A method for optimizing the layout of sensors in railway station roof structures

By using particle swarm optimization algorithm and minimum mode confidence criterion principle in the structure monitoring of railway passenger stations, the sensors are optimized and arranged, and the problems of fuzzy monitoring point positions and inaccurate data acquisition are solved, thus achieving efficient, scientific and accurate structural health monitoring Data collection.

CN116186999BActive Publication Date: 2024-11-08CHINA RAILWAY FIRST SURVEY & DESIGN INST GRP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211679367.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-26
Publication Date
2024-11-08
Estimated Expiration
2042-12-26

AI Technical Summary

Technical Problem

There are problems in the structural monitoring of railway passenger stations with unreasonable sensor arrangement, fuzzy monitoring point locations and inaccurate data collection, resulting in high monitoring costs and incomplete data, making it difficult to meet the needs of structural health monitoring.

Method used

The Midas Gen file is exported using the structural design software 3D3S Design to perform load condition analysis, and combined with the particle swarm optimization algorithm and the principle of minimum mode confidence criterion, screen sensors for member strain, node deformation, support displacement, and structural acceleration. Optimize the layout and clarify the specific location and coordinates of the monitoring points.

Benefits of technology

It improves the efficiency and scientificity of sensor layout, clarifies the location of strain monitoring points, enhances the rationality of acceleration sensor layout, simplifies the operation process of engineers, and improves the accuracy and coverage of monitoring data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116186999B_ABST
    Figure CN116186999B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for optimizing the layout of sensors in the roof truss structure of a railway passenger station. The existing layout of health monitoring points in railway station buildings has the problems of fuzzy monitoring point positions and low efficiency of the layout method. The present invention exports the data of 3D3S Design, uses Midas Gen to perform load condition analysis, and screens and optimizes the layout of sensors for rod strain, node deformation, support displacement, and structural acceleration based on the analysis results. The present invention can intuitively display the rods that need to be strain monitored, and clearly display the positions of strain monitoring points along the length direction of the rods, solving the problem of fuzzy strain monitoring points in traditional methods; the layout method of acceleration sensors based on the particle swarm optimization algorithm and the principle of minimum modal confidence criterion improves the rationality and scientificity of the layout of monitoring points.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of building structures, and in particular relates to a method for optimizing the arrangement of sensors for a railway passenger station roof truss structure. Background Art

[0002] With the continuous emergence and rapid development of emerging technologies such as big data, cloud computing and artificial intelligence, my country's infrastructure construction and operation and maintenance are at a critical moment of industrial modernization and information transformation and upgrading. As a key node in the railway transportation network, my country's railway passenger station buildings are huge in size, expensive in construction and operation costs, and their safety is related to the national economy and people's livelihood. However, in terms of operation and maintenance and monitoring, there are still problems of "focusing on construction, but not maintenance" and "focusing on data, but not analysis", which makes subsequent emergency repairs costly and difficult to detect potential problems.

[0003] The objects of health monitoring of railway passenger stations include the main structure of roof and canopy, track layer structure and other key components and supports. The main monitoring contents include overall structural response (deformation, dynamic characteristics, etc.), local response (stress strain, internal force, etc.), load and environment (wind, rain, snow, temperature and humidity, etc.). Therefore, for railway station structures with large spans, multiple styles and complex loads, the actual monitoring items and monitoring contents are numerous, and the types and number of sensors required also increase accordingly. Too many sensors will not only increase the monitoring cost, but also affect the identification of key data, while too few sensors will cause incomplete data and make it difficult to meet the monitoring requirements. How to use reasonable sensors and optimize their layout to accurately collect as much structural response information as possible is the key to building a railway station structure monitoring database and model database.

