Method for obtaining ultimate stability performance curve of concrete structure and application thereof

Abstract

The application belongs to the technical field of building anti-collapse, and particularly relates to a method for obtaining a limit stability performance curve of a concrete structure and application thereof. The method comprises: taking limit loads of different concrete structures as training samples, training a first neural network model to obtain a limit load prediction model; determining a stable load range based on the limit load; then taking performance curves constituted by different loads and corresponding column top limit displacements of different concrete structures under different column removal conditions within the stable load range of the concrete structures as training samples, training a second neural network model to obtain a performance curve prediction model; after the limit load prediction model predicts the limit load of a to-be-predicted concrete structure, the performance curve prediction model outputs the performance curve of the to-be-predicted concrete structure according to the to-be-predicted concrete structure and the corresponding limit load. The application can efficiently and accurately obtain the limit stability performance curve of the concrete structure.

Description

Technical Field

[0001] This application belongs to the field of building collapse resistance technology, and in particular relates to a method for obtaining the ultimate stability performance curve of a concrete structure and its application. Background Technology

[0002] The collapse of concrete structures often causes serious casualties and huge economic losses. Therefore, it is of great significance to accurately assess and predict the ultimate stability of concrete structures.

[0003] Currently, displacement sensors are typically installed at key locations in concrete structures to monitor the displacement response of the concrete structure in real time, and the stress state of the concrete structure is determined accordingly.

[0004] However, for different concrete structural systems, the same level of displacement may correspond to completely different damage states or safety margins. Therefore, relying solely on structural displacement as a single indicator makes it difficult to accurately determine the current state of the concrete structure relative to its ultimate stability, and also makes it difficult to fully reflect the impact of subsequent impacts on the ultimate stability of the current concrete structure. Summary of the Invention

[0005] The purpose of this application is to overcome the shortcomings of the prior art and provide a method for obtaining the ultimate stability performance curve of a concrete structure, which can efficiently and accurately obtain the ultimate stability performance curve of the concrete structure.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] A method for obtaining the ultimate stability performance curve of a concrete structure includes the following steps:

[0008] Step 1: Use finite element analysis to remove columns from different concrete structures within the gravity load bearing range to obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads. The ultimate load refers to the maximum load that the concrete structure can still recover to a stable state after a column is removed.

[0009] Step 2: Using the ultimate loads of different concrete structures as training samples, the first neural network model is trained to obtain the ultimate load prediction model.

[0010] Step 3: Based on the ultimate load, determine the range of stable load; then, use the performance curves of different concrete structures under different column removal conditions, which are composed of different loads within their own range of stable load and the corresponding ultimate displacement of the column top, as training samples to train the second neural network model and obtain the performance curve prediction model.

[0011] Step 4: After the ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted, the performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the concrete structure to be predicted and the corresponding ultimate load.

[0012] Preferably, step 1 further includes the following sub-steps:

[0013] Step 11: Obtain the structural performance parameters of different concrete structures; the structural performance parameters include structural geometric parameters and material performance parameters.

[0014] Step 12: Using finite element analysis, starting from the self-load of the current concrete structure, continuously increase the gravity load applied to the current concrete structure until the current concrete structure collapses, and obtain the gravity load bearing range of the current concrete structure.

[0015] Step 13: Within the range of gravity load bearing capacity, remove columns according to different column removal strategies under each gravity load value, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal; at the same time, determine the ultimate load of each concrete structure during the column removal process.

[0016] Preferably, in step 13, obtaining the maximum displacement of the column top when the concrete structure stabilizes after each column removal also includes the following sub-steps:

[0017] S1, obtain the column top sway displacement of the column being extracted this time at the first sampling time, the second sampling time and the third sampling time after the column extraction time; the first sampling time, the second sampling time and the third sampling time refer to the time intervals Δt, 2Δt and 3Δt from the column extraction time this time, respectively.

[0018] S2, if the column top swaying displacement of the same column decreases sequentially at the first, second, and third sampling times, then at the fourth sampling time after the third sampling time, the column top swaying displacement of the extracted column is obtained and S3 is executed. Otherwise, it is determined that the concrete structure cannot be stable after the column is extracted, and the concrete structure does not have the maximum column top displacement when it is stable after the column is extracted. The fourth sampling time refers to the time interval ΔT from the third sampling time.

[0019] S3 is the maximum displacement of the column top swaying among the columns removed this time, which is taken as the maximum displacement of the column top when the concrete structure is stable after the column is removed.

[0020] Preferably, step 2 further includes the following: binding the structural performance parameters of different concrete structures with the corresponding ultimate loads to training samples of the first training set, wherein the ultimate load is the sample label; training the first neural network model using the training samples of the first training set to obtain the ultimate load prediction model; the ultimate load prediction model predicts the ultimate load of the corresponding concrete structure based on the structural performance parameters.

[0021] Preferably, the ultimate load in step 3 is obtained from the finite element analysis performed in step 1, or from the ultimate load prediction model in step 2.

[0022] Preferably, step 3 further includes the following sub-steps:

[0023] Step 31: The load range formed by the self-load and ultimate load of the current concrete structure is denoted as the stable load range of the current concrete structure.

[0024] Step 32: Based on the stable load range, set the load increment Δa of the current concrete structure within the stable load range;

[0025] Step 33: Starting from the self-load of the current concrete structure, increment the gravity load value after each increment, as well as the corresponding self-load and ultimate load, as the load point for removing columns in the current concrete structure.

