A method for online monitoring of bolt preload in steel pipes

Abstract

This invention relates to the field of preload monitoring technology, and more particularly to an online monitoring method for bolt preload in steel pipes. First, an original digital signal sequence is acquired based on collected electromagnetic resonance and acoustic reflection signals. After preprocessing and truncation, a time-series segment set is obtained. A structural kernel mapping coding mechanism is introduced to construct a set of structural features. Then, based on the set of structural features, an asymmetric topological graph is constructed, the local topological perturbation intensity of the nodes is calculated, and converted into a topological tension field in the time domain. Based on the preprocessed digital signal sequence, the structural kernel features of the preprocessed digital signal sequence are obtained, the structural perturbation quantization is calculated, and combined with the topological tension field in the time domain, the perturbation inversion structural response function is obtained, and the predicted value of the actual preload is calculated. Finally, the predicted value of the actual preload is corrected to obtain the monitoring result. This method solves the technical problem of non-contact, continuous, and high-precision online monitoring of bolt preload inside closed steel pipes.

Description

Technical Field

[0001] This invention relates to the field of preload monitoring technology, and in particular to an online monitoring method for bolt preload in steel pipes. Background Technology

[0002] In large steel structure projects such as bridges, wind turbine towers, and high-voltage transmission pipelines, high-strength bolts are typically used to connect components to ensure the overall stability and safety of the structure. After installation, the preload level of the bolts directly affects the fatigue resistance and load transfer capacity of the connection node; therefore, monitoring the bolt preload is a crucial aspect of structural safety maintenance. Currently used methods for bolt preload detection include strain gauge methods, ultrasonic echo methods, fiber optic grating sensing methods, and image recognition methods, but most of these methods have significant limitations. In practical applications, the internal space of enclosed steel pipe structures is limited, and bolts are invisible and inaccessible, making it impossible to effectively deploy traditional contact sensors. Furthermore, significant electromagnetic interference, temperature variations, and mechanical vibrations in the environment can severely affect sensor stability and signal acquisition accuracy. Existing methods mostly only acquire single tension states at specific points in time, failing to meet the requirement of continuous monitoring of tension degradation trends. They also lack in-depth modeling capabilities for the nonlinear relationship between bolt structural response characteristics and tension, resulting in insufficient prediction accuracy and reliability in multi-interference environments. Summary of the Invention

[0003] This invention provides an online monitoring method for bolt preload in steel pipes, to solve the technical problem of non-contact, continuous, and high-precision online monitoring of bolt preload inside closed steel pipes.

[0004] The present invention provides an online monitoring method for the preload of bolts inside steel pipes, comprising the following steps:

[0005] S1. Based on the acquired electromagnetic resonance signal and acoustic reflection signal, the original digital signal sequence is obtained; the original digital signal sequence is preprocessed to obtain the preprocessed digital signal sequence; based on the preprocessed digital signal sequence, a time-series segment set is obtained; based on the time-series segment set, a structure kernel mapping coding mechanism is introduced to construct a set of structural features;

[0006] S2. Based on the set of structural features, construct an asymmetric topological graph and calculate the structural correlation degree between nodes in the asymmetric topological graph; based on the structural correlation degree, calculate the local topological perturbation intensity of the nodes and convert the local topological perturbation intensity into a topological tension field in the time domain;

[0007] S3. Based on the preprocessed digital signal sequence, obtain the structural kernel features of the preprocessed digital signal sequence and calculate the structural perturbation quantization; based on the structural perturbation quantization and the topological tension field in the time domain, obtain the perturbation inversion structural response function; based on the perturbation inversion structural response function, obtain the actual preload prediction value; correct the actual preload prediction value to obtain the corrected bolt preload; based on the corrected bolt preload, obtain the monitoring results.

[0008] Preferably, S1 specifically includes:

[0009] In the structure kernel mapping coding mechanism, based on the time segments in the time segment set, the instantaneous curvature, local oscillation rate and modulation speed factor of the time segment are calculated as the structure kernel features of the time segment, which constitute the set of structural features of the time segment.