[0004] Scholars have conducted relevant research on the optimization layout methods and applications of sensors in railway passenger stations. The classic sensor optimization layout method (such as the effective independence method) is a typical reverse deletion method, that is, starting from all measurement points, gradually deleting the degrees of freedom that contribute the least to the structural performance. However, when the structure has thousands of degrees of freedom, its computational efficiency is extremely low. Therefore, the sensor layout method combined with modern intelligent algorithms such as genetic optimization and particle swarm optimization has been widely used in building structures with large scales and numbers of degrees of freedom. Although many studies on the layout of health monitoring points for railway station buildings have been carried out at home and abroad, the actual monitoring plans are mostly undertaken by design units or related commercial monitoring companies, and they still have the following problems:

[0005] 1) The location of strain monitoring points is fuzzy. The existing monitoring scheme is mainly carried out by relevant personnel relying on design software (such as: 3D3S, etc.), through the calculation and analysis of the macroscopic bar system model of the structure, to extract the places with large stress and strain and arrange monitoring points. However, this method can often only reflect the approximate position of the monitoring point in the structural system. In the three-dimensional space system, the local position of the monitoring point in the length direction of the bar cannot be indicated, and it is impossible to get a clear conclusion whether it is at the end of the bar or in the middle of the span. Furthermore, in the case of a three-dimensional entity, the coordinate information of the monitoring point in the circumferential direction of the bar section is even more difficult to express clearly. Therefore, the measurement point arrangement scheme designed based on the above method is relatively rough, and the monitoring position is fuzzy.

[0006] 2) Lack of sensor layout methods. The current "Structural Health Monitoring System Design Standard" provides relatively complete acceleration sensor layout guidelines and methods. However, in actual engineering, existing monitoring schemes are still based mainly on empirical judgment, the monitoring methods are too simple, and sensor optimization is not considered, which leads to inaccurate sensor data collection and hinders subsequent structural damage and performance evaluation. The main reason for this problem is that the methods given in the specifications require engineers to have certain structural dynamics concepts and numerical analysis capabilities, and the number of degrees of freedom of large and complex structures is huge, and it is time-consuming and laborious to directly use traditional methods for analysis. Therefore, how to enable more engineers to apply the above-mentioned standard methods to practice in an efficient and convenient way is the basis and key to station health monitoring. Summary of the invention

[0007] In order to make up for the shortcomings of the existing technology, the present invention provides a method for optimizing the layout of railway passenger station roof truss structure sensors, so as to achieve reasonable layout of station roof truss structure health monitoring points, improve the efficiency and convenience of the layout plan, and provide necessary technical support for relevant engineering personnel.

[0008] In order to achieve the above object, the technical solution adopted by the present invention is:

[0009] A method for optimizing the arrangement of sensors for a railway passenger station roof truss structure, characterized in that it comprises the following steps:

[0010] Step 1: Export a file that can be directly read by Midas Gen from the structural design software 3D3S Design, use Midas Gen to perform load condition analysis, and export the required data;

[0011] Step 2: According to the code provisions and the analysis results of the Midas model, the sensors of member strain, node deformation, support displacement, and structural acceleration are screened and optimally arranged in the order of the program modules;

[0012] Step 3: Export the screening and optimization layout results of various types of sensors and related analysis data.

[0013] Further, in step 1, the exported Midas Gen analysis data includes:

[0014] 1) The section name of each component (SectionName.txt);

[0015] 2) Stress values ​​of each cross-section member under all load combination conditions

[0016] 3) Roof truss node displacement data (.xlsx);

[0017] 4) Roof truss support node displacement data (.xlsx);

[0018] 5) Various modes of roof truss nodes (.xlsx).

[0019] Furthermore, in the step 2, the step of screening the rod strain specifically includes:

[0020] Step 1: Determine the interaction parameters, including file import and export paths and stress screening limits;

[0021] Step 2: Read the file "SectionName.txt" to obtain the section names of all components;

[0022] Step 3: Loop through the files Obtain stress data of various rods;

[0023] Step 4: Calculate the combined stress σ of the member based on the table data and formula (1): con ; Among them, σ N is the axial stress; is the stress caused by the bending moment around the y-axis; is the stress caused by the bending moment around the z-axis;

[0024]

[0025] Step 5: Determine whether the combined stress of the rod exceeds the stress screening limit and obtain the exceeding limit rod;

[0026] Step 6: Screen the over-limit members with the largest combined stress, that is, keep the group of over-limit members with the largest combined stress at the same position;

[0027] Step 7: Save all screening results in Excel document format to the specified directory.

