A method for testing the seismic performance of steel-concrete seismic resistant structures

Abstract

The present invention relates to the technical field of structural component testing, and specifically to a method for testing the seismic performance of a steel-concrete seismic structure, including: obtaining a scaled model of the seismic structure, obtaining an axial stress curve of the model, obtaining an abnormal score of the axial stress value in the axial stress curve; obtaining the possibility of strain localization of a data point in the axial stress curve; obtaining a characteristic value of strain localization of a data point in the axial stress curve; obtaining an influence coefficient of the position of a strain gauge in the model; obtaining abnormal parameters of a data point in the axial stress curve corresponding to the strain gauge in the model according to the abnormal score, the possibility of strain localization of a data point, and the influence coefficient; obtaining a strain localization data point according to the abnormal parameters and then obtaining the performance parameters of the seismic structure. The present invention improves the accuracy of data positioning and identification when strain localization occurs, thereby improving the accuracy and authenticity of evaluating the seismic performance of the seismic structure.

Description

Technical Field

[0001] The invention relates to the technical field of structural component testing, and in particular to a method for testing the seismic performance of a steel-concrete seismic resistant structure. Background Art

[0002] Steel-concrete composite structure is an independent structural type that buries steel in reinforced concrete. Since steel is added to reinforced concrete, the steel has its inherent strength and ductility. The three-in-one work of steel, steel bars and concrete makes the steel-concrete structure have the advantages of greater bearing capacity, greater rigidity and better earthquake resistance than traditional reinforced concrete structures.

[0003] When testing the seismic performance of steel-concrete seismic structures, experimenters usually set up dynamic test equipment to evaluate the performance of the structure under dynamic loads. In the process of obtaining the axial stress of the support of the seismic structure, due to the lateral load of the earthquake, some parts of the structure may be subjected to large lateral forces, thereby causing the formation of strain localization. Strain localization refers to a phenomenon in which the strain (or deformation) of the main body is concentrated in a local narrow area (or called a shear band) when the main body is destroyed. Therefore, the seismic performance of the seismic structure is evaluated by monitoring the corresponding simulated earthquake amplitude and time formed by strain localization.

[0004] The isolation forest algorithm is usually used to obtain data outliers, and the corresponding relevant data of the formation of strain localization is located according to the outliers, so as to judge the overall seismic performance of the structure; however, for seismic resistant structures, the ultimate stresses at different positions are different, so the amplitude and time required for strain localization, as well as the impact on the seismic performance of the entire seismic resistant structure are different, so the axial stresses at different positions are different; at the same time, since the simulated earthquake is a nonlinear change, there is a certain degree of data fluctuation in the monitored axial stress, and the emergence of strain localization will also cause the data to convex. It is not accurate to obtain data outliers using only the isolation forest algorithm, which in turn affects the accuracy of the seismic performance test of the entire seismic resistant structure. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a method for testing the seismic performance of a steel-concrete seismic resistant structure.

[0006] The seismic performance testing method of a steel-concrete seismic resistant structure of the present invention adopts the following technical scheme:

[0007] An embodiment of the present invention provides a method for testing the seismic performance of a steel-concrete seismic resistant structure, the method comprising the following steps:

[0008] Acquire several models of earthquake-resistant structures scaled in equal proportion, install several strain gauges on each model, and acquire several axial stress curves of each model under a preset earthquake load value according to the several strain gauges installed on each model, wherein the axial stress curve comprises several data points, each data point corresponding to a time and an axial stress value; perform anomaly detection on the axial stress curve to obtain an anomaly score of each axial stress value in each axial stress curve;

[0009] Preset the local range of each data point in each axial stress curve; obtain the slope of each data point in each axial stress curve, and obtain the possibility of strain localization for each data point in each axial stress curve according to the local range, the slope of the data point and the axial stress value corresponding to the data point;

[0010] Obtain the vertical distance from each strain gauge in each model to the seismic simulation platform; obtain the angle value between the support where each strain gauge in each model is located and the horizontal plane of the seismic simulation platform; obtain the strain localization performance characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model based on the vertical distance, the angle value of the horizontal plane and the possibility of strain localization of the data point; obtain the characteristic fitting curve corresponding to each strain gauge in each model based on the strain localization performance characteristic value; obtain the slope of each point in the characteristic fitting curve corresponding to each strain gauge in each model, and obtain the influence coefficient of the position of each strain gauge in each model based on the slope of each point in the characteristic fitting curve;

[0011] According to the abnormal score, the possibility of strain localization of the data point and the influence coefficient, the abnormal parameters of each data point in the axial stress curve corresponding to each strain gauge in each model are obtained; according to the size of the abnormal parameters, several strain localization data points of the axial stress curve corresponding to each strain gauge in each model are obtained; according to the time corresponding to the strain localization data point, the performance parameters of the seismic resistant structure are obtained.

[0012] Furthermore, the possibility of strain localization at each data point in each axial stress curve is obtained according to the local range, the slope of the data point and the axial stress value corresponding to the data point, and the specific steps include the following:

[0013] Record any one model as a first model, record several axial stress curves of the first model under corresponding earthquake load values ​​as several axial stress curves of the first model, and record any one axial stress curve of the first model as a target axial stress curve;

[0014] ;

[0015] In the formula, The target axial stress curve The axial stress value corresponding to the data point is The target axial stress curve The maximum value of the axial stress value corresponding to the data point in the local range of the data points; and The specific method of obtaining is as follows: The local range of the data points is recorded as the first local range, and the The variance of the axial stress values ​​corresponding to all the data points on the left side of the data point is recorded as , the first local range The variance of the axial stress values ​​corresponding to all the data points to the right of the data point is recorded as ; The target axial stress curve The number of data points in the local range of data points, The target axial stress curve In the local range of data points The slope of the data point, The target axial stress curve In the local range of data points The slope of the data point, To take the absolute value, is a preset hyperparameter. The target axial stress curve The probability of strain localization for each data point.

[0016] Furthermore, the method of obtaining the strain localization characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model according to the vertical distance, the angle value of the horizontal plane and the possibility of strain localization of the data point includes the following specific steps:

[0017] Any model is recorded as the first model;

[0018] ;

[0019] In the formula, For the first model The angle between the support where the strain gauge is located and the horizontal plane of the earthquake simulation platform, For the first model The vertical distance from the strain gauge to the earthquake simulation platform; for function; and The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , the first axial stress curve The axial stress value corresponding to the data point is recorded as ; For the first model The axial stress curve corresponding to the strain gauge The strain localization performance eigenvalues ​​for each data point.

[0020] Furthermore, the characteristic fitting curve corresponding to each strain gauge in each model is obtained according to the characteristic value of strain localization, and the specific steps include the following:

[0021] Construct a two-dimensional coordinate system to obtain the first model The strain localization characteristic value of each data point in the axial stress curve corresponding to the strain gauge is The order value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The abscissa of the data point in the axial stress curve corresponding to the strain gauge in the two-dimensional coordinate system; The strain localization characteristic value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The ordinate of the data point in the axial stress curve corresponding to the strain gauge is in the two-dimensional coordinate system. All data points in the axial stress curve corresponding to the strain gauge are mapped to the two-dimensional coordinate system to obtain a scatter plot, which is recorded as the first scatter plot. The first scatter plot is fitted to obtain a fitting curve, which is recorded as the first model in the first model. Characteristic fitting curve corresponding to each strain gauge.

[0022] Furthermore, the influence coefficient of each strain gauge position in each model is obtained according to the slope of each point in the characteristic fitting curve, and the specific steps include the following:

[0023] In the first model, get a plurality of strain gauges adjacent to a strain gauge; the first The specific method for obtaining several strain gauges adjacent to the first strain gauge is: The support or support beam directly connected to the support where the strain gauge is located is recorded as The adjacent structure of the pillar where the strain gauge is located The strain gauges in all adjacent structures of the pillar where the strain gauge is located are used as the strain gauges adjacent to each other;

[0024] ;

[0025] In the formula, For the first model and the The number of adjacent strain gauges, For the first model The characteristic fitting curve corresponding to the strain gauge The slope of the point, For the first model and the The strain gauge adjacent to The characteristic fitting curve corresponding to the strain gauge The slope of the point, is the number of points in the characteristic fitting curve; For the first model The influence coefficient of the location of each strain gauge is To take the absolute value.

[0026] Furthermore, the method of obtaining the abnormal parameters of each data point in the axial stress curve corresponding to each strain gauge in each model according to the abnormal score, the possibility of strain localization generated by the data point and the influence coefficient includes the following specific steps:

[0027] ;

[0028] In the formula, The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , is the first axial stress curve The abnormal fraction of axial stress values ​​corresponding to the data points is For the first model and the The strain gauge adjacent to The influence coefficient of the location of each strain gauge is For the first model and the The number of adjacent strain gauges, is the normalization function, For the first model The axial stress curve corresponding to the strain gauge The outlier parameter of each data point.

[0029] Furthermore, the method of obtaining a plurality of strain localization data points of the axial stress curve corresponding to each strain gauge in each model according to the size of the abnormal parameter includes the following specific steps:

[0030] A first threshold is preset, and the abnormal parameter of each data point in the axial stress curve corresponding to each strain gauge in each model is obtained, and the data point whose abnormal parameter is greater than the first threshold is used as the strain localization data point.

[0031] Furthermore, the method of obtaining the performance parameters of the earthquake-resistant structure according to the time corresponding to the strain localization data point includes the following specific steps:

[0032] Obtain the first strain localization data point of the axial stress curve corresponding to each strain gauge in the first model, record it as the target strain localization data point, record the time corresponding to each target strain localization data point as the first time, and compare each first time with The ratio of is recorded as the second time, is a preset second value; the average value of all second times is recorded as the reference time of the first model, and the reference time of each model is obtained;

[0033] ;

[0034] In the formula, For the The reference time of the model, For the The seismic load value corresponding to each model; is the number of earthquake load values ​​corresponding to all models, It is the performance parameter of earthquake-resistant structure.

[0035] Furthermore, the abnormality detection of the axial stress curve to obtain the abnormality score of each axial stress value in each axial stress curve includes the following specific steps:

[0036] Any model is recorded as the first model, several axial stress curves of the first model under the corresponding seismic load value are recorded as several axial stress curves of the first model, and any axial stress curve of the first model is recorded as the target axial stress curve; the target axial stress curve is input into the isolation forest algorithm for anomaly detection, and the anomaly score of each axial stress value in the target axial stress curve is obtained as output.

[0037] Furthermore, the specific steps of presetting the local range of each data point in each axial stress curve include the following:

[0038] Any model is recorded as the first model, several axial stress curves of the first model under the corresponding seismic load value are recorded as several axial stress curves of the first model, and any axial stress curve of the first model is recorded as the target axial stress curve; any data point in the target axial stress curve is recorded as the target data point; in the target axial stress curve, the neighborhood radius is centered at the target data point. The range of is recorded as the local range of the target data point, is a preset first value.

[0039] The beneficial effects of the technical solution of the present invention are as follows: after obtaining several models of earthquake-resistant structures scaled in proportion, the present invention obtains several axial stress curves of each model under a preset earthquake load value, performs preliminary anomaly detection on the axial stress curve, obtains the anomaly score of each axial stress value in each axial stress curve, obtains the possibility of strain localization for each data point in each axial stress curve through the local range of the data point in the axial stress curve, the slope of the data point and the axial stress value corresponding to the data point, obtains the vertical distance from each strain gauge in the model to the earthquake simulation platform, the angle between the support where each strain gauge in the model is located and the horizontal plane of the earthquake simulation platform and the possibility of strain localization for the data point, obtains The strain localization characteristic value of each data point in the axial stress curve is obtained. When obtaining the strain localization characteristic value, the whipping effect of the analysis model is used to make the stress influence between adjacent positions in the subsequent analysis model more accurate. The abnormal parameters of the data point are obtained through the influence coefficient, abnormal score and possibility of strain localization of each strain gauge position in the model. Then, the performance parameters of the seismic resistant structure are obtained by combining the strain localization data points of multiple models, which reduces the misjudgment of strain localization caused by data fluctuations generated by the stress data itself, improves the accuracy of data positioning and identification when strain localization occurs, and thus improves the accuracy and authenticity of the evaluation of the seismic performance of the seismic resistant structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 A flowchart of the steps of a method for testing the seismic performance of a steel-concrete seismic resistant structure provided by one embodiment of the present invention;

[0042] Figure 2 A flow chart for obtaining performance parameters of an earthquake-resistant structure provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the seismic performance testing method of a steel-concrete seismic resistant structure proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0044] Unless defined otherwise, 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 belongs.

[0045] The specific scheme of the seismic performance testing method of a steel-concrete seismic resistant structure provided by the present invention is described in detail below with reference to the accompanying drawings.

[0046] See also Figure 1 , which shows a flowchart of a method for testing the seismic performance of a steel-concrete seismic resistant structure provided by an embodiment of the present invention, the method comprising the following steps:

[0047] Step S001, obtaining several proportionally scaled models of earthquake-resistant structures, installing several strain gauges on each model, and obtaining several axial stress curves of each model under a preset earthquake load value based on the several strain gauges installed on each model.

[0048] Specifically, several models of earthquake-resistant structures are obtained in equal proportion, and any one of the models is recorded as the first model; the first model is placed on an earthquake simulation platform of a dynamic experiment; strain gauges are installed at the pillar positions of the first model, the middle positions of various support beams and other structures, and the intersection positions of the structures; a seismic load sequence is preset, and the seismic load sequence includes several seismic load values. In this embodiment, the seismic load sequence is taken as Describe; obtain several axial stress curves of the first model under the first seismic load value in the seismic load sequence through several strain gauges on the first model, and obtain several axial stress curves of each model under the corresponding seismic load value.

[0049] It should be noted that the ratio of the equal-proportional scaling in this embodiment is , each model corresponds to a unique seismic load value, for example, the first model corresponds to the first seismic load value, and the second model corresponds to the second seismic load value; the seismic load represents the vibration intensity of the seismic simulation platform of the dynamic experiment, and different seismic load values ​​should be gradually increased until they exceed the seismic resistance of the model, so as to facilitate the subsequent analysis of the seismic performance of the model. There is no fixed limit on the seismic load value here, and it can be adjusted according to the specific implementation situation during specific implementation. The proportions of several models of seismic structure scaled proportionally are the same; the sampling intervals of several strain gauges are all 0.1s, and the total acquisition time is 1min. Each strain gauge corresponds to an axial stress curve, and each axial stress curve contains several data points. The axial stress curves are all two-dimensional curves, and each data point in the axial stress curve corresponds to a time and an axial stress value.

[0050] Furthermore, the axial stress curve is subjected to anomaly detection to obtain the anomaly score of each axial stress value in each axial stress curve, as follows:

[0051] Several axial stress curves of the first model under corresponding seismic load values ​​are recorded as several axial stress curves of the first model, and any axial stress curve of the first model is recorded as a target axial stress curve; the target axial stress curve is input into an isolation forest algorithm for anomaly detection, and the anomaly score of each axial stress value in the target axial stress curve is obtained as output.

[0052] It should be noted that using the isolation forest algorithm to obtain the abnormal score of each value in the curve is an existing method of the isolation forest algorithm, which will not be described in detail in this embodiment.

[0053] At this point, the abnormal score of each axial stress value in each axial stress curve is obtained.

[0054] Step S002, presetting the local range of each data point in each axial stress curve; obtaining the slope of each data point in each axial stress curve, and obtaining the possibility of strain localization for each data point in each axial stress curve according to the local range, the slope of the data point and the axial stress value corresponding to the data point.

[0055] It should be noted that the above-mentioned relevant stress data at different positions in the model are obtained. The intensity of the simulated earthquake is fixed during the dynamic experiment, but the change of its amplitude is nonlinear, which leads to a large difference between the axial stress data at the corresponding position and other data. Therefore, the anomaly score directly obtained by the isolation forest algorithm cannot accurately represent the characteristics of strain localization. It is also necessary to determine the strain localization in combination with the performance of this point in all data.

[0056] It should be noted that due to the different amplitudes and periods of the simulated earthquakes, the lateral force generated on the model structure and the inertia of the structure itself continuously offset each other. However, in this process, the simulated earthquakes are constantly changing, so the collected axial stress has more obvious fluctuations, which are normal. When strain localization occurs, the structural rigidity at that location increases sharply, which increases the axial stress at that location. In addition, the baseline of the axial stress data at subsequent time points will rise due to the occurrence of strain localization, and the fluctuation of the axial stress is more disordered and chaotic than before the strain localization occurs. Therefore, the possibility of strain localization is obtained through local analysis.

[0057] Specifically, the local range of each data point in each axial stress curve is preset as follows:

[0058] Any data point in the target axial stress curve is recorded as the target data point; in the target axial stress curve, the target data point is taken as the center, and the neighborhood radius is The range of is recorded as the local range of the target data point, is a preset first value. Give a narrative.

[0059] It should be noted that if the target data point is on both sides of the target axial stress curve, the local range of the target data point may exceed the range of the target axial stress curve. If the local range exceeds the range of the target axial stress curve, this embodiment will not perform analysis.

[0060] Furthermore, the slope of each data point in each axial stress curve is obtained, and the possibility of strain localization at each data point in each axial stress curve is obtained according to the local range, the slope of the data point and the axial stress value corresponding to the data point. As an embodiment, the specific calculation method is:

[0061] It should be noted that obtaining the slope of a data point on a curve is an existing method, which will not be described in detail in this embodiment;

[0062] ;

[0063] In the formula, The target axial stress curve The axial stress value corresponding to the data point is The target axial stress curve The maximum value of the axial stress value corresponding to the data point in the local range of the data points; and The specific method of obtaining is as follows: The local range of the data points is recorded as the first local range, and the The variance of the axial stress values ​​corresponding to all the data points on the left side of the data point is recorded as , the first local range The variance of the axial stress values ​​corresponding to all the data points to the right of the data point is recorded as ; The target axial stress curve The number of data points in the local range of data points, The target axial stress curve In the local range of data points The slope of the data point, The target axial stress curve In the local range of data points The slope of the data point, To take the absolute value, is a preset hyperparameter, the purpose of which is to prevent the denominator from being 0. To narrate, The target axial stress curve The probability of strain localization for each data point.

[0064] It should be noted that Indicates The axial stress value corresponding to the data point is normalized. The larger the value, the greater the The greater the possibility of a data point producing a sharp increase, the higher the possibility of strain localization. It represents the ratio of the variance of the stress value on the right to the variance of the stress value on the left. The variance of the stress data itself reflects the degree of disorder of the data on the left and right sides. The larger the variance, the more disordered the data fluctuation. When strain localization occurs at the corresponding time of a data point, the variance on the right side must be greater than that on the left side. Therefore, the larger the value of the fraction, the greater the possibility of strain localization. Indicates The slope difference between any two adjacent data points in the local range of data points. When strain localization occurs, there will be a steep increase in the slope between two data points in the local range. If strain localization occurs in the local range but the location is not in the first At the data point, The possibility of strain localization at the data point is high, so the Indicates the slope change position and the The distance relationship between the data points. The smaller the value, the The greater the possibility of strain localization in each data point, the greater the possibility of strain localization in each data point; traversing all data points in the local range, we can get If strain localization occurs, the larger the cumulative value, the greater the possibility of strain localization. Multiplying the above formulas, we get , indicating the The probability of strain localization at each data point in time.

[0065] Thus, the possibility of strain localization at each data point in each axial stress curve is obtained.

[0066] Step S003, obtaining the strain localization characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model; obtaining the characteristic fitting curve corresponding to each strain gauge in each model according to the strain localization characteristic value; obtaining the influence coefficient of the position of each strain gauge in each model according to the slope of each point in the characteristic fitting curve.

[0067] It should be noted that the axial stress at different positions of the seismic structure model is different. For example, the axial stress at the intersection of multiple pillars is smaller than the axial stress at the middle of the pillar, or the axial stress of the pillar at the high point of the seismic structure will increase due to the whip effect. The actual axial stress performance at different positions in the structure is affected by the conduction between the various structural points of the seismic structure. The more vertical the pillars of the seismic structure model are to the ground, and the farther the vertical distance from the platform surface, the stronger the whip effect, and the greater the stress impact on the surrounding positions when the strain localization occurs. Therefore, the angle between the pillars at each position and the platform and the vertical distance to the platform are obtained; at the same time, several other positions closest to the current acquisition position under the same experimental conditions are obtained, and their stress data and strain localization possibility are obtained. At this time, for the stress data of the current position, the degree of influence of the current position on the axial stress of other surrounding positions is judged according to the difference between the data and possibility of other positions close to the surrounding at the same time node.

[0068] Specifically, the vertical distance from each strain gauge in each model to the seismic simulation platform is obtained; the angle between the support where each strain gauge in each model is located and the horizontal plane of the seismic simulation platform is obtained, as follows:

[0069] Obtain the vertical distance from each strain gauge in the first model to the seismic simulation table; obtain the angle value between the pillar where each strain gauge is located and the horizontal plane of the seismic simulation table in the first model. It should be noted that in step S001, the model is placed on the seismic simulation table of the dynamic experiment; and strain gauges are installed at the pillar position of the model, the middle position of various support beams and other structures, and the intersection position of the structure. Here, it is necessary to obtain the vertical distance from each strain gauge to the seismic simulation table and the angle value between the pillar where the strain gauge is located and the horizontal plane of the seismic simulation table, so as to analyze the influence between the axial stresses at different positions of the seismic resistant structure model.

[0070] Furthermore, according to the vertical distance, the angle value of the horizontal plane and the possibility of strain localization of the data point, the strain localization performance characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model is obtained. As an embodiment, the specific calculation method is:

[0071] ;

[0072] In the formula, For the first model The angle between the support where the strain gauge is located and the horizontal plane of the earthquake simulation platform, For the first model The vertical distance from the strain gauge to the earthquake simulation platform; it should be noted that the angle value here only analyzes the pillar position of the model, and does not analyze the support beam and other structures, that is, The first strain gauge is a strain gauge at the support position; for Function, used for normalization; and The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , the first axial stress curve The axial stress value corresponding to the data point is recorded as ; For the first model The axial stress curve corresponding to the strain gauge The strain localization performance eigenvalues ​​for each data point.

[0073] It should be noted that Indicates The probability of strain localization at a data point is related to the The product value of the axial stress values ​​corresponding to the data points, which represents the characteristic value of the stress data of any data point at any position; combined with the normalized Get the strain localization characteristic value of the data point, Indicates The whip effect at the support position where the strain gauge is located is Indicates The difference between the angle between the support where the strain gauge is located and the platform and the vertical angle is smaller. The more vertical the support where the strain gauge is located is to the ground, the greater the whip effect will be. Therefore, the fractional expression is used. The farther the position is from the platform, the greater the whip effect will be. The larger the value, the stronger the whip effect. We can use the Euclidean norm to convert it into the same dimension and use Function normalization, using this value to weight the stress data, thus obtaining .

[0074] Furthermore, according to the characteristic value of strain localization, the characteristic fitting curve corresponding to each strain gauge in each model is obtained, as follows:

[0075] Construct a two-dimensional coordinate system to obtain the first model The strain localization characteristic value of each data point in the axial stress curve corresponding to the strain gauge is The order value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The abscissa of the data point in the axial stress curve corresponding to the strain gauge in the two-dimensional coordinate system; The strain localization characteristic value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The ordinate of the data point in the axial stress curve corresponding to the strain gauge is in the two-dimensional coordinate system. All data points in the axial stress curve corresponding to the strain gauge are mapped to the two-dimensional coordinate system to obtain a scatter plot, which is recorded as the first scatter plot. The first scatter plot is fitted to obtain a fitting curve, which is recorded as the first model in the first model. The characteristic fitting curve corresponding to each strain gauge; in this embodiment, the least square method is used to fit the first scatter plot, wherein the fitting curve is a quintic polynomial curve.

[0076] Furthermore, the slope of each point in the characteristic fitting curve corresponding to each strain gauge in each model is obtained, and the influence coefficient of the position of each strain gauge in each model is obtained according to the slope of each point in the characteristic fitting curve. As an embodiment, the specific calculation method is:

[0077] It should be noted that obtaining the slope of each point in the fitting curve is an existing method, which will not be described in detail in this embodiment.

[0078] In the first model, get a plurality of strain gauges adjacent to a strain gauge; the first The specific method for obtaining several strain gauges adjacent to the first strain gauge is: The support or support beam directly connected to the support where the strain gauge is located is recorded as The adjacent structure of the pillar where the strain gauge is located The strain gauges in all adjacent structures of the pillar where the strain gauge is located are used as the strain gages adjacent to each other.

[0079] ;

[0080] In the formula, For the first model and the The number of adjacent strain gauges, For the first model The characteristic fitting curve corresponding to the strain gauge The slope of the point, For the first model and the The strain gauge adjacent to The characteristic fitting curve corresponding to the strain gauge The slope of the point, is the number of midpoints of the characteristic fitting curve; it should be noted that the number of midpoints of the characteristic fitting curves corresponding to different strain gauges is the same; For the first model The influence coefficient of the location of each strain gauge.

[0081] It should be noted that Indicates the first model The strain gauge and The first-order derivative difference of the strain localization characteristic value of adjacent strain gauges at the same time is smaller, indicating that the The greater the influence of each strain gauge on the adjacent position, the fractional expression is used to traverse all adjacent positions and all times, thus obtaining , The larger the value, the greater the The greater the influence of each strain gauge on the axial stress in other adjacent locations.

[0082] At this point, the influence coefficient of each strain gauge location in each model is obtained.

[0083] Step S004: Obtain the abnormal parameters of each data point in the axial stress curve corresponding to each strain gauge in each model according to the abnormal score, the possibility of strain localization generated by the data point and the influence coefficient; obtain a number of strain localization data points of the axial stress curve corresponding to each strain gauge in each model according to the size of the abnormal parameters; obtain the performance parameters of the seismic resistant structure according to the time corresponding to the strain localization data point.

[0084] It should be noted that in the isolated tree, the position characteristics of the current point in the isolated tree are first related to its own path length. The shorter the path length, the greater the position abnormality characteristics. In order to eliminate the influence of the randomness of the isolated tree on the path length, it is also necessary to analyze the position of the data point in the isolated tree and calculate its abnormal characteristics about the position in the isolated tree through the possibility of strain localization generated by the axial stress data point and the degree of influence of other structural positions on the current position. In the above calculation process, only the influence relationship between the nearest sampling positions is considered. Therefore, for the data at the current structural position, when judging whether it is affected by other positions, only the influence coefficients calculated by these nearest sampling positions are taken to judge its influence on the current position, thereby obtaining its abnormal parameters.

[0085] Specifically, according to the abnormal score, the possibility of strain localization generated by the data point and the influence coefficient, the abnormal parameter of each data point in the axial stress curve corresponding to each strain gauge in each model is obtained. As an embodiment, the specific calculation method is:

[0086] ;

[0087] In the formula, The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , is the first axial stress curve The abnormal fraction of axial stress values ​​corresponding to the data points is For the first model and the The strain gauge adjacent to The influence coefficient of the location of each strain gauge is For the first model and the The number of adjacent strain gauges, is a linear normalization function, and the normalized object is All data points in the axial stress curve corresponding to the strain gauge , For the first model The axial stress curve corresponding to the strain gauge The outlier parameter of each data point.

[0088] It should be noted that Indicates The adjacent strain gauge pair The stress influence degree of the location of each strain gauge. The larger the value, the higher the authenticity of the strain localization generated by the data. The anomaly score is adjusted in combination with the possibility of strain localization to obtain the anomaly parameter. The larger the value of the anomaly parameter, the higher the possibility of strain localization.

[0089] Furthermore, according to the size of the abnormal parameters, several strain localization data points of the axial stress curve corresponding to each strain gauge in each model are obtained, as follows:

[0090] A first threshold is preset. In this embodiment, the first threshold is described as 0.7. The abnormal parameters of each data point in the axial stress curve corresponding to each strain gauge in each model are obtained, and the data points with abnormal parameters greater than the first threshold are used as strain localization data points. It should be noted that the data points less than or equal to the first threshold are not analyzed here.

[0091] It should be noted that the above-mentioned several strain localization data points of the axial stress curve corresponding to each strain gauge in each model are obtained. At this time, for the entire seismic structure model, the seismic performance of all models under different seismic load values ​​is comprehensively judged.

[0092] Specifically, the performance parameters of the earthquake-resistant structure are obtained according to the time corresponding to the strain localization data point. As an embodiment, the specific calculation method is:

[0093] Obtain the first strain localization data point of the axial stress curve corresponding to each strain gauge in the first model, record it as the target strain localization data point, record the time corresponding to each target strain localization data point as the first time, and compare each first time with The ratio of is recorded as the second time, is a preset second value. The description is that the total duration of any axial stress curve is normalized by this; the average value of all second times is recorded as the reference time of the first model, and the reference time of each model is obtained.

[0094] ;

[0095] In the formula, For the The reference time of the model, For the The seismic load value corresponding to each model; it should be noted again that each model corresponds to only one seismic load value; is the number of earthquake load values ​​corresponding to all models, It is the performance parameter of earthquake-resistant structure.

[0096] The larger the performance parameter value of the earthquake-resistant structure, the stronger the earthquake-resistant performance of the earthquake-resistant structure. Figure 2 , Figure 2 This is a flow chart of obtaining performance parameters of the earthquake-resistant structure in this embodiment.

[0097] It should be noted that through Weight the seismic load values ​​and traverse all models to obtain the performance parameters of the seismic resistant structure , The larger the value, the better the seismic performance of the seismic-resistant structure.

[0098] At this point, the performance parameters of the seismic structure are obtained by analyzing multiple seismic structure models at different seismic load values, and the seismic performance test of the steel concrete seismic structure is completed.

[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for testing the seismic performance of a steel-concrete seismic structure, characterized in that: The method comprises the following steps: Acquire several models of earthquake-resistant structures that are scaled in proportion, install several strain gauges on each model, and acquire several axial stress curves of each model under a preset earthquake load value according to the several strain gauges installed on each model, wherein the axial stress curve includes several data points, and each data point corresponds to a time and an axial stress value; perform anomaly detection on the axial stress curve to obtain an anomaly score of each axial stress value in each axial stress curve; Preset the local range of each data point in each axial stress curve; obtain the slope of each data point in each axial stress curve, and obtain the possibility of strain localization for each data point in each axial stress curve according to the local range, the slope of the data point and the axial stress value corresponding to the data point; The method of obtaining the possibility of strain localization at each data point in each axial stress curve according to the local range, the slope of the data point and the axial stress value corresponding to the data point includes the following specific steps: Record any one model as a first model, record several axial stress curves of the first model under corresponding earthquake load values ​​as several axial stress curves of the first model, and record any one axial stress curve of the first model as a target axial stress curve; ； In the formula, The target axial stress curve The axial stress value corresponding to the data point is The target axial stress curve The maximum value of the axial stress value corresponding to the data point in the local range of the data points; and The specific method of obtaining is as follows: The local range of the data points is recorded as the first local range, and the The variance of the axial stress values ​​corresponding to all the data points on the left side of the data point is recorded as , the first local range The variance of the axial stress values ​​corresponding to all the data points to the right of the data point is recorded as ; The target axial stress curve The number of data points in the local range of data points, The target axial stress curve In the local range of data points The slope of the data point, The target axial stress curve In the local range of data points The slope of the data point, To take the absolute value, is a preset hyperparameter. The target axial stress curve The probability of strain localization at each data point; Obtain the vertical distance from each strain gauge in each model to the seismic simulation platform; obtain the angle between the support where each strain gauge is located and the horizontal plane of the seismic simulation platform in each model; obtain the strain localization performance characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model based on the vertical distance, the angle value of the horizontal plane and the possibility of strain localization of the data point; The method of obtaining the strain localization characteristic value of each data point in the axial stress curve corresponding to each strain gauge in each model according to the vertical distance, the angle value of the horizontal plane and the possibility of strain localization of the data point includes the following specific steps: Any model is recorded as the first model; ； In the formula, For the first model The angle between the support where the strain gauge is located and the horizontal plane of the earthquake simulation platform, For the first model The vertical distance from the strain gauge to the earthquake simulation platform; for function; and The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , the first axial stress curve The axial stress value corresponding to the data point is recorded as ; For the first model The axial stress curve corresponding to the strain gauge The strain localization performance eigenvalue of each data point; According to the characteristic value of strain localization, the characteristic fitting curve corresponding to each strain gauge in each model is obtained; the slope of each point in the characteristic fitting curve corresponding to each strain gauge in each model is obtained, and the influence coefficient of the position of each strain gauge in each model is obtained according to the slope of each point in the characteristic fitting curve; The specific steps of obtaining the influence coefficient of each strain gauge position in each model according to the slope of each point in the characteristic fitting curve are as follows: In the first model, get a plurality of strain gauges adjacent to a strain gauge; the first The specific method for obtaining several strain gauges adjacent to the first strain gauge is: The support or support beam directly connected to the support where the strain gauge is located is recorded as The adjacent structure of the pillar where the strain gauge is located The strain gauges in all adjacent structures of the pillar where the strain gauge is located are used as the strain gauges adjacent to each other; ； In the formula, For the first model and the The number of adjacent strain gauges, For the first model The characteristic fitting curve corresponding to the strain gauge The slope of the point, For the first model and the The strain gauge adjacent to The characteristic fitting curve corresponding to the strain gauge The slope of the point, is the number of points in the characteristic fitting curve; For the first model The influence coefficient of the location of each strain gauge is To take the absolute value; According to the anomaly score, the possibility of strain localization generated by the data point and the influence coefficient, the anomaly parameter of each data point in the axial stress curve corresponding to each strain gauge in each model is obtained; The specific steps of obtaining the abnormal parameters of each data point in the axial stress curve corresponding to each strain gauge in each model according to the abnormal score, the possibility of strain localization generated by the data point and the influence coefficient are as follows: ； In the formula, The specific method of obtaining is as follows: The axial stress curve corresponding to the strain gauge is recorded as the first axial stress curve. The possibility of strain localization at a data point is denoted as , is the first axial stress curve The abnormal fraction of axial stress values ​​corresponding to the data points is For the first model and the The strain gauge adjacent to The influence coefficient of the location of each strain gauge is For the first model and the The number of adjacent strain gauges, is the normalization function, For the first model The axial stress curve corresponding to the strain gauge The abnormal parameters of the data points; According to the size of the abnormal parameter, several strain localization data points of the axial stress curve corresponding to each strain gauge in each model are obtained; according to the time corresponding to the strain localization data point, the performance parameters of the seismic resistant structure are obtained; The method of obtaining the performance parameters of the earthquake-resistant structure according to the time corresponding to the strain localization data point includes the following specific steps: Obtain the first strain localization data point of the axial stress curve corresponding to each strain gauge in the first model, record it as the target strain localization data point, record the time corresponding to each target strain localization data point as the first time, and compare each first time with The ratio of is recorded as the second time, is a preset second value; the average value of all second times is recorded as the reference time of the first model, and the reference time of each model is obtained; ； In the formula, For the The reference time of the model, For the The seismic load value corresponding to each model; is the number of earthquake load values ​​corresponding to all models, It is the performance parameter of earthquake-resistant structure.

2. The method for testing the seismic performance of a steel-concrete seismic structure according to claim 1, characterized in that: The specific steps of obtaining the characteristic fitting curve corresponding to each strain gauge in each model according to the characteristic value of strain localization performance are as follows: Construct a two-dimensional coordinate system to obtain the first model The strain localization characteristic value of each data point in the axial stress curve corresponding to the strain gauge is The order value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The abscissa of the data point in the axial stress curve corresponding to the strain gauge in the two-dimensional coordinate system; The strain localization characteristic value of the data point in the axial stress curve corresponding to the strain gauge is taken as the The ordinate of the data point in the axial stress curve corresponding to the strain gauge is in the two-dimensional coordinate system. According to the abscissa and ordinate, the All data points in the axial stress curve corresponding to the strain gauge are mapped to the two-dimensional coordinate system to obtain a scatter plot, which is recorded as the first scatter plot. The first scatter plot is fitted to obtain a fitting curve, which is recorded as the first model in the first model. Characteristic fitting curve corresponding to each strain gauge.

3. The method for testing the seismic performance of a steel-concrete seismic structure according to claim 1, characterized in that: The specific steps of obtaining a number of strain localization data points of the axial stress curve corresponding to each strain gauge in each model according to the size of the abnormal parameter are as follows: A first threshold is preset, and the abnormal parameter of each data point in the axial stress curve corresponding to each strain gauge in each model is obtained, and the data point whose abnormal parameter is greater than the first threshold is used as the strain localization data point.

4. The method for testing the seismic performance of a steel-concrete seismic structure according to claim 1, characterized in that: The abnormality detection of the axial stress curve to obtain the abnormality score of each axial stress value in each axial stress curve includes the following specific steps: Record any one model as a first model, record several axial stress curves of the first model under corresponding earthquake load values ​​as several axial stress curves of the first model, and record any one axial stress curve of the first model as a target axial stress curve; The target axial stress curve is input into the isolation forest algorithm for anomaly detection, and the anomaly score of each axial stress value in the target axial stress curve is obtained as output.

5. The method for testing the seismic performance of a steel-concrete seismic structure according to claim 1, characterized in that: The specific steps of presetting the local range of each data point in each axial stress curve include the following: Record any one model as a first model, record several axial stress curves of the first model under corresponding earthquake load values ​​as several axial stress curves of the first model, and record any one axial stress curve of the first model as a target axial stress curve; Any data point in the target axial stress curve is recorded as the target data point; in the target axial stress curve, the target data point is taken as the center, and the neighborhood radius is The range of is recorded as the local range of the target data point, is a preset first value.

Images