Pull-out test method for rock anchor foundation of mountain photovoltaic power station

Abstract

This invention discloses a data processing technology field, specifically a method for pull-out testing of rock anchor foundations in mountainous photovoltaic power stations. The method includes obtaining initial geological parameters of the rock anchor foundation to be tested in the mountainous photovoltaic power station; using the initial geological parameters as constraints, performing stepped loading on the anchor body of the rock anchor foundation to be tested through a preset stepped incremental loading strategy to obtain a preload response time series; using the preload response time series as a prerequisite, conducting a closed-loop incremental static load pull-out test according to a preset static load step; inverting the anchor-rock interface parameters based on the joint parameters of the static load-displacement curve and time-spectrum characteristics; constructing a bearing capacity function based on posterior samples of the inverted anchor-rock interface parameters; obtaining the bearing capacity distribution through Monte Carlo sampling; and calculating a reliability index based on the bearing capacity distribution and load. Through the above technical solution, by inverting the joint parameters of the static load-displacement curve and time-spectrum characteristics and quantifying their uncertainty, the accuracy and reliability of the reliability index can be improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method for testing the pull-out resistance of rock anchor foundations in mountain photovoltaic power stations. Background Technology

[0002] As the photovoltaic energy industry expands into mountainous and hilly areas, rock anchor foundations have become the core foundation form for fixing the supports of mountain photovoltaic power stations due to their suitability for shallow rock strata and undulating terrain. Their pull-out bearing capacity directly determines the long-term stability and operational safety of the power station structure, making precise and efficient pull-out testing technology crucial.

[0003] Existing testing methods are limited and have strong limitations. Traditional static load tests can only obtain the macroscopic load-displacement response and cannot capture the microscopic signs of initial damage at the anchor-rock interface. They are prone to overlooking hidden problems such as interface debonding and local cracks. Moreover, they are mostly destructive tests and are not suitable for large-scale engineering testing scenarios. Summary of the Invention

[0004] The purpose of this invention is to provide a pull-out test method for rock anchor foundations of mountain photovoltaic power stations, so as to solve the technical problems existing in related technologies.

[0005] To achieve the above objectives, the present invention provides a method for testing the pull-out resistance of rock anchor foundations for mountain photovoltaic power stations, comprising:

[0006] Obtain the initial geological parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station;

[0007] Using initial geological parameters as constraints, the anchor body of the rock anchor foundation under test is subjected to stepped loading through a preset stepped incremental loading strategy. At the same time, the initial contact and time-varying behavior of the rock anchor interface of the rock anchor foundation under test are obtained to obtain the preload response time series.

[0008] Using the preload response time series as a prerequisite, a closed-loop incremental static load pull-out test was conducted on the rock anchor foundation under test according to the preset static load stage to determine the static load force-displacement curve and time spectrum characteristics.

[0009] The anchor-rock interface parameters of the rock anchor foundation under test are obtained by inverting the joint parameters of the static load displacement curve and time-frequency characteristics.

[0010] Based on the posterior samples of the inverted anchor rock interface parameters, a bearing capacity function is constructed. The posterior samples are then substituted into the bearing capacity function using the Monte Carlo sampling method to calculate the bearing capacity distribution. The reliability index is then calculated based on the bearing capacity distribution and the load.

[0011] Optionally, obtaining the initial geological parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station includes:

[0012] Point cloud data of the rock anchor foundation to be tested was obtained by a 3D laser scanner, and a digital elevation model of the rock anchor foundation to be tested in the mountain photovoltaic power station was constructed based on the point cloud data. The terrain parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station were then extracted from the digital elevation model.

[0013] The near-surface geological bedding and joint distribution characteristics of the rock anchor foundation under test were obtained by shallow ground radar and shallow core sampling methods. Rock samples were collected simultaneously to carry out unconfined compressive strength tests to obtain the UCS interval.

[0014] Temperature and crack data are obtained by monitoring the temperature and strain profile data of the rock anchor foundation under test using distributed optical fiber sensors.

[0015] Topographic parameters, near-surface geological bedding and joint distribution characteristics, UCS intervals, temperature data, and fracture data were determined as initial geological parameters.

[0016] Optionally, the step loading process involves applying a preset stepped incremental loading strategy to the anchor body of the rock anchor foundation under test, while simultaneously acquiring the initial contact and time-varying behavior of the anchor-rock interface of the rock anchor foundation under test, to obtain a preload response time series, including:

[0017] According to the preset stepped incremental loading strategy, the anchor body of the rock foundation under test is loaded in steps using hydraulic jacks. At the same time, strain data, load data, displacement data and temperature data of the initial contact and time-varying behavior of the rock anchor interface of the rock foundation under test are collected during each step incremental loading process.

[0018] The strain data, load data, displacement data, and temperature data are corrected to obtain the preload response time series.

[0019] Optionally, the step of using the preload response time series as a prerequisite, conducting a closed-loop incremental static load pull-out test on the rock anchor foundation under test according to the preset static load order, and determining the static load force-displacement curve and time-frequency characteristics includes:

[0020] Based on the low-delay real-time flow in the preload response time series, dynamic compensation is performed on each static load step in the closed-loop incremental static load pull-out test using the displacement-force dual closed-loop control method to obtain the static load force-displacement curve.

[0021] Short pulse and linear frequency sweep excitation signals are superimposed on the rock anchor foundation under test in the closed-loop incremental static load pull-out test, which is at the target static load level. Acceleration and strain signals are acquired at the same time, and short-time Fourier transform or wavelet transform is performed on the acceleration and strain signals to obtain the time-frequency characteristics.

[0022] Optionally, the step of dynamically compensating each static load stage in the closed-loop incremental static load pull-out test based on the low-delay real-time flow in the preload response time series using a displacement-force dual closed-loop control method to obtain the static load force-displacement curve includes:

[0023] For each static load stage in the closed-loop incremental static load pull-out test, the displacement amount is obtained by dynamically compensating the real-time displacement component in the low-delay real-time flow and the a priori displacement baseline of the geological and environmental parameters in the initial geological parameters through the displacement-force dual closed-loop control method.

[0024] The displacement is processed by time differentiation to obtain the displacement rate. Anomaly protection is determined based on the displacement rate. Paired data of force and displacement generated throughout the process are obtained and plotted as static load-displacement curve.

[0025] Optionally, the step of inverting the anchor-rock interface parameters of the rock anchor foundation under test based on the static load-displacement curve and time-frequency characteristic joint parameters to obtain the inverted anchor-rock interface parameters includes:

[0026] Based on the static load displacement curve and time spectrum characteristics, a forward model of the anchor-rock interface parameters of the rock anchor foundation to be tested is constructed.

[0027] The parameters of the anchor-rock interface are obtained by performing joint parameter inversion on the forward model using Bayesian inversion or physical constraint-based machine learning regression.

[0028] The above technical solution uses initial geological parameters as constraints and employs a pre-defined stepped incremental loading strategy to apply stepped loading to the anchor body of the rock anchor foundation under test. Simultaneously, it acquires the initial contact and time-varying behavior of the anchor-rock interface, obtaining a preload response time series. This preload response time series is then used as a prerequisite for conducting closed-loop incremental static load pull-out tests, thereby determining the static load-displacement curve and time-frequency spectrum characteristics. This allows for the quantitative inversion of anchor-rock interface parameters without extensive destructive sampling, enabling probabilistic prediction of the remaining bearing capacity. It not only accurately measures the ultimate bearing capacity of the rock anchor foundation but also detects early signs of interface damage through time-frequency response analysis. By combining the static load-displacement curve and time-frequency spectrum characteristics for parameter inversion, the accuracy and reliability of reliability indicators can be improved.

[0029] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0030] Figure 1 This is a schematic diagram illustrating a pull-out test method for a rock anchor foundation of a mountain photovoltaic power station according to an exemplary embodiment of the present invention. Detailed Implementation

[0031] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, so as to provide a better understanding of the concept of the present invention, the technical problem solved, the technical features constituting the technical solution, and the technical effects brought about.

[0032] The inventors discovered that traditional static load tests can only obtain the macroscopic load-displacement response and cannot capture the microscopic signs of initial damage at the anchor rock interface, easily overlooking hidden problems such as interface debonding and local cracks. Traditional static load tests are mostly destructive tests and are not suitable for large-scale engineering testing scenarios.

[0033] In view of this, the present invention provides a pull-out test method for rock anchor foundations of mountain photovoltaic power stations to solve the technical problems existing in the above-mentioned related technologies.

[0034] like Figure 1 As shown, Figure 1 This is a schematic diagram illustrating a pull-out test method for a rock anchor foundation of a mountain photovoltaic power station according to an exemplary embodiment of the present invention. (Refer to...) Figure 1 The method includes;

[0035] S101: Obtain the initial geological parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station;

[0036] S102: Using initial geological parameters as constraints, the anchor body of the rock anchor foundation under test is subjected to stepped loading through a preset stepped incremental loading strategy. At the same time, the initial contact and time-varying behavior of the rock anchor interface of the rock anchor foundation under test are obtained to obtain the preload response time series.

[0037] S103: Using the preload response time series as a prerequisite, a closed-loop incremental static load pull-out test is conducted on the rock anchor foundation under test according to the preset static load step to determine the static load force-displacement curve and time spectrum characteristics.

[0038] S104: The anchor-rock interface parameters of the rock anchor foundation under test are inverted based on the static load displacement curve and time-frequency characteristic combined parameters.

[0039] S105: Based on the inverted anchor rock interface parameters, a posterior sample is sampled to construct a bearing capacity function. The posterior sample is then substituted into the bearing capacity function using the Monte Carlo sampling method to calculate the bearing capacity distribution. The reliability index is then calculated based on the bearing capacity distribution and the load.

[0040] The above technical solution uses initial geological parameters as constraints and employs a pre-defined stepped incremental loading strategy to apply stepped loading to the anchor body of the rock anchor foundation under test. Simultaneously, it acquires the initial contact and time-varying behavior of the anchor-rock interface, obtaining a preload response time series. This preload response time series is then used as a prerequisite for conducting closed-loop incremental static load pull-out tests, thereby determining the static load-displacement curve and time-frequency spectrum characteristics. This allows for the quantitative inversion of anchor-rock interface parameters without extensive destructive sampling, enabling probabilistic prediction of the remaining bearing capacity. It not only accurately measures the ultimate bearing capacity of the rock anchor foundation but also detects early signs of interface damage through time-frequency response analysis. By combining the static load-displacement curve and time-frequency spectrum characteristics for parameter inversion, the accuracy and reliability of reliability indicators can be improved.

[0041] To enable those skilled in the art to better understand the pull-out test method for rock anchor foundations of mountain photovoltaic power stations provided by the present invention, the above steps are illustrated in detail below.

[0042] For example, a mountain photovoltaic power station can be a photovoltaic power station built in mountainous terrain or areas with exposed bedrock / shallow overburden. The anchor rock foundation to be tested can be a structure where holes are drilled in the bedrock, anchor bolts are inserted, and anchoring mortar is poured in. Relying on the bond between the anchor bolts and the bedrock, and the shear bearing capacity of the rock mass, it resists wind loads, snow loads, uplift forces, and horizontal thrusts from the photovoltaic support, transferring the load of the superstructure to the stable bedrock. Initial geological parameters can be the geological parameters before conducting the uplift test. In this embodiment of the invention, the initial geological parameters of the rock anchor to be tested are first obtained and used in subsequent processing steps.

[0043] In possible ways, obtaining the initial geological parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station includes:

[0044] Point cloud data of the rock anchor foundation to be tested was obtained by a 3D laser scanner, and a digital elevation model of the rock anchor foundation to be tested in the mountain photovoltaic power station was constructed based on the point cloud data. The terrain parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station were then extracted from the digital elevation model.

[0045] Near-surface layering and joint distribution characteristics of the rock anchor foundation under test were obtained by shallow ground radar and shallow core sampling methods. Rock samples were collected simultaneously to carry out unconfined compressive strength tests to obtain the UCS interval.

[0046] Temperature and crack data are obtained by monitoring the temperature and strain profile data of the rock anchor foundation under test using distributed optical fiber sensors.

[0047] Topographic parameters, near-surface layer stratification characteristics, joint distribution characteristics, UCS intervals, temperature data, and fracture data were determined as the initial geological parameters.

[0048] It should be understood that a 3D laser scanner can be used to quickly acquire massive amounts of 3D coordinate points on the surface of an object or scene through laser ranging and angle encoding. In this embodiment of the invention, point cloud data of the rock anchor foundation to be tested is acquired using a 3D laser scanner, and a digital elevation model is constructed based on the point cloud data. The 3D laser scanner can be a RIEGL VZ-400, and the digital elevation model can be a 3D spatial data model that digitally discretizes the elevation morphology of the land surface using limited terrain elevation data. Then, the terrain parameters of the rock anchor foundation to be tested can be extracted from the digital elevation model. These terrain parameters can be numerical indicators that quantitatively express the surface morphology, spatial geometric features, and hydrodynamic characteristics. In this embodiment of the invention, the terrain parameters may include slope parameters, slope curvature, and relative elevation differences.

[0049] Shallow ground-penetrating radar (SPR) transmits high-frequency electromagnetic waves into the ground and receives reflected signals. Based on the reflection characteristics of electromagnetic waves at different rock and soil bodies or media interfaces, it interprets shallow geological information such as the shallow strata structure, rock mass integrity, joint and fracture distribution, cavities or weak zones, grout density, and anchorage interface state around the rock anchor foundation without drilling or damaging the rock mass. Shallow core sampling involves drilling shallow-depth boreholes near the rock anchor foundation or in the test area using a small drilling rig. In-situ rock and soil core samples are obtained through drilling, allowing direct observation of the core samples' lithology, weathering degree, fracture development, cementation / grouting quality, and stratigraphic interfaces. Laboratory mechanical tests can also be conducted on the core samples. Near-surface layer stratification and joint distribution parameters are numerical indicators that quantitatively describe the layered structure and spatial development of joints and fractures in the near-surface rock and soil mass of a site, based on data obtained through near-surface exploration methods such as shallow ground-penetrating radar and shallow core sampling. Near-surface layer stratification characteristics can be used to describe the vertical sequence and interface characteristics of near-surface rock and soil masses at a site. Joint distribution characteristics can be used to describe the development and spatial distribution of joints and fissures within near-surface rock masses. In this embodiment of the invention, the near-surface layer stratification characteristics and joint distribution characteristics can be obtained by using shallow ground-penetrating radar and shallow-hole coring methods to test the rock anchor foundation under test. During the shallow-hole coring test, rock samples can also be collected simultaneously for unconfined compressive strength testing to obtain the UCS (Unconfined Compressive Strength) range. That is, in the same operational step of obtaining in-situ near-surface rock cores using the shallow-hole coring method, representative intact rock core samples are selected and unconfined compressive strength tests are conducted in the laboratory according to standard specifications. Simultaneously, the uniaxial unconfined compressive strength index of the rock samples is obtained, and the UCS range is derived.

[0050] Subsequently, temperature and strain profile data of the rock anchor foundation under test can be monitored using distributed optical fiber sensors to obtain temperature and fracture data. The distributed optical fiber sensors can be OFDR / OTDR type with a spatial resolution of 0.5m; however, this embodiment of the invention does not specify a particular type. Temperature data must be collected continuously for at least 48 hours to establish a temperature baseline. Then, topographic parameters, near-surface geological bedding and joint distribution characteristics, UCS intervals, and the temperature and fracture data are determined as initial geological parameters.

[0051] In this embodiment of the invention, all devices and the acquisition system are synchronized with GPS or a master clock. The output fields include point cloud / profile, crack statistics, temperature baseline, and displacement / strain baseline, which can constitute the input for subsequent data correction and inversion priors.

[0052] For example, a pre-set stepped incremental loading strategy can be a standardized loading control scheme for static load pull-out or compressive strength tests in geotechnical engineering, where fixed load increments, loading levels, duration of each level, and termination criteria are pre-defined according to the test plan, and the load is applied step by step. The initial contact and time-varying behavior of the anchor-rock interface can refer to the deformation and deterioration behavior of the anchor bolt or anchoring grout interface with the bedrock, evolving over time under long-term load, environmental disturbance, and creep, from the initial mechanical state at the start of loading. In this embodiment of the invention, using initial geological parameters as constraints, the anchor body can undergo graded preload response under controlled conditions. Simultaneously, the initial contact and deformation of the anchor-rock interface of the rock anchor foundation under test are obtained during the test, and the preload response time series is obtained based on the initial contact and time-varying behavior. In the specific implementation process, the anchor body of the rock anchor foundation under test is preloaded in stages according to a preset stepped incremental loading strategy. Simultaneously, the preload response time series is recorded using a synchronous sampling method, with the sampling frequency set to simultaneously capture static and dynamic information. By preloading the anchor body in stages and recording the initial contact and time-varying behavior of the anchor-rock interface, time series data during the preloading process are collected. This event series data can include strain data, load data, displacement data, and temperature data. The recording and analysis of this data can reveal the interaction between the anchor body and the rock mass, providing preliminary information about the interface bonding state, frictional characteristics, etc., to obtain the response characteristics of the anchor body under different loading conditions, providing a basis for subsequent static load tests.

[0053] In one possible manner, the anchor body of the rock anchor foundation under test is subjected to stepped loading processing through a preset stepped incremental loading strategy, while the initial contact and time-varying behavior of the anchor-rock interface of the rock anchor foundation under test are obtained to obtain the preload response time series, including:

[0054] According to the preset stepped incremental loading strategy, the anchor body of the rock foundation under test is loaded in steps using hydraulic jacks. At the same time, strain data, load data, displacement data and temperature data of the initial contact and time-varying behavior of the rock anchor interface of the rock foundation under test are collected during each step incremental loading process.

[0055] The strain data, load data, displacement data, and temperature data are corrected to obtain the preload response time series.

[0056] It should be understood that the preset stepped incremental loading strategy can have increments of 20→40→60→80→100 kN, with each increment holding for 5–10 minutes, and can be adjusted according to the engineering design. Furthermore, the anchor is subjected to stepped loading using a hydraulic jack equipped with a standardized load sensor. The load sensor in the hydraulic jack has an accuracy ≤0.5%FS and is model HBM. Simultaneously, an LVDT (Linear Variable Differential Transformer) and a head displacement gauge are installed on top of the anchor body, and DFOS (Distributed Fiber Optic Sensing) is used to record the strain profile along the anchor hole. In the dynamic channel, an accelerometer samples at ≥2000 Hz, and in the static channel, it records at 1 Hz. The LVDT is a high-precision displacement sensor with a range of 0–30 mm and a resolution of 0.01 mm.

[0057] The on-site main control system performs clock synchronization, preliminary temperature compensation, zero drift rejection, and filtering. This system is a PLC (Programmable Logic Controller) / real-time DAQ (Data Acquisition). Zero drift rejection and filtering can be static path low-pass filtering or dynamic path band-pass / notch filtering. The processed data is divided into a low-latency real-time stream and a high-resolution batch stream. Strain, load, displacement, and temperature data are recorded at each stage, and the strain, load, displacement, and temperature data are corrected to obtain the preload response time series.

[0058] For example, the preload response time series can include data such as initial response, overall slope deformation, and environmental impact. In subsequent processing, a closed-loop incremental static load pull-out test can be performed on the rock anchor foundation under test according to the preset static load stage to obtain the static load-displacement curve and time-frequency characteristics. In this embodiment of the invention, the closed-loop incremental static load pull-out test can be conducted under controlled conditions, allowing for limit tests and avoiding damage to the rock anchor foundation under test. The static load-displacement curve is a relationship curve plotted with the applied static load value on the vertical axis and the measured displacement value after dynamic compensation at the corresponding time on the horizontal axis. It is the core test result curve used to characterize the bearing and deformation characteristics of the anchor-rock interface. The time-frequency characteristics are comprehensive features of the signal amplitude, energy, and frequency components changing over time after performing time-domain and frequency-domain analysis on the original high-frequency vibration / impact signals collected by dynamic channels such as accelerometers and DFOS during the test. These features can also be used to identify dynamic events such as micro-fractures in the rock anchor or rock mass, transient interface slippage, impacts, and vibration modes.

[0059] In one possible approach, the method of using the preload response time series as a prerequisite, conducting a closed-loop incremental static load pull-out test on the rock anchor foundation under test according to a preset static load step, and determining the static load-displacement curve and time-frequency characteristics includes:

[0060] Based on the low-delay real-time flow in the preload response time series, dynamic compensation is performed on each static load step in the closed-loop incremental static load pull-out test using the displacement-force dual closed-loop control method to obtain the static load force-displacement curve.

[0061] Short pulse and linear frequency sweep excitation signals are superimposed on the rock anchor foundation under test in the closed-loop incremental static load pull-out test, which is at the target static load level. Acceleration and strain signals are acquired at the same time, and short-time Fourier transform or wavelet transform is performed on the acceleration and strain signals to obtain the time-frequency characteristics.

[0062] It should be understood that low-latency real-time stream refers to a continuous data stream in the rock anchor static load test control system, characterized by extremely low transmission and processing latency throughout the entire process from sensor raw signal acquisition, transmission, and preprocessing to closed-loop control / online analysis, meeting the requirements of hard real-time performance. Dynamic compensation refers to the real-time correction algorithm and processing flow in the rock anchor static load test, which removes error components such as environmental interference, overall slope topographic displacement, equipment baseline drift, and structural rigid body displacement from the total displacement measured by the sensors, thus obtaining the pure mechanical deformation or true slip of the anchor-rock interface. In this embodiment of the invention, when determining the static load force-displacement curve, dynamic compensation can be performed on each static load stage in the closed-loop incremental static load pull-out test using the displacement-force dual closed-loop control method based on the low-latency real-time stream in the preload response time series.

[0063] When determining the time-frequency characteristics, short-pulse and linearly swept excitation signals can be simultaneously superimposed onto the closed-loop incremental static load pull-out test, and acceleration and strain signals can be acquired. The time-frequency characteristics can then be obtained based on these acceleration and strain signals. The short-pulse excitation signal can be a controllable excitation input signal with extremely short duration, narrow time domain width, and instantaneously concentrated energy; the linearly swept excitation signal can be a sinusoidal excitation signal whose frequency changes linearly and continuously with time.

[0064] To determine the time-frequency characteristics, a combination of short-pulse and linearly swept excitation signals is used. Short-time Fourier transforms or wavelet transforms are performed on the acquired acceleration and strain signals to obtain the time-frequency characteristics. Small-amplitude dynamic disturbances are superimposed onto the static load test to capture potential nonlinear characteristics and damage at the interface. Time-frequency analysis of dynamic response signals such as acceleration and strain can effectively identify early signs of interface damage, such as interface adhesion failure or reduced friction. This dynamic disturbance can detect potential failure risks in advance, thereby improving the accuracy and reliability of the test.

[0065] In the specific operation process, at the target static load level of the closed-loop incremental static load pull-out test, the sensitive section is identified by superimposed controlled short-term dynamic disturbance, and the disturbance response is analyzed in time and frequency to extract damage-sensitive features. Then, the time-frequency characteristics are obtained based on the damage-sensitive features.

[0066] The specific process is as follows: A short pulse and narrow-band frequency sweep are applied to the target static load level using an instrumented impact hammer or a low-power electromagnetic / electric vibration exciter. The impact hammer model can be PCB Piezotronics 086C. The short pulse can last 10–100 ms, with the peak amplitude limited to no more than 2–5% of the load at that level. The narrow-band frequency sweep can be 10–500 Hz and last for several seconds. The high-frequency bands of the dynamic accelerometer and DFOS are simultaneously sampled at ≥5 kHz. The time-frequency spectrum E(t,f) is calculated offline using short-time Fourier transform and continuous wavelet transform in MATLAB or Python. Peak frequency, bandwidth, energy ratio, second harmonic ratio, and other indicators are extracted and combined to form a spectral damage index, outputting the time-frequency spectrum characteristics.

[0067] In one possible manner, the static load force-displacement curve is obtained by dynamically compensating each static load stage in the closed-loop incremental static load pull-out test using a displacement-force dual closed-loop control method based on the low-delay real-time flow in the preload response time series, including:

[0068] For each static load stage in the closed-loop incremental static load pull-out test, the displacement amount is obtained by dynamically compensating the real-time displacement component in the low-delay real-time flow and the a priori displacement baseline of the geological and environmental parameters in the initial geological parameters through the displacement-force dual closed-loop control method.

[0069] The displacement is processed by time differentiation to obtain the displacement rate. Anomaly protection is determined based on the displacement rate. Paired data of force and displacement generated throughout the process are obtained and plotted as static load-displacement curve.

[0070] It should be understood that the closed-loop incremental static load pull-out test employs a displacement-force dual closed-loop control, and performs real-time compensation for slope baseline displacement and temperature at each static load stage to obtain more comparable and repeatable static load-displacement curves. The pull-out performance of the anchor body is tested through incremental loading. After each increase in force, a constant load is maintained and the anchor body displacement is recorded to obtain the static load-displacement curve. The closed-loop control method ensures the accuracy of the loading process and provides real-time compensation for slope deformation.

[0071] In the specific implementation process, a closed-loop incremental static load pull-out test of displacement-force dual closed-loop control is carried out under the condition of low-delay real-time flow correction to obtain the static load force-displacement curve.

[0072] The low-latency real-time stream is input into the real-time controller. The controller executes a force / displacement dual closed-loop strategy and performs dynamic compensation at each stage based on the real-time displacement component of the low-latency real-time stream and the displacement baseline of the geological and environmental priors in the initial geological parameters. That is, the formula for calculating the displacement is Uref,comp=Uref-[Ubase(t)﹣ Ubase(t0)], where Uref,comp is the displacement, Uref is the original experimental reference displacement, Ubase(t) is the reference displacement at time t, and Ubase(t0) is the reference displacement at time t0. Example of force increment step size ΔF=10 kN / step, load holding for 5 min. The controller is either a Siemens S7 + Real-time IO (Siemens S7 series PLC + real-time input / output module) or a National Instruments controller. The displacement is then processed by time differentiation to obtain the displacement rate. When the displacement rate exceeds the preset displacement rate, or when there is a sudden increase in strain, the system can enter hold / retreat mode and record the anomaly. Throughout the process, the paired data of force and displacement are recorded in real time, and the paired data of force and displacement are plotted as a static load displacement curve.

[0073] For example, joint parameter inversion can be a back analysis method that simultaneously integrates multiple types and physical dimensions of experimental observation data in static or dynamic rock anchor tests, and identifies and optimizes the solution of key model parameters based on the anchor-rock interface mechanical constitutive model and numerical model. In this embodiment of the invention, after obtaining the static load-displacement curve and time-spectrum characteristics, the anchor-rock interface parameters of the rock anchor foundation under test can be inverted based on the joint parameters of the static load-displacement curve and time-spectrum characteristics to obtain the inverted anchor-rock interface parameters. These inverted anchor-rock interface parameters may include data such as bond strength, friction coefficient, effective anchorage length, and nonlinear stiffness.

[0074] Specifically, the joint parameter inversion employs either a physics-prior-driven Bayesian inversion or a physics-constrained machine learning regression. During the inversion process, uncertainties in the forward model are sampled and estimated, and confidence intervals for the parameters are output, resulting in the inversion parameter set and its uncertainty distribution. After obtaining the static load-displacement curve and time-frequency characteristics, joint parameter inversion is performed using the physical model and prior knowledge to calculate the parameters of the anchor-rock interface. Through advanced computational methods such as Bayesian inference, the actual parameters of the anchor body and rock mass can be accurately estimated, and uncertainties can be quantified. This is significant for improving the reliability of test results and providing a more accurate assessment of the anchor body's bearing capacity.

[0075] In one possible manner, the inversion of the anchor-rock interface parameters of the rock anchor foundation under test based on the combined parameters of the static load-displacement curve and the time-frequency characteristic includes:

[0076] Based on the static load displacement curve and time spectrum characteristics, a forward model of the anchor-rock interface parameters of the rock anchor foundation to be tested is constructed.

[0077] The parameters of the anchor-rock interface are obtained by performing joint parameter inversion on the forward model using Bayesian inversion or physical constraint-based machine learning regression.

[0078] It should be understood that the joint parameter inversion can be implemented using the following method. First, a forward model can be constructed based on the static load-displacement curve and time-frequency characteristics. This forward model can be a forward static axial bond-slip model and a simplified dynamic frequency response model. The parameter set in the forward static axial bond-slip model includes average bond strength, effective anchorage length, and nonlinear stiffness. The simplified dynamic frequency response model can be an equivalent multi-degree-of-freedom or parametric damping model.

[0079] Exemplarily, after obtaining the inverted anchor-rock interface parameters, posterior samples can be extracted from the inverted anchor-rock interface parameters to construct a bearing capacity function, and the posterior samples can be brought into the bearing function through the Monte Carlo sampling method for calculation to obtain the bearing capacity distribution. The bearing capacity distribution and the load polarity are calculated to obtain the reliability index, where the load is the load during design. Specifically, through the Monte Carlo or Bayesian update method, uncertainty is incorporated into the analysis, and the probabilistic reliability index is output and the graded implementation adjustment suggestions are formed accordingly; based on the inverted anchor-rock interface parameters obtained by inversion, reliability analysis and remaining bearing capacity prediction are carried out to evaluate the safety of the anchor body under actual working conditions. This step quantifies the uncertainty through methods such as Monte Carlo simulation, generates the failure probability and the reliability index, and provides a basis for subsequent reinforcement design or maintenance decision-making. According to the prediction results, implementation adjustment suggestions are formulated to ensure the safety and reliability of the anchor body throughout its service life.

[0080] Furthermore, taking the posterior of the inverted anchor-rock interface parameters as the input and combining the on-site time-varying / deterioration information reflected in the calibrated preload-response time series, Monte Carlo reliability analysis is carried out and engineering disposal suggestions are generated; Monte Carlo sampling is performed on the bearing capacity function with the posterior samples; the failure probability Pf = P(R < S) and the reliability index are statistically calculated, and graded suggestions G are given according to the predefined decision threshold; if a re-inspection is carried out according to the suggestions and a new preload response time series is collected subsequently, the new preload response time series is merged with the historical calibrated preload response time series and the recalculation of the joint parameter inversion module based on physical prior constraints is triggered to update the inverted anchor-rock interface parameters, so that the whole process forms an adaptive closed loop.

[0081] Specifically, the posterior samples can be a series of values obtained by the Bayesian inversion method, representing the possible values of the anchor-rock interface parameters. It is not a single definite value, but a sample set, and the distribution law of this set depicts the uncertainty of the parameters.

[0082] The specific process is as follows: Input: A posterior sample set of inverted anchor-rock interface parameters obtained from Bayesian inversion. This sample set contains tens of thousands of possible parameter combinations, each representing a possible state under existing observation data. Constructing the bearing capacity function: Establishing a physical formula to map the anchor-rock interface parameters to the ultimate pull-out bearing capacity R of the anchor bolt. Monte Carlo sampling: Substituting the posterior sample set from step one by one into the bearing capacity function R from step two. In this way, a corresponding bearing capacity R will be calculated for each set of parameters. Ultimately, what is obtained is no longer a single bearing capacity value, but a probability distribution of bearing capacity. Calculating the reliability index: Comparing the calculated bearing capacity distribution with the known design load S to obtain the reliability index. In another possible approach, repeated static load or extended tests can be conducted on some of the rock anchor foundations to be tested, and the results of the retest observations can be compared with the inverted anchor-rock interface parameters. If the comparison deviation exceeds the limit, the correction information will be fed back to the joint parameter inversion to update the forward model, thereby closing the data flow and forming an adaptive improvement mechanism to obtain the corrected inverted anchor-rock interface parameters and corresponding adjustment suggestions.

[0083] By combining static load testing, small-amplitude dynamic disturbance testing, and distributed in-situ sensing with the above technical solution, and utilizing physical prior-driven joint inversion, this method can quantify interface parameters and provide confidence intervals without increasing destructive sampling. This allows for probabilistic prediction of the remaining bearing capacity, enabling not only accurate measurement of the anchor's ultimate bearing capacity but also early detection of initial signs of interface damage through time-frequency response analysis. Through this multi-detection method, the present invention effectively improves the comprehensive evaluation of anchor performance and avoids interface damage problems that may be overlooked by single static load testing.

[0084] In the specific implementation process, the specific operations are as follows:

[0085] For a slope of a mountain photovoltaic power station, with bedrock composed of moderately weathered granite, a non-destructive pull-out assessment of a single φ22×3.5 m anchor is proposed to predict its remaining bearing capacity and determine whether reinforcement is necessary. The design load is S=120 kN. The anchor body outer diameter is d=22 mm, and the effective anchorage length L is in mm. The cross-sectional area of ​​the anchor body steel is As=πd² / 4. The average interfacial bond strength is denoted as τb. The bond-type pull-out bearing capacity Rbond is calculated as: Rbond=τb•(πd)•Leff(N), where πd and Leff are both in mm and the result is in N. The ultimate tensile strength of the steel reinforcement is Rsteel: Rsteel=As•fy(N), where fy=500 MPa. The actual pull-out bearing capacity is the minimum of the two values: R=min(Rbond,Rsteel). Numerical sampling (MCMC+MonteCarlo) is used for both the inversion and reliability calculations. MCMC parameters: 20,000 samples, with the first 5,000 used for burn-in. Monte Carlo sample size: 10,000. Design / verification action value: Maximum uplift force S = 120 kN. This design is specifically applied to: well-weathered rock, typical mountain photovoltaic conditions; Condition: moderate slope angle, bedrock is moderately weathered granite (UCS approximately 60 MPa); Anchor specifications: anchor length 3.5 m (L = 3500 mm), diameter 22 mm, grouting strength 35 MPa; Objective: to assess and predict the remaining uplift bearing capacity of the anchor body and propose implementation and adjustment suggestions. Specific steps are as follows:

[0086] Initial geological parameter set acquisition: A RIEGL VZ-400 laser scanner was used to acquire slope point clouds and generate a DEM around the test site. Subsequently, two shallow drills were performed near the anchor positions, with a core sampling depth of 3m. Unconfined compression tests were conducted indoors, and a DFOS (Digital Ground Radar System) with a spatial resolution of 0.5m was deployed along the anchor holes. Temperature and strain were continuously recorded for 48 hours to obtain the temperature baseline and strain limit. These data were combined to form the initial geological parameter set for subsequent processing. The point cloud accuracy of the 3D laser scanning was 5mm / m, the penetration depth of the shallow ground radar was 3m, and the resolution was 0.1m; the spatial resolution of the distributed fiber optic temperature / strain sensors was 0.5m.

[0087] The initial geological parameters obtained include topographic parameters, near-surface layer stratification characteristics and joint distribution characteristics, UCS intervals, temperature data, and fracture data. Specifically, they may include descriptions of rock mass thickness / fracture density, local deformation baselines: displacement baseline 0.0–0.6 mm, temperature baseline 12.5 ± 0.8 ℃.

[0088] The anchor body was preloaded in stages, and the in-situ response was recorded. This was used to record the initial contact and time-varying behavior of the anchor-rock interface under the constraints of initial geological parameters, resulting in a preload response time series. Loading scheme: Stepped preload: 20kN→40kN→60kN→80kN→100kN, with each stage recorded for 10 minutes to eliminate creep effects. Sensing and sampling: The spatial resolution of the DFOS was 0.5m, and the static sampling frequency was 1Hz; the range of the head LVDT was 10mm, the resolution was 0.01mm, and the frequency was 1Hz; the accuracy of the load sensor was 0.5%FS, and the sampling frequency was 1Hz. Results: At the 100kN stage, the head displacement was 1.15mm, and the strain distribution of the DFOS was smooth from 0–2.6m, without any local abrupt increases.

[0089] A closed-loop incremental static load pull-out test was used to obtain the static load-displacement curve, taking into account the overall slope deformation and environmental influences. Static load program: displacement-force dual closed-loop control, with force increasing by 10 kN per step, up to 180 kN. Each load was held for 5 minutes, and slope displacement and temperature compensation were performed. Static characteristics: linear region before yielding, elastic stiffness kel≈12 kN / mm, maximum recorded peak value Fmax,meas=185 kN, corresponding to a displacement of 15.5 mm.

[0090] Small-amplitude dynamic disturbance superposition and time-frequency response acquisition are used to identify interface nonlinearity and damage-sensitive spectral characteristics, obtaining time-frequency spectral features. Dynamic excitation: At target static load levels of 80kN, 120kN, and 160kN, short pulses and linear frequency sweeps are superimposed. The short pulses are equivalent to the peak force of an impact hammer, with a peak value of 1–2kN and a duration of 50ms. The linear frequency sweep is 10–500Hz with an amplitude of 0.5kN. Sampling: Accelerometer and strain gauges are used for dynamic sampling at 2000Hz. Processing: A short-time Fourier transform window of 0.1s with 50% overlap and an FFT length of 1024 is applied. Subsequently, at the 160kN level, the energy in a certain frequency band of 120–150Hz shows a 5% broadening and a slight 1.8% peak frequency shift.

[0091] A joint parameter inversion based on physical prior constraints is used to jointly invert the anchor rock interface parameters from the static load-displacement curve and the time-spectrum characteristics, thus obtaining the inverted anchor rock parameters. Prior settings: τb ~ N (1.2 MPa, 0.252), Leff ~ U (500, 3500) mm, nonlinear stiffness parameter knl ~ N (0.0, 1.0). Inversion method: A forward model is constructed, using the Bayesian MCMC method. The likelihood function consists of the residuals of the static load-displacement curve and the root mean square error of the time-spectrum characteristics.

[0092] ;

[0093] in, It is the likelihood function (the probability density related to the observation residuals). is the standard deviation of the observation noise for static load observation (or a certain order), is the static load residual, is the spectrum residual, is the standard deviation of the spectrum feature observation noise.

[0094] Inversion result: τb = 1.25 MPa (σ = 0.18), Leff = 2600 mm (σ = 300 mm). Calculated by the formula: Rbond = 1.25×(π×22)×2600 = 224.6 kN, Rsteel = Asfy = 380.1×500 = 190.1 kN.

[0095] Therefore, R = min(224.6, 190.1) = 190.1 kN. Among them, the steel limit is the limit, and the mean value of the posterior distribution is about 188 - 190 kN, σ ≈ 14 kN.

[0096] Reliability assessment and prediction of the remaining bearing capacity are carried out and implementation adjustment suggestions are generated, which are used to evaluate the reliability based on the observation uncertainty and obtain the reliability index and suggestions.

[0097] Calculation method: Use the posterior parameter distribution for Monte Carlo sampling. Calculate R each time, calculate the failure probability Pf = P(R < S), and convert it to the reliability index β = -Φ-1(Pf). Take S = 120 kN. Results: The mean value of the posterior R is 188 kN, and the standard deviation is 14 kN; Pf ≈ 1.7×10-6, β ≈ 4.6. Suggestions are obtained: Since Rsteel is close to the limit, it is recommended to increase anti-corrosion and regular inspections during long-term operation; immediate reinforcement is not required, but it is recommended to arrange a re-inspection plan for local anchor points. As can be seen from the above, through the joint inversion method introducing physical prior constraints, the present invention can jointly calculate the accurate parameters of the anchor body and the rock mass based on static load and dynamic data, effectively quantify the uncertainty, and greatly improve the accuracy and reliability of the test results.

[0098] In another embodiment, the specific implementation process is as follows:

[0099] When the rock anchor foundation to be measured is a poorly weathered rock or a slope with more developed local fractures, the specific situation is as follows: Conditions: Weakly weathered gneiss / schist, UCS ≈ 30 MPa, with large joint zones. Anchor specifications: Anchor length 3.0 m, diameter 22 mm, grouting strength 30 MPa. Objective: Use this method to identify early interface deterioration and predict the remaining bearing capacity under poor geological conditions, and put forward reinforcement suggestions.

[0100] Procedure: First, acquire the initial geological parameter set. Data: Point cloud laser scanning; deploy more shallow drills to identify joints at 2m intervals; DFOS spatial resolution is 0.5m; initial displacement baseline is 0.0–1.2mm. Initial geological parameters include: joint angle, joint spacing (0.2–0.6m), and initial local micro-displacement peak value (1.2mm). Second, perform graded preloading of the anchor body and record the in-situ response to record the initial contact and time-varying behavior of the anchor-rock interface, obtaining the preload response time series.

[0101] Preload: 20→40→60→80→100kN, with a holding time of 10min per step. Preload response time series: In the 80–100kN step, DFOS showed a sudden increase in local strain in the 0.6–1.2m range, indicating local bond degradation.

[0102] Closed-loop incremental static load pull-out test was used to obtain the static load-displacement curve. Control: Displacement-force dual closed loop, force increment of 10 kN / step, up to 160 kN or significant slippage. Static load-displacement curve: Maximum recorded value Fmax,meas = 140 kN, corresponding to a displacement of 18.3 mm; the curve shows a rapid decrease in stiffness after 120 kN. Small-amplitude dynamic disturbance superposition and time-frequency response acquisition were used to identify interface nonlinearity and damage spectrum characteristics, obtaining the time-frequency spectrum characteristics. Dynamic excitation was the same as the previous implementation process.

[0103] Time-spectral characteristics: At the 120 kN level, the energy significantly decreases near the dominant frequency of 130 Hz and a new secondary peak appears at 70–90 Hz. Combined with the sudden increase in local strain in DFOS, this indicates viscous partial fracture. Joint parameter inversion based on physical prior constraints is used to obtain the inverted rock-anchor interface parameters. Prior: τb ~ N(0.95, 0.252), Leff ~ U(300, 3000) mm. Inversion then yields τb = 0.95 MPa (σ = 0.22), Leff = 2000 mm (σ = 350 mm). Calculation: Rbond = 0.95 × (π × 22) × 2000 = 131.3 kN, Rsteel = 190.1 kN. Therefore, R ≈ 131.3 kN, and the posterior σ ≈ 18 kN.

[0104] Reliability assessment and recommendation generation yielded reliability indices and implementation adjustment suggestions. Monte Carlo simulation of 10,000 cycles, S = 120 kN. The posterior RRR (Relative Residual Ratio) mean is 130 kN, σ = 18 kN. Calculation: Pf = Φ((120-130) / 18) = Φ(-0.556) = 0.29, β ≈ 0.56. Recommendation: Due to the large Pf, it is recommended to implement local reinforcement of the anchor (e.g., extend the anchorage length by 0.5 m or add casing grouting) and repeat the testing on adjacent anchor points; simultaneously, local reinforcement grouting should be performed in the 0.6–1.2 m section.

[0105] Calculation process: Static load displacement curve fitting and stiffness estimation: The elastic stiffness kel is obtained by fitting the linear segment before the load using linear regression. For example, least squares fitting is performed in the 0–60kN range: kel = ΔF / Δu.

[0106] Time-frequency feature extraction: Perform STFT (Short-Time Fourier Transform) on the dynamic signal: the short-time energy spectrum E(t,f) is obtained, and the peak frequency fp and bandwidth BW are used as features. Spectral impairment index. Calculate using the following formula:

[0107] ;

[0108] in, The spectral bandwidth (in Hz) for the current window (or the disturbance response at the current static load level). The spectral bandwidth (Hz) is the baseline or reference state. The weight is denoted as , and its value is . =1, The baseline peak frequency (Hz) is used. The peak frequency (Hz) of the current window.

[0109] Joint inversion likelihood term: Static load residuals are calculated using the following formula:

[0110] ;

[0111] in, Let F be the standard deviation of the i-th static load observation point divided by the observation noise (units are the same as F), reflecting the combined uncertainty of measurement error and model mismatch term. The measured load at the i-th observation point (unit: N or kN). This represents the predicted load (N or kN) of the forward model for the i-th observation point under parameter set θ.

[0112] The spectral residual is calculated using the following formula:

[0113] ;

[0114] in, The observation uncertainty (dimensionless) of the j-th DI can be estimated from the sample standard deviation of the DI obtained from repeated field perturbations, or given from the baseline perturbation statistics. The measured spectral impairment index for the j-th disturbance event / feature. Let DI be the prediction of the j-th event / feature by the forward dynamic model under parameter θ. It is a dynamic observation index (dimensionless) that can be indexed by disturbance event number (first impact, second impact, etc.) or by different spectral characteristics (peak frequency, bandwidth, etc.).

[0115] Monte Carlo reliability calculation: Draw N times θ(n) (n=1...N) from the posterior a posteriori, and calculate the corresponding... .

[0116] Failure probability Calculate using the following formula.

[0117] ;

[0118] in, This represents the number of samples. Let N be the bearing capacity obtained from the nth sampling. This refers to the action value or design load (N). This is an indicator function that takes the value 1 if the condition within the parentheses is true, and 0 otherwise.

[0119] Reliability index Calculate using the following formula:

[0120] ;

[0121] in, It is the inverse CDF of the standard normal distribution; its input is the probability value (0–1), and its output is the corresponding z value (dimensionless).

[0122] As can be seen from the above, compared to traditional testing methods that typically only focus on the ultimate bearing capacity of the anchor body while neglecting the changes and uncertainties that may occur under actual working conditions, this application provides a more comprehensive safety assessment by evaluating the reliability of the remaining bearing capacity and failure probability. This helps engineering designers understand the performance of the anchor body under different loads in a timely manner and reduces potential safety risks.

[0123] By combining static load testing, small-amplitude dynamic disturbance testing, and distributed in-situ sensing with the above technical solution, and utilizing physical prior-driven joint inversion, interface parameters can be quantified and confidence intervals provided without increasing destructive sampling. This allows for probabilistic prediction of the remaining bearing capacity, enabling accurate measurement of the anchor's ultimate bearing capacity and early detection of initial signs of interface damage through time-frequency response analysis. This multi-detection method effectively improves the comprehensive evaluation of anchor performance, avoiding interface damage issues that might be overlooked in a single static load test. By introducing a joint inversion method with physical prior constraints, this invention can jointly calculate the precise parameters of the anchor and rock mass based on the static load-displacement curve and time-frequency characteristics, effectively quantifying uncertainties and greatly improving the accuracy and reliability of the test results.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for testing the pull-out resistance of rock anchor foundations for mountain photovoltaic power stations, characterized in that, include: Obtain the initial geological parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station; Using initial geological parameters as constraints, the anchor body of the rock anchor foundation under test is subjected to stepped loading through a preset stepped incremental loading strategy. At the same time, the initial contact and time-varying behavior of the rock anchor interface of the rock anchor foundation under test are obtained to obtain the preload response time series. Using the preload response time series as a prerequisite, a closed-loop incremental static load pull-out test was conducted on the rock anchor foundation under test according to the preset static load stage to determine the static load force-displacement curve and time spectrum characteristics. The anchor-rock interface parameters of the rock anchor foundation under test are obtained by inverting the joint parameters of the static load displacement curve and time-frequency characteristics. Based on the posterior samples of the inverted anchor rock interface parameters, a bearing capacity function is constructed. The posterior samples are then substituted into the bearing capacity function using the Monte Carlo sampling method to calculate the bearing capacity distribution. The reliability index is then calculated based on the bearing capacity distribution and the load.

2. The pull-out test method for rock anchor foundations of mountain photovoltaic power stations according to claim 1, characterized in that, The acquisition of initial geological parameters for the rock anchor foundation of the mountain photovoltaic power station includes: Point cloud data of the rock anchor foundation to be tested was obtained by a 3D laser scanner, and a digital elevation model of the rock anchor foundation to be tested in the mountain photovoltaic power station was constructed based on the point cloud data. The terrain parameters of the rock anchor foundation to be tested in the mountain photovoltaic power station were then extracted from the digital elevation model. The near-surface layering and joint distribution characteristics of the rock anchor foundation under test were obtained by shallow ground radar and shallow core sampling methods. Rock samples were collected simultaneously to carry out unconfined compressive strength tests to obtain the UCS interval. Temperature and crack data are obtained by monitoring the temperature and strain profile data of the rock anchor foundation under test using distributed optical fiber sensors. Topographic parameters, near-surface layer stratification characteristics, joint distribution characteristics, UCS intervals, temperature data, and fracture data were determined as the initial geological parameters.

3. The pull-out test method for rock anchor foundations of mountain photovoltaic power stations according to claim 1, characterized in that, The process involves applying a stepped loading strategy to the anchor body of the rock anchor foundation under test, while simultaneously acquiring the initial contact and time-varying behavior of the anchor-rock interface to obtain the preload response time series, including: According to the preset stepped incremental loading strategy, the anchor body of the rock foundation under test is loaded in steps using hydraulic jacks. At the same time, strain data, load data, displacement data and temperature data of the initial contact and time-varying behavior of the rock anchor interface of the rock foundation under test are collected during each step incremental loading process. The strain data, load data, displacement data, and temperature data are corrected to obtain the preload response time series.

4. The pull-out test method for rock anchor foundations of mountain photovoltaic power stations according to claim 1, characterized in that, The process of using the preload response time series as a prerequisite, conducting closed-loop incremental static load pull-out tests on the rock anchor foundation under test according to the preset static load order, and determining the static load-displacement curve and time-frequency characteristics includes: Based on the low-delay real-time flow in the preload response time series, dynamic compensation is performed on each static load step in the closed-loop incremental static load pull-out test using the displacement-force dual closed-loop control method to obtain the static load force-displacement curve. Short pulse and linear frequency sweep excitation signals are superimposed on the rock anchor foundation under test in the closed-loop incremental static load pull-out test, which is at the target static load level. Acceleration and strain signals are acquired at the same time, and short-time Fourier transform or wavelet transform is performed on the acceleration and strain signals to obtain the time-frequency characteristics.

5. The pull-out test method for rock anchor foundations of mountain photovoltaic power stations according to claim 4, characterized in that, The method involves dynamically compensating each static load stage in the closed-loop incremental static load pull-out test using a displacement-force dual closed-loop control method based on the low-delay real-time flow in the preload response time series, to obtain the static load force-displacement curve, including: For each static load stage in the closed-loop incremental static load pull-out test, the displacement amount is obtained by dynamically compensating the real-time displacement component in the low-delay real-time flow and the a priori displacement baseline of the geological and environmental parameters in the initial geological parameters through the displacement-force dual closed-loop control method. The displacement is processed by time differentiation to obtain the displacement rate. Anomaly protection is determined based on the displacement rate. Paired data of force and displacement generated throughout the process are obtained and plotted as static load-displacement curve.

6. The pull-out test method for rock anchor foundations of mountain photovoltaic power stations according to claim 1, characterized in that, The anchor-rock interface parameters of the rock anchor foundation under test are obtained by inverting the joint parameters of the static load displacement curve and time-frequency characteristics, including: Based on the static load displacement curve and time spectrum characteristics, a forward model of the anchor-rock interface parameters of the rock anchor foundation to be tested is constructed. The parameters of the anchor-rock interface are obtained by performing joint parameter inversion on the forward model using Bayesian inversion or physical constraint-based machine learning regression.

Images