[0028] Furthermore, in step 2, the node deformation screening step specifically includes:

[0029] Step 1: Determine the interaction parameters, including file import and export paths and displacement screening limits;

[0030] Step 2: Read the file "roof truss node displacement data.xlsx" to obtain the displacement of all nodes;

[0031] Step 3: Substitute the displacement components U in the three directions x , U y , U z Sort the data from large to small;

[0032] Step 4: Determine whether the displacement of the node in each direction exceeds the corresponding displacement screening limit, and obtain the node with excessive displacement;

[0033] Step 5: Filter out the nodes with the largest displacement, that is, keep the group with the largest displacement in the same direction among the nodes with the same displacement;

[0034] Step 6: All screening results are saved in the Excel document format to the specified directory.

[0035] Furthermore, in step 2, the step of screening support deformation specifically includes:

[0036] Step 1: Determine the interaction parameters, including file import and export paths, and support displacement screening limits in the x, y, z, and space directions;

[0037] Step 2: Read the file "roof truss support node displacement data.xlsx" to obtain the displacement of all support nodes;

[0038] Step 3: Calculate the spatial displacement U of each node according to formula (2): xyz , where U x , U y , U z is the displacement value of the node in three directions;

[0039]

[0040] Step 4: Connect the U x , U y , U z , U xyz The four components are sorted from large to small;

[0041] Step 5: Determine whether the displacement of the support node in each direction exceeds the corresponding displacement screening limit, and obtain the support displacement exceeding limit node;

[0042] Step 6: Filter out the over-limit nodes with the largest displacement, that is, retain the group with the largest displacement in the same direction among the over-limit nodes;

[0043] Step 7: Save all screening results in Excel document format to the specified directory.

[0044] Furthermore, in step 2, the step of screening and optimizing the structural acceleration specifically includes:

[0045] Step 1: Determine the file import and export paths, as well as the parameters of the particle swarm optimization algorithm;

[0046] Step 2: Read the file "roof truss node various order modes.xlsx", transform and generate the modal information matrix in three directions ([U x ]、[U y ]、[U z ]);

[0047] Step 3: Execute the particle swarm optimization algorithm to obtain the optimal layout of the structural acceleration sensor;

[0048] Step 4: Save all optimization layout results in Excel document format to the specified directory.

[0049] Furthermore, the parameters of the particle swarm optimization algorithm in step 1 include:

[0050] 1) Particle swarm size N;

[0051] 2) The number of iterations K;

[0052] 3) Particle dimension D, i.e. the number of acceleration sensors planned to be deployed;

[0053] 4) Individual learning factor c1 and group learning factor c2;

[0054] 5) Maximum inertia weight ω max With the minimum inertia weight ω min ;

[0055] 6) Maximum speed limit υ max With the minimum speed limit υ min ;

[0056] 7) Maximum position limit x max With the minimum position limit x min , default x max is the particle dimension D, x min is 1.

[0057] Furthermore, the specific steps of the particle swarm optimization algorithm in step 3 include:

[0058] Step 1): Read the particle swarm parameters and complete initialization;

[0059] Step 2): Randomly initialize the position and velocity of the particle swarm, that is, in the D-dimensional population, according to the position limit [x min ,x max ] and speed limit [v min,v max ] Randomly select the position X of the nth particle nd and the initial value of speed V nd , which is expressed as follows:

[0060] X nd ={x n1 ,x n2 ,......,x nD} (3)

[0061] V nd = {v n1 ,v n2 ,......,v nD} (4)

[0062] Step 3): Calculate the optimal P of the particle swarm individuals through the fitness function nd,pbest and the group optimal P d,gbest , which is expressed as follows:

[0063] P nd,pbest ={p n1 ,p n2 ,......,p nD} (5)

[0064] P d,gbest ={p 1,gbest ,p 2,gbest ,......,p D,gbest} (6)

[0065] Where n is the particle number, n = 1, 2, ..., N; d is the particle dimension number, d = 1, 2, ..., D; 1 ≤ j ≤ D;

[0066] Step 4): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value. If not, update the particle position and velocity according to equations (7) and (8).

