A method for on-line identification of multi-axis tilt angles of a power plant support hanger

Abstract

The present application relates to the technical field of power plant support hanger measurement, in particular to a kind of power plant support hanger multi-axis inclination angle online identification method, comprising the following steps: S1, construct connection structure atlas and extract response characteristics, carry out time domain stability analysis to initial inclination angle sequence, identify inertia inclination direction and floating interval, output initial attitude envelope model;S2, combine real-time inclination data with envelope model, separate out flexible disturbance and rigid pose change quantity;S3, project rigid change quantity to main direction, generate inclination angle identification result and add confidence label for state early warning.The present application, by dynamic modeling, rigid-flexible decoupling and state evaluation fusion, realizes the high-precision, anti-interference, early-warning online identification of the multi-axis inclination angle of power plant support hanger, effectively improves the intelligentization and reliability level of structure monitoring.

Description

Technical Field

[0001] This invention relates to the field of power plant support and hanger measurement technology, and in particular to an online identification method for multi-axis tilt angles of power plant supports and hangers. Background Technology

[0002] In thermal power plants and large-scale energy equipment systems, supports and hangers are widely used to support and limit the thermal expansion and contraction and operational vibration of key components such as pipelines and flues. Their structural safety is directly related to the operational stability and lifespan of the entire equipment. As power plants develop towards larger scale, higher temperature and pressure, and more complex operating conditions, supports and hangers may experience varying degrees of tilting deformation, abnormal stress, and thermal vibration coupling response during operation. In particular, attitude changes in multiple axes not only affect the structural body but may also trigger cascading failures such as pipeline displacement and connection instability. Therefore, achieving online high-precision identification and status assessment of the multi-axis tilt angles of supports and hangers has become a key technical requirement for improving the level of intelligent operation monitoring in power plants.

[0003] Existing methods for monitoring the attitude of supports and hangers mostly rely on static sensing, limited point measurement, or manual periodic inspection. These methods are difficult to adapt to the actual scenarios of multi-directional coupling, dynamic stress changes, and non-standard installation of support and hanger structures. On the one hand, existing methods generally rely on preset calibration or fixed reference benchmarks, which cannot solve the problem of unclear initial attitude of irregular structures. On the other hand, tilt angle data often contains high-frequency disturbances and sensing errors. Existing technologies lack effective differentiation between rigid deformation and flexible interference, which can easily lead to false alarms or missed alarms and cannot support a reliable online early warning mechanism. In addition, existing methods usually only output numerical results and lack judgment and classification strategies for the credibility of identification, resulting in delayed back-end operation and maintenance response and low level of intelligence. Summary of the Invention

[0004] This invention provides an online identification method for the multi-axis tilt angle of power plant supports and hangers. It constructs a dynamic initial attitude envelope model based on the characteristics of the connection structure and the operation response law, integrates inertial measurement data and rigid-flexible decoupling algorithm, accurately extracts the real tilt angle change of the structure in the main response direction, and introduces a disturbance confidence label mechanism to realize the reliable quantification and hierarchical early warning of attitude identification results.

[0005] A method for online identification of multi-axis tilt angles of power plant supports and hangers includes the following steps:

[0006] S1, based on the connection structure diagram of the support and the installation components, combined with the thermal or vibration response characteristics of the structural components, performs time-domain stability analysis on the collected initial tilt angle sequence, identifies the inertial tilt direction of the structure and its corresponding stable floating range, and outputs a dynamic initial attitude envelope model.

[0007] S2, based on the real-time tilt data collected by the inertial measurement unit, combined with the dynamic initial attitude envelope model, the rigid-flexible component separation algorithm is used to perform differential decomposition on the structural tilt angle change to obtain the flexible disturbance compensation vector and the rigid pose change.

[0008] S3 performs vector projection of the rigid pose change and the structural inertial tilt direction to generate the three-axis principal direction tilt angle identification result under the relative reference, and adds a flexible disturbance confidence label to the identification result to guide the judgment of the support and hanger operation status and the optimization of early warning strategy.

