CN116882268A - Data-driven tunnel fire smoke development prediction method and intelligent control system - Google Patents
Data-driven tunnel fire smoke development prediction method and intelligent control system Download PDFInfo
- Publication number
- CN116882268A CN116882268A CN202310708482.5A CN202310708482A CN116882268A CN 116882268 A CN116882268 A CN 116882268A CN 202310708482 A CN202310708482 A CN 202310708482A CN 116882268 A CN116882268 A CN 116882268A
- Authority
- CN
- China
- Prior art keywords
- tunnel
- temperature
- data
- temperature sensors
- deep learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000000779 smoke Substances 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000011161 development Methods 0.000 title claims abstract description 25
- 238000013135 deep learning Methods 0.000 claims abstract description 37
- 238000012502 risk assessment Methods 0.000 claims abstract description 30
- 238000002474 experimental method Methods 0.000 claims abstract description 28
- 238000004088 simulation Methods 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000012795 verification Methods 0.000 claims abstract description 11
- 238000013507 mapping Methods 0.000 claims abstract description 9
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 claims description 28
- 239000003546 flue gas Substances 0.000 claims description 28
- 238000012544 monitoring process Methods 0.000 claims description 10
- 239000003570 air Substances 0.000 claims description 7
- 230000006378 damage Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 230000010354 integration Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000011217 control strategy Methods 0.000 claims description 4
- 239000012080 ambient air Substances 0.000 claims description 3
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 239000012530 fluid Substances 0.000 claims description 3
- 230000007787 long-term memory Effects 0.000 claims description 3
- 238000013517 stratification Methods 0.000 claims description 3
- 230000006403 short-term memory Effects 0.000 claims 1
- 238000011156 evaluation Methods 0.000 abstract description 4
- 230000001276 controlling effect Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 4
- 238000009423 ventilation Methods 0.000 description 4
- 230000007480 spreading Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G06Q50/265—Personal security, identity or safety
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Fire Alarms (AREA)
Abstract
The invention discloses a data-driven tunnel fire smoke development prediction method and an intelligent control system, wherein the method comprises the following steps: a temperature sensor is longitudinally arranged in the model experiment tunnel, and the position coordinates and the temperature time sequence data of the temperature sensor form a verification data set; performing numerical simulation on the model experiment tunnel to construct a simulation tunnel model, wherein the position coordinates and the temperature time sequence data of the temperature sensor form a deep learning data set; verifying whether the simulated tunnel model is successfully constructed or not through a verification data set; training by adopting a deep learning data set to obtain a deep learning prediction model; inputting the temperature time sequence data into a deep learning prediction model, and mapping to obtain temperature time sequence data of temperature sensors arranged vertically; and performing risk assessment. According to the invention, a small amount of tunnel fixed sensor data is mapped to obtain flow field distribution in the vertical direction of the whole tunnel space, the required execution time is short, and super-real-time prediction and effective evaluation of tunnel fire smoke development by using limited sensors in the tunnel are realized.
Description
Technical Field
The invention relates to a prediction method for fire smoke development in a tunnel based on data driving, and also relates to an intelligent control system using the prediction method for fire smoke development in the tunnel based on data driving.
Background
Highway tunnels are very common in cities and have become an important part of highway traffic in China. The tunnel structure is long and narrow, has a large aspect ratio and a small cross section, and is relatively confined in space compared with a common building. Once a fire accident occurs in a tunnel, the accident of group death and group injury is easy to occur. The tunnel fire smoke development parameters are obtained through a small number of tunnel fixed temperature sensors, the tunnel fire is rapidly alarmed, and powerful guidance can be provided for emergency response of accidents and formulation of rescue strategies.
The current stage tunnel fire smoke development forecasting method mainly comprises three methods: the first is that temperature data are collected in real time through a sensor arranged on a tunnel ceiling, but the defect is that the sensor data can only reflect the situation of the local position of the sensor data; the second is through the physical model forecast, but the precondition that the physical model obtains certain flue gas development information at present is the accurate key parameter of input, and often difficult to confirm the input parameter accurately in the actual fire; the third is CFD field simulation reconstruction prediction, but a simple fire scene of tens of minutes often takes more than ten hours or even days to reproduce by using the CFD method, and the requirement of calculation speed required by real-time prediction cannot be met from the view point of the current computer level.