[0010] Preferably, S2 specifically includes:

[0011] In asymmetric topological graphs, based on the set of structural features of time-series segments, structural dissimilarity is introduced to calculate the structural correlation between nodes.

[0012] Preferably, S2 specifically includes:

[0013] In asymmetric topological graphs, the local topological perturbation intensity of a node is calculated based on the structural correlation between a node and its neighboring nodes, combined with the local oscillation rate in the set of structural features of a time-series segment.

[0014] Preferably, S3 specifically includes:

[0015] Based on the preprocessed digital signal sequence, a structure kernel mapping coding mechanism is introduced to calculate the instantaneous curvature, local oscillation rate, and modulation speed factor of the preprocessed digital signal sequence, thereby obtaining the structure kernel features of the preprocessed digital signal sequence.

[0016] Preferably, S3 specifically includes:

[0017] Based on the structural kernel characteristics of the preprocessed digital signal sequence, an energy distribution amplification factor is introduced to calculate the quantization of structural perturbation.

[0018] Preferably, S3 specifically includes:

[0019] The product of the structural perturbation quantization and the topological tension field in the time domain is used as the perturbation-propagation coupling coefficient, and a signal acceleration term is introduced to form the perturbation inversion structural response function.

[0020] Preferably, S3 specifically includes:

[0021] Based on the perturbation-inverted structural response function, combined with the structural kernel characteristics and structural perturbation quantization of the preprocessed digital signal sequence, a tension prediction estimation formula is constructed to calculate the predicted value of the actual preload.

[0022] The beneficial effects of the technical solution of the present invention are:

[0023] 1. By performing local differentiation, modulation rate factor calculation, and local oscillation rate calculation on a time series set with a fixed time step, structural features such as instantaneous curvature, local oscillation rate, and modulation rate factor are introduced to effectively extract the micro-deformation features of the digital signal caused by tension. This solves the problems of slow response and low recognition ability of conventional methods to weak tension drift and improves the analytical resolution of the true state of the bolt.

[0024] 2. By establishing an asymmetric topological graph with structural feature sets as node features, and using the morphological similarity kernel function and modulation direction consistency weighting factor to construct structural correlation, the potential energy flow of tension disturbance is further calculated to form local topological disturbance intensity. This effectively characterizes the diffusion characteristics of tension changes in the signal structure space, and can identify potential structural relaxation trends in advance, thus realizing early warning modeling of bolt preload degradation process.

[0025] 3. The product of the structural perturbation quantization and the topological tension field in the time domain is introduced as the perturbation-propagation coupling coefficient, and combined with the signal acceleration term to form the perturbation inversion structural response function. The physical driving relationship between the micro-perturbation signal and the macro-bolt preload is theoretically established, which improves the physical interpretability and engineering transferability. Attached Figure Description

[0026] Figure 1 This is a flowchart of an online monitoring method for bolt preload in steel pipes according to the present invention. Detailed Implementation

[0027] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0029] The following description, in conjunction with the accompanying drawings, details the specific scheme of the online monitoring method for bolt preload in steel pipes provided by this invention.

[0030] See attached document Figure 1 The diagram illustrates a flowchart of an online monitoring method for bolt preload in a steel pipe according to an embodiment of the present invention. The method includes the following steps:

[0031] S1. Based on the acquired electromagnetic resonance signal and acoustic reflection signal, the original digital signal sequence is obtained; the original digital signal sequence is preprocessed to obtain the preprocessed digital signal sequence; based on the preprocessed digital signal sequence, a time-series segment set is obtained; based on the time-series segment set, a structure kernel mapping coding mechanism is introduced to construct a set of structural features.

