CN114861458A - Quick evaluation method for bridge uplink vehicle safety based on mapping relation agent model - Google Patents
Quick evaluation method for bridge uplink vehicle safety based on mapping relation agent model Download PDFInfo
- Publication number
- CN114861458A CN114861458A CN202210590453.9A CN202210590453A CN114861458A CN 114861458 A CN114861458 A CN 114861458A CN 202210590453 A CN202210590453 A CN 202210590453A CN 114861458 A CN114861458 A CN 114861458A
- Authority
- CN
- China
- Prior art keywords
- mapping relation
- bridge
- agent model
- model
- optimization
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention discloses a method for quickly evaluating the safety of an on-bridge vehicle based on a mapping relation agent model, which comprises the following steps of 1, recording the service performance evolution of a basic structure; step 2, constructing a multipoint mapping relation required by the agent model; step 3, constructing a mapping relation agent model; and 4, evaluating the driving safety and comfort based on the agent model. The invention discloses a multi-target parallel optimization new algorithm capable of simultaneously considering system global search and local development and combined with a vehicle-rail-bridge coupling vibration model, a mapping relation agent model for efficiently calculating driving safety parameters under a complex service condition is built, and a bridge driving safety evaluation device is formed, so that the fast and accurate evaluation of the bridge driving safety is realized.
Description
Technical Field
The invention relates to the technical field of train driving safety, in particular to a quick evaluation method for on-bridge driving safety based on a mapping relation agent model.
Background
The operation mileage of the high-speed railway in China reaches 4 kilometers, and the average percentage of bridges is 58%. The line mileage under special geology and complex climatic conditions such as soft soil, high temperature, frost heaving and the like is longer and longer, and the phenomena of deformation of an offline foundation structure and performance degradation of an interlayer structure are more and more prominent. For high-speed railway bridges, attention should be paid to the safety and comfort of trains running on the tracks.
Since the performance evolution of the infrastructure is a complex nonlinear problem, if various additional deformations, interlayer connection failures and other nonlinear factors are directly considered in the dynamics of the vehicle-infrastructure system, huge calculation scale is caused, and the safety of the vehicle running on the bridge under the complex environment is difficult to evaluate quickly in real time. Based on the agent model technology, the method for efficiently predicting the complex response of the real system can solve the problems of long time consumption, high calculation cost and the like of dynamic analysis numerical simulation to a certain extent by simulating the mapping relation between input and output, but the method is generally based on prediction optimization of specific response of the real system, is difficult to accurately describe the change rule of the global trend of the system, and needs to improve the optimization efficiency and the final precision.
Disclosure of Invention
The invention aims to provide a method for quickly evaluating the safety of a vehicle running on a bridge based on a mapping relation agent model, which is used for establishing the agent model for efficiently solving the traffic safety parameters under the condition of information updating, so that the quick evaluation and prejudgment of the safety performance of the vehicle running under the complex service condition are realized, and the safety service of a bridge-track system is guaranteed.
The invention is realized by the following technical scheme:
a method for evaluating the safety of a vehicle running on a bridge based on a mapping relation agent model comprises the following steps,
and 4, evaluating the driving safety and comfort based on the agent model.
In the prior art, the performance of a basic structure is evolved into a complex nonlinear problem, various nonlinear factors such as additional deformation and interlayer connection failure cannot be combined with vehicle-basic structure system dynamics, namely, the driving safety on a bridge in a complex environment cannot be accurately evaluated in real time, therefore, an applicant develops a new multi-target parallel optimization algorithm which can simultaneously consider system global search and local development and combines the algorithm with a vehicle-rail-bridge coupling vibration model to build a mapping relation agent model for efficiently calculating driving safety parameters under a complex service condition to form an on-bridge driving safety evaluation device, so that the on-bridge driving safety can be quickly and accurately evaluated; the specific operation steps are as follows:
firstly, recording and obtaining the service state of a bridge-track structure along a high-speed rail by a detection monitoring module, such as information of bridge pier settlement and the like, so as to obtain evolution data of the service performance of a basic structure, and predicting the influence rule of the train dynamic performance by using a train-track-bridge coupling vibration model based on the obtained evolution data, thereby constructing a multi-point mapping relation required by an agent model; and extracting initial modeling samples based on the multipoint mapping relation, establishing an initial Kriging proxy model, and then completing the construction of the mapping relation proxy model according to an EF-MSE algorithm.