In addition, the current fire smoke ventilation control measures basically enable smoke to be rapidly diffused by controlling the opening of the fan, but the wind speed of the fan cannot be regulated and controlled according to the real dynamic change of the fire smoke in the tunnel at the present stage, and the harm degree of the smoke to personnel cannot be truly reflected only by limited sensor data due to various harm mechanisms of the fire smoke to human bodies. Therefore, the method for monitoring and controlling the smoke spreading of tunnel fire cannot accurately rescue trapped people.
Disclosure of Invention
The first object of the invention is to provide a data-driven tunnel fire smoke development prediction method, which can realize reliable prediction and effective evaluation of tunnel fire smoke development by using limited sensors in a tunnel.
The first object of the invention is achieved by the following technical measures: the method for predicting the development of the fire smoke based on the data-driven tunnel is characterized by comprising the following steps of:
s1, arranging a plurality of temperature sensors in a model experiment tunnel along the longitudinal direction, obtaining temperature time sequence data of the positions of the temperature sensors under different working conditions, establishing a coordinate system in the model experiment tunnel, recording position coordinates of the temperature sensors, and forming a verification data set with the temperature time sequence data acquired by the temperature sensors at the corresponding positions;
s2, carrying out numerical simulation on the model experiment tunnel to construct a simulation tunnel model, corresponding to the model experiment tunnel, arranging a plurality of temperature sensors in the simulation tunnel along the longitudinal direction of the model experiment tunnel, arranging a plurality of temperature sensors in the vertical direction of the model experiment tunnel, arranging a plurality of groups of temperature sensors in the vertical direction of the model experiment tunnel, obtaining temperature time sequence data of the temperature sensors in the longitudinal direction of the model experiment tunnel under different working conditions through simulation, establishing a coordinate system in the model experiment tunnel, recording position coordinates of the temperature sensors, and forming a deep learning data set with the temperature time sequence data of the temperature sensors at corresponding positions;
s3, verifying whether the simulated tunnel model is successfully constructed or not through a verification data set, and if so, turning to step S4; if the construction is unsuccessful, turning to step S2;
s4, processing the deep learning data set into a training set, a verification set and a test set for training the deep learning prediction model, and training, verifying and testing the deep learning prediction model to obtain a trained deep learning prediction model;
s5, inputting temperature time sequence data acquired by the temperature sensors arranged in the longitudinal direction into a trained deep learning prediction model, and mapping to obtain temperature time sequence data of the temperature sensors arranged in the vertical direction, so as to obtain vertical temperature distribution of the simulated tunnel;
and S6, performing risk assessment according to the vertical temperature distribution of the simulated tunnel to obtain risk indexes at all positions of the simulated tunnel, and controlling to start the exhaust device if the risk indexes are higher than a set value, and repeating the steps S5 and S6 until the risk indexes do not exceed the set value.
According to the invention, limited sensor data in the tunnel is utilized, a small amount of tunnel fixed sensor data is mapped to obtain flow field distribution in the vertical direction of the whole tunnel space in a data driving mode, the required execution time is far less than the execution time of CFD in long and large tunnel simulation and the actual spreading time of early smoke of tunnel fire, super real-time prediction is realized, the purpose of monitoring the environmental parameter in the tunnel in real time is expanded, and the intelligent control of tunnel smoke can be realized by feeding back the risk analysis to a control system, so that the reliable prediction and effective evaluation of the development of tunnel fire smoke by using the limited sensor in the tunnel can be realized.
The risk assessment specifically comprises the following steps:
calculating the average temperature of a flue gas layer and the thickness of the flue gas layer:
the integral ratio of the flue gas temperature is:
the percentage of integration of fresh air temperature is:
sum of integration ratio:
γ＝γ up +γ low formula (3)
When gamma is minimum, corresponding h int The height of the interface of the flue gas layer is the average temperature of the flue gas layer at the moment:
wherein: t (T) avg The average temperature (K) of the flue gas layer is T (z), the temperature distribution (K) of the flue gas in the vertical direction of the tunnel is T, and h is the tunnel height (m);
second, the dimensionless temperature ratio delta T is calculated cf /ΔT avg ：
ΔT avg ＝T avg -T a Formula (5)
ΔT cf ＝T c -T f Formula (6)
Wherein: t (T) c For tunnel top temperature (K), T f Is the ground temperature (K), T a Is the ambient air average temperature (K);
judging whether the smoke is better in layering or not, and calculating a risk assessment index I 1 ：
If I 1 Less than 0 indicates that the smoke layering in the area is better;
if I 1 Greater than 0 indicates poorer stratification of the flue gas in the region;
fourth step of setting a flue gas temperature threshold T Threshold value Calculating a risk assessment index I 2 ：
Comprehensive consideration risk assessment index I 1 And I 2 And obtaining the risk index of each position of the simulated tunnel.
