A foundation pit construction risk judgment method and system based on multi-parameter correlation analysis

Abstract

The present application relates to the technical field of information retrieval analysis, in particular to a foundation pit construction risk judgment method and system based on multi-parameter correlation analysis, comprising the following steps: collecting the deformation of the vibrating string stress meter and the elastic modulus calculation energy storage of the support member, combining the external work difference to obtain the dissipation gap and generate the dissipation deviation, performing partial derivation on the water level variable to extract multi-dimensional sensitive features, establishing weighted ranging combined with the deviation vector to obtain the span and early warning boundary, judging the level limit trigger signal and outputting the risk judgment result, in the present application, the equivalent energy storage and work difference value are extracted to quantify the energy dissipation gap and establish a two-dimensional mapping system, the low-risk elements are removed to generate the dissipation deviation to deeply mine the water level sensitive features, the ranging structure is established combined with the safety limit deviation to calculate the penalty span and the danger limit, and the lagging disadvantage is overcome to realize the panoramic dynamic early warning of the potential cumulative damage effect under the multi-dimensional complex interactive evolution of the foundation pit periphery.

Description

Technical Field

[0001] This invention relates to the field of information retrieval and analysis technology, and in particular to a method and system for determining the construction risk of foundation pits based on multi-parameter correlation analysis. Background Technology

[0002] The field of information retrieval and analysis technology encompasses core aspects such as feature extraction and multi-dimensional indicator comparison of various datasets. This field establishes underlying data tables to find logical correspondences between time series and spatial distributions from numerical variables across different dimensions, integrating historical discrete feature values ​​with real-time collected indicator values ​​to construct a fundamental relational mapping structure. Among these, the traditional method for determining the risk of foundation pit construction based on multi-parameter correlation analysis refers to the specific process of numerically defining the safety status of the surrounding soil and support structure during the foundation pit excavation stage. This typically involves drilling holes in the soil around the foundation pit and lowering inclinometer probes to read horizontal displacement values ​​at different depths; binding vibrating wire stress gauges to the main reinforcing bars of reinforced concrete supports to read stress deformation values; and using a plumb bob to measure the liquid level in groundwater observation wells. Subsequently, on-site personnel input the independent physical quantity values ​​read by various sensors into an electronic spreadsheet, plotting a line graph of a single variable changing over time, and comparing each independent physical quantity value with the static fixed limit values ​​given in the construction safety drawings and specifications.

[0003] Traditional risk assessment processes for foundation pit construction often involve reading independent physical quantities on-site, filling them into spreadsheets, drawing single-variable line graphs, and independently comparing single variables with static fixed limits. This results in the fragmentation of discrete, multi-dimensional, heterogeneous parameters collected by different types of sensors, making it difficult to uncover the deep logical correspondence between time-series characteristics and spatial distribution characteristics. Relying solely on isolated static indicators for surface definition easily overlooks the complex cumulative destructive effects generated by the synergistic evolution of the resistance of the surrounding soil and the stress on the supporting components. This leads to serious lag biases and risks of misjudgment and underreporting in the overall risk level assessment. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a method for determining the construction risk of foundation pits based on multi-parameter correlation analysis, comprising the following steps: S1: Collect stress deformation data of vibrating wire stress gauge and elastic modulus data of supporting component, calculate equivalent elastic energy storage, extract horizontal displacement data of inclinometer and perform integral calculation with soil resistance to extract external work, and differentially calculate equivalent elastic energy storage and external work to obtain the energy dissipation characteristic difference. S2: Extract the energy dissipation feature difference and the corresponding underlying sensor plane spatial coordinates, write them into a two-dimensional plane data table according to the spatial coordinate mapping relationship, extract the differences between adjacent elements and compare them with the preset risk warning benchmark threshold, remove elements below the threshold, and generate a plane energy dissipation deviation. S3: Extract the water level elevation variable from the plane energy dissipation deviation, solve the partial derivative, extract the sensitive features, obtain the absolute deviation vector between the static safety limit coordinates and the multi-source fusion coordinates, perform Hadamard operation, and establish multi-source sensitive weighted ranging. S4: Squaring the multi-source sensitive weighted ranging elements and accumulating them, taking the square root to obtain the penalty distance span, comparing it with the failure limit distance threshold, and triggering an instruction when the penalty distance span is greater than the threshold to obtain the cross-domain attenuation warning boundary. S5: Based on the cross-domain attenuation early warning boundary marker, retrieve the risk triggering rules, determine the danger level boundary, trigger the audible and visual alarm activation level signal, and output the foundation pit construction risk assessment result.

[0005] As a further aspect of the present invention, the energy dissipation characteristic difference includes the hysteresis dissipation value of the support system, the plastic deformation potential of the retaining structure, and the difference in irreversible work done by the soil; the planar energy dissipation deviation includes the extreme value of the local grid residual, the distortion gradient of adjacent measuring points, and the regional stress imbalance index; the multi-source sensitive weighted ranging includes the water level fluctuation sensitive offset, the mechanical coupling deformation vector, and the spatial instability span coordinates; the cross-domain attenuation early warning boundary includes the state degradation critical contour, the ultimate bearing failure red line, and the safety redundancy depletion bottom marker; and the foundation pit construction risk assessment result includes the disaster evolution early warning level, the emergency control intervention plan, and the sound and light linkage execution strategy.

[0006] As a further aspect of the present invention, the risk warning benchmark threshold refers to a reference warning threshold used to determine whether the difference in energy dissipation between adjacent spatial locations constitutes an anomaly, eliminating normal minor fluctuations and extracting risk characteristics.

[0007] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Collect the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component on site; perform a square operation on the stress deformation data of the vibrating wire stress gauge on site to obtain the square value of deformation; divide the square value of deformation by the elastic modulus data of the supporting component to generate equivalent elastic energy storage. S102: Obtain the horizontal displacement data of the inclinometer probe and the soil resistance, perform a multiplication operation on the horizontal displacement data of the inclinometer probe and the soil resistance to obtain the local work, and perform a definite integral operation on the local work along the depth direction to generate the external work. S103: Call the equivalent elastic energy storage and the external work done, subtract the external work from the equivalent elastic energy storage and perform a differential operation to extract the energy difference value, extract the absolute value of the energy difference value as a feature term, perform format recombination on the feature term to obtain the energy dissipation feature difference.

[0008] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Call the energy dissipation feature difference and extract the corresponding underlying sensor plane spatial coordinates, convert the underlying sensor plane spatial coordinates into matrix row and column index parameters, and write the energy dissipation feature difference into the array coordinate nodes in sequence according to the spatial coordinate mapping relationship to establish a two-dimensional plane data table. S202: Call the two-dimensional plane data table, extract the energy dissipation feature difference for adjacent elements in the two-dimensional plane data table, introduce the node pore water pressure equivalent energy, structural damage reduction energy and environmental thermal radiation energy, and calculate and extract the difference elements; S203: Call the difference element to obtain the preset risk warning benchmark threshold, perform a numerical comparison operation between the difference element and the preset risk warning benchmark threshold, identify redundant elements whose values ​​are lower than the preset risk warning benchmark threshold, remove the redundant elements from the data array, extract the retained nodes, and generate a planar energy dissipation deviation.

[0009] As a further aspect of the present invention, the operation to extract the difference elements specifically involves: ; in, Represents the elements of difference. This represents the difference in energy dissipation characteristics at the first node. This represents the difference in energy dissipation characteristics between adjacent nodes. Represents the pore water pressure equivalent energy at the node. Represents structural damage reduction energy. It represents environmental thermal radiation energy.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Extract the water level elevation variable in the mechanical coupling equation based on the plane energy dissipation deviation, perform partial derivative calculation on the water level elevation variable to obtain the partial derivative feature matrix, extract the node scalar data of the partial derivative feature matrix, and generate sensitive features; S302: Obtain the static safety limit coordinates and the multi-source fusion coordinates, subtract the multi-source fusion coordinates from the static safety limit coordinates, perform difference operation to extract the spatial dimension difference term, calculate the absolute value of the spatial dimension difference term and reorganize it into a one-dimensional array to generate an absolute deviation vector; S303: Call the absolute deviation vector and the sensitive feature, perform a Hadamard product operation on the absolute deviation vector and the sensitive feature according to the corresponding positions of the elements to extract the weighted ranging components, perform dimensionality reduction and aggregation transformation on the weighted ranging components, and establish a multi-source sensitive weighted ranging.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Call the multi-source sensitive weighted ranging, extract the internal elements of the multi-source sensitive weighted ranging, perform numerical squaring operation on each internal element to obtain the corresponding element squared item, perform cumulative summation operation on all the element squared items to generate a cumulative sum; S402: Call the cumulative sum, input the cumulative sum to the square root operation logic unit, control the square root operation logic unit to perform square root numerical operation on the cumulative sum to obtain distance feature items, perform feature extraction on the distance feature items, and generate penalty distance span; S403: Call the penalty distance span, obtain the preset failure limit distance threshold, compare the penalty distance span with the failure limit distance threshold, and trigger the corresponding over-limit instruction when the penalty distance span is greater than the failure limit distance threshold to obtain the cross-domain attenuation warning boundary.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Call the cross-domain attenuation warning boundary marker, access the on-site remote control terminal device, retrieve the pre-stored risk triggering rules in the on-site remote control terminal device, map and combine the cross-domain attenuation warning boundary marker with the pre-stored risk triggering rules, and generate a rule response sequence. S502: Invoke the rule response sequence, parse the threshold interval within the rule response sequence, perform numerical range matching between the cross-domain attenuation warning boundary and the threshold interval, extract the corresponding level identifier based on the interval matching status, and obtain the danger level boundary; S503: Call the hazard level boundary, extract the audible and visual alarm activation level signal corresponding to the hazard level boundary, send the audible and visual alarm activation level signal to the on-site remote control terminal equipment to trigger hardware alarm action, integrate the sending records, and obtain the foundation pit construction risk assessment result.

[0013] A foundation pit construction risk assessment system based on multi-parameter correlation analysis includes: The energy dissipation feature extraction module is used to achieve S1: collecting the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component, calculating the equivalent elastic energy storage, extracting the horizontal displacement data of the inclinometer and performing an integral operation with the soil resistance to extract the external work, and differentially calculating the equivalent elastic energy storage and the external work to obtain the energy dissipation feature difference. The planar energy deviation generation module is used to implement S2: extract the energy dissipation feature difference and the corresponding underlying sensor planar spatial coordinates, write them into a two-dimensional planar data table according to the spatial coordinate mapping relationship, extract the differences between adjacent elements and compare them with the preset risk warning benchmark threshold, remove elements below the threshold, and generate planar energy dissipation deviation. The multi-source sensitive weighted ranging module is used to implement S3: extract the water level elevation variable from the plane energy dissipation deviation, solve the partial derivative, extract sensitive features, obtain the absolute deviation vector between the static safety limit coordinates and the multi-source fusion coordinates, perform Hadamard operation, and establish multi-source sensitive weighted ranging; The penalty distance and boundary trigger module is used to implement S4: squaring the multi-source sensitive weighted ranging elements and accumulating them, taking the square root to obtain the penalty distance span, comparing it with the failure limit distance threshold, and triggering an instruction when the penalty distance span is greater than the threshold to obtain the cross-domain attenuation warning boundary; The risk assessment and alarm output module is used to implement S5: based on the cross-domain attenuation warning boundary marker, retrieve the risk triggering rules, determine the danger level boundary, trigger the start level signal of the audible and visual alarm, and output the risk assessment result of the foundation pit construction.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the difference between the equivalent elastic energy storage of the supporting components and the external work done is comprehensively extracted to quantify the energy dissipation characteristic gap. A two-dimensional mapping structure system is established by integrating the planar spatial coordinates of multi-source bottom-level sensors. Low-risk fluctuation elements are accurately eliminated to generate planar energy dissipation deviation. Multi-dimensional sensitive features are deeply mined by performing partial derivative calculation on the water level elevation variable. A multi-source sensitive weighted ranging structure is established by combining the absolute deviation vector of the static safety limit. The cumulative penalty distance span is dynamically calculated and the boundaries of complex danger levels are accurately defined. This overcomes the drawbacks of the delayed judgment caused by the isolated static comparison mode and realizes comprehensive monitoring and early warning of the potential cumulative damage effect under the heterogeneous collaborative interaction and evolutionary operation state around the foundation pit. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] Please see Figure 1 This invention provides a method for determining the construction risk of foundation pits based on multi-parameter correlation analysis, comprising the following steps: S1: Collect the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component on site. Square the stress deformation data and divide it by the elastic modulus data of the supporting component to extract the equivalent elastic energy storage. Extract the horizontal displacement data of the inclinometer probe and perform multiplication and integration operations with the soil resistance to extract the external work. Subtract the external work from the equivalent elastic energy storage and perform a differential operation to obtain the energy dissipation characteristic difference. S2: Call the energy dissipation feature difference and extract the corresponding underlying sensor plane spatial coordinates. Write the energy dissipation feature difference into the two-dimensional plane data table according to the spatial coordinate mapping relationship. Extract the energy dissipation feature difference for adjacent corresponding elements in the two-dimensional plane data table, perform a subtraction operation to extract the difference element, compare the difference element with the preset risk warning benchmark threshold and remove elements below the preset risk warning benchmark threshold to generate plane energy dissipation deviation. S3: Extract the water level elevation variable in the mechanical coupling equation based on the plane energy dissipation deviation, perform partial derivative calculation to extract sensitive features, obtain the static safety limit coordinates and multi-source fusion coordinates, perform difference calculation to extract the absolute deviation vector, and perform Hadamard product operation on the absolute deviation vector and sensitive features to establish multi-source sensitive weighted ranging. S4: Perform squaring operations on the internal elements of the multi-source sensitive weighted ranging and accumulate them to extract the cumulative sum. Perform square root operation on the cumulative sum to extract the penalty distance span. Compare the penalty distance span with the failure limit distance threshold. When the penalty distance span is greater than the failure limit distance threshold, trigger the corresponding instruction operation to obtain the cross-domain attenuation warning boundary. S5: Based on the cross-domain attenuation early warning boundary marker, retrieve the pre-stored risk triggering rules in the on-site remote control terminal equipment, determine the danger level boundary according to the pre-stored risk triggering rules, and trigger the corresponding audible and visual alarm activation level signal on the on-site remote control terminal equipment to send an action, and output the foundation pit construction risk assessment result.

[0020] The differences in energy dissipation characteristics include the hysteresis dissipation value of the support system, the plastic deformation potential of the retaining structure, and the difference in irreversible work done by the soil. Planar energy dissipation deviations include the extreme values ​​of local grid residuals, the distortion gradient of adjacent measuring points, and the regional stress imbalance index. Multi-source sensitive weighted ranging includes the water level fluctuation sensitive offset, the mechanical coupling deformation vector, and the spatial instability span coordinates. Cross-domain attenuation early warning markers include the critical contour of state degradation, the ultimate bearing failure red line, and the safety redundancy depletion bottom marker. The results of foundation pit construction risk assessment include the disaster evolution early warning level, the emergency control intervention plan, and the sound and light linkage execution strategy.

[0021] Please see Figure 2 The specific steps of S1 are as follows: S101: Collect the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component on site. Perform a square operation on the stress deformation data of the vibrating wire stress gauge on site to obtain the square value of deformation. Divide the square value of deformation by the elastic modulus data of the supporting component to generate the equivalent elastic energy storage. Twelve vibrating wire stress gauges, connected via signal transmission cables using a serial communication protocol, are positioned inside the reinforcement cage of the underground continuous wall in the foundation pit. Vibration frequency signals ranging from 400Hz to 3000Hz are collected as raw physical parameters. The received continuous time-series signal is converted from analog to digital at a sampling frequency of 1000Hz, transforming the analog signal into a discrete digital sequence. A Kalman filter algorithm is applied to the discrete digital sequence to eliminate high-frequency vibration noise caused by environmental construction machinery. State variables and covariance matrices are initialized, with the diagonal elements of the process noise covariance matrix set to 0.01 and the measurement noise covariance matrix to 0.5, acquiring the current observation input. The Kalman gain parameter is calculated, the difference between the current observation value and the previous state prediction value is multiplied by the Kalman gain, and accumulated to the predicted state, completing the update of the optimal state estimate and outputting the denoised clean frequency data. Based on the factory calibration coefficients of the vibrating wire stress gauges, the following formula is used... The micro-strain values ​​were calculated and recorded as the stress-deformation data from the vibrating wire stress gauge. The initial elastic modulus test data of the C35 grade diaphragm wall concrete at the current elevation and corresponding depth were extracted from the foundation pit construction engineering survey report and recorded as 31500. During the continuous loading stage of construction, the ultrasonic wave velocity test results of the concrete were extracted and used with the formula... (Where 2450 is the measured density of concrete), calculate and obtain the elastic modulus data of the dynamic support component under the current load-bearing state, and update the value to 29800.

[0022] Obtain the micro-strain values ​​within the computation buffer unit and perform the calculation. This value is determined to be the square of the deformation. The updated elastic modulus data of the support component, 29800, is retrieved, and a high-precision floating-point division operation is performed. Obtain the quotient. Perform dimensional transformation and scaling, through calculation. This generates equivalent elastic energy storage. Experimental data shows that when the sampling frequency is set to 1000Hz and the noise variance of the filtering process is set to 0.01, the stress variation coefficient is reduced to 1.2%, which improves the data smoothness by 23.5% compared with the traditional moving average filtering method, effectively suppressing the sudden data jump interference caused by the operation of heavy crawler cranes in the foundation pit.

[0023] Table 1. Stress and Energy Storage Parameters of Foundation Pit at Different Depths ; As shown in Table 1, by calculating the parameters at different depth levels, the changes in the energy storage level caused by the deformation under tension and compression of the deep structure are reflected.

[0024] S102: Obtain the horizontal displacement data of the inclinometer probe and the soil resistance, perform a multiplication operation on the horizontal displacement data of the inclinometer probe and the soil resistance to obtain the local work, and perform a definite integral operation on the local work along the depth direction to generate the external work; The inclinometer probe is guided into a 40m deep inclinometer tube pre-embedded outside the foundation pit via guide wheels. The probe is then raised from bottom to top in fixed 0.5m increments. At each measurement node, the probe pauses for 3 seconds, and the voltage signal from the tilt sensor is collected using an internal microelectromechanical biaxial accelerometer. The formula is then used to... The voltage signal is converted into a gravitational acceleration component, and the tilt angle of the probe at various depths is further calculated using an arcsine function. The formula is then applied. The cumulative horizontal offset at each depth was calculated segment by segment from the fixed bottom towards the borehole opening, and recorded as the horizontal displacement data of the inclinometer probe. At a depth of 20m, the horizontal displacement data was recorded as 18.6. Using soil geological borehole sampling data and triaxial shear test reports, the internal friction angle (15°) and cohesion (24°) of the soil at the current depth were extracted. Applying Rankine's earth pressure theory, the active earth pressure coefficient was calculated to be 0.588. The self-weight stress obtained by multiplying the total unit weight of the overlying soil layer at this depth by the depth was extracted. Combining this with the active earth pressure coefficient and the cohesion reduction term, the horizontal soil resistance value at this depth was calculated to be 145.2.

[0025] For the node at a depth of 20m, the horizontal displacement data of 18.6 and the soil resistance value of 145.2 were extracted, and a floating-point multiplication operation was performed. The product result is determined as the mechanical work consumed by the interaction between soil deformation and resistance within that local area, defined as the local work quantity. The set of local work quantities corresponding to all depth nodes is obtained, establishing discrete data pairs between depth coordinates and local work quantities. Simpson's numerical integration method is applied to perform definite integral operations on the above discrete data pairs along a depth interval of 40m, using the formula... This method generates external work that characterizes the energy exerted on the surrounding soil layers by the overall deformation of the foundation pit. Displacements in the range of 0 to 15 are classified as slight deformations; displacements in the range of 15 to 30 are classified as significant deformations; and displacements above 30 are classified as extremely strong deformations. Experimental data show that when the integration step size is set to 0.5, the integration calculation residual is stable within 0.05%, which is 14.2% more accurate than the trapezoidal rule integration method.

[0026] S103: Call the equivalent elastic energy storage and external work, subtract the external work from the equivalent elastic energy storage and perform a differential operation to extract the energy difference value, extract the absolute value of the energy difference value as a feature term, perform format reorganization on the feature term to obtain the energy dissipation feature difference; Extract the equivalent elastic energy storage value of 781.4 and the external work value of 84520.5 from the buffer. Perform a unit area conversion operation on the equivalent elastic energy storage and calculate... To obtain localized equivalent elastic energy storage, a similar area weighting coefficient is used to scale the external work performed, extracting the converted value of 156.40 for the external work within the same depth influence range. This triggers the execution of a differential subtraction instruction. This value represents the net energy surplus state between the release of internal elastic energy and resistance to external earth pressure, and is defined as the energy difference. The sign bit flag within the arithmetic logic unit is retrieved, and the absolute value of the energy difference is extracted. This is then encapsulated as a feature item of the spatial node.

[0027] For feature sequence containing multiple acquisition depths, a contiguous static memory space is allocated. The unique sensor hardware identifier, timestamp, and 3D coordinate variables carried by each feature are read. These are sorted in ascending order by timestamp, with the Z-axis depth value of the 3D coordinates as a secondary sorting criterion. The originally one-dimensional randomly distributed feature items are converted into a structure array arranged in a matrix format according to depth and time, and then reformatted. Each structure occupies a fixed 32 bytes: the first 8 bytes store the timestamp, the middle 12 bytes store the 3D coordinates, and the last 12 bytes store the high-precision absolute value of the feature item. The core variables representing energy loss differences are extracted from the reformatted structure array to obtain the energy dissipation feature gap. Experimental data shows that when using a 32-byte fixed-length structure for reformatting and setting the control area coefficient to 0.25, the data retrieval latency is reduced to 2 milliseconds, and the processing efficiency is improved by 41.8%.

[0028] Please see Figure 3 The specific steps of S2 are as follows: S201: Call the energy dissipation feature difference and extract the corresponding underlying sensor plane spatial coordinates, convert the underlying sensor plane spatial coordinates into matrix row and column index parameters, write the energy dissipation feature difference into the array coordinate nodes in sequence according to the spatial coordinate mapping relationship, and establish a two-dimensional plane data table. The restructured structure array is read, and specific data fields are parsed to obtain the energy dissipation characteristic difference of 38.95. The pointer offset is used to locate the 3D coordinate variable memory segment, and the Z-axis depth value is extracted for planar dimensionality reduction. The X-axis coordinate value of 45.5 and the Y-axis coordinate value of 20.0 are extracted to form the bottom-layer sensor planar spatial coordinates representing the horizontal projection position of the current measuring point. An orthogonal reference grid for the foundation pit area is established, with the origin coordinate set to 0.0, and the basic resolution of the grid cells in the X direction and the Y direction set to 0.5. Through calculation... Get the integer index of the X-axis; by calculation Obtain the integer index of the Y-axis. Convert the above integer coordinate pairs into matrix row and column index parameters suitable for memory addressing of two-dimensional arrays.

[0029] Allocate a 2D double-precision floating-point array of size 200 rows by 300 columns in random access memory, and assign all initial elements the placeholder value -999.0. Calculate the absolute memory offset address using row number 40 and column number 91 according to the mapping relationship, using the formula: Jump to the offset address, overwrite the initial placeholder value with the energy dissipation characteristic difference of 38.95, and write it sequentially to the specified array coordinate nodes. Repeat the above coordinate transformation and address writing instructions for all sensor nodes. After traversal, scan the two-dimensional array space; for blank nodes with placeholders, apply the inverse distance weight interpolation algorithm. By filling in the blank data, a two-dimensional planar data table with full spatial coverage was established. Experimental data shows that when the grid resolution is set to 0.5 and the interpolation search radius is set to 5, the data table reconstruction error rate is only 3.4%.

[0030] S202: Call the two-dimensional planar data table, extract the energy dissipation characteristic differences between adjacent elements within the table, and introduce the nodal pore water pressure equivalent energy, structural damage reduction energy, and environmental thermal radiation energy, using the following formula: ; The operation extracts the difference elements; where, Represents the elements of difference. This represents the difference in energy dissipation characteristics at the first node. This represents the difference in energy dissipation characteristics between adjacent nodes. Represents the pore water pressure equivalent energy at the node. Represents structural damage reduction energy. Represents environmental thermal radiation energy; Read the constructed two-dimensional planar data table. Initialize the row and column pointers, and set the analysis window to a 3x3 grid. Place the center anchor point at row 40 and column 91, and extract its energy dissipation characteristic difference value. As the first node, based on the four-neighbor rule, locate the 39th row and 91st column of the adjacent node above, and extract the energy dissipation characteristic difference value. As an adjacent node, the pore water pressure value of 115.5 collected by the field piezometer is extracted and calculated. The equivalent energy of pore water pressure at the nodes was calculated. By calling up the historical displacement monitoring curves of the foundation pit support structure and substituting the ratio 0.12, the structural damage reduction energy reflecting the degree of material deterioration is calculated. The ambient temperature difference of 12.5°C measured by the infrared temperature sensor is obtained. Calculate... Quantitative acquisition of environmental thermal radiation energy .

[0031] Substitute the obtained parameters into the formula to calculate and extract the difference elements: ; The advantage of this formula lies in its ability to vector-add the absolute energy difference between nodes to a corrected energy term generated by coupling multiple physical fields (water pressure, structural damage, and thermal radiation). This addresses the significant error of a single mechanical energy index in environments with abundant groundwater and drastic temperature variations. Experimental data demonstrate that under complex geological conditions with a water pressure gradient of 20 kPa / m, the formula achieves a risk area identification accuracy of 92.4%, representing an 18.7% improvement compared to the traditional difference method without multi-field coupling correction. This numerical result visually quantifies the abrupt change in local stress concentration, providing a benchmark for subsequent redundant data removal.

[0032] Table 2 Multidimensional Energy Parameters of Grid Nodes ; As shown in Table 2, the combined input of multi-physics energy parameters provides a complete multi-source data input benchmark for the calculation of differential elements.

[0033] S203: Call the difference element, obtain the preset risk warning benchmark threshold, perform a numerical comparison operation between the difference element and the preset risk warning benchmark threshold, identify redundant elements whose values ​​are lower than the preset risk warning benchmark threshold, remove the redundant elements from the data array, extract the retained nodes, and generate planar energy dissipation deviation. Extract the calculated set of difference elements column by column from the memory stack. Call the configuration file to parse the risk warning baseline threshold set for the silty clay layer foundation pit. Extract energy difference distribution samples from 100 historical safe construction cycles, and calculate the expected value (12.4) and standard deviation (3.5) of this normally distributed dataset. Apply statistical formulas. The upper limit value of the abnormal control is obtained. This upper limit is then added to the safety redundancy margin of 2.1 set due to the impact of heavy rainfall, and the calculation is performed. The preset risk warning baseline threshold of 25.0 is established. A high-concurrency numerical comparator is started, sending the currently extracted difference element of 40.16 to input port A and the preset risk warning baseline threshold of 25.0 to input port B. A floating-point numerical comparison operation is then performed.

[0034] When the comparator determines the logic Upon establishment, the node's coordinates and values ​​are pushed into the high-risk feature retention queue. When the difference element of another node is read to be 18.4, the decision logic is... The algorithm identifies low-value elements as background noise representing a safe region with a smooth energy transition and marks them as redundant. A memory pointer redirection operation is performed to sever the redundant elements from their original linked lists, releasing their occupied storage space and removing them from the data array. After traversal, all elements in the high-risk feature retention queue are summarized and remapped to a blank canvas matrix with the same scale as the pit plane. Two-dimensional spatial smoothing filtering based on a Gaussian kernel function is applied to the remaining non-zero nodes to generate a planar energy dissipation bias. Experimental data shows that when the baseline threshold is set to 25.0, the data volume of the data array is compressed by 78.5%.

[0035] Please see Figure 4 The specific steps of S3 are as follows: S301: Extract the water level elevation variable in the mechanical coupling equation based on the plane energy dissipation deviation, perform partial derivative calculation on the water level elevation variable to obtain the partial derivative feature matrix, extract the node scalar data of the partial derivative feature matrix, and generate sensitive features; From the region of high-valued planar energy dissipation deviation, the three-dimensional coordinates and strain parameters carried by the nodes are read. A pre-constructed three-dimensional consolidation mechanics coupling equation based on Biot's consolidation theory is loaded. The real-time water level elevation variable reflecting the current location of the groundwater seepage free surface is extracted, and its value at 45.5 on the X-axis and 20.0 on the Y-axis is recorded as -12.3. A Taylor expansion is introduced to linearize the stress tensor expression in the mechanics coupling equation.

[0036] For this function mapping, the water level elevation variable of -12.3 is used as the independent variable, with a small step size of 0.001. The partial derivatives are calculated using the central difference method. The micro-sensitivity gradient value of the spatial coordinate point with respect to water level changes is obtained. The above difference calculation is iterated within a two-dimensional planar grid to construct a partial derivative feature matrix. The specific value 3.8 is extracted as the node scalar data using a double-loop index. These scalar data are normalized according to their spatial weights in the matrix, and the calculation is performed. This method obtains dimensionless relative change rates, generating sensitive features characterizing the degree to which mechanics is affected by groundwater fluctuations. Experimental data show that when the difference step size is set to 0.001, the calculation truncation error of the partial derivative characteristic matrix is ​​limited to the order of 10 to the power of -5.

[0037] S302: Obtain the static safety limit coordinates and multi-source fusion coordinates, subtract the multi-source fusion coordinates from the static safety limit coordinates, perform difference operation to extract the spatial dimension difference term, calculate the absolute value of the spatial dimension difference term and reorganize it into a one-dimensional array to generate the absolute deviation vector; Access the finite element simulation database of the foundation pit design drawings and extract the set of node coordinates calculated using the strength reduction method up to the point of non-convergence when the model becomes unstable. Select the spatial locations on the failure sliding surface where the maximum principal strain exceeds 0.05, and record the coordinates of one key node as X-axis 46.0, Y-axis 21.5, Z-axis -14.0, establishing it as the static safety limit coordinates. Activate the on-site BeiDou satellite navigation and total station linkage positioning service to obtain the actual deformation monitoring location of this key node at the current moment, recording the coordinates as X-axis 45.8, Y-axis 20.9, Z-axis -13.5, establishing it as the multi-source fused coordinates.

[0038] The static safety limit coordinates and multi-source fused coordinate components are loaded into the subtractor, and the difference operation is performed to obtain the spatial dimension difference term. The absolute value instruction set is called to calculate... , , Allocate a continuous single-precision floating-point array of capacity 3, and push the three absolute values ​​sequentially into the memory address, reassembling them into a one-dimensional array. Extract the memory starting address of this one-dimensional array and generate an absolute deviation vector representing the spatial span between the current spatial state and the absolute unstable state. Experimental data show that when the total station weight is set to 0.7 and the satellite weight is set to 0.3, the root mean square error of positioning is reduced to 2 mm, and the stability is improved by 56.4%.

[0039] S303: Call the absolute deviation vector and sensitive features, perform Hadamard product operation on the corresponding positions of the elements of the absolute deviation vector and sensitive features to extract the weighted ranging components, perform dimensionality reduction and aggregation transformation on the weighted ranging components, and establish multi-source sensitive weighted ranging; Read the one-dimensional absolute deviation vector generated in the previous stage, whose internal elements are 0.2, 0.6, and 0.5 respectively. Call the sensitive feature set for the corresponding node in three directions, obtained through normalization processing, and extract the values ​​to 0.73, 0.85, and 0.42. Start the tensor operation core of the digital signal processor, align and load the absolute deviation vector with the sensitive features, and perform the Hadamard product operation: , , The set of these three product values ​​is extracted as a weighted ranging component.

[0040] For this one-dimensional array structure containing three components, a dimensionality reduction and aggregation transformation is performed. The weighting parameters for the hazard impact factors in each direction are set as follows: 0.2 for the X direction, 0.5 for the Y direction, and 0.3 for the Z direction. Each weighted ranging component is multiplied by the corresponding directional impact factor and then summed. The single-dimensional numerical output representing the imminent severity of comprehensive space hazards was used to establish a multi-source sensitive weighted ranging system. Experimental data show that, with a high weighting tendency of the Y-direction influence factor set at 0.5, this weighted ranging system achieved a recall rate of 95.5% for early warning of sudden changes in support axial force.

[0041] Please see Figure 5 The specific steps of S4 are as follows: S401: Call the multi-source sensitive weighted ranging, extract the internal elements of the multi-source sensitive weighted ranging, perform numerical squaring operation on each internal element to obtain the corresponding element squared item, perform cumulative summation operation on all element squared items to generate cumulative sum; The calculated composite scalar value of 0.3472 is extracted from the register bus and used as the input to the calculation module for calling the multi-source sensitive weighted ranging. A pointer is used to traverse and obtain the set of multi-source sensitive weighted ranging values ​​from all sensor nodes within the same monitoring time window. The internal elements of the multi-source sensitive weighted ranging are extracted sequentially, such as the first node element 0.3472, the second node element 0.4125, the third node element 0.2568, etc., for a total of 100 valid sample points.

[0042] Initialize a double-precision floating-point accumulator, and perform a numerical square operation on the first extracted internal element 0.3472. Obtain the square of the corresponding element. Repeat this process to calculate. as well as As the loop instructions continue to execute, newly calculated squared terms are continuously pushed into the input of the adder. This performs an accumulation and summation operation. The Kahan summation algorithm is incorporated into this process to compensate for truncation errors. After 100 iterations, the accumulator outputs a total value of 15.4265, generating the cumulative sum. Experimental data shows that when using the Kahan summation algorithm to process 100 floating-point data points, the accumulation error is controlled to less than one ten-thousandth, and the precision retention rate is improved by 100%.

[0043] S402: Call the cumulative sum, input the cumulative sum to the square root operation logic unit, control the square root operation logic unit to perform square root operation on the cumulative sum to obtain the distance feature item, perform feature extraction on the distance feature item, and generate the penalty distance span; The cumulative sum of 15.4265 stored in main memory is retrieved. The status word of the floating-point unit inside the central processing unit is configured to prepare for the execution of a non-linear square root instruction. The cumulative sum is input as an operand to the hardware-level square root operation logic unit. This unit is controlled to use the Newton-Raphson iterative algorithm to select an initial guess value. Perform the first round of iterative calculations. ,Right now The square root numerical calculation is performed continuously until the absolute value of the difference between the guesses of two adjacent iterations is less than the preset precision. After 5 iterations, the final converged result is extracted. This value is obtained as the distance feature term.

[0044] Feature extraction is performed on the acquired distance features. A penalty coefficient is introduced for safety correction. The environmental sensitivity assessment report surrounding the project is retrieved; the environmental risk rating is determined to be extremely high, and the corresponding nonlinear amplification parameter of 1.35 is extracted. Multiplication operations are then performed. The corrected value was calculated. This spatial metric value, which amplifies the risk of extremely high-risk scenarios, was then encapsulated as a penalty distance span. Experimental data shows that after introducing a penalty coefficient of 1.35, the advance warning time was advanced by 45 minutes.

[0045] S403: Call the penalty distance span, obtain the preset failure limit distance threshold, compare the penalty distance span with the failure limit distance threshold, and trigger the corresponding over-limit instruction when the penalty distance span is greater than the failure limit distance threshold, and obtain the cross-domain attenuation warning boundary. Extract the penalty distance span of 5.3022 with amplified environmental risk from the buffer. Initiate an access request to the read-only memory to obtain the preset limit constraint parameters. Set the constraint value of 4.5 as the failure limit distance threshold when the overall comprehensive displacement of the foundation pit support structure exceeds 4.5, and extract this constraint value as the failure limit distance threshold. Activate the hardware interrupt function of the comparator, load the penalty distance span of 5.3022 to the non-inverting input, load the failure limit distance threshold of 4.5 to the inverting input, and perform a value comparison.

[0046] After comparison using a differential amplifier circuit, the logic conditions were confirmed. Established. When the penalty distance exceeds the failure limit distance threshold, the control logic unit triggers the corresponding over-limit instruction. The over-limit difference is calculated. A hexadecimal alarm data packet containing the specified value is generated. This alarm data packet is pushed to the front of the priority processing queue, and the warning boundary definition table is extracted. Based on the 0.5 to 1.0 range to which the over-limit difference value of 0.8022 belongs, the corresponding boundary identification code is assigned, and the cross-domain attenuation warning boundary in the format A-Critical-08 is obtained. Experimental data shows that under the set failure limit distance threshold of 4.5, the abnormal response delay is reduced from 2 hours to less than 50 milliseconds.

[0047] Please see Figure 6 The specific steps of S5 are as follows: S501: Call the cross-domain attenuation warning boundary marker, access the on-site remote control terminal device, retrieve the pre-stored risk triggering rules in the on-site remote control terminal device, map and combine the cross-domain attenuation warning boundary marker with the pre-stored risk triggering rules, and generate a rule response sequence. Invoke the cross-domain attenuation warning boundary marker A-Critical-08, which carries critical information. Establish a network Socket connection based on the TCP / IP protocol suite and send a handshake request to the on-site remote control terminal device. After a successful three-way handshake, establish a stable encrypted data channel and access the local database of the on-site remote control terminal device. Issue an SQL query command, using the boundary marker identification code within the cross-domain attenuation warning boundary marker as the index primary key, to retrieve the pre-stored risk triggering rules stored in the flash memory of the on-site remote control terminal device. Extract the rule body file containing the emergency response plan, personnel evacuation instruction set, and equipment shutdown sequence.

[0048] The cross-domain attenuation warning boundary marker A-Critical-08 is combined with the extracted pre-stored risk triggering rules through field mapping. The location parameter of A-zone in the boundary marker is parsed to locate the regional power control node in the rule body file. The critical parameter in the boundary marker is parsed to extract the address of the highest-level all-staff broadcast evacuation recording in the rule body file. The over-limit parameter 08 is parsed to match the start delay time of the automatic grouting reinforcement machine to 0 seconds. All the parsed and interrelated specific execution instructions are assembled into a continuous XML file, arranged in logical order as follows: Step 1: Cut off the construction power to A-zone; Step 2: Start the all-staff broadcast; Step 3: Immediately start the grouting machine. A rule response sequence is generated from this XML file containing ordered execution tasks. Experimental data shows that when using XML format for mapping and combination and transmitting within the local area network, the data packet parsing time is only 15 milliseconds.

[0049] S502: Call the rule response sequence, parse the threshold range within the rule response sequence, perform numerical range matching between the cross-domain attenuation warning boundary and the threshold range, extract the corresponding level identifier based on the range matching status, and obtain the danger level boundary. Read the generated XML-formatted rule response sequence file from the cache. Load the file tree structure using the Document Object Model parser, traverse the attribute tags of each node, and parse the threshold ranges used to define the intensity of alarms within the rule response sequence. Extract the over-limit difference value of 0.8022 contained in the previous cross-domain attenuation warning boundary. Input this over-limit difference value of 0.8022 into the boundary judgment logic of each interval, and perform numerical range matching.

[0050] After comparison and judgment, it conforms to the logical expression. Based on the return result of this logical judgment, the interval matching status is locked as a warning interval. According to the preset dictionary mapping table, the level identifier number 3 corresponding to the warning interval is extracted. Combining this level identifier with the area attributes, the final danger level boundary, which includes both color and level attributes, is obtained and set as orange - Level 3 warning. Experimental data shows that after adopting a step-by-step matching algorithm with four threshold intervals, the false alarm rate was reduced by 89.5%.

[0051] Table 3 Risk Warning Response Level Table ; As shown in Table 3, by setting strictly corresponding stepped threshold ranges, a clear decision logic basis is provided for the level driving of subsequent audible and visual alarm devices.

[0052] S503: Call the danger level boundary, extract the activation level signal of the audible and visual alarm corresponding to the danger level boundary, send the activation level signal of the audible and visual alarm to the on-site remote control terminal equipment to trigger the hardware alarm action, integrate the sending records, and obtain the foundation pit construction risk assessment result; The generated hazard level boundary, i.e., orange - level three warning, is invoked. The device driver mapping table is queried to extract the corresponding audible and visual alarm activation level signal configuration parameters. The obtained parameters are: a continuous 5-volt high-level output from driver pin GPIO_14 illuminates the orange rotating warning light; a 2Hz square wave pulse signal with a 50% duty cycle output from driver pin GPIO_15 drives a high-pitched buzzer to emit an intermittent alarm sound. These audible and visual alarm activation level signals are packaged into a data frame based on the Modbus-RTU protocol. This data frame is then transmitted at a baud rate of 9600 to the programmable logic controller (PLC) within the field remote control terminal device via the RS485 industrial bus.

[0053] After parsing the data frame, the programmable logic controller (PLC) closes the corresponding solid-state relay, connecting the 220-volt AC power supply circuit of the high-power audible and visual alarm on site. This successfully triggers the hardware alarm action, causing the alarm light to rotate and flash, and the buzzer to emit a piercing long sound. The system simultaneously captures key elements such as the start time of the event, the extreme limit of 0.8022, the physical address of the alarm device, and the executed linkage rules. All these elements are integrated and converted into a structured record, which is then inserted into the alarm history form of the MySQL database on the site management server for persistent storage. Based on this complete record and the on-site audible and visual feedback, the risk assessment result for the entire foundation pit construction phase, involving high water pressure and uneven settlement, is obtained, confirming a true positive alarm event indicating that the local support system is approaching its limit state. Experimental data shows that when the level signal transmission baud rate is set to 9600 and the Modbus-RTU protocol is used to issue commands, the end-to-end delay from receiving the command to the relay closing is only 25 milliseconds, achieving zero-delay safety interception of foundation pit hazards.

[0054] Please see Figure 7 A foundation pit construction risk assessment system based on multi-parameter correlation analysis includes: The energy dissipation feature extraction module is used to achieve S1: collecting the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component, calculating the equivalent elastic energy storage, extracting the horizontal displacement data of the inclinometer and performing an integral operation with the soil resistance to extract the external work, and differentially calculating the equivalent elastic energy storage and the external work to obtain the energy dissipation feature difference. The planar energy deviation generation module is used to implement S2: extract the energy dissipation feature difference and the corresponding underlying sensor planar spatial coordinates, write them into a two-dimensional planar data table according to the spatial coordinate mapping relationship, extract the differences between adjacent elements and compare them with the preset risk warning benchmark threshold, remove elements below the threshold, and generate planar energy dissipation deviation. The multi-source sensitive weighted ranging module is used to implement S3: extract the water level elevation variable from the plane energy dissipation deviation, solve the partial derivative, extract sensitive features, obtain the absolute deviation vector between the static safety limit coordinates and the multi-source fusion coordinates, perform Hadamard operation, and establish multi-source sensitive weighted ranging; The penalty distance and boundary triggering module is used to implement S4: square multi-source sensitive weighted ranging elements and accumulate them, take the square root to obtain the penalty distance span, compare it with the failure limit distance threshold, and trigger an instruction when the penalty distance span is greater than the threshold to obtain the cross-domain attenuation warning boundary; The risk assessment and alarm output module is used to implement S5: based on the cross-domain attenuation warning boundary marker, retrieve the risk triggering rules, determine the danger level boundary, trigger the start level signal of the audible and visual alarm, and output the risk assessment result of the foundation pit construction.

[0055] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A method for determining the construction risk of foundation pits based on multi-parameter correlation analysis, characterized in that, Includes the following steps: S1: Collect stress deformation data of vibrating wire stress gauge and elastic modulus data of supporting component, calculate equivalent elastic energy storage, extract horizontal displacement data of inclinometer and perform integral calculation with soil resistance to extract external work, and differentially calculate equivalent elastic energy storage and external work to obtain the energy dissipation characteristic difference. S2: Extract the energy dissipation feature difference and the corresponding underlying sensor plane spatial coordinates, write them into a two-dimensional plane data table according to the spatial coordinate mapping relationship, extract the differences between adjacent elements and compare them with the preset risk warning benchmark threshold, remove elements below the threshold, and generate a plane energy dissipation deviation. S3: Extract the water level elevation variable from the plane energy dissipation deviation, solve the partial derivative, extract the sensitive features, obtain the absolute deviation vector between the static safety limit coordinates and the multi-source fusion coordinates, perform Hadamard operation, and establish multi-source sensitive weighted ranging. S4: Squaring the multi-source sensitive weighted ranging elements and accumulating them, taking the square root to obtain the penalty distance span, comparing it with the failure limit distance threshold, and triggering an instruction when the penalty distance span is greater than the threshold to obtain the cross-domain attenuation warning boundary. S5: Based on the cross-domain attenuation early warning boundary marker, retrieve the risk triggering rules, determine the danger level boundary, trigger the audible and visual alarm activation level signal, and output the foundation pit construction risk assessment result.

2. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The energy dissipation characteristic differences include the hysteresis dissipation value of the support system, the plastic deformation potential of the retaining structure, and the difference in irreversible work done by the soil. The planar energy dissipation deviations include the extreme values ​​of local grid residuals, the distortion gradient of adjacent measuring points, and the regional stress imbalance index. The multi-source sensitive weighted ranging includes the water level fluctuation sensitive offset, the mechanical coupling deformation vector, and the spatial instability span coordinates. The cross-domain attenuation early warning markers include the critical contour of state degradation, the ultimate bearing failure red line, and the safety redundancy depletion bottom marker. The foundation pit construction risk assessment results include the disaster evolution early warning level, the emergency control intervention plan, and the sound and light linkage execution strategy.

3. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The aforementioned risk warning benchmark threshold refers to the reference warning threshold used to determine whether the difference in energy dissipation between adjacent spatial locations constitutes an anomaly, eliminating normal minor fluctuations and extracting risk characteristics.

4. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Collect the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component on site; perform a square operation on the stress deformation data of the vibrating wire stress gauge on site to obtain the square value of deformation; divide the square value of deformation by the elastic modulus data of the supporting component to generate equivalent elastic energy storage. S102: Obtain the horizontal displacement data of the inclinometer probe and the soil resistance, perform a multiplication operation on the horizontal displacement data of the inclinometer probe and the soil resistance to obtain the local work, and perform a definite integral operation on the local work along the depth direction to generate the external work. S103: Call the equivalent elastic energy storage and the external work done, subtract the external work from the equivalent elastic energy storage and perform a differential operation to extract the energy difference value, extract the absolute value of the energy difference value as a feature term, perform format recombination on the feature term to obtain the energy dissipation feature difference.

5. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Call the energy dissipation feature difference and extract the corresponding underlying sensor plane spatial coordinates, convert the underlying sensor plane spatial coordinates into matrix row and column index parameters, and write the energy dissipation feature difference into the array coordinate nodes in sequence according to the spatial coordinate mapping relationship to establish a two-dimensional plane data table. S202: Call the two-dimensional plane data table, extract the energy dissipation feature difference for adjacent elements in the two-dimensional plane data table, introduce the node pore water pressure equivalent energy, structural damage reduction energy and environmental thermal radiation energy, and calculate and extract the difference elements; S203: Call the difference element to obtain the preset risk warning benchmark threshold, perform a numerical comparison operation between the difference element and the preset risk warning benchmark threshold, identify redundant elements whose values ​​are lower than the preset risk warning benchmark threshold, remove the redundant elements from the data array, extract the retained nodes, and generate a planar energy dissipation deviation.

6. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 5, characterized in that, The specific operation for extracting the difference elements is as follows: ； in, Represents the elements of difference. This represents the difference in energy dissipation characteristics at the first node. This represents the difference in energy dissipation characteristics between adjacent nodes. Represents the pore water pressure equivalent energy at the node. Represents structural damage reduction energy. It represents environmental thermal radiation energy.

7. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Extract the water level elevation variable in the mechanical coupling equation based on the plane energy dissipation deviation, perform partial derivative calculation on the water level elevation variable to obtain the partial derivative feature matrix, extract the node scalar data of the partial derivative feature matrix, and generate sensitive features; S302: Obtain the static safety limit coordinates and the multi-source fusion coordinates, subtract the multi-source fusion coordinates from the static safety limit coordinates, perform difference operation to extract the spatial dimension difference term, calculate the absolute value of the spatial dimension difference term and reorganize it into a one-dimensional array to generate an absolute deviation vector; S303: Call the absolute deviation vector and the sensitive feature, perform a Hadamard product operation on the absolute deviation vector and the sensitive feature according to the corresponding positions of the elements to extract the weighted ranging components, perform dimensionality reduction and aggregation transformation on the weighted ranging components, and establish a multi-source sensitive weighted ranging.

8. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Call the multi-source sensitive weighted ranging, extract the internal elements of the multi-source sensitive weighted ranging, perform numerical squaring operation on each internal element to obtain the corresponding element squared item, perform cumulative summation operation on all the element squared items to generate a cumulative sum; S402: Call the cumulative sum, input the cumulative sum to the square root operation logic unit, control the square root operation logic unit to perform square root numerical operation on the cumulative sum to obtain distance feature items, perform feature extraction on the distance feature items, and generate penalty distance span; S403: Call the penalty distance span, obtain the preset failure limit distance threshold, compare the penalty distance span with the failure limit distance threshold, and trigger the corresponding over-limit instruction when the penalty distance span is greater than the failure limit distance threshold to obtain the cross-domain attenuation warning boundary.

9. The method for determining the construction risk of foundation pits based on multi-parameter correlation analysis according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Call the cross-domain attenuation warning boundary marker, access the on-site remote control terminal device, retrieve the pre-stored risk triggering rules in the on-site remote control terminal device, map and combine the cross-domain attenuation warning boundary marker with the pre-stored risk triggering rules, and generate a rule response sequence. S502: Invoke the rule response sequence, parse the threshold interval within the rule response sequence, perform numerical range matching between the cross-domain attenuation warning boundary and the threshold interval, extract the corresponding level identifier based on the interval matching status, and obtain the danger level boundary; S503: Call the hazard level boundary, extract the audible and visual alarm activation level signal corresponding to the hazard level boundary, send the audible and visual alarm activation level signal to the on-site remote control terminal equipment to trigger hardware alarm action, integrate the sending records, and obtain the foundation pit construction risk assessment result.

10. A foundation pit construction risk assessment system based on multi-parameter correlation analysis, characterized in that, The system is used to implement the method for determining the risk of foundation pit construction based on multi-parameter correlation analysis as described in any one of claims 1-9, including: The energy dissipation feature extraction module is used to achieve S1: collecting the stress deformation data of the vibrating wire stress gauge and the elastic modulus data of the supporting component, calculating the equivalent elastic energy storage, extracting the horizontal displacement data of the inclinometer and performing an integral operation with the soil resistance to extract the external work, and differentially calculating the equivalent elastic energy storage and the external work to obtain the energy dissipation feature difference. The planar energy deviation generation module is used to implement S2: extract the energy dissipation feature difference and the corresponding underlying sensor planar spatial coordinates, write them into a two-dimensional planar data table according to the spatial coordinate mapping relationship, extract the differences between adjacent elements and compare them with the preset risk warning benchmark threshold, remove elements below the threshold, and generate planar energy dissipation deviation. The multi-source sensitive weighted ranging module is used to implement S3: extract the water level elevation variable from the plane energy dissipation deviation, solve the partial derivative, extract sensitive features, obtain the absolute deviation vector between the static safety limit coordinates and the multi-source fusion coordinates, perform Hadamard operation, and establish multi-source sensitive weighted ranging; The penalty distance and boundary trigger module is used to implement S4: squaring the multi-source sensitive weighted ranging elements and accumulating them, taking the square root to obtain the penalty distance span, comparing it with the failure limit distance threshold, and triggering an instruction when the penalty distance span is greater than the threshold to obtain the cross-domain attenuation warning boundary; The risk assessment and alarm output module is used to implement S5: based on the cross-domain attenuation warning boundary marker, retrieve the risk triggering rules, determine the danger level boundary, trigger the start level signal of the audible and visual alarm, and output the risk assessment result of the foundation pit construction.

Images