[0009] Optionally, S1 includes:

[0010] S11. Based on the connection type, support method, spatial arrangement and material properties between the support and its installation components, a connection structure diagram of the support system is constructed. The connection structure diagram includes the force transmission path, stiffness coupling relationship and structural degree of freedom characteristics of each support point. At the same time, combined with the response characteristics of structural components under thermal load and mechanical vibration in the actual operating environment, the micro deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation are extracted.

[0011] S12, perform time-domain stability analysis on the initial tilt angle sequence collected by the inertial measurement unit deployed on the support and hanger. Using the sliding window trend judgment, fluctuation extreme value identification and drift rate calculation method, identify the steady-state tilt segment with dynamic offset that does not contain flexible disturbance and has a period less than the set time threshold (5 seconds). Combine the inertial feature direction information extracted from the connection structure map to determine the main direction of structural inertial tilt, and determine its stable floating interval threshold in the main tilt direction. Output a dynamic initial attitude envelope model for support and hanger attitude recognition.

[0012] Optionally, S11 includes:

[0013] S111, based on design drawings, BIM models, or on-site structural identification, extracts the core connection parameters between supports and installation components, including connection type (rigid, semi-rigid, flexible), support method (single-point suspension, guide sliding, fixed support), spatial layout (three-dimensional coordinates and relative distribution of support points), and component material properties (elastic modulus E). c Coefficient of thermal expansion α c Damping coefficient ζ c );

[0014] S112, based on the structural topology, all support and hanger nodes are represented as graph G = (N, L), where N represents the connection node of the support and hanger, and L = {l ij} represents a rigid or flexible connection path between nodes, where each edge l ij Associate a force transfer function;

[0015] S113, Construct the stiffness coupling matrix K on the node connection graph G. global This represents the force-displacement response relationship between nodes along different axes;

[0016] S114, regarding the stiffness coupling matrix K global Eigenvalue decomposition is performed to identify the direction of the principal deformation degree of freedom with the weakest stiffness response, which is used as the principal response axis where the support is most prone to tilting.

[0017] S115, targeting the main response axis, combines material thermodynamics and vibration response models to extract the micro-deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation.

[0018] Optionally, S114 includes:

[0019] S1141, using linear algebra methods, the stiffness coupling matrix K global It can be decomposed into the product of a set of eigenvector matrices Q and eigenvalue diagonal matrices ∑;

[0020] S1142, in the set of stiffness response eigenvalues ​​(σ1,σ2,σ3), find the eigenvector corresponding to the smallest eigenvalue as the direction of the weakest stiffness response, which corresponds to the main response axis of the support and hanger most prone to tilting.

[0021] Optionally, S12 includes:

[0022] S121, the original three-axis tilt angle vector Θ(t) = (θ) acquired by the inertial measurement unit (IMU) x (t),θ y (t),θ z (t)} is partitioned in the time domain using a sliding window method, with a window length of Δt and a step size of δt. Each sliding window is defined as W. n =[t n ,t n +Δt], extract the average tilt angle vector within each window;

[0023] S122, in each window, calculate the flexible disturbance intensity A of the local angle change. n With attitude change rate r n As a characteristic quantity for judging stability, if A n th And r n <r th Then it is a stable window, where A th r th The set threshold values ​​for disturbance and change;

[0024] S123, will the window W that meets the conditions n The corresponding tilt angle vector ​Collected as a stable dataset

[0025] S124, perform principal component analysis on the tilt angle vectors in the stable set to extract the principal tilt direction vector d. tilt Calculate the maximum fluctuation range of the projected value sequence along the main direction within the stable window, and extract the stable fluctuation interval threshold Δ. tilt And construct a dynamic initial attitude envelope model.

[0026] Optionally, S2 includes:

[0027] S21. Compare the three-axis tilt angle data collected in real time by the inertial measurement unit with the dynamic initial attitude envelope model. Based on the main tilt direction defined in the dynamic initial attitude envelope model, perform directional projection analysis on the current attitude and extract the tilt angle change in that direction.

[0028] S22, the rigid-flexible component separation algorithm is used to process the obtained tilt angle change.

