Method for predicting effective area of natural ventilation of railway station elevated waiting hall
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENYANG JIANZHU UNIVERSITY
- Filing Date
- 2026-04-10
- Publication Date
- 2026-07-10
Smart Images
Figure CN122365062A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of predicting the natural ventilation effect in railway station waiting halls, and more specifically, to a method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations. Background Technology
[0002] Elevated waiting halls in railway stations are densely populated, generating a large amount of polluted air from passenger metabolism. Mechanical ventilation consumes enormous amounts of energy, making natural ventilation even more crucial. Due to technological limitations and passenger flow considerations, many elevated waiting halls in recent years have been constructed as large, open spaces, making it difficult to optimize natural ventilation through adjustments to spatial geometry. Therefore, the key challenge in the design of natural ventilation for elevated waiting halls is to rationally utilize the openings within the hall and scientifically match the combination and relative positions of air inlets and outlets to ensure effective natural ventilation covering the activity area of passengers. Machine learning, represented by Artificial Neural Networks (ANNs), has gradually emerged as an important method for addressing these architectural challenges, demonstrating accuracy and high evaluation efficiency.
[0003] Therefore, we urgently need to develop a method for predicting the natural ventilation effect of elevated waiting halls in railway stations based on ANN. Summary of the Invention
[0004] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.
[0005] Therefore, the purpose of this invention is to propose a method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations.
[0006] To achieve the above objectives, the present invention provides a method for predicting the effective area of natural ventilation in an elevated waiting hall of a railway station. The elevated waiting hall is a rectangular structure, with its two short sides connected to corresponding entrance spaces. These entrance spaces are located on both sides of the railway line. The elevated waiting hall is situated above the railway station platform. The elevated waiting hall contains various waiting areas and passenger passageways between them. Other ancillary facilities, besides the waiting areas and passenger passageways, are arranged inside the elevated waiting hall or on the two long sides of its exterior walls. The elevated waiting hall has an axis. The linear spatial characteristics; the longitudinal natural ventilation of the elevated waiting hall of the railway station is limited by the entrance space, and the lateral natural ventilation of the elevated waiting hall of the railway station is related to the layout of other ancillary facilities. The prediction method includes: Step S1: Extracting multiple spatial patterns of the elevated waiting hall of the railway station; wherein, the multiple spatial patterns include: side interface unobstructed mode, upper side interface semi-obstructed mode, lower side interface semi-obstructed mode, and side interface fully obstructed mode; the side interface unobstructed mode is: dividing the outer wall of the elevated waiting hall of the railway station vertically from bottom to top into the first layer. The other ancillary facilities are arranged inside the elevated waiting hall of the railway station; the other ancillary facilities divide the elevated waiting hall of the railway station into multiple parts; the side windows are opened on the first floor and / or the second floor; the elevated waiting hall of the railway station is naturally ventilated through the side windows opened on the exterior wall; the upper part of the side interface is semi-obstructed in the following manner: the exterior wall of the elevated waiting hall of the railway station is divided into a first floor and a second floor vertically from bottom to top; the other ancillary facilities are arranged on the second floor; the side windows are opened on the first floor; the elevated waiting hall of the railway station is naturally ventilated through the side windows. The partial obstruction mode of the lower side interface is as follows: the outer wall of the elevated waiting hall of the railway station is divided vertically from bottom to top into a first layer and a second layer; other ancillary facilities are arranged on the first layer; the side windows are arranged on the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows; the full obstruction mode of the side interface is as follows: other ancillary facilities are arranged along the vertical length and height of the outer wall of the elevated waiting hall of the railway station; no side windows are opened on the outer wall of the elevated waiting hall of the railway station; and the side windows opened on the first layer are low side windows; the side windows opened on the second layer are high side windows;
[0007] Step S2: Based on the aforementioned multiple spatial models, a numerical simulation method using computational fluid dynamics is selected to construct m sets of natural ventilation simulation experimental models for the elevated waiting halls of the railway stations. The opening types of the natural ventilation simulation experimental models include one or a combination of the following: ground opening A, roof opening B, low-profile side window C on the windward side, high-profile side window D on the windward side, low-profile side window E on the leeward side, and high-profile side window F on the leeward side. The number of openings corresponding to ground opening A, roof opening B, low-profile side window C on the windward side, high-profile side window D on the windward side, low-profile side window E on the leeward side, and high-profile side window F is [missing information]. The horizontal distance between the ground opening A and the windward side of the outer wall of the elevated waiting hall of the railway station is L1; the horizontal distance between the roof opening B and the windward side of the outer wall of the elevated waiting hall of the railway station is L2; the height of the windward low-side window C from the ground of the elevated waiting hall of the railway station is H1; the height of the windward high-side window D from the ground of the elevated waiting hall of the railway station is H2; the height of the leeward low-side window E from the ground of the elevated waiting hall of the railway station is H3; the height of the leeward high-side window F from the ground of the elevated waiting hall of the railway station is H4; and the ground opening A is expressed as A. L1 The roof opening B is expressed as B. L2 The windward low-side window C is expressed in the form of C. H1 The windward high side window D is expressed in the form of D. H2 The leeward low-side window E is expressed as E H3 The leeward side window F is expressed as F H4 Step S3: Calculate the actual value of the effective area ratio of natural ventilation for m groups of natural ventilation simulation experimental models; Step S4: Obtain the dataset; wherein, the dataset includes: the opening type expression form corresponding to m groups of natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each group of natural ventilation simulation experimental models; m is a positive integer; Step S5: Preprocess the dataset; Step S6: Divide the preprocessed dataset into a training set and a test set; Step S7: Build a natural ventilation prediction model for elevated waiting halls of railway stations; wherein, the natural ventilation prediction model is an artificial neural network; the natural ventilation prediction... The input to the model is a training set or a test set; the output of the natural ventilation prediction model is the predicted value of the effective area ratio of natural ventilation corresponding to the training set or the test set; Step S8: Use the training set to train and optimize the natural ventilation prediction model, and save the final optimized natural ventilation prediction model; Step S9: Input the test set into the final optimized natural ventilation prediction model to obtain the predicted value of the effective area ratio of natural ventilation corresponding to the test set; Step S10: Evaluate the prediction performance of the natural ventilation prediction model based on the predicted value of the effective area ratio of natural ventilation corresponding to the test set and the actual value of the effective area ratio of natural ventilation corresponding to the test set.