[0026] Step 34: At each load point, remove columns from the current concrete structure according to different column removal strategies, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal.

[0027] Step 35: Take the maximum value of the maximum displacement of the column top obtained by the current concrete structure at the same load point among all column removal strategies, and take it as the ultimate displacement of the column top of the current concrete structure at the corresponding load point.

[0028] Step 36: Using each load point of the current concrete structure as the vertical axis and the corresponding column top ultimate displacement as the horizontal axis, fit the performance curve of the current concrete structure.

[0029] Step 37: The structural performance parameters of several concrete structures and the corresponding performance curves within the stable load range are used as training samples of the second training set. The second neural network model is trained to obtain the performance curve prediction model. The performance curve prediction model predicts the performance curve of the concrete structure within the corresponding stable load range based on the structural performance parameters of the concrete structure and the stable load range.

[0030] Preferably, step 4 further includes the following sub-steps:

[0031] Step 41: The ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted based on the structural performance parameters of the concrete structure to be predicted.

[0032] Step 42: Obtain the stability load range of the concrete structure to be predicted based on the ultimate load of the concrete structure to be predicted.

[0033] Step 43: The performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the structural performance parameters and stable load range of the concrete structure to be predicted.

[0034] This application also provides a method for safety assessment of a construction scheme, including the following steps:

[0035] Step 01: According to the construction plan, obtain the maximum column top displacement fitted by the finite element analysis of the concrete structure to be constructed; at the same time, use the method described above for obtaining the limit stability performance curve of a concrete structure to obtain the performance curve of the concrete structure to be constructed.

[0036] Step 02: If the maximum column top displacement is to the left of the performance curve of the concrete structure to be constructed, the current construction plan is deemed safe; otherwise, the current construction plan is deemed unsafe and the construction plan is adjusted.

[0037] This application also provides a method for safety assessment of damaged concrete structures, including the following steps:

[0038] Step 01': Obtain the initial concrete structure before damage. Use the method described above for obtaining the ultimate stability performance curve of a concrete structure to obtain the performance curve of the initial concrete structure.

[0039] Step 02': Obtain the maximum column top displacement of the damaged concrete structure relative to the initial concrete structure. If the maximum column top displacement is to the left of the performance curve of the initial concrete structure, the current damaged concrete structure is determined to be safe; otherwise, the current damaged concrete structure is determined to be dangerous.

[0040] This application also provides a system for obtaining the ultimate stability performance curve of a concrete structure, comprising: a first data module, a first training module, a second training module, and a performance curve prediction module; the first data module is used to perform column extraction on different concrete structures within the gravity load bearing range using finite element analysis, obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads, and then send them to the first training module and the second training module respectively; the first training module uses the ultimate loads of different concrete structures as training samples to train a first neural network model to obtain an ultimate load prediction model, which is then sent to the performance curve prediction module; the second training module determines the stable load range based on the ultimate load; then, under different column extraction conditions, the performance curves formed by different loads and corresponding column top ultimate displacements within their own stable load range for different concrete structures are used as training samples to train a second neural network model to obtain a performance curve prediction model, which is then sent to the performance curve prediction module; the performance curve prediction module outputs the corresponding performance curve according to the concrete structure to be predicted; each module is programmed or configured to execute the steps of the method for obtaining the ultimate stability performance curve of a concrete structure as described above.

[0041] The beneficial effects of this application are as follows:

[0042] (1) The method for obtaining the ultimate stability performance curve of this application is to efficiently and accurately predict the ultimate load of the concrete structure by training the ultimate load prediction model; and then using the trained performance curve prediction model to efficiently and accurately output the performance curve of the concrete structure to be predicted based on the concrete structure to be predicted and the corresponding ultimate load.

[0043] (2) In the method for obtaining the limit stability performance curve of this application, a lot of computing power is saved and efficiency is improved in the process of data sample acquisition and model training:

[0044] ① In obtaining the maximum displacement of the column top when the concrete structure stabilizes after each column removal, the trend of the column top swaying displacement within a very short time interval after column removal is used to determine whether the concrete structure after this column removal has the ability to recover to a stable state. Only if it has the ability to recover to a stable state will this application spend ΔT time to wait for it to recover to a stable state before collecting the corresponding maximum displacement of the column top; otherwise, this application will not spend too much time blindly waiting after column removal. An unstable concrete structure is unsafe, and the displacement of any column top in an unstable concrete structure is constantly changing, so the data of an unstable concrete structure has no value for subsequent use. This can significantly shorten the time spent on the initial data collection and processing while ensuring the quality of the collected data, and improve the efficiency of subsequent performance curve construction.

[0045] ②If the ultimate load of the training samples used when training the second neural network model is obtained through the ultimate load prediction model, then the order of training the ultimate load prediction model first and then training the performance curve prediction model in this application not only improves the training efficiency of the performance curve prediction model, but also shortens the overall time spent obtaining the ultimate stability performance curve.

[0046] (3) The performance curve acquisition method of this application introduces finite element analysis only in the early stage of training sample acquisition. The obtained physical characteristic parameters such as the ultimate displacement of concrete structure make the trained model subject to the constraints of the structural physical mechanism on the basis of data driving, thereby effectively reducing the prediction error caused by multi-parameter coupling and improving the accuracy and robustness of the prediction of ultimate stability performance curve. In formal use, the trained ultimate load prediction model and performance curve prediction model can be used to realize the accurate and efficient prediction of various new and specific ultimate stability performance curves of concrete structures in one stop. It not only has high prediction flexibility and applicability, but also avoids a large number of nonlinear dynamic time history analysis calculations in the prior art, and greatly improves the prediction efficiency.