[0029] Optionally, S21 includes:

[0030] S211, the three-axis tilt angle vector Θ at the current moment is acquired in real time by the inertial measurement unit (IMU). seq Extract the steady-state reference attitude Θ0={θ} from the dynamic initial attitude envelope model. x,0 ,θ y,0 ,θ z,0};

[0031] S212, extract the direction vector d of the principal deformation degree of freedom from the constructed dynamic initial attitude envelope model. soft ;

[0032] S213, by calculating the tilt difference between the current attitude and the initial attitude and projecting it onto the direction of the principal deformation degree of freedom, the real-time attitude offset of the structure in the principal response axis is extracted.

[0033] Optionally, S22 includes:

[0034] S221, extract the real-time attitude offset Δθ along the main response axis. soft (t) Samples are taken over a continuous time period to form a sequence of tilt angle changes in the direction of the principal deformation degrees of freedom, S. θ ;

[0035] S222, for the tilt angle change sequence S θ An adaptive signal decomposition algorithm with multi-scale noise control capability (CEEMDAN) is used to decompose the tilt angle variation sequence S. θ It is decomposed into multiple intrinsic mode components;

[0036] S223, by performing frequency analysis on each intrinsic modal component obtained from the decomposition, the dominant frequency value of each component is calculated, and based on the set frequency threshold, the intrinsic modal components below the frequency threshold are classified as the rigid response part representing the actual attitude change of the structure, and the intrinsic modal components above the frequency threshold are classified as the flexible disturbance part reflecting environmental disturbance and vibration interference.

[0037] Optionally, S3 includes:

[0038] S31, the rigid response part obtained after rigid-flexible separation is projected onto the three principal deformation direction vectors respectively to obtain the rigid tilt angle change value under each principal direction;

[0039] S32, based on the flexible perturbation part, evaluates the perturbation energy and frequency concentration in each time period and gives the confidence label γ(t) of the flexible perturbation;

[0040] S33 combines the change in rigid tilt angle with the corresponding confidence label to form the structural state evaluation vector S. tilt (t), the structural state evaluation vector is used as the input index for the operation status analysis of the support and hanger. Based on the offset threshold and confidence level set by the working condition, the support and hanger structure is classified into multiple levels, including normal, early warning and alarm.

[0041] Optionally, S33 includes:

[0042] S331, Set the three-axis attitude offset threshold T for structural safety. x T y T z The confidence levels for the flexible perturbation were set as follows: when γ(t) ≥ 0.85, the confidence level for the flexible perturbation is high confidence; when 0.65 ≤ γ(t) < 0.85, the confidence level for the flexible perturbation is medium confidence; and when γ(t) < 0.65, the confidence level for the flexible perturbation is low confidence.

[0043] S332 defines a multi-level classification rule for the operating status of supports and hangers, specifically including:

[0044] Normal state: The attitude offset in all main directions is less than the corresponding attitude offset threshold and the confidence level is medium or high confidence.

[0045] Warning status: The attitude deviation in any direction is greater than 80% of its corresponding attitude deviation threshold or the confidence level is medium or low confidence.

[0046] Alarm status: The attitude deviation in any direction exceeds its corresponding attitude deviation threshold and the confidence level is not lower than medium confidence, or three or more warning status trigger events are detected within a 10-minute sliding time window.

[0047] The beneficial effects of this invention are:

[0048] This invention, by constructing a dynamic initial attitude envelope model based on the connection structure spectrum and thermal-vibration response characteristics, can accurately identify the direction of the main deformation degrees of freedom of supports and hangers in complex environments and their steady-state tilt characteristics without the need for manual initial calibration. It effectively solves the problems of unclear initial state and lack of tilt angle reference in the prior art, and significantly improves the structural adaptability and modeling accuracy of tilt identification.

[0049] This invention introduces a rigid-flexible component separation algorithm to decouple the real-time tilt angle change signal into a rigid component representing the actual structural response and a flexible component representing external disturbances. This achieves effective filtering and dynamic compensation for environmental interference, avoids interference from non-structural factors such as vibration and thermal fluctuations on the identification accuracy, and improves the stability and reliability of the tilt monitoring system.

