A high-speed rail wind speed instrument precision improvement method based on CFD

Abstract

The application discloses a high-speed rail wind speed instrument precision improving method based on CFD, specifically: under a preset wind speed condition, a scene of an ultrasonic wind speed instrument working on a high-speed rail power supply catenary support is simulated by using CFD, and a wind speed distribution cloud picture is acquired; the cloud picture is divided to determine the wind direction under the condition that the ultrasonic wind speed instrument is most obviously affected by the catenary support shielding; the observed wind speed of the ultrasonic wind speed instrument is acquired; a high-speed rail wind speed instrument physical revision model is established, taking the preset wind speed-catenary support structure as input and the error between the preset wind speed and the observed wind speed as output; the revision model is optimized in parameters by using a Bayesian optimization method, the optimal parameters of the support structure that minimize the error are found, and the revised wind speed under the optimal parameters is acquired; the revised wind speed is corrected by using an XG-Boost model, and the corrected wind speed is acquired. The method of combining the catenary support structure revision and artificial intelligence algorithm improves the real-time prediction accuracy of high-speed rails under extreme gale conditions.

Description

Technical Field

[0001] This invention relates to a method for improving the accuracy of high-speed rail anemometers based on CFD, belonging to the field of railway safety technology. Background Technology

[0002] Strong winds are one of the most significant meteorological hazards threatening the safety of high-speed rail operations. Railway accidents caused by strong winds occur frequently both domestically and internationally. High-speed rail, due to its lightweight nature and high speed, is even more sensitive to strong winds. To ensure the safety of high-speed rail operations, both domestic and international high-speed rail systems have established strong wind forecasting (early warning) systems, which rely on wind speed and direction measurements along the railway lines. This places higher demands on wind measurement along high-speed rail lines.

[0003] Compared with traditional mechanical wind measurement, ultrasonic wind measurement equipment adopts an integrated design, with no moving parts, no mechanical wear, no risk of parts falling off during the wind measurement process, and a long service life. Ultrasonic wind measurement has a fast response speed, high measurement accuracy, high resolution, and can measure high-frequency pulsating components in wind speed. It adopts random error identification technology to ensure low discrete error in measurement even in strong winds, resulting in more stable output and lower maintenance costs, making it the preferred choice for railway wind measurement.

[0004] Because ultrasonic anemometers along high-speed railways are typically mounted on the overhead contact line supports, and these supports are tall and large, they can obstruct wind speed readings under certain wind conditions, leading to measurement errors. High-speed trains are extremely sensitive to strong winds; if the instantaneous wind speed reaches the overturning threshold, it can cause a train to derail in just 1-2 seconds. Therefore, reducing the errors caused by support obstruction through scientific and accurate methods is urgently needed. A review of existing literature on improving the measurement accuracy of ultrasonic anemometers on railways revealed no similar research. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a method for improving the accuracy of high-speed rail anemometers based on CFD, reducing the measurement error caused by the obstruction of the railway power supply contact network support on the anemometer, and making the real-time forecast and early warning of high-speed rail more accurate.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] A method for improving the accuracy of high-speed rail anemometers based on CFD includes the following steps:

[0008] Step 1: Under preset wind speed conditions, use computational fluid dynamics simulation to simulate the scenario of an ultrasonic anemometer working on the support of the high-speed railway power supply contact network, and obtain the wind speed distribution cloud map under the scenario.

[0009] Step 2: Divide the wind speed distribution cloud map according to the preset division rules, determine the obstructed area and the obstructed area in the wind speed distribution cloud map, and thus determine the wind direction when the ultrasonic anemometer is most significantly affected by the obstruction of the contact wire support.

[0010] Step 3: Obtain the observed wind speed of the ultrasonic anemometer using the time difference method, that is, calculate the pulse channel average wind speed when the ultrasonic anemometer is most significantly affected by the obstruction of the contact wire support under the scenario in Step 1.

