Structural fire hybrid test method and system

Abstract

The application discloses a structural fire mixed test method and system, and relates to the field of structural fire resistance test. The method comprises the following steps: predicting the temperature of a physical substructure through a temperature field space-time distribution model; predicting the boundary displacement of the physical substructure by combining the temperature of the physical substructure and the boundary restoring force of the physical substructure at the previous moment through a structural fire reaction mixed simulation agent model; controlling the temperature of a thermal environment simulation system to reach the predicted temperature; controlling an actuator to make the physical substructure realize the predicted boundary displacement; collecting the boundary restoring force of the physical substructure and feeding back to the structural fire reaction mixed simulation agent model. The application predicts the temperature of each part of the physical substructure at different moments and the boundary displacement of the physical substructure by using the method of constructing or training a model. When the test is carried out, the development and change of the fire can be simulated, and the thermal-mechanical coupling analysis is not needed, so that the efficiency is improved, and the real-time performance of numerical calculation is ensured.

Description

Technical Field

[0001] This invention relates to the field of structural fire resistance testing, and in particular to a method and system for mixed structural fire testing. Background Technology

[0002] Fires and secondary fires following earthquakes severely impact the mechanical properties of structures, causing localized damage and even progressive collapse, resulting in enormous economic losses and casualties. Studying the overall response of structures under fire can help improve structural design theory and reduce or prevent major casualties.

[0003] There are usually two methods for simulating structural fire response: numerical simulation and physical testing. In terms of numerical simulation, the restoring force characteristics of the material need to be known in advance. Due to the strong nonlinearity that may occur during structural fire, the existing numerical simulation results are distorted and cannot truly reflect the disaster mechanism of the structure. In terms of physical testing, due to the limitations of fire test sites, equipment and funding, the existing research on the fire resistance performance of structures mainly focuses on individual components, simplified substructures or scaled-down specimens, and it is difficult to effectively consider the boundary constraint effect of non-fired parts on the fire-affected structure, and cannot consider the influence of the redistribution of internal forces in the structure and the real constraint of non-fired components on the fire-affected components. Therefore, combining numerical simulation with physical testing, i.e. fire-resistant hybrid testing, is an effective way to simulate the overall fire response of the structure. However, there are still limiting factors in the existing fire-resistant hybrid testing: (1) It is often assumed that only columns or part of the substructure are in the fire-affected state and the hybrid test is carried out with them as physical substructures, ignoring the spread of fire, i.e., the non-fire-affected parts of the structure are also in a high-temperature state, which seriously affects the mechanical properties of the components and the boundary conditions of the physical substructure. (2) When the numerical substructure considers the effect of temperature, a thermo-mechanical coupling analysis is required, which will significantly increase the calculation time and is not conducive to carrying out fire-resistant hybrid tests. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for mixed structural fire testing.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A method for mixed structural fire testing, the method comprising:

[0007] Step 1.1: Obtain the spatiotemporal distribution model of the temperature field of the physical structure; the spatiotemporal distribution model of the temperature field is used to predict the temperature of the physical substructure at the current moment; the physical structure includes several physical substructures;

[0008] Step 1.2: Obtain the structural fire response hybrid simulation proxy model; the structural fire response hybrid simulation proxy model is used to predict the boundary displacement of the physical substructure at the current moment based on the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment;

[0009] Step 1.3: Using the predicted temperature of the physical substructure at the current moment as the target temperature, control the thermal environment simulation system to make the temperature in the thermal environment simulation system reach the target temperature;

[0010] Step 1.4: Take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement;

[0011] Step 1.5: Collect the boundary restoring force of the physical substructure at the current moment, and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model in Step 1.2.

[0012] Optionally, the spatiotemporal distribution model of the temperature field is established through a machine learning algorithm; the input of the spatiotemporal distribution model of the temperature field includes the location of the fire source, the time history of the fire source temperature, and structural information, the structural information including the floor height, span, beam and column cross-sectional dimensions, and window location; the output of the spatiotemporal distribution model of the temperature field includes the temperature of the calculated points of the high-temperature part and the fire-affected part of the structure.