It should be noted that the EF-MSE algorithm includes an EF function and an MSE function, and the MSE function is used to take the point with larger prediction uncertainty as an update sample.
In step 3, the mapping relation agent model is constructed as follows:
firstly, extracting an initial modeling sample by adopting a Latin hypercube test based on a multipoint mapping relation, and establishing an initial Kriging agent model;
and secondly, performing parallel iterative optimization on the whole dynamic response change trend and the safety limit value region according to an EF-MSE algorithm until the optimization convergence requirement is met, and finally completing the construction of a mapping relation agent model.
The step 3 of constructing the mapping relation agent model specifically comprises the following steps,
3.1, generating initial sample variables required by modeling in a given design space;
step 3.2, acquiring actual response of the existing sample variable through physical methods such as numerical analysis and the like, and constructing a Kriging agent model aiming at the corresponding sample pair;
3.3, respectively calculating the global trend of the current Kriging model and the deviation condition of the prediction and the actual response of the threshold region;
step 3.4, checking whether the Kriging model is converged, if the set convergence condition is met, stopping iteration, and outputting the current model as the optimal agent model; otherwise, executing the next step and continuing optimization updating;
step 3.5, solving two point adding criteria based on an EF function and an MSE function through an intelligent optimization algorithm, and simultaneously obtaining a plurality of parallel new samples;
step 3.6, performing correlation analysis between new samples and between the new samples and the existing samples by using a Gaussian criterion to delete redundant samples to obtain final optimized samples;
wherein, the expression of the Gaussian criterion is as follows:
in the formula, theta i The parameters of the correlation function can be solved through maximum likelihood estimation;
and 3.7, adding the optimized sample into the existing sample variable, and turning to the step 3.2, so that the Kriging proxy model is continuously updated until convergence.
In the step 3.1, a Latin hypercube test design method which can enable the samples to be randomly distributed in the whole design space is adopted for sampling.
In said step 3.5, the EF function has the form:
in the formula (I), the compound is shown in the specification,
s (x) are respectively the predicted mean and standard deviation estimation of the Kriging model; z is a specified threshold; z is a radical of + 、z - Respectively, above and below the threshold value, and is usually denoted as z ± ＝z±μ；
Is composed of
The probability density of (a) of (b),
on the basis, the optimization and updating of the proxy model can be realized by solving the sample point corresponding to the EF (x) maximum value:
x new ＝argmax{EF(x)}。
in said step 3.5, the form of the MSE function is as follows:
in the formula, x new Is a new sample variable for optimization;
mean square error estimates representing predictions of relative real system responses; arg is the mapping relationship between independent variables and dependent variables.
In the step 1, the service state of the bridge-track structure along the high-speed rail comprises bridge pier settlement, beam end corner, gap, plate bottom void and upwarp state.
The system also comprises a performance evolution data detection and monitoring module, which is used for recording the service state of the bridge-track structure along the high-speed rail;
the train response prediction module is used for inputting the measured data into the mapping relation agent model and outputting a vehicle response value and a change trend in real time;
and the quick evaluation and early warning module is used for analyzing the response value and the change trend of the vehicle and outputting the evaluation and early warning results of the driving safety and comfort in real time.
Through the matching use of the performance evolution data detection monitoring module, the train response prediction module and the rapid evaluation and early warning module, detection personnel in the field can more efficiently and accurately evaluate and early warn the driving safety on the bridge, and the safety service of the bridge-track system is guaranteed.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the EF-MSE algorithm provided by the invention can perform parallel optimization on global search and local development in sequence optimization iteration, adaptively improve the multi-target optimization precision of the model, remarkably reduce the number of times of adding points, greatly improve the convergence rate, show good stability, and calculate the dynamic response characteristics of the on-bridge train with low calculation amount and physical mechanism explanation by the proxy model with high precision and high efficiency;
2. the on-bridge vehicle safety evaluation device formed based on the method can be used for quickly evaluating and prejudging the safety performance of vehicles under complex service conditions, promotes the gradual formation of intelligent evaluation and early warning systems for the driving safety on high-speed bridges, and powerfully ensures the safe service of a bridge-track system.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a flow chart of the evaluation method of the present invention;
FIG. 2 is a flowchart of a proxy model for mapping relationships constructed based on the EF-MSE method;
FIG. 3 is a block diagram of the driving safety rapid evaluation device according to the present invention;
FIG. 4 is the acceleration response of the train vehicle in the working condition of the combination of pier settlement and plate bottom emptying in the embodiment 1;
FIG. 5 shows the vertical force response of the train wheel track under the combined working conditions of pier settlement and plate bottom emptying in example 1;
FIG. 6 is a mapping relationship between pier settlement and the maximum reduction of vertical force of wheel tracks under the condition of plate bottom void in example 1;
fig. 7 is a mapping relationship between pier settlement and the maximum variation of the vertical acceleration under the plate bottom void condition in example 1;
FIG. 8 is a process of global precision optimization of two-dimensional nonlinear multi-modal functions in example 2;
FIG. 9 is a two-dimensional nonlinear multi-modal function threshold precision optimization process according to example 2;
FIG. 10 is a two-dimensional nonlinear cubic function global accuracy optimization process in example 2;
FIG. 11 is a two-dimensional non-linear cubic function threshold accuracy optimization process of example 2.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention. It should be noted that the present invention is in practical development and use.
