A method and system for intelligent detection of pile foundation bearing capacity

Abstract

The present invention discloses an intelligent detection method and system for pile foundation bearing capacity, which relates to the field of pile foundation detection technology. By injecting energy pulses and collecting the original signal package Ds containing force and acceleration time-series signals and the pile body foundation parameter set Pk, and calculating the injected effective energy Ein, the accuracy of the source data is ensured. According to the original signal package Ds and the pile body foundation parameter set Pk, the energy decoupling vector Edec is calculated and generated, achieving a deep insight into the complex bearing characteristics of the pile foundation. According to the energy decoupling vector Edec, the characteristic feature vector Vx is calculated and generated, laying a standardized feature foundation for subsequent intelligent evaluation. The pile foundation wall end coupling coefficient Ψ is calculated by integrating multi-dimensional information and compared with the preset pile foundation wall end coupling coefficient threshold Ψth to generate a judgment conclusion J, thereby realizing the automation and intelligence of the evaluation. The self-model is iteratively optimized according to the historical database HIS, so that the model can continuously improve the detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of pile foundation detection, and in particular to an intelligent detection method and system for pile foundation bearing capacity. Background Art

[0002] In civil engineering and infrastructure construction, pile foundations are an important form of foundation structure. Their primary function is to transfer the load of the superstructure to a stable bearing stratum deep within the ground, making them a critical link in ensuring the safety of engineering structures. Under various complex geological conditions, pile foundations, as a structural form capable of effectively transferring enormous loads to a stable bearing stratum deep underground, play an irreplaceable and core role. Therefore, how to scientifically and accurately assess whether a single pile foundation can bear the designed predetermined load, that is, to test its "bearing capacity," has become the cornerstone for ensuring the safety and stability of the entire project. This process is not only a key link in project quality acceptance, but also a lifeline for preventing future structural risks at the source. This has given rise to and developed the specific technical field of pile foundation bearing capacity testing.

[0003] However, the inherent shortcomings and deficiencies of traditional testing methods used to determine pile foundation bearing capacity are becoming increasingly apparent in the face of increasingly complex engineering requirements. For example, the industry-standard static load test method, while reliable, is cumbersome to prepare and requires a large number of counterweights or anchor reaction systems, resulting in extremely long testing cycles and high costs. Furthermore, only a small number of piles can typically be sampled for inspection within a project, resulting in extremely low inspection coverage and a significant number of project piles becoming "blind spots" in quality control. Another widely used high-strain dynamic testing method, while more efficient, relies heavily on empirical theories and parameter selection, such as the Case method. The accuracy of the test results is closely tied to the analyst's experience and subjective judgment, resulting in significant human uncertainty. Furthermore, both static load and high-strain methods typically only provide an overall bearing capacity value for the pile foundation, making it difficult to accurately distinguish and quantify how much of this bearing capacity comes from friction on the pile sidewalls and how much comes from the end support at the pile tip. This makes it difficult to effectively decouple "side friction" from "end bearing capacity."

[0004] The shortcomings of these traditional methods are further magnified when dealing with pile foundation projects in complex geological conditions. When the design requires the pile foundation to penetrate a soft soil layer and a medium-hard intermediate layer, ultimately resting on deep bedrock to form an "end-bearing pile," misjudgment is highly likely. Because the intermediate hard layer provides strong friction, traditional testing methods may indicate that the pile meets the design's total bearing capacity requirements. However, in reality, the pile foundation is "suspended" by the intermediate layer and does not actually contact the required bedrock, essentially becoming a "friction pile." This fundamental misjudgment of bearing properties poses a serious safety hazard. In the short term, this "pseudo-qualified" pile foundation may remain stable. However, over time and under long-term loads, the unstable bearing stratum may experience creep or uneven settlement, ultimately leading to cracking and tilting of the superstructure. In extreme conditions, it may even cause catastrophic structural failure, resulting in immeasurable loss of life and property. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for intelligent detection of pile foundation bearing capacity, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for intelligent detection of pile foundation bearing capacity, comprising the following steps:

[0007] S1, inject energy pulses into the pile foundation through the force sensor, collect the original signal package Ds through the high-frequency acceleration sensor, and obtain the pile foundation parameter set Pk from the construction record through the API interface;

[0008] The energy pulse injection method is as follows: the pile foundation is struck in a standardized manner by a free-falling hammer built into the force sensor, the energy pulse is injected into the pile foundation, and the time-course signal F(t) of the energy pulse is synchronously collected;

[0009] The original signal package Ds is collected by installing two high-frequency acceleration sensors symmetrically at the top of the pile to synchronously collect the pile-top acceleration time-history signal a(t) of the energy pulse. The impact force time-history signal F(t) and the pile-top acceleration time-history signal a(t) are packaged to generate the original signal package Ds, where t represents the time process variable from the moment the free-fall hammer performs a standardized impact on the pile foundation, starting from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse.

[0010] The pile body basic parameter set Pk includes: pile body length L, pile body diameter d, pile body cross-sectional area Ap, pile body material elastic modulus Em and pile body material density ρ;

[0011] S2. Calculate the velocity time-history signal v(t) based on the original signal packet Ds, and calculate the injected effective energy Ein by combining the original signal packet Ds and the velocity time-history signal v(t);

[0012] The velocity time history signal v(t) is calculated as follows: Based on the pile top acceleration time history signal a(t) in the original signal package Ds, the velocity time history signal v(t) is calculated through integration operation;

[0013] Among them, the calculation expression of the velocity time history signal v(t) is as follows:

[0014] ;

[0015] Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears;

[0016] The injection effective energy Ein is calculated as follows: the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds are multiplied, and the product is integrated to obtain the injection effective energy Ein.

[0017] Among them, the calculation expression of injected effective energy Ein is as follows:

[0018] ;

[0019] Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse;

[0020] S3. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), the pile wall friction dissipation energy Ef and the pile end effective reflection energy Ep are calculated to generate the energy decoupling vector Edec;

[0021] S4. Based on the injected effective energy Ein and the energy decoupling vector Edec, the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing are calculated to generate the characteristic vector Vx;

[0022] S5. Based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, the pile foundation wall end coupling coefficient Ψ is calculated and compared with a preset pile foundation wall end coupling coefficient threshold Ψth to generate a judgment conclusion J;

[0023] S6. Perform verification feedback actions according to the judgment conclusion J and collect feedback information dfa, construct a feedback data set dfb, add the feedback data set dfb to the historical database HIS, and perform iterative optimization based on the historical database HIS.

