A method for rapid non-destructive detection of concrete impermeability by electric flux

Abstract

The application discloses a kind of for concrete impermeability performance electric flux rapid nondestructive testing method, comprising: arranging detection patch on the surface of concrete, constructs multiple spatial distribution different structural guiding potential disturbance field;Impose short time bipolar electric excitation sequence, gather current response sequence and recovery voltage sequence;Current response sequence and recovery voltage sequence are carried out multi-scale time-frequency decomposition, and construct response fusion tensor in combination with electrode topological coding;The disturbance field electric response fusion tensor is input into improved TPA-LSTM model, and the output three kinds of dynamic response gate result;Pore group dynamics inverter is constructed, and three kinds of pore group dynamics parameters are solved according to dynamic response result;According to three kinds of pore group dynamics parameters, calculate equivalent electric flux, and output concrete impermeability performance detection result.The application is analyzed electric response signal by structural guiding potential disturbance field and improved TPA-LSTM model, realizes the electric flux rapid nondestructive testing of concrete impermeability performance.

Description

Technical Field

[0001] This invention relates to the field of concrete material performance testing technology, and in particular to a rapid non-destructive testing method for electrical flux in concrete to improve its impermeability. Background Technology

[0002] As a crucial structural material, concrete's impermeability directly impacts the durability and service life of a structure. When external moisture, chloride ions, or other corrosive media penetrate the concrete's pore structure, they can easily lead to problems such as steel reinforcement corrosion, structural cracking, and durability degradation. Therefore, accurate testing of concrete's impermeability is of paramount importance. Current engineering testing methods commonly employ electrical flux testing, resistivity testing, and water seepage testing to evaluate concrete's impermeability. Among these, the electrical flux testing method, which involves applying voltage across a concrete specimen and measuring the current change to calculate the amount of charge passing through the concrete, indirectly reflects its impermeability. This method is widely used in engineering testing.

[0003] However, traditional electrical flux testing methods typically require long-term electrical tests on standard specimens, resulting in lengthy testing cycles. Moreover, these methods largely rely on laboratory conditions, making it difficult to achieve rapid in-situ testing of existing concrete structures. Existing methods often only provide overall conductivity or overall electrical flux results, failing to distinguish the migration dynamics of different pore structures within the concrete. For example, they may reveal the response differences between different types of pores, such as connected pore groups, confined pore groups, and stagnant pore groups, during the electromigration process. In actual engineering environments, factors such as the water content of the concrete surface, the distribution of reinforcing steel, and electrode contact conditions can all interfere with the electrical response signal, thereby affecting the stability and accuracy of the test results.

[0004] To address the aforementioned issues, current technologies for evaluating the impermeability of concrete using electrical methods still lack a detection technique capable of non-destructively acquiring information on the migration dynamics of multiple types of pores within concrete in a short time. Existing detection methods struggle to effectively separate the electrical response characteristics of different pore structures under complex structural conditions, and also find it difficult to establish a stable correlation between structural disturbances, electrical response signals, and pore migration dynamics, resulting in insufficient interpretability and reliability of the detection results.

[0005] Therefore, how to provide a rapid and non-destructive testing method for the electrical flux of concrete to improve its impermeability is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a rapid, non-destructive testing method for the electrical flux of concrete to assess its impermeability. This invention acquires the electrical response signal generated by concrete under short-term bipolar electrical excitation by constructing a structure-guided potential perturbation field. It then extracts time-mode features from the electrical response sequence using multi-scale time-frequency decomposition and an improved TPA-LSTM model. Furthermore, it utilizes a pore group dynamics inversion generator to solve for the migration dynamic parameters of connected, confined, and stagnant pore groups, thereby obtaining the equivalent electrical flux of the concrete and outputting the impermeability test results. This invention fully utilizes structural perturbation electric field excitation, temporal deep learning analysis, and pore group dynamics inversion technology to achieve precise separation of the migration behavior of different pore structures within concrete. It possesses advantages such as short detection time, no structural damage, strong anti-interference capability, and high interpretability of the test results.

[0007] A rapid non-destructive testing method for electrical flux in concrete impermeability according to an embodiment of the present invention includes: Detection patches are placed on the concrete surface to be tested, and multiple spatially distributed structural directional potential disturbance fields are constructed by controlling the various electrode combinations. Under the influence of various structural directional potential disturbance fields, a short-time bipolar electric excitation sequence is applied to the concrete to be tested, and the current response sequence and recovery voltage sequence corresponding to each structural directional potential disturbance field are collected. Multi-scale time-frequency decomposition is performed on the current response sequence and the recovery voltage sequence. The obtained multi-scale components are synchronously mapped with the electrode topology code to generate a three-dimensional index of structure-time-frequency domain and construct a structure topology-driven perturbation field electric response fusion tensor. The TPA-LSTM model is improved by fusing the perturbation field electrical response with tensor input. The structural index dimension, time series dimension and frequency domain feature dimension are jointly weighted by structural constraint attention mechanism to obtain the dynamic response gating results corresponding to fast conduction response, medium-speed polarization release response and slow stagnation recovery response. A pore group dynamics inversion device is constructed. Based on the dynamic response gating results, a pore group response characteristic tensor is constructed and input into the pore group dynamics inversion device together with the field distribution information of the structure-guided potential disturbance field. The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group inside the concrete to be tested are solved. Based on the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group, the equivalent electrical flux of the concrete to be tested is determined, and the test results of the impermeability of the concrete to be tested are output.

[0008] Optionally, the detection patch includes a central excitation electrode, a ring detection electrode, a sector orientation electrode, and a guard electrode.

