Anchor cable loosening ultrasonic identification method and test system for multi-anchor cable system

Abstract

The invention discloses a multi-anchor-cable system anchor cable loosening ultrasonic identification method and a test system, and relates to the technical field of structure health monitoring. The technical problems that in the prior art, the randomness of a contact interface of an anchorage device and an anchor bearing plate is high, signal universality is poor, and the tensioning force of a single prestressed tendon of a multi-anchor-cable system cannot be recognized are solved. According to the invention, parallel detection of a multi-anchor cable system, positioning of a loosening position of a single anchor cable and quantitative evaluation of the loosening degree are realized by constructing a multi-ultrasonic signal receiving and transmitting channel, designing a double-ring piezoelectric ultrasonic signal acquisition device, building a test system adaptive to different scenes, optimizing a signal processing method and introducing a neural network model. The method and the system are suitable for monitoring anchoring systems of newly-built and existing prestressed concrete structures, have the advantages of high detection efficiency, high identification precision, strong universality and the like, and can be widely applied to safety monitoring of large civil structures such as bridges, dams, tunnels and the like.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, and in particular to an ultrasonic identification method and testing system for anchor cable loosening in a multi-anchor cable system. Background Technology

[0002] Prestressed concrete structures, by pre-applying internal stress to concrete or reinforced concrete, alter the distribution of internal stress within the components, making the structure less prone to cracking or delaying cracking under service loads, thereby significantly improving the structure's load-bearing capacity and durability. As an important form of bridge and large-scale civil engineering structure, prestressed concrete structures fully utilize the material advantages of concrete, steel reinforcement, and prestressing tendons, resulting in significantly improved adaptability, spanning capacity, and economy.

[0003] In practical engineering, post-tensioning is one of the most commonly used methods for applying prestress, but it involves a large amount of construction work, complex working conditions, and diverse anchorage forms. Effectively identifying the prestressing force acting on concrete members is crucial for structural safety assessment and crack prevention. However, prestressed anchorage systems generally exhibit characteristics such as large loads, small deformations, and a large number of tendons. This is especially true in multi-hole anchorage systems, which contain multiple independent prestressing tendons, making quantitative identification of the tension force in each strand even more challenging.

[0004] To ensure that the performance of post-tensioned prestressed structures meets design requirements, scholars and engineers at home and abroad have proposed a variety of prestress monitoring methods, including pressure ring method, vibration method, magnetoelastic method, sonic elastic method and strain method, but these methods all have certain limitations.

[0005] In recent years, piezoelectric smart materials (PZT) have been applied in structural health monitoring due to their advantages of lightweight, high sensitivity, strong anti-interference, and low cost, giving rise to piezoelectric impedance and piezoelectric wave methods. The piezoelectric impedance method is suitable for high-frequency small deformation detection, but it can only identify stress changes and is significantly affected by temperature, which is not conducive to long-term monitoring. The piezoelectric wave method identifies the structural state by detecting the propagation characteristics of ultrasonic signals. It can use the correlation between reflected waves, transmitted waves and contact pressure at rough contact interfaces to estimate the tension of the entire prestressed tendon. However, existing piezoelectric active sensing technology is significantly affected by the randomness of interface roughness and contact pressure in anchor head-anchor plate contact interface testing, resulting in unstable signal response characteristics and poor method versatility. At the same time, existing detection methods based on acoustic or piezoelectric principles generally focus on the identification of the entire prestressed tendon, making it difficult to achieve independent tension identification and ultrasonic identification of a single prestressed tendon in a multi-anchor system. Summary of the Invention

[0006] The purpose of this invention is to provide an ultrasonic identification method and testing system for anchor cable loosening in a multi-anchor cable system. This method aims to solve the technical problems of high randomness at the contact interface between the anchor and the anchor plate, poor signal universality, and inability to identify the tension of a single prestressed tendon in a multi-anchor cable system. It enables parallel detection of multi-anchor cable systems, location of loosening of a single anchor cable, and quantitative assessment of the degree of loosening.

[0007] To achieve the above objectives, the present invention provides an ultrasonic identification method for anchor cable loosening in a multi-anchor cable system, comprising the following steps: Step 1: Construct multiple ultrasonic signal transmission and reception channels for the anchor cable anchoring system: Multiple independent ultrasonic signal propagation paths are formed within the anchor head. When the ultrasonic signal reaches the test interface, it generates reflected and transmitted waves as it propagates through the interface. The characteristic changes of the reflected and transmitted waves characterize the interface contact pressure and the stress state of a single prestressed tendon. Step 2, Signal Transmission and Acquisition: A dual-ring piezoelectric ultrasonic signal acquisition device is used. Ultrasonic signals are transmitted through the piezoelectric probe of the upper ring. Depending on the test scenario, the piezoelectric probe of the lower ring is selected to receive transmitted waves, the piezoelectric probe of the upper ring is selected to receive reflected waves, or the transmission and reception of ultrasonic signals can be completed by only using the piezoelectric probe of the upper ring. Step 3, Signal Processing and Feature Extraction: After denoising and filtering preprocessing of the acquired reflected wave signal and / or transmitted wave signal, the signal feature parameters are extracted. The signal feature parameters include, but are not limited to, one or more of the following: energy density, wavelet packet energy, peak amplitude, arrival time, energy ratio, dominant frequency offset, and propagation delay. Step 4: Neural Network Model Training: Collect multi-channel ultrasonic signals under different tension conditions, extract signal feature parameters to construct a sample dataset, and after normalization, denoising, and dimensionality reduction preprocessing, input it into a multilayer feedforward neural network or convolutional neural network for training to establish a nonlinear mapping relationship between signal feature parameters and anchor cable stress level and loosening coefficient. Step 5: Anchor cable loosening identification and assessment: Input the characteristic parameters of the measured ultrasonic signal into the trained neural network model, determine whether the anchor cable is loose by feature matching, quantify the degree of loosening by the loosening coefficient output by the model, locate the loosening location, and generate a multi-anchor cable status distribution map and alarm information.

[0008] Preferably, the dual-ring piezoelectric ultrasonic signal acquisition device in step two includes an upper test ring and a lower test ring; The test ring consists of a magnetic ring, an upper skeleton ring, and an ultrasonic transmitting probe. The magnetic ring is located at the top of the upper skeleton ring. An upper wire groove is formed along the axial direction at the top of the upper skeleton ring. An upper connector groove is formed on the outer side of the upper skeleton ring. An upper wire head is set in the upper connector groove. An upper cylindrical cavity matching the size of the ultrasonic transmitting probe is uniformly formed along the circumference at the bottom of the upper skeleton ring. The upper cylindrical cavity and the upper connector groove are connected to the upper wire groove. The ultrasonic transmitting probe is embedded in the upper cylindrical cavity. A thrust spring, a probe rear cover plate, an upper piezoelectric element, and a probe front cover plate are coaxially assembled in sequence along the axial direction inside the ultrasonic transmitting probe. The two ends of the thrust spring abut against the upper cylindrical cavity and the probe rear cover plate, respectively. The test lower ring consists of an isolation pad ring, a lower skeleton ring, a lower piezoelectric plate, and a filling head. The upper surface of the test lower ring and the lower surface of the anchor form a test interface. The top surface of the lower skeleton ring is provided with several pad ring mounting grooves, and the isolation pad ring is placed in the pad ring mounting grooves. The bottom of the lower skeleton ring is machined with several lower cylindrical cavities, which are arranged vertically opposite to the pad ring mounting grooves and the upper cylindrical cavities. The lower piezoelectric plate is installed at the top of the lower cylindrical cavity to receive ultrasonic waves passing through the test interface, and the filling head is installed at the bottom of the lower cylindrical cavity. The side of the lower skeleton ring is provided with a lower connector groove, and a lower wire groove is opened on the circumference of the bottom surface of the lower skeleton ring. The lower wire groove is connected to the lower connector groove and the lower cylindrical cavity, and a lower wire head is provided in the lower connector groove.

[0009] Preferably, the signal transmission and acquisition in step two includes two modes, namely: Transmitted / Reflected Wave Joint Testing Mode: Applicable to newly constructed prestressed structures. An upper test ring is installed above the anchor, and a lower test ring is set between the anchor and the anchor plate. Ultrasonic signals are emitted through the upper test ring, and reflected waves are received by the upper test ring and transmitted waves are received by the lower test ring. Reflected wave test mode: Applicable to existing anchorage structures, only the test ring is installed above the anchorage, and the transmission of ultrasonic signals and reception of reflected waves are completed through the test ring.

[0010] Preferably, the signal processing in step three employs the following method: (1) Transmission wave analysis: For the swept-frequency signal energy: the energy density of the transmitted signal is calculated through the frequency domain energy distribution to reflect the interface pressure change; for the wavelet packet energy: the transmitted wave signal is decomposed into multiple scales to extract the energy characteristics of each frequency band and construct an energy ratio index to identify load changes; by integrating multiple feature indices, a feature database of the transmitted wave is established through parameters such as energy amplitude ratio, time difference of arrival, and frequency band energy distribution.

[0011] (2) Analysis of reflected waves: One of the following methods is employed: Direct pulse method (signal peak): extracting the amplitude and arrival time of the first peak wave to quickly determine changes in contact stiffness and interface pressure; Time inversion method (focusing signal peak): reversing the time of the received signal to achieve energy focusing and peak enhancement, suitable for low signal-to-noise ratio scenarios; combining the time inversion method and the direct pulse method to achieve synchronous discrimination of signal peak characteristics and enhance the ability to distinguish signals from multiple anchor cables.

[0012] Preferably, the neural network model training step in step four includes: (1) Sample acquisition: Multi-channel reflected and transmitted wave signals were acquired under different tension conditions, and time-domain and frequency-domain characteristic parameters were extracted; (2) Data preprocessing: Normalization, noise reduction, wavelet dimensionality reduction and other processing are performed to ensure sample balance and stability; (3) Feature input layer design: Wavelet packet energy, peak amplitude, time delay, energy ratio, etc. are selected as input features; (4) Model training and validation: Supervised learning is adopted, and the network weights are optimized through the backpropagation algorithm. (5) Model output: Output the stress level or loosening coefficient of a single anchor cable to realize the mapping relationship from signal to force.

[0013] Preferably, in step five, the automatic identification steps are as follows: (1) Signal input and matching: Input the characteristic values ​​of reflected and transmitted waves of each channel and perform similarity matching with the feature templates in the database; (2) Loosening judgment: If a certain channel feature deviates from the training mean threshold, the anchor cable is judged to be loose; (3) Quantification of loosening degree: The loosening degree is quantitatively evaluated by the percentage of interfacial contact pressure and stress loss corresponding to the loosening coefficient. It is divided into five levels: normal, slightly loose, moderately loose, significantly loose and severely loose. The value range of the loosening coefficient is 0~1. (4) Visualization and alarm output: The system automatically generates a multi-anchor cable status distribution map, locates the loosening position and marks the loosening level, and realizes intelligent monitoring and early warning of the health status of the multi-anchor cable system.

[0014] An ultrasonic testing system for anchor cable loosening in a multi-anchor cable system includes a dual-ring piezoelectric ultrasonic signal acquisition device, a multi-channel signal receiver and transmitter, and a computer. The dual-ring piezoelectric ultrasonic signal acquisition device includes an upper test ring and a lower test ring, which are respectively connected to the multi-channel signal receiver and transmitter via signal lines. The multi-channel signal receiver and transmitter is communicatively connected to the computer. The computer has a built-in signal processing module, a neural network model training module, and a loosening identification and evaluation module.

[0015] The upper test ring consists of a magnetic ring, an upper skeleton ring, and an ultrasonic transmitting probe. The thrust spring inside the ultrasonic transmitting probe is used to keep the piezoelectric element in a compacted state. The isolation pad ring of the lower test ring is used to achieve high-sensitivity reception and electromagnetic isolation. The multi-channel signal receiver and transmitter is used to achieve synchronous excitation, acquisition, and transmission of ultrasonic signals.

[0016] Preferably, the signal processing module in the computer is used to perform signal denoising, filtering, feature extraction and normalization; the neural network model training module is used to construct and train a multilayer feedforward neural network or a convolutional neural network; and the loosening identification and evaluation module is used to locate the loosening position of the anchor cable, quantify the degree of loosening and provide visualization output and alarm.

[0017] Preferably, the testing system is suitable for detecting loosening of multiple anchor cables in bridges, dams, tunnels, slope anchoring systems, foundation pit support, underground structures, wind turbine tower foundations, and ground anchoring systems.

[0018] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: (1) Realize the overall detection and condition differentiation of multi-anchor cable system Traditional ultrasonic testing often targets entire anchor bundles, making it difficult to quickly identify problems in complex anchorage groups. This invention utilizes a testing system with multi-point excitation and multi-channel synchronous acquisition to achieve parallel testing of multi-anchor systems. It can distinguish the differences in ultrasonic propagation response among individual anchors, significantly improving testing efficiency and positioning accuracy.

[0019] (2) Quantitative identification of loosening of each anchor cable based on wave energy and time-frequency characteristics This invention introduces comprehensive indicators such as wave energy attenuation coefficient, characteristic frequency band energy ratio, dominant frequency offset, and propagation delay variation to construct a multi-dimensional loosening identification feature system. Compared with traditional methods that rely solely on amplitude changes, this method can maintain high identification accuracy even when signal energy attenuation is not significant or environmental interference is substantial.

[0020] (3) Establish the coupling relationship between multi-channel signal characteristics and the mechanical state of each anchor cable. This invention realizes the coupled mapping between "ultrasonic response-loosening location-loosening degree" through a quantitative relationship model between measured ultrasonic signals and anchor cable tension. It can accurately identify the loosening state of multiple steel strands at different degrees, providing a reliable basis for long-term safety monitoring of anchor cable groups.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of ultrasonic propagation at a rough contact interface in Embodiment 1 of the ultrasonic system for anchor cable loosening in a multi-anchor cable system according to the present invention. Figure 2 This is a schematic diagram of the overall structure of the device according to Embodiment 1 of the present invention; Figure 3 This is a disassembly diagram of the ultrasonic identification device for anchor cable loosening according to Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the test upper ring structure in Embodiment 1 of the present invention, wherein (a) is a disassembled structural diagram of the test upper ring, and (b) is a perspective view of the test upper ring structure; Figure 5 This is a schematic diagram of the ultrasonic transmitting probe structure according to Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the test lower ring structure of Embodiment 1 of the present invention, wherein (a) is a disassembled structural diagram of the test lower ring, and (b) is a perspective view of the test lower ring structure; Figure 7 This is a schematic diagram of the combined transmission / reflection wave test mode of Embodiment 1 of the present invention; wherein (a) is an exploded view of the sensor installation of the test system under the combined transmission / reflection wave test mode; and (b) is a diagram of the propagation path of the test ultrasonic signal under the combined transmission / reflection wave test mode. Figure 8 This is a schematic diagram of the signal acquisition system of the test system under the combined transmission wave / reflection wave test mode of Embodiment 1 of the present invention; Figure 9 This is a schematic diagram of the neural network construction in Embodiment 1 of the present invention; Figure 10 This is a linear sweep frequency signal transmission diagram of Embodiment 1 of the present invention; Figure 11 The above are characteristic diagrams of ultrasonic testing for the gradual loosening of a single anchor cable in Embodiment 1 of the present invention, wherein (a) is the reflected waveform received by the upper ring under the gradual loosening of a single cable; and (b) is the waveform of wavelet packet energy changing with load under the gradual loosening of a single cable. Figure 12 This is a schematic diagram of the reflected wave test mode in Embodiment 2 of the present invention; wherein (a) is an exploded view of the sensor installation of the test system in the reflected wave test mode; and (b) is a diagram of the propagation path of the ultrasonic signal in the reflected wave test mode. Figure 13This is a schematic diagram of the signal acquisition system of the test system in the reflected wave test mode of Embodiment 2 of the present invention; Figure 14 This is a pulse signal transmission diagram of Embodiment 2 of the present invention; Figure 15 This is a schematic diagram of ultrasonic propagation at a rough contact interface in Embodiment 2 of the present invention. (a) is a waveform diagram acquired using the time-reversal method; (b) is a Fourier spectrum diagram of (a). Figure 16 This is a convergence curve diagram of the mean square error training, testing, and verification of the multi-anchor cable loosening identification neural network in Embodiment 2 of the present invention. Figure 17 This is a bar chart for identifying loosening of multiple anchor cables in Embodiment 2 of the present invention.

[0024] Figure Labels 1. Test upper ring; 11. Magnetic ring; 12. Upper skeleton ring; 13. Ultrasonic transmitting probe; 131. Probe rear cover plate; 132. Upper piezoelectric element; 133. Probe front cover plate; 134. Thrust spring; 14. Upper lead wire head; 15. Upper lead wire groove; 16. Upper cylindrical cavity; 17. Upper connector wire groove; 2. Anchor; 3. Test lower ring; 31. Isolation pad ring; 32. Lower skeleton ring; 33. Lower cylindrical cavity; 34. Pad ring mounting groove; 35. Lower lead wire groove; 36. Lower lead wire head; 37. Filler head; 38. Lower piezoelectric element; 39. Lower connector wire groove; 4. Anchor plate; 5. Steel strand; 6. Multi-channel signal receiver and transmitter; 7. Computer. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] like Figure 1As shown, this invention constructs multiple ultrasonic signal transceiver channels at the contact interface between the anchor head and the anchor plate to improve the directivity and repeatability of ultrasonic signal propagation. When an ultrasonic signal TX is emitted from PZT1 at one end of the interface, the signal propagates through the rough contact interface and is divided into a reflected wave RX1 and a transmitted wave RX2. The reflected wave is received by the transmitting end PZT1, and the transmitted wave is received by the other end PZT2. Under external load, the contact area of ​​the uneven surface of the interface increases with increasing stress, causing changes in the amplitude ratio, phase difference, and propagation delay of the reflected and transmitted waves. By analyzing the reflected and transmitted wave signals, the interface contact pressure and the stress state of a single prestressed tendon can be indirectly reflected, achieving unidirectional dissipation and stress characteristic characterization.

[0028] Example 1: Ultrasonic Testing and Loosening Identification of Newly Built Structural Anchoring Systems This embodiment is applied to the construction period of new structures. By installing the upper test ring 1 and the lower test ring 3 on the anchor 2, the reflection-transmission combined ultrasonic identification of the tension force and loosening state of the prestressed tendons under the anchor is realized.

[0029] (I) Preparation and setup of the testing device 1. Composition and fabrication of the testing device (1) Overall structure like Figure 2 and Figure 3 As shown, taking a five-cable anchoring system as an example (S1-S5), a multi-cable anchor loosening ultrasonic identification device mainly consists of an upper test ring 1, an anchor 2, a lower test ring 3, an anchor plate 4, and steel strands 5. The steel strands 5 pass sequentially through the anchor 2 and the anchor plate 4, with the anchor 2 and the anchor plate 4 jointly bearing the tension of the steel strands 5. The upper test ring 1 is installed above the anchor 2, and the lower test ring 3 is positioned between the anchor 2 and the anchor plate 4, forming an ultrasonic propagation interface.

[0030] (2) Test the structure of the upper ring 1 like Figure 4 and Figure 5 As shown, the test upper ring 1 consists of a magnetic ring 11, an upper skeleton ring 12, and an ultrasonic transmitting probe 13. The magnetic ring 11 is located on the top of the upper skeleton ring 12. An upper guide wire groove 15 is formed along the axial direction on the top of the upper skeleton ring 12. An upper connector wire groove 17 is provided on the outer side of the upper skeleton ring 12, and an upper guide wire head 14 is provided in the upper connector wire groove 17. An upper cylindrical cavity 16 matching the size of the ultrasonic transmitting probe 13 is uniformly formed along the circumference at the bottom of the upper skeleton ring 12. The upper cylindrical cavity 16 and the upper connector wire groove 17 are both connected to the upper guide wire groove 15.

[0031] An ultrasonic transmitting probe 13 is embedded in an upper cylindrical cavity 16. Inside the ultrasonic transmitting probe 13, a thrust spring 134, a probe rear cover plate 131, an upper piezoelectric element 132, and a probe front cover plate 133 are coaxially assembled in sequence along the axial direction. The two ends of the thrust spring 134 abut against the upper cylindrical cavity 16 and the probe rear cover plate 131, respectively. The thrust spring 134 is used to maintain the piezoelectric element in a compacted state to prevent it from breaking under pressure. The signal line of the upper piezoelectric element 132 is led out through the upper cylindrical cavity and the upper wire groove 15 to the upper wire tip 14. The upper test ring 1 can serve as both a transmitter and a receiver.

[0032] (3) Test the structure of the lower ring 3 like Figure 6 As shown, the lower test ring 3 consists of an isolation pad ring 31, a lower skeleton ring 32, a lower piezoelectric plate 38, and a filling head 37. The upper surface of the lower test ring 3 and the lower surface of the anchor 2 form a test interface. The top surface of the lower skeleton ring 32 has several pad mounting grooves 34, and the isolation pad ring 31 is placed within these grooves. The bottom of the lower skeleton ring 32 has several lower cylindrical cavities 33. The lower piezoelectric plate 38 is installed at the top of each lower cylindrical cavity 33 to receive ultrasonic waves transmitted through the test interface. The filling head 37 is installed at the bottom of each lower cylindrical cavity 33. The lower cylindrical cavities 33, the pad mounting grooves 34, and the upper cylindrical cavity 16 are arranged vertically opposite each other.

[0033] The lower frame ring 32 has a lower connector groove 3 on its side, and a lower lead wire 36 is installed inside the lower connector groove 3. A lower lead wire groove 35 is opened on the bottom circumference of the lower frame ring 32, and the lower lead wire groove 35 is connected to the lower connector groove 3 and the lower cylindrical cavity 33. The signal line is led out to the lower lead wire 36 through the lower cylindrical cavity 33 and the lower lead wire groove 35 to achieve high-sensitivity reception and electromagnetic isolation.

[0034] By coordinating the upper and lower rings, a stable and controllable signal transmission and feedback path is constructed, laying the foundation for multi-channel ultrasonic testing.

[0035] Component manufacturing requirements: (1) Magnetic ring / skeleton ring / probe housing / isolation pad ring / filler head: precision machined from 304 stainless steel to ensure strength, durability and corrosion resistance. (2) Steel strand 5, anchor 2, anchor plate 4: selected according to GB / T 14370 and GB / T 5224 to maintain structural compatibility with the actual anchoring system in the project.

[0036] 2. Test the upper ring. Assemble the upper ring 1 as follows: (1) Fix the magnetic ring 11 to the outer skeleton ring to achieve quick adsorption installation; (2) Assemble the front and rear cover plates onto both sides of the PZT; (3) The thrust spring 134 presses the PZT tightly against the metal surface to ensure acoustic coupling; (4) The wires are laid along the wiring trough to a unified interface; (5) The outer layer of the signal line is shielded to reduce strong electromagnetic interference at the construction site.

[0037] This structure is detachable and reusable, making it suitable for multiple data collections during repeated tensioning during construction.

[0038] 3. Test the assembly of lower ring 3, the steps are as follows: (1) The isolation pad ring 31 is embedded into the lower ring skeleton to isolate external mechanical disturbances; (2) Secure the PZT to the bottom of the filling head 37 with high-damping acoustic adhesive; (3) After assembly, a transmission calibration test is performed to check the signal amplitude, noise ratio and consistency of each channel to ensure measurement accuracy.