Example 1
As shown in fig. 1, the present embodiment includes the steps of,
taking the Jing Hu high-speed rail ascending section (K1308+ 733.919-K1309 +159.71375) as an example, the data detailed in Table 1 are measured by recording the actual performance evolution data of the infrastructure through the detection monitoring module.
TABLE 1 Jinghush high-speed iron base structure Performance evolution actual measurement data
Note: and the plate bottom void is a longitudinal rail plate bottom void corresponding to the pier top position of the pier in the table.
the method is characterized in that a 32m simply supported beam bridge of a Jinghu high-speed rail, a CRTS II type plate ballastless track and a CRH2 train are taken as research objects, and aiming at the typical working condition of bridge pier settlement and bed plate void, the additional irregularity of the track is solved and input into a train-rail-bridge coupling vibration model as system excitation based on a bridge-track deformation mapping model, so that the train dynamic response under the typical working condition is obtained.
As shown in fig. 4 and 5, since only the rail is vertically deformed due to the bridge pier settlement and the plate bottom void, when the train passes through the settlement and void region, the vertical acceleration of the train body and the vertical force of the wheel rail are significantly changed, and the transverse acceleration and the transverse force of the wheel rail are not basically changed, which indicates that the combined action of the bridge pier settlement and the slab void of the rail mainly affects the vertical vibration of the train-rail-bridge system, and the influence on the transverse dynamic performance is small.
The measurement and calculation results show that typical diseases of the high-speed railway foundation structure in the soft soil area are uneven settlement of piers and rail plate bottom void, and obvious influence is generated on the vertical dynamic performance of the train. Therefore, the driving safety and comfort of the high-speed rail operated in the subsidence area are comprehensively considered, the driving speed is 350km/h as a calculation working condition, the combined action of subsidence and emptying is considered in the mapping relation, and the vertical force and vertical acceleration change of the train wheel track caused by the combined action are correspondingly considered.
Wherein, the settlement and the emptying ranges are respectively 0-30 mm and 0-6 m, and the emptying position is selected to settle adjacent piers which are more unfavorable for the deformation of the track.
the overall accuracy of the model is based on a plurality of sample pairs uniformly distributed in the design space, and the average error between the response prediction result and the rail bridge finite element simulation result is calculated for testing.
For the description precision of the threshold contour, calculating the deviation of the numerical simulation result and the threshold value through a plurality of sample pairs uniformly extracted from the threshold value contour line simulated by the current agent model for checking;
firstly, selecting a combination working condition of 20 different bridge pier settlement and plate bottom emptying by adopting Latin hypercube sampling, and obtaining corresponding train wheel-rail vertical force variation through a train-rail-bridge coupling vibration model to form an initial sample pair so as to establish an initial Kriging agent model; and secondly, performing parallel iterative optimization on the whole dynamic response change trend and the safety limit value region according to two point adding criteria of the EF function and the MSE function until the optimization convergence requirement is met, and finally completing the construction of the mapping relation agent model. The overall accuracy of the model is based on 20 sample pairs uniformly distributed in the design space, and the average error between the response prediction result and the rail bridge finite element simulation result is calculated for testing. And for the description precision of the threshold contour, calculating the deviation of the numerical simulation result and the threshold value through 20 sample pairs uniformly extracted from the threshold value contour line simulated by the current agent model, and checking. In this example, the global prediction average absolute error is less than 5% and the threshold prediction deviation is less than 0.1% as the accuracy convergence condition.