[0032] By installing electromagnetic excitation coils and ultrasonic transducers on the outer wall of the steel pipe, high-frequency electromagnetic excitation signals and pulsed acoustic signals are emitted to the internal bolt structure of the pipe in a non-contact manner. The resulting electromagnetic resonance signals and acoustic reflection signals are collected by electromagnetic induction coils and ultrasonic receivers deployed outside the steel pipe, respectively. After analog-to-digital conversion, a raw digital signal sequence containing time stamps is formed. The internal bolt structure of the pipe is the stress- and signal-response part composed of the bolt body and the contact interface between it and the inner wall of the steel pipe or connecting components. The bolt body includes the screw part and the nut part. The electromagnetic excitation coils and ultrasonic transducers are pre-embedded by professionals in the ferromagnetic resonant structure and acoustic reflection cavity of the internal bolt structure of the pipe. Under the actual preload of the bolts, the internal bolt structure of the pipe will undergo resonance and echo responses related to axial stress coupling under the action of electromagnetic fields and acoustic waves.

[0033] The original digital signal sequence is preprocessed, including denoising and normalization, to obtain the preprocessed digital signal sequence. ,in, Indicates the current time The preprocessed digital signal sequence is then truncated using an existing window function to obtain a set of time-series segments with fixed time steps. ,in, It is the total number of segments in a time-series segment set with a fixed time step; It is a time-series segment index; It is the current time. The Each time segment. The preprocessing method is a technique well-known to those skilled in the art and will not be described in detail here.

[0034] Based on a set of time series segments with a fixed time step, a structure kernel mapping coding mechanism is introduced to construct a set of structural features for the time series segments. ,in, , , This represents the structural features in the structural feature set constructed through the structural kernel mapping coding mechanism, specifically the structural kernel features of a time series segment. The structural kernel mapping coding mechanism extracts multi-dimensional statistical features, including the first derivative, second derivative, root mean square value, and local curvature of the time series segment, through local differentiation operations, modulation filter banks, and normalization operations, further constructing the structural feature set. The specific calculation process is as follows:

[0035] Instantaneous curvature The calculation formula is:

[0036] ,

[0037] in, Indicates the current time The The instantaneous curvature of a time segment; It is the current time. The The first derivative of each time segment is obtained by first-order difference calculation; It is the current time. The The second derivative of each time segment is obtained through second-order difference calculation. The formula for calculating instantaneous curvature reflects the waveform response sensitivity under structural disturbances by quantifying the degree of local deformation of the digital signal of the time segment at a certain instant.

[0038] Local oscillation rate The calculation formula is:

[0039] ,

[0040] in, Indicates the current time The The local oscillation rate of each time segment is used to measure the sliding window. The deviation trend of the amplitude of the internal digital signal relative to the mean; Based on the current time Centered on, with a length of The average value of the preprocessed digital signal sequence within the sliding window; It is a very small constant, usually 1 / 2. ; It is the length of the semi-sliding window used for local sliding window integration calculations, representing the scale of local response identification in the time domain. It reflects the time span within which the intensity of digital signal oscillations is considered associated with a certain tension perturbation, determined by the short-time Fourier transform. The dominant frequency of the preprocessed digital signal sequence of each time segment is determined by inverse calculation, with a reference range of values. ; It is a time integral variable; Indicates time The A time sequence segment; It is the inverse deviation normalization term, used to measure the current time. The reciprocal of the magnitude deviation of a time segment from the average value of the preprocessed digital signal sequence within the sliding window; It is a quantification term of local oscillation intensity, which describes the degree of local stress release under structural disturbance.

[0041] Modulation rate factor The calculation formula is:

[0042] ,

[0043] in, Indicates the current time The The modulation rate factor of each time segment is used to capture the periodic digital signal change response caused by deformation after the bolt is stressed. It is the projection of the modulation rate of the preprocessed digital signal sequence under a specific period. The preprocessed digital signal sequence dominates the modulation period, which can be obtained from the current time using Fourier transform. The a time sequence segment The main peak frequency is extracted and calculated in reverse.

[0044] S2. Based on the set of structural features, construct an asymmetric topological graph and calculate the structural correlation degree between nodes in the asymmetric topological graph; based on the structural correlation degree, calculate the local topological perturbation intensity of the nodes and convert the local topological perturbation intensity into a topological tension field in the time domain.

[0045] By combining the three types of structural features from the structural feature set of each time series segment as node features, i.e., the structural feature set of each time series segment, an asymmetric topological graph is constructed. ,in, It is a set of nodes, representing all time-series segments. It is a set of edges. The nodes in the set of nodes. Indicates the first The structural state of each time segment, and the weights of the edges in the edge set. Reflecting the synergy or heterogeneity between two structural states, i.e., nodes and nodes Synergy or heterogeneity between them Indicates the first The structural state of each time segment. The formula for calculating structural correlation is introduced as follows:

[0046] ,

[0047] in, It is a node and nodes The strength of the topological connections between them, i.e., the degree of structural correlation; This is the scaling factor for normalizing structural dissimilarity, i.e., the Gaussian kernel bandwidth parameter. It is set based on the Gaussian kernel scale determined by the mean square error of the structural kernel features, with a reference value range of [range to be specified]. The mean squared error of the structural kernel features is calculated by squaring the deviations of the structural kernel feature values ​​of each time segment from the mean during the calculation of the distribution range of the structural kernel features of time segments. It reflects the degree of dispersion of the features. It is the first The first time segment and the first The phase steering angle of each time segment is calculated based on cosine similarity. Indicates the first The instantaneous curvature of a time segment; Indicates the first Modulation rate factor for each time segment; Indicates the degree of structural difference; The kernel function representing the morphological similarity between two nodes; This represents the modulation direction consistency weighting factor, used to strengthen the edge connection probability between nodes with similar modulation trends.

[0048] The structural correlation calculation formula realizes the weighted connection between signal microstructures based on local response morphology, so as to highlight the asymmetric feature propagation mechanism caused by structural stress.

[0049] Furthermore, the tension perturbation potential energy flow of each node in the asymmetric topology is calculated using the following formula:

[0050] ,

[0051] in, It is a node The local topological perturbation intensity, i.e. the tension perturbation potential energy flow of the node, is used to reflect the degree of non-equilibrium of the morphological change trend in the topological structure. It is a node The neighborhood of a node is defined as the neighborhood of nodes in an asymmetric topological graph. A set of nodes connected by edges; It is an index of the neighboring nodes; It is a node and neighboring nodes The strength of the topological connections between them, i.e., the degree of structural correlation; Indicates the first The local oscillation rate of a time series segment. Tension perturbation potential energy flow reflects the intensity of oscillation perturbation propagation within the local topological range of a certain structural state. The larger the value of tension perturbation potential energy flow, the more likely there are signs of structural imbalance.

[0052] To transform the flow of tension perturbation potential energy at nodes into a continuous-time function for subsequent integration operations, the existing weighted kernel density estimation method is used to estimate the local topological perturbation intensity. Topological tension field converted to the time domain That is, the potential field for the propagation of tension perturbation, which participates in the calculation of the subsequent perturbation inversion structural response function.

[0053] S3. Based on the preprocessed digital signal sequence, obtain the structural kernel features of the preprocessed digital signal sequence and calculate the structural perturbation quantization; based on the structural perturbation quantization and the topological tension field in the time domain, obtain the perturbation inversion structural response function; based on the perturbation inversion structural response function, obtain the actual preload prediction value; correct the actual preload prediction value to obtain the corrected bolt preload; based on the corrected bolt preload, obtain the monitoring results.