[0047] (4) The ultimate stability performance curve obtained by the performance curve acquisition method of this application is accurate while prioritizing safety. Therefore, the safety factor of various applications using the ultimate stability performance curve of this application is extremely high. This is because the training samples used in the process of obtaining the ultimate load prediction model and the performance curve prediction model have already excluded data that collapsed or could not be restored to a stable state; and the acquisition of the ultimate load will also predict that the performance curve of the concrete structure will end at the ultimate load point, ensuring that the concrete structure can still be restored to a stable state after at least one column is removed, further improving the safety factor of the left region of the ultimate stability performance curve and the application value of the ultimate stability performance curve.

[0048] (5) The application of the performance curve of concrete structure in this application can not only evaluate the safety of the construction plan, but also evaluate the safety of the damaged concrete structure. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating a method for obtaining the ultimate stability performance curve of a concrete structure according to this application.

[0050] Figure 2 This is a schematic diagram showing the difference between the limiting stability performance curve of this application and the simulation results. Detailed Implementation

[0051] To make the technical solution of this application clearer and more explicit, the application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Solutions derived by those skilled in the art through equivalent substitution and conventional reasoning of the technical features of the technical solution of this application without creative effort all fall within the protection scope of this application.

[0052] like Figure 1 The flowchart shown in this application illustrates a method for obtaining the ultimate stability performance curve of a concrete structure, comprising:

[0053] Step 1: Use finite element analysis to remove columns from different concrete structures within the gravity load bearing range to obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads. The ultimate load refers to the maximum load that the concrete structure can still recover to a stable state after a column is removed.

[0054] The stable state of a concrete structure refers to the absence of collapse or vibration.

[0055] Step 2: Using the ultimate loads of different concrete structures as training samples, the first neural network model is trained to obtain the ultimate load prediction model.

[0056] Step 3: Based on the ultimate load, determine the range of stable load; then, use the performance curves of different concrete structures under different column removal conditions, which are composed of different loads within their own stable load range and the corresponding ultimate displacement of the column top, as training samples to train the second neural network model and obtain the performance curve prediction model.

[0057] Step 4: After the ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted, the performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the concrete structure to be predicted and the corresponding ultimate load.

[0058] Step 1 also includes the following sub-steps:

[0059] Step 11: Obtain the structural performance parameters of different concrete structures; the structural performance parameters include structural geometric parameters and material performance parameters.

[0060] In step 11, at least k specific concrete structures are obtained for each type of concrete structure, where k is a positive integer. These are classified according to their load-bearing systems, including frame structures, tube structures, frame-shear wall structures, etc. It should be noted that even concrete structures of the same type (e.g., all frame structures) may have different specific structures due to varying engineering requirements.

[0061] The structural geometric parameters and material performance parameters of different types of concrete structures are not entirely the same. Besides the beam and column cross-sectional dimensions, steel bar diameter, yield strength, ultimate strength, and ultimate strain that must be considered for each type of concrete structure, some types of concrete structures also require prestressing tendons. This directly leads to differences in material performance parameters: the dimensions, strength, and diameter of the prestressing tendons need to be considered; while concrete structures that do not require prestressing tendons do not need to obtain these material performance parameters at all. For example, some types of concrete structures require infill walls, in which case the required material performance parameters include the compressive strength and ultimate strength of the infill walls, which are not considered for concrete structures that do not require infill walls.

[0062] In this embodiment, the structural geometric parameters of the concrete structure include: column height, column width, beam height, beam width, beam reinforcement diameter, column reinforcement diameter, reinforcement elastic modulus, reinforcement yield strength, ultimate strain, concrete compressive strength, steel strand elastic modulus, steel strand prestress, steel strand yield strength, steel strand diameter, steel strand ultimate strain, steel strand initial position, infill wall compressive strength, and infill wall ultimate strain.

[0063] Step 12: Using finite element analysis, starting from the self-load of the current concrete structure, continuously increase the gravity load applied to the current concrete structure until the current concrete structure collapses, and obtain the gravity load bearing capacity of the current concrete structure.

[0064] The lower limit of the gravity load bearing capacity is the current self-load of the concrete structure, and the upper limit is the gravity load value before the gravity load value at the time of collapse.

[0065] Optionally, gravity loads can be applied uniformly to the beams and columns of the existing concrete structure.

[0066] The gravity load in this application must be applied to the columns, not just the beams, which is different from the application of gravity loads in the prior art.

[0067] Optionally, gravity loads can also be applied primarily to specific beams and columns.

[0068] For example, if the gravity load of the current concrete structure under normal use is P, and the self-load of the current concrete structure is 0.3P, then starting from 0.3P, the gravity load applied to the current concrete structure will increase in increments of 0.1P. When the gravity load is 3P, the current concrete structure collapses. At the time of collapse, the gravity load value is 3P, and the previous gravity load value is 2.9P. Therefore, the gravity load bearing range of the current concrete structure is [0.3P, 2.9P].

[0069] In real life, renovations often involve moving a lot of furniture into a building, or adding floors to a self-built house, all of which result in changes in the gravity load applied to the concrete structure. Even if there is nothing inside the house, the concrete structure still bears its own load.