[0008] Preferably, in step S2, the roof of the elevated waiting hall of the railway station corresponding to the natural ventilation simulation experimental model is a flat roof, and the dimensions of the elevated waiting hall of the railway station are length... Meters, width meters, height Meters; the ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F are all square in shape; the area of each individual opening corresponding to the ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F is... square meters; , , and All are positive numbers.
[0009] Preferably, It is 200; It is 100; It is 20; =1; It is 14.
[0010] Preferably, in step S2, the values of L1 and L2 are both in the range of 10-90 m, and the interval between the values of L1 and L2 is 10 m; the values of H1 and H3 are both in the range of 1-9 m, the values of H2 and H4 are both in the range of 10-18 m, and the interval between the values of H1, H3, H2 and H4 is 1 m.
[0011] Preferably, in step S3, the actual value of the effective area ratio of natural ventilation is: the ratio of the effective area of natural ventilation in the area where the height of people in the elevated waiting hall of the railway station is 1.5m to the ground area of the elevated waiting hall of the railway station; the effective area of natural ventilation is: the area of the area where the wind speed is higher than 0.1m / s in the area where the height of people in the elevated waiting hall of the railway station is 1.5m.
[0012] Preferably, step S3 specifically involves: calculating the actual value of the effective area ratio of natural ventilation in the natural ventilation simulation experiment model by using the wind speed cloud map of the elevated waiting hall of the railway station at a height of 1.5m.
[0013] Preferably, step S6 specifically involves: using the Latin cube sampling method to divide the preprocessed dataset into a training set and a test set.
[0014] Preferably, m is 475; the training set data includes: the opening type expression forms corresponding to 329 natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each natural ventilation simulation experimental model; the test set data includes: the opening type expression forms corresponding to 146 natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each natural ventilation simulation experimental model.
[0015] Preferably, the artificial neural network is composed of a multilayer perceptron; the multilayer perceptron is composed of an input layer, a hidden layer and an output layer; the hidden layer is located between the input layer and the output layer.
[0016] The beneficial effects of this invention are:
[0017] (1) The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations provided by the present invention takes the performance of natural ventilation as the target and establishes a natural ventilation prediction model applicable to elevated waiting halls of railway stations. By combining CFD simulation and machine learning, the method can efficiently predict the proportion of effective area of natural ventilation in elevated waiting halls of railway stations under various opening types and location combinations, which significantly improves the prediction accuracy and efficiency. By integrating interface opening design parameters, the natural ventilation prediction model of elevated waiting halls of railway stations can predict the proportion of effective area of natural ventilation in real time, thereby guiding the refined design of opening type, opening location and opening opening and closing strategy.
[0018] (2) The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations provided by the present invention can quickly evaluate the natural ventilation performance of different opening combinations by inputting opening design parameters (i.e., opening type and opening location). Specifically, the predicted value of the effective area ratio of natural ventilation output by the natural ventilation prediction model of the elevated waiting hall of railway stations can be used to assist in the construction of the natural ventilation system of the elevated waiting hall of railway stations and improve the reliability of natural ventilation in the elevated waiting hall of railway stations. The purpose of the natural ventilation prediction model of the elevated waiting hall of railway stations is to establish a relationship between the independent variables (opening type and opening location) and the dependent variables (the effective area ratio of natural ventilation corresponding to each group of opening types and opening locations), provide a scientific quantitative tool for the natural ventilation design of elevated waiting halls of railway stations, and support multi-objective optimization decision-making in the scheme stage.
[0019] (3) The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations provided by this invention, in a macro sense, realizes the transformation from experience-driven to data-driven design, and promotes the low-carbon and intelligent development of green transportation hubs. It provides quantifiable performance feedback for designers of elevated waiting halls of railway stations, so that the natural ventilation strategy of elevated waiting halls of railway stations can be evaluated and optimized in the early stage of the scheme.
[0020] Additional aspects and advantages of the invention will become apparent from the description which follows, or may be learned by practice of the invention. Attached Figure Description
[0021] Figure 1 A flowchart illustrating a method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to an embodiment of the present invention is shown.
[0022] Figure 2 A schematic diagram of the spatial structure of an elevated waiting hall in a railway station according to an embodiment of the present invention is shown;
[0023] Figure 3 A schematic diagram of the unobstructed side interface of an elevated waiting hall in a railway station according to an embodiment of the present invention is shown.
[0024] Figure 4 A schematic diagram of the upper semi-obstructed side interface of an elevated waiting hall in a railway station according to an embodiment of the present invention is shown.
[0025] Figure 5 A schematic diagram of the lower half-obstruction mode of the side interface of an elevated waiting hall in a railway station according to an embodiment of the present invention is shown.
[0026] Figure 6 This invention illustrates a structural schematic diagram of a fully obscured side interface of an elevated waiting hall in a railway station according to an embodiment of the present invention.
[0027] Figure 7 A schematic diagram of a fluid computational domain according to an embodiment of the present invention is shown;
[0028] Figure 8 The diagram illustrates the opening variable settings for ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F, respectively, according to an embodiment of the present invention.
[0029] Figure 9 A schematic table illustrating the number of corresponding groups in the training set and test set in sample data (i.e., dataset) according to an embodiment of the present invention is shown.
[0030] Figure 10 A schematic table showing a summary of a natural ventilation prediction model for an elevated waiting hall in a railway station according to an embodiment of the present invention (i.e., including the following: average sum of errors and relative errors obtained after training the model using the training set, and average sum of errors and relative errors obtained after verifying the model performance using the test set);
[0031] Figure 11The diagram shows the data point distribution corresponding to the fitting relationship between the predicted value of the effective area ratio of natural ventilation and the actual value of the effective area ratio of natural ventilation in the test set according to an embodiment of the present invention.
[0032] Figure 12 A schematic table illustrating the importance analysis of independent variables according to an embodiment of the present invention is shown. Detailed Implementation
[0033] To better understand the above-mentioned objects, features, and advantages of the present invention, such as Figures 1 to 12 As shown in the accompanying drawings and specific embodiments, the present invention will be further described in detail below. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.
[0034] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.
[0035] Figure 1 A flowchart illustrating a method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations, according to an embodiment of the present invention, is shown. Figure 2 As shown, the elevated waiting hall of the railway station has a rectangular structure, with its two shorter sides connected to corresponding entrance spaces. The entrance spaces are located on both sides of the railway line. The elevated waiting hall is situated above the railway station platform. Various waiting areas and passenger passageways are arranged within the elevated waiting hall. Other ancillary facilities, besides the waiting areas and passenger passageways, are located inside the elevated waiting hall or on the two longer sides of its exterior walls. The elevated waiting hall has an axial spatial characteristic. The longitudinal (longer direction) natural ventilation of the elevated waiting hall is limited by the entrance spaces, while the lateral (shorter direction) natural ventilation is related to the layout of the other ancillary facilities. Figure 1 As shown, the method for predicting the effective area of natural ventilation in the elevated waiting hall of the railway station includes:
[0036] Step S1: Extract multiple spatial modes of the elevated waiting hall of the railway station; wherein, the multiple spatial modes include: side interface unobstructed mode, upper side interface semi-obstructed mode, lower side interface semi-obstructed mode, and side interface fully obstructed mode.
[0037] The unobstructed side interface mode is as follows: the outer wall of the elevated waiting hall of the railway station is vertically divided into a first layer and a second layer from bottom to top; the other ancillary facilities are arranged inside the elevated waiting hall of the railway station; the other ancillary facilities divide the elevated waiting hall of the railway station into multiple parts; the side windows are opened on the first layer and / or the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows opened on the outer wall;
[0038] The upper semi-obstruction mode of the side interface is as follows: the outer wall of the elevated waiting hall of the railway station is divided into a first layer and a second layer vertically from bottom to top; the other ancillary facilities are arranged on the second layer; the side windows are opened on the first layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows.
[0039] The lower semi-obstructed mode of the side interface is as follows: the outer wall of the elevated waiting hall of the railway station is divided into a first layer and a second layer vertically from bottom to top; the other ancillary facilities are arranged on the first layer; the side windows are arranged on the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows;
[0040] The side interface full-coverage mode is as follows: the other ancillary facilities are arranged along the vertical direction of the outer wall of the elevated waiting hall of the railway station in a continuous length and height; no side windows are opened on the outer wall of the elevated waiting hall of the railway station; and the side windows opened on the first floor are low side windows; the side windows opened on the second floor are high side windows.
[0041] Step S2: Based on multiple spatial models, the numerical simulation method of computational fluid dynamics is selected to build m sets of natural ventilation simulation experimental models for elevated waiting halls of railway stations; the opening types of the natural ventilation simulation experimental models include one or a combination of the following: ground opening A, roof opening B, low side window on the windward side C, high side window on the windward side D, low side window on the leeward side E, and high side window on the leeward side F.
[0042] The number of openings corresponding to ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F are all... indivual;
[0043] The horizontal distance between the ground opening A and the windward side of the outer wall of the elevated waiting hall of the railway station is L1, and the horizontal distance between the roof opening B and the windward side of the outer wall of the elevated waiting hall of the railway station is L2.
[0044] The windward low-side window C is at a height of H1 from the ground of the elevated waiting hall of the railway station; the windward high-side window D is at a height of H2 from the ground of the elevated waiting hall of the railway station; the leeward low-side window E is at a height of H3 from the ground of the elevated waiting hall of the railway station; and the leeward high-side window F is at a height of H4 from the ground of the elevated waiting hall of the railway station.
[0045] And the ground opening A is expressed as A L1 The roof opening B is expressed as B. L2 The windward low-side window C is expressed in the form of C. H1 The windward high side window D is expressed in the form of D. H2 The leeward low-side window E is expressed as E H3 The leeward side window F is expressed as F H4 ;
[0046] Step S3: Calculate the actual value of the effective area ratio of natural ventilation for m groups of natural ventilation simulation experimental models; specifically, the numerical simulation method of Computational Fluid Dynamics (CFD) is selected to calculate the actual value of the effective area ratio of natural ventilation for m groups of natural ventilation simulation experimental models.
[0047] Step S4: Obtain the dataset; wherein the dataset includes: the opening type expression form corresponding to m groups of natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each group of natural ventilation simulation experimental models; m is a positive integer;
[0048] Step S5: Preprocess the dataset;
[0049] Step S6: Divide the preprocessed dataset into a training set and a test set;
[0050] Step S7: Build a natural ventilation prediction model for elevated waiting halls in railway stations; wherein, the natural ventilation prediction model is an artificial neural network; the input of the natural ventilation prediction model is a training set or a test set; the output of the natural ventilation prediction model is the predicted value of the effective area ratio of natural ventilation corresponding to the training set or the test set.
[0051] Step S8: Train and optimize the natural ventilation prediction model using the training set, and save the final optimized natural ventilation prediction model;
[0052] Step S9: Input the test set into the final optimized natural ventilation prediction model to obtain the predicted value of the effective area ratio of natural ventilation corresponding to the test set;
[0053] Step S10: Evaluate the predictive performance of the natural ventilation prediction model based on the predicted value of the effective area ratio of natural ventilation in the test set and the actual value of the effective area ratio of natural ventilation in the test set.
[0054] In this embodiment, the method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations provided by the present invention takes natural ventilation performance as the objective guide, establishes a natural ventilation prediction model applicable to elevated waiting halls of railway stations, and uses a combination of CFD simulation and machine learning to efficiently predict the proportion of the effective area of natural ventilation in elevated waiting halls of railway stations under various combinations of opening types and locations, significantly improving prediction accuracy and efficiency; by integrating interface opening design parameters, the natural ventilation prediction model of elevated waiting halls of railway stations can predict the proportion of the effective area of natural ventilation in real time, thereby guiding the refined design of opening types, opening locations and opening opening and closing strategies;
[0055] In this embodiment, the natural ventilation prediction model for elevated waiting halls in railway stations is fitted using a training set. Specifically, the model is trained using the input data (opening locations and types) and the target value (actual value of the effective area ratio for natural ventilation) from the training set. This process is based on the backpropagation algorithm, which minimizes the loss function by continuously adjusting the weights and biases in the model, thereby enabling it to accurately fit the training set data. The test set data is not used in training and is used to measure the difference between the predicted and actual effective area ratios for natural ventilation output by the model.
[0056] The following specific embodiment illustrates the method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to the present invention. The elevated waiting hall is a cuboid structure, with its two short sides connected to corresponding entrance spaces. These entrance spaces are located on both sides of the railway line. The elevated waiting hall is situated above the railway station platform. Various waiting areas and passenger passageways are arranged within the elevated waiting hall. Other ancillary facilities, besides the waiting areas and passenger passageways, are arranged inside the elevated waiting hall or on the two long sides of its exterior walls. The elevated waiting hall has an axial spatial characteristic. The longitudinal (longer direction) natural ventilation of the elevated waiting hall is limited by the entrance spaces, while the lateral (shorter direction) natural ventilation is related to the layout of these other ancillary facilities. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations is implemented through the following steps:
[0057] Step S1: Extract multiple spatial modes of the elevated waiting hall of the railway station; among which, the multiple spatial modes include: side interface unobstructed mode, upper side interface semi-obstructed mode, lower side interface semi-obstructed mode and side interface fully obstructed mode.
[0058] The unobstructed side facade design involves dividing the exterior wall of the elevated waiting hall of the railway station vertically into a first layer and a second layer from bottom to top; other ancillary facilities are arranged inside the elevated waiting hall; these ancillary facilities divide the elevated waiting hall into multiple parts; the side windows are located on the first layer and / or the second layer; the elevated waiting hall is naturally ventilated through the side windows on the exterior wall. The semi-obstructed upper side facade design involves dividing the exterior wall of the elevated waiting hall vertically into a first layer and a second layer from bottom to top; other ancillary facilities are arranged on the second layer; the side windows are located on the first layer. The elevated waiting hall of the railway station is naturally ventilated through the side windows; the lower part of the side interface is partially obscured as follows: the outer wall of the elevated waiting hall is vertically divided into a first layer and a second layer from bottom to top; the other ancillary facilities are arranged on the first layer; the side windows are arranged on the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows; the side interface is fully obscured as follows: the other ancillary facilities are arranged along the entire length and height of the outer wall of the elevated waiting hall; no side windows are opened on the outer wall of the elevated waiting hall; and the side windows opened on the first layer are low side windows; the side windows opened on the second layer are high side windows.
[0059] In step S1, as Figure 2 As shown, in the elevated waiting hall mode of railway stations, the station building has entrance halls on both sides of the railway line, and the waiting halls are elevated above the platform level. The waiting halls are mostly single spaces with large spans and high ceilings. The interior of the waiting hall is mainly divided into waiting areas for different train numbers, with passenger passages from the entrance hall to the ticket gates interspersed within. Other ancillary facilities such as shops, restaurants, and restrooms are also centrally located.
[0060] Elevated waiting halls in railway stations typically feature a large, axial space. Longitudinal natural ventilation is limited by the entrance space (i.e., the entrance hall), while lateral natural ventilation is closely related to the layout of other ancillary facilities. Based on the location of these facilities and their impact on natural ventilation, railway station waiting halls can be categorized into four typical spatial patterns. For example... Figure 3 As shown, one category is other ancillary facilities ( Figure 3 The gray blocks (as depicted in the image) are typically single-layered blocks within the large spaces of railway station waiting halls, dividing the waiting hall into multiple sections. The exterior walls can have high and low side windows for natural ventilation, creating an unobstructed side facade. For example... Figure 4 As shown, the second category is other ancillary facilities ( Figure 4 The gray blocks (in the image) are arranged only along the two-story longitudinal exterior wall of the railway station waiting hall, allowing for natural ventilation through low side windows on the first-floor exterior wall, creating a semi-obscured upper side facade; (e.g.) Figure 5 As shown, the third category is other ancillary facilities ( Figure 5The gray blocks (in the image) are arranged only along the first floor of the longitudinal exterior wall of the railway station waiting hall, allowing for natural ventilation via high side windows on the second floor exterior wall, creating a semi-obscured lower side facade; (e.g.) Figure 6 As shown, the fourth category consists of other ancillary facilities that are set along the longitudinal exterior walls of the railway station waiting hall, which have the greatest impact on the natural ventilation of the railway station waiting hall. Side windows cannot be opened on the longitudinal exterior walls on both sides, resulting in a full occlusion mode of the side interface.
[0061] II. Step S2: Based on multiple spatial models, a numerical simulation method using computational fluid dynamics is selected to build m sets of natural ventilation simulation experimental models for elevated waiting halls of railway stations. The opening types in the natural ventilation simulation experimental models include one or a combination of the following: ground opening A, roof opening B, low side window on the windward side C, high side window on the windward side D, low side window on the leeward side E, and high side window on the leeward side F. The number of openings corresponding to ground opening A, roof opening B, low side window on the windward side C, high side window on the windward side D, low side window on the leeward side E, and high side window on the leeward side F is [missing information]. The horizontal distance between the ground opening A and the windward side of the outer wall of the elevated waiting hall of the railway station is L1; the horizontal distance between the roof opening B and the windward side of the outer wall of the elevated waiting hall of the railway station is L2; the height of the windward low-side window C from the ground of the elevated waiting hall of the railway station is H1; the height of the windward high-side window D from the ground of the elevated waiting hall of the railway station is H2; the height of the leeward low-side window E from the ground of the elevated waiting hall of the railway station is H3; the height of the leeward high-side window F from the ground of the elevated waiting hall of the railway station is H4; and the ground opening A is expressed as A. L1 The roof opening B is expressed as B. L2 The windward low-side window C is expressed in the form of C. H1 The windward high side window D is expressed in the form of D. H2 The leeward low-side window E is expressed as E H3 The leeward side window F is expressed as F H4 ;
[0062] The natural ventilation simulation model corresponds to a flat roof in the elevated waiting hall of a railway station, and the dimensions of the elevated waiting hall are [length missing]. Meters, width meters, height Meters; the ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F are all square in shape; the area of each individual opening corresponding to the ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F is... square meters; It is 200; It is 100; It is 20; =1;
[0063] The number of openings is 14. There are 14 openings for each type: ground opening A, roof opening B, low side window on the windward side C, high side window on the windward side D, low side window on the leeward side E, and high side window on the leeward side F. For example, C6B... 90 There are 14 low-lying side windows (C) facing the wind and 14 roof openings (B). If it includes three types of openings, there are 14 openings of each type. This number of 14 is based on extensive surveys and design specifications for railway stations.
[0064] The values of L1 and L2 are both in the range of 10-90 m, and the interval between the values of L1 and L2 is 10 m; the values of H1 and H3 are both in the range of 1-9 m, the values of H2 and H4 are both in the range of 10-18 m, and the interval between the values of H1, H3, H2 and H4 is 1 m.
[0065] In step S2, a survey and analysis of elevated waiting halls in Chinese railway stations was first conducted. Based on a survey of 67 railway stations in 11 provinces and municipalities in northern my country, there are 59 large railway stations, of which 81.4% are elevated waiting halls. 95.8% of the elevated waiting halls have a rectangular floor plan, with spatial dimensions typically around 200m long, 100m wide, and 20m high. 94.0% of the elevated waiting halls have flat roofs, often with skylights. Regarding side openings in the waiting halls, the "Code for Design of Railway Passenger Station Buildings" and related standards do not explicitly specify the area of natural ventilation openings. Field surveys show that there are generally 1-2 low and high side windows between the ticket gates in the waiting hall, with each opening area between 0.5-0.8 square meters. The "Code for Design of Railway Passenger Station Buildings" requires that large stations with a maximum capacity of ≥8000 people should have 28 ticket gates on both sides. Based on this, the baseline area of the high side windows, low side windows, top windows, and ground openings in the basic experimental model of the elevated waiting hall was set at 1m², with a total of 14 openings evenly distributed on each ticket gate side. Preliminary simulation experiments revealed that, with the same opening area, compared to horizontal and vertical vents, square vents achieve a more uniform indoor airflow distribution, reduce local vortices, and contribute to an overall improvement in the ventilation efficiency of the waiting hall.
[0066] Based on the above survey, statistics, and analysis, a natural ventilation simulation experimental model for elevated waiting halls in railway stations was constructed: the station building and the elevated waiting hall have a rectangular floor plan, the roof of the elevated waiting hall is a flat roof, and the dimensions of the elevated waiting hall are 200m long, 100m wide, and 20m high. The side surfaces can have low and high side windows, the top surface can have skylights, and the bottom surface can have ground-level openings. The openings are square, with an area of 1m² per opening, and 14 openings of each type are evenly distributed along the long axis of the waiting hall.
[0067] In step S2, the basic setup for the simulation experiment corresponding to the natural ventilation simulation model of the elevated waiting hall of the railway station is carried out.
[0068] (1) Simulation Experiment Platform: The numerical simulation method of Computational Fluid Dynamics (CFD) was selected. A parameterized experimental model and its external computational domain were established through the Geometry preprocessor. A full hexahedral mesh was established by calling Mesh, and the mesh was gradually refined from the external computational domain to the internal mesh of the waiting hall. Mesh independence verification was performed to ensure that the mesh accuracy of the key areas at the openings met the computational specifications. The mesh was transferred to the solver Fluent, and boundary conditions and simulation settings were set in Fluent, including the temperature and heat transfer coefficient of the enclosure structure, the wind speed and temperature of the air inlet, the return temperature of the air outlet, and the heat generated by the people in the room. Experimental variables were input and assigned values, and the equations were solved iteratively. When the residual of the equation is less than one-thousandth, the simulation results are considered to be closest to the steady state. At this time, the CFD Post postprocessor was called for post-processing to extract the natural ventilation evaluation index of the waiting hall reference surface, which is used to describe and judge the natural ventilation effect of the waiting hall.
[0069] (2) Simulation Experiment Boundary Conditions: ① Setting of the Computational Domain. To avoid flow field distortion due to an excessively small simulation area and unnecessary computational load due to an excessively large simulation area, in accordance with the provisions of the "Standard for Calculation of Green Performance of Civil Buildings" (JGJ / T449-2018), the calculation domain for this simulation experiment is set as follows: the height of the waiting hall is 'a', the vertical distance from the top of the building to the upper boundary of the calculation domain is 6.6a, the horizontal distance from the outer edge of the building on the outflow side to the side boundary of the calculation domain is 13.2a, and the horizontal distance from the outer edge of the building in other directions to the side boundary of the calculation domain is 6.6a (e.g., ...). Figure 7 (As shown). Calculations show that the blockage rate in the incoming flow direction is 2.94%, which meets the requirement of a blockage rate less than 3%. ② Setting of wind boundary conditions. The average wind speed of 3 m / s during the transitional season in Shenyang, a typical cold-region city, is used as the reference wind speed at 10 m above the ground; railway stations are mostly located in urban environments, and the ground roughness index is taken as 0.22. The average air temperature is taken as 22℃, and the kinematic viscosity at one atmosphere is... Reynolds number It belongs to fully turbulent flow, based on the Reynolds-averaged Navier-Stokes equations (RANS) model, and adopts... A Realizable turbulence model was used for CFD simulation. Air was assumed to be an incompressible ideal gas, and the Boussinesq approximation was applied. Temperature-induced changes in airflow density were considered, and a pressure-velocity coupling method was used to solve the steady-state airflow field. ③ Thermal boundary conditions were set. Based on survey and measurement results, the inlet air temperature, the interior wall temperature of the building, and the interior roof temperature were simplified and set to constant values of 22℃, 28℃, and 30℃, respectively. According to the classification of labor intensity in the "Air Conditioning Design Manual," passengers are considered to be engaged in very light physical labor, with a human body heat output of 134W / person at 26℃. The maximum number of people gathered in the waiting hall was calculated to be 10,000, assuming a uniform distribution of personnel.
[0070] In step S2, the variables for the simulation experiment corresponding to the natural ventilation simulation model of the elevated waiting hall of the railway station are set again.
[0071] (1) Independent variables in the simulation experiment: ① Opening variables. This study selected ground opening A, roof opening B, low side window C on the windward side and high side window D on the windward side, low side window E on the leeward side and high side window F on the leeward side as independent variables for the openings of the waiting hall interface. For example Figure 8 As shown, the variables for openings A and B are their horizontal distances from the windward sidewall, L1 and L2 respectively, ranging from 10 to 90 meters with a 10-meter interval. The variables for openings C, D, E, and F are their heights from the ground (the floor of the elevated waiting hall in the railway station), H1, H2, H3, and H4 respectively; C and E range from 1 to 9 meters, and D and F range from 10 to 18 meters with a 1-meter interval. Each opening's independent variable is expressed as a combination of the opening type and the values of H and L, such as A... 30 This indicates a ground opening, 30m horizontally from the windward side, F 12This indicates a high-side window on the leeward side, 12m above the ground. ② Opening combination independent variable. Theoretically, there are three types of opening combinations in elevated waiting halls: First, two-opening combinations. There are 15 combinations of any two openings on the waiting hall interface. After removing invalid CD and EF combinations on the same interface, 13 combinations of two openings are entered into the simulation experiment: CA, CB, CE, CF, DA, DB, DE, DF, AB, AE, AF, BE, and BF. In the actual simulation experiment, these 13 opening combinations are simulated parametrically one by one. Second, three-opening combinations. Any three openings on the waiting hall interface can form 20 combinations. To reduce the workload of the simulation experiment and directly approach the optimal solution, we adopt the optimal combination of two openings under three typical spatial modes: unobstructed side interface, high side semi-obstruction, and low side semi-obstruction. We introduce a third opening as an independent variable for the three-opening combination to explore the impact of the third opening on natural ventilation. Third, four-opening combinations. Any four openings on the interface can form 15 combinations. We used an optimal combination of three openings as a basis, introducing a fourth opening as an independent variable in the four-opening combination to explore the impact of the fourth opening on natural ventilation. Independent variables for two or more opening combinations are expressed as combinations of the opening variable and its H and L values, such as C6F. 18 The independent variable represents the combination of openings consisting of a low side window (6m above the ground on the windward side) and a high side window (18m above the ground on the leeward side).
[0072] In this specific implementation, the natural ventilation prediction model for elevated waiting halls in railway stations only predicts two types of opening combinations, not combinations with three or four openings. The 475 sets of experiments input into the natural ventilation prediction model for elevated waiting halls in railway stations consist of data from 13 combinations: CA, CB, CE, CF, DA, DB, DE, DF, AB, AE, AF, BE, and BF.
[0073] III. Step S3: Calculate the actual value of the effective area ratio of natural ventilation for the m groups of natural ventilation simulation experimental models. In Step S3, the actual value of the effective area ratio of natural ventilation is: the ratio of the effective area of natural ventilation in the area where the height of passengers in the elevated waiting hall of the railway station is 1.5m to the ground area of the elevated waiting hall of the railway station; the effective area of natural ventilation is: the area of the area where the wind speed is higher than 0.1m / s in the area where the height of passengers in the elevated waiting hall of the railway station is 1.5m. Specifically, Step S3 involves: calculating the actual value of the effective area ratio of natural ventilation for the natural ventilation simulation experimental model using the wind speed cloud map at the height of passengers in the elevated waiting hall of the railway station at 1.5m.
[0074] In step S3, the dependent variable is set for the simulation experiment corresponding to the natural ventilation simulation model of the elevated waiting hall of the railway station. Specifically, this specific embodiment mainly addresses the problems of uneven indoor air environment and insufficient effective natural ventilation area in the elevated waiting hall. Therefore, the proportion of effective natural ventilation area at a height of 1.5m is selected as the evaluation index, i.e., the dependent variable of the simulation experiment.
[0075] Among them, the effective area ratio of natural ventilation is the actual value, that is, the ratio of the ground area that meets the specified wind speed requirements to the area of the waiting hall. At present, there is no clear regulation on the wind speed limit for natural ventilation in railway station buildings. Among the relevant domestic and foreign standards, ISO7730 stipulates that the summer comfortable wind speed is 0.1~0.3m / s, ASHRAE55-2017 stipulates that the minimum indoor healthy wind speed is 0.15m / s, WELL requires that the instantaneous wind speed in the work area be ≤0.3m / s, the "Code for Design of Heating, Ventilation and Air Conditioning of Civil Buildings" (GB 50736-2012) stipulates that the wind speed should not be greater than 0.5m / s under the cooling condition in the short-stay area
[11] , the "Standard for Evaluation of Indoor Thermal and Humid Environment of Civil Buildings" (GB / T 50785-2012) requires that the minimum wind speed in the space without artificial cold and heat sources be 0.1m / s.
[0076] Based on the above standards, the effective wind speed for natural ventilation in the waiting hall's activity area was set to be above 0.1 m / s to ensure a balance between ventilation effectiveness and human comfort. After simulation, CFD-Post was used for post-processing to export a wind speed cloud map at a height of 1.5 m in the waiting hall. Areas with wind speeds above 0.1 m / s were defined as the effective natural ventilation area. Image analysis software was used to statistically analyze the percentage of areas with wind speeds above 0.1 m / s, representing the actual percentage of the effective natural ventilation area.
[0077] IV. Step S4: Obtain the dataset; wherein, the dataset includes: the opening type expression form corresponding to m groups of natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each group of natural ventilation simulation experimental models; m is 475.
[0078] 5. Step S5: Preprocess the dataset.
[0079] VI. Step S6: The preprocessed dataset is divided into a training set and a test set using the Latin cube sampling method. The training set includes: the opening type representations for 329 natural ventilation simulation experimental models and the actual value of the effective natural ventilation area ratio calculated for each model; the test set includes: the opening type representations for 146 natural ventilation simulation experimental models and the actual value of the effective natural ventilation area ratio calculated for each model.
[0080] VII. Step S7: Construct a natural ventilation prediction model for elevated waiting halls in railway stations; wherein the natural ventilation prediction model is an artificial neural network; the artificial neural network consists of a multilayer perceptron; the multilayer perceptron consists of an input layer, a hidden layer, and an output layer; the hidden layer is located between the input layer and the output layer. The input of the natural ventilation prediction model is a training set or a test set; the output of the natural ventilation prediction model is the predicted value of the effective area ratio of natural ventilation corresponding to the training set or the test set.
[0081] 8. Step S8: Use the training set to train and optimize the natural ventilation prediction model, and save the final optimized natural ventilation prediction model.
[0082] 9. Step S9: Input the test set into the final optimized natural ventilation prediction model to obtain the predicted value of the effective area ratio of natural ventilation corresponding to the test set.
[0083] 10. Step S10: Evaluate the predictive performance of the natural ventilation prediction model based on the predicted value of the effective area ratio of natural ventilation corresponding to the test set and the actual value of the effective area ratio of natural ventilation corresponding to the test set.
[0084] In steps S4 to S10 above, since the impact of openings on ventilation is non-linear, linear regression analysis cannot reveal its complete internal logic. Therefore, we choose an Artificial Neural Network (ANN) and train it with a Multilayer Perceptron (MLP) structure using the backpropagation algorithm. By approximating the complex input-output relationship through multilayer non-linear mapping, we effectively capture the deep correlation between opening layout and ventilation efficiency.
[0085] Furthermore, a natural ventilation prediction model for elevated waiting halls in railway stations is trained to accurately predict opening types, relative distances, and ventilation performance. The opening types and locations are digitally reduced in dimensionality. The input layer includes six types of openings (A, B, C, D, E, and F) and their horizontal spacings L1 and L2, as well as vertical height differences H1, H2, H3, and H4. The output layer corresponds to the percentage of effective ventilation area at a height of 1.5m.
[0086] like Figure 9As shown, 479 sets of high-precision sample data (i.e., datasets) were generated through CFD simulation experiments, covering different opening combinations. After standardization, the data was used to train and evaluate the predictive performance of the artificial neural network. The dataset was divided into a training set (69.3%, 329 sets) and a test set (30.7%, 146 sets) using Latin cube sampling. The training set was used to train and optimize the natural ventilation prediction model, fit the patterns between data, and determine the model's weights and other parameters. The test set data was not used for training but was used to evaluate the generalization ability of the natural ventilation prediction model and assess its predictive performance. The mean error and relative error are commonly used to evaluate the prediction quality of the artificial neural network. The test error characterizes the generalization ability of the ANN; a smaller test error indicates better training quality.
[0087] like Figure 10 As shown, the natural ventilation prediction model was trained using the training set, with an average error of 2.119 and a relative error of 0.148. When the prediction performance of the natural ventilation prediction model was tested using the test set, the average error was 1.337 and the relative error was 0.226. The test error of the ANN in this specific embodiment meets the accuracy requirements.
[0088] Cross-validation by dividing the dataset into training and test sets shows that artificial neural networks (ANNs) meet engineering accuracy requirements. Figure 11 As shown, the data point distribution plot corresponding to the fitting relationship between the predicted and actual values of the effective area ratio of natural ventilation in the test set is used to evaluate the prediction effect of the natural ventilation prediction model. It is found that the data points are basically distributed along the 45° line, indicating that the predicted values are highly close to the actual values. The residual values are randomly distributed near the 0 line without obvious patterns, indicating that the natural ventilation prediction model has a good fit. The residuals meet the assumptions of independence and normality, further verifying that the natural ventilation prediction model has high prediction accuracy and stability.
[0089] Furthermore, in this specific embodiment, we conducted an importance analysis of the independent variables (including: relative vertical distance between the air inlet and outlet, relative horizontal distance between the air inlet and outlet, air inlet type, and air outlet type), such as... Figure 12 As shown, the relative vertical distance between the air inlet and outlet contributes the most to ventilation performance (100%), followed by the air inlet category at 99.2%, the relative horizontal distance between the air inlet and outlet at 86.0%, and the air outlet category at 69.4%. This provides a reliable quantitative basis for optimizing natural ventilation design.
[0090] In this specific embodiment, the purpose of this natural ventilation prediction model is to establish a relationship between independent and dependent variables, providing a scientific quantitative tool for the natural ventilation design of elevated waiting halls in railway stations, and supporting multi-objective optimization decisions during the design phase. This natural ventilation prediction model can quickly respond to changes in ventilation performance under different opening combinations. Designers can predict ventilation effects using the model, and adjust opening types and relative positions based on design and ventilation objectives, effectively supporting the dynamic optimization and decision-making of natural ventilation schemes for elevated waiting halls in railway stations. Based on this natural ventilation prediction model, the design team can dynamically adjust the opening configuration of elevated waiting halls in railway stations, achieving real-time prediction and optimization of ventilation performance. By inputting different combinations of opening types and positions, the system outputs the predicted effective ventilation area ratio under the corresponding working conditions within 3 seconds, significantly improving the efficiency of scheme comparison. Furthermore, it can be integrated with a digital twin platform to embed the natural ventilation prediction model into the building information model, achieving synchronous updates of physical space and virtual predictions.
[0091] There are many ways to predict the effective area of natural ventilation in the elevated waiting hall of a railway station. The above-described method is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for predicting the effective area of natural ventilation in an elevated waiting hall of a railway station, wherein the elevated waiting hall is a cuboid structure, and its two short sides are respectively connected to corresponding entrance spaces; the entrance spaces are respectively located on both sides of the railway line; the elevated waiting hall is built above the railway station platform; various waiting areas and passenger passages between the waiting areas are arranged within the elevated waiting hall; other ancillary facilities besides the waiting areas and passenger passages are arranged inside the elevated waiting hall or on the two long sides of the exterior wall of the elevated waiting hall; the elevated waiting hall has an axial spatial characteristic; characterized in that... The longitudinal natural ventilation of the elevated waiting hall of the railway station is limited by the entrance space, and the lateral natural ventilation of the elevated waiting hall is related to the layout of other ancillary facilities. The prediction method includes: Step S1: Extract multiple spatial modes of the elevated waiting hall of the railway station; wherein, the multiple spatial modes include: side interface unobstructed mode, upper side interface semi-obstructed mode, lower side interface semi-obstructed mode, and side interface fully obstructed mode. The unobstructed side interface mode is as follows: the outer wall of the elevated waiting hall of the railway station is vertically divided into a first layer and a second layer from bottom to top; the other ancillary facilities are arranged inside the elevated waiting hall of the railway station; the other ancillary facilities divide the elevated waiting hall of the railway station into multiple parts; the side windows are opened on the first layer and / or the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows opened on the outer wall; The upper semi-obstruction mode of the side interface is as follows: the outer wall of the elevated waiting hall of the railway station is divided into a first layer and a second layer vertically from bottom to top; the other ancillary facilities are arranged on the second layer; the side windows are opened on the first layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows. The lower semi-obstructed mode of the side interface is as follows: the outer wall of the elevated waiting hall of the railway station is divided into a first layer and a second layer vertically from bottom to top; the other ancillary facilities are arranged on the first layer; the side windows are arranged on the second layer; the elevated waiting hall of the railway station is naturally ventilated through the side windows; The side interface full-coverage mode is as follows: the other ancillary facilities are arranged along the vertical direction of the outer wall of the elevated waiting hall of the railway station, and no side windows are opened on the outer wall of the elevated waiting hall of the railway station; The side windows on the first floor are low side windows; the side windows on the second floor are high side windows. Step S2: Based on the various spatial patterns, a numerical simulation method of computational fluid dynamics is selected to build m sets of natural ventilation simulation experimental models for the elevated waiting halls of the railway stations; wherein, the opening types of the natural ventilation simulation experimental models include one or a combination of the following: ground opening A, roof opening B, low side window on the windward side C, high side window on the windward side D, low side window on the leeward side E, and high side window on the leeward side F; The number of openings corresponding to ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F are all... indivual; The horizontal distance between the ground opening A and the windward side of the outer wall of the elevated waiting hall of the railway station is L1, and the horizontal distance between the roof opening B and the windward side of the outer wall of the elevated waiting hall of the railway station is L2. The windward low-side window C is at a height of H1 from the ground of the elevated waiting hall of the railway station; the windward high-side window D is at a height of H2 from the ground of the elevated waiting hall of the railway station; the leeward low-side window E is at a height of H3 from the ground of the elevated waiting hall of the railway station; and the leeward high-side window F is at a height of H4 from the ground of the elevated waiting hall of the railway station. And the ground opening A is expressed as A L1 The roof opening B is expressed as B. L2 The windward low-side window C is expressed in the form of C. H1 The windward high side window D is expressed in the form of D. H2 The leeward low-side window E is expressed as E H3 The leeward side window F is expressed as F H4 ; Step S3: Calculate the actual value of the effective area ratio of natural ventilation for the m groups of natural ventilation simulation experimental models; Step S4: Obtain the dataset; wherein the dataset includes: the opening type expression form corresponding to m groups of natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each group of natural ventilation simulation experimental models; m is a positive integer; Step S5: Preprocess the dataset; Step S6: Divide the preprocessed dataset into a training set and a test set; Step S7: Build a natural ventilation prediction model for elevated waiting halls in railway stations; wherein, the natural ventilation prediction model is an artificial neural network; the input of the natural ventilation prediction model is a training set or a test set; the output of the natural ventilation prediction model is the predicted value of the effective area ratio of natural ventilation corresponding to the training set or the test set. Step S8: Use the training set to train and optimize the natural ventilation prediction model, and save the final optimized natural ventilation prediction model; Step S9: Input the test set into the final optimized natural ventilation prediction model to obtain the predicted value of the effective area ratio of natural ventilation corresponding to the test set; Step S10: Evaluate the predictive performance of the natural ventilation prediction model based on the predicted value of the effective area ratio of natural ventilation corresponding to the test set and the actual value of the effective area ratio of natural ventilation corresponding to the test set.
2. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 1, characterized in that, In step S2, the roof of the elevated waiting hall of the railway station corresponding to the natural ventilation simulation experimental model is a flat roof, and the dimensions of the elevated waiting hall of the railway station are length... Meters, width meters, height rice; The ground opening A, roof opening B, windward low side window C, windward high side window D, leeward low side window E, and leeward high side window F are all square in shape; the area of each individual opening is... square meters; , , and All are positive numbers.
3. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 2, characterized in that, It is 200; It is 100; It is 20; =1; It is 14.
4. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 1, characterized in that, In step S2, the values of L1 and L2 are both in the range of 10-90 m, and the interval between the values of L1 and L2 is 10 m. The values of H1 and H3 are both in the range of 1-9 m, and the values of H2 and H4 are both in the range of 10-18 m. The interval between the values of H1, H3, H2 and H4 is 1 m.
5. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 1, characterized in that, In step S3, the actual value of the effective area ratio of natural ventilation is: the ratio of the effective area of natural ventilation in the area where the height of passengers in the elevated waiting hall of the railway station is 1.5m to the floor area of the elevated waiting hall of the railway station; The effective area for natural ventilation is defined as the area in the elevated waiting hall of a railway station where the height of passengers is 1.5m and the wind speed is higher than 0.1m / s.
6. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 5, characterized in that, Step S3 specifically involves: using the wind speed cloud map of the elevated waiting hall of the railway station at a height of 1.5m, calculating the actual value of the effective area ratio of natural ventilation in the natural ventilation simulation experiment model.
7. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 1, characterized in that, Step S6 specifically involves using the Latin cube sampling method to divide the preprocessed dataset into a training set and a test set.
8. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to claim 7, characterized in that, m is 475; the training set data includes: the opening type expression forms corresponding to 329 natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each natural ventilation simulation experimental model; the test set data includes: the opening type expression forms corresponding to 146 natural ventilation simulation experimental models and the actual value of the effective area ratio of natural ventilation calculated by each natural ventilation simulation experimental model.
9. The method for predicting the effective area of natural ventilation in elevated waiting halls of railway stations according to any one of claims 1 to 8, characterized in that, The artificial neural network is composed of a multilayer perceptron; the multilayer perceptron is composed of an input layer, a hidden layer and an output layer; the hidden layer is located between the input layer and the output layer.