[0067]

[0068]

[0069] Among them, r1 and r2 are random numbers in the interval [0, 1]; is the position vector of particle n in the dth dimension in the kth iteration; is the velocity vector of particle n in the dth dimension in the kth iteration; is the historical optimal position of particle n in the dth dimension in the kth iteration, that is, the optimal solution searched by the nth particle (individual) after the kth iteration; is the historical optimal position of the swarm in the dth dimension in the kth iteration, that is, the optimal solution in the entire particle swarm after the kth iteration; ω is the inertia weight coefficient, which is calculated as follows.

[0070]

[0071] Among them, k is the current iteration number.

[0072] Step 5): According to step 3), update the individual optimal position and fitness value, the group optimal position and fitness value;

[0073] Step 6): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value; if not, repeat steps 3) to 5) until the conditions are met.

[0074] Furthermore, the specific steps of the fitness function in step 3) include:

[0075] Step ①: Read data information, i.e., modal matrix [ndof×nModel], modal order nModel, particle displacement matrix [D×N]; where ndof is the number of structural degrees of freedom in the modal matrix; nModel is the modal order;

[0076] Step ②: Generate a new modal matrix [D×nModel];

[0077] Step 3: According to formula (10), solve the MAC matrix [nModel×nModel] of the new modal matrix;

[0078]

[0079] Among them, MAC ij is the element in the i-th row and j-th column of the MAC matrix; Φ i , Φ j Represent the modal vectors of the i-th and j-th order respectively.

[0080] Step ④: Select the maximum value max(MAC) of the non-main diagonal elements of the MAC matrix through loop traversal;

[0081] Step ⑤: Construct the fitness function Fitness according to formula (11), and iterate to obtain the optimal fitness value;

[0082] Fitness = 1 - max(MAC) (11).

[0083] Beneficial effects of the present invention:

[0084] 1) The present invention can intuitively display the rods that need to be strain monitored, clearly display the position of the strain monitoring points along the length direction of the rods, and their specific coordinates on the cross section of the component, solving the problem of fuzzy strain monitoring points in the traditional layout method;

[0085] 2) The acceleration sensor layout method of the present invention based on the particle swarm optimization algorithm and the minimum modal confidence criterion principle improves the rationality and scientificity of the monitoring point layout and makes up for the shortcomings of the sensor layout method dominated by experience;

[0086] 3) The operation method of the present invention is simple, and on the basis of meeting the needs of engineering personnel for formulating monitoring point layout plans, the efficiency of sensor screening and optimization is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 A flow chart of selecting and optimizing sensors for the roof truss structure of a railway passenger station in the method of the present invention;

[0088] Figure 2 The calculation results of the roof structure model of a station;

[0089] Figure 3 This is a schematic diagram of the arrangement of strain monitoring points of a station house truss obtained based on the method of the present invention;

[0090] Figure 4 A schematic diagram of the layout of the displacement monitoring points of a station building frame obtained based on the method of the present invention;

[0091] Figure 5 A schematic diagram of the arrangement of monitoring points for displacement of a station's building frame support obtained based on the method of the present invention;

[0092] Figure 6 The schematic diagram of the layout of the acceleration monitoring points of the building truss of a certain station and the fitness curve obtained based on the method of the present invention are shown in FIG. DETAILED DESCRIPTION

[0093] The present invention is described in detail below in conjunction with specific implementation modes.

[0094] The present invention is used for optimizing and rationally disposing the sensors for monitoring the health of the roof structure of a railway passenger station, such as Figure 1 As shown, the present invention comprises the following steps:

[0095] Step 1: Export the (.mgt) file that can be directly read by Midas Gen from the structural design software 3D3S Design. With the help of Midas Gen's convenient data post-processing capabilities, perform model load condition analysis and export the required data, including:

[0096] 1) The section name of each component (SectionName.txt);

[0097] 2) Stress values ​​of each cross-section member under all load combination conditions

[0098] 3) Roof truss node displacement data (.xlsx);

[0099] 4) Roof truss support node displacement data (.xlsx);

[0100] 5) Each order mode of the roof truss node (.xlsx);

[0101] Step 2: According to the code provisions and the analysis results of the Midas model, the sensors of member strain, node deformation, support displacement, and structural acceleration are screened and optimally arranged in the order of the program modules; the screening basis and methods mainly include stress screening limit, (vertical, horizontal, spatial) displacement screening limit, particle swarm optimization algorithm and minimum modal confidence criterion method;

[0102] Step 3: Export the screening and optimization layout results of various types of sensors and related analysis data. The screening and optimization results mainly include rod strain monitoring points (.xlsx), roof truss node displacement monitoring points (.xlsx), support displacement monitoring points (.xlsx) and acceleration monitoring points (.xlsx).

[0103] Step 2 includes four modules. The specific steps of each module are as follows:

[0104] (I) Steps for selecting rod strain monitoring points:

[0105] Step 1: Determine the file import and export paths and stress screening limit interaction parameters;

[0106] Step 2: Read the file "SectionName.txt" to obtain the section names of all components;

[0107] Step 3: Loop through the files Obtain stress data of various rods;

[0108] Step 4: Calculate the combined stress σ based on the table data using the following formula (1): con Among them, σ N is the axial stress; is the stress caused by the bending moment around the y-axis; is the stress caused by the bending moment about the z-axis.

[0109]

[0110] Step 5: Determine whether the combined stress of the rod exceeds the stress screening limit and obtain the exceeding limit rod;

[0111] Step 6: Screen the over-limit members with the largest combined stress, that is, keep the group of over-limit members with the largest combined stress at the same position;

[0112] Step 7: Save all screening results in Excel document format to the specified directory.

[0113] (II) Steps for selecting node deformation monitoring points:

[0114] Step 1: Determine the interaction parameters, including file import, export path and displacement screening limit interaction parameters;

[0115] Step 2: Read the file "roof truss node displacement data.xlsx" to obtain the displacement of all nodes;

[0116] Step 3: Substitute the displacement components U in the three directions x , U y , U z Sort the data from large to small;

[0117] Step 4: Determine whether the displacement of the node in each direction exceeds the corresponding displacement screening limit, and obtain the node with excessive displacement;

[0118] Step 5: Filter out the nodes with the largest displacement, that is, keep the group with the largest displacement in the same direction among the nodes with the same displacement;

[0119] Step 6: All screening results are saved in the Excel document format to the specified directory.

[0120] (III) Screening steps for support displacement monitoring points:

[0121] Step 1: Determine the interaction parameters, including file import and export paths, and support displacement screening limits in the x, y, z, and space directions;

[0122] Step 2: Read the file "roof truss support node displacement data.xlsx" to obtain the displacement of all support nodes;

[0123] Step 3: Calculate the spatial displacement U of each node according to formula (2): xyz Among them, U x , U y , U z is the displacement value of the node in three directions;

[0124]

[0125] Step 4: Connect the U x , U y , U z , U xyz The four components are sorted from large to small;

[0126] Step 5: Determine whether the displacement of the support node in each direction exceeds the corresponding displacement screening limit, and obtain the support displacement exceeding limit node;

[0127] Step 6: Filter out the over-limit nodes with the largest displacement, that is, retain the group with the largest displacement in the same direction among the over-limit nodes;

[0128] Step 7: Save all screening results in Excel document format to the specified directory.

[0129] (IV) Acceleration monitoring point optimization steps:

[0130] Step 1: Determine the file import and export paths, as well as the parameters of the particle swarm optimization algorithm, including: particle swarm size N, recommended value is 20-1000; number of iterations K, recommended value is 50-100; individual learning factor c1 and group learning factor c2, recommended value is [0, 4]; maximum inertia weight ω max With the minimum inertia weight ω min , the recommended value is [0.4, 2]; the maximum speed limit υ max With the minimum speed limit υ min , the recommended value is 10% to 20% of the particle variation range; the maximum position limit x max With the minimum position limit x min , default x max is the particle dimension D (the number of acceleration sensors planned to be arranged), x min is 1;

[0131] Step 2: Read the file "roof truss node various order modes.xlsx", transform and generate the modal information matrix in three directions ([U x ]、[U y ]、[U z ]);

[0132] Step 3: Execute the particle swarm optimization algorithm to obtain the optimal layout of the structural acceleration sensor. The steps of the particle swarm optimization algorithm are as follows:

[0133] Step 1): Read the particle swarm parameters and complete initialization;

[0134] Step 2): Randomly initialize the position and velocity of the particle swarm, that is, in the D-dimensional population, according to the position limit [x min ,x max ] and speed limit [v min ,v max ] Randomly select the position X of the nth particle nd and the initial value of speed V nd , which is expressed as follows:

[0135] Xnd ={x n1 ,x n2 ,......,x nD} (3)

[0136] V nd = {v n1 ,v n2 ,......,v nD} (4)

[0137] Step 3): Calculate the optimal P of the particle swarm individuals through the fitness function nd,pbest and the group optimal P d,gbest , which is expressed as follows:

[0138] P nd,pbest ={p n1 ,p n2 ,......,p nD} (5)

[0139] P d,gbest ={p 1,gbest ,p 2,gbest ,......,p D,gbest} (6)

[0140] Where n is the particle number, n = 1, 2, ..., N; d is the particle dimension number, d = 1, 2, ..., D; 1 ≤ j ≤ D;

[0141] The fitness function calculation steps are as follows:

[0142] Step ①: Read data information, i.e., modal matrix [ndof×nModel], modal order nModel, particle displacement matrix [D×N]; where ndof is the number of structural degrees of freedom in the modal matrix, and nModel is the modal order;

[0143] Step ②: Generate a new modal matrix [D×nModel];

[0144] Step 3: According to the following formula (7), solve the MAC matrix [nModel×nModel] of the new modal matrix;

[0145]

[0146] Among them, MAC ij is the element in the i-th row and j-th column of the MAC matrix; Φ i , Φ j Represent the modal vectors of the i-th and j-th order respectively.

[0147] Step ④: Select the maximum value max(MAC) of the non-main diagonal elements of the MAC matrix through loop traversal;

[0148] Step ⑤: Construct the fitness function Fitness according to the following formula (8), and iterate to obtain the fitness value;

[0149] Fitness=1-max(MAC) (8)

[0150] Step 4): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value. If not, update the particle position and velocity according to the following equations (9) and (10).

[0151]

[0152]

[0153] Among them, r1 and r2 are random numbers in the interval [0,1]; is the position vector of particle n in the dth dimension in the kth iteration; is the velocity vector of particle n in the dth dimension in the kth iteration; is the historical optimal position of particle n in the dth dimension in the kth iteration, that is, the optimal solution searched by the nth particle (individual) after the kth iteration; is the historical optimal position of the swarm in the dth dimension in the kth iteration, that is, the optimal solution in the entire particle swarm after the kth iteration; ω is the inertia weight coefficient, which is calculated as follows:

[0154]

[0155] Among them, k is the current iteration number.

[0156] Step 5): According to step 3), update the individual optimal position and fitness value, the group optimal position and fitness value;

[0157] Step 6): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value. If not, repeat steps 4) to 6) until the conditions are met.

[0158] Step 4: Save all optimization layout results in Excel document format to the specified directory.

[0159] The specific embodiments are as follows:

[0160] The design scale of a station building is 180m in length and 273m in maximum width. The material is Q355B steel, and the rods are all tubular sections. The structural model of the station building contains 5108 nodes, 13344 units, and 26 section types. The load conditions of the station building structure include constant (live) load, roof live load, wind load, snow load, temperature and earthquake action. The software calculates the stress cloud diagram of the beam unit (constant load + live load + roof live load), displacement cloud diagram (spatial displacement U xyz ), modal cloud diagram (first order), such as Figure 2 The analysis software integrating the technology of the present invention optimizes the layout of various types of monitoring points (bar strain, node displacement, support displacement, acceleration) of the roof. The schematic diagram of the arrangement of strain monitoring points, the schematic diagram of the arrangement of displacement monitoring points, the schematic diagram of the support displacement monitoring points and the schematic diagram of the acceleration monitoring points are shown in the figure below. Figure 3 , Figure 4 , Figure 5 , Figure 6 shown.

[0161] Comparison of monitoring point layout ( Figures 3 to 6 ) and software analysis results ( Figure 2 ), it can be seen that based on the method of the present invention, the sensor screening and optimal arrangement of the structural member strain, node deformation, support displacement, and structural acceleration of the railway station house frame can be realized. Figure 3 It can be seen that this method can not only determine the rods that need to be strain monitored, but also clarify the position of the strain monitoring points along the length of the rods and their specific coordinates on the cross section of the component, solving the problem of fuzzy strain monitoring points in the traditional layout method. Figure 6 It can be seen that the acceleration sensor layout method based on the particle swarm optimization algorithm and the principle of minimum modal confidence criterion improves the rationality and scientificity of the monitoring point layout, and makes up for the shortcomings of the sensor layout method dominated by experience. Based on the invention method, the screening and optimization layout of the measuring points can be realized quickly and accurately, avoiding the signal interference and cost waste caused by the redundant sensor layout, and maximizing the coverage and sensitivity of the structural response.

[0162] In the description of the present invention, unless otherwise clearly specified and limited, the terms "set", "install", "connect", "connect", and "fix" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0163] The content of the present invention is not limited to the embodiments listed, and any equivalent changes made to the technical solution of the present invention by ordinary technicians in this field after reading the description of the present invention are covered by the claims of the present invention.

Claims

1. A method for optimizing the arrangement of sensors for a railway passenger station roof structure, characterized in that: The steps include: Step 1: Export a file that can be directly read by MidasGen from the structural design software 3D3SDesign, use MidasGen to perform load condition analysis, and export the required data; Step 2: According to the code provisions and the analysis results of the Midas model, the sensors of member strain, node deformation, support displacement, and structural acceleration are screened and optimally arranged in the order of the program modules; Step 3: Export the screening and optimization results of various types of sensors and related analysis data; In step 1, the exported Midas Gen analysis data includes: 1) Section name of each component SectionName.txt; 2) Stress values ​​of each cross-section member under all load combination conditions; 3) Roof truss node displacement data.xlsx; 4) Roof truss support node displacement data.xlsx; 5) Each order mode of the roof truss node.xlsx; In the step 2, the step of screening the rod strain specifically includes: Step 1: Determine the interaction parameters, including file import and export paths and stress screening limits; Step 2: Read the file "SectionName.txt" to obtain the section names of all components; Step 3: Circularly read the stress values ​​of each cross-section component under all load combination conditions to obtain the stress data of various rods; Step 4: Calculate the combined stress σ of the member based on the table data according to the following formula (1): con ; Among them, σ N is the axial stress; is the stress caused by the bending moment around the y-axis; is the stress caused by the bending moment around the z-axis; Step 5: Determine whether the combined stress of the rod exceeds the stress screening limit and obtain the exceeding limit rod; Step 6: Screen the over-limit members with the largest combined stress, that is, keep the group of over-limit members with the largest combined stress at the same position; Step 7: Save all screening results in Excel document format to the specified directory; In the step 2, the node deformation screening step specifically includes: Step 1: Determine the interaction parameters, including file import and export paths and displacement screening limits; Step 2: Read the file "roof truss node displacement data.xlsx" to obtain the displacement of all nodes; Step 3: Substitute the displacement components U in the three directions x , U y , U z Sort the data from large to small; Step 4: Determine whether the displacement of the node in each direction exceeds the corresponding displacement screening limit, and obtain the displacement exceeding limit node; Step 5: Filter out the nodes with the largest displacement, that is, retain the group of nodes with the largest displacement in the same direction among the nodes with the same displacement; Step 6: All screening results are saved in the Excel document format to the specified directory; In the step 2, the step of screening the support deformation specifically includes: Step 1: Determine the interaction parameters, including file import and export paths, and support displacement screening limits in the x, y, z, and space directions; Step 2: Read the file "roof truss support node displacement data.xlsx" to obtain the displacement of all support nodes; Step 3: Calculate the spatial displacement U of each node according to formula (2): xyz , where U x , U y , U z is the displacement value of the node in three directions; Step 4: Connect the U x , U y , U z , U xyz The four components are sorted from large to small; Step 5: Determine whether the displacement of the support node in each direction exceeds the corresponding displacement screening limit, and obtain the support displacement exceeding limit node; Step 6: Filter out the over-limit nodes with the largest displacement, that is, retain the group with the largest displacement in the same direction among the over-limit nodes; Step 7: Save all screening results in Excel document format to the specified directory; In step 2, the screening and optimization step of the structural acceleration specifically includes: Step 1: Determine the file import and export paths, as well as the parameters of the particle swarm optimization algorithm; Step 2: Read the file "roof truss node each order mode.xlsx", and transform and generate the modal information matrix in three directions [U x ]、[U y ]、[U z ]; Step 3: Execute the particle swarm optimization algorithm to obtain the optimal layout of the structural acceleration sensor; Step 4: Save all optimization layout results in Excel document format to the specified directory.

2. The method for optimizing the arrangement of sensors for the roof truss structure of a railway passenger station according to claim 1, characterized in that: The parameters of the particle swarm optimization algorithm include: 1) Particle swarm size N; 2) The number of iterations K; 3) Particle dimension D, i.e. the number of acceleration sensors planned to be deployed; 4) Individual learning factor c1 and group learning factor c2; 5) Maximum inertia weight ω max With the minimum inertia weight ω min ; 6) Maximum speed limit υ max With the minimum speed limit υ min ; 7) Maximum position limit x max With the minimum position limit x min , default x max is the particle dimension D, x min is 1.

3. The method for optimizing the arrangement of sensors for the roof truss structure of a railway passenger station according to claim 1, characterized in that: The specific steps of the particle swarm optimization algorithm include: Step 1): Read the particle swarm parameters and complete initialization; Step 2): Randomly initialize the position and velocity of the particle swarm, that is, in the D-dimensional population, according to the position limit [x min ,x max ] and speed limit [v min ,v max ] Randomly select the position X of the nth particle nd and the initial value of speed V nd , which is expressed as follows: X nd ={x n1 ,x n2 ,......,x nD } (3) V nd ={v n1 ,v n2 ,......,v nD } (4) Step 3): Calculate the optimal P of the particle swarm through the fitness function nd,pbest and the group optimal P d,gbest , which is expressed as follows: P nd,pbest ={p n1 ,p n2 ,......,p nD } (5) P d,gbest ={p 1,gbest ,p 2,gbest ,......,p D,gbest } (6) Where n is the particle number, n = 1, 2, ..., N; d is the particle dimension number, d = 1, 2, ..., D; 1≤d≤D; Step 4): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value. If not, update the particle position and velocity according to the following equations (7) and (8). Among them, r1 and r2 are random numbers in the interval [0,1]; is the position vector of particle n in the dth dimension in the kth iteration; is the velocity vector of particle n in the dth dimension in the kth iteration; is the historical optimal position of particle n in the dth dimension in the kth iteration, that is, the optimal solution searched by the nth particle after the kth iteration; is the historical optimal position of the swarm in the dth dimension in the kth iteration, that is, the optimal solution in the entire particle swarm after the kth iteration; ω is the inertia weight coefficient, c1 is the individual learning factor, and c2 is the group learning factor; ω max is the maximum inertia weight, ω min is the minimum inertia weight; it is calculated as follows; Where k is the current iteration number; Step 5): According to step 3), update the individual optimal position and fitness value, the group optimal position and fitness value; Step 6): Determine whether the number of iterations K is reached. If so, exit the optimization process and output the optimal solution and fitness value; if not, repeat steps 3) to 5) until the conditions are met.

4. The method for optimizing the arrangement of sensors for the roof truss structure of a railway passenger station according to claim 3, characterized in that: The calculation step of the fitness function includes: Step ①: Read data information, i.e., modal matrix [ndof×nModel], modal order nModel, particle displacement matrix [D×N]; where ndof is the number of structural degrees of freedom in the modal matrix; nModel is the modal order; Step ②: Generate a new modal matrix [D×nModel]; Step 3: According to formula (10), solve the MAC matrix [nModel×nModel] of the new modal matrix; Among them, MAC ij is the element in the i-th row and j-th column of the MAC matrix; Φ i , Φ j Represent the modal vectors of the i-th and j-th order respectively; Step ④: Select the maximum value max(MAC) of the non-main diagonal elements of the MAC matrix through loop traversal; Step ⑤: Construct the fitness function Fitness according to formula (11), and iterate to obtain the optimal fitness value; Fitness = 1 - max(MAC) (11).