[0013] Optionally, before establishing the spatiotemporal distribution model of the temperature field, the following steps are also included:

[0014] Establish a Cartesian coordinate system for the structure and record the fire source position h0;

[0015] Fire dynamics simulation tools were used to simulate the development process of structural fires, record the time history of fire source temperature, and the time history of fire temperature at key locations in the structure.

[0016] Based on the temperature distribution at the fire site and the high-temperature mechanical properties of the structural materials, the structure is divided into a normal temperature section, a high-temperature section, and a fire-affected section.

[0017] Optionally, the steps for establishing the hybrid simulation proxy model of structural fire response specifically include:

[0018] Step 2.1: Obtain the influencing factors of the structural fire mixed test;

[0019] Step 2.2: Based on the aforementioned influencing factors, establish a spatiotemporal distribution model of the temperature field;

[0020] Step 2.3: Perform structural thermo-mechanical coupling analysis to obtain structural fire response and obtain a structural fire response database;

[0021] Step 2.4: Based on the aforementioned structural fire response database, iterative training is conducted using machine learning algorithms to obtain multiple surrogate models;

[0022] Step 2.5: Based on the multiple agent models, the final agent model is obtained by averaging the models.

[0023] Optionally, step 2.2 specifically includes:

[0024] Based on the influencing factors, a sample space is established; an experimental design method is used to sample the sample space; and a corresponding spatiotemporal distribution model of the temperature field is established based on the sample points of different influencing factors.

[0025] Optionally, the final agent model M in step 2.5 is represented as follows:

[0026] M = w1m1 + w2m2 + ... + w p m p ,

[0027] In the formula, M is the final model, m i and w i These are the i-th trained prediction model and its corresponding weights.

[0028] Optionally, the step of performing structural thermo-mechanical coupling analysis to obtain the structural fire response specifically includes:

[0029] Step 3.1: Based on the current temperature field of the structure, solve the governing equations to obtain the nodal displacements of the structure at the current moment;

[0030] Step 3.2: Based on the nodal displacements, perform linear static analysis on the elastic main structure to obtain the stress state of the structure at the current moment;

[0031] Step 3.3: Using the yield strength of the material as the criterion, if the stress of the structural member is greater than the yield strength, proceed to step 3.4; otherwise, proceed to step 3.5.

[0032] Step 3.4: Using a room-temperature refined numerical substructure model, perform static analysis directly based on the nodal displacements;

[0033] Step 3.5: Using the fire-affected and high-temperature numerical substructure models, and based on the current temperatures of the fire-affected and high-temperature numerical substructures and the nodal displacements, perform static analysis based on thermal stress analysis.

[0034] Step 3.6: Through the static analysis in Steps 3.4 and 3.5, the structural restoring force of the substructure at the current moment is obtained, and the nonlinear correction force of the substructure at the current moment is calculated based on the structural restoring force.

[0035] Optionally, the governing equations are as follows:

[0036]

[0037] In the formula, K dyn The dynamic stiffness is denoted by F; F is the equivalent load. It is a nonlinear correction force;

[0038] K dyn F, The calculation formulas are as follows:

[0039]

[0040]

[0041]

[0042] In the formula, γ and δ are the adjustment parameters of the Newmark-β method, K is the structural stiffness matrix, Δt is the integration step size, k is the time step, R() is the structural restoring force, and F G Let T be the gravity load on the structure, d be the temperature, d be the nodal displacement, M be the mass matrix, and C be the mass matrix of the structure, W. kin The determined structural damping coefficient matrix.

[0043] Optionally, the governing equations are obtained from the pseudo-dynamic equations using the Newmark-β method, and the specific pseudo-dynamic equations are as follows:

[0044]

[0045] In the formula, R() is the structural restoring force, and F G Let T be the gravity load on the structure, d be the temperature, d be the nodal displacement, M be the mass matrix, and C be the mass matrix of the structure, W. kin The determined structural damping coefficient matrix.

[0046] The present invention also provides a structural fire hybrid testing system, comprising:

[0047] The temperature prediction module is used to predict the temperature of the physical substructure at the current moment using a temperature field spatiotemporal distribution model.

[0048] The boundary displacement prediction module is used to predict the boundary displacement of the physical substructure at the current moment by combining the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment with the structural fire response hybrid simulation proxy model.