[0011] Step 4: Establish a physical revision model for the high-speed rail anemometer, with the preset wind speed and contact wire support structure as input and the error between the preset wind speed and the observed wind speed as output.

[0012] Step 5: Use the Bayesian optimization method to optimize the parameters of the physical revision model of the high-speed rail anemometer established in Step 4, find the optimal parameters of the contact wire support structure that minimize the error between the preset wind speed and the observed wind speed, and obtain the revised wind speed under the optimal parameters.

[0013] Step 6: Use the Extreme Gradient Boosting (XG-Boost) model to further refine the revised wind speed and obtain the final corrected wind speed.

[0014] In a preferred embodiment of the present invention, in step 1, the wind speed conditions include wind speed magnitude and wind speed direction, with the wind speed magnitude set to 10-30 m / s and the wind speed direction set to 0-360°.

[0015] As a preferred embodiment of the present invention, in step 2, the division rule is set to divide the wind speed distribution cloud map according to the magnitude of wind speed energy, or to divide the wind speed distribution cloud map according to the isolinear values ​​of a preset number of layers.

[0016] The wind direction when the ultrasonic anemometer is most significantly obstructed by the overhead contact line support, i.e., when the overhead contact line support and the ultrasonic anemometer are in a straight line.

[0017] As a preferred embodiment of the present invention, in step 4, the contact wire support structure includes the height of the contact wire support fixing device and the length of the long rod and the cross rod of the contact wire support. The height of the contact wire support fixing device is the distance of the contact wire support fixing device from the ground. The contact wire support is L-shaped, with one end parallel to the ground being the cross rod and the other end being the long rod.

[0018] In a preferred embodiment of the present invention, in step 6, a preset wind speed is used as the target value, and a revised wind speed is used as the feature value. A training set and a validation set are constructed, and the extreme gradient boosting model XG-Boost is trained and validated using the training set and the validation set to obtain the optimal parameters of the extreme gradient boosting model XG-Boost, including the learning rate, the maximum tree depth, and the L1 regularization weight term. The trained extreme gradient boosting model XG-Boost is then used to further correct the revised wind speed.

[0019] A computer device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the steps of the CFD-based method for improving the accuracy of a high-speed rail anemometer.

[0020] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the CFD-based method for improving the accuracy of high-speed rail anemometers.

[0021] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0022] This invention studies the measurement error of anemometers caused by the overhead contact line support structure along high-speed railways, and reduces this error by modifying the structure of the contact line support structure and combining it with artificial intelligence algorithms. This improves the accuracy of real-time forecasting and early warning for high-speed railways in extreme wind conditions, ensuring the reliability and safety of railway operation under strong wind conditions. Attached Figure Description

[0023] Figure 1 This is a model diagram of the anemometer based on CREO according to the present invention;

[0024] Figure 2 This is a flowchart of a method for improving the accuracy of a high-speed rail anemometer based on CFD, according to the present invention.

[0025] Figure 3 This is a flowchart of the extraction and error analysis of wind speed measurement using an ultrasonic anemometer based on CFD for railway catenary support obstruction.

[0026] Figure 4 This is a flowchart of the Bayesian optimization physical model and the XG-Boost model compensation process;

[0027] Figure 5 This is a wind speed cloud map caused by the anemometer being obstructed by the power supply contact network support. Detailed Implementation

[0028] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0029] like Figure 2 As shown, this invention proposes a CFD-based high-speed rail safety protection method. This method studies the measurement error of anemometers caused by the overhead contact line supports along the high-speed rail line, and reduces this error by modifying the structure of the contact line supports and combining it with artificial intelligence algorithms. This achieves the goal of improving the accuracy of real-time forecasting and early warning for high-speed rail in extreme wind conditions. The specific process is as follows:

[0030] S1. Using computational fluid dynamics (CFD) simulation, under preset wind speed conditions, obtain a cloud map of the wind speed distribution after the air passes through the anemometer on the railway power supply contact network support. Figure 3 As shown.

