Method for predicting static behavior of post-fire bridge by using FDS-Kriging model

Abstract

The invention discloses a method for predicting static behaviors of a bridge after a fire by using an FDS-Kriging model. The method specifically comprises the following steps: S1, determining main parameters influencing fire simulation; s2, performing fire simulation according to the calculated parameter values, and judging whether the flame height is similar to the 3D area or not; s3, determining a heat source heat release rate curve; s4, determining the amplitude modulation range of the bridge temperature field distribution, the correction period and the interval; determining to-be-corrected parameters corrected by the bridge finite element model; s5, performing sensitivity analysis on the to-be-corrected parameters obtained in the step S4, designing and constructing a training sample after analysis, establishing a Kriging model, and performing regression prediction on the measured value through the Kriging model to obtain a predicted value; s6, correcting the finite element structure analysis model by using the predicted value obtained in the step S5; and S7, verifying a correction effect generated due to the correction index through a bridge static load test.

Description

Technical Field

[0001] The invention belongs to the technical field of bridge damage detection, and in particular relates to a method for predicting the static behavior of a bridge after a fire by using an FDS-Kriging model. Background Art

[0002] With the rapid development of computer performance and the maturity of finite element theory, finite element analysis has become an effective means to solve the mechanical performance and state evaluation of bridge structures, wind and earthquake resistance and fire simulation, etc. A reasonable and accurate finite element model is the basis for bridge structure analysis and state evaluation. However, due to the complex structure of the actual bridge and the uncertainty of related parameters, there is often a certain error between the actual bridge and the initial finite element model of the bridge. The initial finite element model cannot well reflect the actual boundary conditions and material parameters of the structure, and cannot accurately characterize the actual mechanical health state of the bridge structure. In addition, after a fire occurs in a bridge, the structural performance and material properties of the bridge often degrade to varying degrees. Therefore, it is necessary to modify the original finite element model of the bridge, identify and optimize the uncertain parameters of the bridge structure state, so as to correctly evaluate the actual bearing performance of the bridge after the fire.

[0003] At present, the modification of finite element models is mostly limited to bridges in normal operation, and less consideration is given to bridges that have suffered extreme or sudden accidents. For example, when a fire occurs, the degradation of the material properties and structural properties of the bridge at the local location of the fire source is far greater than that of other parts of the bridge. There are also different degrees of differences in the local areas of the same component of the bridge, that is, there may also be differences between the parameters of the local components. Therefore, it is unreasonable to directly use the finite element modification method of the overall component parameters of the bridge to modify the numerical model of the bridge after the fire. Summary of the invention

[0004] In order to solve the above-mentioned defects of the existing finite element model correction method, the present invention provides a method for predicting the static behavior of a bridge after a fire using the FDS-Kriging model, so as to correctly evaluate the actual bearing capacity of the bridge after a fire.

[0005] The technical solution adopted by the present invention is:

[0006] A method for predicting the static behavior of a bridge after a fire using an FDS-Kriging model, characterized in that it specifically comprises the following steps:

[0007] S1. Use the initial setting of the 3D area boundary of the FDS fire simulation, and calculate and determine the values ​​of airflow and dry airflow under standard conditions, the required air volume during combustion, and the maximum heat release rate of the fire source, so as to determine the main parameters affecting the fire simulation;

[0008] S2. Use the parameter values ​​calculated in step S1 to perform fire simulation and determine whether the flame height is similar to the 3D area:

[0009] If not similar, return to step S1 to continue correcting the parameter values;

[0010] If they are similar, the maximum heat release rate of the fire source is obtained;

[0011] S3, using the maximum heat release rate of the fire source finally obtained in step S2 to draw a heat source release curve, thereby determining a heat release rate curve of the heat source;

[0012] S4. Use FDS bridge fire temperature field simulation to determine the bridge temperature field distribution, correction cycle and interval modulation range; establish a bridge structure analysis model, determine the input value and response value of the natural frequency through bridge dynamic load test, and thus determine the parameters to be corrected for the bridge finite element model correction;

[0013] S5, performing sensitivity analysis on the parameters to be corrected obtained in step S4, designing and constructing training samples after the analysis, establishing a Kriging model, and performing regression prediction on the measured values ​​through the Kriging model to obtain predicted values;

[0014] S6, using the predicted value obtained in step S5 to correct the finite element structural analysis model;

[0015] S7. Verify the correction effect caused by the correction index through bridge static load test.

[0016] Furthermore, the fire was caused by the fall of a tanker truck, which led to the fire of the bridge superstructure.

[0017] Furthermore, the mathematical model of the heat release rate curve of the heat source selected in step S3 is:

[0018]

[0019] Where Q is the fire heat release rate; a is the fire growth coefficient; t is the fire occurrence time; t max is the time when the fire reaches the maximum heat release rate; t d The time to maintain the maximum heat release rate; Q max is the maximum heat release rate of fire.

[0020] Furthermore, in step S7, the root mean square error RMSE is used to evaluate the prediction results of the bridge static load test, and the formula is as follows:

[0021]

[0022] Among them, y i is the actual measured value on site. is the predicted value in this paper, N is the calculation sample size, RMSE represents the average deviation between the predicted value and the measured value, and the smaller the RMSE value, the more accurate the predicted value.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0024] 1. The method of the present invention combines FDS technology with the Kriging model, uses FDS technology to perform fire numerical simulation, and can divide the range of different damage degrees on the damaged bridge according to different temperature ranges, thereby determining the local beam section parameter selection range.

[0025] 2. The method of the present invention constructs a response surface model with high fitting accuracy for beam sections with different damage degrees to realize a correction method for the initial finite element model, which is used to correct the finite element model of the fire bridge, so that the corrected finite element model is closer to the actual state of the fire bridge.

[0026] 3. The method of the present invention combines the dynamic load test results of the bridge to predict the static behavior of the fire bridge, which has practical engineering significance for the evaluation of the bearing capacity and repair of the bridge after the fire. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Calculate the flow chart for the tanker fire model;

[0028] Figure 2 This is the flow chart of finite element correction of fire bridge based on FDS-Kriging model. DETAILED DESCRIPTION

[0029] The fire in this embodiment was caused by the tanker truck falling and catching fire, which led to the fire of the bridge superstructure.

[0030] Reference Figure 1 and Figure 2 The present invention provides a method for predicting the static behavior of a bridge after a fire using the FDS-Kriging model, which specifically comprises the following steps:

[0031] S1. Use the initial settings of the 3D area boundary of the FDS fire simulation, and calculate and determine the values ​​of airflow and dry airflow under standard conditions, the air volume required during combustion, and the maximum heat release rate of the fire source, so as to determine the main parameters affecting the fire simulation.

[0032] Among them, the main parameters include:

[0033] 1. Calculation area and grid size: determine the accuracy and computational cost of the simulation.

[0034] 2. Boundary conditions and material properties: affect the spread of fire and heat transfer on solid surfaces.

[0035] 3. Chemical reaction type: determines the transformation of chemical substances and the release of energy in the fire.

[0036] 4. Combustion surface parameters: such as power, heat release rate, etc., directly affect the intensity and spread speed of the fire.

[0037] 5. Air volume: affects the oxygen supply and smoke diffusion in the fire.

[0038] 6. Maximum heat release rate of fire source: determines the intensity and spread speed of fire and is a key parameter in fire simulation.

[0039] S2. Use the parameter values ​​calculated in step S1 to perform fire simulation and determine whether the flame height is similar to the 3D area:

[0040] If not similar, return to step S1 to continue correcting the parameter values;

[0041] If they are similar, the maximum heat release rate of the fire source is obtained;

[0042] Among them, the similarity conditions meet the following points: geometric similarity, physical process similarity, boundary condition similarity, initial condition similarity and meshing similarity.

[0043] S3, using the maximum heat release rate of the fire source finally obtained in step S2 to draw a heat source release curve, thereby determining a heat release rate curve of the heat source;

[0044] Specifically, the mathematical model of the heat release rate curve of the heat source selected in step S3 is:

[0045]

[0046] Where Q is the fire heat release rate; a is the fire growth coefficient; t is the fire occurrence time; t max is the time when the fire reaches the maximum heat release rate; t d The time to maintain the maximum heat release rate; Q max is the maximum heat release rate of fire.

[0047] S4. Use FDS bridge fire temperature field simulation to determine the bridge temperature field distribution, correction cycle and interval modulation range; establish a bridge structure analysis model, determine the input value and response value of the natural frequency through bridge dynamic load test, and thus determine the parameters to be corrected for the bridge finite element model correction;

[0048] S5, performing sensitivity analysis on the parameters to be corrected obtained in step S4, designing and constructing training samples after the analysis, establishing a Kriging model, and performing regression prediction on the measured values ​​through the Kriging model to obtain predicted values;

[0049] Among them, the Kriging model is a spatial interpolation method based on statistics. It considers the autocorrelation between spatial data, uses the linear combination of known data points to predict the value of unknown points, and gives the prediction error. It is suitable for data analysis and optimization problems in many fields such as geography, environment, and engineering.

[0050] S6, using the predicted value obtained in step S5 to correct the finite element structural analysis model;

[0051] S7. Verify the correction effect caused by the correction index through bridge static load test.

[0052] Specifically, in step S7, the root mean square error RMSE is used to evaluate the prediction results of the bridge static load test, and the formula is as follows:

[0053]

[0054] Among them, y i is the actual measured value on site. is the predicted value in this paper, N is the calculation sample size, RMSE represents the average deviation between the predicted value and the measured value, and the smaller the RMSE value, the more accurate the predicted value.

[0055] Finally, it should be noted that the above examples are only specific embodiments of the present invention. Obviously, the present invention is not limited to the above examples, and many variations are possible. All variations that can be directly derived or associated with the content disclosed by a person skilled in the art should be considered as the protection scope of the present invention.

Claims

1. A method for predicting the static behavior of a bridge after fire using the FDS-Kriging model, characterized in that: The specific steps include: S1. Use the initial setting of the 3D area boundary of the FDS fire simulation, and calculate and determine the values ​​of airflow and dry airflow under standard conditions, the required air volume during combustion, and the maximum heat release rate of the fire source, so as to determine the main parameters affecting the fire simulation; S2. Use the parameter values ​​calculated in step S1 to perform fire simulation and determine whether the flame height is similar to the 3D area: If not similar, return to step S1 to continue correcting the parameter values; If they are similar, the maximum heat release rate of the fire source is obtained; S3, using the maximum heat release rate of the fire source finally obtained in step S2 to draw a heat source release curve, thereby determining a heat release rate curve of the heat source; S4. Use FDS bridge fire temperature field simulation to determine the bridge temperature field distribution, correction cycle and interval modulation range; establish a bridge structure analysis model, determine the input value and response value of the natural frequency through bridge dynamic load test, and thus determine the parameters to be corrected for the bridge finite element model correction; S5, performing sensitivity analysis on the parameters to be corrected obtained in step S4, designing and constructing training samples after the analysis, establishing a Kriging model, and performing regression prediction on the measured values ​​through the Kriging model to obtain predicted values; S6, using the predicted value obtained in step S5 to correct the finite element structural analysis model; S7. Verify the correction effect caused by the correction index through bridge static load test.

2. The method for predicting the static behavior of a bridge after fire using the FDS-Kriging model as claimed in claim 1, characterized in that: The fire was caused by a tanker truck falling and catching fire, which led to the bridge superstructure catching fire.

3. The method for predicting the static behavior of a bridge after fire using the FDS-Kriging model as claimed in claim 1, characterized in that: The mathematical model of the heat release rate curve of the heat source selected in step S3 is: Where Q is the fire heat release rate; a is the fire growth coefficient; t is the fire occurrence time; t max is the time when the fire reaches the maximum heat release rate; t d The time to maintain the maximum heat release rate; Q max is the maximum heat release rate of fire.

4. The method for predicting the static behavior of a bridge after fire using the FDS-Kriging model as claimed in claim 1, characterized in that: In step S7, the root mean square error RMSE is used to evaluate the prediction results of the bridge static load test, and the formula is as follows: Among them, y i is the actual measured value on site. is the predicted value in this paper, N is the calculation sample size, RMSE represents the average deviation between the predicted value and the measured value, and the smaller the RMSE value, the more accurate the predicted value.

Images