[0024] S31. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), a pile wall friction dissipation model is constructed. By performing basic physical analysis and calculation on the pile body basic parameter set Pk, the propagation velocity c of the stress wave in the pile body material and the theoretical round-trip time t of the reflected wave at the pile end are obtained. p , and multiply the pile cross-sectional area Ap by the product operation to obtain the pile mechanical impedance Zp, perform the square operation on the velocity time history signal v(t) and construct the time process variable t=0 to the time process variable t=t p The integral term of simulating the dynamic dissipation process is used to calculate the friction dissipation energy Ef of the pile wall;

[0025] The pile wall friction dissipation model expression is as follows:

[0026] ;

[0027] ;

[0028] Where c represents the propagation velocity of stress wave in the pile material, and the specific calculation method is: , t p It represents the theoretical round trip time of the first pile end reflection wave, which is calculated as follows: , The variable t represents the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse in the interval [0, t p ] is an integral operation, exp represents the natural exponential function, and Gs represents the equivalent shear modulus of the soil around the pile wall preset by professionals in this field.

[0029] Preferably, S32, based on the mechanical impedance Zp of the pile body and in combination with the velocity time history signal v(t), an effective reflection model of the pile end is constructed, the velocity time history signal v(t) is squared and a time process variable t=t is constructed. p When the time process variable t=△t, simulate the integral term of the received reflected energy pulse effect and calculate the effective reflected energy Ep at the pile end;

[0030] Among them, △t represents the moment when the reflected energy pulse is completely received for the first time. By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the theoretical round-trip time t of the first pile end reflected wave is p The time from the moment the reflected energy pulse is received to the moment the reflected energy pulse is completely received is △t', and the theoretical round trip time t of the first pile end reflection wave is t p The time △t of the first complete reception of the reflected energy pulse is obtained by adding them together. The effective reflection model expression of the pile end is as follows:

[0031] ;

[0032] ;

[0033] Where, The variable t representing the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse is related to the interval [t p , △t], tanh represents the hyperbolic tangent function, and Zr represents the equivalent mechanical impedance of the soil below the pile end preset by professionals in this field;

[0034] According to the friction dissipation energy Ef of the pile wall and the effective reflected energy Ep of the pile end, the energy decoupling vector Edec=[Ef, Ep] is generated.

[0035] Preferably, S4 includes S41;

[0036] S41. Based on the injected effective energy Ein and the energy decoupling vector Edec, a friction-dominated analysis algorithm and a pile-end bearing-dominated analysis algorithm are constructed;

[0037] By calculating the ratio of the pile wall friction dissipation energy Ef to the injected effective energy Ein, and the ratio of the pile wall friction dissipation energy Ef to the sum of the energy decoupling vector Edec, a friction-dominant analysis algorithm is constructed, and the pile wall friction-dominant index Xf is calculated.

[0038] The friction-dominated analysis algorithm is expressed as follows:

[0039] ;

[0040] In the formula, ε represents the regularization coefficient to prevent the denominator from being zero, specifically 1*10 -8 ;

[0041] By calculating the ratio of the effective reflected energy Ep at the pile end to the effective injected energy Ein, and the ratio of the effective reflected energy Ep at the pile end to the sum of the energy decoupling vector Edec, a pile end bearing dominant analysis algorithm is constructed, and the pile end bearing dominant index Xp is calculated.

[0042] The dominant analysis algorithm expression of pile end bearing is as follows:

[0043] ;

[0044] According to the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing, the characteristic vector Vx=[Xf,Xp] is generated.

[0045] Preferably, S5 includes S51 and S52;

[0046] S51. Based on the characteristic vector Vx, the injected effective energy Ein, and the energy decoupling vector Edec, a pile foundation-wall end coupling algorithm is constructed. By calculating the ratio of the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep to the injected effective capacity Ein, the characteristic vector Vx is quantitatively evaluated using the natural exponential function exp. Finally, the quality of the pile end effective reflected energy Ep is evaluated using the hyperbolic tangent function tanh, and the pile foundation-wall end coupling coefficient Ψ is calculated.

[0047] ;

[0048] Where λ represents the modulation coefficient of the preset quantitative evaluation of the wall end bearing behavior, β represents the modulation coefficient of the preset quality evaluation of the effective reflected energy Ep of the pile end, and v(t p , △t) represents the theoretical round trip time t of the first pile end reflection wave p The velocity time history signal within the time interval from the moment △t when the reflected energy pulse is completely received for the first time, max represents the maximum value.

[0049] Preferably, S52, comparing the pile foundation wall end coupling coefficient Ψ with a pile foundation wall end coupling coefficient threshold Ψth;

[0050] If the pile foundation wall end coupling coefficient Ψ ≥ the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be a qualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is qualified;

[0051] If the pile foundation wall end coupling coefficient Ψ is less than the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be an unqualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is abnormal.

[0052] Preferably, S6 includes S61;

[0053] S61. Execute verification feedback action according to the judgment conclusion J;

[0054] If the conclusion J is that the current pile foundation bearing capacity is qualified, the manual verification feedback action is performed according to the preset verification period T and the feedback information dfa is collected;

[0055] If the conclusion J is that the current pile foundation bearing capacity is abnormal, the staff is arranged to perform manual verification feedback actions on the current pile foundation and collect feedback information dfa;

[0056] The feedback information dfa is the detection correctness of the judgment conclusion J. After the judgment conclusion J performs the manual verification feedback action, if the detection result is correct, the feedback information dfa is recorded as true; if the detection result is wrong, the feedback information dfa is recorded as false;

[0057] Combine the feedback information dfa to construct a feedback data set dfb, and add the feedback data set dfb to the historical database HIS, where the feedback data set dfb = [Ds, Pk, Ein, Edec, Vx, Ψ, J, dfa];

[0058] Perform iterative optimization based on the historical database HIS. If the percentage of true feedback information dfa in the historical database HIS is greater than or equal to the expected accuracy rate true_P preset by the construction party, no iterative optimization is required. Otherwise, iterative optimization is performed.

[0059] Among them, the iterative optimization operation is to use the gradient descent supervised learning method to optimize and adjust the equivalent shear modulus Gs of the soil around the pile wall in the pile wall friction dissipation model and the equivalent mechanical impedance Zr of the soil below the pile end in the pile end effective reflection model. After executing the gradient descent supervised learning method once, the new feedback data set dfb_new is simulated and calculated according to the historical database HIS and a historical database copy HIS_new is generated. The new feedback information dfa_new in the new feedback data set dfb_new is tested. If the proportion of the new feedback information dfa_new in the historical database copy HIS_new being true is ≥ the expected accuracy true_P preset by the construction party, the historical database copy HIS_new is destroyed and this round of iterative optimization is ended. Otherwise, the gradient descent supervised learning method is continued.