[0031] S101. Obtain the geometric model of the anemometer implanted in the high-speed rail catenary support based on CREO, such as... Figure 1 As shown; based on FLUENT preset boundary conditions, turbulence model and slip mesh model.

[0032] Considering that high-speed trains will begin to limit speed when the ambient wind speed exceeds 15 m / s, and will be prohibited from entering the section when the ambient wind speed exceeds 30 m / s, specific wind warning and control measures are shown in Table 1. The wind speed and wind direction are set as boundary conditions for the speed limit settings, with the wind speed set to 10-30 m / s and the ambient wind direction set as boundary conditions of 0-360°.

[0033] Table 1. Warning and Control Measures for Strong Winds

[0034]

[0035] S102. Under preset wind speed conditions, use fluid dynamics CFD to simulate the scenario of an ultrasonic anemometer working on a contact wire support, and obtain a cloud map of the wind speed distribution under this scenario, such as... Figure 5 As shown.

[0036] S103. The wind speed conditions are preset. The wind speed conditions can be set to include wind speed magnitude and wind speed direction, etc. When using the CFD simulation method to obtain the cloud map of wind speed distribution, the railway ultrasonic anemometer needs to be simulated under the preset wind speed conditions. It should be noted that the cloud map of wind speed distribution is obtained by making a cross section on the horizontal plane of the ultrasonic anemometer. Therefore, the simulated wind speed of the anemometer can be obtained through the cloud map of wind speed distribution.

[0037] S104. After obtaining the cloud map of wind speed distribution, the cloud map is divided according to a preset division rule to determine the obstructed and unobstructed areas in the cloud map. The obstructed area is the area where the anemometer is obstructed by the contact wire support. In addition, the division rule is preset and can be set by those skilled in the art according to specific design requirements. For example, the division rule can be set to divide according to the magnitude of wind speed energy, or the cloud map can be divided according to a preset number of contour lines, such as dividing the cloud map into 20, 25, or 30 layers, etc. When dividing the cloud map according to the preset number of contour lines, the difference between the multiple layers after division can also be set. For example, when dividing the cloud map according to a preset 20-layer contour line, the difference between the layers determined by the 20-layer contour line is less than 5%. At this time, the influence of the obstruction effect can be effectively analyzed. Since the obstruction effect is most obvious when the obstruction is in a straight line with the anemometer when optimizing the contact wire support structure, this invention assumes it to be a specific wind direction.

[0038] S105. Wind tunnel simulation experiments were conducted to verify the accuracy of CFD simulation experiments on anemometers along the railway line.

[0039] S2. Obtain the anemometer's observation values ​​using the time-difference method, that is, calculate the average wind speed of the ultrasonic anemometer pulse channel in the CFD simulation wind speed distribution under the direction of the obstruction and the anemometer's straight wind direction (hereinafter referred to as the specific direction), and analyze its error with the preset wind speed.

[0040] S3. Based on the revised obstruction relationship between the overhead contact line support and the anemometer, the accuracy of wind measurement on the high-speed rail overhead contact line support of the anemometer has been improved, including:

[0041] S301. Based on the cloud map of wind speed distribution obtained in the previous step, visualize the occlusion relationship between the contact wire support and the anemometer under specific wind conditions. By modifying the height of the fixing device and the length of the long rod and crossbar of the contact wire support, establish a physical revision model for the high-speed rail anemometer, with the ambient wind speed and support structure as input and the error between the preset wind speed and the observed wind speed as output.

[0042] S302. The parameters of the physical model of the high-speed railway anemometer are optimized by Bayesian optimization (BO) to obtain the optimal physical model of the high-speed railway anemometer, determine the parameters of the contact wire support under the optimal conditions, and obtain the revised wind speed.

[0043] A Gaussian process model is established to estimate the objective function, i.e., the error between the preset wind speed and the observed wind speed. The Gaussian process model can learn the distribution of the objective function using existing data and predict the objective function value corresponding to a new combination of design parameters. Bayesian optimization is then used to find the design parameter combination that minimizes the objective function in the parameter space. This process includes selecting a new combination of design parameters, evaluating the objective function value, and updating the Gaussian process model, until a stopping condition is met (e.g., convergence to a sufficiently small objective function value).

[0044] Bayesian optimization is a highly effective global optimization algorithm, primarily consisting of two core components: a probabilistic surrogate model and a data collection function. Its framework is an iterative process: maximizing the data collection function to select the next most promising evaluation point x. Then, the objective function value y is evaluated based on the selected evaluation point. The newly obtained (x, y) is added to the historical evaluation point set, and the probabilistic surrogate model is updated. This iterative process yields the optimal solution, as shown in the following formula:

[0045]

[0046] Where f represents the unknown objective function (or the parameters in the parametric model); D 1:t ={(x1,y1),(x2,y2),...,(x t ,y t )} represents the observed set, x t Let y represent the decision vector. t =f(x) t )+ε t Represents the observed value, ε t p(D) represents the observation error; 1:t |f) represents the likelihood distribution of y, which is also called "noise" due to the error in the observation; p(f) represents the prior probability distribution of f, that is, the hypothesis about the state of the unknown objective function; p(D 1:t p(D) represents the marginal likelihood distribution or "evidence" of marginalized f. Because this marginal likelihood involves the product and integral of probability density functions, it is usually difficult to obtain an explicit analytical expression. In Bayesian optimization, this marginal likelihood is mainly used to optimize hyperparameters; 1:t |t) represents the posterior probability distribution of f, which describes the confidence of the unknown objective function after correcting the prior with the observed dataset.

[0047] After obtaining the optimized contact wire support parameter combination, verification experiments or numerical simulations are conducted to evaluate the actual effect of the contact wire support on the anemometer. The optimization results are verified by simulating the measurement error under specific wind speed conditions, and finally the parameters of the contact wire support under the optimal conditions are determined, while the revised wind speed is obtained.

[0048] S303. The Extreme Gradient Boosting (XG-Boost) algorithm is used to further correct the wind speed in the optimal physical model of the high-speed rail anemometer, thereby compensating for the observation error of the high-speed rail anemometer, basically eliminating the obstruction effect of the catenary support, and ensuring the reliability and safety of railway operation under high wind conditions. The implementation steps of the XG-Boost model are as follows: Figure 4 As shown.

[0049] A training set and a test set are constructed by using a preset wind speed as the target value and the wind speed to be compensated as the feature value. Feature engineering is then performed to add lag features, moving averages, and differencing sequences to enhance the model's ability to represent sequences.

[0050] Estimated output of gradient boosting tree model The predicted score f of all trees k The relationship between (x) can be expressed as:

[0051]

[0052] The XG-Boost algorithm uses m trees (where fm is a tree). k Let Γ represent the space of the regression tree (xk). i This represents the feature of the i-th wind speed sample, and each leaf node j generates a prediction score f. k (x), also known as leaf weight, leaf weight ω j Let j be the regression value of all samples at leaf node j in the tree. If the number of leaves in a tree is T, it can be represented as j∈{1,2,...T}.

[0053] The objective function plays a crucial role in machine learning problems, and its optimization process continues until the reduction of the objective function reaches a finite state. To approximate the set of functions used in the model, the following regularized objective function is defined:

[0054]

[0055] Where φ represents the value of the loss function, and n is the given wind speed sample. This indicates how well the model's training loss function fits the training data. The regularization term represents the complexity of the tree. Here, γ is the tree's partition degree, and λ is the regularization supercoefficient.

[0056] The optimal parameters of the XG-Boost model, including the learning rate, maximum tree depth, and L1 regularization weight (alpha), were obtained. Cross-validation and other methods were used to select the best hyperparameter combination. The trained XG-Boost model was then used to correct the entire wind speed to be compensated, resulting in a sequence closer to the preset wind speed. Finally, based on the physical revision model of the high-speed rail anemometer and the XG-Boost compensation method, the measurement error caused by the obstruction of the anemometer by the railway power supply contact network support was reduced, thereby improving the accuracy of real-time high-speed rail forecasts.

[0057] Based on the same inventive concept, this application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the aforementioned method for improving the accuracy of a CFD-based high-speed rail anemometer.

[0058] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned method for improving the accuracy of a CFD-based high-speed rail anemometer.

[0059] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0060] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0061] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0062] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0063] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A method for improving the accuracy of high-speed rail anemometers based on CFD, characterized in that, Includes the following steps: Step 1: Under preset wind speed conditions, use computational fluid dynamics simulation to simulate the scenario of an ultrasonic anemometer working on the support of the high-speed railway power supply contact network, and obtain the wind speed distribution cloud map under the scenario. Step 2: Divide the wind speed distribution cloud map according to the preset division rules, determine the obstructed area and the obstructed area in the wind speed distribution cloud map, and thus determine the wind direction when the ultrasonic anemometer is most significantly affected by the obstruction of the contact wire support. Step 3: Obtain the observed wind speed of the ultrasonic anemometer using the time difference method, that is, calculate the pulse channel average wind speed when the ultrasonic anemometer is most significantly affected by the obstruction of the contact wire support under the scenario in Step 1. Step 4: Establish a physical revision model for the high-speed rail anemometer, with the preset wind speed and contact wire support structure as input and the error between the preset wind speed and the observed wind speed as output. The contact wire support structure includes the height of the contact wire support fixing device and the length of the long rod and cross rod of the contact wire support. The height of the contact wire support fixing device is the distance of the contact wire support fixing device from the ground. The contact wire support is L-shaped, with one end parallel to the ground being the cross rod and the other end being the long rod. Step 5: Use the Bayesian optimization method to optimize the parameters of the physical revision model of the high-speed rail anemometer established in Step 4, find the optimal parameters of the contact wire support structure that minimize the error between the preset wind speed and the observed wind speed, and obtain the revised wind speed under the optimal parameters. Step 6: Use the Extreme Gradient Boosting (XG-Boost) model to further refine the revised wind speed to obtain the final corrected wind speed; The specific correction process is as follows: Using the preset wind speed as the target value and the revised wind speed as the feature value, a training set and a validation set are constructed. The extreme gradient boosting model XG-Boost is trained and validated using the training set and the validation set to obtain the optimal parameters of the extreme gradient boosting model XG-Boost, including the learning rate, the maximum tree depth, and the L1 regularization weight term. The trained extreme gradient boosting model XG-Boost is then used to further refine the revised wind speed.

2. The method for improving the accuracy of high-speed rail anemometers based on CFD according to claim 1, characterized in that, In step 1, the wind speed conditions include wind speed magnitude and wind speed direction. The wind speed magnitude is set to 10-30 m / s, and the wind speed direction is set to 0-360°.

3. The method for improving the accuracy of high-speed rail anemometers based on CFD according to claim 1, characterized in that, In step 2, the division rule is set to divide the wind speed distribution cloud map according to the magnitude of wind speed energy, or to divide the wind speed distribution cloud map according to the isolinear values ​​of a preset number of layers. The wind direction when the ultrasonic anemometer is most significantly obstructed by the overhead contact line support, i.e., when the overhead contact line support and the ultrasonic anemometer are in a straight line.

4. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the CFD-based high-speed rail anemometer accuracy improvement method as described in any one of claims 1 to 3.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the CFD-based method for improving the accuracy of high-speed rail anemometers as described in any one of claims 1 to 3.

Images