[0009] Optionally, constructing multiple spatially distributed structure-guided potential perturbation fields includes: A multi-level path is established between the central excitation electrode of the detection patch and each ring detection electrode, and energized sequentially from near to far to form a multi-depth radial potential perturbation field. The conduction path is switched sequentially between the central excitation electrode of the detection patch and the electrodes in different sector orientations, so that the current line rotates continuously in the circumferential direction, generating a rotating sector potential disturbance field. Simultaneously, a constraint potential with opposite polarity and controlled amplitude is applied to the guard electrode to restrict the lateral diffusion of current and form a boundary constraint type potential disturbance field. A time-division composite excitation is inserted between the multi-depth radial potential perturbation field, the rotating sector potential perturbation field, and the boundary-constrained potential perturbation field. Different field types are alternately superimposed in a preset time window by a fast electronic switching matrix to construct an interleaved depth-azimuth hybrid potential perturbation field. The multi-depth radial potential disturbance field, the rotating sector potential disturbance field, the boundary-constrained potential disturbance field, and the staggered depth-azimuth hybrid potential disturbance field are combined according to a preset order from shallow to deep and from clockwise to counterclockwise to obtain multiple structure-guided potential disturbance fields with different spatial distributions and unique index identifiers.

[0010] Optionally, the acquisition of the current response sequence and recovery voltage sequence corresponding to the structural guided potential disturbance field includes: Under the influence of the first structural guiding potential disturbance field, the positive pulse excitation stage is entered. A positive pulse voltage of preset amplitude is applied to the concrete to be tested through the central excitation electrode. During the positive pulse excitation stage, the corresponding current response signal is continuously collected to form a positive current response sequence. After the positive pulse excitation stage ends, the static stage begins. During the preset static time, the excitation voltage is stopped while the detection circuit remains connected. The recovery voltage signal generated inside the concrete is continuously acquired to form the static stage recovery voltage sequence. After the static stage, the reverse pulse excitation stage begins. A reverse pulse voltage with opposite polarity and corresponding amplitude to the positive pulse voltage is applied to the concrete to be tested through the central excitation electrode. During the reverse pulse excitation stage, the corresponding current response signal is continuously acquired to form a reverse current response sequence. After the reverse pulse excitation stage ends, the open circuit recovery stage begins. The excitation voltage circuit is disconnected within a preset recovery time and the detection electrode is kept in an open circuit state. The recovery voltage signal generated inside the concrete is continuously acquired to form the voltage sequence of the open circuit recovery stage. The forward current response sequence corresponding to the forward pulse excitation stage, the resting stage recovery voltage sequence corresponding to the resting stage, the reverse current response sequence corresponding to the reverse pulse excitation stage, and the open circuit recovery stage voltage sequence corresponding to the open circuit recovery stage are integrated in chronological order and correlated with the corresponding structure-guided potential disturbance field to obtain the current response sequence and recovery voltage sequence corresponding to each structure-guided potential disturbance field.

[0011] Optionally, the generation of the structure-time-frequency domain three-dimensional index, constructing the structure topology-driven perturbation field electrical response fusion tensor, includes: For each structure-guided potential disturbance field, the corresponding forward current response sequence, reverse current response sequence, and recovery voltage sequence are spliced ​​together in chronological order to obtain the complete electrical response sequence under the structure-guided potential disturbance field. The complete electrical response sequence is then denoised and smoothed. The complete electrical response sequence is divided into multiple time windows at two different time scales. In each time window, the amplitude features, energy features and main frequency band features that characterize the time scale are extracted to obtain the corresponding multi-scale components. Assign a time scale number and a frequency band number to each multi-scale component, and assign a unique electrode topology code to each multi-scale component based on the electrode combination method used when generating the multi-scale component, so that each multi-scale component carries a structure index, a time index, and a frequency domain index simultaneously. The multi-scale components carrying structure index, time index, and frequency index are classified and arranged in the order of priority of structure index, second time index, and third frequency index. A corresponding data block is created for each type of structure index in the data storage space, and multi-scale components with different time indexes and frequency indexes under the same structure index are grouped into the same data block. Based on the organization of the data blocks, a structure topology-driven perturbation field electrical response fusion tensor is constructed. The perturbation field electrical response fusion tensor has three dimensions: structure index, time index, and frequency domain index, and uses the amplitude characteristics of multi-scale components as unit data.

[0012] Optionally, obtaining the dynamic response gating results corresponding to the fast conduction response, medium-speed polarization release response, and slow stagnation recovery response includes: An improved TPA-LSTM model is constructed, which includes a structure encoding layer, a time-mode long short-term memory layer, and a multi-branch dynamic gated output layer. The structure encoding layer receives the fused tensor of the perturbation field electrical response. The output of the time-mode long short-term memory layer is connected to the output of the structure encoding layer, and the output of the multi-branch dynamic gated output layer is connected to the output of the time-mode long short-term memory layer. The perturbation field electrical response fusion tensor is input into the structure coding layer in the order of structure index dimension, time series dimension and frequency domain feature dimension. The time series and frequency domain features under each structure index are encoded to obtain the electrical response feature sequence carrying the structure index information. In the temporal pattern long short-term memory layer, the electrical response feature sequence is updated sequentially in the time sequence dimension to generate the hidden state sequence corresponding to each time step. The temporal pattern attention weight is calculated based on the hidden state sequence to obtain the weighted temporal pattern feature sequence. Based on the weighted temporal pattern feature sequence, a structural constraint attention mechanism is introduced to jointly assign weights to the temporal pattern features corresponding to different structural indices, resulting in a fusion feature representation that simultaneously reflects the structural index dimension, the time series dimension, and the frequency domain feature dimension. In the multi-branch dynamic gating output layer, the fused feature representation is input into three parallel gating branches to obtain the gating output corresponding to the fast turn-on response, the gating output corresponding to the medium-speed polarization release response, and the gating output corresponding to the slow stagnation recovery response. The three gating outputs are used as the dynamic response gating results for the fast turn-on response, the medium-speed polarization release response, and the slow stagnation recovery response, respectively.