[0050] This invention combines the triaxial rigid tilt angle change value with the flexible disturbance confidence label into a structural state evaluation vector, and introduces a multi-level classification criterion jointly driven by the degree of offset and the recognition confidence. It establishes an intelligent monitoring mechanism that integrates attitude recognition, confidence evaluation and state early warning, which can realize real-time classification judgment and linkage early warning of the operating status of the support and hanger, and has good practicality and engineering promotion value. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a schematic diagram of the identification method flow according to an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of the posture change recognition process according to an embodiment of the present invention. Detailed Implementation

[0054] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art may employ other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0055] like Figures 1-2 As shown, a method for online identification of multi-axis tilt angles of power plant supports includes the following steps:

[0056] S1, based on the connection structure diagram of the support and the installation components, combined with the thermal or vibration response characteristics of the structural components, performs time-domain stability analysis on the collected initial tilt angle sequence, identifies the inertial tilt direction of the structure and its corresponding stable floating range, and outputs a dynamic initial attitude envelope model.

[0057] S2, based on the real-time tilt data collected by the inertial measurement unit, combined with the dynamic initial attitude envelope model, the rigid-flexible component separation algorithm is used to perform differential decomposition on the structural tilt angle change to obtain the flexible disturbance compensation vector and the rigid pose change.

[0058] S3 performs vector projection of the rigid pose change and the structural inertial tilt direction to generate the three-axis principal direction tilt angle identification result under the relative reference, and adds a flexible disturbance confidence label to the identification result to guide the judgment of the support and hanger operation status and the optimization of early warning strategy.

[0059] S1 includes:

[0060] S11. Based on the connection type, support method, spatial arrangement and material properties between the support and the installation components, a connection structure map of the support system is constructed. The connection structure map includes the force transmission path, stiffness coupling relationship and structural degree of freedom characteristics of each support point. At the same time, combined with the response characteristics of structural components under thermal load and mechanical vibration in the actual operating environment, the micro deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation are extracted.

[0061] S12, perform time-domain stability analysis on the initial tilt angle sequence collected by the inertial measurement unit deployed on the support and hanger. Using the sliding window trend judgment, fluctuation extreme value identification and drift rate calculation method, identify the steady-state tilt segment with dynamic offset that does not contain flexible disturbance and has a period less than the set time threshold (5 seconds). Combine the inertial feature direction information extracted from the connection structure map to determine the main direction of structural inertial tilt, and determine its stable floating interval threshold in the main tilt direction. Output a dynamic initial attitude envelope model for support and hanger attitude recognition.

[0062] S11 includes:

[0063] S111, based on design drawings, BIM models, or on-site structural identification, extracts the core connection parameters between supports and installation components, including connection type (rigid, semi-rigid, flexible), support method (single-point suspension, guide sliding, fixed support), spatial layout (three-dimensional coordinates and relative distribution of support points), and component material properties (elastic modulus E). c Coefficient of thermal expansion αc Damping coefficient ζ c );

[0064] S112, Based on the structural topology, all support nodes are represented as graph G = (N, L), where N = {n1, n2, ..., n} m} represents the connection node of the support and hanger, L = {l ij} represents a rigid or flexible connection path between nodes, where each edge l ij Associate a force transfer function;

[0065] The force transmission function is expressed as:

[0066]

[0067] Among them, T ij f is the deformation transfer amount from node i to j. ij K represents the force between nodes. ij The equivalent stiffness of the path depends on the shape and material stiffness of the component;

[0068]

[0069] Among them, E ij Let A be the elastic modulus of the path component. ij Let L be the cross-sectional area. ij This represents the length of the connection path.

[0070] S113, Construct the stiffness coupling matrix K on the node connection graph G. global This represents the force-displacement response relationship between nodes along different axes, expressed as:

[0071]

[0072] Where, d ij Let be the unit direction vector from node i to node j;

[0073] S114, regarding the stiffness coupling matrix K global Eigenvalue decomposition is performed to identify the direction of the principal deformation degree of freedom with the weakest stiffness response, which is used as the principal response axis where the support is most prone to tilting.

[0074] S115, focusing on the main response axis, combines material thermodynamics and vibration response models to extract the micro-deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation, expressed as follows:

[0075]

[0076] Where, θ T The change in tilt angle is caused by thermal expansion, ΔT is the temperature difference, and L is the angle of inclination. cH is the length of the component. c The height from the support point to the center of gravity;

[0077]

[0078] Among them, D vib Let f be the power spectral density of the acceleration signal at frequency f. min ,f max [This refers to the perturbation frequency set in practical applications (0.2Hz-10Hz).]

[0079] S114 includes:

[0080] S1141, using linear algebra methods, the stiffness coupling matrix K global It can be decomposed into the product of a set of eigenvector matrices Q and eigenvalue diagonal matrices ∑, expressed as:

[0081] K g ·Q=Q·∑;

[0082] Where Q = [q1, q2, q3] is the orthogonal eigenvector matrix, representing the three principal directions, and ∑ = diag(σ1, σ2, σ3) is the stiffness response eigenvalue of the corresponding principal direction;

[0083] S1142, in the set of stiffness response eigenvalues ​​(σ1, σ2, σ3), find the eigenvector corresponding to the smallest eigenvalue as the direction of the weakest stiffness response, which corresponds to the principal response axis where the support is most prone to tilting, expressed as:

[0084] σ min =min(σ1,σ2,σ3),d soft =q i whereσ i =σ min ;

[0085] Where, d soft The direction vector of the main deformation degree of freedom, σ min This is the direction with the weakest stiffness response.

[0086] S12 includes:

[0087] S121, the original three-axis tilt angle vector Θ(t) = (θ) acquired by the inertial measurement unit (IMU) x (t),θ y (t),θ z (t)} is partitioned in the time domain using a sliding window method, with a window length of Δt = 5s and a step size of δt. Each sliding window is defined as W. n =[t n ,t n+Δt], extract the average tilt angle vector within each window, represented as:

[0088]

[0089] in, Let be the local mean tilt angle of the nth window, and N be the total number of windows;

[0090] S122, in each window, calculate the flexible disturbance intensity A of the local angle change. n With attitude change rate r n As a characteristic quantity for judging stability, if A n th And r n <r th Then it is a stable window, where A th r th The set threshold values ​​for disturbance and change are expressed as:

[0091] A n =max(Θ(t))-min(Θ(t)),t∈W n ;

[0092]

[0093] in, Let A be the average tilt angle within the (n-1)th sliding window. th =0.2, r th =0.05 is the set threshold for disturbance and change;

[0094] S123, will the window W that meets the conditions n The corresponding tilt angle vector Collected as a stable dataset

[0095] S124, perform principal component analysis on the tilt angle vectors in the stable set to extract the principal tilt direction vector d. tilt Calculate the maximum fluctuation range of the projected value sequence along the main direction within the stable window, and extract the stable fluctuation interval threshold Δ. tilt And construct a dynamic initial attitude envelope model. Represented as:

[0096]

[0097] S2 includes:

[0098] ​S21. Compare the three-axis tilt angle data collected in real time by the inertial measurement unit with the dynamic initial attitude envelope model. Based on the main tilt direction defined in the dynamic initial attitude envelope model, perform directional projection analysis on the current attitude and extract the tilt angle change in that direction.

[0099] S22, the rigid-flexible component separation algorithm is used to process the obtained tilt angle change.

[0100] S21 includes:

[0101] S211, the three-axis tilt angle vector Θ at the current moment is acquired in real time by the inertial measurement unit (IMU). seq Extract the steady-state reference attitude Θ0={θ} from the dynamic initial attitude envelope model. x,0 ,θ y,0 ,θ z,0};

[0102] S212, extract the direction vector d of the principal deformation degree of freedom from the constructed dynamic initial attitude envelope model. soft ;

[0103] S213, by calculating the tilt difference between the current attitude and the initial attitude, and projecting it onto the direction of the principal deformation degree of freedom, the real-time attitude offset of the structure in the principal response axis is extracted and expressed as:

[0104] ΔΘ(t)=Θ seq -Θ0;

[0105] Δθ soft (t)=ΔΘ(t)·d soft ;

[0106] Where, Δθ soft ΔΘ(t) represents the real-time attitude offset of the support in the main response axis, and ΔΘ(t) represents the tilt difference between the current attitude and the initial attitude.

[0107] S22 includes:

[0108] S221, extract the real-time attitude offset Δθ along the main response axis. soft (t) Samples are taken over a continuous time period to form a sequence of tilt angle changes in the direction of the principal deformation degrees of freedom, S. θ , represented as:

[0109] S θ ={Δθ soft (t1),Δθ soft (t2),...,Δθ soft (tN)};

[0110] Where N is the total number of sampling points;

[0111] S222, for the tilt angle change sequence S θ An adaptive signal decomposition algorithm with multi-scale noise control capability (CEEMDAN) is used to decompose the tilt angle variation sequence S. θ Decomposed into multiple intrinsic mode components, represented as:

[0112]

[0113] Among them, c j (t) is the j-th component, reflecting the fluctuation characteristics at different frequencies, r(t) is the residual term, including the trend or slow change component, and M is the total number of modes obtained by decomposition;

[0114] S223, by performing frequency analysis on each intrinsic modal component obtained from the decomposition, the dominant frequency value of each component is calculated, and based on the set frequency threshold, the intrinsic modal components below the frequency threshold are classified as the rigid response component representing the actual attitude change of the structure, and the intrinsic modal components above the frequency threshold are classified as the flexible disturbance component reflecting environmental disturbances and vibration interference, expressed as:

[0115]

[0116] Among them, f j f is the dominant frequency value of the j-th component. th =0.2 is the frequency threshold. This represents the true rigid pose change of the structure. It is a flexible disturbance or environmental interference component;

[0117]

[0118] Among them, P j (f) represents the power spectral density of the j-th modal component at frequency f, where f min f max These represent the minimum and maximum frequency ranges for spectral analysis (0.01Hz–10Hz).

[0119] S3 includes:

[0120] S31, the rigid response portion obtained after rigid-flexible separation is projected onto the three principal deformation direction vectors to obtain the rigid tilt angle change value in each principal direction, expressed as:

[0121]

[0122] in, The recognition results are shown for the x, y, and z principal axes, respectively. Let d be the rigid pose change vector at the current moment.x d y d z It is the unit vector of the principal deformation direction of the structure;

[0123] S32, based on the flexible perturbation component, evaluates the perturbation energy and frequency concentration in each time period, and gives the confidence label γ(t) of the flexible perturbation, expressed as:

[0124]

[0125] Among them, E flex (t) represents the signal energy of the flexible component within the current window, E total (t) represents the total energy at the tilt angle. It is a flexible tilt angle perturbation sequence;

[0126] S33, combine the rigid tilt angle change value with the corresponding confidence label to form a structural state assessment vector. The structural state assessment vector serves as the input indicator for the analysis of the operating status of the support and hanger. Based on the offset threshold and confidence level set for the working conditions, the support and hanger structure is classified into multiple levels, including normal, early warning, and alarm.

[0127] S33 includes:

[0128] S331, Set the three-axis attitude offset threshold T for structural safety. x =2.0, T y =2.5, T z =3.0, and set the confidence level of the flexible disturbance, where when γ(t)≥0.85, the confidence level of the flexible disturbance is high confidence, when 0.65≤γ(t)<0.85, the confidence level of the flexible disturbance is medium confidence, and when γ(t)<0.65, the confidence level of the flexible disturbance is low confidence.

[0129] S332 defines a multi-level classification rule for the operating status of supports and hangers, specifically including:

[0130] Normal state: The attitude offset in all main directions is less than the corresponding attitude offset threshold and the confidence level is medium or high confidence.

[0131] Warning status: The attitude deviation in any direction is greater than 80% of its corresponding attitude deviation threshold or the confidence level is medium or low confidence.

[0132] Alarm status: The attitude deviation in any direction exceeds its corresponding attitude deviation threshold and the confidence level is not lower than medium confidence, or three or more warning status trigger events are detected within a 10-minute sliding time window.

[0133] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0134] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for online identification of multi-axis tilt angles of power plant supports, characterized in that, Includes the following steps: S1, based on the connection structure diagram of the support and the installation components, combined with the thermal or vibration response characteristics of the structural components, performs time-domain stability analysis on the collected initial tilt angle sequence, identifies the inertial tilt direction of the structure and its corresponding stable floating range, and outputs a dynamic initial attitude envelope model. S2, based on the real-time tilt data collected by the inertial measurement unit, combined with the dynamic initial attitude envelope model, the rigid-flexible component separation algorithm is used to perform differential decomposition on the structural tilt angle change to obtain the flexible disturbance compensation vector and the rigid pose change. S3 performs vector projection of the rigid pose change and the structural inertial tilt direction to generate the three-axis principal direction tilt angle identification result under the relative reference, and adds a flexible disturbance confidence label to the identification result to guide the judgment of the support and hanger operation status and the optimization of early warning strategy.

2. The method for online identification of multi-axis tilt angle of power plant supports according to claim 1, characterized in that, S1 includes: S11. Based on the connection type, support method, spatial arrangement and material properties between the support and its installation components, a connection structure diagram of the support system is constructed. The connection structure diagram includes the force transmission path, stiffness coupling relationship and structural degree of freedom characteristics of each support point. At the same time, combined with the response characteristics of structural components under thermal load and mechanical vibration in the actual operating environment, the micro deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation are extracted. S12, perform time-domain stability analysis on the initial tilt angle sequence collected by the inertial measurement unit deployed on the support and hanger. Using the sliding window trend judgment, fluctuation extreme value identification and drift rate calculation method, identify the steady-state tilt segment with no flexible disturbance and dynamic offset with a period less than the set time threshold. Combine the inertial feature direction information extracted from the connection structure map to determine the main direction of structural inertial tilt, and determine its stable floating interval threshold in the main tilt direction. Output a dynamic initial attitude envelope model for support and hanger attitude recognition.

3. The method for online identification of multi-axis tilt angle of power plant supports according to claim 2, characterized in that, S11 includes: S111, based on design drawings, BIM models or on-site structural identification, extracts the core connection parameters between supports and installation components, including connection type, support method, spatial arrangement and component material properties; In S112, all the support and hanger nodes are represented as a graph G=(N, L) based on the structural topological relationship, where N represents the connection nodes of the support and hanger, L={l ij} represents the rigid or flexible connection path between nodes, and each edge l ij is associated with a force transmission function. S113, constructing a stiffness coupling matrix K for the node connection graph G global , representing the force-displacement response relationship between nodes in different axial directions; S114, the stiffness coupling matrix K global Eigen decomposition is performed to identify the direction of the main deformation freedom degree with the weakest stiffness response as the main response axis of the support hanger most prone to tilting. S115, targeting the main response axis, combines material thermodynamics and vibration response models to extract the micro-deformation trend caused by temperature rise and the dynamic disturbance parameters under vibration excitation.

4. The method for online identification of multi-axis tilt angle of power plant supports according to claim 3, characterized in that, S114 includes: S1141, by linear algebra method, the stiffness coupling matrix K global is decomposed into a product of a set of eigenvector matrices Q and a diagonal matrix of eigenvalues∑; S1142, in the set of stiffness response eigenvalues ​​(σ1,σ2,σ3), find the eigenvector corresponding to the smallest eigenvalue as the direction of the weakest stiffness response, which corresponds to the main response axis of the support and hanger most prone to tilting.

5. The method for online identification of multi-axis tilt angle of power plant supports according to claim 4, characterized in that, S12 includes: S121, the original three-axis tilt angle vector Θ(t) = {θ x (t),θ y (t),θ z (t)} collected by the inertial measurement unit is time-domain segmented in a sliding window manner, the window length Δt is set, the step is δt, each sliding window is defined as W n = [t n , t n + Δt], and the average tilt angle vector is extracted in each window; S122, in each window, calculate the local angle change flexibility perturbation intensity A n with the attitude change rate r n As a judgment stability characteristic quantity, if A n <A th And r n <r th , it is a stable window, wherein A th , r th The set perturbation and change threshold; S123, will the window W that meets the conditions n The corresponding tilt angle vector Collected as a stable dataset S124, perform principal component analysis on the tilt angle vectors in the stable set to extract the principal tilt direction vector d. tilt Calculate the maximum fluctuation range of the projected value sequence along the main direction within the stable window, and extract the stable fluctuation interval threshold Δ. tilt And construct a dynamic initial attitude envelope model.

6. The method for online identification of multi-axis tilt angle of power plant supports according to claim 5, characterized in that, S2 includes: S21. Compare the three-axis tilt angle data collected in real time by the inertial measurement unit with the dynamic initial attitude envelope model. Based on the main tilt direction defined in the dynamic initial attitude envelope model, perform directional projection analysis on the current attitude and extract the tilt angle change in that direction. S22, the rigid-flexible component separation algorithm is used to process the obtained tilt angle change.

7. The method for online identification of multi-axis tilt angle of power plant supports according to claim 6, characterized in that, S21 includes: S211, the three-axis tilt angle vector Θ at the current moment is acquired in real time by the inertial measurement unit. seq Extract the steady-state reference attitude Θ0={θ} from the dynamic initial attitude envelope model. x,0 ,θ y,0 ,θ z,0 }; S212, extract the direction vector d of the principal deformation degree of freedom from the constructed dynamic initial attitude envelope model. soft ; S213, by calculating the tilt difference between the current attitude and the initial attitude and projecting it onto the direction of the principal deformation degree of freedom, the real-time attitude offset of the structure in the principal response axis is extracted.

8. The method for online identification of multi-axis tilt angle of power plant supports according to claim 7, characterized in that, S22 includes: S221, extract the real-time attitude offset Δθ along the main response axis. soft (t) Samples are taken over a continuous time period to form a sequence of tilt angle changes in the direction of the principal deformation degrees of freedom, S. θ ; S222, for the tilt angle change sequence S θ An adaptive signal decomposition algorithm with multi-scale noise control capability is used to decompose the tilt angle variation sequence S. θ It is decomposed into multiple intrinsic mode components; S223, by performing frequency analysis on each intrinsic modal component obtained from the decomposition, the dominant frequency value of each component is calculated, and based on the set frequency threshold, the intrinsic modal components below the frequency threshold are classified as the rigid response part representing the actual attitude change of the structure, and the intrinsic modal components above the frequency threshold are classified as the flexible disturbance part reflecting environmental disturbance and vibration interference.

9. The method for online identification of multi-axis tilt angle of power plant supports according to claim 8, characterized in that, S3 includes: S31, the rigid response part obtained after rigid-flexible separation is projected onto the three principal deformation direction vectors respectively to obtain the rigid tilt angle change value under each principal direction; S32, based on the flexible perturbation part, evaluates the perturbation energy and frequency concentration in each time period and gives the confidence label γ(t) of the flexible perturbation; S33 combines the change in rigid tilt angle with the corresponding confidence label to form the structural state evaluation vector S. tilt (t), the structural state evaluation vector is used as the input index for the operation status analysis of the support and hanger. Based on the offset threshold and confidence level set by the working condition, the support and hanger structure is classified into multiple levels, including normal, early warning and alarm.

10. The method for online identification of multi-axis tilt angle of power plant supports according to claim 9, characterized in that, S33 includes: S331, Set the three-axis attitude offset threshold T for structural safety. x T y T z The confidence levels for the flexible perturbation were set as follows: when γ(t) ≥ 0.85, the confidence level for the flexible perturbation is high confidence; when 0.65 ≤ γ(t) < 0.85, the confidence level for the flexible perturbation is medium confidence; and when γ(t) < 0.65, the confidence level for the flexible perturbation is low confidence. S332 defines a multi-level classification rule for the operating status of supports and hangers, specifically including: Normal state: The attitude offset in all main directions is less than the corresponding attitude offset threshold and the confidence level is medium or high confidence. Warning status: The attitude deviation in any direction is greater than 80% of its corresponding attitude deviation threshold or the confidence level is medium or low confidence. Alarm status: The attitude deviation in any direction exceeds its corresponding attitude deviation threshold and the confidence level is not lower than medium confidence, or three or more warning status trigger events are detected within a 10-minute sliding time window.

Images