[0039] 4. Test mode setup The newly constructed structure can accommodate both the upper and lower rings simultaneously; therefore, this embodiment adopts the following approach. Figure 7 The combined transmitted / reflected wave testing system shown in (a) is suitable for newly constructed prestressed structures. During construction, a lower test ring 3 is embedded between the anchor 2 and the anchor plate 4, while an upper test ring 1 is installed above the anchor 2. Figure 7 As shown in (b), a contact interface is formed between anchor 2 and test lower ring 3. Excitation signals (TX1~TX5) are emitted by PZT in the upper ring, and the reflected waves are received by the upper ring. ~ The transmitted wave is received by the lower ring. ~ The reflected ultrasound (upper ring) and transmitted ultrasound (lower ring) form multiple sets of signals, enhancing the distinguishability of loosening identification.

[0040] (II) Signal excitation, acquisition and processing process 1. Test Condition Design The multi-anchor system consists of five independently tensioned steel strands 5 (S1~S5). To train and validate the ultrasonic signal-based multi-anchor loosening detection neural network (ANN), this embodiment sets up various typical working conditions in the tensioning system of the five anchors. The anchoring force of each anchor can be independently adjusted using a jack. Let the upper limit of the tension force of a single steel strand 5 under normal tensioning conditions be... F 0 (120kN in this embodiment), and the stress level of each steel strand 5 is divided into 1.00. F 0, 0.90 F 0, 0.75 F 0, 0.50 F 0 and 0.25 F The five levels (as shown in Table 1) correspond to normal, slightly loose, moderately loose, significantly loose, and severely loose states, respectively.

[0041] Table 1 Stress Levels and Condition Grades

[0042] Based on the different failure characteristics of multi-anchor cable systems, five typical working conditions were designed, including single anchor loosening, double anchor loosening, multiple anchor loosening, overall loosening, and normal control, totaling approximately 60 groups. Each working condition corresponds to a different combination of anchoring stresses. By adjusting the tension of the anchorages and conducting ultrasonic excitation tests, the response signals under each working condition were collected for neural network training and verification.

[0043] (1) Single-strand loosening condition The simulation of a typical situation where the anchoring force of a single anchor cable decreases in the field is used to train the network's sensitivity to single-point anomalies. Only one steel strand 5 is selected from different grades from L1 to L4, while the other four strands remain at L0, for a total of 20 working conditions (as shown in Table 2).

[0044] Table 2 Single-strand loosening conditions

[0045] (2) Double-root loosening condition A typical scenario of decreased anchoring force in two anchor cables was simulated to train the network's ability to identify simultaneous loosening at multiple points. Two steel strands (5) were randomly selected to loosen simultaneously, while the other three remained normal. Three loosening levels (L1, L2, L3) for common combinations were selected (as shown in Table 3).

[0046] Table 3. Double-root loosening condition

[0047] (3) Multiple roots (3~4 roots) loosening at the same time Simulating the complex scenario of simultaneous loosening of multiple anchors is used for learning high-dimensional coupling features of the network. Three or four steel strands are randomly combined (as shown in Table 4); each strand is assigned a different loosening level (e.g., L1~L3 randomly distributed).

[0048] Table 4. Multiple Wires Loosening Simultaneously

[0049] (4) Overall loosening condition Simulate the overall stress drop or synchronous loosening to assess the network's ability to identify global loosening. Five steel strands drop at the same level simultaneously (as shown in Table 5).

[0050] Table 5 Overall Loosening Conditions

[0051] (5) Normal and control operating conditions All L0 (normal tension state) are used for network normalization and output reference. There is a total of 1 set of control conditions.

[0052] The above-mentioned systematic working condition design can cover the main state space of multi-anchor cable systems under different loosening modes, providing data support for the high-precision training and wide applicability of neural network models.

[0053] 2. Signal stimulation Building such Figure 8 The acquisition system shown utilizes a multi-channel signal receiver / transmitter 6 to achieve synchronous acquisition, with the signal type selected as follows: Figure 10 The linear sweep frequency signal shown has an amplitude of 8V and a frequency range from 5Hz to 1MHz. After power amplification, it is sent to the test upper loop 1 PZT to realize the propagation of longitudinal waves along the anchor cable. Software control enables multi-channel automatic excitation and delay control to obtain signal responses with different incident angles and paths, enhancing detection resolution. The data is input into computer 7 via a control program for signal analysis and processing, enabling the identification and quantitative calculation of the force differences among the cables in the multi-cable system.

[0054] 3. Signal reception and acquisition The upper and lower PZT loops of each anchor cable serve as receiving terminals. The upper PZT loop primarily receives reflected wave signals, while the lower PZT loop receives transmitted wave signals. The signals are acquired by a multi-channel synchronous sampling device at a sampling rate of 8.93 MS / s to ensure waveform accuracy and time synchronization. Figure 11 (a) in the figure is the reflection waveform diagram collected by the test ring 1. After the collected data is preprocessed, it is automatically uploaded to the host computer for storage and analysis.

[0055] 4. Signal Processing and Feature Extraction The acquired raw signals are processed through the following steps: (1) Denoising and filtering: Wavelet threshold denoising and bandpass filtering (30~300kHz) are used to remove mechanical and environmental noise; (2) Feature extraction: Multi-scale features are extracted through wavelet packet decomposition and variational mode decomposition: transmitted wave energy E t Energy of reflected waves E r .like Figure 11 (b) shows the variation of reflected wave energy with load under different loads during single-root progressive loosening. As the load increases, the reflected wave energy decreases, indicating that energy can be used as an important indicator of load magnitude. (3) Feature normalization: The z-score method is used to eliminate the amplitude differences between different channels, resulting in a standardized feature matrix that can be input into a machine learning model. .

[0056] (III) Loosening Identification and Result Output 1. Neural Network Training and Recognition Model Construction To ensure that the training samples for the neural network model cover typical working conditions and possess statistical significance, this embodiment performs multiple independent samplings on the aforementioned approximately 60 sets of typical working conditions. A preferred approach is to perform 50 independent samplings on each set of working conditions (collecting a set of original waveforms and extracting features after each loading / stabilization), totaling 3000 samples. This dataset is then divided into a 70%:15%:15% (training / validation / test) ratio, i.e., 2100 for training, 450 for validation, and 450 for testing. The training set is used for model parameter learning, the validation set for hyperparameter tuning and early stopping, and the test set for final performance evaluation. To ensure a balanced distribution of each working condition across all sets, stratified splitting is used to ensure that each group exists proportionally in training / validation / testing, thereby obtaining a reliable estimate of the recognition performance for each group. Figure 9 As shown, the extracted characteristic parameters (transmitted wave energy) are used. E t Energy of reflected waves E r Using the input layer of the neural network (e.g., 5 strands), and the actual tension of each steel strand 5 as the output layer, a multi-input single-output mapping relationship is established. This is achieved by inputting ultrasonic signal features (...). E t , E r A neural network model for input and output was established. The network uses two hidden layers (10⁻⁸ structure), and the training algorithm is Levenberg–Marquardt (LM).

[0057] 2. Loosening identification and condition assessment Using a trained network model, the system inputs characteristic parameters of the field-acquired signals for identification and calculation, outputting the stress estimate and loosening level for each steel strand. The system constructs loosening criteria based on the output prestress values, achieving five levels of identification as shown in Table 1: "Normal State - Slight Loosening - Moderate Loosening - Significant Loosening - Severe Loosening". When the prestress loss value is 50% of the normal tension value, the system automatically determines it as a significant loosening state and issues an alarm.

[0058] 3. Results Output and System Updates The system outputs predicted and actual prestress loss values, which are automatically stored in the system database. It can dynamically update benchmarks based on long-term monitoring data, compare and adaptively correct prediction and test results from different batches and environmental conditions, thereby improving the accuracy and versatility of long-term monitoring.

[0059] Example 2: Ultrasonic Testing and Loosening Identification of Existing Structure Anchoring Systems This embodiment is applicable to existing projects or rapid inspection scenarios. Since it is impossible to install the lower ring below anchor 2, this embodiment relies solely on reflected ultrasonic signals to complete the ultrasonic identification and quantitative assessment of the tension and loosening status of the prestressed tendons under the anchor in the existing structural anchoring system.

[0060] (I) Preparation and setup of the testing device 1. Composition and fabrication of the testing device The testing device consists of a test upper ring 1, an anchor 2, an anchor plate 4, and steel strands 5. The test upper ring 1 includes a magnetic ring 11, an upper skeleton ring 12, and an ultrasonic transmitting probe 13. The magnetic ring 11, the upper skeleton ring 12, the ultrasonic transmitting probe 13 housing, and the isolation pad ring 31 are all precision-machined from 304 stainless steel to ensure structural strength and corrosion resistance. Each anchor cable is equipped with a piezoelectric element for emitting pulses and receiving reflections. The piezoelectric elements, anchor 2, anchor plate 4, and steel strands 5 are selected and procured according to national standards GB / T14370 and GB / T 5224 to ensure compatibility between the testing system and the actual anchoring system in the project.

[0061] 2. Test the assembly of ring 1. Only the upper ring components are retained. (According to...) Figure 4 and Figure 5 The structure shown comprises the magnetic ring 11, the skeleton ring, the front and rear cover plates of the probe, the thrust spring 134, the piezoelectric element, and the wire wiring groove, assembled sequentially. Spring loading ensures a tight fit between the piezoelectric element and the metal contact surface, guaranteeing stable acoustic energy transmission and preventing damage from overvoltage. All electrical signal lines are centrally encapsulated at the lead interface, using shielded wires to reduce external electromagnetic interference.

[0062] 3. Test mode setup To adapt to existing structural testing scenarios, test ring 1 is installed only above anchor 2, and a setup is constructed as follows: Figure 12 The reflected wave testing system shown in (a) uses reflected waves for detection. Its modular installation and reusability, achieved through a quick-assembly / disassembly structure, improves testing efficiency.

[0063] (II) Signal excitation, acquisition and processing process 1. Test Condition Design The same 60 working conditions were used as in Example 1. Each sample recorded the working condition ID, the true tension of 5 steel strands, the corresponding transmit / receive waveforms, etc. The loading / force measuring equipment was the same as above.

[0064] 2. Signal stimulation Since the test upper ring 1 is installed only above anchor 2, the excitation and reception share the same PZT. A contact interface is formed between anchor 2 and anchor plate 4. The excitation signal (TX1~TX5) is emitted by the PZT in the upper ring, and the reflected wave is received by the upper ring. ~ ),like Figure 12 As shown in (b) of the diagram.

[0065] like Figure 13 As shown, the multi-channel signal receiver / transmitter 6 is used to complete signal excitation and echo acquisition. The signal type selection is as follows. Figure 14 The pulse wave shown is amplified and fed into the 1PZT test loop to propagate longitudinal waves along the anchor cable. Software control enables multi-channel automatic excitation and delay control to obtain signal responses with different incident angles and paths, enhancing detection resolution. Finally, changes in the reflected signal characteristics are used to identify the anchor cable's loosening status.

[0066] 3. Signal reception and acquisition Each anchor cable's upper PZT loop serves as a receiver. Signals are acquired via a multi-channel synchronous sampling device to ensure waveform accuracy and time synchronization. Time reversal is applied to the received signals to achieve energy focusing and peak enhancement. Figure 15 (a) in the figure is a waveform obtained by time inversion method. The collected data is automatically uploaded to the host computer for storage and analysis after preprocessing.

[0067] 4. Signal Processing and Feature Extraction The acquired raw signals are processed through the following steps: (1) Envelope Extraction and Time-Frequency Analysis: The envelope signal is extracted using Hilbert transform and combined with short-time Fourier transform to obtain the spectrum, such as... Figure 15 As shown in (b); (2) Feature extraction: the peak value of the signal after time reversal p k Directly used as extracted features; (3) Feature normalization: The z-score method is used to eliminate the amplitude differences between different channels, resulting in a standardized feature matrix that can be input into a machine learning model. ,in, This represents the peak value of the time-reversed signal.

[0068] (III) Loosening Identification and Result Output 1. Neural Network Training and Recognition Model Construction Similar to Example 1, this example performs multiple independent samplings on the aforementioned approximately 60 typical working conditions. A preferred approach is to perform 50 independent samplings on each working condition, totaling 3000 samples; this dataset is then divided into 70%:15%:15% (training / validation / testing), i.e., 2100 for training, 450 for validation, and 450 for testing. The extracted feature parameters ( p k Using the actual tension of each steel strand 5 as the input layer of the neural network, a multi-input single-output mapping relationship is established. Tables 6 and 7 show input-output examples for five typical working conditions.

[0069] Table 6. Peak values ​​of time-reversed signals extracted under tension conditions A~E (input layer)

[0070] Table 7 Tension values ​​of each steel strand under typical working conditions A~E (output layer)

[0071] Same as in Example 1, input ultrasound signal characteristics ( p k A neural network model for input and output was established. The network uses two hidden layers (10⁻⁸ structure), and the training algorithm is Levenberg–Marquardt (LM). Figure 16 The figure shows the convergence curves of the mean square error (MSE) training, testing, and validation of the neural network for identifying loose anchor cables in this invention. As can be seen from the figure, the network converges after approximately 40 iterations, and the errors on the training, testing, and validation sets all steadily decrease, indicating that the model has good fitting ability and generalization performance, and can be used to identify the loose state of a single anchor cable in a multi-anchor cable system.

[0072] 2. Loosening identification and condition assessment Using the trained network model, the characteristic parameters of the field-collected signals are input for identification and calculation, and the stress estimate and loosening level of each steel strand 5 are output. The judgment rules are the same as in Example 1.

[0073] 3. Results Output and System Updates The test results are presented in a bar chart, such as... Figure 17 As shown, the predicted and actual values ​​of prestress loss can be compared, and significant loosening warning lines can be marked and automatically stored in the system database. The system can dynamically update the benchmark based on long-term monitoring data, compare the predicted results and test results under different batches and environmental conditions, and make adaptive corrections to improve the accuracy and versatility of long-term monitoring.

[0074] (iv) Scope of application and promotion The ultrasonic identification method and testing system for loosening of multi-anchor cable systems proposed in this invention are not only applicable to in-service testing of anchorage systems for bridges, dams, tunnels and slopes, but can also be widely applied to foundation pit support, underground structures, wind turbine tower foundations and ground anchorage systems.

[0075] The system has the advantages of simple structure, convenient installation, stable signal and high identification accuracy. It can realize rapid, non-destructive and quantifiable health diagnosis of complex multi-anchor systems, and has significant engineering promotion value and economic benefits.

[0076] The remaining technical features in the above embodiments can be flexibly selected by those skilled in the art to meet different specific practical needs according to actual circumstances. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims. In the above description, numerous specific details have been set forth to provide a thorough understanding of the present invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to implement the present invention. In other instances, to avoid obscuring the present invention, well-known techniques, such as specific construction details, operating conditions, and other technical conditions, have not been specifically described.

[0077] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for ultrasonic identification of anchor cable loosening in a multi-anchor cable system, characterized in that, The steps are as follows: Step 1: Construct multiple ultrasonic signal transmission and reception channels for the anchor cable anchoring system: Set up an ultrasonic signal propagation path at the interface between the anchor head and the anchor plate so that reflected waves and transmitted waves are generated when the ultrasonic signal propagates through this interface. The characteristic changes of reflected waves and transmitted waves are used to characterize the interface contact pressure and the stress state of a single prestressed tendon. Step 2, Signal Transmission and Acquisition: A dual-ring piezoelectric ultrasonic signal acquisition device is used. Ultrasonic signals are transmitted through the piezoelectric probe of the upper ring. Depending on the test scenario, the piezoelectric probe of the lower ring is selected to receive transmitted waves, the piezoelectric probe of the upper ring is selected to receive reflected waves, or the transmission and reception of ultrasonic signals can be completed by only using the piezoelectric probe of the upper ring. Step 3, Signal Processing and Feature Extraction: After denoising and filtering preprocessing of the acquired reflected wave signal and / or transmitted wave signal, the signal feature parameters are extracted. The signal feature parameters include one or more of the following: energy density, wavelet packet energy, peak amplitude, arrival time, energy ratio, dominant frequency offset, and propagation delay. Step 4: Neural Network Model Training: Collect multi-channel ultrasonic signals under different tension conditions, extract signal feature parameters to construct a sample dataset, and after normalization, denoising, and dimensionality reduction preprocessing, input it into a multilayer feedforward neural network or convolutional neural network for training to establish a nonlinear mapping relationship between signal feature parameters and anchor cable stress level and loosening coefficient. Step 5: Anchor cable loosening identification and assessment: After training, the measured signals are input into the neural network model to realize the automatic identification of multi-anchor cable systems.

2. The ultrasonic identification method for anchor cable loosening in a multi-anchor cable system according to claim 1, characterized in that: The dual-ring piezoelectric ultrasonic signal acquisition device described in step two includes an upper test ring and a lower test ring; The test ring consists of a magnetic ring, an upper skeleton ring, and an ultrasonic transmitting probe. The magnetic ring is located at the top of the upper skeleton ring. An upper wire groove is formed along the axial direction at the top of the upper skeleton ring. An upper connector groove is formed on the outer side of the upper skeleton ring. An upper wire head is set in the upper connector groove. An upper cylindrical cavity matching the size of the ultrasonic transmitting probe is uniformly formed along the circumference at the bottom of the upper skeleton ring. The upper cylindrical cavity and the upper connector groove are connected to the upper wire groove. The ultrasonic transmitting probe is embedded in the upper cylindrical cavity. A thrust spring, a probe rear cover plate, an upper piezoelectric element, and a probe front cover plate are coaxially assembled in sequence along the axial direction inside the ultrasonic transmitting probe. The two ends of the thrust spring abut against the upper cylindrical cavity and the probe rear cover plate, respectively. The test lower ring consists of an isolation pad ring, a lower skeleton ring, a lower piezoelectric plate, and a filling head. The upper surface of the test lower ring and the lower surface of the anchor form a test interface. The top surface of the lower skeleton ring is provided with several pad ring mounting grooves, and the isolation pad ring is placed in the pad ring mounting grooves. The bottom of the lower skeleton ring is machined with several lower cylindrical cavities, which are arranged vertically opposite to the pad ring mounting grooves and the upper cylindrical cavities. The lower piezoelectric plate is installed at the top of the lower cylindrical cavity to receive ultrasonic waves passing through the test interface, and the filling head is installed at the bottom of the lower cylindrical cavity. The side of the lower skeleton ring is provided with a lower connector groove, and a lower wire groove is opened on the circumference of the bottom surface of the lower skeleton ring. The lower wire groove is connected to the lower connector groove and the lower cylindrical cavity, and a lower wire head is provided in the lower connector groove.

3. The ultrasonic identification method for anchor cable loosening in a multi-anchor cable system according to claim 2, characterized in that: Step two involves signal transmission and acquisition in two modes: Transmitted / Reflected Wave Joint Testing Mode: Applicable to newly constructed prestressed structures. An upper test ring is installed above the anchor, and a lower test ring is set between the anchor and the anchor plate. Ultrasonic signals are emitted through the upper test ring, and reflected waves are received by the upper test ring and transmitted waves are received by the lower test ring. Reflected wave test mode: Applicable to existing anchorage structures, only the test ring is installed above the anchorage, and the ultrasonic signal transmission and reflected wave reception are completed through the test ring.

4. The ultrasonic identification method for anchor cable loosening in a multi-anchor cable system according to claim 3, characterized in that: The signal processing in step three employs the following methods: (1) Transmission wave analysis: For frequency sweep signal energy: the energy density of the transmitted signal is calculated through the frequency domain energy distribution to reflect the interface pressure change; for wavelet packet energy: the transmitted wave signal is decomposed into multiple scales to extract the energy characteristics of each frequency band and construct an energy ratio index to identify load changes; by integrating multiple feature indices, a feature database of the transmitted wave is established. (2) Analysis of reflected waves: The direct pulse method is used to extract the peak amplitude and arrival time of the first wave, which is used to quickly determine the changes in contact stiffness and interface pressure; or the time inversion method is used to reverse the time of the received signal to achieve energy focusing and peak enhancement, which is suitable for low signal-to-noise ratio scenarios; or the time inversion method and the direct pulse method are combined to achieve synchronous discrimination of signal peak characteristics.

5. The ultrasonic identification method for anchor cable loosening in a multi-anchor cable system according to claim 4, characterized in that: Step four, the neural network model training steps, include: (1) Sample acquisition: Multi-channel reflected and transmitted wave signals were acquired under different tension conditions, and time-domain and frequency-domain characteristic parameters were extracted; (2) Data preprocessing: Normalization, noise reduction, and wavelet dimensionality reduction are performed; (3) Feature input layer design: Select one or more of the following as input features: energy density, wavelet packet energy, peak amplitude, arrival time, energy ratio, dominant frequency offset and propagation delay; (4) Model training and validation: Supervised learning is adopted, and the network weights are optimized through the backpropagation algorithm. (5) Model output: Output the stress level or loosening coefficient of a single anchor cable to realize the mapping relationship from signal to force.

6. The ultrasonic identification method for anchor cable loosening in a multi-anchor cable system according to claim 5, characterized in that: In step five, the automatic recognition process is as follows: (1) Signal input and matching: Input the characteristic values ​​of reflected and transmitted waves of each channel and perform similarity matching with the feature templates in the database; (2) Loosening judgment: If a certain channel feature deviates from the training mean threshold, the anchor cable is judged to be loose; (3) Quantification of loosening degree: The loosening degree is quantitatively evaluated by the percentage of interfacial contact pressure and stress loss corresponding to the loosening coefficient. It is divided into five levels: normal, slightly loose, moderately loose, significantly loose and severely loose. The value range of the loosening coefficient is 0~1. (4) Visualization and alarm output: The system automatically generates a multi-anchor cable status distribution map, locates the loosening position and marks the loosening level, and realizes intelligent monitoring and early warning of the health status of the multi-anchor cable system.

7. An ultrasonic testing system for anchor cable loosening in a multi-anchor cable system, used to perform the ultrasonic identification method for anchor cable loosening in a multi-anchor cable system as described in any one of claims 1-6, characterized in that, The device includes a dual-ring piezoelectric ultrasonic signal acquisition device, a multi-channel signal receiver and transmitter, and a computer. The dual-ring piezoelectric ultrasonic signal acquisition device includes an upper test ring and a lower test ring, which are respectively connected to the multi-channel signal receiver and transmitter via signal lines. The multi-channel signal receiver and transmitter is communicatively connected to the computer. The computer has a built-in signal processing module, a neural network model training module, and a loosening identification and evaluation module.

8. The ultrasonic testing system for anchor cable loosening in a multi-anchor cable system according to claim 7, characterized in that, The upper test ring consists of a magnetic ring, an upper skeleton ring, and an ultrasonic transmitting probe. The thrust spring inside the ultrasonic transmitting probe is used to keep the piezoelectric element in a compacted state. The isolation pad ring of the lower test ring is used to achieve high-sensitivity reception and electromagnetic isolation. The multi-channel signal receiver and transmitter is used to achieve synchronous excitation, acquisition, and transmission of ultrasonic signals.

9. The ultrasonic testing system for anchor cable loosening in a multi-anchor cable system according to claim 7, characterized in that, The signal processing module in the computer is used to perform signal denoising, filtering, feature extraction and normalization; the neural network model training module is used to construct and train a multilayer feedforward neural network or a convolutional neural network; the loosening identification and evaluation module is used to locate the loosening position of the anchor cable, quantify the degree of loosening and provide visualization output and alarm.

10. The ultrasonic testing system for anchor cable loosening in a multi-anchor cable system according to claim 7, characterized in that, It is applicable to the detection of loosening of multiple anchor cables in bridges, dams, tunnels, slope anchoring systems, foundation pit support, underground structures, wind turbine tower foundations and ground anchoring systems.

Images