The invention combines the dangerous indexes at each position of the simulated tunnel with the number positions of trapped people and the number types of blocked vehicles to judge the harm degree of fire smoke in the tunnel to the people.
In the steps S1 and S2, a coordinate system is constructed by taking the elevation of the tunnel ground as a reference, and the origin of the coordinate system is the position of the fire source.
The deep learning prediction model is composed of a multi-layer convolutional neural network, a long-term memory network and other deep learning algorithms, temperature time sequence data acquired by temperature sensors longitudinally arranged along a tunnel are used as input layers, temperature time sequence data of the temperature sensors vertically arranged along the tunnel, which correspond to input layer samples, are used as output layers, and the deep learning prediction model is built.
The temperature sensors longitudinally arranged are arranged on the tunnel ceiling at equal intervals along the central axis of the tunnel, and are 10-100 cm away from the ceiling, and the intervals are not more than 10-40 m.
The second object of the invention is to provide an intelligent control system using the data-driven tunnel fire smoke development prediction method.
The second object of the present invention is achieved by the following technical measures: an intelligent control system using the data-driven tunnel fire smoke development prediction method is characterized by comprising the following steps:
the on-line monitoring module is used for acquiring the position coordinates of the temperature sensors longitudinally arranged and the temperature time sequence data of the fire smoke acquired by the temperature sensors in real time;
the prediction module is used for inputting the data acquired by the data acquisition module into the deep learning prediction model, and mapping to obtain the vertical temperature distribution of the simulated tunnel;
the risk assessment module is used for carrying out risk assessment according to the vertical temperature distribution of the simulated tunnel;
and the intelligent control module is used for controlling the air exhaust device according to the risk assessment result, providing fire control strategy reference and controlling the fire control device.
The on-line monitoring module comprises a data experimental unit and a data simulation unit, wherein the data experimental unit comprises a plurality of temperature sensors and a host connected with the temperature sensors, and the host consists of a central processor, a coordinate system construction module, an alarm module and a display module which are respectively connected with the central processor; the data simulation unit is used for obtaining a data set by simulating a calculation fluid simulation method according to the simulated tunnel geometric parameters and the simulated accident working condition parameters.
Compared with the prior art, the invention has the following remarkable effects:
the method comprises the steps of mapping a small amount of tunnel fixed sensor data to obtain flow field distribution in the vertical direction of a whole tunnel space in a data driving mode by utilizing limited sensor data in the tunnel, wherein the required execution time is far less than the execution time of CFD in long and large tunnel simulation and the actual spreading time of early smoke of tunnel fire, so that ultra-real-time prediction can be realized, and reliable prediction and effective evaluation of the development of the smoke of the tunnel fire can be realized.
The method expands the application of the real-time monitoring data of the environmental parameters in the tunnel, provides corresponding risk assessment indexes, combines the number and the positions of trapped people, researches and judges the harmfulness in time and space, feeds back the harmfulness to a control system, realizes intelligent control of tunnel smoke, and has important engineering significance for fire control of tunnel fire.
According to the invention, tunnel fire situation prediction, research and judgment and control functions are integrated on one system, so that trapped people can be accurately rescued, and a certain fire control strategy reference is provided.
Drawings
The invention will now be described in further detail with reference to the drawings and to specific examples.
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a schematic diagram of data set extraction and partitioning in accordance with the method of the present invention;
FIG. 3 is a block diagram of a predictive flow of the method of the present invention;
FIG. 4 is a block diagram of a risk assessment flow for the method of the present invention;
fig. 5 is a schematic diagram of the composition structure of the intelligent control system of the present invention.
Detailed Description
The present invention will be described in detail with reference to the following examples and the accompanying drawings to help those skilled in the art to better understand the inventive concept of the present invention, but the scope of the claims of the present invention is not limited to the following examples, and it is within the scope of the present invention to those skilled in the art that all other examples obtained without making creative efforts are included in the scope of the present invention.
