Pile foundation depth detection method based on electrical method temperature sensing device and random forest
Patent Information
- Application Number
- CN202510985001.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-17
- Publication Date
- 2025-10-31
Smart Images
Figure CN120867348A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of pile foundation testing technology, and particularly relates to a pile foundation depth detection method based on an electrical resistivity tomography (EDM) device and random forest. Background Technology
[0002] Pile foundations are crucial structures in buildings, bearing the weight of the structure and transferring it to the deep ground. Their depth information directly affects the safety of construction. With the development of urban underground space, the demand for pile foundation testing is increasing, especially the determination of pile depth, which has become a critical step in underground engineering construction. Traditional testing methods, such as low-strain methods, high-strain methods, static load tests, and core drilling, have limitations in practical applications, making it difficult to simultaneously meet the requirements of efficiency, non-destructive testing, and accuracy. In recent years, electrical resistivity tomography (EDT) technology, due to its non-destructive nature and sensitivity to underground electrical media, has been widely used in geological exploration and engineering testing.
[0003] However, the application of existing electrical resistivity tomography (EDT) techniques in pile foundation depth detection still faces challenges such as insufficient accuracy and limited applicability.
[0004] Electrical resistivity tomography (ERT) techniques mainly include two categories: surface ERT and borehole ERT. Surface ERT infers subsurface structural information by placing electrodes on the surface and utilizing changes in the current field, and is widely used in shallow geological exploration. However, in pile foundation depth detection, surface ERT is limited by penetration depth and resolution, making it difficult to accurately detect the location of deep piles, and is easily affected by urban environmental noise. In contrast, borehole ERT, by placing electrodes in boreholes, directly measures signals deep underground, significantly improving detection depth and resolution, and is suitable for deep structure detection.
[0005] Existing borehole electrical resistivity techniques include resistivity logging and cross-hole electrical resistivity methods.
[0006] Resistivity logging: This method measures the resistivity change of the medium surrounding the borehole by arranging electrodes in a single borehole. It is mainly used for stratum stratification and groundwater detection, but it is difficult to accurately distinguish the interface between the pile foundation and the surrounding medium in pile foundation depth detection.
[0007] Trans-hole electrical resistivity method: By measuring the electric field between two or more boreholes, a three-dimensional distribution of underground resistivity can be constructed. However, multiple boreholes are required for pile foundation testing, resulting in high construction costs and complex operation.
[0008] Therefore, developing an efficient, accurate, and adaptable method for detecting pile foundation depth has significant technical and engineering value. Summary of the Invention
[0009] The purpose of this invention is to address the aforementioned technical problems by providing a method for detecting pile foundation depth based on an electrical resistivity tomography (EDM) device and random forest.
[0010] In view of this, the present invention provides a method for detecting the depth of pile foundations based on an electrical resistivity tomography (EDT) device and a random forest, comprising the following steps:
[0011] S1, Drill holes next to the pile foundation to determine the location and depth of the holes;
[0012] S2, Arrange electrode pairs AB and MN along the borehole, set the spacing and fix the electrodes;
[0013] S3: Apply current, measure the potential difference signal at different depths, and record the data;
[0014] S4, calculate the apparent resistivity distribution and generate a curve;
[0015] S5 inputs the apparent resistivity data into the random forest model and outputs the pile foundation depth and uncertainty.
[0016] Preferably, in step S2, the electrodes are arranged along the drilling depth direction, with electrode B located deep in the drilling hole, electrode MN located above electrode B, and electrode A located shallow in the drilling hole.
[0017] Preferably, in step S3, the AB electrode is used to apply a constant current, and the MN electrode is used to record voltage signals at different depths.
[0018] Preferably, in step S3, the electrode pair covers the entire borehole depth range by stepping and moving to obtain continuous electric field distribution data.
[0019] Preferably, after the electrode pair moves to the deepest point, it is measured again in reverse to cross-validate and superimpose the data, thereby improving the signal-to-noise ratio.