[0049] The temperature simulation module is used to control the thermal environment simulation system by taking the predicted temperature of the physical substructure at the current moment as the target temperature, so that the temperature in the thermal environment simulation system reaches the target temperature.

[0050] The displacement simulation module is used to take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement.

[0051] The acquisition module is used to acquire the boundary restoring force of the physical substructure at the current moment and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model.

[0052] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0053] This invention provides a method and system for hybrid structural fire testing. The method includes: predicting the temperature of a physical substructure at the current moment using a temperature field spatiotemporal distribution model; predicting the boundary displacement of the physical substructure at the current moment using a hybrid structural fire response simulation proxy model, combining the temperature of the physical substructure at the current moment with the boundary restoring force of the physical substructure at the previous moment; controlling a thermal environment simulation system to bring the temperature in the thermal environment simulation system to the predicted temperature; controlling an actuator to act on the physical substructure to achieve the predicted boundary displacement; and collecting the boundary restoring force of the physical substructure at the current moment and feeding it back to the hybrid structural fire response simulation proxy model. Compared to existing technologies that only consider the structural parts exposed to fire, neglecting the spread and development of the fire, and requiring thermo-mechanical coupling analysis when considering the temperature effect on the numerical substructure, which significantly increases computation time, this invention uses a spatiotemporal temperature field distribution model to predict the temperature of each part of the physical substructure at different times. This allows for the simulation of fire development and changes during the experiment, taking into account the boundary constraint effect of non-fire-exposed parts on the fire-exposed structure, and more accurately simulating the fire response of the structure. In addition, this invention uses a hybrid simulation proxy model of structural fire response to predict the boundary displacement of the physical substructure. During the experiment, thermo-mechanical coupling analysis is not required, avoiding cumbersome finite element calculations, improving efficiency, and ensuring the real-time performance of numerical calculations. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 A flowchart of a structural fire hybrid test provided in an embodiment of the present invention;

[0056] Figure 2Another structural fire hybrid test flowchart provided in this embodiment of the invention;

[0057] Figure 3 A flowchart illustrating the establishment of a temperature field spatiotemporal model provided in an embodiment of the present invention;

[0058] Figure 4 An example diagram illustrating the use of the LSTM method to construct a temperature spatiotemporal distribution model, provided in an embodiment of the present invention.

[0059] Figure 5 Flowchart for establishing the structural fire response prediction model provided in this embodiment of the invention;

[0060] Figure 6 A flowchart for structural fire response simulation provided in an embodiment of the present invention;

[0061] Figure 7 This is a structural block diagram of a fire mixing test system provided in an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Currently, there are still limiting factors in existing fire-resistant hybrid tests: (1) It is often assumed that only columns or parts of the substructure are in a fire-exposed state, and hybrid tests are carried out using these as physical substructures, ignoring the spread of fire. That is, the non-fire-exposed parts of the structure are also in a high-temperature state, which seriously affects the mechanical properties of the components and the boundary conditions of the physical substructure. (2) When the temperature effect is considered in the numerical substructure, thermo-mechanical coupling analysis needs to be carried out, which will seriously increase the calculation time and is not conducive to carrying out fire-resistant hybrid tests. (3) Existing fire-resistant hybrid tests are mainly based on static analysis, but after the fire occurs and develops, it will cause large nonlinearity in local components and generate dynamic effects.

[0064] The purpose of this invention is to provide a method and system for mixed structural fire testing.

[0065] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] See Figure 1 This invention provides a method for mixed structural fire testing, specifically including:

[0067] Step 1.1: Obtain the spatiotemporal distribution model of the temperature field of the physical structure; the spatiotemporal distribution model of the temperature field is used to predict the temperature of the physical substructure at the current moment; the physical structure includes several physical substructures;

[0068] Step 1.2: Obtain the structural fire response hybrid simulation proxy model; the structural fire response hybrid simulation proxy model is used to predict the boundary displacement of the physical substructure at the current moment based on the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment;

[0069] Step 1.3: Using the predicted temperature of the physical substructure at the current moment as the target temperature, control the thermal environment simulation system to make the temperature in the thermal environment simulation system reach the target temperature;

[0070] Step 1.4: Take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement;

[0071] Step 1.5: Collect the boundary restoring force of the physical substructure at the current moment, and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model in Step 1.2.

[0072] In fire tests, displacement is an important aspect of structural performance evaluation. This invention achieves the prediction of structural displacement through the above steps, and can obtain the mechanical performance degradation law of the structure under fire, providing a basis for establishing corresponding fire-resistant design methods.

[0073] See Figure 2 In some embodiments, the fire-resistant mixed test method may further include the following steps:

[0074] ① Determine the fire source temperature model according to the current "Code for Fire Protection Design of Buildings";

[0075] ②At t k At any given moment, the established spatiotemporal distribution model of the temperature field is used to predict the temperature field of the structure at the current moment, and then the prediction is sent to the corresponding substructure.

[0076] ③ Under the structural temperature field predicted in step ②, the physical substructure t is used. k-1 The boundary restoring force at any given time is used to predict the displacement response of the structure through the established structural fire response hybrid simulation proxy model, and the boundary displacement of the physical substructure is sent to the physical experimental device.

[0077] ④ In the physical experimental setup, thermal boundary conditions and displacement boundary conditions are implemented. For the thermal boundary conditions, the temperature of the physical substructure predicted in step ② is used as the target quantity. A suitable temperature field controller (such as a proportional-integral controller) is used to make the temperature in the thermal environment simulation system equal to the target temperature.

[0078] For displacement boundary conditions, taking the predicted physical substructure boundary displacement in step ③ as the target, a suitable boundary coordination strategy is adopted to realize the target displacement on the physical substructure through an actuator; thus, the boundary restoring force of the physical substructure is obtained.

[0079] ⑤ Feedback the boundary restoring forces of the physical substructures to the hybrid simulation proxy model of structural fire response;

[0080] ⑥ Repeat steps ② to ⑤ until the simulation ends.

[0081] See Figure 3 In some embodiments, the spatiotemporal distribution model of the temperature field is established through a machine learning algorithm; the input of the spatiotemporal distribution model of the temperature field includes the location of the fire source, the time history of the fire source temperature, and structural information, the structural information including the floor height, span, beam and column cross-sectional dimensions, and window location; the output of the spatiotemporal distribution model of the temperature field includes the calculated temperatures of the high-temperature part and the fire-affected part of the structure.

[0082] In some embodiments, prior to establishing the spatiotemporal distribution model of the temperature field, the method further includes:

[0083] Establish a Cartesian coordinate system for the structure and record the fire source position h0;

[0084] Fire dynamics simulation tools were used to simulate the development process of structural fires, record the time history of fire source temperature, and the time history of fire temperature at key locations in the structure.

[0085] Based on the temperature distribution at the fire site and the high-temperature mechanical properties of the structural materials, the structure is divided into a normal temperature section, a high-temperature section, and a fire-affected section.

[0086] In some embodiments, the process of establishing the spatiotemporal distribution model of the temperature field may further include:

[0087] 1. Establish a Cartesian coordinate system for the structure and record the fire source position h0.

[0088] 2. Based on the principles of fire dynamics, the Fire Dynamics Simulator (FDS) tool is used to obtain the development process of structural fire, record the time history of fire source temperature, and record the time history of fire temperature at key locations in the structure.

[0089] 3. Based on the temperature distribution at the fire scene and the high-temperature mechanical properties of the structural materials, the structure is divided into a normal temperature section, a high-temperature section, and a fire-affected section.

[0090] 4. Using the spatial-temporal distribution of structural fire field temperature obtained in step 2, and taking the fire source location, fire source temperature time history, and structural information as inputs, and the temperatures of calculation points in the high-temperature and fire-affected parts of the structure as outputs, establish a spatial-temporal distribution model of fire field temperature for the fire-affected and high-temperature parts. Machine learning algorithms such as LSTM, CNN, NARX, and support vector machines can be used, as well as Kriging models and multinomial models. An example of using the LSTM method to construct a spatial-temporal temperature distribution model is shown below. Figure 4 As shown.

[0091] See Figure 5 In some embodiments, the steps for establishing the structural fire response hybrid simulation proxy model specifically include:

[0092] Step 2.1: Obtain the influencing factors of the structural fire mixed test;

[0093] Step 2.2: Based on the aforementioned influencing factors, establish a spatiotemporal distribution model of the temperature field;

[0094] Step 2.3: Perform structural thermo-mechanical coupling analysis to obtain structural fire response and obtain a structural fire response database;

[0095] Step 2.4: Based on the aforementioned structural fire response database, iterative training is conducted using machine learning algorithms to obtain multiple surrogate models;

[0096] Step 2.5: Based on the multiple agent models, the final agent model is obtained by averaging the models.

[0097] In some embodiments, obtaining the influencing factors may specifically include:

[0098] Based on the actual situation, the uncertainties in physical experiments and numerical calculations during the implementation of the mixed experiment are analyzed, and the influencing factors of the structural response are obtained, such as the fire source temperature model, the fire source location, and the structural cross-sectional dimensions.

[0099] In some embodiments, step 2.2 specifically includes:

[0100] Based on the influencing factors, a sample space is established; an experimental design method is used to sample the sample space; and a corresponding spatiotemporal distribution model of the temperature field is established based on the sample points of different influencing factors.

[0101] In some embodiments, step 2.2 may further include:

[0102] Determine the levels of influencing factors to obtain the sample space of parameters;

[0103] Sampling was performed in the sample space using experimental design methods;

[0104] Different spatiotemporal distribution models of the temperature field were obtained for different sample points of influencing factors.

[0105] In some embodiments, step 2.4 may further include:

[0106] Different machine learning algorithms (such as LSTM, CNN, NARX, recurrent neural networks, etc.) are selected and iterative training is carried out to obtain multiple surrogate models m. i .

[0107] In some embodiments, the final proxy model M in step 2.5 is represented as follows:

[0108] M = w1m1 + w2m2 + ... + w p m p (1)

[0109] In the formula, M is the final model, m i and w i These are the i-th trained prediction model and its corresponding weights.

[0110] In some embodiments, the final surrogate model M for predicting the mixed experimental structure response in step 2.5 can also be represented as follows:

[0111] M = w1m1 + w2m2 + ... + w p m p =w T m, (2)

[0112] In the formula, M is the final model, m i and w i Let be the i-th trained prediction model and its corresponding weights, and w and m be the weight coefficient vector and model vector, respectively. The weight coefficient vector can then be obtained through optimization, specifically as follows:

[0113] w = min(yw) T m) T (yw T m), (3)

[0114] In the formula, y represents the measured structural fire response.

[0115] In some embodiments, the step of performing structural thermo-mechanical coupling analysis to obtain the structural fire response specifically includes:

[0116] Step 3.1: Based on the current temperature field of the structure, solve the governing equations to obtain the nodal displacements of the structure at the current moment;

[0117] Step 3.2: Based on the nodal displacements, perform linear static analysis on the elastic main structure to obtain the stress state of the structure at the current moment;

[0118] Step 3.3: Using the yield strength of the material as the criterion, if the stress of the structural member is greater than the yield strength, proceed to step 3.4; otherwise, proceed to step 3.5.

[0119] Step 3.4: Using a room-temperature refined numerical substructure model, perform static analysis directly based on the nodal displacements;

[0120] Step 3.5: Using the fire-affected and high-temperature numerical substructure models, and based on the current temperatures of the fire-affected and high-temperature numerical substructures and the nodal displacements, perform static analysis based on thermal stress analysis.

[0121] Step 3.6: Through the static analysis in Steps 3.4 and 3.5, the structural restoring force of the substructure at the current moment is obtained, and the nonlinear correction force of the substructure at the current moment is calculated based on the structural restoring force.

[0122] In some embodiments, the governing equations are specifically as follows:

[0123]

[0124] In the formula, K dyn The dynamic stiffness is denoted by F; F is the equivalent load. It is a nonlinear correction force;

[0125] K dyn F, The calculation formulas are as follows:

[0126]

[0127]

[0128]

[0129] In the formula, γ and δ are the adjustment parameters of the Newmark-β method, K is the structural stiffness matrix, Δt is the integration step size, k is the time step, R() is the structural restoring force, and F G Let T be the gravity load on the structure, d be the temperature, d be the nodal displacement, M be the mass matrix, and C be the mass matrix of the structure, W. kin The structural damping coefficient matrix is ​​determined by the following formula:

[0130]

[0131] In the formula, K is the structural stiffness matrix, ε is a very small positive number, η is the mass damping coefficient, and α and β are the mass and stiffness damping coefficients, respectively, which are calculated from the first two frequencies of the structure.

[0132] It should be noted that the governing equation shown in equation (4) is obtained by combining the static equilibrium of the structure with the dynamic relaxation method. The static-dynamic transformation during the fire occurrence and development process can be considered. That is, when the temperature of the physical substructure is lower than the critical temperature, it is equivalent to static analysis, avoiding iterative solution. As the temperature rises, after the dynamic effect is generated, the structural analysis is transformed from static to dynamic.

[0133] K in the governing equation (4) dyn The pseudo-dynamic stiffness and the equivalent load F are related to the selected step-by-step integration method. The nonlinear correction force is related to the boundary restoring force of the physical substructure and the nonlinear state of the numerical substructure. The above formulas (5), (6) and (7) are obtained by solving the pseudo-dynamic equations when the step-by-step integration method used is the Newmark-β method.

[0134] Furthermore, when a structure has not entered a nonlinear state, the formula (4) used to calculate the nodal displacements does not include a nonlinear correction force. Only after this structure enters a nonlinear state will the formula (4) used to calculate the nodal displacements have a nonlinear correction force.

[0135] In some embodiments, the governing equations are obtained from pseudo-dynamic equations using the Newmark-β method, and the pseudo-dynamic equations are as follows:

[0136]

[0137] In the formula, R() is the structural restoring force, and F G Let T be the gravity load on the structure, d be the temperature, d be the nodal displacement, M be the mass matrix, and C be the mass matrix of the structure, W. kin The determined structural damping coefficient matrix.

[0138] Structural restoring force includes boundary restoring force.

[0139] In some embodiments, the governing equations can also be obtained by solving the pseudo-dynamic equations using other stepwise integration methods (such as the central difference method).

[0140] See Figure 6 In some embodiments, the step of performing structural thermo-mechanical coupling analysis to obtain the structural fire response may further include:

[0141] 1. An elastic principal structure model is established in MATLAB to solve the overall structural response. A room-temperature refined numerical substructure model is established in OpenSees to analyze its static restoring force. Fire-exposed and high-temperature substructure models are established in ABAQUS to analyze their static restoring forces at high temperatures. Data exchange between different computing platforms is achieved through a TCP / IP-based socket communication mechanism. It should be noted that other computational analysis software can be used instead of MATLAB, OpenSees, and ABAQUS (Step 1 is an overview, which introduces which software is used and when it is used. For example, a room-temperature refined numerical substructure model is only established after the structure enters nonlinearity).

[0142] 2. Solve the control equations in MATLAB to obtain the nodal displacements of the overall structure. See the previous example for the specific process.

[0143] 3. Send the overall structural nodal displacements obtained from solving the governing equations to the elastic master structure in MATLAB to complete the linear static analysis.

[0144] 4. Using the material's yield strength as the criterion, determine if any new structural components have entered a nonlinear state. If the calculated stress of a component is greater than its yield strength, then a component has entered a nonlinear state. In this case, establish a room-temperature refined numerical substructure model in OpenSees and proceed to step 5; otherwise, if the calculated stress of a component is less than its yield strength, proceed to step 5.

[0145] 5. Substructure Analysis. Using both the fire source temperature model and the fire field temperature spatiotemporal distribution model, and taking the current time step and component position coordinates, the temperatures of the fire-affected numerical substructure and the high-temperature numerical substructure at the current moment are calculated. Thermal stress analysis is then performed in ABAQUS. The displacements corresponding to the boundary degrees of freedom of the physical substructure obtained from solving the governing equations are transmitted to the numerical model of the physical substructure in ABAQUS via a Socket communication mechanism. Based on the thermal stress analysis, static analysis of the components is then completed using nodal displacements. For the normal-temperature numerical substructure in OpenSees, static analysis is directly performed using nodal displacements.

[0146] 6. Read the structural restoring forces of different refined substructures in the calculation and analysis software ABAQUS and OpenSees, and transmit them to the coordinator MATLAB through the Socket communication mechanism to calculate the nonlinear correction forces of the substructures.

[0147] 7. Repeat steps 2 through 6 until the simulation ends.

[0148] The present invention also provides a structural fire hybrid testing system, specifically comprising:

[0149] The temperature prediction module is used to predict the temperature of the physical substructure at the current moment using a temperature field spatiotemporal distribution model.

[0150] The boundary displacement prediction module is used to predict the boundary displacement of the physical substructure at the current moment by combining the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment with the structural fire response hybrid simulation proxy model.

[0151] The temperature simulation module is used to control the thermal environment simulation system by taking the predicted temperature of the physical substructure at the current moment as the target temperature, so that the temperature in the thermal environment simulation system reaches the target temperature.

[0152] The displacement simulation module is used to take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement.

[0153] The acquisition module is used to acquire the boundary restoring force of the physical substructure at the current moment and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model.

[0154] In some embodiments, the acquisition module is also used to acquire the temperature in the thermal environment simulation system in order to monitor and control the temperature in the thermal environment simulation system.

[0155] See Figure 7 In some embodiments, the test system for conducting mixed structural fire tests may further include the following:

[0156] The hybrid testing system consists of a numerical calculation module, a coordination module, and a physics module. The numerical calculation module includes sub-modules for structural model parameters, temperature field spatiotemporal distribution, and structural fire response prediction, used to achieve structural response prediction based on a machine learning surrogate model. The coordination module includes sub-modules for communication, actuator control, and temperature field control, used to implement data interaction and test control algorithms. The physics module includes a thermal environment simulation system, a physical boundary simulation loading module, and a data acquisition sub-module, used for thermal and mechanical boundary simulation, and also includes data acquisition functionality.

[0157] In summary, the present invention has the following advantages:

[0158] By taking into account the temperature distribution of the structure in the fire field caused by fire spread, the fire response of the structure can be simulated more accurately.

[0159] The boundary conditions of the physical specimen are obtained through structural calculations, which can reveal the actual mechanical property degradation law of the physical specimen under fire.

[0160] The real-time hybrid test for fire resistance enables real-time interaction between the physical substructure and the numerical part, and can effectively take into account the influence of the strain rate of the material at high temperatures.

[0161] The deep learning-based surrogate model avoids the cumbersome finite element calculations, has high computational efficiency, and ensures the real-time performance of numerical calculations.

[0162] It can simulate the entire process of structural fires, especially considering the conversion from static to dynamic fires.

[0163] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0164] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for mixed structural fire testing, characterized in that, The method includes: Step 1.1: Obtain the spatiotemporal distribution model of the temperature field of the physical structure; the spatiotemporal distribution model of the temperature field is used to predict the temperature of the physical substructure at the current moment; the physical structure includes several physical substructures; Step 1.2: Obtain the structural fire response hybrid simulation proxy model; the structural fire response hybrid simulation proxy model is used to predict the boundary displacement of the physical substructure at the current moment based on the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment; Step 1.3: Using the predicted temperature of the physical substructure at the current moment as the target temperature, control the thermal environment simulation system to make the temperature in the thermal environment simulation system reach the target temperature; Step 1.4: Take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement; Step 1.5: Collect the boundary restoring force of the physical substructure at the current moment, and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model in Step 1.2; The steps for establishing the hybrid simulation proxy model for structural fire response specifically include: Step 2.1: Obtain the influencing factors of the structural fire mixed test; Step 2.2: Based on the aforementioned influencing factors, establish a spatiotemporal distribution model of the temperature field; Step 2.3: Perform structural thermo-mechanical coupling analysis to obtain structural fire response and obtain a structural fire response database; Step 2.4: Based on the aforementioned structural fire response database, iterative training is conducted using machine learning algorithms to obtain multiple surrogate models; Step 2.5: Based on the multiple agent models, the final agent model is obtained by averaging the models. The process of obtaining structural fire response through structural thermo-mechanical coupling analysis specifically includes: Step 3.1: Based on the temperature field of the structure at the current moment, solve the governing equations to obtain the nodal displacements of the structure at the current moment; Step 3.2: Based on the nodal displacements, perform linear static analysis on the elastic main structure to obtain the stress state of the structure at the current moment; Step 3.3: Using the yield strength of the material as the criterion, if the stress of the structural member is greater than the yield strength, proceed to step 3.4; otherwise, proceed to step 3.

5. Step 3.4: Using a room-temperature refined numerical substructure model, perform static analysis directly based on the nodal displacements; Step 3.5: Using the fire-affected and high-temperature numerical substructure models, and based on the current temperatures of the fire-affected and high-temperature numerical substructures and the nodal displacements, perform static analysis based on thermal stress analysis. Step 3.6: Through the static analysis in Steps 3.4 and 3.5, the structural restoring force of the substructure at the current moment is obtained, and the nonlinear correction force of the substructure at the current moment is calculated based on the structural restoring force.

2. The structural fire mixed test method according to claim 1, characterized in that, The spatiotemporal distribution model of the temperature field is established through machine learning algorithms; the inputs of the spatiotemporal distribution model of the temperature field include the location of the fire source, the time history of the fire source temperature, and structural information, including the floor height, span, beam and column cross-sectional dimensions, and window locations; the outputs of the spatiotemporal distribution model of the temperature field include the calculated temperatures of the high-temperature parts and the fire-affected parts of the structure.

3. The structural fire mixed test method according to claim 2, characterized in that, Before establishing the spatiotemporal distribution model of the temperature field, the following steps are also included: Establish a Cartesian coordinate system for the structure and record the fire source position h0; Fire dynamics simulation tools were used to simulate the development process of structural fires, record the time history of fire source temperature, and the time history of fire temperature at key locations in the structure. Based on the temperature distribution at the fire site and the high-temperature mechanical properties of the structural materials, the structure is divided into a normal temperature section, a high-temperature section, and a fire-affected section.

4. The structural fire mixed test method according to claim 1, characterized in that, Step 2.2 specifically includes: Based on the influencing factors, a sample space is established; an experimental design method is used to sample the sample space; and a corresponding spatiotemporal distribution model of the temperature field is established based on the sample points of different influencing factors.

5. The structural fire mixed test method according to claim 1, characterized in that, The final proxy model M in step 2.5 is represented as follows: ， In the formula, M This is the final model. m i and w i They are the first i A trained prediction model and its corresponding weights.

6. The structural fire mixed test method according to claim 1, characterized in that, The governing equations are as follows: ， In the formula, K dyn The dynamic stiffness is denoted by F; F is the equivalent load. It is a nonlinear correction force; K dyn F, The calculation formulas are as follows: ， ， ， In the formula, and Here, K is the adjustment parameter for the Newmark-β method, and K is the structural stiffness matrix. For the integration step size, k Let R be the time step, R() be the structural restoring force, and F be the time step. G For the structure's gravity load, T Let d be the temperature, d be the nodal displacement of the structure, M be the structural mass matrix, and C be the structural kinetic energy. W kin The determined structural damping coefficient matrix.

7. The structural fire mixed test method according to claim 6, characterized in that, The governing equations are obtained from the pseudo-dynamic equations using the Newmark-β method. The specific pseudo-dynamic equations are as follows: ， In the formula, R() is the structural restoring force, and F G For the structure's gravity load, T Let d be the temperature, d be the nodal displacement of the structure, M be the structural mass matrix, and C be the structural kinetic energy. W kin The determined structural damping coefficient matrix.

8. A structural fire mixed testing system for implementing the structural fire mixed testing method of claim 1, characterized in that, include: The temperature prediction module is used to predict the temperature of the physical substructure at the current moment using a temperature field spatiotemporal distribution model. The boundary displacement prediction module is used to predict the boundary displacement of the physical substructure at the current moment by combining the temperature of the physical substructure at the current moment and the boundary restoring force of the physical substructure at the previous moment with the structural fire response hybrid simulation proxy model. The temperature simulation module is used to control the thermal environment simulation system by taking the predicted temperature of the physical substructure at the current moment as the target temperature, so that the temperature in the thermal environment simulation system reaches the target temperature. The displacement simulation module is used to take the predicted boundary displacement of the physical substructure at the current moment as the target displacement, and control the actuator to act on the physical substructure so that the physical substructure achieves the target displacement. The acquisition module is used to acquire the boundary restoring force of the physical substructure at the current moment and feed the boundary restoring force of the physical substructure at the current moment back to the structural fire response hybrid simulation proxy model.

Images