[0060] An intelligent detection system for pile foundation bearing capacity, comprising a pile foundation data acquisition module, a pile foundation data processing module, an energy decoupling module, a bearing characteristic analysis module, a pile foundation wall end comprehensive evaluation module, and a decision execution and iterative optimization module;

[0061] The pile foundation data acquisition module performs standardized tapping on the pile foundation through the free-falling hammer built into the force sensor to inject energy pulses, and synchronously collects the impact force time-history signal F(t). By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the pile top acceleration time-history signal a(t) is synchronously collected, and the impact force time-history signal F(t) and the pile top acceleration time-history signal a(t) are packaged to generate an original signal package Ds. The pile body foundation parameter set Pk is obtained from the construction record through the API interface. The pile body foundation parameter set Pk includes the pile body length L, the pile body diameter d, the pile body cross-sectional area Ap, the pile body material elastic modulus Em, and the pile body material density ρ.

[0062] The pile foundation data processing module calculates the velocity time history signal v(t) based on the pile top acceleration time history signal a(t) in the original signal package Ds through integration operation;

[0063] Among them, the calculation expression of the velocity time history signal v(t) is as follows:

[0064] ;

[0065] Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears;

[0066] Perform product operation on the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds, integrate the product operation result, and calculate the injected effective energy Ein;

[0067] Among them, the calculation expression of injected effective energy Ein is as follows:

[0068] ;

[0069] Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse;

[0070] The energy decoupling module calculates the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep based on the pile body basic parameter set Pk and the velocity time history signal v(t), and generates the energy decoupling vector Edec;

[0071] The bearing characteristic analysis module calculates the pile wall friction dominant index Xf and the pile end bearing dominant index Xp based on the injected effective energy Ein and the energy decoupling vector Edec, and generates the characteristic vector Vx;

[0072] The comprehensive evaluation module of the pile foundation wall end calculates the pile foundation wall end coupling coefficient Ψ based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, and compares it with the preset pile foundation wall end coupling coefficient threshold Ψth to generate the judgment conclusion J;

[0073] The decision execution and iterative optimization module performs verification feedback actions according to the judgment conclusion J and collects feedback information dfa, constructs a feedback data set dfb, adds the feedback data set dfb to the historical database HIS, and performs iterative optimization based on the historical database HIS.

[0074] The present invention provides a method and system for intelligent detection of pile foundation bearing capacity, which has the following beneficial effects:

[0075] (1) Through standardized energy pulse injection and high-frequency synchronous acquisition, the original signal package Ds containing the impact force time-history signal F(t) and the pile top acceleration time-history signal a(t) is obtained, and combined with the pile body foundation parameter set Pk, a solid data foundation is laid for its core analysis. Its technical breakthrough is that it no longer regards the pile foundation as a whole "black box", but creatively adopts an energy decoupling model to quantitatively decompose the complex dynamic response into the pile wall friction dissipation energy Ef and the pile end effective reflection energy Ep, thereby being able to gain a deep insight into the real bearing mode of the pile foundation from the physical root. On this basis, a standardized characteristic vector Vx is further generated through an objective algorithm and fused into a decisive pile foundation wall end coupling coefficient Ψ. Finally, by introducing an iterative optimization mechanism driven by a historical database HIS, continuous self-improvement of detection accuracy is achieved. In summary, the efficiency and objectivity of detection are improved, and through its deep physical insight and adaptive learning ability, it provides an unprecedented, profound and reliable technical means to ensure the foundation safety of major projects.

[0076] (2) Through the standardized tapping method and the synchronous high-frequency acquisition of the tapping force time-history signal F(t) and the pile top acceleration time-history signal a(t), the input source of each test is guaranteed to be consistent and traceable. The resulting raw signal package Ds provides a high-quality, high-fidelity data foundation for all subsequent analyses. Next, the injected effective energy Ein of each tap is accurately calculated based on the measured raw signal package Ds and the calculated velocity time-history signal v(t). This eliminates the uncertainty caused by the difference between the equipment and the actual working conditions, and achieves accurate quantification of the "input". On this basis, by constructing the pile end effective reflection model and the pile wall friction dissipation model, the complex energy response is quantitatively decoupled into two independent physical components: the pile wall friction dissipation energy Ef and the pile end effective reflection energy Ep, and an energy decoupling vector Edec is generated, which enables us to fundamentally understand how the pile foundation bears load, not just whether it can bear load, and provides a clearer technical perspective for accurately revealing its true bearing characteristics.

[0077] (3) The energy decoupling vector Edec is converted into the dimensionless pile-wall friction dominant index Xf and the pile-end bearing dominant index Xp through a specific algorithm. This "indexing" process effectively shields the accidental fluctuations of the energy size of a single impact, allowing the generated characteristic vector Vx to be fairly and stably compared under different working conditions, laying a standardized characteristic foundation for subsequent intelligent evaluation. Subsequently, the characteristic vector Vx and other multi-dimensional information are comprehensively calculated into a decisive pile-wall coupling coefficient Ψ, and based on its comparison with the threshold Ψth, an objective and repeatable judgment conclusion J is automatically generated. This replaces the fuzzy process of traditional methods that relies on manual experience for interpretation, and realizes the automation and objectivity of decision-making. Finally, by introducing feedback information dfa to construct the feedback data set dfb, and using machine learning methods such as gradient descent to iteratively optimize the system model, the system can learn from each engineering verification and continuously self-correct and evolve. This closed-loop learning mechanism ensures that the system becomes increasingly adaptable to the geological conditions of a specific site, and its long-term detection accuracy continues to improve. This closed-loop learning mechanism ensures that the system can adapt to the geological conditions of a specific site and continuously improve its long-term detection accuracy, which is a feature not possessed by traditional static detection methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 This is a schematic diagram of the steps of an intelligent detection method for pile foundation bearing capacity according to the present invention;

[0079] Figure 2 This is a schematic diagram of a block diagram of an intelligent detection system for pile foundation bearing capacity according to the present invention;

[0080] Figure 3 Schematic diagram of the data processing flow for calculating the coupling coefficient Ψ of the pile foundation wall end;

[0081] Figure 4 It is the variation curve of the coupling coefficient Ψ at the pile foundation wall end. DETAILED DESCRIPTION

[0082] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0083] Example 1

[0084] The present invention provides an intelligent detection method for pile foundation bearing capacity. Figure 1 , including the following steps:

[0085] S1, inject energy pulses into the pile foundation through the force sensor, collect the original signal package Ds through the high-frequency acceleration sensor, and obtain the pile foundation parameter set Pk from the construction record through the API interface;

[0086] The energy pulse injection method is as follows: the pile foundation is struck in a standardized manner by a free-falling hammer built into the force sensor, the energy pulse is injected into the pile foundation, and the time-course signal F(t) of the energy pulse is synchronously collected;

[0087] The original signal package Ds is collected by installing two high-frequency acceleration sensors symmetrically at the top of the pile to synchronously collect the pile-top acceleration time-history signal a(t) of the energy pulse. The impact force time-history signal F(t) and the pile-top acceleration time-history signal a(t) are packaged to generate the original signal package Ds, where t represents the time process variable from the moment the free-fall hammer performs a standardized impact on the pile foundation, starting from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse.