[0013] Optionally, the step of solving for the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the retained pore group within the concrete to be tested includes: A pore group dynamics inversion module is constructed, which includes a pore group response reconstruction module, a structural constraint dynamics decomposition module, and a multi-perturbation field consistency correction module connected in sequence. The dynamic response gating results of the fast conduction response, medium-speed polarization release response and slow stagnation recovery response corresponding to each structure-guided potential disturbance field are input into the pore group response reconstruction module in a one-to-one correspondence with the field distribution information of each structure-guided potential disturbance field. In the pore group response reconstruction module, the dynamic response gating results are reorganized according to the arrangement of the structural index dimension, time series dimension and frequency domain feature dimension to construct the pore group response feature tensor. The pore group response feature tensor is input into the structural constraint dynamics decomposition module. The field distribution information of the structure-guided potential perturbation field is used as the structural constraint. The pore group response features under different structural indices are decomposed to obtain the initial dynamic parameters of the connected pore group, the initial dynamic parameters of the confined pore group, and the initial dynamic parameters of the stagnant pore group under each structure-guided potential perturbation field. The initial dynamic parameters of the connected hole group, the confined hole group, and the stagnant hole group are input into the multi-perturbation field consistency correction module. Cross-field comparison and constraint correction are performed on the initial dynamic parameters corresponding to different structural directional potential perturbation fields to eliminate local deviations under single perturbation field conditions. The dynamic parameters of the connected hole group, the confined hole group, and the stagnant hole group after multi-perturbation field consistency correction are obtained. The dynamic parameters of the connected pore group, corrected by the consistency of the multi-perturbation field, are determined as the migration rate of the connected pore group; the dynamic parameters of the confined pore group, corrected by the consistency of the multi-perturbation field, are determined as the polarization release rate of the confined pore group; and the dynamic parameters of the stagnant pore group, corrected by the consistency of the multi-perturbation field, are determined as the return migration rate of the stagnant pore group.

[0014] Optionally, determining the equivalent electrical flux of the concrete to be tested and outputting the test results of the concrete's impermeability includes: The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group are obtained. The three types of rate parameters are organized according to the number of the structural directional potential disturbance field to form a set of pore group dynamic parameters of the concrete to be tested. The set of pore group dynamic parameters is input into the equivalent electric flux calculation module. In the equivalent electric flux calculation module, the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group are comprehensively calculated according to the pre-calibrated weight relationship to obtain the equivalent electric flux value of the concrete to be tested. Based on the current response sequence, the current values ​​corresponding to each time sampling point are accumulated hourly, and the current accumulation is calculated by combining the current sampling time interval. The current accumulation is used as the reference value of current flux. The equivalent electrical flux value is checked for consistency with the electrical flux reference value. When the difference between the two is less than the preset error threshold, the equivalent electrical flux value is determined as the final electrical flux result of the concrete to be tested. The final electrical flux result is compared with the preset impermeability grade division interval. The impermeability grade of the concrete to be tested is determined according to the corresponding interval, and the impermeability performance test result of the concrete to be tested is output.

[0015] The beneficial effects of this invention are: This invention constructs multiple spatially distributed structural directional potential perturbation fields on the surface of the concrete to be tested, and acquires the current response sequence and recovery voltage sequence of the concrete under short-term bipolar electrical excitation conditions. This allows the dynamic information of the electromigration behavior inside the concrete to be obtained in a short time. Unlike traditional electrical flux tests that require long-term energization and rely on laboratory specimens, this invention can perform rapid testing on the actual structural surface without sampling or structural damage, improving testing efficiency and reducing the impact of the testing process on the structural body, thus achieving rapid and non-destructive evaluation of the impermeability of concrete.

[0016] This invention performs multi-scale time-frequency decomposition on the electrical response sequence and constructs a structural topology-driven perturbation field electrical response fusion tensor by combining electrode topology encoding. This allows the detection data to simultaneously contain structural information, temporal evolution information, and frequency domain distribution information. By improving the TPA-LSTM model and introducing a structural constraint attention mechanism, the invention jointly learns the electrical response characteristics under different time scales and different structural perturbation conditions, identifying three types of electrical response modes: fast conduction response, medium-speed polarization release response, and slow retention recovery response. This achieves effective separation of the migration dynamics characteristics of different pore structures inside concrete and improves the precision of electrical response analysis.

[0017] This invention constructs a pore group dynamics inversion device. By comprehensively analyzing the fast conduction response, medium-speed polarization release response, and slow retention recovery response, the migration rate of connected pore groups, the polarization release rate of confined pore groups, and the return rate of retained pore groups are solved respectively. Based on the dynamic parameters, the equivalent electrical flux of concrete is determined. Through multi-perturbation field consistency correction, the influence of factors such as surface water content, electrode contact conditions, and reinforcement distribution on the test results can be effectively reduced, improving the stability and reliability of the impermeability test results, and enabling the obtained electrical flux results to more accurately reflect the true impermeability of the concrete interior. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a rapid non-destructive testing method for electrical flux in concrete to improve its impermeability, as proposed in this invention. Figure 2 This is a schematic diagram of the improved TPA-LSTM model structure for a rapid non-destructive testing method for the electrical flux of concrete for impermeability proposed in this invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] refer to Figure 1 and Figure 2 A rapid non-destructive testing method for electrical flux in concrete to improve its impermeability includes: Detection patches are placed on the concrete surface to be tested, and multiple spatially distributed structural directional potential disturbance fields are constructed by controlling the various electrode combinations. Under the influence of various structural directional potential disturbance fields, a short-time bipolar electric excitation sequence is applied to the concrete to be tested, and the current response sequence and recovery voltage sequence corresponding to each structural directional potential disturbance field are collected. Multi-scale time-frequency decomposition is performed on the current response sequence and the recovery voltage sequence. The obtained multi-scale components are synchronously mapped with the electrode topology code to generate a three-dimensional index of structure-time-frequency domain and construct a structure topology-driven perturbation field electric response fusion tensor. The TPA-LSTM model is improved by fusing the perturbation field electrical response with tensor input. The structural index dimension, time series dimension and frequency domain feature dimension are jointly weighted by structural constraint attention mechanism to obtain the dynamic response gating results corresponding to fast conduction response, medium-speed polarization release response and slow stagnation recovery response. A pore group dynamics inversion device is constructed. Based on the dynamic response gating results, a pore group response characteristic tensor is constructed and input into the pore group dynamics inversion device together with the field distribution information of the structure-guided potential disturbance field. The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group inside the concrete to be tested are solved. Based on the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group, the equivalent electrical flux of the concrete to be tested is determined, and the test results of the impermeability of the concrete to be tested are output.

[0021] In this embodiment, the detection patch includes a central excitation electrode, a ring detection electrode, a sector orientation electrode, and a guard electrode.

[0022] In this embodiment, constructing multiple spatially distributed structure-guided potential perturbation fields includes: A multi-level path is established between the central excitation electrode of the detection patch and each ring detection electrode, and energized sequentially from near to far to form a multi-depth radial potential perturbation field. The conduction path is switched sequentially between the central excitation electrode of the detection patch and the electrodes in different sector orientations, so that the current line rotates continuously in the circumferential direction, generating a rotating sector potential disturbance field. Simultaneously, a constraint potential with opposite polarity and controlled amplitude is applied to the guard electrode to restrict the lateral diffusion of current and form a boundary constraint type potential disturbance field. A time-division composite excitation is inserted between the multi-depth radial potential perturbation field, the rotating sector potential perturbation field, and the boundary-constrained potential perturbation field. Different field types are alternately superimposed in a preset time window by a fast electronic switching matrix to construct an interleaved depth-azimuth hybrid potential perturbation field. The multi-depth radial potential disturbance field, the rotating sector potential disturbance field, the boundary-constrained potential disturbance field, and the staggered depth-azimuth hybrid potential disturbance field are combined according to a preset order from shallow to deep and from clockwise to counterclockwise to obtain multiple structure-guided potential disturbance fields with different spatial distributions and unique index identifiers.

[0023] In this embodiment, the acquisition of the current response sequence and recovery voltage sequence corresponding to the perturbation field of the structure's directional potential includes: Under the influence of the first structural guiding potential disturbance field, the positive pulse excitation stage is entered. A positive pulse voltage of preset amplitude is applied to the concrete to be tested through the central excitation electrode. During the positive pulse excitation stage, the corresponding current response signal is continuously collected to form a positive current response sequence. The positive pulse voltage of preset amplitude is an adjustable DC pulse voltage in the range of 20 volts to 40 volts. After the positive pulse excitation stage ends, the static stage begins. During the preset static time, the excitation voltage is stopped while the detection circuit is connected, and the recovery voltage signal generated inside the concrete is continuously acquired to form a static stage recovery voltage sequence. The preset static time is between 1 and 5 seconds. After the static stage, the reverse pulse excitation stage begins. A reverse pulse voltage with opposite polarity and corresponding amplitude to the positive pulse voltage is applied to the concrete to be tested through the central excitation electrode. During the reverse pulse excitation stage, the corresponding current response signal is continuously acquired to form a reverse current response sequence. After the reverse pulse excitation stage ends, the open circuit recovery stage begins. The excitation voltage circuit is disconnected within a preset recovery time and the detection electrode is kept in an open circuit state. The recovery voltage signal generated inside the concrete is continuously acquired to form an open circuit recovery stage voltage sequence. The preset recovery time is 8 to 20 seconds. The forward current response sequence corresponding to the forward pulse excitation stage, the resting stage recovery voltage sequence corresponding to the resting stage, the reverse current response sequence corresponding to the reverse pulse excitation stage, and the open circuit recovery stage voltage sequence corresponding to the open circuit recovery stage are integrated in chronological order and correlated with the corresponding structure-guided potential disturbance field to obtain the current response sequence and recovery voltage sequence corresponding to each structure-guided potential disturbance field.

[0024] In this embodiment, the generation of the structure-time-frequency domain three-dimensional index and the construction of the structure topology-driven perturbation field electrical response fusion tensor include: For each structure-guided potential disturbance field, the corresponding forward current response sequence, reverse current response sequence, and recovery voltage sequence are spliced ​​together in chronological order to obtain the complete electrical response sequence under the structure-guided potential disturbance field. The complete electrical response sequence is then denoised and smoothed. The complete electrical response sequence is divided into multiple time windows at two different time scales. In each time window, the amplitude features, energy features and main frequency band features that characterize the time scale are extracted to obtain the corresponding multi-scale components. Assign a time scale number and a frequency band number to each multi-scale component, and assign a unique electrode topology code to each multi-scale component based on the electrode combination method used when generating the multi-scale component, so that each multi-scale component carries a structure index, a time index, and a frequency domain index simultaneously. The multi-scale components carrying structure index, time index, and frequency index are classified and arranged in the order of priority of structure index, second time index, and third frequency index. A corresponding data block is created for each type of structure index in the data storage space, and multi-scale components with different time indexes and frequency indexes under the same structure index are grouped into the same data block. Based on the organization of the data blocks, a structure topology-driven perturbation field electrical response fusion tensor is constructed. The perturbation field electrical response fusion tensor has three dimensions: structure index, time index, and frequency domain index, and uses the amplitude characteristics of multi-scale components as unit data.

[0025] In this embodiment, obtaining the dynamic response gating results corresponding to the fast conduction response, medium-speed polarization release response, and slow stagnation recovery response includes: An improved TPA-LSTM model is constructed, which includes a structure encoding layer, a time-mode long short-term memory layer, and a multi-branch dynamic gated output layer. The structure encoding layer receives the fused tensor of the perturbation field electrical response. The output of the time-mode long short-term memory layer is connected to the output of the structure encoding layer, and the output of the multi-branch dynamic gated output layer is connected to the output of the time-mode long short-term memory layer. The perturbation field electrical response fusion tensor is input into the structure coding layer in the order of structure index dimension, time series dimension and frequency domain feature dimension. The time series and frequency domain features under each structure index are encoded to obtain the electrical response feature sequence carrying the structure index information. In the temporal pattern long short-term memory layer, the electrical response feature sequence is updated sequentially along the time series dimension to generate hidden state sequences corresponding to each time step. Then, temporal pattern attention weights are calculated based on these hidden state sequences to obtain a weighted temporal pattern feature sequence, where: Generate the hidden state sequence corresponding to each time step, specifically as follows: In the long short-term memory layer of the time pattern, the electrical response feature vectors of each time step are input sequentially according to the time sequence. The input features of the current time step, the hidden state of the previous time step, and the memory state are jointly updated to obtain the hidden state corresponding to the current time step. The hidden states of each time step are arranged in chronological order to form a hidden state sequence that reflects the time evolution characteristics of the electrical response. The temporal pattern attention weights are calculated based on the hidden state sequence, specifically as follows: The importance of the hidden states at each time step in the hidden state sequence is evaluated. The corresponding weight coefficient is calculated based on the contribution of each hidden state to the overall time response features. The hidden states at each time step are weighted according to their corresponding weights to obtain the weighted time pattern feature sequence. Based on the weighted temporal pattern feature sequence, a structure-constrained attention mechanism is introduced to jointly assign weights to the temporal pattern features corresponding to different structure indices, resulting in a fusion feature representation that simultaneously reflects the structure index dimension, the time series dimension, and the frequency domain feature dimension. The structure-constrained attention mechanism refers to: The structural index information is determined based on the electrode topology coding corresponding to the structure-guided potential perturbation field. The structural index information is then jointly associated with the weighted time-mode feature sequence. The structural weight coefficients are calculated for data belonging to different structural indices. When calculating the structural weight coefficients, the corresponding time-mode features are weighted according to the contribution of each structure-guided potential perturbation field to the change in electrical response. The time-mode features corresponding to each structural index are weighted and fused according to the structural weight coefficients, and combined with the corresponding frequency domain features to generate a fused feature representation that simultaneously reflects the structural index dimension, the time series dimension, and the frequency domain feature dimension. In the multi-branch dynamic gating output layer, the fused feature representation is input into three parallel gating branches to obtain the gating output corresponding to the fast turn-on response, the gating output corresponding to the medium-speed polarization release response, and the gating output corresponding to the slow stagnation recovery response. The three gating outputs are used as the dynamic response gating results for the fast turn-on response, the medium-speed polarization release response, and the slow stagnation recovery response, respectively. The fast turn-on response, the medium-speed polarization release response, and the slow stagnation recovery response are the three types of electrical response time mode features output by the improved TPA-LSTM model after learning the time mode features of the fused tensor of the electrical response of the perturbation field.

[0026] In this embodiment, the calculation of the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the retained pore group inside the concrete to be tested includes: A pore group dynamics inversion module is constructed, which includes a pore group response reconstruction module, a structural constraint dynamics decomposition module, and a multi-perturbation field consistency correction module connected in sequence. The dynamic response gating results of the fast conduction response, medium-speed polarization release response, and slow stagnation recovery response corresponding to each structure-guided potential disturbance field are input into the pore group response reconstruction module in a one-to-one correspondence with the field distribution information of each structure-guided potential disturbance field. In the pore group response reconstruction module, the dynamic response gating results are reorganized according to the arrangement of the structural index dimension, time series dimension, and frequency domain feature dimension to construct the pore group response feature tensor. Specifically, the construction of the pore group response feature tensor is as follows: The dynamic response gating results of fast conduction response, medium-speed polarization release response and slow stagnation recovery response are marked with structural index according to the structural guidance potential disturbance field number, so that each group of dynamic response gating results corresponds to a unique structural index. The dynamic response features of each time step are arranged in chronological order under the same structural index to form a data set in the time series dimension; The frequency domain features extracted from each time series data within the corresponding time window are synchronously mapped so that the response features of each time step simultaneously contain structural index information, time series information, and frequency domain feature information. The pore group response feature tensor is generated by arranging and stacking the structural index dimension, time series dimension and frequency domain feature dimension in three dimensions to characterize the electrical response characteristics of different pore groups. The pore group response feature tensor is input into the structural constraint dynamics decomposition module. Using the field distribution information of the structure-guided potential perturbation field as the structural constraint, the pore group response features under different structural indices are decomposed to obtain the initial dynamic parameters of the connected pore group, the confined pore group, and the stagnant pore group under each structure-guided potential perturbation field. Specifically, the initial dynamic parameters of the connected pore group, the confined pore group, and the stagnant pore group under each structure-guided potential perturbation field are as follows: Based on the field distribution information of the structure-guided potential disturbance field, the response features corresponding to different structure indices in the pore group response feature tensor are grouped. Under each structural index, the feature components corresponding to the fast conduction response are extracted and determined as the initial parameters of the connected pore group dynamics, the feature components corresponding to the medium-speed polarization release response are extracted and determined as the initial parameters of the confined pore group dynamics, and the feature components corresponding to the slow retention recovery response are extracted and determined as the initial parameters of the stagnant pore group dynamics. The three types of pore group dynamics initial parameters under each structural index are output. The initial dynamic parameters of the connected hole group, confined hole group, and stagnant hole group are input into the multi-perturbation field consistency correction module. Cross-field comparison and constraint correction are performed on the initial dynamic parameters corresponding to different structural directional potential perturbation fields to eliminate local deviations under single perturbation field conditions. This yields the connected hole group, confined hole group, and stagnant hole group dynamic parameters after multi-perturbation field consistency correction. Specifically, the cross-field comparison and constraint correction of the initial dynamic parameters corresponding to different structural directional potential perturbation fields involves: The initial dynamic parameters of the connected pore group, the confined pore group, and the stagnant pore group corresponding to each perturbation field are classified and collected according to the number of the structure-guided potential perturbation field. The initial dynamic parameters of the same type of pore group obtained under different structurally guided potential perturbation fields are compared laterally to identify anomalous parameters that deviate from the overall trend. The abnormal parameters are adjusted according to the preset consistency constraint rules, and the dynamic initial parameters corresponding to each disturbance field are comprehensively corrected. The dynamic parameters of the connected hole group, the confined hole group, and the stagnant hole group after multi-disturbance field consistency correction are output. The dynamic parameters of the connected pore group, corrected by the consistency of the multi-perturbation field, are determined as the migration rate of the connected pore group; the dynamic parameters of the confined pore group, corrected by the consistency of the multi-perturbation field, are determined as the polarization release rate of the confined pore group; and the dynamic parameters of the stagnant pore group, corrected by the consistency of the multi-perturbation field, are determined as the return migration rate of the stagnant pore group.

[0027] In this embodiment, determining the equivalent electrical flux of the concrete to be tested and outputting the test results of the concrete's impermeability includes: The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group are obtained. The three types of rate parameters are organized according to the number of the structural directional potential disturbance field to form a set of pore group dynamic parameters of the concrete to be tested. The set of pore group dynamic parameters is input into the equivalent electric flux calculation module. In this module, the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group are comprehensively calculated according to a pre-calibrated weighting relationship to obtain the equivalent electric flux value of the concrete to be tested. The pre-calibrated weighting relationship is as follows: the migration rate of the connected pore group has a weighting coefficient of 0.50–0.65; the polarization release rate of the confined pore group has a weighting coefficient of 0.20–0.35; and the return rate of the stagnant pore group has a weighting coefficient of 0.10–0.20. Specifically, the equivalent electric flux value of the concrete to be tested is as follows: The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group are weighted with their respective weighting coefficients to obtain the contribution value of each pore group to the electric flux. The electric flux contribution values ​​corresponding to each pore group are superimposed to obtain the comprehensive electric flux result. The comprehensive electric flux result is averaged according to the number of structural directional potential disturbance fields to obtain the equivalent electric flux value of the concrete to be tested. Based on the current response sequence, the current values ​​corresponding to each time sampling point are accumulated hourly, and the current accumulation is calculated by combining the current sampling time interval. The current accumulation is used as the reference value of current flux. The equivalent electrical flux value is checked for consistency with the electrical flux reference value. When the difference between the two is less than the preset error threshold, the equivalent electrical flux value is determined as the final electrical flux result of the concrete to be tested. The final electrical flux result is compared with the preset impermeability grade division interval. The impermeability grade of the concrete to be tested is determined according to the corresponding interval, and the impermeability performance test result of the concrete to be tested is output.

[0028] Example 1:

[0029] To verify the feasibility of this invention in practice, it was applied to the pier structure of a highway bridge in a coastal area. The pier concrete design strength grade was C40, and the environment was a typical marine salt spray environment. The concrete surface at a height of approximately 4.5 meters in the middle of the pier was selected as the detection area. Detection patches were deployed on the concrete surface, including a central excitation electrode, a ring-shaped detection electrode, a sector orientation electrode, and a guard electrode. Multiple spatially distributed structural directional potential perturbation fields were constructed by controlling different electrode combinations. Under each structural directional potential perturbation field condition, a short-time bipolar electrical excitation sequence was applied to the concrete through the central excitation electrode. This excitation sequence sequentially included a forward pulse excitation stage, a static stage, a reverse pulse excitation stage, and an open-circuit recovery stage. The corresponding current response sequence and recovery voltage sequence were collected at each stage.

[0030] Subsequently, the acquired current response sequence and recovery voltage sequence are subjected to multi-scale time-frequency decomposition, and the obtained multi-scale components are synchronously mapped with electrode topology encoding to generate a three-dimensional index structure containing structural information, time series information, and frequency domain information, thereby constructing a structure topology-driven perturbation field electrical response fusion tensor. This fusion tensor is input into an improved TPA-LSTM model, in which a structural constraint attention mechanism is used to jointly learn the weights of the structural index dimension, time series dimension, and frequency domain feature dimension, thereby obtaining three types of dynamic response gating results: fast conduction response, medium-speed polarization release response, and slow retention recovery response. These three types of dynamic response results are then input into a pore group dynamics inversion module, which includes a pore group response reconstruction module, a structural constraint dynamics decomposition module, and a multi-perturbation field consistency correction module. Through processing by these modules, the migration rate of connected pore groups, the polarization release rate of confined pore groups, and the return rate of retained pore groups are obtained, respectively. Based on these three types of dynamic parameters, the equivalent electrical flux value of the concrete to be tested is calculated.

[0031] In this embodiment, the same test area of ​​the bridge pier is repeatedly tested six times, with each test lasting about 50 to 65 seconds and the test interval being about 3 minutes, in order to reduce the influence of environmental factors. After the bridge test is completed, samples are taken from the vicinity of the same bridge pier to make standard test specimens, and the results are compared using a laboratory standard electrical flux test.

[0032] Table 1 Comparison of Field Test Results and Standard Electrical Flux Test Results

[0033] As shown in Table 1, when six repeated tests were conducted within the same test area of ​​the bridge pier concrete, the single test time of the method of this invention remained between 55 and 63 seconds, with an average test time of approximately 59 seconds. The overall test process was stable and time-efficient. This indicates that under on-site testing conditions, by deploying test patches and applying a short-duration bipolar electrical excitation sequence, the acquisition and subsequent analysis of electrical response signals can be completed in a short time. Compared to the traditional standard electrical flux test, which typically requires several hours, the method of this invention has a significant advantage in testing efficiency and can meet the needs of rapid on-site testing in engineering projects.

[0034] Further analysis of the pore group dynamic parameters reveals that the migration rate of the connected pore group obtained from each test remained stable between 0.72 and 0.75, the polarization release rate of the confined pore group remained stable between 0.45 and 0.47, and the return migration rate of the stagnant pore group remained stable between 0.37 and 0.39. The overall fluctuation range is relatively small, indicating that by collecting electrical response data through the structure-guided potential disturbance field and using the improved TPA-LSTM model for dynamic response gating analysis, the migration dynamic characteristics corresponding to different pore groups inside concrete can be stably identified. The small variation in the dynamic parameters of the three types of pore groups obtained from multiple tests also demonstrates that the method of this invention still has good repeatability and stability under complex on-site environmental conditions.

[0035] Comparing the electrical flux test results, the equivalent electrical flux value calculated by the method of this invention ranges from 1665 coulombs to 1694 coulombs, while the laboratory standard electrical flux test results range from 1708 coulombs to 1730 coulombs. The difference between the two is approximately 30 to 50 coulombs, with a relative error of about 2% to 4%. The overall deviation is small and the trend is consistent. This indicates that the migration rate of connected pore groups, the polarization release rate of confined pore groups, and the return migration rate of retained pore groups obtained through pore group dynamics inversion can effectively characterize the electromigration characteristics inside concrete and accurately estimate the electrical flux level of concrete. In summary, the method of this invention shortens the testing time while ensuring the accuracy of the test results and can complete the assessment of impermeability without damaging the structure, thus possessing high engineering application value.

[0036] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A rapid non-destructive testing method for electrical flux in concrete to assess its impermeability, characterized in that, include: Detection patches are placed on the concrete surface to be tested, and multiple spatially distributed structural directional potential disturbance fields are constructed by controlling the various electrode combinations. Under the influence of various structural directional potential disturbance fields, a short-time bipolar electric excitation sequence is applied to the concrete to be tested, and the current response sequence and recovery voltage sequence corresponding to each structural directional potential disturbance field are collected. Multi-scale time-frequency decomposition is performed on the current response sequence and the recovery voltage sequence. The obtained multi-scale components are synchronously mapped with the electrode topology code to generate a three-dimensional index of structure-time-frequency domain and construct a structure topology-driven perturbation field electric response fusion tensor. The TPA-LSTM model is improved by fusing the perturbation field electrical response with tensor input. The structural index dimension, time series dimension and frequency domain feature dimension are jointly weighted by structural constraint attention mechanism to obtain the dynamic response gating results corresponding to fast conduction response, medium-speed polarization release response and slow stagnation recovery response. A pore group dynamics inversion device is constructed. Based on the dynamic response gating results, a pore group response characteristic tensor is constructed and input into the pore group dynamics inversion device together with the field distribution information of the structure-guided potential disturbance field. The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group inside the concrete to be tested are solved. Based on the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group, the equivalent electrical flux of the concrete to be tested is determined, and the test results of the impermeability of the concrete to be tested are output.

2. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The detection patch includes a central excitation electrode, a ring detection electrode, a sector orientation electrode, and a guard electrode.

3. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The construction of multiple spatially distributed structure-guided potential perturbation fields includes: A multi-level path is established between the central excitation electrode of the detection patch and each ring detection electrode, and energized sequentially from near to far to form a multi-depth radial potential perturbation field. The conduction path is switched sequentially between the central excitation electrode of the detection patch and the electrodes in different sector orientations, so that the current line rotates continuously in the circumferential direction, generating a rotating sector potential disturbance field. Simultaneously, a constraint potential with opposite polarity and controlled amplitude is applied to the guard electrode to restrict the lateral diffusion of current and form a boundary constraint type potential disturbance field. A time-division composite excitation is inserted between the multi-depth radial potential perturbation field, the rotating sector potential perturbation field, and the boundary-constrained potential perturbation field. Different field types are alternately superimposed in a preset time window by a fast electronic switching matrix to construct an interleaved depth-azimuth hybrid potential perturbation field. The multi-depth radial potential disturbance field, the rotating sector potential disturbance field, the boundary-constrained potential disturbance field, and the staggered depth-azimuth hybrid potential disturbance field are combined according to a preset order from shallow to deep and from clockwise to counterclockwise to obtain multiple structure-guided potential disturbance fields with different spatial distributions and unique index identifiers.

4. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The acquisition of the current response sequence and recovery voltage sequence corresponding to the perturbation field of the structure's guided potential includes: Under the influence of the first structural guiding potential disturbance field, the positive pulse excitation stage is entered. A positive pulse voltage of preset amplitude is applied to the concrete to be tested through the central excitation electrode. During the positive pulse excitation stage, the corresponding current response signal is continuously collected to form a positive current response sequence. After the positive pulse excitation stage ends, the static stage begins. During the preset static time, the excitation voltage is stopped while the detection circuit remains connected. The recovery voltage signal generated inside the concrete is continuously acquired to form the static stage recovery voltage sequence. After the static stage, the reverse pulse excitation stage begins. A reverse pulse voltage with opposite polarity and corresponding amplitude to the positive pulse voltage is applied to the concrete to be tested through the central excitation electrode. During the reverse pulse excitation stage, the corresponding current response signal is continuously acquired to form a reverse current response sequence. After the reverse pulse excitation stage ends, the open circuit recovery stage begins. The excitation voltage circuit is disconnected within a preset recovery time and the detection electrode is kept in an open circuit state. The recovery voltage signal generated inside the concrete is continuously acquired to form the voltage sequence of the open circuit recovery stage. The forward current response sequence corresponding to the forward pulse excitation stage, the resting stage recovery voltage sequence corresponding to the resting stage, the reverse current response sequence corresponding to the reverse pulse excitation stage, and the open circuit recovery stage voltage sequence corresponding to the open circuit recovery stage are integrated in chronological order and correlated with the corresponding structure-guided potential disturbance field to obtain the current response sequence and recovery voltage sequence corresponding to each structure-guided potential disturbance field.

5. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The generated structure-time-frequency domain three-dimensional index constructs a structure topology-driven perturbation field electrical response fusion tensor, including: For each structure-guided potential disturbance field, the corresponding forward current response sequence, reverse current response sequence, and recovery voltage sequence are spliced ​​together in chronological order to obtain the complete electrical response sequence under the structure-guided potential disturbance field. The complete electrical response sequence is then denoised and smoothed. The complete electrical response sequence is divided into multiple time windows at two different time scales. In each time window, the amplitude features, energy features and main frequency band features that characterize the time scale are extracted to obtain the corresponding multi-scale components. Assign a time scale number and a frequency band number to each multi-scale component, and assign a unique electrode topology code to each multi-scale component based on the electrode combination method used when generating the multi-scale component, so that each multi-scale component carries a structure index, a time index, and a frequency domain index simultaneously. The multi-scale components carrying structure index, time index, and frequency index are classified and arranged in the order of priority of structure index, second time index, and third frequency index. A corresponding data block is created for each type of structure index in the data storage space, and multi-scale components with different time indexes and frequency indexes under the same structure index are grouped into the same data block. Based on the organization of the data blocks, a structure topology-driven perturbation field electrical response fusion tensor is constructed. The perturbation field electrical response fusion tensor has three dimensions: structure index, time index, and frequency domain index, and uses the amplitude characteristics of multi-scale components as unit data.

6. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The obtained dynamic response gating results for the corresponding fast conduction response, medium-speed polarization release response, and slow stagnation recovery response include: An improved TPA-LSTM model is constructed, which includes a structure encoding layer, a time-mode long short-term memory layer, and a multi-branch dynamic gated output layer. The structure encoding layer receives the fused tensor of the perturbation field electrical response. The output of the time-mode long short-term memory layer is connected to the output of the structure encoding layer, and the output of the multi-branch dynamic gated output layer is connected to the output of the time-mode long short-term memory layer. The perturbation field electrical response fusion tensor is input into the structure coding layer in the order of structure index dimension, time series dimension and frequency domain feature dimension. The time series and frequency domain features under each structure index are encoded to obtain the electrical response feature sequence carrying the structure index information. In the temporal pattern long short-term memory layer, the electrical response feature sequence is updated sequentially in the time sequence dimension to generate the hidden state sequence corresponding to each time step. The temporal pattern attention weight is calculated based on the hidden state sequence to obtain the weighted temporal pattern feature sequence. Based on the weighted temporal pattern feature sequence, a structural constraint attention mechanism is introduced to jointly assign weights to the temporal pattern features corresponding to different structural indices, resulting in a fusion feature representation that simultaneously reflects the structural index dimension, the time series dimension, and the frequency domain feature dimension. In the multi-branch dynamic gating output layer, the fused feature representation is input into three parallel gating branches to obtain the gating output corresponding to the fast turn-on response, the gating output corresponding to the medium-speed polarization release response, and the gating output corresponding to the slow stagnation recovery response. The three gating outputs are used as the dynamic response gating results for the fast turn-on response, the medium-speed polarization release response, and the slow stagnation recovery response, respectively.

7. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The calculation of the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the retained pore group within the concrete to be tested includes: A pore group dynamics inversion module is constructed, which includes a pore group response reconstruction module, a structural constraint dynamics decomposition module, and a multi-perturbation field consistency correction module connected in sequence. The dynamic response gating results of the fast conduction response, medium-speed polarization release response and slow stagnation recovery response corresponding to each structure-guided potential disturbance field are input into the pore group response reconstruction module in a one-to-one correspondence with the field distribution information of each structure-guided potential disturbance field. In the pore group response reconstruction module, the dynamic response gating results are reorganized according to the arrangement of the structural index dimension, time series dimension and frequency domain feature dimension to construct the pore group response feature tensor. The pore group response feature tensor is input into the structural constraint dynamics decomposition module. The field distribution information of the structure-guided potential perturbation field is used as the structural constraint. The pore group response features under different structural indices are decomposed to obtain the initial dynamic parameters of the connected pore group, the initial dynamic parameters of the confined pore group, and the initial dynamic parameters of the stagnant pore group under each structure-guided potential perturbation field. The initial dynamic parameters of the connected hole group, the confined hole group, and the stagnant hole group are input into the multi-perturbation field consistency correction module. Cross-field comparison and constraint correction are performed on the initial dynamic parameters corresponding to different structural directional potential perturbation fields to eliminate local deviations under single perturbation field conditions. The dynamic parameters of the connected hole group, the confined hole group, and the stagnant hole group after multi-perturbation field consistency correction are obtained. The dynamic parameters of the connected pore group, corrected by the consistency of the multi-perturbation field, are determined as the migration rate of the connected pore group; the dynamic parameters of the confined pore group, corrected by the consistency of the multi-perturbation field, are determined as the polarization release rate of the confined pore group; and the dynamic parameters of the stagnant pore group, corrected by the consistency of the multi-perturbation field, are determined as the return migration rate of the stagnant pore group.

8. The rapid non-destructive testing method for electrical flux in concrete impermeability according to claim 1, characterized in that, The process of determining the equivalent electrical flux of the concrete to be tested and outputting the test results of the concrete's impermeability includes: The migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return migration rate of the stagnant pore group are obtained. The three types of rate parameters are organized according to the number of the structural directional potential disturbance field to form a set of pore group dynamic parameters of the concrete to be tested. The set of pore group dynamic parameters is input into the equivalent electric flux calculation module. In the equivalent electric flux calculation module, the migration rate of the connected pore group, the polarization release rate of the confined pore group, and the return rate of the stagnant pore group are comprehensively calculated according to the pre-calibrated weight relationship to obtain the equivalent electric flux value of the concrete to be tested. Based on the current response sequence, the current values ​​corresponding to each time sampling point are accumulated hourly, and the current accumulation is calculated by combining the current sampling time interval. The current accumulation is used as the reference value of current flux. The equivalent electrical flux value is checked for consistency with the electrical flux reference value. When the difference between the two is less than the preset error threshold, the equivalent electrical flux value is determined as the final electrical flux result of the concrete to be tested. The final electrical flux result is compared with the preset impermeability grade division interval. The impermeability grade of the concrete to be tested is determined according to the corresponding interval, and the impermeability performance test result of the concrete to be tested is output.

Images