As shown in fig. 1 to 4, the method for predicting the development of fire smoke based on the data-driven tunnel comprises the following steps:
s1, arranging temperature sensors on a ceiling at equal intervals along a longitudinal central axis in a model experiment tunnel, wherein the distance between the temperature sensors and the ceiling is 20cm, and the arrangement interval is not more than 20 m. Different working conditions refer to different fire source power and tunnel environment conditions, temperature time sequence data of the positions of the temperature sensors are obtained, a coordinate system is established in a model experiment tunnel (the coordinate system is established by taking the elevation of the tunnel ground as a reference, and the origin of the coordinate system is the position of the fire source), the position coordinates of the temperature sensors are recorded, and a verification data set is formed by the position coordinates of the temperature sensors and the temperature time sequence data acquired by the temperature sensors at the corresponding positions;
s2, amplifying the key size of a model experiment tunnel according to a Froude similarity criterion, constructing a simulation tunnel model by numerical simulation based on the key size, corresponding to the model experiment tunnel, arranging temperature sensors at corresponding positions of a ceiling in the longitudinal direction of the simulation tunnel, arranging a plurality of temperature sensors at equal intervals in the vertical direction, arranging the temperature sensors in the vertical direction into a plurality of groups, controlling key fire environment boundary conditions (variable fire source power, variable ventilation speed and variable fire source positions) corresponding to each temperature sensor in the longitudinal direction, simulating m working conditions by simulation to obtain temperature time sequence data of the temperature sensors arranged in the longitudinal direction under different fire source power conditions and tunnel environment conditions, establishing a coordinate system (constructing the coordinate system by taking the elevation of the tunnel ground as a reference, wherein the coordinate origin of the coordinate system is the position of the fire source), recording the position coordinates of the temperature sensors, and forming a deep learning data set with the temperature time sequence data of the temperature sensors at the corresponding positions;
specific:
the method comprises the steps of constructing an x-y-z coordinate system in a tunnel, taking the position of a fire source as an origin (0, 0), wherein x represents the direction along the length of the tunnel and is forward towards an air outlet of the tunnel. y represents a forward direction toward the front of the tunnel in the tunnel width direction. z represents the tunnel height direction, and is forward toward the tunnel top.
Secondly, according to a coordinate system, data sets (x) of n tunnel ceiling fixed temperature sensors are correspondingly recorded 1 ,y 1 ,z 1 ),(x 2 ,y 2 ,z 2 ),……,(x n ,y n ,z n ). The values of the n coordinate points at each moment are extracted in m working conditions respectively, and a temperature time sequence data set (x) acquired by the sensor at the corresponding coordinate position can be obtained 1 ,y 1 ,z 1 ,T 1 ),(x 2 ,y 2 ,z 2 ,T 2 ),……,(x n ,y n ,z n ,T n )。
Third, according to the coordinate system (i.e., (x) n ,y n ) The same) record a dataset of tunnel vertical temperature distribution. Such as at (x) 1 ,y 1 ,z 1 ) Record corresponding to the location (x) 1 ,y 1 ,z’ 1,1 ),(x 2 ,y 2 ,z’ 1,2 ),……,(x n ,y n ,z’ 1,j ) Is a tunnel vertical temperature distribution (x 1 ,y 1 ,z’ 1,1 ,T’ 1,1 ),(x 2 ,y 2 ,z’ 1,2 ,T’ 1,2 ),……,(x n ,y n ,z’ 1,j ,T’ 1,j )。
Taking Deltax as the translation distance of the sensor group in the tunnel length direction (x direction), and translating the coordinate of the sensor group to (x) 1 +Δx,y 1 ,z 1 ),(x 2 +Δx,y 2 ,z 2 ),……,(x n +Δx,y n ,z n ) And correspondingly obtaining time sequence data of the tunnel fixed temperature sensor and the tunnel vertical temperature distribution.
And fifthly, translating the coordinates of the sensor group k times according to the interval delta x, and repeating the steps. Finally, time sequence data of the (k+1) multiplied by m groups of tunnel fixed temperature sensors and corresponding tunnel vertical temperature distribution are obtained.
S3, verifying whether the simulated tunnel model is successfully constructed or not through a verification data set according to the Froude number similarity law, and if so, turning to the step S4; if the construction is unsuccessful, turning to step S2;
s4, processing the deep learning data set into a training set, a verification set and a test set for training the deep learning prediction model, and training, verifying and testing the deep learning prediction model to obtain a trained deep learning prediction model;
the deep learning data set is specifically divided into a 60% training set, a 20% validation set and a 20% test set. Training the preprocessed training set sample, selecting a corresponding loss function and an optimizer, and training for multiple times to obtain a corresponding deep learning prediction model.
S5, inputting temperature time sequence data acquired by the temperature sensors arranged in the longitudinal direction into a trained deep learning prediction model, and mapping to obtain temperature time sequence data of the temperature sensors arranged in the vertical direction, so as to obtain vertical temperature distribution of the simulated tunnel;
the deep learning prediction model is composed of a multi-layer convolutional neural network, a long-term memory network and other deep learning algorithms, the final output result is a prediction result of vertical temperature distribution of the tunnel, and the model structure is not fixed and is a combination of a certain super parameter and a model structure.
And taking temperature time sequence data acquired by temperature sensors longitudinally arranged along the tunnel as an input layer, and taking temperature time sequence data of the temperature sensors longitudinally arranged along the tunnel, which correspond to samples of the input layer, as an output layer, so as to construct the deep learning prediction model.
And carrying out inverse normalization processing on the result output by the trained deep learning prediction model to obtain corresponding vertical temperature distribution of the tunnel.
And S6, performing risk assessment according to the vertical temperature distribution of the simulated tunnel to obtain risk indexes at all positions of the simulated tunnel, and controlling to start the exhaust device if the risk indexes are higher than a set value, and repeating the steps S5 and S6 until the risk indexes do not exceed the set value.
The method specifically comprises the following steps:
the risk assessment specifically comprises the following steps:
calculating the average temperature of the flue gas layer and the thickness of the flue gas layer:
the integral ratio of the flue gas temperature is:
the percentage of integration of fresh air temperature is:
sum of integration ratio:
γ＝γ up +γ low formula (3)
When gamma is minimum, corresponding h int The height of the interface of the flue gas layer is the average temperature of the flue gas layer at the moment:
wherein: t (T) avg The average temperature (K) of the flue gas layer is T (z), the temperature distribution (K) of the flue gas in the vertical direction of the tunnel is T, and h is the tunnel height (m);
calculation of dimensionless temperature ratio DeltaT cf /ΔT avg ：
ΔT avg ＝T avg -T a Formula (5)
ΔT cf ＝T c -T f Formula (6)
Wherein: t (T) c For tunnel top temperature (K), T f Is the ground temperature (K), T a Is the ambient air average temperature (K);
judging whether the smoke is layered or notWell, calculate risk assessment index I 1 ：
If I 1 Less than 0 indicates that the smoke layering in the area is better;
if I 1 Greater than 0 indicates poorer stratification of the flue gas in the region;
setting a flue gas temperature threshold T Threshold value Calculating a risk assessment index I 2 ：
Comprehensive consideration of risk assessment index I 1 And I 2 And obtaining the risk index of each position of the simulated tunnel.
And obtaining the risk indexes of the simulated tunnel at each position through risk assessment, and judging the degree of harm of fire smoke in the tunnel to personnel by combining the risk indexes of the simulated tunnel at each position with the number of trapped personnel and the number of blocking vehicles. If the risk index is higher than the set value, when people or vehicles at the position are damaged to a certain extent, the intelligent control system obtains certain information and converts the information into a signal which can be identified by the fan controller to control the fan. After the fan is started, the fixed temperature sensor of the tunnel ceiling records new data, and the steps S5 and S6 are repeated until the risk index does not exceed the set value, and the risk index is reduced to be within an acceptable range for personnel.
Referring to fig. 5, an intelligent control system using the data-driven tunnel fire smoke development prediction method includes:
the on-line monitoring module is used for acquiring the position coordinates of the temperature sensors longitudinally arranged and the temperature time sequence data of the fire smoke acquired by the temperature sensors in real time;
the prediction module is used for inputting the data acquired by the data acquisition module into the deep learning prediction model, and mapping to obtain the vertical temperature distribution of the simulated tunnel;
the risk assessment module is used for carrying out risk assessment according to the vertical temperature distribution of the simulated tunnel;
and the intelligent control module is used for controlling the air exhaust device according to the risk assessment result, providing fire control strategy reference and controlling the fire control device.
The on-line monitoring module comprises a data experiment unit and a data simulation unit, the data experiment unit comprises a plurality of temperature sensors and a host connected with the temperature sensors, and the host consists of a central processing unit, a coordinate system construction module, an alarm module and a display module which are respectively connected with the central processing unit; the data simulation unit is used for obtaining a data set by simulating a calculation fluid simulation method according to the simulated tunnel geometric parameters and the simulated accident working condition parameters.
