A reasonable smoke exhaust method for highway tunnels

Abstract

The present invention belongs to the technical field of highway tunnel smoke exhaust, in particular to a reasonable highway tunnel smoke exhaust method, step 1: calculating the plume mass flow rate of smoke #imgabs0#, and calculating the smoke generation rate V; step 2: acquiring parameter samples, using fire dynamics simulation software FDS to perform numerical analysis, and obtaining different ambient temperatures T0, smoke spreading range S under the arch, visibility D at a height of 2 m in the tunnel, and smoke exhaust duct flow rate V d , smoke outlet flow rate V0, smoke exhaust efficiency E y , heat removal efficiency E r The theoretical smoke exhaust volume Y under the parameters, this application comprehensively considers the parameter information of ambient temperature, smoke spread range under the arch, visibility at a height of 2 m in the tunnel, smoke exhaust duct flow velocity, smoke exhaust port flow velocity, smoke exhaust efficiency, and heat exhaust efficiency, and can accurately predict the optimal smoke exhaust volume of the highway tunnel.

Description

Technical Field

[0001] The invention relates to the technical field of smoke exhaust in highway tunnels, and in particular to a reasonable smoke exhaust method for highway tunnels. Background Art

[0002] In the calculation and analysis of reasonable smoke exhaust volume in highway tunnel fires published by Zhang Xin and Li Jian, in order to study the reasonable smoke exhaust volume in highway tunnel fires under the coupling effect of longitudinal ventilation and lateral concentrated smoke exhaust, the fire dynamics numerical simulation was used to analyze the influence of different smoke exhaust volumes on six key parameters: smoke spread range, visibility at a height of 2 m, smoke exhaust port flow rate, smoke exhaust duct flow rate, smoke exhaust efficiency, and heat exhaust efficiency. The smoke generation formula for axisymmetric smoke plume was further proposed. , Smoke generation rate formula ;

[0003] Fire smoke control is a key and difficult issue in the fire protection design of highway tunnels. Smoke exhaust volume is a key design parameter of the tunnel smoke exhaust system. Especially for new tunnels and new ventilation and smoke exhaust methods, the reasonable value of the key design parameter of smoke exhaust volume needs to be explored. Summary of the invention

[0004] The invention provides a reasonable smoke exhaust method for a highway tunnel, which solves the problem of how to reasonably design the smoke exhaust volume in the prior art.

[0005] The present invention provides the following technical solutions:

[0006] A reasonable smoke exhaust method for a highway tunnel comprises the following steps:

[0007] Step 1: Calculate the plume mass flow rate of the flue gas using the formula:

[0008]

[0009] in, is the plume mass flow rate of flue gas, in kg / s; is the convection part of the heat release rate of the fire source, in kW, , is the heat release rate of the fire source, is the flame limit height, in m; z is the height from the fuel surface to the bottom of the smoke layer, in m; the coefficients in the formula are obtained through experimental regression, and the exponents in the formula are obtained through dimensional analysis and fire plume theory;

[0010] The formula (1) is used to calculate Substituting into formula (2), the smoke generation rate can be calculated:

[0011]

[0012] Where V is the rate of fire smoke generation, unit: m 3 / s; is the ambient air density, in kg / m 3 , the value is 1.2kg / m 3 ; is the ambient temperature; is the average temperature of the fire smoke layer;

[0013] Step 2: Obtain parameter samples and use fire dynamics simulation software FDS for numerical analysis to obtain different ambient temperatures T0, smoke spread range S under the vault, visibility D at a height of 2 m in the tunnel, and exhaust duct flow velocity V d , smoke outlet flow rate V0, smoke exhaust efficiency E y , heat removal efficiency E r Theoretical smoke exhaust volume Y under parameters;

[0014] Step 3: Based on the input parameters X={T0, S, D, V d , V0, E y ,E r} and the corresponding theoretical smoke exhaust volume Y, and normalize X to establish a smoke exhaust volume knowledge base;

[0015] Step 4: Perform KNN algorithm preprocessing and define the Euclidean distance Dist of the normalized input parameter X in the smoke exhaust knowledge base as:

[0016] In the formula, i and j represent the i-th and j-th samples, and the N samples with the closest Euclidean distance are taken, and their respective weights are calculated according to the size of the Euclidean distance, that is:

[0017]

[0018] in, , They represent the Euclidean distance and smoke exhaust weight of the ith sample among the N samples closest to the input sample X, Represents the theoretical smoke exhaust volume of the ith sample, and uses the weighted average method to calculate the initial predicted value of the corresponding smoke exhaust volume of the input sample for:

[0019] .

[0020] As a further improvement of the above scheme.

[0021] Preferably, the method comprises the following steps:

[0022] Step 5: Further establish a smoke exhaust prediction model based on BP neural network. The BP neural network input is X={T0,S, D, Vd , V0,E y , E r}, the output of BP neural network is the smoke exhaust volume ; The number of hidden layers is 15, and the activation function is the sigmoid function. , x is the neuron input;

[0023] Specifically, normalize the input X:

[0024]

[0025] Input layer to hidden layer:

[0026]

[0027] Wi is the weight vector of the i-th hidden layer node, and b is the bias term;

[0028] w 1i Refers to the weight of the first input to the i-th hidden layer node...w 7i Refers to the weight of the 7th input to the i-th hidden layer node;

[0029] Hidden layer to output layer:

[0030]

[0031] Here i represents the weight from the hidden layer to the output layer, y i represents the hidden layer output, bk is the bias of the output layer;

[0032] Step 6: Use the sample data in the smoke exhaust knowledge base to train the BP neural network to obtain the model node parameters;

[0033] Step 7: Input the sample to be tested into the trained neural network to obtain the smoke exhaust volume predicted by the BP neural network .

[0034] Preferably, the method comprises the following steps:

[0035] Step 8: The formula used to calculate the smoke exhaust volume in KNN is modified to:

[0036]

[0037] The number of samples with the closest Euclidean distance in the KNN used for output weighted calculation is modified according to the number of smoke exhaust knowledge base samples. Assuming that the number of smoke exhaust knowledge base samples is M, then N' in formula (6) is ; The BP neural network updates its weights synchronously with the expansion of the smoke exhaust knowledge base;

[0038] Step 9: Smoke exhaust volume decision, take the final smoke exhaust volume is the weighted sum of the BP neural network output and the KNN output, that is .

[0039] In the present invention, a reasonable smoke exhaust method for a highway tunnel is proposed, which comprehensively considers parameter information of ambient temperature, smoke spreading range under the arch, visibility at a height of 2 m in the tunnel, smoke exhaust duct flow velocity, smoke exhaust port flow velocity, smoke exhaust efficiency, and heat exhaust efficiency, and can accurately predict the optimal smoke exhaust volume of the highway tunnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of the BP neural network structure provided by an embodiment of the present invention;

[0041] Figure 2 A flowchart provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The embodiments of the present invention are described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0043] In the embodiments of the present invention, "and / or" is only a description of the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0044] References to "one embodiment" or "some embodiments" etc. described in this specification mean that a particular feature, structure or characteristic described in conjunction with the embodiment is included in one or more embodiments of the present invention. Thus, the phrases "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. appearing in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0045] Example

[0046] Reference Figure 1-2 , a reasonable smoke exhaust method for a highway tunnel, the specific steps are as follows:

[0047] Step 1: Calculate the mass flow rate of the smoke plume. The theoretical calculation value of the tunnel smoke exhaust volume can be calculated using the smoke generation of the axisymmetric smoke plume:

[0048]

[0049] Among them, in the formula, is the plume mass flow rate of flue gas, in kg / s; is the convection part of the heat release rate of the fire source, in kW, , is the heat release rate of the fire source. is the flame limit height, in m; z is the height from the fuel surface to the bottom of the smoke layer, in m, the coefficients in the formula are obtained through experimental regression, and the exponents in the formula are obtained through dimensional analysis and fire plume theory.

[0050] The formula (1) is used to calculate Substituting into formula (2), the smoke generation rate can be calculated:

[0051]

[0052] Where V is the rate of fire smoke generation, unit: m 3 / s; is the ambient air density, in kg / m 3 , usually taken as =1.2kg / m 3 ; is the ambient temperature; is the average temperature of the fire smoke layer.

[0053] Step 2: Obtain parameter samples. The fire dynamics simulation software FDS was used for numerical analysis to obtain the different ambient temperatures T0, smoke spread range S under the vault, visibility D at a height of 2 m in the tunnel, and flow velocity V in the smoke exhaust duct. d , smoke outlet flow rate V0, smoke exhaust efficiency E y , heat removal efficiency E r The theoretical smoke exhaust volume Y under the parameters. The minimum smoke exhaust volume required by the tunnel should be equal to the fire smoke generation rate V; in actual design, factors such as air leakage and resistance loss along the way need to be considered, and the designed smoke exhaust volume should be greater than the calculated smoke exhaust volume.

[0054] Step 3: Based on the input parameters X={T0, S, D, V d , V0, E y , E r} and the corresponding theoretical smoke exhaust volume Y, and normalize X to establish a smoke exhaust volume knowledge base.

[0055] Step 4: KNN algorithm preprocessing, define the Euclidean distance Dist of the normalized input parameter X in the smoke exhaust knowledge base as:

[0056] In the formula, i and j represent the i-th and j-th samples, and the N samples with the closest Euclidean distance are taken, and their respective weights are calculated according to the size of the Euclidean distance, that is:

[0057]

[0058] in, , They represent the Euclidean distance and smoke exhaust weight of the ith sample among the N samples closest to the input sample X, Represents the theoretical smoke exhaust volume of the ith sample, and uses the weighted average method to calculate the initial predicted value of the corresponding smoke exhaust volume of the input sample for:

[0059] ;

[0060] Step 5: Since the number of samples in the smoke exhaust knowledge base is small at the beginning, the prediction effect of KNN is limited. A smoke exhaust prediction model based on BP neural network is established. The input of BP neural network is X={T0, S, D, V d , V0, Ey, Er}, the output of BP neural network is the smoke exhaust volume ; The number of hidden layers is 15, and the activation function is the sigmoid function. , x is the neuron input;

[0061] Specifically, normalize the input X:

[0062]

[0063] Input layer to hidden layer:

[0064]

[0065] Wi is the weight vector of the i-th hidden layer node, and b is the bias term;

[0066] w 1i Refers to the weight of the first input to the i-th hidden layer node...w 7i Refers to the weight of the 7th input to the i-th hidden layer node;

[0067]

[0068] Here i represents the weight from the hidden layer to the output layer, y i represents the hidden layer output, bk is the bias of the output layer; the schematic diagram of the model is shown in Figure 1 .

[0069] Step 6: Use the sample data in the smoke exhaust knowledge base to train the above BP neural network to obtain the model node parameters.

[0070] Step 7: Input the sample to be tested into the trained neural network to obtain the smoke exhaust volume predicted by the BP neural network .

[0071] Step 8: Exhaust volume knowledge base expansion, KNN correction, BP network update. On the one hand, when new samples are continuously input, they are input into the exhaust volume knowledge base for sample expansion; on the other hand, as the number of samples increases, the formula used to calculate the exhaust volume in KNN is corrected to:

[0072]

[0073] Specifically, the number of samples with the closest Euclidean distance in the KNN used for output weighted calculation is modified according to the number of smoke exhaust knowledge bases. Assuming that the number of samples in the smoke exhaust knowledge base is M, then N' in formula (6) is .

[0074] At the same time, the BP neural network updates its weights synchronously with the expansion of the smoke exhaust knowledge base.

[0075] Step 9: Smoke exhaust volume decision, take the final smoke exhaust volume is the weighted sum of the BP neural network output and the KNN output, that is ; and when the number of samples is large enough, the BP neural network output and the KNN output will tend to be consistent, and the system is in a consistent and stable state.

[0076] In summary, the present invention comprehensively considers the parameter information of ambient temperature, smoke spreading range under the arch, visibility at a height of 2 m in the tunnel, smoke exhaust duct flow velocity, smoke exhaust port flow velocity, smoke exhaust efficiency, and heat exhaust efficiency, and can accurately predict the optimal smoke exhaust volume of highway tunnels, which is of great significance for the construction of urban rail transit.

[0077] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention; the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A reasonable smoke exhaust method for a highway tunnel, characterized in that: The steps include: Step 1: Calculate the plume mass flow rate of flue gas according to the existing formula , calculate the smoke generation rate V; Step 2: Obtain parameter samples and use fire dynamics simulation software FDS for numerical analysis to obtain different ambient temperatures T0, smoke spread range S under the vault, visibility D at a height of 2 m in the tunnel, and exhaust duct flow rate V. d , smoke outlet flow rate V0, smoke exhaust efficiency E y , heat removal efficiency E r The theoretical smoke exhaust volume Y under the parameters; Step 3: Based on the input parameters X={T0, S, D, V d , V0, E y , E r } and the corresponding theoretical smoke exhaust volume Y, and normalize X to establish a smoke exhaust volume knowledge base; Step 4: Perform KNN algorithm preprocessing and define the Euclidean distance Dist of the normalized input parameter X in the smoke exhaust knowledge base as: In the formula, i and j represent the i-th and j-th samples, and the N samples with the closest Euclidean distance are taken, and their respective weights are calculated according to the size of the Euclidean distance, that is: in, , They represent the Euclidean distance and smoke exhaust weight of the ith sample among the N samples closest to the input sample X, Represents the theoretical smoke exhaust volume of the ith sample, and uses the weighted average method to calculate the initial predicted value of the corresponding smoke exhaust volume of the input sample for: ； Step 5: Further establish a smoke exhaust prediction model based on BP neural network. The BP neural network input is X={T0, S,D, V d , V0, E y , E r }, the output of BP neural network is the smoke exhaust volume ; The number of hidden layers is 15, and the activation function is the sigmoid function. , x is the neuron input; Specifically, normalize the input X: Input layer to hidden layer: Wi is the weight vector of the i-th hidden layer node, and b is the bias term; w 1i Refers to the weight of the first input to the i-th hidden layer node...w 7i Refers to the weight of the 7th input to the i-th hidden layer node; Hidden layer to output layer: Here i represents the weight from the hidden layer to the output layer, y i represents the hidden layer output, bk is the bias of the output layer; Step 6: Use the sample data in the smoke exhaust knowledge base to train the BP neural network to obtain the model node parameters; Step 7: Input the sample to be tested into the trained neural network to obtain the smoke exhaust volume predicted by the BP neural network ; Step 8: The formula used to calculate the smoke exhaust volume in KNN is modified to: The number of samples with the closest Euclidean distance in the KNN used for output weighted calculation is modified according to the number of smoke exhaust knowledge base samples. Assuming that the number of smoke exhaust knowledge base samples is M, then N' in formula (6) is ; The BP neural network updates its weights synchronously with the expansion of the smoke exhaust knowledge base; Step 9: Smoke exhaust volume decision, take the final smoke exhaust volume is the weighted sum of the BP neural network output and the KNN output, that is .

Images