AI-based project cost big data intelligent analysis system and method

Abstract

The invention belongs to the technical field of engineering cost, and relates to an AI-based engineering cost big data intelligent analysis system and method.A ground pile coupling model is constructed, rainfall prediction parameters in a construction period serve as time-varying boundary conditions to be input, and transient seepage changes of foundation soil under the rainfall effect are analyzed; the method comprises the following steps: quantifying pore water pressure increment distribution in a pile periphery influence domain of each support pile, predicting a pile body inclination probability, a build-in failure probability and an inter-pile soil loss probability of the support piles based on a reduction relation of pore water pressure increment relative to effective stress of a soil body, and triggering directional parameter adjustment according to a condition that the probability exceeds a threshold value, a support pile design parameter adjustment scheme and corresponding support structure increment cost are formulated, a high rainfall risk working condition is recognized in advance to output risk regulation support cost, the possibility of failure of the support structure is effectively reduced, the safety and stability of an engineering structure are enhanced, and then economic losses caused by potential accidents are reduced.

Description

Technical Field

[0001] The present invention belongs to the field of engineering cost technology, and specifically relates to an AI-based engineering cost big data intelligent analysis system and method. Background Art

[0002] With the acceleration of urbanization and the continued expansion of underground space development, deep foundation pit projects are becoming increasingly deep and complex. Support pile structures, with their strong resistance to lateral pressure and wide construction adaptability, have become the core support method for high-rise buildings, subway hubs, and underground pipeline corridors. However, due to factors such as complex geological conditions, diverse design options, and volatile material prices, cost control of support pile projects has become a key and difficult issue in project cost management.

[0003] However, the existing solutions for the cost of supporting pile projects still have significant limitations when addressing the above challenges, mainly manifested in the following aspects: the current methods mainly rely on design drawings and static geological survey reports, and do not fully consider the time-varying impact of dynamic precipitation on the permeability characteristics of the soil layer during the construction period. As a result, there is a lack of dynamic prediction of the probability of support pile failure caused by precipitation and a quantification mechanism for the additional repair costs caused by seepage damage. This technical gap can easily lead to serious cost overruns of supporting pile projects when precipitation control is inadequate. Summary of the Invention

[0004] In view of this, in order to solve the problems raised in the above background technology, an AI-based engineering cost big data intelligent analysis system and method are proposed.

[0005] The technical solution adopted by the present invention to solve its technical problems is: First, the present invention provides an AI-based engineering cost big data intelligent analysis system, including: a model construction module, a penetration deduction module, a failure probability prediction module, a parameter optimization module, a control cost output module and an original cost output module.

[0006] The model construction module is connected to the penetration deduction module, the penetration deduction module is connected to the failure probability prediction module, the failure probability prediction module is respectively connected to the parameter optimization module and the original cost output module, and the parameter optimization module is connected to the control cost output module.

[0007] The model building module constructs a ground-pile coupling model based on the geological survey parameters of the foundation engineering and the support pile design parameters.

[0008] The infiltration simulation module inputs the predicted precipitation parameters during the construction period as surface boundary conditions into the model, analyzes the transient seepage changes in the foundation soil under the action of precipitation, and quantifies the distribution of pore water pressure increments within the influence area around each support pile.

[0009] The failure probability prediction module predicts the failure probability distribution of each supporting pile based on the reduction relationship between the pore water pressure increment and the effective stress of the soil. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles.

[0010] The parameter optimization module formulates a support pile design parameter adjustment plan when the probability of any sub-item in the failure probability distribution of any support pile exceeds a preset warning threshold. The adjustment plan includes one or more combinations of increasing the pile diameter, deepening the embedment depth, or reducing the pile spacing.

[0011] The cost control output module calculates the incremental cost of the support structure generated by implementing the adjustment plan, integrates it with the original support cost of the foundation project, and outputs the risk-controlled support cost.

[0012] The original cost output module outputs the original support cost of the foundation project when the probability of all sub-items in the failure probability distribution of all support piles does not exceed the threshold.

[0013] In the second aspect, the present invention provides an AI-based intelligent analysis method for engineering cost big data, including: constructing a ground pile coupling model based on geological survey parameters of foundation engineering and support pile design parameters.

[0014] The precipitation prediction parameters during the construction period are input into the model as surface boundary conditions to analyze the transient seepage changes of the foundation soil under the action of precipitation and quantify the distribution of the pore water pressure increment within the influence area around each support pile.

[0015] Based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, the failure probability distribution of each supporting pile is predicted. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles.

[0016] When the probability of any sub-item in the failure probability distribution of any supporting pile exceeds a preset warning threshold, a supporting pile design parameter adjustment plan is formulated, which includes one or more combinations of increasing the pile diameter, deepening the embedding depth, or reducing the pile spacing.

[0017] The incremental cost of the support structure resulting from the implementation of the adjustment plan is calculated and integrated with the original support cost of the foundation project to output the risk-adjusted support cost.

[0018] When the probability of all sub-items in the failure probability distribution of all support piles does not exceed the threshold, the original support cost of the foundation engineering is output.

[0019] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention constructs a ground-pile coupling model, inputs precipitation prediction parameters as time-varying boundary conditions, and uses pore water pressure increment as the starting point to analyze the transient seepage field evolution caused by precipitation infiltration, thereby providing a scientific and systematic analysis method for in-depth understanding of the stress mechanism of the support structure under precipitation conditions, and helping to more comprehensively grasp the engineering risks.

[0020] (2) Based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, the present invention predicts the probability of pile body tilt, the probability of embedded failure and the probability of soil loss between piles, and triggers directional parameter adjustment based on the probability exceeding the threshold value, identifies high precipitation risk conditions in advance to output risk control support cost, effectively reduces the possibility of support structure failure, enhances the safety and stability of the engineering structure, and thus reduces the economic losses caused by potential accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0022] Figure 1 This is a system structure block diagram provided for the first embodiment of the present invention.

[0023] Figure 2 This is a logical diagram of the construction of the ground-pile coupling model in the first embodiment of the present invention.

[0024] Figure 3 This is a flow chart of a method provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] For example 1, please refer to Figure 1 As shown, in the first embodiment of the present invention, an AI-based engineering cost big data intelligent analysis system is provided, including: a model building module, a penetration deduction module, a failure probability prediction module, a parameter optimization module, a control cost output module and an original cost output module.

[0027] The model construction module is connected to the penetration deduction module, the penetration deduction module is connected to the failure probability prediction module, the failure probability prediction module is respectively connected to the parameter optimization module and the original cost output module, and the parameter optimization module is connected to the control cost output module.

[0028] The model building module builds a ground pile coupling model based on geological survey parameters of the foundation engineering and support pile design parameters.

[0029] See also Figure 2 As shown, in a preferred embodiment of the present invention, the construction process of the pile coupling model includes the following steps: A1. Extracting the spatial distribution data of each soil layer and the groundwater level depth in the geological survey parameters, constructing a three-dimensional foundation initial profile, and presetting a saturated permeability index for each soil layer in association with its soil type.

[0030] It should be noted that the above-mentioned soil types include but are not limited to clay, sand, silt, etc. The preset saturated permeability index can be calibrated by conducting variable head permeability tests on various soil types in the early stage of system development, and can be pre-stored in the cloud database and directly extracted and applied.

[0031] A2. Extract the distribution coordinates, pile diameter, and pile tip elevation of each support pile from the support pile design parameters. Position the support pile in the initial foundation profile and set its geometry to a cylindrical structure. Establish material property transition zones for the cross-layer interface sections where the support piles pass through different soil layers.

[0032] A3. Grid the initial foundation profile. Centered around the axis of the support pile, adaptively set the radius based on the pile diameter to define the influence zone around the spherical pile, i.e., the densified zone. Mesh this influence zone, refining it to a higher density than the baseline density of the non-densified zone (i.e., all areas of the initial foundation profile excluding the influence zone around the spherical pile).

[0033] It should be noted that the radius of the influence zone around the pile needs to cover the stress disturbance range caused by the interaction between the supporting pile and the soil layer. Its adaptive setting standard is specifically the product of the pile diameter value and the soil dependence coefficient, where the soil dependence coefficient is calibrated according to the permeability standard of the soil type.

[0034] A4. Based on the initial groundwater level, the foundation profile is divided into saturated and unsaturated zones with initial state markers. A pore water pressure baseline value is assigned to the saturated zone, and a matrix suction baseline value is assigned to the unsaturated zone.

[0035] It should be noted that the above-mentioned saturated zone division standards are areas in the foundation profile whose elevation is lower than or equal to the initial groundwater level, and parts of the soil layer that reach the preset saturated permeability index of the corresponding soil type.

[0036] The infiltration deduction module inputs the precipitation prediction parameters during the construction period into the model as surface boundary conditions, analyzes the transient seepage changes of the foundation soil under the action of precipitation, and quantifies the pore water pressure increment distribution in the influence area around each support pile.

[0037] In a preferred embodiment of the present invention, the quantification of the pore water pressure increment distribution within the influence area around each support pile includes: converting the peak intensity and duration of a single precipitation in the precipitation prediction parameters into a time-varying surface infiltration flux boundary curve according to preset rules, and loading it into the model surface area according to the time series.

[0038] It should be noted that the specific contents of the above-mentioned preset rules include: setting the preset saturated permeability index of the topmost soil layer in the vertical direction of the three-dimensional profile of the foundation as the upper limit threshold of the infiltration flux; when the peak intensity of precipitation does not exceed the saturated permeability index, defining the time-varying surface infiltration flux boundary curve as a constant value curve, and its flux value is equal to the peak intensity of precipitation; conversely, when the peak intensity of precipitation exceeds the saturated permeability index, performing a segmented conversion, including taking the upper limit value of the saturated permeability index as the peak continuation segment of the flux value and the segment after the precipitation ends when the flux value returns to zero.

[0039] The time span of the surface infiltration flux boundary curve is strictly aligned with the duration of precipitation and continues for a preset drainage time after the precipitation ends.

[0040] The saturation state identifier and pore water pressure value of each grid cell in the foundation profile are reassigned time step by time, and the reassignment operation includes: i. triggering the saturation state identifier update when the water content of the grid cell reaches the preset saturation permeability index of its associated soil layer.

[0041] ii. Delete the matrix suction baseline value of the newly saturated grid cell, update the elevation deviation between the cell center elevation and the groundwater level in real time, and quantify the pore water pressure value of the newly saturated grid cell.

[0042] iii. The pore water pressure replenishment value quantified by the water level rise is superimposed on the configured pore water pressure baseline value of the saturated grid cell.

[0043] It should be noted that the pore water pressure value of the above-mentioned new saturated grid unit is specifically quantified as the product of the absolute deviation between the unit center elevation and the real-time updated elevation of the groundwater level and the preset water density.

[0044] The pore water pressure replenishment value quantified by the water level elevation is specifically quantified as the product of the water level elevation and the preset bulk density of water.

[0045] The pore water pressure value of each grid unit of the pile coupling model at each time step during the precipitation process is collected to analyze the transient seepage changes of the foundation soil under the action of precipitation.

[0046] The pore water pressure differences of each grid unit in the influence area around each support pile before and after precipitation are sorted out, and the pore water pressure increment distribution in the influence area around each support pile is quantified.

[0047] In a preferred embodiment of the present invention, the process of obtaining the pore water pressure difference before and after precipitation of the grid unit includes: if the initial state of the grid unit before precipitation is an unsaturated zone, then the difference between the pore water pressure value after precipitation and the matrix suction reference value is used as the pore water pressure difference before and after precipitation, wherein the matrix suction reference value is a negative value.

[0048] If the initial state of the grid unit before precipitation is a saturated zone, the difference between the pore water pressure after precipitation and the pore water pressure baseline value is taken as the pore water pressure difference before and after precipitation.

[0049] The embodiment of the present invention constructs a ground-pile coupling model, inputs precipitation prediction parameters as time-varying boundary conditions, and uses the pore water pressure increment as the starting point to analyze the transient seepage field evolution caused by precipitation infiltration, providing a scientific and systematic analysis method for in-depth understanding of the stress mechanism of the support structure under precipitation conditions, which helps to more comprehensively grasp the engineering risks.

[0050] The failure probability prediction module predicts the failure probability distribution of each supporting pile based on the reduction relationship between the pore water pressure increment and the effective stress of the soil. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles.

[0051] In a preferred embodiment of the present invention, the predicting of the probability of pile body tilt of each supporting pile includes dividing the influence area around the pile into a left area and a right area with the axis of the supporting pile as the boundary.

[0052] A fitting function representing the reduction relationship between the pore water pressure increment and the effective stress of the soil was constructed through regression analysis. The average effective stress reduction index of the left and right regions was calculated respectively by combining the pore water pressure increment of each grid cell in the left and right regions.

[0053] It should be noted that the above-mentioned fitting function for characterizing the reduction relationship between the pore water pressure increment and the effective stress of the soil is constructed through regression analysis. The specific process includes: based on the force balance principle of the soil unit, defining the physical correlation constraints between the pore water pressure increment and the reduction amount of the effective stress of the soil, and setting a mathematical model for characterizing the nonlinear reduction relationship between the two. In the present invention, an exponential decay model function can be used as an example.

[0054] A preset number of observation groups containing pore water pressure increments and their corresponding soil effective stress reduction indices are obtained, and the nonlinear least squares method is applied to iteratively optimize the mathematical model parameters to minimize the sum of squares of the observation group residuals, thereby determining the final fitting function.

[0055] The absolute difference between the average effective stress reduction indexes on the left and right sides is converted into bilateral imbalance and substituted into the preset failure probability evaluation function to quantify the predicted value of the pile body tilt probability.

[0056] It should be noted that the specific conversion process of the above-mentioned bilateral imbalance degree is: using the absolute difference between the average effective stress reduction indexes of the left and right sides as the numerator, and the maximum value of the average effective stress reduction indexes of the left and right sides as the denominator, the ratio operation is expanded to convert and obtain the bilateral imbalance degree.

[0057] It should also be noted that the above-mentioned preset failure probability evaluation function can be exemplified by a sigmoid function, which is based on the fact that the sigmoid function can strictly follow the critical catastrophe theory of soil instability.

[0058] In a preferred embodiment of the present invention, the prediction of the embedment failure probability of each supporting pile includes: taking the bottom surface of the pile end of the supporting pile as the reference plane, expanding symmetrically on both sides along the cross-section of the pile body and extending vertically downward to a preset embedment depth, and defining the pile end bearing area within the influence area around the pile.

[0059] The pore water pressure increment of each grid unit in the pile end bearing zone is extracted, and the gradient component of the pore water pressure increment in the vertical direction of the pile end bearing zone is calculated using the central difference method.

[0060] Referring to the effective stress reduction index corresponding to the maximum pore water pressure increment of the grid unit in the pile end bearing zone, the ratio of the effective stress reduction index to the preset allowable reduction index threshold of the pile end is substituted into the preset failure probability assessment function, and the correction factor is determined in combination with the positive and negative logical relationship of the gradient component to quantify the predicted value of the embedded failure probability of the support pile.

[0061] It should be noted that the specific process of determining the correction factor by combining the positive and negative logical relationship of the gradient component is as follows: if the gradient component is positive, it means that the pore water pressure increases along the depth direction, that is, the soil pressure below the pile end is higher than the interface pressure at the pile end, indicating that the hydraulic jacking effect occurs in the bearing layer at the pile end, weakening the pile-soil interface bond, thus triggering the risk amplification correction. The correction factor value range is set to .

[0062] If the gradient component is negative, it means that the pore water pressure decreases along the depth direction, that is, the soil pressure below the pile end is lower than the interface pressure at the pile end, forming a favorable pressure gradient, enhancing the embedded stability of the pile end, triggering the risk reduction correction, and setting the correction factor value range to .

[0063] The present invention may exemplarily take the median value of the correction factor according to the range of values.

[0064] In a preferred embodiment of the present invention, the prediction of the probability of soil loss between each supporting pile includes: locating adjacent supporting pile pairs and using the line connecting the centers of the two piles as the reference axis, extending a set distance to both sides to form an interaction area between the piles, and the set distance is dynamically calculated based on the proportional relationship between the pile diameter and the pile spacing.

[0065] It should be noted that the above-mentioned method for dynamically calculating the set distance is as follows: obtain the sum of the pile radius of the adjacent supporting pile pairs, where the pile radius is half of the pile diameter value, subtract the pile spacing from the sum of the pile radius, and use the preset ratio of the difference as the set distance, where the preset ratio can be exemplarily taken in the range of .

[0066] The pore water pressure increments of each grid cell in the pile-to-pile interaction zone are traversed, and cells with increment values ​​higher than the regional average are marked as high-pressure cells, and cells with increment values ​​lower than the average are marked as low-pressure cells.

[0067] The direction data of the seepage path generated by the high-pressure unit to the adjacent low-pressure unit are collected. When the angle between the extension direction of the seepage path and the normal direction of the reference axis is less than the preset angle, it is determined to be a soil loss risk path towards the gap between piles.

[0068] The ratio of risk paths to all seepage paths is calculated and substituted into the preset failure probability evaluation function to quantify the probability of soil loss between piles.

[0069] The parameter optimization module formulates a support pile design parameter adjustment plan when the probability of any sub-item in the failure probability distribution of any support pile exceeds a preset warning threshold. The adjustment plan includes one or more combinations of increasing the pile diameter, deepening the embedding depth, or reducing the pile spacing.

[0070] In a preferred embodiment of the present invention, the formulation of a support pile design parameter adjustment plan includes: marking support piles in which the probability of sub-items in the failure probability distribution exceeds a preset warning threshold as risky support piles, and using the distribution position coordinates of the risky support piles as the adjustment position coordinates.

[0071] The failure mode types triggered by each risk support pile and their corresponding adjustment measures are sorted out. If a single risk support pile triggers multiple failure mode types, a combined measure chain is generated according to the preset priority adjustment sequence, where the preset priority adjustment sequence is, in order, increasing the pile diameter, deepening the embedment depth, and reducing the pile spacing.

[0072] It should be noted that the ranking basis of the above-mentioned preset priority adjustment order is to minimize the construction disturbance when increasing the pile diameter and simultaneously improve the bending and shearing capacity. The incremental material cost is linearly controllable and is suitable as the first choice for priority adjustment.

[0073] Deepening the embedment depth requires additional excavation, which is specifically designed to address embedment failure and the limitation of the bearing layer depth. However, compared with reducing the pile spacing, which requires overall pile layout adjustment and triggers chain design changes, the project is easier to control and is suitable as a priority adjustment option.

[0074] The risk level table for each failure mode type stored in the cloud database is accessed to determine the risk level mapped to the probability prediction value corresponding to the triggering failure mode type, and the preset adjustment parameters of the risk level are simultaneously extracted.

[0075] The control cost output module calculates the incremental cost of the support structure generated by implementing the adjustment plan, integrates it with the original support cost of the foundation project, and outputs the risk-controlled support cost.

[0076] In a preferred embodiment of the present invention, the calculation of the incremental cost of the support structure generated by implementing the adjustment scheme includes: analyzing the adjustment parameter items in the adjustment scheme, and extracting the cumulative adjustment amounts corresponding to the pile diameter increase value, the embedment deepening value, and the pile spacing reduction value.

[0077] The unit cost parameter table stored in the cloud database is accessed to determine the incremental cost corresponding to the cumulative adjustment amount of each adjustment parameter item type in the adjustment plan.

[0078] The adjusted parameter item types are summarized to obtain the incremental cost of the support structure generated by the adjustment scheme.

[0079] The original cost output module outputs the original support cost of the foundation project when all sub-item probabilities in the failure probability distribution of all support piles do not exceed a threshold value.

[0080] The embodiment of the present invention predicts the probability of pile body inclination, embedment failure and soil loss between piles of support piles based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, and triggers directional parameter adjustment based on the probability exceeding the threshold value, identifies high precipitation risk conditions in advance to output risk-controlled support costs, effectively reduces the possibility of failure of the support structure, enhances the safety and stability of the engineering structure, and thus reduces the economic losses caused by potential accidents.

[0081] It should also be added that a cloud database is applied during the execution of the present invention, wherein all parameters in the cloud database are implanted during system development.

[0082] For example 2, please refer to Figure 3 As shown, in the second embodiment of the present invention, an AI-based intelligent analysis method for engineering cost big data is provided, including: constructing a ground pile coupling model according to the geological survey parameters of the foundation engineering and the support pile design parameters.

[0083] The precipitation prediction parameters during the construction period are input into the model as surface boundary conditions to analyze the transient seepage changes of the foundation soil under the action of precipitation and quantify the distribution of the pore water pressure increment within the influence area around each support pile.

[0084] Based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, the failure probability distribution of each supporting pile is predicted. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles.

[0085] When the probability of any sub-item in the failure probability distribution of any supporting pile exceeds a preset warning threshold, a supporting pile design parameter adjustment plan is formulated, which includes one or more combinations of increasing the pile diameter, deepening the embedding depth, or reducing the pile spacing.

[0086] The incremental cost of the support structure resulting from the implementation of the adjustment plan is calculated and integrated with the original support cost of the foundation project to output the risk-adjusted support cost.

[0087] When the probability of all sub-items in the failure probability distribution of all support piles does not exceed the threshold, the original support cost of the foundation engineering is output.

[0088] The embodiment of the present invention provides an AI-based intelligent analysis method for engineering cost big data. Its implementation principle and technical effects are the same as those of the aforementioned system embodiment. For the sake of brief description, for matters not mentioned in the method embodiment, please refer to the corresponding content in the aforementioned system embodiment.

[0089] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. An AI-based intelligent analysis system for engineering cost big data, characterized by: include: Construct a ground-pile coupling model based on the geological survey parameters of the foundation engineering and the design parameters of the support piles; The predicted precipitation parameters during the construction period are input into the model as surface boundary conditions to analyze the transient seepage changes of the foundation soil under the action of precipitation and quantify the pore water pressure increment distribution in the influence area around each support pile. Based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, the failure probability distribution of each supporting pile is predicted. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles. When the probability of any sub-item in the failure probability distribution of any supporting pile exceeds the preset warning threshold, a supporting pile design parameter adjustment plan is formulated, which includes one or more combinations of increasing the pile diameter, deepening the embedment depth, or reducing the pile spacing; Calculate the incremental cost of the support structure resulting from the implementation of the adjustment plan, integrate it with the original support cost of the foundation project, and output the risk-adjusted support cost; When the probability of all sub-items in the failure probability distribution of all support piles does not exceed the threshold, the original support cost of the foundation engineering is output.

2. The AI-based engineering cost big data intelligent analysis system according to claim 1 is characterized by: The construction process of the ground-pile coupling model includes the following steps: A1. Extract the spatial distribution data of each soil layer and the groundwater level depth from the geological survey parameters, construct an initial 3D foundation profile, and preset saturated permeability indicators for each soil layer, correlating them with its soil type. A2. Extract the distribution coordinates, pile diameter, and pile tip elevation of each support pile from the support pile design parameters. Position the support pile in the initial foundation section and set its geometry to a cylindrical structure. Establish material property transition zones for the intersection sections where the support piles pass through different soil layers. A3. Mesh the initial foundation cross-section. Centered around the support pile axis, adaptively set the radius based on the pile diameter to define the influence area around the spherical pile, i.e., the densification area. Mesh refinement is then performed within this influence area, requiring the mesh density to be higher than the baseline density of the non-densification area. A4. Based on the initial groundwater level, the foundation profile is divided into saturated and unsaturated zones with initial state markers. A pore water pressure baseline value is assigned to the saturated zone, and a matrix suction baseline value is assigned to the unsaturated zone.

3. The AI-based engineering cost big data intelligent analysis system according to claim 2 is characterized by: The quantification of the pore water pressure increment distribution within the influence area around each support pile includes: The peak intensity and duration of a single rainfall in the precipitation prediction parameters are converted into a time-varying surface infiltration flux boundary curve according to a preset rule, and loaded into the model surface area according to the time series; Reassigning the saturation state identifier and pore water pressure value of each grid cell in the foundation section at each time step, wherein the reassignment operation includes: i. triggering the update of the saturation state identifier when the water content of the grid cell reaches the preset saturation permeability index of the associated soil layer; ii. Delete the matrix suction baseline value of the newly saturated grid cell, update the elevation deviation between the cell center elevation and the groundwater level in real time, and quantify the pore water pressure value of the newly saturated grid cell; iii. The pore water pressure supplement value quantified by the water level rise is superimposed on the configured pore water pressure baseline value of the saturated grid cell; The pore water pressure value of each grid unit of the pile coupling model at each time step during the precipitation process is collected to analyze the transient seepage changes of the foundation soil under the action of precipitation; The pore water pressure differences of each grid unit in the influence area around each support pile before and after precipitation are sorted out, and the pore water pressure increment distribution in the influence area around each support pile is quantified.

4. The AI-based engineering cost big data intelligent analysis system according to claim 3 is characterized by: The process of obtaining the pore water pressure difference before and after the precipitation action on the grid unit includes: If the initial state of the grid unit before precipitation is an unsaturated zone, the difference between the pore water pressure value after precipitation and the matrix suction reference value is taken as the pore water pressure difference before and after precipitation, where the matrix suction reference value is a negative value. If the initial state of the grid unit before precipitation is a saturated zone, the difference between the pore water pressure after precipitation and the pore water pressure baseline value is taken as the pore water pressure difference before and after precipitation.

5. The AI-based engineering cost big data intelligent analysis system according to claim 3 is characterized by: The prediction of the probability of the pile body tilting of each supporting pile includes: The left and right areas of the influence area around the pile are divided based on the axis of the supporting pile. A fitting function representing the reduction relationship between pore water pressure increment and soil effective stress was constructed through regression analysis. The average effective stress reduction index of the left and right regions was calculated by combining the pore water pressure increment of each grid cell in the left and right regions. The absolute difference between the average effective stress reduction indexes on the left and right sides is converted into bilateral imbalance and substituted into the preset failure probability evaluation function to quantify the predicted value of the pile body tilt probability.

6. The AI-based engineering cost big data intelligent analysis system according to claim 4 is characterized by: The prediction of the embedment failure probability of each supporting pile includes: Taking the bottom surface of the supporting pile end as the reference plane, the pile end bearing zone within the influence area around the pile is defined by extending symmetrically on both sides along the cross section of the pile body and vertically downward to a preset embedding depth; Extracting the pore water pressure increment of each grid cell in the pile end bearing zone, and calculating the gradient component of the pore water pressure increment in the vertical direction of the pile end bearing zone using the central difference method; Referring to the effective stress reduction index corresponding to the maximum pore water pressure increment of the grid unit in the pile end bearing zone, the ratio of the effective stress reduction index to the preset allowable reduction index threshold of the pile end is substituted into the preset failure probability assessment function, and the correction factor is determined in combination with the positive and negative logical relationship of the gradient component to quantify the predicted value of the embedded failure probability of the support pile.

7. The AI-based engineering cost big data intelligent analysis system according to claim 4 is characterized by: The prediction of soil loss probability between each supporting pile includes: Position adjacent support pile pairs and use the line connecting the centers of the two piles as the reference axis, extending a set distance to both sides to form an interaction zone between the piles. The set distance is dynamically calculated based on the ratio between the pile diameter and the pile spacing. The pore water pressure increment of each grid cell in the pile interaction zone is traversed, and the cells with increment values ​​higher than the regional average level are marked as high-pressure cells, and the cells with increment values ​​lower than the average level are marked as low-pressure cells; The direction data of the seepage path generated by the high-pressure unit to the adjacent low-pressure unit are collected. When the angle between the extension direction of the seepage path and the normal direction of the reference axis is less than the preset angle, it is determined to be a soil loss risk path towards the gap between the piles; The ratio of risk paths to all seepage paths is calculated and substituted into the preset failure probability evaluation function to quantify the probability of soil loss between piles.

8. The AI-based engineering cost big data intelligent analysis system according to claim 1 is characterized by: The formulation of the support pile design parameter adjustment plan includes: Mark the support piles with sub-item probabilities exceeding the preset warning threshold in the failure probability distribution as risky support piles, and use the distribution position coordinates of the risky support piles as the adjustment position coordinates; Organize the failure mode types triggered by each risk support pile and their corresponding adjustment measures. If a single risk support pile triggers multiple failure mode types, generate a combined measure chain according to the preset priority adjustment sequence, where the preset priority adjustment sequence is, in order, increasing the pile diameter, deepening the embedment depth, and reducing the pile spacing; The risk level table for each failure mode type stored in the cloud database is accessed to determine the risk level mapped to the probability prediction value corresponding to the triggering failure mode type, and the preset adjustment parameters of the risk level are simultaneously extracted.

9. The AI-based engineering cost big data intelligent analysis system according to claim 1 is characterized by: The calculation of the incremental cost of the supporting structure resulting from the implementation of the adjustment plan includes: Analyze the adjustment parameters in the adjustment plan and extract the cumulative adjustment values ​​corresponding to the pile diameter increase, embedment deepening value, and pile spacing reduction value; Accessing a unit cost parameter table stored in a cloud database to determine the incremental cost corresponding to the cumulative adjustment amount of each adjustment parameter item type in the adjustment plan; The adjusted parameter item types are summarized to obtain the incremental cost of the support structure generated by the adjustment scheme.

10. An AI-based intelligent analysis method for engineering cost big data, characterized in that: include: Construct a ground-pile coupling model based on the geological survey parameters of the foundation engineering and the design parameters of the support piles; The predicted precipitation parameters during the construction period are input into the model as surface boundary conditions to analyze the transient seepage changes of the foundation soil under the action of precipitation and quantify the pore water pressure increment distribution in the influence area around each support pile. Based on the reduction relationship between the pore water pressure increment and the effective stress of the soil, the failure probability distribution of each supporting pile is predicted. The failure probability distribution includes the probability of pile body tilt, the probability of embedded failure, and the probability of soil loss between piles. When the probability of any sub-item in the failure probability distribution of any supporting pile exceeds a preset warning threshold, a supporting pile design parameter adjustment plan is formulated, the adjustment plan including one or more combinations of increasing the pile diameter, deepening the embedding depth, or reducing the pile spacing; Calculate the incremental cost of the support structure resulting from the implementation of the adjustment plan, integrate it with the original support cost of the foundation project, and output the risk-adjusted support cost; When the probability of all sub-items in the failure probability distribution of all support piles does not exceed the threshold, the original support cost of the foundation engineering is output.

Images