The whole intelligent control system comprises the following working processes: when a fire occurs in the tunnel, a temperature sensor in the on-line monitoring module detects an abnormal rise in temperature, and this information is transmitted to the server to determine that the fire occurs. But at present mainly data of limited temperature sensors distributed longitudinally of the tunnel are collected. And inputting the data into a data-driven-based super-real-time prediction module, and mapping the smoke key parameter data longitudinally distributed in the tunnel into space-time distribution data in the vertical distribution of the tunnel after the prediction module is trained through a database established in advance. And then, the risk assessment module calculates the risk index of the smoke key parameters in space and time through the distribution of the input smoke key parameters in space and time, and combines the number position of trapped people and the number type of blocked vehicles to judge the harm degree of the smoke to the people in space and time. If the risk level is within the acceptable range, continuing to monitor and forecast. And if the dangerous degree exceeds the threshold range, starting the intelligent control module. And the ventilation and smoke discharging parameters are regulated and controlled through algorithms such as PID intelligent control and the like. The system forecast and control result can be used as the reference for controlling fire control decision of evacuation and longitudinal ventilation.
Claims (8)
1. The method for predicting the development of the fire smoke based on the data-driven tunnel is characterized by comprising the following steps of:
s1, arranging a plurality of temperature sensors in a model experiment tunnel along the longitudinal direction, obtaining temperature time sequence data of the positions of the temperature sensors under different working conditions, establishing a coordinate system in the model experiment tunnel, recording position coordinates of the temperature sensors, and forming a verification data set with the temperature time sequence data acquired by the temperature sensors at the corresponding positions;
s2, carrying out numerical simulation on the model experiment tunnel to construct a simulation tunnel model, corresponding to the model experiment tunnel, arranging a plurality of temperature sensors in the simulation tunnel along the longitudinal direction of the model experiment tunnel, arranging a plurality of temperature sensors in the vertical direction of the model experiment tunnel, arranging a plurality of groups of temperature sensors in the vertical direction of the model experiment tunnel, obtaining temperature time sequence data of the temperature sensors in the longitudinal direction of the model experiment tunnel under different working conditions through simulation, establishing a coordinate system in the model experiment tunnel, recording position coordinates of the temperature sensors, and forming a deep learning data set with the temperature time sequence data of the temperature sensors at corresponding positions;
s3, verifying whether the simulated tunnel model is successfully constructed or not through a verification data set, and if so, turning to step S4; if the construction is unsuccessful, turning to step S2;
s4, processing the deep learning data set into a training set, a verification set and a test set for training the deep learning prediction model, and training, verifying and testing the deep learning prediction model to obtain a trained deep learning prediction model;
s5, inputting temperature time sequence data acquired by the temperature sensors arranged in the longitudinal direction into a trained deep learning prediction model, and mapping to obtain temperature time sequence data of the temperature sensors arranged in the vertical direction, so as to obtain vertical temperature distribution of the simulated tunnel;
and S6, performing risk assessment according to the vertical temperature distribution of the simulated tunnel to obtain risk indexes at all positions of the simulated tunnel, and controlling to start the exhaust device if the risk indexes are higher than a set value, and repeating the steps S5 and S6 until the risk indexes do not exceed the set value.
2. The data-driven tunnel-based fire smoke development prediction method according to claim 1, wherein: the risk assessment specifically comprises the following steps:
calculating the average temperature of a flue gas layer and the thickness of the flue gas layer:
the integral ratio of the flue gas temperature is:
the percentage of integration of fresh air temperature is:
sum of integration ratio:
γ＝γ up +γ low formula (3)
When gamma is minimum, corresponding h int The height of the interface of the flue gas layer is the average temperature of the flue gas layer at the moment:
wherein: t (T) avg The average temperature (K) of the flue gas layer is T (z), the temperature distribution (K) of the flue gas in the vertical direction of the tunnel is T, and h is the tunnel height (m);
second, the dimensionless temperature ratio delta T is calculated cf /ΔT avg ：
ΔT avg ＝T avg -T a Formula (5)
ΔT cf ＝T c -T f Formula (6)
Wherein: t (T) c For tunnel top temperature (K), T f Is the ground temperature (K), T a Is the ambient air average temperature (K);
judging whether the smoke is better in layering or not, and calculating a risk assessment index I 1 ：
If I 1 Less than 0 indicates that the smoke layering in the area is better;
if I 1 Greater than 0 indicates poorer stratification of the flue gas in the region;
fourth step of setting a flue gas temperature threshold T Threshold value Calculating a risk assessment index I 2 ：
Comprehensive consideration risk assessment index I 1 And I 2 And obtaining the risk index of each position of the simulated tunnel.