[0070] Step 13: Within the range of gravity load bearing capacity, remove columns according to different column removal strategies under each gravity load value, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal; at the same time, determine the ultimate load of each concrete structure during the column removal process.

[0071] It should be noted that although the gravity load bearing capacity of the current concrete structure has been determined in step 12, the 2.9P gravity load applied to the current concrete structure can only ensure that the current concrete structure is stable under static conditions without external impact, but this does not fall under the category of "safe and stable" in the field of anti-collapse.

[0072] We apply external impact to the concrete structure in the form of column removal.

[0073] This application requires the determination of a new ultimate load within the gravity load bearing range, so that the current concrete structure can still return to a stable state after at least one least important column is removed, within the range of the current self-load and ultimate load of the concrete structure.

[0074] In actual column removal, gravity loads P and 2P are applied to the same concrete structure, and then the same column removal strategy is used under both gravity loads (for example, only the No. 1 middle column of the current concrete structure is removed). The concrete structure with gravity load P applied can recover to a stable state after the column is removed, while the concrete structure with gravity load 2P applied cannot recover to a stable state after the column is removed (it keeps oscillating).

[0075] Similarly, when applying a 2P gravity load to a concrete structure (the current gravity load tolerance range of the concrete structure is [0.3P, 2.9P]), different column removal strategies will directly affect whether the concrete structure can return to a stable state after column removal. For example, removing only the No. 1 corner column of the current concrete structure will cause the current concrete structure to collapse or vibrate continuously; however, removing only the less important No. 1 edge column of the current concrete structure will allow the current concrete structure to return to a stable state; even when applying a 2.8P gravity load, removing only the No. 1 edge column of the current concrete structure is necessary to prevent the current concrete structure from returning to a stable state, while when applying a 2.7P gravity load, removing only the No. 1 edge column of the current concrete structure will allow the current concrete structure to return to a stable state. Based on this situation, the ultimate load of the current concrete structure can be determined to be 2.7P. Within the gravity load range of (2.7P, 2.9P), the current concrete structure is still unsafe because any external impact will prevent it from returning to a stable state. Therefore, the range formed by the ultimate load and the upper limit of the gravity load bearing range is directly defined as the unsafe range. As long as the gravity load applied to the current concrete structure is within this range, it is unsafe, so there is no need to process the data within this range further.

[0076] In step 13, the column removal strategy will differ depending on the concrete structure.

[0077] For the same concrete structure, the different column removal strategies can be explained as follows:

[0078] ① Only draw one pillar: Even if the pillars are of the same type (e.g., they are all corner pillars), as long as they are not the same pillar, they belong to different pillar drawing strategies.

[0079] ② Draw K1 pillars at the same time, where K1 is a positive integer greater than 1: As long as the K1 pillars drawn this time are not exactly the same as the K1 pillars drawn at the same time in a previous time, they belong to different pillar drawing strategies.

[0080] ③ Only one column is removed at a time without replacement, until the current concrete structure can no longer be restored to a stable state. This is one round of column removal. Only when the total number of columns removed in this round is exactly the same as the total number of columns removed in a previous round, and the order of removal is the same, is it considered the same column removal strategy. Otherwise, different rounds are different column removal strategies.

[0081] The three situations mentioned above (① to ③) are different column removal strategies for the same concrete structure.

[0082] In step 13, if the current concrete structure cannot be stabilized after a column is removed, then any column top displacement in the current concrete structure cannot be used.

[0083] Column removal is a momentary impact action. After the column is removed, the current concrete structure will shake. If the concrete structure stops shaking after a period of time, it means that the current concrete structure can recover stability. If the current concrete structure continues to shake (i.e., the shaking displacement does not converge), it means that the current concrete structure cannot recover stability after this column removal.

[0084] In the same concrete structure, the top of the column that is removed each time is usually the point of greatest sway relative to other columns at the same time, and it is usually also the point of greatest displacement of the column top after it finally stabilizes. Therefore, when the displacement of the top of the removed column no longer changes, the displacement of the tops of other columns in the same concrete structure also no longer changes, and the concrete structure returns to a stable state. If a column can never return to a stable state, then the concrete structure in which that column is located will always be in an unstable state.

[0085] The process of "obtaining the maximum displacement of the column top when the concrete structure stabilizes after each column removal" also includes the following steps:

[0086] S1, obtain the column top sway displacement of the column being extracted this time at the first sampling time, the second sampling time and the third sampling time after the column extraction time; the first sampling time, the second sampling time and the third sampling time refer to the time intervals Δt, 2Δt and 3Δt from the column extraction time this time, respectively.

[0087] S2, if the column top swaying displacement of the same column decreases sequentially at the first, second, and third sampling times, then at the fourth sampling time after the third sampling time, the column top swaying displacement of the extracted column is obtained and S3 is executed. Otherwise, it is determined that the concrete structure cannot be stabilized after the column is extracted, and the concrete structure does not have the maximum column top displacement when it is stable after the column is extracted. The fourth sampling time refers to the time interval ΔT from the third sampling time.

[0088] S3 is the maximum displacement of the column top swaying among the columns removed this time, which is taken as the maximum displacement of the column top when the concrete structure is stable after the column is removed.

[0089] The value of Δt is usually small, while the value of ΔT is larger than Δt. In this embodiment, Δt = 0.1s and ΔT = 100s. ΔT is generally determined by technicians based on the longest stabilization time of the concrete structure after being subjected to impact.

[0090] As can be seen from S1 to S3 of this application, the trend of the column top swaying displacement in a short period of time after the column is removed is used to determine whether the concrete structure can be restored to a stable state after the column is removed. If the column top swaying displacement decreases continuously in a short period of time after the column is removed, it means that the concrete structure can be restored to a stable state after the column is removed, and this application can spend ΔT time to wait for it to be restored to a stable state; otherwise, it means that the concrete structure cannot be restored to a stable state after the column is removed, and there is no need to spend time waiting for an unstable concrete structure.

[0091] Unstable concrete structures are unsafe, and the displacement of any column top in an unstable concrete structure is constantly changing. Therefore, data on unstable concrete structures are not valuable for subsequent use.

[0092] This approach ensures the quality of the collected data while significantly reducing the time spent on initial data collection and processing, thereby improving the efficiency of constructing subsequent performance curves.

[0093] Step 2 also includes the following:

[0094] The structural performance parameters of different concrete structures are bound to the corresponding ultimate loads to form training samples in the first training set, where the ultimate load is the sample label. The first neural network model is trained using the training samples in the first training set to obtain the ultimate load prediction model. The ultimate load prediction model predicts the ultimate load of the corresponding concrete structure based on the structural performance parameters.

[0095] The ultimate load in step 3 can be obtained in step 1 or obtained using the ultimate load prediction model in step 2.

[0096] Step 3 also includes the following sub-steps:

[0097] Step 31: The load range consisting of the self-load and ultimate load of the current concrete structure is denoted as the stable load range of the current concrete structure.

[0098] Step 32: Based on the stable load range, set the load increment △a of the current concrete structure within the stable load range.

[0099] In step 32, for example, if the current stable load range of the concrete structure is [0.3P, 2.7P], and if the technician wants the gravity load of the current concrete structure to eventually reach 2.7P from 0.3P through 24 increments, then the load increment Δa of the current concrete structure can be set to 0.1P.

[0100] Step 33: Starting from the current self-load of the concrete structure, increment the gravity load value after each increment, as well as the corresponding self-load and ultimate load, as the load point for removing columns from the current concrete structure.

[0101] Step 34: At each load point, remove columns from the current concrete structure according to different column removal strategies, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal.

[0102] Step 35: Take the maximum value of the column top displacement obtained by the current concrete structure at the same load point among all column removal strategies, and take it as the ultimate displacement of the column top of the current concrete structure at the corresponding load point.

[0103] Step 36: Using each load point of the current concrete structure as the vertical axis and the corresponding column top ultimate displacement as the horizontal axis, fit the performance curve of the current concrete structure.

[0104] Step 37: The structural performance parameters of several concrete structures and the corresponding performance curves within the stable load range are used as training samples of the second training set. The second neural network model is trained to obtain the performance curve prediction model. The performance curve prediction model predicts the performance curve of the concrete structure within the corresponding stable load range based on the structural performance parameters of the concrete structure and the stable load range.

[0105] For a concrete structure, the ultimate displacement of the column top at a certain load point represents the maximum displacement of the column top at that load point under the corresponding gravity load. This displacement is determined by various column removal strategies, and if the concrete structure can return to a stable state after column removal, we retain the maximum displacement of the column top at that stable state. We then take the maximum value among these maximum displacements as the ultimate displacement of the column top at that load point. If we are informed (or obtain the result through pre-fitting using finite element analysis) that the maximum displacement of the column top after column removal at that load point exceeds the corresponding ultimate displacement, it indicates that the concrete structure may not converge after column removal (i.e., it may continue to shake), or even that the concrete structure has collapsed.

[0106] Based on the above explanation, the points on the performance curve correspond to the ultimate displacement of the column top of the current concrete structure under a certain gravity load. The column at which this ultimate displacement is located is the most likely place for the current concrete structure to collapse. The premise for obtaining this ultimate displacement is that the concrete structure can return to a stable state after the column is removed. Therefore, the concrete structure at the ultimate displacement of the column top on the performance curve will not collapse, nor will the concrete structure at the maximum displacement of the column top below the performance curve collapse. However, this is not necessarily the case for the area to the right of the performance curve. Therefore, the concrete structure performance curve obtained in this application is also the ultimate performance curve of the concrete structure.

[0107] Steps 1 to 3 are the training process for the ultimate load prediction model and the performance curve prediction model; step 4 is the formal use of these two models to predict the performance curve of the concrete structure to be predicted.

[0108] Step 4 includes the following:

[0109] Step 41: The ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted based on the structural performance parameters of the concrete structure to be predicted.

[0110] Step 42: Obtain the stable load range of the concrete structure to be predicted based on the ultimate load of the concrete structure to be predicted.

[0111] Step 43: The performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the structural performance parameters and stable load range of the concrete structure to be predicted.

[0112] like Figure 2 As shown, this diagram illustrates the difference between the ultimate stability performance curve of this application and the simulation results. The performance curve obtained using the performance curve acquisition method of this application is a completely new performance curve of the concrete structure to be predicted, represented by the red dashed line. Using the finite element analysis method, by continuously changing the gravity load on this new concrete structure to be predicted and by using various column removal strategies, the fitted performance curve is the simulation result curve, represented by the blue curve. It can be seen that the two curves basically overlap, indicating that the performance curve of the concrete structure to be predicted obtained by the performance curve acquisition method of this application has extremely high accuracy and high robustness.

[0113] The method for obtaining the ultimate stability performance curve in this application efficiently and accurately predicts the ultimate load of a concrete structure by training an ultimate load prediction model; then, using the trained performance curve prediction model, it efficiently and accurately outputs the performance curve of the concrete structure to be predicted based on the concrete structure and the corresponding ultimate load. Compared to obtaining the performance curve entirely using traditional finite element calculations, this method improves efficiency by approximately 70 times.

[0114] In the method for obtaining the limit stability performance curve in this application, a significant amount of computing power is saved and efficiency is improved during the data sample acquisition and model training process before S4:

[0115] ① In the process of obtaining the maximum displacement of the column top when the concrete structure stabilizes after each column removal, the trend of the column top swaying displacement within a very short time interval after the column removal is used to determine whether the concrete structure after the current column removal has the ability to recover to a stable state. Only if it has the ability to recover to a stable state will this application spend △T time to wait for it to recover to a stable state before collecting the corresponding maximum displacement of the column top. Otherwise, this application will not spend too much time blindly waiting after the column removal.

[0116] ②If the ultimate load of the training samples used when training the second neural network model is obtained through the ultimate load prediction model, then the order of training the ultimate load prediction model first and then training the performance curve prediction model in this application not only improves the training efficiency of the performance curve prediction model, but also shortens the overall time spent obtaining the ultimate stability performance curve.

[0117] Current technologies for obtaining performance curves of concrete structures require finite element analysis based on a physical model to obtain a specific performance curve for a concrete structure, which presents the following problems:

[0118] ① Because reaching the ultimate limit state typically requires extensive and computationally intensive nonlinear dynamic time-history analysis, and due to computational limitations, existing ultimate stability performance curves are often used to represent the ultimate stability performance curves of concrete structures of the same type. However, even within the same type of concrete structure, specific structures can differ. For example, if two concrete structures of the same type, such as concrete structure A, have two more decorative columns than concrete structure B, these two decorative columns, which were originally not primary load-bearing, may become crucial in restoring the concrete structure to a stable state after an impact (or column removal). Therefore, this "representation" inherently suffers from low accuracy and poor flexibility. Thus, existing technologies not only require significant time and computational resources to obtain the ultimate stability performance curve of a specific concrete structure, but also lack flexibility, failing to provide a fast and timely solution.

[0119] ② Furthermore, concrete structures involve a wide variety of parameters, including component dimensions, reinforcement types, material strength, and prestress levels. These parameters have obvious coupling relationships and have a complex impact on the overall structural performance and collapse resistance. If the concrete structure is complex (i.e., the data dimension is high), it will further prolong the time required to obtain the ultimate stability performance curve of a specific concrete structure using existing technologies, and the demand for computing power will also increase. The computational efficiency in high-dimensional parameter spaces is also difficult to meet the needs of rapid engineering evaluation.

[0120] ③ If existing technologies are used to directly predict structural response to obtain the ultimate stability performance curve, but the coupling relationship between parameters is not considered, and some non-critical data are reduced when the concrete structure is complex (i.e., data dimensionality is reduced), it will be difficult to fully reflect the physical and mechanical properties of the concrete structure (for example, for different concrete structural systems, the same displacement level may correspond to completely different damage states or safety margins. Therefore, relying solely on the single indicator of structural displacement makes it difficult to accurately determine the current state of the concrete structure relative to its ultimate stability, and also makes it difficult to fully reflect the impact of subsequent impacts on the ultimate stability of the current concrete structure). The prediction accuracy will be even lower.

[0121] The performance curve acquisition method of this application introduces finite element analysis only in the early training sample acquisition stage. The obtained physical characteristic parameters such as the ultimate displacement of the concrete structure make the trained model subject to the constraints of the structural physical mechanism on the basis of data driving, thereby effectively reducing the prediction error caused by multi-parameter coupling and improving the accuracy and robustness of the prediction of the ultimate stability performance curve. In formal use, the trained ultimate load prediction model and performance curve prediction model can be used to realize accurate and efficient prediction of the ultimate stability performance curve of various new and specific concrete structures in one stop. It not only has extremely high prediction flexibility and applicability, but also avoids a lot of nonlinear dynamic time history analysis calculations in existing technologies, thus greatly improving prediction efficiency.

[0122] The ultimate stability performance curve obtained by the performance curve acquisition method of this application is accurate while prioritizing safety. Therefore, the safety factor for various applications using the ultimate stability performance curve of this application is extremely high. This is because the training samples used in the training process for obtaining the ultimate load prediction model and the performance curve prediction model have already excluded data that could collapse or fail to recover to a stable state. Furthermore, the acquisition of the ultimate load will predict that the performance curve of the concrete structure will terminate at the ultimate load point, ensuring that the concrete structure can recover to a stable state even after at least one column is removed. This further enhances the safety factor in the left region of the ultimate stability performance curve and the application value of the ultimate stability performance curve.

[0123] This application also provides a first application of the performance curve of concrete structures, namely a method for safety assessment of a construction scheme, including steps 01 to 02:

[0124] Step 01: According to the construction plan, obtain the maximum column top displacement fitted by the finite element analysis of the concrete structure to be constructed; at the same time, use the method described above for obtaining the limit stability performance curve of a concrete structure to obtain the performance curve of the concrete structure to be constructed.

[0125] Step 02: If the maximum column top displacement is to the left of the performance curve of the concrete structure to be constructed, the current construction plan is deemed safe; otherwise, the current construction plan is deemed unsafe and the construction plan is adjusted.

[0126] In the performance curve of a concrete structure, the ultimate displacement at the top of the column increases with the increase of gravity load. In the first application of this application, the maximum displacement at the top of the column after construction is compared with the performance curve of the concrete structure to be constructed. The performance curve is used as the ultimate performance of the concrete structure to be constructed to evaluate the safety of the construction plan. Once the maximum displacement at the top of the column is to the right of the performance curve of the concrete structure under construction, it means that the concrete structure obtained after construction according to the current construction plan will have a great hidden danger to its safety.

[0127] For example, if renovations are underway downstairs and the construction plan involves removing a column, but our calculated maximum column top displacement indicates this will cause the balcony floor upstairs to tilt slightly downwards. After this construction, is the concrete structure upstairs still safe? If the maximum column top displacement is to the left of the performance curve, it means that although there is this slight tilt, the overall concrete structure upstairs is still safe. After the upstairs and downstairs parties agree on compensation, the downstairs party can proceed with the construction plan, and the upstairs party can still live normally. Conversely, if the maximum column top displacement is on or to the right of the performance curve, it means the overall concrete structure upstairs is no longer safe. Once the construction plan is carried out, regardless of the compensation offered by the downstairs party, the upstairs residents must be evacuated as quickly as possible.

[0128] This application also provides a second application to the performance curves of concrete structures, namely, a method for assessing the safety of damaged concrete structures, comprising the following steps:

[0129] Step 01': Obtain the initial concrete structure before damage. Use the method described above for obtaining the ultimate stability performance curve of a concrete structure to obtain the performance curve of the initial concrete structure.

[0130] Step 02': Obtain the maximum column top displacement of the damaged concrete structure relative to the initial concrete structure. If the maximum column top displacement is to the left of the performance curve of the initial concrete structure, the current damaged concrete structure is determined to be safe; otherwise, the current damaged concrete structure is determined to be dangerous.

[0131] The second application of this application is the safety assessment of damaged concrete structures, which is crucial for the reuse of damaged concrete structures, as it relates to whether damaged buildings can be directly re-inhabited.

[0132] This application also provides a system for obtaining the ultimate stability performance curve of a concrete structure, comprising:

[0133] First data module, first training module, second training module, performance curve prediction module;

[0134] The first data module is used to perform column extraction on different concrete structures within the gravity load bearing range using finite element analysis, and to obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads, which are then sent to the first training module and the second training module respectively.

[0135] The first training module uses the ultimate loads of different concrete structures as training samples to train the first neural network model and obtain the ultimate load prediction model, which is then sent to the performance curve prediction module.

[0136] The second training module determines the range of stable loads based on the ultimate load; then, it uses the performance curves formed by different loads and corresponding ultimate displacements at the top of the column under different column removal conditions for different concrete structures as training samples to train the second neural network model and obtain the performance curve prediction model, which is then fed into the performance curve prediction module.

[0137] The performance curve prediction module outputs the corresponding performance curve based on the concrete structure to be predicted.

[0138] Each module is programmed or configured to perform the steps of a method for obtaining the ultimate stability performance curve of a concrete structure as described above.

[0139] This application also provides a computer-readable storage medium storing a computer program programmed or configured to perform a method for obtaining the ultimate stability performance curve of a concrete structure as described above.

[0140] This application also provides a computer program product, including a computer program / instructions, which are executed by a processor to implement the steps of the method for obtaining the ultimate stability performance curve of a concrete structure as described above.

[0141] The technologies, shapes, and structures not described in detail in this application are all well-known technologies. It should also be noted that the above are merely preferred embodiments of this application and are not intended to limit the scope of this application. The components or steps in the embodiments of this application can be decomposed and / or recombined, and these decompositions and / or recombinations should be considered as equivalent solutions of this application and should all fall within the protection scope of this application.

Claims

1. A method for obtaining the ultimate stability performance curve of a concrete structure, characterized in that, Includes the following steps: Step 1: Use finite element analysis to remove columns from different concrete structures within the gravity load bearing range to obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads. The ultimate load refers to the maximum load that the concrete structure can still recover to a stable state after a column is removed. Step 2: Using the ultimate loads of different concrete structures as training samples, the first neural network model is trained to obtain the ultimate load prediction model. Step 3: Based on the ultimate load, determine the range of stable load; then, use the performance curves of different concrete structures under different column removal conditions, which are composed of different loads within their own range of stable load and the corresponding ultimate displacement of the column top, as training samples to train the second neural network model and obtain the performance curve prediction model. Step 4: After the ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted, the performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the concrete structure to be predicted and the corresponding ultimate load. Step 1 also includes the following sub-steps: Step 11: Obtain the structural performance parameters of different concrete structures; Structural performance parameters include structural geometric parameters and material performance parameters; Step 12: Using finite element analysis, starting from the self-load of the current concrete structure, continuously increase the gravity load applied to the current concrete structure until the current concrete structure collapses, and obtain the gravity load bearing range of the current concrete structure. Step 13: Within the range of gravity load bearing capacity, remove columns according to different column removal strategies under each gravity load value, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal; at the same time, determine the ultimate load of each concrete structure during the column removal process. Step 13, obtaining the maximum displacement of the column top when the concrete structure stabilizes after each column removal, also includes the following sub-steps: S1, obtain the column top sway displacement of the column being extracted this time at the first sampling time, the second sampling time and the third sampling time after the column extraction time; the first sampling time, the second sampling time and the third sampling time refer to the time intervals Δt, 2Δt and 3Δt from the column extraction time this time, respectively. S2, if the column top swaying displacement of the same column decreases sequentially at the first, second, and third sampling times, then at the fourth sampling time after the third sampling time, the column top swaying displacement of the extracted column is obtained and S3 is executed. Otherwise, it is determined that the concrete structure cannot be stable after the column is extracted, and the concrete structure does not have the maximum column top displacement when it is stable after the column is extracted. The fourth sampling time refers to the time interval ΔT from the third sampling time. S3 is the maximum displacement of the column top swaying among the columns removed this time, which is taken as the maximum displacement of the column top when the concrete structure is stable after the column is removed.

2. The method for obtaining the ultimate stability performance curve of a concrete structure according to claim 1, characterized in that, Step 2 also includes the following: binding the structural performance parameters of different concrete structures with the corresponding ultimate loads to training samples of the first training set, where the ultimate load is the sample label; training the first neural network model using the training samples of the first training set to obtain the ultimate load prediction model; the ultimate load prediction model predicts the ultimate load of the corresponding concrete structure based on the structural performance parameters.

3. The method for obtaining the ultimate stability performance curve of a concrete structure according to any one of claims 1 or 2, characterized in that, The ultimate load in step 3 is obtained from the finite element analysis in step 1, or from the ultimate load prediction model in step 2.

4. The method for obtaining the ultimate stability performance curve of a concrete structure according to claim 3, characterized in that, Step 3 also includes the following sub-steps: Step 31: The load range formed by the self-load and ultimate load of the current concrete structure is denoted as the stable load range of the current concrete structure. Step 32: Based on the stable load range, set the load increment Δa of the current concrete structure within the stable load range; Step 33: Starting from the self-load of the current concrete structure, increment the gravity load value after each increment, as well as the corresponding self-load and ultimate load, as the load point for removing columns in the current concrete structure. Step 34: At each load point, remove columns from the current concrete structure according to different column removal strategies, and obtain the maximum displacement of the column top when the concrete structure is stable after each column removal. Step 35: Take the maximum value of the maximum displacement of the column top obtained by the current concrete structure at the same load point among all column removal strategies, and take it as the ultimate displacement of the column top of the current concrete structure at the corresponding load point. Step 36: Using each load point of the current concrete structure as the vertical axis and the corresponding column top ultimate displacement as the horizontal axis, fit the performance curve of the current concrete structure. Step 37: The structural performance parameters of several concrete structures and the corresponding performance curves within the stable load range are used as training samples of the second training set. The second neural network model is trained to obtain the performance curve prediction model. The performance curve prediction model predicts the performance curve of the concrete structure within the corresponding stable load range based on the structural performance parameters of the concrete structure and the stable load range.

5. The method for obtaining the ultimate stability performance curve of a concrete structure according to claim 1, characterized in that, Step 4 also includes the following sub-steps: Step 41: The ultimate load prediction model predicts the ultimate load of the concrete structure to be predicted based on the structural performance parameters of the concrete structure to be predicted. Step 42: Obtain the stability load range of the concrete structure to be predicted based on the ultimate load of the concrete structure to be predicted. Step 43: The performance curve prediction model outputs the performance curve of the concrete structure to be predicted based on the structural performance parameters and stable load range of the concrete structure to be predicted.

6. A method for safety assessment of a construction scheme, characterized in that, Includes the following steps: Step 01: According to the construction plan, obtain the maximum column top displacement fitted by the finite element analysis of the concrete structure to be constructed; at the same time, use the method for obtaining the ultimate stability performance curve of the concrete structure as described in any one of claims 1-5 to obtain the performance curve of the current concrete structure to be constructed. Step 02: If the maximum column top displacement is to the left of the performance curve of the concrete structure to be constructed, the current construction plan is deemed safe; otherwise, the current construction plan is deemed unsafe and the construction plan is adjusted.

7. A method for safety assessment of damaged concrete structures, characterized in that, Includes the following steps: Step 01': Obtain the initial concrete structure before damage by using the method for obtaining the ultimate stability performance curve of a concrete structure as described in any one of claims 1-5 to obtain the performance curve of the initial concrete structure. Step 02': Obtain the maximum column top displacement of the damaged concrete structure relative to the initial concrete structure. If the maximum column top displacement is to the left of the performance curve of the initial concrete structure, the current damaged concrete structure is determined to be safe; otherwise, the current damaged concrete structure is determined to be dangerous.

8. A system for obtaining the ultimate stability performance curve of a concrete structure, characterized in that, include: First data module, first training module, second training module, performance curve prediction module; The first data module is used to perform column extraction on different concrete structures within the gravity load bearing range using finite element analysis, and to obtain the maximum displacement of the column top and the ultimate load of each concrete structure under different gravity loads, which are then sent to the first training module and the second training module respectively. The first training module uses the ultimate loads of different concrete structures as training samples to train the first neural network model and obtain the ultimate load prediction model, which is then sent to the performance curve prediction module. The second training module determines the stable load range based on the ultimate load; then, it uses the performance curves formed by different loads and corresponding column top ultimate displacements within the stable load range of different concrete structures under different column removal conditions as training samples to train the second neural network model and obtain a performance curve prediction model. The performance curve prediction model is then fed into the performance curve prediction module; the performance curve prediction module outputs the corresponding performance curve according to the concrete structure to be predicted; each module is programmed or configured to execute the steps of the method for obtaining the ultimate stability performance curve of a concrete structure as described in any one of claims 1-5.

Images