[0020] Preferably, in step S4, the formula for calculating the apparent resistivity is:
[0021]
[0022] Where K is a geometric factor, determined by the relative positions of the electrodes, ΔU MN Let be the potential difference measured between the electrodes MN and I be the current intensity of the constant current applied to AB by the power supply electrode.
[0023] Preferably, the formula for calculating K is:
[0024]
[0025] Where, r xy Let x be the distance between points x and y.
[0026] Preferably, the training process of the random forest model includes:
[0027] Construct a dataset and obtain sample data of known pile foundation depths through numerical simulation;
[0028] Feature extraction is performed, with input features including apparent resistivity and potential difference along the borehole depth;
[0029] Bootstrap sampling is used to generate multiple subset datasets, multiple decision trees are trained, and hyperparameters are optimized through cross-validation.
[0030] Preferably, when the model outputs an estimated value of the pile foundation depth, the standard deviation of the prediction result is calculated using the inter-tree variance, providing a quantitative basis for engineering decision-making.
[0031] Preferably, the electrode is integrally formed from a corrosion-resistant and highly conductive metal material and is fixed inside the borehole by a cable.
[0032] The beneficial effects of this invention are:
[0033] This invention simplifies the construction process by applying the Wenner device electrical resistivity method to a single borehole environment, enabling real-time prediction of pile foundation conditions and shortening the detection cycle. The optimized electrode arrangement can accurately adapt to the pile foundation detection requirements, effectively capturing the electrical differences between the pile foundation and the surrounding medium, while avoiding signal problems caused by improper distance. The derived apparent resistivity calculation method suitable for borehole environments overcomes the limitations of traditional formulas in this environment and can accurately reflect the electrical characteristics of the underground medium. The introduction of random forest technology to process electrical resistivity data not only provides accurate pile foundation depth prediction but also performs uncertainty assessment, reducing engineering decision-making risks and significantly improving the accuracy and reliability of detection. Attached Figure Description
[0034] Figure 1 This is a schematic diagram of the Wenner device for detecting pile foundation depth according to the present invention;
[0035] Figure 2 This is a flowchart of the pile foundation depth prediction process of the present invention;
[0036] Figure 3 This is a graph showing the apparent resistivity versus depth under different hole spacings according to the present invention.
[0037] Figure 4 This is a schematic diagram of the random forest model structure of the present invention;
[0038] Figure 5 This is a schematic diagram illustrating the exploratory effect of different borehole spacings of a dipole-dipole device on the depth of the pile foundation.
[0039] Figure 6 A schematic diagram illustrating the exploratory effect of different borehole spacings on pile foundation depth using the Wenner method;
[0040] Figure 7A schematic diagram illustrating the effect of different electrical conductivity values of dipole-dipole pile foundations on the detectability of pile foundation depth.
[0041] Figure 8 A schematic diagram illustrating the effect of different pile foundation electrical conductivity on the detectability of pile foundation depth in Wenner.
[0042] Figure 9 This is a schematic diagram showing the prediction error corresponding to different leaf sizes in the random forest method.
[0043] Figure 10 This is a schematic diagram showing the change of the root mean square error as a function of the number of trees during the training phase.
[0044] Figure 11 This is a schematic diagram of the prediction results for random forest pile foundations. Detailed Implementation
[0045] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0046] It should be noted that all directional and positional terms used in this invention, such as "up," "down," "left," "right," "front," "back," "vertical," "horizontal," "inner," "outer," "top," "lower," "lateral," "longitudinal," and "center," are only used to explain the relative positional relationships and connections between components in a specific state (as shown in the accompanying drawings). They are merely for the convenience of describing the invention and do not require the invention to be constructed and operated in a specific orientation; therefore, they should not be construed as limitations on the invention. Furthermore, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.
[0047] In the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0048] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0049] like Figures 1-4 As shown, a method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest includes the following steps:
[0050] S1, Drill holes next to the pile foundation to determine the location and depth of the holes;
[0051] S2, Arrange electrode pairs AB and MN along the borehole, set the spacing and fix the electrodes;
[0052] S3: Apply current, measure the potential difference signal at different depths, and record the data;
[0053] S4, calculate the apparent resistivity distribution and generate a curve;
[0054] S5 inputs the apparent resistivity data into the random forest model and outputs the pile foundation depth and uncertainty.
[0055] Compared to existing technologies, this application simplifies the construction process by applying the Wenner device electrical resistivity method to a single borehole environment. It enables real-time prediction of pile foundations, shortens the detection cycle, and the optimized electrode arrangement can accurately adapt to the pile foundation detection requirements. It can effectively capture the electrical differences between the pile foundation and the surrounding medium, while avoiding signal problems caused by improper distance. The derived apparent resistivity calculation method suitable for the borehole environment overcomes the limitations of traditional formulas in this environment and can accurately reflect the electrical characteristics of the underground medium. The introduction of random forest technology to process electrical resistivity data not only provides accurate prediction of pile foundation depth but also performs uncertainty assessment, reduces engineering decision-making risks, and significantly improves the accuracy and reliability of detection.
[0056] As a preferred example of this application, in step S2, the electrodes are arranged along the drilling depth direction, with electrode B located deep in the borehole, electrode MN located above electrode B, and electrode A located shallow in the borehole. This arrangement is suitable for a single borehole environment and can work with borehole electrical resistivity tomography to more accurately detect deep information of the pile foundation, providing more reliable basic data for the random forest model, helping to improve the accuracy of pile foundation depth prediction, and also helping to simplify the electrode arrangement process during construction.
[0057] Additionally, it should be noted that the borehole is located to the side of the pile foundation being tested (e.g., ...). Figure 1 As shown in the figure, the distance between the center of the electrode and the edge of the pile foundation is a key parameter in the detection design. The reference value is that this distance is equal to 1 meter. The selection of this distance is based on ensuring that the electrode can effectively detect the electrical difference between the pile foundation and the surrounding medium, while avoiding signal saturation due to being too close or signal attenuation due to being too far. The specific distance can be adjusted according to the pile foundation type, geological conditions, and experimental verification. The drilling depth should be slightly greater than the expected pile foundation depth, preferably 10%-20% greater than the expected pile foundation depth, to ensure coverage of the entire pile foundation and acquisition of bottom characteristic signals.
[0058] The electrode is integrally formed from a corrosion-resistant, highly conductive metal material (such as stainless steel or copper) and fixed inside the borehole by a cable to ensure good contact with the surrounding medium.
[0059] As a preferred example of this application, in step S3, the AB electrode is used to apply a constant current, and the MN electrode is used to record voltage signals at different depths. The reference distance between the power supply electrodes AB is 4.5 meters. The reference distance between the measuring electrodes MN is 1.5 meters; the reference distance between electrode B and electrode M is 1.5 meters. Stable current and accurate voltage measurement provide a guarantee for obtaining effective information about the deep foundation of the pile, making the subsequent apparent resistivity calculation more accurate, and thus making the prediction of the random forest model more reliable. Moreover, this process does not require damage to the pile foundation structure, which meets the needs of quality inspection of completed projects.
[0060] It should be noted that the Wenner device generates a local electric field through the AB electrodes, and the MN electrodes measure the changes in the electric field, enabling it to sensitively capture resistivity anomalies caused by the bottom of the pile foundation. The electrode distance from the reference value is based on the penetration depth theory of electrical resistivity tomography and the size characteristics of the pile foundation. Through experiments and numerical simulations, the signal resolution and signal-to-noise ratio are ensured to be optimal (e.g., ...). Figure 3 (As shown).
[0061] As a preferred example of this application, in step S3, the electrode pair covers the entire borehole depth range by stepping and moving (each movement distance can be set to 0.5 meters or 1 meter) to obtain continuous electric field distribution data. Continuous data acquisition can comprehensively reflect the pile foundation information at different depths, which helps to improve the adaptability to different pile foundation types and geological conditions. At the same time, continuous data also provides the possibility for real-time prediction, shortens the detection cycle, and makes the uncertainty assessment of the prediction results more accurate.
[0062] As a preferred example of this application, after the electrode pair moves to the deepest point, it is measured again in reverse to cross-validate and superimpose the data, thereby improving the signal-to-noise ratio. This operation further improves the data quality, makes the deep information of the pile foundation clearer, enhances the accuracy of the random forest model prediction, and reduces the impact of data errors on the prediction results under different pile foundation types and geological conditions.
[0063] A constant current is applied through the AB electrodes, and the current intensity can be set according to geological conditions and equipment capabilities; the current must be kept stable to ensure the reliability of the measurement data.
[0064] The MN electrode records voltage signals at different depths. After each measurement, the electrode pair moves downwards along the borehole, with the step size adjustable according to the required detection accuracy (e.g., 0.5 meters or 1 meter). After reaching the deepest point, the electrode is reversed and measured again. The data are cross-validated and superimposed to improve the signal-to-noise ratio. The voltage value (ΔU) at each depth point is recorded. MN The potential difference data along the borehole depth is generated by combining the electrode positions with the potential difference data.
[0065] As a preferred example of this application, due to the apparent resistivity (ρ) a The apparent resistivity reflects the electrical characteristics of the underground medium and is a core parameter for pile foundation depth detection. Due to the special characteristics of the borehole environment (limited space, non-homogeneous medium), the traditional apparent resistivity formula for surface electrical methods is not applicable. Therefore, in step S4, a calculation formula for the apparent resistivity applicable to the borehole Wenner device is proposed:
[0066]
[0067] Where K is a geometric factor, determined by the relative positions of the electrodes, ΔU MN Let be the potential difference measured between the electrodes and MN, and I be the current intensity of the constant current applied to AB by the power supply electrode.
[0068] The geometric factor K is determined by the relative positions of the electrodes and can be derived using the mirror method (i.e., introducing virtual power supply electrodes A' and B' with the same current intensity at the mirror points of power supply electrodes A and B (with the ground as the plane of symmetry)).
[0069]
[0070] Where, r xy The distance between points x and y is used to calculate the apparent resistivity distribution along the borehole depth based on the measurement data, forming an apparent resistivity curve for subsequent analysis.
[0071] Random forest is an ensemble learning algorithm that constructs multiple decision trees and integrates their predictions, enabling it to handle high-dimensional, nonlinear data. Therefore, this invention utilizes the random forest model to analyze apparent resistivity data and predict pile foundation depth (e.g., ...). Figure 4 (as shown);
[0072] As a preferred example of this application, the training process of the random forest model includes:
[0073] A dataset was constructed, and sample data of known pile foundation depths were obtained through numerical simulation, including pile diameter, length, electrical conductivity, and observation device parameters.
[0074] Feature extraction is performed, with input features including apparent resistivity and potential difference along the borehole depth;
[0075] Multiple subset datasets were generated using Bootstrap sampling, 100 decision trees were trained, and hyperparameters (such as tree depth and leaf size) were optimized through cross-validation.
[0076] As a preferred example of this application, when the model outputs an estimated value of the pile foundation depth, the standard deviation of the prediction result is calculated through the inter-tree variance to provide a quantitative basis for engineering decision-making;
[0077] Random forests can capture the nonlinear relationship between apparent resistivity (or potential difference) and pile depth. They have good tolerance for noisy data and outliers. They provide feature importance analysis to help understand which depth points have the greatest contribution to prediction.
[0078] In addition, the following verification analyses were performed in this application:
[0079] 1. Sensitivity analysis of borehole electrical resistivity to pile depth:
[0080] To ensure building safety while also considering detection effectiveness during construction, boreholes are typically installed approximately 1 meter alongside the pile foundation. Therefore, three borehole spacings (distance between the borehole and the pile foundation) of 0.5m, 1m, and 1.5m were designed to investigate the detectability of pile foundation depth using borehole electrical resistivity tomography (ERT) under different conditions. The simulation was set with a pile foundation depth of 8m. Due to the volume effect of ERT, the impact of pile diameter on detection effectiveness was also tested, with three pile diameters set at 0.4m, 1m, and 2m. The pile foundation conductivity was set to 1 S / m, and the background conductivity was set to 0.01 S / m.
[0081] Figure 5 The effects of different pile diameters and borehole spacings on pile depth detection using a dipole-dipole setup are demonstrated. The first row of piles has a diameter of 0.4m, the second row has a diameter of 1m, and the third row has a diameter of 2m. When the pile diameter is fixed, it is evident that the smaller the borehole spacing, the more significant the changes in the potential difference curve and apparent resistivity curve become at the point where the pile disappears. As the borehole spacing increases to 1.5m, the magnitude of these changes weakens. When the pile diameter changes from 0.4m to 2m, the magnitude of the changes in both the potential difference curve and apparent resistivity increases, which is related to the volume effect of the pile, indicating that piles with larger diameters are more detectable.
[0082] Figure 6The study demonstrates the effectiveness of the Wenner device for detecting pile depth under different pile diameters and borehole spacings. Similar to the dipole-dipole device, the smaller the borehole spacing and the larger the pile diameter, the more significant the detection effect of the Wenner device. Furthermore, compared to the dipole-dipole device, the potential difference curve of the Wenner device exhibits a characteristic of first decreasing and then increasing, which corresponds to the pile depth. Considering construction conditions, boreholes should be placed as close to the pile as possible; however, too close a distance may compromise building safety. Therefore, considering all factors, placing boreholes 1m away from the pile yields the highest efficiency.
[0083] The difference in electrical conductivity between the pile foundation and the background is also a factor affecting the pile foundation detection effect. Therefore, with a fixed borehole spacing of 1m and a background conductivity of 0.01S / m, the detection effect of the dipole-dipole device and the Wenner device on the pile foundation depth was simulated under different pile foundation diameters (0.4m, 1m, 2m) and different pile foundation electrical conductivity (0.05S / m, 0.1S / m, 0.5S / m, 1S / m).
[0084] Simulation results show ( Figure 7 , Figure 8 The first to third columns correspond to pile diameters of 0.4m, 1m, and 2m, respectively. A higher pile conductivity, meaning a greater difference in conductivity between the pile and the background, results in a more significant change in the simulated response curve, which is more beneficial for detection. When the pile conductivity is set to 0.05 S / m, which is only 5 times the background conductivity, the change in the simulated response curve is not very obvious, indicating that in actual detection, changes in pile type affect the detection effect of pile depth.
[0085] 2. Pile foundation depth prediction based on random forest:
[0086] After obtaining electrical response data, the inversion method can be used to predict the pile foundation depth. However, the observation data in pile foundation testing is only obtained from the well, the amount of data is small, the inversion model space is large, the non-uniqueness of the solution is high, and the inversion is time-consuming, resulting in poor timeliness in engineering construction. This application uses forward numerical simulation to generate samples and uses the random forest method to train the model. Finally, the trained model is used for testing.
[0087] During sample generation, the pile conductivity was set to logarithmically vary from 0.05 S / m to 10 S / m, the pile depth to linearly vary from 8 m to 30 m, and the pile diameter to linearly vary from 0.4 m to 2 m, generating a total of 1000 samples. Sensitivity analysis results showed that the Wenner apparatus was more sensitive to the location of the pile depth; therefore, the Wenner apparatus was selected for numerical simulation of the samples. Simulating 1000 samples took a total of 20.35 hours.
[0088] First, we tested the effect of different leaf sizes on prediction error using 50 trees, with leaf sizes set to 5, 10, 20, 50, and 100. From the perspective of prediction error... Figure 9 A leaf size of 5 or 10 is suitable. A larger leaf size can increase the model's generalization ability; a leaf size of 10 was subsequently chosen. During the training phase, the number of trees was set to 100, using the observed potential difference as input and the pile depth as the predicted value. To simulate actual observation, Gaussian noise with a standard deviation of 0.2 was added to each potential difference data point. The entire training process took 0.62 seconds.
[0089] The training process is as follows Figure 10 As shown, in the initial stage, the number of trees involved in training was small, and the root mean square error was large. When the number of trees reached 20, the error was significantly improved. Finally, after training with 100 trees, the root mean square error was less than 10⁻³. Figure 11 The training results are shown, demonstrating a very high degree of agreement between the predicted and actual pile foundation depths, with an overall prediction error of less than 0.08m. It should be noted that when constructing samples, different pile diameters and electrical conductivity may correspond to the same pile foundation depth, which increases the complexity of the prediction. In this case, the random forest also exhibits good adaptability.
[0090] By numerically simulating the potential difference and apparent resistivity response of different pile foundation types under borehole measurement conditions using dipole-dipole and Wenner devices, and estimating the pile foundation depth using random forest, the following conclusions were obtained:
[0091] (1) For both the borehole dipole-dipole device and the Wenner device, the potential difference and apparent resistivity response are sensitive to the depth of the pile foundation. Among them, the simulated response of the Wenner device has a better correspondence with the pile foundation depth.
[0092] (2) The diameter and conductivity of the pile foundation have a significant impact on the simulation response. In particular, the difference between the conductivity of the pile foundation and the background conductivity determines the detectability of the pile foundation depth by electrical method. The simulation shows that when the conductivity of the pile foundation is 0.05 S / m and the background conductivity is 0.01 S / m, the electrical method device can still effectively detect the pile foundation depth.
[0093] (3) Random forests can predict the depth of pile foundations well, have good robustness to noise, and can adapt to different types of pile foundations. In practical applications, the depth of pile foundations can be obtained in real time through pre-trained models.
[0094] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.
Claims
1. A method for detecting pile foundation depth based on an electrical resistivity tomography (EDM) device and random forest, characterized in that: Includes the following steps: S1, Drill holes next to the pile foundation to determine the location and depth of the holes; S2, Arrange electrode pairs AB and MN along the borehole, set the spacing and fix the electrodes; S3: Apply current, measure the potential difference signal at different depths, and record the data; S4, calculate the apparent resistivity distribution and generate a curve; S5 inputs the apparent resistivity data into the random forest model and outputs the pile foundation depth and uncertainty.
2. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 1, characterized in that: In step S2, the electrodes are arranged along the drilling depth direction, with electrode B located deep in the drilling hole, electrode MN located above electrode B, and electrode A located shallow in the drilling hole.
3. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 2, characterized in that: In step S3, the AB electrode is used to apply a constant current, and the MN electrode is used to record voltage signals at different depths.
4. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 3, characterized in that: In step S3, the electrode pair covers the entire borehole depth range by stepping and moving to obtain continuous electric field distribution data.
5. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 4, characterized in that: After the electrode pair moves to the deepest point, it is measured again in reverse to cross-validate and superimpose the data, thereby improving the signal-to-noise ratio.
6. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 5, characterized in that: In step S4, the formula for calculating the apparent resistivity is: Where K is a geometric factor, determined by the relative positions of the electrodes, ΔU MN Let be the potential difference measured between the electrodes MN and I be the current intensity of the constant current applied to AB by the power supply electrode.
7. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 6, characterized in that: The formula for calculating K is: Where, r xy Let x be the distance between points x and y.
8. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 7, characterized in that: The training process of the random forest model includes: Construct a dataset and obtain sample data of known pile foundation depths through numerical simulation; Feature extraction is performed, with input features including apparent resistivity and potential difference along the borehole depth; Bootstrap sampling is used to generate multiple subset datasets, multiple decision trees are trained, and hyperparameters are optimized through cross-validation.
9. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 8, characterized in that: When the model outputs an estimated value of the pile foundation depth, it calculates the standard deviation of the prediction result using the inter-tree variance, providing a quantitative basis for engineering decision-making.
10. The method for detecting pile foundation depth based on an electrical resistivity tomography (EDT) device and random forest according to claim 1, characterized in that: The electrode is integrally formed from a corrosion-resistant, highly conductive metal material and is fixed inside the borehole by a cable.