3. The data-driven tunnel-based fire smoke development prediction method according to claim 2, wherein: and combining the dangerous indexes at each position of the simulated tunnel with the number positions of trapped people and the number types of blocked vehicles, and judging the harm degree of fire smoke in the tunnel to the people.
4. The data-driven tunnel-based fire smoke development prediction method according to claim 3, wherein: in the steps S1 and S2, a coordinate system is constructed with the elevation of the tunnel ground as a reference, and the origin of coordinates of the coordinate system is the location of the fire source.
5. The data-driven tunnel-based fire smoke development prediction method according to claim 4, wherein: the deep learning prediction model is composed of a multi-layer convolutional neural network, a long-term memory network, a short-term memory network and other deep learning algorithms, temperature time sequence data acquired by temperature sensors longitudinally arranged along a tunnel are used as an input layer, temperature time sequence data of the temperature sensors longitudinally arranged along the tunnel, which correspond to samples of the input layer, are used as an output layer, and the deep learning prediction model is built.
6. The data-driven tunnel-based fire smoke development prediction method according to claim 5, wherein: the temperature sensors longitudinally arranged are arranged on the tunnel ceiling at equal intervals along the central axis of the tunnel, and are 10-100 cm away from the ceiling, and the intervals are not more than 10-40 m.
7. An intelligent control system using the data-driven tunnel fire smoke development prediction method according to any one of claims 1 to 6, characterized by comprising:
the on-line monitoring module is used for acquiring the position coordinates of the temperature sensors longitudinally arranged and the temperature time sequence data of the fire smoke acquired by the temperature sensors in real time;
the prediction module is used for inputting the data acquired by the data acquisition module into the deep learning prediction model, and mapping to obtain the vertical temperature distribution of the simulated tunnel;
the risk assessment module is used for carrying out risk assessment according to the vertical temperature distribution of the simulated tunnel;
and the intelligent control module is used for controlling the air exhaust device according to the risk assessment result, providing fire control strategy reference and controlling the fire control device.
8. The intelligent control system of claim 7, wherein: the on-line monitoring module comprises a data experiment unit and a data simulation unit, wherein the data experiment unit comprises a plurality of temperature sensors and a host connected with the temperature sensors, and the host consists of a central processor, a coordinate system construction module, an alarm module and a display module which are respectively connected with the central processor; the data simulation unit is used for obtaining a data set by simulating a calculation fluid simulation method according to the simulated tunnel geometric parameters and the simulated accident working condition parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310708482.5A CN116882268B (en) | 2023-06-15 | 2023-06-15 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310708482.5A CN116882268B (en) | 2023-06-15 | 2023-06-15 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116882268A true CN116882268A (en) | 2023-10-13 |
| CN116882268B CN116882268B (en) | 2024-02-06 |
Family
ID=88261276
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310708482.5A Active CN116882268B (en) | 2023-06-15 | 2023-06-15 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116882268B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102030006B1 (en) * | 2019-06-05 | 2019-11-11 | 주식회사 리트코이엔에스 | An INTEGRATED FIREFIGHTING INTELLIGENT SYSTEM OF TUNNEL FOR TRAFFIC |
| CN113743015A (en) * | 2021-09-07 | 2021-12-03 | 同济大学 | Fire scene data acquisition method, medium and electronic device |
| CN113902963A (en) * | 2021-12-10 | 2022-01-07 | 交通运输部公路科学研究所 | Method and device for evaluating fire detection capability of tunnel |
| CN114741974A (en) * | 2022-05-12 | 2022-07-12 | 重庆交通大学 | Highway tunnel fire disaster growth period parameter identification and prediction method |
| CN114880935A (en) * | 2022-05-13 | 2022-08-09 | 西南交通大学 | Tunnel fire advanced prediction method |
| KR102441871B1 (en) * | 2022-04-01 | 2022-09-13 | 주식회사 맥서브 | Smoke guide system in tunnel with longitudinal ventilation |
| CN115526123A (en) * | 2022-08-25 | 2022-12-27 | 重庆大学 | Tunnel fire smoke ceiling jet flow development forecasting method and system based on data assimilation |
| CN115905998A (en) * | 2022-11-14 | 2023-04-04 | 国网上海市电力公司 | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm |
-
2023
- 2023-06-15 CN CN202310708482.5A patent/CN116882268B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102030006B1 (en) * | 2019-06-05 | 2019-11-11 | 주식회사 리트코이엔에스 | An INTEGRATED FIREFIGHTING INTELLIGENT SYSTEM OF TUNNEL FOR TRAFFIC |
| CN113743015A (en) * | 2021-09-07 | 2021-12-03 | 同济大学 | Fire scene data acquisition method, medium and electronic device |
| CN113902963A (en) * | 2021-12-10 | 2022-01-07 | 交通运输部公路科学研究所 | Method and device for evaluating fire detection capability of tunnel |
| KR102441871B1 (en) * | 2022-04-01 | 2022-09-13 | 주식회사 맥서브 | Smoke guide system in tunnel with longitudinal ventilation |
| CN114741974A (en) * | 2022-05-12 | 2022-07-12 | 重庆交通大学 | Highway tunnel fire disaster growth period parameter identification and prediction method |
| CN114880935A (en) * | 2022-05-13 | 2022-08-09 | 西南交通大学 | Tunnel fire advanced prediction method |
| CN115526123A (en) * | 2022-08-25 | 2022-12-27 | 重庆大学 | Tunnel fire smoke ceiling jet flow development forecasting method and system based on data assimilation |
| CN115905998A (en) * | 2022-11-14 | 2023-04-04 | 国网上海市电力公司 | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm |
Non-Patent Citations (4)
| Title |
|---|
| XIAONING ZHANG等: "Smart real-time forecast of transient tunnel fires by a dual-agent deep learning model", 《TUNNELLING AND UNDERGROUND SPACE TECHNOLOGY》, vol. 129, pages 1 - 14 * |
| YINGLI LIU等: "Experimental study on synergistic effect of exhaust vent layout and exhaust rate on performance of ceiling central smoke extraction in road tunnel fires", 《INTERNATIONAL JOURNAL OF THERMAL SCIENCES 》, vol. 183, pages 1 - 13 * |
| 朱鹏浩等: "基于多传感器融合的隧道智能巡检系统", 《科学技术与工程》, vol. 23, no. 02, pages 648 - 655 * |
| 阳东等: "隧道顶部障碍物下游火灾烟气密度跃变特性", 《中国安全科学学报》, vol. 33, no. 02, pages 68 - 74 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116882268B (en) | 2024-02-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ren et al. | Incorporating online monitoring data into fast prediction models towards the development of artificial intelligent ventilation systems | |
| Bastani et al. | Contaminant source identification within a building: toward design of immune buildings | |
| Wu et al. | An intelligent tunnel firefighting system and small-scale demonstration | |
| CN113379267B (en) | Urban fire event processing method, system and storage medium based on risk classification prediction | |
| CN114021501B (en) | Fire temperature field reconstruction method, system, computer equipment, medium and terminal | |
| CN112199882B (en) | Simulation method for evacuation process of people in fire, evacuation early warning method and system | |
| CN102999664A (en) | Method for identifying position of indoor pollution source | |
| CN114117617B (en) | Rapid prediction method for earthquake response of three-span gallery type RC frame building | |
| CN114019110B (en) | Workplace gas detector end cloud integration platform based on big data | |
| CN115310361B (en) | Underground coal mine dust concentration prediction method and system based on WGAN-CNN | |
| CN114662344B (en) | Atmospheric pollution source tracing prediction method and system based on continuous online observation data | |
| CN115392708A (en) | Fire risk assessment and early warning method and system for building fire protection | |
| CN114880935A (en) | Tunnel fire advanced prediction method | |
| Stathopoulos | Wind loads on low buildings: in the wake of Alan Davenport's contributions | |
| Christodoulou et al. | A BIM-based framework for forecasting and visualizing seismic damage, cost and time to repair | |
| CN116882268B (en) | Data-driven tunnel fire smoke development prediction method and intelligent control system | |
| CN114519304A (en) | Multi-target fire scene temperature prediction method based on distributed optical fiber temperature measurement | |
| Liu et al. | Real-time monitoring and prediction method of commercial building fire temperature field based on distributed optical fiber sensor temperature measurement system | |
| CN108364098B (en) | Method for measuring influence of weather characteristics on user sign-in | |
| CN114139243A (en) | BIM-based bridge fire emergency rescue method, terminal and storage medium | |
| CN113536433A (en) | BIM platform-based dynamic escape route optimization system for evacuation after disaster of building | |
| JP4209354B2 (en) | Diffusion state prediction method and diffusion state prediction system | |
| CN111178756A (en) | Multiple linear regression fire risk assessment method based on environmental big data | |
| Xue | The road tunnel fire detection of multi-parameters based on BP neural network | |
| JP2005264671A (en) | Dynamic simulation device and program for road tunnel ventilation control |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |