A method for detecting stress of weld seams of upper and lower sections of a contact fluidized bed

Abstract

The invention provides a method for detecting weld stress of an upper and lower section of a contact fluidized bed, which belongs to the technical field of fluidized bed installation and use, and comprises: respectively applying vibration excitations of different frequencies and amplitudes to an upper section and a lower section of the contact fluidized bed; using an optical fiber interferometer to monitor the weld stress changes of the upper and lower section of the fluidized bed; drawing a plurality of weld stress distribution diagrams corresponding to vibration excitations of different frequencies according to measured weld stress change data; performing Bayesian inference on the plurality of weld stress distribution diagrams, predicting the coordinates of the highest point of weld stress by establishing a probability model, and recording the predicted highest point of weld stress as a flying point; performing self-supervised clustering on a plurality of flying points to obtain a plurality of clustered flying points as the highest points of weld stress; determining a weak area of ​​the weld according to the position of the highest point of weld stress; and using a finite element analysis method to perform stress analysis on the weak area of ​​the weld to obtain a weld stress distribution diagram for representing a weld stress detection result.

Description

Technical Field

[0001] The invention belongs to the technical field of fluidized bed installation and use, and in particular relates to a method for detecting stress of welds of upper and lower sections of a contact fluidized bed. Background Art

[0002] As a highly efficient fluidized reactor, the contact fluidized bed is widely used in the fields of chemical industry, energy and environmental protection. Its typical structure consists of two upper and lower segment components, which are connected together by welding to form a whole. During operation, the fluidized bed will be subjected to complex vibrations and external forces. These mechanical loads will cause stress concentration in the weld area, which will seriously affect the service life and safety of the equipment. Therefore, accurate evaluation of the stress state of the contact fluidized bed weld has become one of the key technologies to ensure the reliability of equipment operation.

[0003] At present, the commonly used weld stress detection methods mainly include strain gauge measurement, X-ray diffraction and finite element analysis. Although the strain gauge measurement method is simple and intuitive, it can only obtain local stress information and cannot reflect the overall stress state; the X-ray diffraction method can accurately measure the surface residual stress, but cannot evaluate the deep stress distribution; the finite element analysis method can simulate the stress field of the weld, but it needs to determine the appropriate boundary conditions and material parameters in advance, and the results have great uncertainty. These methods have their own advantages and disadvantages, and it is difficult to achieve a comprehensive assessment of the stress of complex structural welds. Summary of the invention

[0004] In view of this, the present invention provides a method for detecting stress of welds of upper and lower sections of a contact fluidized bed, which can solve the technical problem that existing methods for detecting stress of welds of upper and lower sections of a contact fluidized bed are difficult to achieve comprehensive evaluation of stress of welds of complex structures.

[0005] The present invention is achieved in that:

[0006] A first aspect of the present invention provides a method for detecting stress of a weld of an upper and lower section of a contact fluidized bed, comprising the following steps:

[0007] S10, applying vibration excitations of different frequencies and amplitudes to the upper and lower groups of the contact fluidized bed respectively;

[0008] S20, using a fiber optic interferometer to monitor the stress changes of the welds of the upper and lower sections of the fluidized bed;

[0009] S30, drawing a plurality of weld stress distribution diagrams corresponding to vibration excitations of different frequencies according to the measured weld stress change data;

[0010] S40, performing Bayesian inference on the plurality of weld stress distribution diagrams, predicting the coordinates of the highest point of weld stress by establishing a probability model, and recording the predicted highest point of weld stress as a flying point;

[0011] S50, using the DBSCAN algorithm to perform self-supervisory clustering on the multiple flying points to obtain multiple clustered flying points as the highest stress points of the weld;

[0012] S60, determining a weak area of ​​the weld according to the position of the highest stress point of the weld;

[0013] S70. Use a finite element analysis method to perform stress analysis on the weak area of ​​the weld to obtain a weld stress distribution diagram for representing the weld stress detection result.

[0014] On the basis of the above technical solution, the weld stress detection method of the upper and lower sections of a contact fluidized bed of the present invention can also be improved as follows:

[0015] Wherein, step S30 specifically includes:

[0016] Step 1, discretizing the time-stress variation curve obtained in step S20 to obtain stress values ​​at different positions on the weld at each moment;

[0017] Step 2: Convert these discrete stress data into continuous weld stress distribution diagrams by interpolation or fitting, and draw corresponding weld stress distribution diagrams for vibration excitation conditions of different frequencies.

[0018] Wherein, step S40 specifically includes:

[0019] Step 1: Establish a probabilistic model to describe the relationship between the stress distribution of the weld and the location and amplitude parameters of the stress concentration point;

[0020] Step 2: Using the stress distribution data obtained in step S30, the model parameters are inferred and updated through the Bayesian formula, and finally the probability distribution of the highest stress point of the weld is obtained;

[0021] Step 3: The peak point coordinates of the probability distribution are used as the predicted highest stress point of the weld, which is recorded as a flying point.

[0022] Wherein, step S50 specifically includes:

[0023] Step 1: Perform cluster analysis on the multiple flying point coordinates predicted in step S40 using the DBSCAN algorithm, wherein the key parameters of the DBSCAN algorithm include the radius threshold Eps and the minimum sample number MinPts;

[0024] Step 2: By reasonably setting Eps and MinPts, the DBSCAN algorithm automatically identifies the clustering structure in the data and divides the flying points into different cluster centers, which represent the key high stress areas of the weld stress.

[0025] Wherein, step S70 specifically includes:

[0026] Step 1: Establish a finite element model and take the weak area of ​​the weld determined in step S60 as the analysis object;

[0027] Step 2: Through finite element analysis, the detailed stress distribution of the weak area can be obtained, including the location and amplitude of the stress peak;

[0028] Step 3: Draw the finite element analysis results into a stress distribution cloud diagram to intuitively display the weld stress detection results.

[0029] In step S10, the vibration excitation F of the upper section assembly u (t) and the vibration excitation F of the lower section component l The frequency of (t) satisfies |f u -f l |≥5Hz to ensure that the two vibration excitations are independent of each other and produce obvious differences in stress response.

[0030] In step S10, the vibration excitation amplitude A of the upper section component u and the vibration excitation amplitude A of the lower section component l Satisfaction|A u -A l |≥20N to ensure that there is enough difference in the intensity of the two vibration excitations to induce obvious stress concentration effect.

[0031] The probability model p(σ(x, y, t)|θ) assumes a normal distribution where μ=μ(x, y, t, θ), σ 2 =σ 2 (x, y, t, θ).

[0032] Furthermore, in step S70, the finite element analysis uses 8-node hexahedral units for meshing.

[0033] Furthermore, in step S70, the finite element analysis uses large deformation theory to more accurately describe the large deformation and plastic behavior that may occur in the weld area.

[0034] Compared with the prior art, the method for detecting stress of welds of upper and lower sections of a contact fluidized bed provided by the present invention has the following beneficial effects:

[0035] 1. Integrate multiple advanced testing and analysis methods. The present invention combines vibration excitation, optical interferometry, Bayesian inference, density clustering and finite element analysis to make up for the limitations of existing single methods and can more comprehensively and accurately evaluate the stress state of the weld.

[0036] 2. Realize overall performance monitoring of welds in complex structures. Compared with traditional local strain measurement or simulation analysis, the method of the present invention can obtain overall performance information through vibration induction, and accurately predict stress concentration areas in combination with probability analysis, providing a reliable basis for integrity assessment of complex structures.

[0037] 3. Improve the reliability and repeatability of the test results. The use of light interferometry and probability model analysis can significantly improve the accuracy and stability of weld stress detection and reduce the influence of human factors and instrument measurement errors. At the same time, the clustering algorithm can automatically identify high stress areas and improve the objectivity and repeatability of the test results.

[0038] In summary, the solution of the present invention solves the technical problem that the existing weld stress detection methods for the upper and lower sections of the contact fluidized bed are difficult to achieve a comprehensive assessment of the weld stress of a complex structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0040] Figure 1 A flow chart of the method provided by the present invention; DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0042] like Figure 1 As shown, it is a flow chart of a weld stress detection method for an upper and lower section pair of a contact fluidized bed provided by the present invention. The method comprises the following steps: wherein, the following steps are included:

[0043] S10, applying vibration excitations of different frequencies and amplitudes to the upper and lower groups of the contact fluidized bed respectively;

[0044] S20, using a fiber optic interferometer to monitor the stress changes of the welds of the upper and lower sections of the fluidized bed;

[0045] S30, drawing a plurality of weld stress distribution diagrams corresponding to vibration excitations of different frequencies according to the measured weld stress change data;

[0046] S40, performing Bayesian inference on a plurality of weld stress distribution diagrams, predicting the coordinates of the highest point of weld stress by establishing a probability model, and recording the predicted highest point of weld stress as a flying point;

[0047] S50, using the DBSCAN algorithm to perform self-supervised clustering on multiple flying points to obtain multiple clustered flying points as the highest stress points of the weld;

[0048] S60, determining the weak area of ​​the weld according to the position of the highest stress point of the weld;

[0049] S70. Use a finite element analysis method to perform stress analysis on the weak area of ​​the weld to obtain a weld stress distribution diagram for representing the weld stress detection result.

[0050] The specific implementation methods of the above steps are described in detail below:

[0051] The specific implementation method of step S10 is: first, it is necessary to prepare a contact fluidized bed device, which consists of two components, an upper section and a lower section. For the upper section component, it is necessary to install a vibration excitation device that can generate a specific frequency and amplitude, such as an electromagnetic vibrator or a piezoelectric vibration device. Similarly, it is also necessary to install a vibration excitation device that can generate different frequencies and amplitudes for the lower section component. In this way, different vibration excitations can be applied to the upper and lower sections respectively. The frequency and amplitude of the vibration excitation can be adjusted by controlling the parameters of the voltage, current or drive signal.

[0052] By applying different vibration excitations to the upper and lower sections, different stress distributions can be generated on the fluidized bed. This is because the vibration of the upper and lower sections will be transmitted through the stress of the welded part, resulting in changes in the stress state at the weld. By applying different vibration excitations, different weld stress responses can be obtained, providing basic data for subsequent stress detection and analysis.

[0053] The specific implementation method of step S20 is: using a fiber optic interferometer to monitor the stress changes of the upper and lower sections of the fluidized bed on the weld. The fiber optic interferometer is a high-precision optical measuring instrument that can detect small displacements and strains through changes in interference fringes. Specifically, the measuring beam of the fiber optic interferometer can be focused on the key position of the fluidized bed weld, and the optical path changes caused by the small deformation of the weld surface can be used to detect the changes in weld stress.

[0054] Through continuous monitoring by the fiber interferometer, the curve of weld stress change over time under different vibration excitation conditions can be obtained. These data provide basic data support for subsequent stress analysis and prediction. It is necessary to select appropriate measurement points to ensure that the stress change information at the weld can be accurately captured.

[0055] The specific implementation of step S30 is: based on the weld stress change data obtained in step S20, a corresponding weld stress distribution diagram can be drawn for vibration excitations of different frequencies.

[0056] Specifically, the time-stress variation curve can be discretized to obtain the stress values ​​at different positions on the weld at each moment. Then, these discrete stress data can be converted into a continuous stress distribution diagram by interpolation or fitting. In this way, the stress distribution state on the weld under vibration excitation of different frequencies can be clearly observed.

[0057] These stress distribution diagrams provide an intuitive data basis for subsequent stress analysis and prediction. Based on these stress distribution diagrams, the weak areas of the weld and the locations where stress concentration may occur can be preliminarily determined.

[0058] The specific implementation method of step S40 is: perform Bayesian inference on the multiple weld stress distribution maps obtained in step S30, and establish a probability model to predict the coordinates of the highest point of the weld stress.

[0059] Bayesian inference is an analysis method based on probability statistics theory, which can use known data to infer the probability distribution of unknown parameters. In this case, the stress distribution diagram of the weld can be regarded as a probability distribution determined by certain unknown parameters (such as the location and amplitude of the stress concentration point). Through Bayesian inference, the probability distribution of these unknown parameters can be estimated based on the existing stress distribution data.

[0060] Specifically, a probability model needs to be established first to describe the relationship between the weld stress distribution and the unknown parameters. Then, the stress distribution data obtained in step S30 is used to infer and update the model parameters through the Bayesian formula, and finally the probability distribution of the highest point of weld stress is obtained. The peak point coordinates of this probability distribution are used as the predicted highest point of weld stress, which is recorded as a "flying point".

[0061] The purpose of this step is to use statistical methods to predict the possible highest point of weld stress based on limited stress distribution data, which provides an important reference for subsequent stress analysis.

[0062] The specific implementation of step S50 is: using the DBSCAN algorithm to perform self-supervised clustering on the multiple "flying points" predicted in step S40, and obtaining multiple cluster centers as the highest points of weld stress.

[0063] DBSCAN is a density-based clustering algorithm that can automatically discover clusters of any shape in data without specifying the number of clusters in advance. In this case, the coordinates of the multiple "flying points" obtained in step S40 are used as input data to perform cluster analysis using the DBSCAN algorithm.

[0064] The key parameters of the DBSCAN algorithm include: radius threshold Eps, which indicates the maximum distance between two data points; minimum sample number MinPts, which indicates the minimum number of samples required to form a cluster. By setting these two parameters reasonably, the DBSCAN algorithm can automatically identify the clustering structure in the data and divide the "flying points" into different cluster centers.

[0065] The purpose of this step is to further analyze and process the "flying points" obtained in step S40, and identify the true highest points of weld stress by clustering. The multiple cluster centers in the clustering results represent the key high stress areas of the weld stress.

[0066] The specific implementation of step S60 is: determining the weak area of ​​the weld according to the position of the highest stress point of the weld identified in step S50.

[0067] Specifically, the coordinates of the multiple cluster centers obtained in step S50 can be plotted on the weld stress distribution diagram obtained in step S30. In this way, the specific locations of these high stress areas on the weld can be intuitively seen. These high stress areas are defined as weak areas of the weld, which need to be focused on and analyzed.

[0068] By identifying the weak areas of the weld, a clear research object and direction are provided for subsequent stress analysis. More in-depth finite element analysis can be carried out on these weak areas to further evaluate the stress state and damage risk of the weld.

[0069] The specific implementation method of step S70 is: using the finite element analysis method to perform stress analysis on the weak area of ​​the weld determined in step S60, and obtaining a weld stress distribution diagram for representing the detection results.

[0070] Finite element analysis is a numerical simulation method widely used in engineering analysis, which can accurately simulate the physical behaviors of structures or materials such as stress and deformation under external loads. In this case, a corresponding finite element model can be established, and the weak area of ​​the weld determined in step S60 is used as the analysis object.

[0071] Through finite element analysis, the detailed stress distribution of these weak areas can be obtained, including the location and amplitude of the stress peak. These results can be directly drawn into stress distribution cloud maps to intuitively display the detection results of weld stress.

[0072] The purpose of this step is to use advanced numerical simulation methods to conduct in-depth analysis of key weld weak areas and obtain more accurate and detailed stress state information, which provides a reliable basis for the final weld stress detection results.

[0073] In order to better understand the method of the present invention, the specific implementation of the method of the present invention is described below in combination with a specific formula:

[0074] Step S10: Apply vibration excitation of different frequencies and amplitudes to the upper and lower sections of the contact fluidized bed respectively.

[0075] In this step, a contact fluidized bed device is first prepared, which consists of an upper section component and a lower section component. For the upper section component, a device that can generate vibration excitation with a specific frequency and amplitude, such as an electromagnetic vibrator or a piezoelectric vibration device, needs to be installed. Similarly, the lower section component also needs to be installed with a device that can generate vibration excitation with different frequencies and amplitudes.

[0076] Assume that the vibration excitation of the upper component is F u (t) = A u sin(2πf u t), where A u represents the amplitude, f u Indicates frequency. The vibration excitation of the lower component is F l (t) = A l sin(2πf l t), where A l represents the amplitude, f l Indicates frequency. By adjusting the voltage, current or drive signal parameters, different vibration excitations can be applied to the upper and lower segments.

[0077] The purpose of this is that the vibration of the upper and lower components will transfer stress through the welded part, causing the stress state at the weld to change. Different vibration excitations will produce different weld stress responses, providing basic data for subsequent stress detection and analysis.

[0078] Step S20: Using a fiber optic interferometer to monitor the stress changes of the welds of the upper and lower sections of the fluidized bed

[0079] In this step, a fiber optic interferometer is used to monitor the stress changes of the fluidized bed weld. The fiber optic interferometer is a high-precision optical measuring instrument that can detect tiny displacements and strains through changes in interference fringes.

[0080] Specifically, the measurement beam of the fiber interferometer is focused on the key position of the fluidized bed weld, and the optical path change caused by the slight deformation of the weld surface is used to monitor the change of weld stress. Assuming that the weld surface displacement is u(x, y, t), the optical path difference Δx can be expressed as:

[0081] Δx=2u(x,y,t)

[0082] According to the working principle of fiber interferometer, the displacement u(x, y, t) and the change ΔI of interference fringes satisfy the following relationship:

[0083]

[0084] Among them, I0 is the light intensity and λ is the light wavelength.

[0085] By continuously monitoring the changes in these interference fringes, the stress change curve of the weld under different vibration excitation conditions can be obtained. These stress change data provide basic support for subsequent stress analysis and prediction. It is necessary to select appropriate measurement points to ensure that the stress change information at the weld can be accurately captured.

[0086] Step S30: Based on the measured weld stress change data, draw multiple weld stress distribution diagrams corresponding to vibration excitations of different frequencies

[0087] In this step, the time-stress variation curve obtained in step S20 is first discretized to obtain the stress values ​​at different positions on the weld at each moment. Assume that the discrete stress data is σ(x i ,y j , t k ), where (x i ,y j ) represents the coordinate position on the weld, t k Indicates a point in time.

[0088] Then, these discrete stress data are converted into continuous weld stress distribution diagrams by interpolation or fitting. Corresponding weld stress distribution diagrams are drawn for vibration excitation conditions of different frequencies.

[0089] Assume that the weld stress distribution function obtained by interpolation or fitting is σ(x, y, t) = f(x, y, t, A u , f u , A l , f l ), where (x, y) represents any coordinate position on the weld, t represents time, and A u , f u , A l , f l Represent the amplitude and frequency of vibration excitation in the upper and lower segments respectively.

[0090] These stress distribution diagrams provide an intuitive data basis for subsequent stress analysis and prediction, and can preliminarily determine the weak areas of the weld and the locations where stress concentration may occur.

[0091] Step S40: Perform Bayesian inference on the multiple weld stress distribution diagrams to predict the coordinates of the highest point of weld stress by establishing a probability model

[0092] In this step, a probability model is established to predict the coordinates of the highest point of weld stress using the Bayesian inference method.

[0093] Assume that the weld stress distribution σ(x, y, t) can be expressed by some unknown parameters θ = [θ1, θ2, …, θ n ], then the following probability model can be established:

[0094]

[0095] Among them, μ(x, y, t, θ) and σ 2 (x, y, t, θ) represent the expectation and variance of the stress distribution, respectively, which are related to the unknown parameter θ.

[0096] Then, the stress distribution data σ(x i ,y j , t k ), the model parameters θ are inferred and updated through the Bayesian formula:

[0097]

[0098] Among them, p(θ) is the prior probability distribution of the parameter, p(σ(x i ,y j , t k )|θ) is the likelihood function, p(σ(x i ,y j , t k )) is the marginal probability.

[0099] Through Bayesian inference, we can finally get the probability distribution of the highest stress point of the weld. The peak point coordinates of this probability distribution are used as the predicted highest stress point of the weld, recorded as the "flying point". This provides an important reference for subsequent stress analysis.

[0100] Step S50: Use the DBSCAN algorithm to perform self-supervisory clustering on the multiple flying points to obtain multiple clustered flying points as the highest stress points of the weld.

[0101] In this step, the DBSCAN algorithm is used to perform cluster analysis on the coordinates of the multiple "flying points" predicted in step S40 to identify the key high stress areas of the weld stress.

[0102] DBSCAN is a density-based clustering algorithm that can automatically discover clusters of any shape in the data without specifying the number of clusters in advance. The key parameters of the algorithm include:

[0103] 1. Radius threshold E, which represents the maximum distance between two data points;

[0104] 2.Minimum number of samples MinPts, which indicates the minimum number of samples required to form a cluster.

[0105] Given a set of "flying point" coordinates {(x i ,y i )}, the specific steps of the DBSCAN algorithm are as follows:

[0106] 1. For each "flying point" (x i ,y i ), calculate the number of samples N in its E neighborhood i ;

[0107] 2. If N i ≥MinPts, then (x i ,y i ) are marked as "core points";

[0108] 3. For each "core point", mark all "unvisited" points in its ∈ neighborhood as "core points";

[0109] 4. Repeat step 3 until no new "core point" can be found;

[0110] 5. All points that are directly or indirectly density-reachable from the "core point" are divided into the same cluster;

[0111] 6. Mark the points that are not classified into any cluster as "noise points".

[0112] By properly setting the ∈ and MinPts parameters, the DBSCAN algorithm can automatically identify the clustering structure in the data and divide the "flying points" into different cluster centers. These cluster centers represent the key high stress areas of the weld stress.

[0113] Step S60: Determine the weak area of ​​the weld according to the position of the highest stress point of the weld

[0114] In this step, the coordinates of the multiple cluster centers obtained in step S50 are plotted on the weld stress distribution diagram obtained in step S30. In this way, the specific locations of these high stress areas on the weld can be intuitively determined, and they can be defined as weak areas of the weld that require special attention and analysis.

[0115] Assume that the cluster center coordinates obtained in step S50 are These coordinate points can be marked on the stress distribution diagram σ(x, y, t) in step S30 to determine the weak area of ​​the weld, which provides a clear research object and direction for subsequent finite element analysis.

[0116] Step S70: Use the finite element analysis method to perform stress analysis on the weak area of ​​the weld to obtain a weld stress distribution diagram for representing the weld stress detection result.

[0117] In this step, a finite element analysis method is used to perform stress analysis on the weak area of ​​the weld determined in step S60 to obtain a weld stress distribution diagram.

[0118] Specifically, we first need to establish a finite element model and take the weak area of ​​the weld as the analysis object. Assume that the geometric model of this area is The boundary is On this basis, the following finite element equation can be established:

[0119]

[0120] Among them, σ ij is the stress tensor, ε ij is the strain tensor, b i For physical strength, t i is the surface force, u i is the displacement vector. By solving the equation with finite element method, the detailed stress distribution in the weak area of ​​the weld can be obtained, including the position and amplitude of the stress peak.

[0121] Finally, the finite element analysis results are plotted into a stress distribution cloud diagram to intuitively display the weld stress detection results, which provides a reliable basis for the final weld stress detection.

[0122] It should be noted that the variables involved are explained as follows:

[0123]

[0124] Specifically, the principle of the present invention is:

[0125] 1. Vibration excitation induces stress concentration effect. By applying different vibration excitations to the upper and lower components, a local stress concentration area can be generated at the weld. This is because vibration will transfer stress through the weld, causing the stress state at the weld to change. Different vibration characteristics will cause different stress responses, providing basic data for subsequent stress detection and analysis.

[0126] 2. Measure weld stress changes by optical interference. Fiber optic interferometers can accurately monitor changes in the optical path caused by tiny displacements on the weld surface, thereby obtaining a curve of stress changes over time at the weld. This provides detailed data support for evaluating the overall stress state of the weld, making up for the limitation that traditional strain measurement can only obtain local information.

[0127] 3. Bayesian inference predicts the highest stress point. By establishing a probability model to describe the relationship between weld stress distribution and unknown parameters (such as the location and amplitude of stress concentration points), Bayesian inference technology can be used to estimate the probability distribution of these unknown parameters based on measured data. This can predict the possible highest point of weld stress, providing an important reference for determining weak areas.

[0128] 4. DBSCAN clustering identifies high stress areas. The DBSCAN algorithm can automatically discover clusters of any shape in the data without pre-specifying the number of clusters. Using the predicted "flying point" coordinates as input data, DBSCAN clustering can more accurately identify the key high stress areas on the weld, providing a clear research object for subsequent finite element analysis.

[0129] 5. Finite element analysis to evaluate weak areas. Finally, the finite element method is used to perform stress analysis on the identified weak areas of the weld to obtain more detailed and reliable stress distribution information. This provides a reliable basis for the final test results and provides strong support for the safety assessment and life prediction of the equipment.

[0130] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for detecting stress of welds of upper and lower sections of a contact fluidized bed, characterized in that: The following steps are involved: S10, applying vibration excitations of different frequencies and amplitudes to the upper and lower groups of the contact fluidized bed respectively; S20, using a fiber optic interferometer to monitor the stress change of the welds of the upper and lower sections of the fluidized bed, and obtaining a curve of the weld stress change over time under different vibration excitation conditions, recorded as a time-stress change curve; S30, drawing a plurality of weld stress distribution diagrams corresponding to vibration excitations of different frequencies according to the measured weld stress change data; S40, performing Bayesian inference on the plurality of weld stress distribution diagrams, predicting the coordinates of the highest point of weld stress by establishing a probability model, and recording the predicted highest point of weld stress as a flying point; S50, using the DBSCAN algorithm to perform self-supervisory clustering on the multiple flying points to obtain multiple clustered flying points as the highest stress points of the weld; S60, determining a weak area of ​​the weld according to the position of the highest stress point of the weld; S70. Use a finite element analysis method to perform stress analysis on the weak area of ​​the weld to obtain a weld stress distribution diagram for representing the weld stress detection result.

2. A method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: Step S30 specifically includes: Step 1, discretizing the time-stress variation curve obtained in step S20 to obtain stress values ​​at different positions on the weld at each moment; Step 2: Convert these discrete stress data into continuous weld stress distribution diagrams by interpolation or fitting, and draw corresponding weld stress distribution diagrams for vibration excitation conditions of different frequencies.

3. The method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: Step S40 specifically includes: Step 1: Establish a probabilistic model to describe the relationship between the stress distribution of the weld and the location and amplitude parameters of the stress concentration point; Step 2: Using the stress distribution data obtained in step S30, the model parameters are inferred and updated through the Bayesian formula, and finally the probability distribution of the highest stress point of the weld is obtained; Step 3: The peak point coordinates of the probability distribution are used as the predicted highest stress point of the weld, which is recorded as a flying point.

4. The method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: Step S50 specifically includes: Step 1: Perform cluster analysis on the multiple flying point coordinates predicted in step S40 using the DBSCAN algorithm, wherein the key parameters of the DBSCAN algorithm include the radius threshold Eps and the minimum sample number MinPts; Step 2: By reasonably setting Eps and MinPts, the DBSCAN algorithm automatically identifies the clustering structure in the data and divides the flying points into different cluster centers, which represent the key high stress areas of the weld stress.

5. The method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: Step S70 specifically includes: Step 1: Establish a finite element model and take the weak area of ​​the weld determined in step S60 as the analysis object; Step 2: Through finite element analysis, the detailed stress distribution of the weak area can be obtained, including the location and amplitude of the stress peak; Step 3: Draw the finite element analysis results into a stress distribution cloud diagram to intuitively display the weld stress detection results.

6. The method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: In step S10, the vibration excitation F of the upper section assembly u (t) and the vibration excitation F of the lower section component l The frequency of (t) satisfies |f u -f l |≥5Hz, to ensure that the two vibration excitations are independent of each other and produce obvious stress response differences; where F u (t) = A u sin(2πf u t), A u represents the amplitude of the vibration excitation of the upper component, f u Indicates the frequency of vibration excitation of the upper component; F l (t) = A l sin(2πf l t), A l represents the amplitude of the vibration excitation of the lower component, f l Indicates the frequency of vibration excitation of the lower section component.

7. A method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: In step S10, the vibration excitation amplitude A of the upper section component u and the vibration excitation amplitude A of the lower section component l Satisfaction|A u -A l |≥20N to ensure that there is enough difference in the intensity of the two vibration excitations to induce obvious stress concentration effect.

8. The method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 1, characterized in that: The probability model p(σ(x, y, t)|θ) assumes a normal distribution Where μ=μ(x, y, t, θ), σ 2 =σ 2 (x, y, t, θ).

9. A method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 5, characterized in that: In the step S70, the finite element analysis uses 8-node hexahedral units for meshing.

10. A method for detecting stress of welds of upper and lower sections of a contact fluidized bed according to claim 9, characterized in that: In step S70, the finite element analysis uses large deformation theory to more accurately describe the large deformation and plastic behavior that may occur in the weld area.

Images