The precision test result shows that for the whole design space, the average absolute error of the wheel-rail vertical force variation predicted by the agent mapping model is 2.86%, and the prediction deviation of the region near the safety limit value is only 0.81%, so that the optimization target is met, and the agent model optimized based on the EF-MSE algorithm can better realize the prediction requirements of the global trend and the threshold contour. The results of the simulation of the converged physiological model for the amount of vertical force variation in the wheel-rail are shown in fig. 6.
Similarly, a mapping relation between pier settlement and the maximum variation of the vertical acceleration under the plate bottom void condition is constructed based on an EF-MSE optimization algorithm. FIG. 7 is a mapping relationship of the surrogate model simulation after convergence, comparing the numerical simulation results, the average deviation of the model overall prediction is 1.78%, and the threshold region prediction deviation is 0.18%, which satisfies the convergence requirement.
In addition, for other various basic structure performance degradation phenomena such as beam end corners, separation joints, arching and the like, the driving power response under the corresponding working condition can be rapidly output through training the mapping relation based on the proxy model technology, so that large-scale vehicle-rail-bridge coupling calculation is avoided.
Step 4, evaluating the driving safety and comfort based on the agent model; and (3) replacing complex time-consuming track bridge calculation by constructed mapping relation agent model analysis, outputting a train dynamic response result, and evaluating train driving safety and comfort indexes and early warning in real time.
By adopting a driving safety evaluation method based on the proxy model, the complex time-consuming track bridge calculation is replaced by the constructed mapping relation, and the train dynamic response of the Shanghai high-speed rail ascending section (K1308+ 733.919-K1309 +159.71375) under different working conditions is accurately output in real time, as shown in Table 2.
TABLE 2 solving train Power response based on mapping agent model
As can be seen from Table 2, when the train passes through No. 441 bridge piers from No. 440 bridge piers, the driving safety and comfort are out of limit, the maximum wheel weight load shedding rate reaches 0.646, and the maximum vertical acceleration is 1.746m/s 2 At the moment, the uneven settlement and the emptying are respectively 27mm and 2.5m, and the train about to pass through the bridge span should send out speed reduction early warning in time so as to ensure the safety and the comfort of the train. Therefore, the driving safety evaluation method based on the agent model realizes the rapid and accurate evaluation and early warning of the driving safety and provides a foundation for guaranteeing the long-term operation safety of the train.
Example 2
As shown in fig. 2, this embodiment is based on embodiment 1, wherein the step 3 of constructing the mapping relationship agent model specifically includes the following steps,
3.1, generating initial sample variables required by modeling in a given design space; and sampling by adopting a Latin hypercube test design method which can ensure that the samples are randomly distributed in the whole design space.
Step 3.2, acquiring actual response of the existing sample variable through physical methods such as numerical analysis and the like, and constructing a Kriging agent model aiming at the corresponding sample pair;
3.3, respectively calculating the global trend of the current Kriging model and the deviation condition of the prediction and the actual response of the threshold region;
step 3.4, checking whether the Kriging model is converged, if the set convergence condition is met, stopping iteration, and outputting the current model as the optimal agent model; otherwise, executing the next step and continuing optimization updating;
step 3.5, solving two point adding criteria based on an EF function and an MSE function through an intelligent optimization algorithm, and simultaneously obtaining a plurality of parallel new samples; preferably, the optimization algorithm adopts a genetic algorithm with strong global optimization and high search efficiency;
step 3.6, performing correlation analysis between new samples and between the new samples and the existing samples by using a Gaussian criterion to delete redundant samples to obtain final optimized samples;
wherein, the expression of the Gaussian criterion is as follows:
in the formula, theta i The parameters of the correlation function can be solved through maximum likelihood estimation;
and 3.7, adding the optimized sample into the existing sample variable, and turning to the step 3.2, so that the Kriging proxy model is continuously updated until convergence.