[0054] To simulate the response change process of the preprocessed digital signal sequence under tension reduction or stress change conditions, a structural perturbation quantization quantity is introduced. The calculation formula is:

[0055] ,

[0056] in, It is in time Quantization of structural disturbance at the location; This is the disturbance response adjustment factor, a coefficient used to represent the control of the overall disturbance response strength. It is determined with reference to existing propagation loss models, and the reference value range is [range missing]. ; It is a preprocessed digital signal sequence In time The instantaneous curvature; It is a preprocessed digital signal sequence In time Modulation rate factor; It is a preprocessed digital signal sequence In time Local oscillation rate; structural kernel features of the preprocessed digital signal sequence , , The calculation method and steps of the S1 structure kernel mapping coding mechanism are based on the structure kernel features of the temporal segments. , , The calculation method is the same; It is the reference oscillation rate, used to represent the baseline oscillation amplitude of the bolt structure under healthy conditions. It is obtained by averaging healthy state samples retrieved from the existing database. Expressing the geometric or dynamic response intensity of a preprocessed digital signal sequence under structural perturbation; It is an energy distribution amplification factor, introduced by the local oscillation rate. The nonlinear amplification effect is used to enhance the sensitivity to large oscillation responses. The formula for calculating the quantization of structural perturbations reflects the sensitivity of microstructural changes to digital signals.

[0057] Furthermore, the perturbation-inverted structural response function is defined. :

[0058] ,

[0059] in, It is the perturbation-inverted structural response function, used to characterize the time... The total energy response of a bolted structure after being disturbed; It is a time integral variable; It is time Quantization of structural perturbations; Indicates time The topological tension field in the time domain; It is time The preprocessed digital signal sequence; It is the signal acceleration term, that is, the acceleration in the preprocessed digital signal sequence. It is the second derivative and is used to estimate the intensity of digital signal fluctuations and deformation rate. It is sensitive to digital signal oscillations and is calculated by the finite difference numerical method. It is the length of the time window of the preprocessed digital signal sequence, that is, the total sampling time range of the original digital signal sequence, which is determined according to specific requirements; It is the disturbance-propagation coupling coefficient, used to describe the likelihood that a local disturbance in the structure will cause a system-level response in the topological propagation structure; It describes how local structural disturbances propagate at a specific intensity and cause acceleration changes in a preprocessed digital signal sequence.

[0060] Further, in the tension prediction stage, the structural kernel features of the preprocessed digital signal sequence, the quantization of structural perturbations, and the perturbation-inverted structural response function are integrated to construct the tension prediction estimation formula as follows:

[0061] ,

[0062] in, It is the predicted value of the current actual preload of the bolt; It is the mapping scaling factor, used to map the integral result from energy or structural quantity to engineering tension unit, and can be determined through experimental fitting such as least squares method, support vector machine or neural network regression model; Indicates the oscillation acceleration of the signal; It is a time window weighting function used to enhance the weight of the digital signal in the middle of the time window. To avoid noise caused by end truncation, cosine window weighting is introduced. It is the energy weighting factor for topological inversion, used to balance the structural response function in the perturbation inversion. The weighting in tension prediction estimation is set according to the type of bolt structure inside the pipe, with a reference range of values. ; It serves as a reference benchmark for the energy corresponding to structural tension, which can be taken as the structural response energy of bolts under standard preload, such as the perturbation inversion structural response function. The maximum value; This is the morphological perturbation term of the preprocessed digital signal sequence, used to describe the morphological response intensity in the digital signal attributable to tension changes under the current bolt structure stability level; the numerator term... This reflects the combined amplitude of the modulation trend and local oscillation intensity of the preprocessed digital signal sequence, i.e., the intensity of the digital signal disturbance response caused by changes in the bolt structure state; denominator term It is a structural disturbance suppression term, used to characterize whether the current digital signal change is due to a real change in the bolt structure or due to environmental disturbances or digital signal noise. It is a combination of signal oscillation acceleration and time window weighting function, used to reflect the second-order change of digital signal, that is, signal oscillation acceleration, within the time window. Dynamic contribution; It is the energy normalization term of the topological perturbation inversion response, used to reflect the total response of the digital signal under structural perturbation.

[0063] To convert the actual preload prediction value into the bolt preload force in the context of actual bolt structure engineering, an existing multi-physical parameter collaborative calibration model is introduced to correct the actual preload prediction value, resulting in the corrected bolt preload force. The multi-physical parameter collaborative calibration model compares real-time operating parameters such as temperature, frequency drift, and reflection delay collected by existing sensors with information from a preset calibration database. It dynamically adjusts the predicted value of the actual preload to correct deviations caused by factors such as thermal expansion and contraction of materials and disturbances in the propagation path of digital signals, thereby improving the accuracy of tension prediction and estimation and its engineering applicability.

[0064] The corrected bolt preload Compared with the safety threshold preset based on expert experience method , If a comparison is made, This indicates that the current bolt preload is within the design safety range, and the recorded status is normal; if This indicates that there is some tension decay in the bolts, but it does not yet pose a structural risk. The system issues a warning signal, and maintenance personnel are advised to pay close attention. This indicates that the bolt preload has fallen below the minimum safety threshold, indicating a potential risk of failure. An alarm command is immediately issued and pushed to the maintenance terminal through the monitoring platform, suggesting immediate repair or replacement.

[0065] In summary, a method for online monitoring of bolt preload in steel pipes has been developed.

[0066] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0067] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0068] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method of monitoring the preload of a steel pipe bolt on line, characterized by, The method comprises the following steps: S1. Based on the collected electromagnetic resonance signal and acoustic reflection signal, an original digital signal sequence is obtained; the original digital signal sequence is preprocessed to obtain a preprocessed digital signal sequence; based on the preprocessed digital signal sequence, a time sequence segment set is obtained; based on the time sequence segments in the time sequence segment set, a structural kernel mapping coding mechanism is introduced, the instantaneous curvature, local oscillation rate and modulation speed factor of the time sequence segments are calculated as the structural kernel features of the time sequence segments, and a structural feature set of the time sequence segments is constructed; S2. Based on the structural feature set, an asymmetric topological graph is constructed, and the structural correlation degree between nodes in the asymmetric topological graph is calculated; based on the structural correlation degree, the local topological disturbance intensity of the nodes is calculated, and the local topological disturbance intensity is converted into a topological tension field in the time domain; S3. Based on the preprocessed digital signal sequence, structural kernel features of the preprocessed digital signal sequence are obtained, and a structural disturbance quantization amount is calculated; the product of the structural disturbance quantization amount and the topological tension field in the time domain is taken as a disturbance-propagation coupling coefficient, and a signal acceleration term is introduced to obtain a disturbance inversion structural response function; Based on the disturbance inversion structural response function, the structural kernel features of the preprocessed digital signal sequence and the structural disturbance quantization amount are combined to construct a tension prediction estimation formula, and an actual pretightening force prediction value is obtained; The actual pretightening force prediction value is corrected to obtain a corrected bolt pretightening force; Based on the corrected bolt pretightening force, a monitoring result is obtained.

2. The method of claim 1, wherein the method comprises: The S2 specifically comprises: In the asymmetric topological graph, based on the structural feature set of the time sequence segments, a structural difference degree is introduced, and the structural correlation degree between nodes is calculated.

3. The method of claim 2, wherein the method is characterized by: The S2 specifically comprises: In the asymmetric topological graph, based on the structural correlation degree between the nodes and the neighborhood nodes, in combination with the local oscillation rate in the structural feature set of the time sequence segments, the local topological disturbance intensity of the nodes is calculated.

4. The method of claim 1, wherein the method is characterized by: The S3 specifically comprises: Based on the preprocessed digital signal sequence, a structural kernel mapping coding mechanism is introduced, the instantaneous curvature, local oscillation rate and modulation speed factor of the preprocessed digital signal sequence are calculated, and the structural kernel features of the preprocessed digital signal sequence are obtained.

5. The method of claim 4, wherein the method is characterized by: The S3 specifically comprises: Based on the structural kernel features of the preprocessed digital signal sequence, an energy distribution amplification factor is introduced, and the structural disturbance quantization amount is calculated.

Images