[0088] The pile body basic parameter set Pk includes: pile body length L, pile body diameter d, pile body cross-sectional area Ap, pile body material elastic modulus Em and pile body material density ρ;

[0089] S2. Calculate the velocity time-history signal v(t) based on the original signal packet Ds, and calculate the injected effective energy Ein by combining the original signal packet Ds and the velocity time-history signal v(t);

[0090] The velocity time history signal v(t) is calculated as follows: Based on the pile top acceleration time history signal a(t) in the original signal package Ds, the velocity time history signal v(t) is calculated through integration operation;

[0091] Among them, the calculation expression of the velocity time history signal v(t) is as follows:

[0092] ;

[0093] Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears;

[0094] The injection effective energy Ein is calculated as follows: the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds are multiplied, and the product is integrated to obtain the injection effective energy Ein.

[0095] Among them, the calculation expression of injected effective energy Ein is as follows:

[0096] ;

[0097] Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse;

[0098] S3. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), the pile wall friction dissipation energy Ef and the pile end effective reflection energy Ep are calculated to generate the energy decoupling vector Edec;

[0099] S4. Based on the injected effective energy Ein and the energy decoupling vector Edec, the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing are calculated to generate the characteristic vector Vx;

[0100] S5. Based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, the pile foundation wall end coupling coefficient Ψ is calculated and compared with a preset pile foundation wall end coupling coefficient threshold Ψth to generate a judgment conclusion J;

[0101] S6. Perform verification feedback actions according to the judgment conclusion J and collect feedback information dfa, construct a feedback data set dfb, add the feedback data set dfb to the historical database HIS, and perform iterative optimization based on the historical database HIS.

[0102] In this embodiment, the pile foundation is subjected to standardized knocking by a free-falling hammer built into a force sensor to inject an energy pulse, and the force sensor is used to synchronously collect the knocking force time-history signal F(t) of the energy pulse. At the same time, two high-frequency acceleration sensors are installed at symmetrical positions on the pile top to synchronously collect the pile top acceleration time-history signal a(t) under the action of the energy pulse. The collected knocking force time-history signal F(t) and the pile top acceleration time-history signal a(t) will be packaged together to generate the original signal package Ds. This standardized knocking and synchronous high-frequency acquisition method ensures that the input source of each detection is consistent and traceable, providing a high-quality, high-fidelity data cornerstone for all subsequent analyses. In this process, the pile body foundation parameter set Pk is automatically obtained from the construction record through the API interface. This parameter set records in detail the key physical information including pile body length L, pile body diameter d, pile body cross-sectional area Ap, pile body material elastic modulus Em and pile body material density ρ. Next, during the data processing phase, the velocity time-history signal v(t) is precisely calculated through integration based on the pile top acceleration time-history signal a(t) in the original signal packet Ds. Finally, this velocity time-history signal v(t) is multiplied with the impact force time-history signal F(t) in the original signal packet Ds, and the product is integrated again to accurately calculate the injected effective energy Ein. This method accurately calculates the injected effective energy Ein for each impact based on the measured signal, eliminating the uncertainty caused by differences between the equipment and actual working conditions, achieving precise quantification of the input, and providing unprecedented insight into the fundamental bearing behavior of the pile foundation. Subsequently, based on the pile body foundation parameter set Pk and the velocity time-history signal v(t), a physical model is constructed to quantitatively decouple the complex energy response into two independent physical components: the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep. This generates the energy decoupling vector Edec, resolving the significant safety hazard caused by the inability of traditional methods to effectively distinguish bearing modes. This method clearly identifies the true bearing behavior of the pile foundation from its physical roots. Based on the injected effective energy Ein and the energy decoupling vector Edec, the dimensionless dominant index of pile-wall friction, Xf, and the dominant index of pile-end bearing, Xp, are calculated and combined into a characteristic vector Vx. These two intuitive indices can avoid the risk of misjudging a friction pile as a qualified end-bearing pile under complex geological conditions when it is actually suspended by an intermediate hard layer, thereby avoiding the catastrophic consequences such as uneven structural settlement and even long-term failure caused by such a misjudgment. Furthermore, by integrating the characteristic vector Vx, the injected effective energy Ein, and the energy decoupling vector Edec, a pile-base-wall coupling algorithm is used to calculate the final pile-base-wall coupling coefficient, Ψ. This coefficient is then compared with a preset threshold value, Ψth, to automatically generate a clear judgment conclusion, J.Finally, according to the judgment conclusion J, the corresponding verification feedback action is executed. After collecting the feedback information dfa, the feedback data set dfb is constructed and stored in the historical database HIS to drive the subsequent iterative optimization of the system model, providing a solid and reliable technical guarantee for the fundamental safety of the project.

[0103] Example 2

[0104] This embodiment is explained in Example 1, please refer to Figure 1 and Figure 3 ,Specifically: S3 includes S31 and S32;

[0105] S31. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), a pile wall friction dissipation model is constructed. By performing basic physical analysis and calculation on the pile body basic parameter set Pk, the propagation velocity c of the stress wave in the pile body material and the theoretical round-trip time t of the reflected wave at the pile end are obtained. p , and multiply the pile cross-sectional area Ap by the product operation to obtain the pile mechanical impedance Zp, perform the square operation on the velocity time history signal v(t) and construct the time process variable t=0 to the time process variable t=t p The integral term of simulating the dynamic dissipation process is used to calculate the friction dissipation energy Ef of the pile wall;

[0106] The pile wall friction dissipation model expression is as follows:

[0107] ;

[0108] ;

[0109] Where c represents the propagation velocity of stress wave in the pile material, and the specific calculation method is: , t p It represents the theoretical round trip time of the first pile end reflection wave, which is calculated as follows: , The variable t represents the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse in the interval [0, t p ] is an integral operation, exp represents a natural exponential function, and Gs represents the equivalent shear modulus of the soil around the pile wall preset by professionals in this field;

[0110] S32. Based on the mechanical impedance Zp of the pile body and the velocity time history signal v(t), an effective reflection model of the pile end is constructed. The velocity time history signal v(t) is squared and the time process variable t=t is constructed. p When the time process variable t=△t, simulate the integral term of the received reflected energy pulse effect and calculate the effective reflected energy Ep at the pile end;

[0111] Among them, △t represents the moment when the reflected energy pulse is completely received for the first time. By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the theoretical round-trip time t of the first pile end reflected wave is p The time from the moment the reflected energy pulse is received to the moment the reflected energy pulse is completely received is △t', and the theoretical round trip time t of the first pile end reflection wave is t p The time △t of the first complete reception of the reflected energy pulse is obtained by adding them together. The effective reflection model expression of the pile end is as follows:

[0112] ;

[0113] ;

[0114] Where, The variable t representing the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse is related to the interval [t p , △t], tanh represents the hyperbolic tangent function, and Zr represents the equivalent mechanical impedance of the soil below the pile end preset by professionals in this field;

[0115] According to the friction dissipation energy Ef of the pile wall and the effective reflected energy Ep of the pile end, the energy decoupling vector Edec=[Ef, Ep] is generated;

[0116] S4 includes S41;

[0117] S41. Based on the injected effective energy Ein and the energy decoupling vector Edec, a friction-dominated analysis algorithm and a pile-end bearing-dominated analysis algorithm are constructed;

[0118] By calculating the ratio of the pile wall friction dissipation energy Ef to the injected effective energy Ein, and the ratio of the pile wall friction dissipation energy Ef to the sum of the energy decoupling vector Edec, a friction-dominant analysis algorithm is constructed, and the pile wall friction-dominant index Xf is calculated.

[0119] The friction-dominated analysis algorithm is expressed as follows:

[0120] ;

[0121] In the formula, ε represents the regularization coefficient to prevent the denominator from being zero, specifically 1*10 -8 ;

[0122] By calculating the ratio of the effective reflected energy Ep at the pile end to the effective injected energy Ein, and the ratio of the effective reflected energy Ep at the pile end to the sum of the energy decoupling vector Edec, a pile end bearing dominant analysis algorithm is constructed, and the pile end bearing dominant index Xp is calculated.

[0123] The dominant analysis algorithm expression of pile end bearing is as follows:

[0124] ;

[0125] According to the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing, the characteristic vector Vx=[Xf,Xp] is generated;

[0126] The specific example of generating the energy decoupling vector Edec is as follows:

[0127] Pile foundation parameter set Pk = [L = 20m, d = 0.8m, Ap = 0.5027m 2 , Em=3.5*10 10 Pa, ρ=2500kg / m 3 ];

[0128] Injection effective energy Ein=30000J;

[0129] Equivalent shear modulus Gs=5*10 7 Pa, equivalent mechanical impedance Zr = 2.5 * 10 7 Ns / m;

[0130] The moment when the reflected energy pulse is completely received for the first time is △t=0.0112s;

[0131] Calculate the propagation velocity c of the stress wave in the pile material:

[0132] ;

[0133] Calculate the theoretical round trip time t of the first pile tip reflection wave p :

[0134] ;

[0135] Calculate the mechanical impedance Zp of the pile body:

[0136]

[0137] Calculate the friction dissipation energy Ef of the pile wall:

[0138] ;

[0139] Calculate the effective reflected energy Ep at the pile end:

[0140] ;

[0141] Energy decoupling vector Edec=[4500J, 22500J].

[0142] In this embodiment, a pile wall friction dissipation model is constructed based on the pile body basic parameter set Pk and the velocity time history signal v(t). By performing basic physical analysis on the pile body basic parameter set Pk, the propagation velocity c of the stress wave in the pile body material and the theoretical round-trip time t of the reflected wave at the pile end are calculated. p , and the pile body mechanical impedance Zp is calculated in combination with the pile body cross-sectional area Ap. Then, the velocity time history signal v(t) is squared and a time process variable from t=0 to t=t is constructed. p The integral term is used to simulate the dynamic dissipation process and finally calculate the pile wall friction dissipation energy Ef. Based on the pile body mechanical impedance Zp and the velocity time history signal v(t), the effective reflection model of the pile end is constructed. By performing a square operation on the velocity time history signal v(t) and constructing a pThe integral term from the moment Δt to the first complete reception of the reflected energy pulse is used to simulate the effect of receiving the reflected energy pulse, thereby calculating the effective reflected energy Ep at the pile end. Finally, the calculated pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep are combined to generate the energy decoupling vector Edec. By constructing the pile wall friction dissipation model and the pile end effective reflection model, the pile foundation is no longer considered an indivisible whole with regard to bearing capacity. Instead, the complex dynamic energy response is successfully quantitatively decomposed from its physical roots into two independent components with clear physical meanings: the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep. This energy decoupling provides a clear physical perspective for deep insight into the true bearing mode of the pile foundation, that is, how it can bear loads, rather than simply whether it can bear loads. Based on the calculated injected effective energy Ein and the energy decoupling vector Edec, friction-dominated analysis algorithms and pile end load-dominated analysis algorithms are constructed. By calculating the ratio of the pile wall friction dissipation energy Ef to the injected effective energy Ein, as well as the ratio of the pile wall friction dissipation energy Ef to the sum of the components of the energy decoupling vector Edec, a friction-dominant analysis algorithm is constructed, and the pile wall friction-dominant index Xf is calculated. Similarly, by calculating the ratio of the pile end effective reflected energy Ep to the injected effective energy Ein, as well as the ratio of the pile end effective reflected energy Ep to the sum of the components of the energy decoupling vector Edec, a pile end bearing-dominant analysis algorithm is constructed, and the pile end bearing-dominant index Xp is calculated. Finally, a characteristic vector Vx is generated based on the pile wall friction-dominant index Xf and the pile end bearing-dominant index Xp. The characteristic vector Vx quantitatively evaluates the relative proportions of the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep in the current situation, and calculates the dominant indices of the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep in the pile foundation. This indexation process eliminates the difference in absolute energy values ​​caused by different single-strike forces, more purely reflects the intrinsic bearing characteristics of the pile foundation, and enables the final generated characteristic vector Vx to be fairly and stably compared under different working conditions, laying a standardized feature foundation for the subsequent realization of automated and intelligent evaluation decisions.

[0143] Example 3

[0144] This embodiment is explained in Example 2, please refer to Figure 1 、 Figure 3 and Figure 4 ,Specifically: S5 includes S51 and S52;

[0145] S51. Based on the characteristic vector Vx, the injected effective energy Ein, and the energy decoupling vector Edec, a pile foundation-wall end coupling algorithm is constructed. By calculating the ratio of the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep to the injected effective capacity Ein, the characteristic vector Vx is quantitatively evaluated using the natural exponential function exp. Finally, the quality of the pile end effective reflected energy Ep is evaluated using the hyperbolic tangent function tanh, and the pile foundation-wall end coupling coefficient Ψ is calculated.

[0146] ;

[0147] Where λ represents the modulation coefficient of the preset quantitative evaluation of the wall end bearing behavior, β represents the modulation coefficient of the preset quality evaluation of the effective reflected energy Ep of the pile end, and v(t p , △t) represents the theoretical round trip time t of the first pile end reflection wave p The velocity time history signal in the time interval from △t to the moment when the reflected energy pulse is first completely received, max represents the maximum value;

[0148] S52, comparing the pile foundation wall end coupling coefficient Ψ with the pile foundation wall end coupling coefficient threshold Ψth;

[0149] If the pile foundation wall end coupling coefficient Ψ ≥ the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be a qualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is qualified;

[0150] If the pile foundation wall end coupling coefficient Ψ is less than the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be an unqualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is abnormal;

[0151] S6 includes S61;

[0152] S61. Execute verification feedback action according to the judgment conclusion J;

[0153] If the conclusion J is that the current pile foundation bearing capacity is qualified, the manual verification feedback action is performed according to the preset verification period T and the feedback information dfa is collected;

[0154] If the conclusion J is that the current pile foundation bearing capacity is abnormal, the staff is arranged to perform manual verification feedback actions on the current pile foundation and collect feedback information dfa;

[0155] The feedback information dfa is the detection correctness of the judgment conclusion J. After the judgment conclusion J performs the manual verification feedback action, if the detection result is correct, the feedback information dfa is recorded as true; if the detection result is wrong, the feedback information dfa is recorded as false;

[0156] Combine the feedback information dfa to construct a feedback data set dfb, and add the feedback data set dfb to the historical database HIS, where the feedback data set dfb = [Ds, Pk, Ein, Edec, Vx, Ψ, J, dfa];

[0157] Perform iterative optimization based on the historical database HIS. If the percentage of true feedback information dfa in the historical database HIS is greater than or equal to the expected accuracy rate true_P preset by the construction party, no iterative optimization is required. Otherwise, iterative optimization is performed.

[0158] Among them, the iterative optimization operation is to use the gradient descent supervised learning method to optimize and adjust the equivalent shear modulus Gs of the soil around the pile wall in the pile wall friction dissipation model and the equivalent mechanical impedance Zr of the soil below the pile end in the pile end effective reflection model. After executing the gradient descent supervised learning method once, the new feedback data set dfb_new is simulated and calculated according to the historical database HIS and a historical database copy HIS_new is generated. The new feedback information dfa_new in the new feedback data set dfb_new is tested. If the proportion of the new feedback information dfa_new in the historical database copy HIS_new being true is ≥ the expected accuracy true_P preset by the construction party, the historical database copy HIS_new is destroyed and this round of iterative optimization is ended. Otherwise, the gradient descent supervised learning method is continued.

[0159] In this embodiment, based on the characteristic vector Vx, the system starts the pile foundation wall end coupling algorithm to calculate a key comprehensive indicator - the pile foundation wall end coupling coefficient Ψ. The information of the three dimensions is multiplied and integrated: the first is the basic energy evaluation item that reflects the overall energy interaction efficiency ; Second, the core quantitative evaluation item of the difference in bearing characteristics between the pile end and the pile wall is amplified by using the natural exponential function ; Third, the signal quality evaluation item introduced to ensure the authenticity of the signal, which uses the hyperbolic tangent function to evaluate the peak velocity of the reflected wave at the pile end The calculated pile foundation-wall coupling coefficient Ψ is then objectively compared with the preset pile foundation-wall coupling coefficient threshold Ψth, automatically generating a clear judgment conclusion J, replacing the fuzzy interpretation process that relies on manual experience in traditional methods. Based on the judgment conclusion J, the system triggers differentiated manual verification actions and collects the most critical feedback information dfa on whether the detection is correct or not. The feedback information dfa, together with all the process data of the detection, will be fully packaged into a feedback data set dfb and stored in the historical database HIS. This ever-growing database is the core of the system's intelligence. Once the system detects that its accuracy on historical data is lower than the preset expected value, it will automatically trigger an iterative optimization program based on the gradient descent supervised learning method. This program does not adjust the evaluation algorithm, but instead accurately and automatically fine-tunes two core preset parameters in the physical model - the equivalent shear modulus Gs of the soil around the pile wall and the equivalent mechanical impedance Zr of the soil below the pile end, so that they are more in line with the actual geological conditions of the construction site. Not only does it achieve automated and objective assessment and decision-making through a multi-dimensional information fusion algorithm, it also builds a closed-loop adaptive learning mechanism based on real-world feedback. This ability to learn from the historical database (HIS) and iteratively optimize its own model using methods such as gradient descent transforms it from a static detection tool into an intelligent detection system that continuously learns, automatically adapts to specific site geological conditions, and continuously improves detection accuracy.

[0160] Example 4

[0161] An intelligent detection system for pile foundation bearing capacity, please refer to Figure 2 Specifically, it includes pile foundation data acquisition module, pile foundation data processing module, energy decoupling module, bearing characteristic analysis module, pile foundation wall end comprehensive evaluation module and decision execution and iterative optimization module;

[0162] The pile foundation data acquisition module performs standardized tapping on the pile foundation through the free-falling hammer built into the force sensor to inject energy pulses, and synchronously collects the impact force time-history signal F(t). By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the pile top acceleration time-history signal a(t) is synchronously collected, and the impact force time-history signal F(t) and the pile top acceleration time-history signal a(t) are packaged to generate an original signal package Ds. The pile body foundation parameter set Pk is obtained from the construction record through the API interface. The pile body foundation parameter set Pk includes the pile body length L, the pile body diameter d, the pile body cross-sectional area Ap, the pile body material elastic modulus Em, and the pile body material density ρ.

[0163] The pile foundation data processing module calculates the velocity time history signal v(t) based on the pile top acceleration time history signal a(t) in the original signal package Ds through integration operation;

[0164] Among them, the calculation expression of the velocity time history signal v(t) is as follows:

[0165] ;

[0166] Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears;

[0167] Perform product operation on the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds, integrate the product operation result, and calculate the injected effective energy Ein;

[0168] Among them, the calculation expression of injected effective energy Ein is as follows:

[0169] ;

[0170] Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse;

[0171] The energy decoupling module calculates the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep based on the pile body basic parameter set Pk and the velocity time history signal v(t), and generates the energy decoupling vector Edec;

[0172] The bearing characteristic analysis module calculates the pile wall friction dominant index Xf and the pile end bearing dominant index Xp based on the injected effective energy Ein and the energy decoupling vector Edec, and generates the characteristic vector Vx;

[0173] The comprehensive evaluation module of the pile foundation wall end calculates the pile foundation wall end coupling coefficient Ψ based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, and compares it with the preset pile foundation wall end coupling coefficient threshold Ψth to generate the judgment conclusion J;

[0174] The decision execution and iterative optimization module performs verification feedback actions according to the judgment conclusion J and collects feedback information dfa, constructs a feedback data set dfb, adds the feedback data set dfb to the historical database HIS, and performs iterative optimization based on the historical database HIS.

[0175] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent detection method for pile foundation bearing capacity, characterized by: The following steps are involved: S1, inject energy pulses into the pile foundation through the force sensor, collect the original signal package Ds through the high-frequency acceleration sensor, and obtain the pile foundation parameter set Pk from the construction record through the API interface; The energy pulse injection method is as follows: the pile foundation is struck in a standardized manner by a free-falling hammer built into the force sensor, the energy pulse is injected into the pile foundation, and the time-course signal F(t) of the energy pulse is synchronously collected; The original signal package Ds is collected by installing two high-frequency acceleration sensors symmetrically at the top of the pile to synchronously collect the pile-top acceleration time-history signal a(t) of the energy pulse. The impact force time-history signal F(t) and the pile-top acceleration time-history signal a(t) are packaged to generate the original signal package Ds, where t represents the time process variable from the moment the free-fall hammer performs a standardized impact on the pile foundation, starting from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse. The pile body basic parameter set Pk includes: pile body length L, pile body diameter d, pile body cross-sectional area Ap, pile body material elastic modulus Em and pile body material density ρ; S2. Calculate the velocity time-history signal v(t) based on the original signal packet Ds, and calculate the injected effective energy Ein by combining the original signal packet Ds and the velocity time-history signal v(t); The velocity time history signal v(t) is calculated as follows: Based on the pile top acceleration time history signal a(t) in the original signal package Ds, the velocity time history signal v(t) is calculated through integration operation; Among them, the calculation expression of the velocity time history signal v(t) is as follows: ； Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears; The injection effective energy Ein is calculated as follows: the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds are multiplied, and the product is integrated to obtain the injection effective energy Ein. Among them, the calculation expression of injected effective energy Ein is as follows: ； Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse; S3. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), the pile wall friction dissipation energy Ef and the pile end effective reflection energy Ep are calculated to generate the energy decoupling vector Edec; S4. Based on the injected effective energy Ein and the energy decoupling vector Edec, the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing are calculated to generate the characteristic vector Vx; S5. Based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, the pile foundation wall end coupling coefficient Ψ is calculated and compared with a preset pile foundation wall end coupling coefficient threshold Ψth to generate a judgment conclusion J; S6. Perform verification feedback actions according to the judgment conclusion J and collect feedback information dfa, construct a feedback data set dfb, add the feedback data set dfb to the historical database HIS, and perform iterative optimization based on the historical database HIS.

2. The intelligent detection method for pile foundation bearing capacity according to claim 1, characterized in that: S3 includes S31 and S32; S31. Based on the pile body basic parameter set Pk and the velocity time history signal v(t), a pile wall friction dissipation model is constructed. By performing basic physical analysis and calculation on the pile body basic parameter set Pk, the propagation velocity c of the stress wave in the pile body material and the theoretical round-trip time t of the reflected wave at the pile end are obtained. p , and multiply the pile cross-sectional area Ap by the product operation to obtain the pile mechanical impedance Zp, perform the square operation on the velocity time history signal v(t) and construct the time process variable t=0 to the time process variable t=t p The integral term of simulating the dynamic dissipation process is used to calculate the friction dissipation energy Ef of the pile wall; The pile wall friction dissipation model expression is as follows: ； ； Where c represents the propagation velocity of stress wave in the pile material, and the specific calculation method is: , t p It represents the theoretical round trip time of the first pile end reflection wave, which is calculated as follows: , The variable t represents the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse in the interval [0, t p ] is an integral operation, exp represents the natural exponential function, and Gs represents the equivalent shear modulus of the soil around the pile wall preset by professionals in this field.

3. The intelligent detection method for pile foundation bearing capacity according to claim 2, characterized in that: S32. Based on the mechanical impedance Zp of the pile body and the velocity time history signal v(t), an effective reflection model of the pile end is constructed. The velocity time history signal v(t) is squared and the time process variable t=t is constructed. p When the time process variable t=△t, simulate the integral term of the received reflected energy pulse effect and calculate the effective reflected energy Ep at the pile end; Among them, △t represents the moment when the reflected energy pulse is completely received for the first time. By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the theoretical round-trip time t of the first pile end reflected wave is p The time from the moment the reflected energy pulse is received to the moment the reflected energy pulse is completely received is △t', and the theoretical round trip time t of the first pile end reflection wave is t p The time △t of the first complete reception of the reflected energy pulse is obtained by adding them together. The effective reflection model expression of the pile end is as follows: ； ； Where, The variable t representing the time from the injection of the energy pulse into the pile foundation to the complete disappearance of the energy pulse is related to the interval [t p , △t], tanh represents the hyperbolic tangent function, and Zr represents the equivalent mechanical impedance of the soil below the pile end preset by professionals in this field; According to the friction dissipation energy Ef of the pile wall and the effective reflected energy Ep of the pile end, the energy decoupling vector Edec=[Ef, Ep] is generated.

4. The intelligent detection method for pile foundation bearing capacity according to claim 3, characterized in that: S4 package S41; S41. Based on the injected effective energy Ein and the energy decoupling vector Edec, a friction-dominated analysis algorithm and a pile-end bearing-dominated analysis algorithm are constructed; By calculating the ratio of the pile wall friction dissipation energy Ef to the injected effective energy Ein, and the ratio of the pile wall friction dissipation energy Ef to the sum of the energy decoupling vector Edec, a friction-dominant analysis algorithm is constructed, and the pile wall friction-dominant index Xf is calculated. The friction-dominated analysis algorithm is expressed as follows: ； In the formula, ε represents the regularization coefficient to prevent the denominator from being zero, specifically 1*10 -8 ; By calculating the ratio of the effective reflected energy Ep at the pile end to the effective injected energy Ein, and the ratio of the effective reflected energy Ep at the pile end to the sum of the energy decoupling vector Edec, a pile end bearing dominant analysis algorithm is constructed, and the pile end bearing dominant index Xp is calculated. The dominant analysis algorithm expression of pile end bearing is as follows: ； According to the dominant index Xf of pile wall friction and the dominant index Xp of pile end bearing, the characteristic vector Vx=[Xf,Xp] is generated.

5. The intelligent detection method for pile foundation bearing capacity according to claim 4, characterized in that: S5 includes S51 and S52; S51. Based on the characteristic vector Vx, the injected effective energy Ein, and the energy decoupling vector Edec, a pile foundation-wall end coupling algorithm is constructed. By calculating the ratio of the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep to the injected effective capacity Ein, the characteristic vector Vx is quantitatively evaluated using the natural exponential function exp. Finally, the quality of the pile end effective reflected energy Ep is evaluated using the hyperbolic tangent function tanh, and the pile foundation-wall end coupling coefficient Ψ is calculated. ； Where λ represents the modulation coefficient of the preset quantitative evaluation of the wall end bearing behavior, β represents the modulation coefficient of the preset quality evaluation of the effective reflected energy Ep of the pile end, and v(t p , △t) represents the theoretical round trip time t of the first pile end reflection wave p The velocity time history signal within the time interval from the moment △t when the reflected energy pulse is completely received for the first time, max represents the maximum value.

6. The intelligent detection method for pile foundation bearing capacity according to claim 5, characterized in that: S51, comparing the pile foundation wall end coupling coefficient Ψ with the pile foundation wall end coupling coefficient threshold Ψth; If the pile foundation wall end coupling coefficient Ψ ≥ the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be a qualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is qualified; If the pile foundation wall end coupling coefficient Ψ is less than the pile foundation wall end coupling coefficient threshold Ψth, the current pile foundation is judged to be an unqualified end-bearing pile, and the judgment conclusion J is that the current pile foundation bearing capacity is abnormal.

7. The intelligent detection method for pile foundation bearing capacity according to claim 6, characterized in that: S6 includes S61; S61. Execute verification feedback action according to the judgment conclusion J; If the conclusion J is that the current pile foundation bearing capacity is qualified, the manual verification feedback action is performed according to the preset verification period T and the feedback information dfa is collected; If the conclusion J is that the current pile foundation bearing capacity is abnormal, the staff is arranged to perform manual verification feedback actions on the current pile foundation and collect feedback information dfa; The feedback information dfa is the detection correctness of the judgment conclusion J. After the judgment conclusion J performs the manual verification feedback action, if the detection result is correct, the feedback information dfa is recorded as true; if the detection result is wrong, the feedback information dfa is recorded as false; Combine the feedback information dfa to construct a feedback data set dfb, and add the feedback data set dfb to the historical database HIS, where the feedback data set dfb = [Ds, Pk, Ein, Edec, Vx, Ψ, J, dfa]; Perform iterative optimization based on the historical database HIS. If the percentage of true feedback information dfa in the historical database HIS is greater than or equal to the expected accuracy rate true_P preset by the construction party, no iterative optimization is required. Otherwise, iterative optimization is performed. Among them, the iterative optimization operation is to use the gradient descent supervised learning method to optimize and adjust the equivalent shear modulus Gs of the soil around the pile wall in the pile wall friction dissipation model and the equivalent mechanical impedance Zr of the soil below the pile end in the pile end effective reflection model. After executing the gradient descent supervised learning method once, the new feedback data set dfb_new is simulated and calculated according to the historical database HIS and a historical database copy HIS_new is generated. The new feedback information dfa_new in the new feedback data set dfb_new is tested. If the proportion of the new feedback information dfa_new in the historical database copy HIS_new being true is ≥ the expected accuracy true_P preset by the construction party, the historical database copy HIS_new is destroyed and this round of iterative optimization is ended. Otherwise, the gradient descent supervised learning method is continued.

8. An intelligent detection system for pile foundation bearing capacity, applied to the intelligent detection method for pile foundation bearing capacity according to any one of claims 1 to 7, characterized in that: It includes pile foundation data acquisition module, pile foundation data processing module, energy decoupling module, bearing characteristics analysis module, pile foundation wall end comprehensive evaluation module and decision execution and iterative optimization module; The pile foundation data acquisition module performs standardized tapping on the pile foundation through the free-falling hammer built into the force sensor to inject energy pulses, and synchronously collects the impact force time-history signal F(t). By installing two high-frequency acceleration sensors at symmetrical positions on the pile top, the pile top acceleration time-history signal a(t) is synchronously collected, and the impact force time-history signal F(t) and the pile top acceleration time-history signal a(t) are packaged to generate an original signal package Ds. The pile body foundation parameter set Pk is obtained from the construction record through the API interface. The pile body foundation parameter set Pk includes the pile body length L, the pile body diameter d, the pile body cross-sectional area Ap, the pile body material elastic modulus Em, and the pile body material density ρ. The pile foundation data processing module calculates the velocity time history signal v(t) based on the pile top acceleration time history signal a(t) in the original signal package Ds through integration operation; Among them, the calculation expression of the velocity time history signal v(t) is as follows: ； Where t1 represents the upper limit of integration, which is the time from the injection of the energy pulse into the pile foundation to the moment when the energy pulse completely disappears; Perform product operation on the impact force time history signal F(t) and the velocity time history signal v(t) in the original signal package Ds, integrate the product operation result, and calculate the injected effective energy Ein; Among them, the calculation expression of injected effective energy Ein is as follows: ； Where t2 represents the upper limit of integration, specifically the time from the start of injecting the energy pulse into the pile foundation to the end of the injection of the energy pulse; The energy decoupling module calculates the pile wall friction dissipation energy Ef and the pile end effective reflected energy Ep based on the pile body basic parameter set Pk and the velocity time history signal v(t), and generates the energy decoupling vector Edec; The bearing characteristic analysis module calculates the pile wall friction dominant index Xf and the pile end bearing dominant index Xp based on the injected effective energy Ein and the energy decoupling vector Edec, and generates the characteristic vector Vx; The comprehensive evaluation module of the pile foundation wall end calculates the pile foundation wall end coupling coefficient Ψ based on the characteristic vector Vx, the injected effective energy Ein and the energy decoupling vector Edec, and compares it with the preset pile foundation wall end coupling coefficient threshold Ψth to generate the judgment conclusion J; The decision execution and iterative optimization module performs verification feedback actions according to the judgment conclusion J and collects feedback information dfa, constructs a feedback data set dfb, adds the feedback data set dfb to the historical database HIS, and performs iterative optimization based on the historical database HIS.

Images