The arithmetic processing of the EF function and the MSE function in the present embodiment,
wherein the form of the EF function is as follows:
in the formula (I), the compound is shown in the specification,
s (x) are respectively the predicted mean and standard deviation estimation of the Kriging model; z is a specified threshold; z is a radical of + 、z - Respectively, above and below the threshold value, and is usually denoted as z ± ＝z±μ；
Is composed of
The probability density of (a) of (b),
on the basis, the optimization and updating of the proxy model can be realized by solving the sample point corresponding to the EF (x) maximum value:
x new ＝argmax{EF(x)}。
the form of the MSE function is as follows:
in the formula, x new Is a new sample variable for optimization;
mean square error estimates representing predictions of relative real system responses; arg is the mapping relationship between independent variables and dependent variables.
By taking two typical complex two-dimensional nonlinear functions as an example, the superiority of the EF-MSE algorithm in the aspects of global trend and threshold search is comparatively analyzed. Both function thresholds are defined as function zeros.
The specific expressions of the two-dimensional functions are as follows:
(1) two-dimensional nonlinear multi-modal function:
x 1 ∈[-4,7]x 2 ∈[-3,8]，
(2) two-dimensional nonlinear cubic function:
because the EF-MSE algorithm is a parallel optimization process, when each optimization function in the convergence criterion meets the convergence requirement, only the optimization process of the optimization function is ended, and the overall optimization process of the algorithm is not ended.
Based on 20 initial samples, the optimization process of the two-dimensional nonlinear function global and threshold prediction precision by adopting different algorithms is shown in fig. 8 to 11.
The convergence results of the different optimization algorithms are shown in table 3.
TABLE 3 comparison of convergence results for different optimization algorithms
It can be found from table 3 that the EF-MSE algorithm optimization can effectively reduce the number of optimization times while maintaining good prediction accuracy, and accordingly, although the EGRA algorithm only needs 7.62 sub-optimization for the cubic function with regular spatial distribution, too early convergence and a large function gradient also result in poor prediction results.
As an optimization algorithm of parallel computer calculation, the total running time of an EF-MSE algorithm program is greatly shortened compared with other algorithms, and the optimization efficiency is remarkably improved; in addition, under the condition that the convergence conditions of the same optimization function are the same, although the number of samples added in one parallel optimization process of the EF-MSE algorithm is twice that of other algorithms, the total number of samples is only slightly larger than that of other algorithms, and the EF-MSE algorithm is not simple superposition of the two algorithms, but has more stable and efficient optimization performance.
Example 3
As shown in fig. 3, on the basis of embodiment 1, the embodiment further includes a performance evolution data detection and monitoring module, which is configured to record a service state of the bridge-track structure along the high-speed rail;
the train response prediction module is used for inputting the measured data into the mapping relation agent model and outputting a vehicle response value and a change trend in real time;
and the quick evaluation and early warning module is used for analyzing the response value and the change trend of the vehicle and outputting the evaluation and early warning results of the driving safety and comfort in real time.
Through the matching use of the performance evolution data detection monitoring module, the train response prediction module and the rapid evaluation and early warning module, detection personnel in the field can more efficiently and accurately evaluate and early warn the driving safety on the bridge, and the safety service of the bridge-track system is guaranteed.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. A quick evaluation method for safety of an on-bridge vehicle based on a mapping relation agent model is characterized by comprising the following steps: comprises the following steps of (a) carrying out,
step 1, recording service performance evolution of a basic structure; recording the service state of the bridge-track structure along the high-speed rail by a detection monitoring module;
step 2, constructing a multipoint mapping relation required by the agent model; according to the service state of the bridge-track structure along the high-speed rail obtained through detection and monitoring, the influence rule of the bridge-track-bridge coupled vibration model on the dynamic performance of the train is predicted based on the train-track-bridge coupled vibration model, and then the multipoint mapping relation required to be constructed by the agent model is clarified;
step 3, constructing a mapping relation agent model; extracting an initial modeling sample based on a multipoint mapping relation, and establishing an initial Kriging agent model; then according to the EF-MSE algorithm, completing the construction of a mapping relation agent model;
and 4, evaluating the driving safety and comfort based on the agent model.
2. The on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 1, characterized in that: in step 3, the mapping relation agent model is constructed as follows:
firstly, extracting an initial modeling sample by adopting a Latin hypercube test based on a multipoint mapping relation, and establishing an initial Kriging agent model;
and secondly, performing parallel iterative optimization on the whole dynamic response change trend and the safety limit value region according to an EF-MSE algorithm until the optimization convergence requirement is met, and finally completing the construction of a mapping relation agent model.
3. The on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 2, characterized in that: the step 3 of constructing the mapping relation agent model specifically comprises the following steps,
3.1, generating initial sample variables required by modeling in a given design space;
step 3.2, acquiring actual response of the existing sample variable through physical methods such as numerical analysis and the like, and constructing a Kriging agent model aiming at the corresponding sample pair;
3.3, respectively calculating the global trend of the current Kriging model and the deviation condition of the prediction and the actual response of the threshold region;
step 3.4, checking whether the Kriging model is converged, if the set convergence condition is met, stopping iteration, and outputting the current model as the optimal agent model; otherwise, executing the next step and continuing optimization updating;
step 3.5, solving two point adding criteria based on an EF function and an MSE function through an intelligent optimization algorithm, and simultaneously obtaining a plurality of parallel new samples;
step 3.6, performing correlation analysis between new samples and between the new samples and the existing samples by using a Gaussian criterion to delete redundant samples to obtain final optimized samples;
wherein, the expression of the Gaussian criterion is as follows:
in the formula, theta i The parameters of the correlation function can be solved through maximum likelihood estimation;
and 3.7, adding the optimized sample into the existing sample variable, and turning to the step 3.2, so that the Kriging proxy model is continuously updated until convergence.
4. The on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 3, characterized in that: in the step 3.1, a Latin hypercube test design method which can enable the samples to be randomly distributed in the whole design space is adopted for sampling.
5. The on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 3, characterized in that: in said step 3.5, the EF function has the form:
in the formula (I), the compound is shown in the specification,
s (x) are respectively the predicted mean and standard deviation estimation of the Kriging model; z is a specified threshold; z is a radical of + 、z - Respectively, above and below the threshold value, and is usually denoted as z ± ＝z±μ；
Is composed of
The probability density of (a) of (b),
on the basis, the optimization and updating of the proxy model can be realized by solving the sample point corresponding to the EF (x) maximum value:
x new ＝argmax{EF(x)}。
6. the on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 3, characterized in that: in said step 3.5, the form of the MSE function is as follows:
in the formula, x new Is a new sample variable for optimization;
a mean square error estimate representing a prediction of the associated true system response; arg is the mapping relationship between independent variables and dependent variables.
7. The on-bridge vehicle safety rapid evaluation method based on the mapping relation agent model according to claim 1, characterized in that: in the step 1, the service state of the bridge-track structure along the high-speed rail comprises bridge pier settlement, beam end corner, gap, plate bottom void and upwarp state.
8. The on-bridge traveling safety evaluation method based on the mapping relation agent model according to any one of claims 1 to 7, characterized in that: also comprises
The performance evolution data detection monitoring module is used for recording the service state of the bridge-track structure along the high-speed rail;
the train response prediction module is used for inputting the measured data into the mapping relation agent model and outputting a vehicle response value and a change trend in real time;
and the quick evaluation and early warning module is used for analyzing the response value and the change trend of the vehicle and outputting the evaluation and early warning results of the driving safety and comfort in real time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210590453.9A CN114861458B (en) | 2022-05-26 | 2022-05-26 | Bridge traffic safety rapid evaluation method based on mapping relation proxy model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210590453.9A CN114861458B (en) | 2022-05-26 | 2022-05-26 | Bridge traffic safety rapid evaluation method based on mapping relation proxy model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114861458A true CN114861458A (en) | 2022-08-05 |
| CN114861458B CN114861458B (en) | 2023-07-14 |
Family
ID=82641111
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210590453.9A Active CN114861458B (en) | 2022-05-26 | 2022-05-26 | Bridge traffic safety rapid evaluation method based on mapping relation proxy model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114861458B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336869A (en) * | 2013-07-05 | 2013-10-02 | 广西大学 | Multi-objective optimization method based on Gaussian process simultaneous MIMO model |
| CN110609975A (en) * | 2019-09-20 | 2019-12-24 | 大连理工大学 | Parallel computing global optimization algorithm based on Kriging agent model |
| CN110728017A (en) * | 2019-08-21 | 2020-01-24 | 西南交通大学 | Method and device for constructing dynamic mapping relation between bridge additional deformation and driving safety |
| CN113505545A (en) * | 2021-06-21 | 2021-10-15 | 西南交通大学 | Rail transit vehicle pneumatic multi-objective optimization method based on improved point adding criterion |
| CN114117840A (en) * | 2021-10-28 | 2022-03-01 | 中国运载火箭技术研究院 | Structural performance prediction method based on simulation and test data hybrid drive |
-
2022
- 2022-05-26 CN CN202210590453.9A patent/CN114861458B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336869A (en) * | 2013-07-05 | 2013-10-02 | 广西大学 | Multi-objective optimization method based on Gaussian process simultaneous MIMO model |
| CN110728017A (en) * | 2019-08-21 | 2020-01-24 | 西南交通大学 | Method and device for constructing dynamic mapping relation between bridge additional deformation and driving safety |
| CN110609975A (en) * | 2019-09-20 | 2019-12-24 | 大连理工大学 | Parallel computing global optimization algorithm based on Kriging agent model |
| CN113505545A (en) * | 2021-06-21 | 2021-10-15 | 西南交通大学 | Rail transit vehicle pneumatic multi-objective optimization method based on improved point adding criterion |
| CN114117840A (en) * | 2021-10-28 | 2022-03-01 | 中国运载火箭技术研究院 | Structural performance prediction method based on simulation and test data hybrid drive |
Non-Patent Citations (3)
| Title |
|---|
| HONGYE GOU .ETC: "In-situ testing and model updating of a long-span cable-stayed railway bridge with hybrid girders subjected to a running train", vol. 253, no. 253, pages 1 - 14 * |
| 勾红叶等: "高速铁路桥梁-轨道体系检测监测与行车安全研究进展", vol. 22, no. 1, pages 1 - 23 * |
| 闫宇智等: "高寒地区高速铁路桥梁下部结构模态参数的变化规律研究", vol. 38, no. 6, pages 47 - 55 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114861458B (en) | 2023-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111256924B (en) | Intelligent monitoring method for expansion joint of large-span high-speed railway bridge | |
| Liu et al. | Dynamic response limit of high-speed railway bridge under earthquake considering running safety performance of train | |
| Yousefikia et al. | Modeling degradation of tracks for maintenance planning on a tram line | |
| Lu et al. | Lifetime deflections of long-span bridges under dynamic and growing traffic loads | |
| CN101697175A (en) | Simulated prediction method for rail transit noise | |
| Peng et al. | A novel bagged tree ensemble regression method with multiple correlation coefficients to predict the train body vibrations using rail inspection data | |
| Zhang et al. | An adaptive surrogate model approach for random vibration analysis of the train–bridge system | |
| Gao et al. | Estimation of rail renewal period in small radius curves: A data and mechanics integrated approach | |
| Xie et al. | Stochastic analysis of dynamic stress amplification factors for slab track foundations | |
| Liu et al. | Review of perspectives on pantograph-catenary interaction research for high-speed railways operating at 400 km/h and above | |
| Wang et al. | Time-variant fatigue reliability assessment of rib-to-deck welded joints using ANN-based methods | |
| Li et al. | A reliability assessment approach for slab track structure based on vehicle-track dynamics and surrogate model | |
| Cho et al. | Reliability analysis for the uncertainties in vehicle and high-speed railway bridge system based on an improved response surface method for nonlinear limit states | |
| Deng et al. | Evaluation index of CRTS III prefabricated slab track cracking condition based on interval AHP | |
| Yan et al. | Modeling and parameter optimization of dynamic characteristic variables of ballast bed during operation for dynamic track stabilizer | |
| You et al. | Frequency response function-based model updating of flexible vehicle body using experiment modal parameter | |
| Yao et al. | Prediction on building vibration induced by moving train based on support vector machine and wavelet analysis | |
| CN114861458B (en) | Bridge traffic safety rapid evaluation method based on mapping relation proxy model | |
| Skrypnyk et al. | On the influence of crossing angle on long-term rail damage evolution in railway crossings | |
| Wang et al. | Study on rail profile optimization based on the nonlinear relationship between profile and wear rate | |
| CN112199838A (en) | Method and device for identifying damage of ballast layer under ballast track sleeper | |
| Ye et al. | Review of integrated full life cycle data management and application of the slab tracks | |
| Huang et al. | Quantification of dynamic track stiffness using machine learning | |
| Shi et al. | Estimation of cement asphalt mortar disengagement degree using vehicle dynamic response | |
| CN106844971B (en) | Rail transit bridge noise inversion prediction method based on inverse boundary element |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |