+530: intensity 0.75, 0.70 (high risk), volatility 0.05 (low). Stake number K1+530 to K1+550: intensity 0.70, 0.68, 0.66 (high -> medium risk), volatility 0.02, 0.02 (low). Overall for K1+500 to K1+550: the risk level fluctuates between medium and high, but the intensity change is relatively gentle (low volatility). Can be merged into interval A. Stake number K2+100 to K2+110: intensity 0.40, 0.95 (low -> high risk), volatility 0.55 (high). Stake number K2+110 to K2+120: intensity 0.95, 0.85 (high risk), volatility 0.10 (low). Overall for K2+100 to K2+120: the risk level jumps from low to high, and high volatility appears. Can be merged into interval B.

[0188] Generate a stake number interval table to summarize the analysis results.

[0189] Table 2 Stake Number Risk Interval Division Table

[0190]

[0191] As shown in Table 2, based on the early warning intensity value and the degree of its sharp change along the line, the line is divided into intervals with different risk characteristics.

[0192] Based on the mileage interval table, the measure mapping generation sub-module calls the parameter prediction trend to judge the risk evolution direction, extracts the grade values of the prevention and control recommendation priorities and the grid cell values of the instability index distribution, divides the support measures into bolt support, pile foundation support, and retaining wall support types according to the priority level, and establishes a one-to-many mapping relationship between the mileage and the measure type and strength, generating a mileage-measure mapping table;

[0193] The module performs subsequent processing based on the mileage interval table (see Table 2) generated by the previous module. Select interval B (mileage K2+100 to K2+120, high risk, high fluctuation) in Table 2 as an example.

[0194] First, call the parameter prediction trend output by the risk threshold warning module for this interval (K2+100 to K2+120). Query to obtain the strength attenuation rate of this interval The prediction trend is "it is expected to maintain a high level above 0.8 within the next monitoring period", indicating that the risk state is persistent.

[0195] Secondly, extract the topographic and geological information generated by the stability weight evaluation module that covers this mileage interval. Query the terrain instability index distribution map and the prevention and control priority coefficient distribution map, and obtain that the average terrain instability index of the grid cells within the range of mileage K2+100 to K2+120 is , and the corresponding prevention and control priority coefficient is mainly 4 (high risk).

[0196] Integrate the information of interval B: high warning intensity, high volatility, prediction trend maintaining a high level, high terrain instability index (0.75), and high prevention and control priority coefficient (level 4). Based on this information, appropriate engineering measures need to be selected. Establish a measure selection rule that associates the risk level, priority coefficient with the specific support measure type and strength: *Priority coefficient 1 or 2 (low / medium-low risk): Generally choose slope protection (such as vegetation, lattice) and simple retaining structures (such as dry stone masonry). *Priority coefficient 3 (medium-high risk): Usually adopt systematic bolt / cable support or small anti-slide piles. *Priority coefficient 4 (high risk): Strong support measures need to be adopted, such as large-diameter anti-slide piles, pile-slab walls, or gravity retaining walls.

[0197] For interval B, its comprehensive evaluation points to the highest risk level (priority 4), so the support measure with the highest strength is selected. Considering the local terrain (assumed to be a relatively steep slope) and geological conditions (assumed to have a weak interlayer) at this location, it is determined to use large-diameter (such as 1.5 meters in diameter) reinforced concrete anti-slide piles for support. Further, based on the predicted trend (the strength decay rate remains at a high level) and the composition of the terrain instability index (assuming that the slope variation range and the density of elevation anomaly points contribute greatly), the design parameters are refined. For example, it is determined that the pile length needs to penetrate the sliding surface to reach the stable bedrock, and the pile spacing is determined according to the force calculation (such as 4 meters).

[0198] Establish a mapping relationship between the pile number interval B (K2+100 to K2+120) and the selected engineering measures (type: large-diameter anti-slide pile, strength: enhanced design, parameters: pile diameter 1.5m, pile length XXm, pile spacing 4m). Repeat this process for all pile number intervals that require measures to generate the final pile number - measure mapping table, which details the specific engineering measures to be implemented for each risk section and their design key points.

[0199] The measure priority ranking sub-module constructs the criterion layer and the scheme layer structure using the analytic hierarchy process (AHP). The ratio of the weight of the support measure to the weight of the drainage measure is set to 3:2. Calculate the weight value of each measure for the risk control of the pile number, and generate the measure priority ranking in descending order of the weight value, and bind the pile number interval to output the segmented project control plan.

[0200] To ensure the effective utilization of resources and the risk control effect, it is necessary to rank the multiple measures (such as support, drainage, and monitoring) that may be involved in the same pile number interval. The analytic hierarchy process (AHP) is used to construct a decision-making model.

[0201] The model structure is set as follows: *Objective layer (Level0): Optimization of the comprehensive control of the pile number interval risk. *Criterion layer (Level1): The main factors affecting the risk control effect, set as: effectiveness of the support measure (C1), effectiveness of the drainage measure (C2), timeliness of the monitoring and early warning (C3). *Scheme layer (Level2): Specific engineering measure options proposed for a specific pile number interval (such as interval B), including: anti-slide pile support (S1), slope intercepting and drainage ditch (S2), deep drainage holes (S3), automated displacement monitoring (S4).

[0202] Construct the judgment matrix of the criterion layer relative to the objective layer. Based on engineering experience and expert consensus, it is considered that for the high-risk and highly volatile interval B, the support measure is the most critical, followed by drainage, and finally monitoring. And, it is clearly required that the weight of the support measure and the weight of the drainage measure The ratio is 3:2. This means that when constructing the judgment matrix, the importance scale of C1 to C2 is set to 3 / 2 = 1.5. By comparing pairwise with other criteria (for example, support is extremely important for monitoring, with a scale of 5; drainage is relatively important for monitoring, with a scale of 3), the judgment matrix is constructed, and by calculating the maximum eigenvalue of the matrix and the corresponding normalized eigenvector, the weight vector of the criterion layer is obtained. The calculation result is . This result meets the requirements.

[0203] Next, for interval B, construct the judgment matrix of each measure in the solution layer with respect to each criterion. For example, with respect to "the effectiveness of the support measure (C1)", the anti-slide pile (S1) is the most directly effective and has the highest weight (such as 0.6); other measures (drainage, monitoring) have an indirect impact on the support effect and have lower weights (such as S2 = 0.1, S3 = 0.2, S4 = 0.1). With respect to "the effectiveness of the drainage measure (C2)", the deep drainage hole (S3) has the best effect (such as 0.5), the intercepting drainage ditch (S2) comes second (such as 0.3), and the roles of support and monitoring are relatively small (S1 = 0.1, S4 = 0.1). With respect to "the timeliness of monitoring and early warning (C3)", the automated monitoring (S4) is the most relevant (such as 0.7), and the relevance of other measures is low (S1 = 0.1, S2 = 0.1, S3 = 0.1). Through calculation, the weight vector of the solution layer under each criterion is obtained.

[0204] Finally, calculate the comprehensive weight of each solution (measure), that is, the weighted sum of its weights under each criterion (the weights are the weights of the criterion layer):

[0205] Comprehensive weight (S1) = ;

[0206] Comprehensive weight (S2) = ;

[0207] Comprehensive weight (S3) = ;

[0208] Comprehensive weight (S4) = ;

[0209] Arrange the calculated comprehensive weight values in descending order: S1(0.350) > S3(0.282) > S4(0.202) > S2(0.166). This indicates that for stake number interval B, the priority order of the measures is: anti-slide pile support > deep drainage hole > automated displacement monitoring > slope intercepting drainage ditch.

[0210] Bind this sorting result to the mileage interval B. Repeat this sorting process for all mileage intervals that require intervention, and finally output a detailed segmented project control plan, which clarifies the combination of measures to be taken for each risk section, their respective priorities, and key design parameters.

[0211] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An integrated safety engineering management system, characterized in that, The system includes: A risk threshold warning module, which is used to receive lithology mutation marks and perform spatial clustering analysis, compare the average value of the sliding window length dynamically adjusted based on the borehole spacing and the geological structure activity cycle for the layer thickness mutation rate and the strength attenuation rate, generate a segmented project warning signal when the threshold is exceeded, use the moving average model regression analysis to determine the window weight coefficient, and generate and transmit the warning signal distribution map and parameter prediction trend to the risk comprehensive control module; A stability weight evaluation module, which is used to calculate the grid slope amplitude change and slope direction dispersion through a digital elevation model, generate a terrain instability index in combination with the elevation anomaly point density, normalize the slope amplitude change value using the maximum-minimum method, output the prevention and control suggestion priority and instability index distribution, and transmit it to the risk comprehensive control module; A risk comprehensive control module, which is used to receive the pile number interval matching the warning signal distribution map, call the parameter prediction trend to judge the risk evolution direction, generate a pile number - measure mapping relationship table in combination with the prevention and control suggestion priority and the instability index distribution, use the analytic hierarchy process to construct the criterion layer weights of support, drainage, and monitoring measures, and output the segmented project control plan bound to the pile number.

2. The comprehensive safety engineering management system according to claim 1, wherein The segmented project warning signal includes a warning level, a warning section range, and a warning trigger type. The warning signal distribution map includes a warning pile number distribution, a spatial risk concentration area, and a warning signal density. The parameter prediction trend includes a layer thickness change trend, a strength evolution trend, and a comprehensive risk change trend. The terrain instability index includes a slope change index, a slope direction dispersion index, and an elevation anomaly index. The prevention and control suggestion priority specifically includes a support priority, a drainage priority, and a monitoring priority. The instability index distribution includes a high-risk instability area, a medium-risk instability area, and a low-risk instability area. The pile number - measure mapping relationship table includes the support measures corresponding to the pile number, the drainage measures corresponding to the pile number, and the monitoring measures corresponding to the pile number. The segmented project control plan includes a segmented support plan, a segmented drainage plan, and a segmented monitoring plan.

3. The comprehensive safety engineering management system according to claim 2, wherein The dynamic adjustment rule of the sliding window length is that when the borehole spacing is less than 10 meters and the tectonic activity cycle is less than 1 year, the window length is set to 5 layers; The normalization formula of the maximum-minimum method is slope amplitude change value = (original value - minimum value) / (maximum value - minimum value); The criterion layer weights include that the support measure weight is dynamically allocated according to the product result of the instability index distribution and the parameter prediction trend.

4. The comprehensive safety engineering management system according to claim 3, characterized in that The risk threshold warning module includes: The spatial clustering sub-module calls the lithology mutation mark, extracts the coordinate data of the marked points, sets the clustering radius threshold based on the inverse proportional relationship between the layer thickness mutation rate and the marked point density, calculates the Euclidean distance between the marked points, merges the marked points with a distance less than the threshold into the same group, and generates a lithology clustering partition according to the spatial distribution of the groups; The inverse proportional relationship sets the clustering radius threshold as clustering radius = 10 meters / (layer thickness mutation rate × 0.1 + 1); The mean comparison sub-module calls the lithology clustering partition, the layer thickness mutation rate, and the intensity attenuation rate, intercepts continuous data segments within the partition according to a preset window length, calculates the arithmetic mean of the mutation rate and the attenuation rate within the window, compares the mean with the engineering risk threshold set based on the regional geological stability classification, and generates an over-threshold interval; The rules for setting the regional geological stability classification are: the threshold for the first-level stable area = 0.3, the threshold for the second-level medium area = 0.5, and the threshold for the third-level high-risk area = 0.7; The early warning generation sub-module calls the over-threshold interval, performs a moving average calculation on the mutation rate and the attenuation rate within the interval according to the time series, sets the window weight coefficient to be 0.7 times that of the previous window determined by historical data regression analysis, and generates an early warning signal distribution map and a parameter prediction trend.

5. The comprehensive safety engineering management system according to claim 4, wherein The stability weight evaluation module includes: The slope feature quantification sub-module obtains digital elevation model data, calculates the slope angle difference between adjacent elevation points, extracts the maximum and minimum slope angle differences within the grid cell, generates the slope variation amplitude, converts the slope direction angle to polar coordinates and then calculates the standard deviation to eliminate the periodic influence, and outputs the slope direction dispersion; The instability index calculation sub-module calls the slope variation amplitude for normalization processing, converts the slope direction dispersion into a positive correlation coefficient through the inverse proportional function f(x)=1 / (x + 1), superimposes the elevation anomaly point density data, and performs a weighted product operation according to the 4:3:3 weight based on the geological parameter sensitivity analysis to generate the terrain instability index at the center point of the grid cell; The geological parameter sensitivity analysis determines the contribution rate weights of the slope variation amplitude, the slope direction dispersion, and the elevation anomaly point density through the principal component analysis method; The prevention and control measure classification sub-module divides the terrain instability index into four types of intervals based on the fuzzy comprehensive evaluation method, defines the vertex parameters of the triangular membership function as the instability index 0.2, 0.5, and 0.8, generates the prevention and control priority coefficient, and outputs the global instability index distribution.

6. The comprehensive safety engineering management system according to claim 5, wherein The risk comprehensive control module includes: The mileage interval analysis sub-module obtains the pixel coordinates and the early warning intensity values of the early warning signal distribution map, converts the horizontal and vertical coordinates into mileage segment identifiers based on the mileage coding rule, calculates the absolute value of the early warning intensity difference between adjacent mileages, compares it with the early warning fluctuation threshold set according to the regional historical disaster frequency, and generates a mileage interval table; The rules for setting the regional historical disaster frequency are: when the disaster frequency ≥ 5 times / year, the fluctuation threshold = 0.5; The measure mapping generation sub-module, based on the mileage interval table, calls the parameter prediction trend to judge the risk evolution direction, extracts the grade values of the prevention and control recommendation priorities and the grid cell values of the instability index distribution, divides the support measures into bolt support, pile foundation support, and retaining wall support types according to the priority level, and establishes a one-to-many mapping relationship between the mileage and the measure type and intensity, and generates a mileage-measure mapping table; The measure priority ranking sub-module constructs the criterion layer and the scheme layer structure using the analytic hierarchy process, sets the ratio of the support measure weight to the drainage measure weight to 3:2, calculates the weight value of each measure for the risk control of the pile number, generates the measure priority ranking in descending order of the weight value, and binds the pile number interval to output the sectional project control plan.

7. The integrated safety engineering management system according to claim 6, characterized in that The system further includes: The geological parameter dynamic monitoring module is used to extract the lithology distribution, layer thickness sequence, and shear strength sequence through the drilling profile, calculate the layer thickness mutation rate and strength attenuation rate of adjacent segments, optimize the threshold generation logic by combining the layer thickness mutation rate and the dynamic coupling relationship of the strength attenuation rate using the coefficient of variation method, identify the lithology mutation abnormal area, generate the lithology mutation mark, and transmit it to the risk threshold warning module.

8. The comprehensive safety engineering management system according to claim 7, characterized in that The lithology mutation mark specifically includes mutation position identification, mutation area boundary, and mutation type classification. The layer thickness mutation rate includes layer thickness variation range, layer thickness variation degree, and local layer thickness difference. The strength attenuation rate includes shear strength variation range, strength decline rate, and abnormal strength gradient.

9. The comprehensive safety engineering management system according to claim 8, wherein The dynamic coupling relationship optimization threshold generation logic is to adjust the coefficient of variation threshold according to the correlation coefficient between the layer thickness mutation rate and the strength attenuation rate. When the correlation coefficient is greater than 0.5, the threshold is increased to 1.8 times the mean value of the coefficient of variation in the current dataset.

10. The integrated safety engineering management system according to claim 9, characterized in that, The geological parameter dynamic monitoring module includes: The lithology sequence extraction sub-module obtains the drilling profile data, divides the lithology horizons using the interval segmentation method based on the mineral composition and particle size parameters in the lithology classification standard, extracts the lithology codes of multiple horizons, and statistically analyzes the layer thickness and shear strength data to generate the layer thickness sequence and the shear strength sequence. The mutation rate calculation sub-module calls the layer thickness sequence, calculates the absolute value of the difference between the current layer thickness and the previous layer thickness in the order of adjacent horizons, unifies the layer thickness unit to meters before performing the difference calculation, divides the difference by the previous layer thickness to obtain the layer thickness mutation rate of multiple horizons, calls the shear strength sequence, calculates the absolute value of the difference between the current strength and the previous strength in the order of adjacent horizons, unifies the shear strength unit to megapascals before performing the difference calculation, and divides the difference by the previous strength to obtain the strength attenuation rate of multiple horizons. The abnormal identification sub-module calls the layer thickness mutation rate and the strength attenuation rate, calculates the mean value and standard deviation of the two sequences, divides the standard deviation by the mean value to obtain the coefficient of variation, and based on the lithology mutation threshold generation rule verified by the historical geological disaster case database, filters the horizon numbers with the coefficient of variation exceeding the threshold to generate the lithology mutation mark. The verification rule of the historical geological disaster case database is that when both the layer thickness mutation rate and the strength attenuation rate exceed the threshold, the coefficient of variation threshold is reduced to 1.2 times the mean value.

Images