Highway tunnel fire time-varying temperature field reconstruction method and system

Abstract

The invention relates to a highway tunnel fire time-varying temperature field reconstruction method, which comprises the following steps: S1, determining vehicle ignition ranges of different fire source powers of a tunnel fire, S2, determining vehicle ignition time near a fire source in a tunnel fire development process, S3, calibrating a tunnel fire scale test multi-fire-source heat release rate, and S4, calculating a tunnel fire scale test multi-fire-source heat release rate. S4, setting a tunnel fire spreading numerical simulation heat release rate curve; S5, establishing a tunnel fire temperature field database; S6, designing a tunnel fire reconstruction temperature field sensor array layout mode; S7, providing an interpolation method for tunnel fire temperature field reconstruction; s9, a reconstruction temperature field cloud picture of different sensor array layout and interpolation method combination is drawn, and S10, the interpolation precision of each layout of tunnel fire disaster temperature field reconstruction is evaluated, the fire spreading distance and time are predicted according to the temperature field reconstruction method, the fire rescue intelligence degree is improved, and the life and property safety of people is guaranteed to the maximum extent.

Description

Technical Field

[0001] The invention belongs to the technical field of tunnel fire safety, and relates to a method and system for reconstructing a time-varying temperature field of a highway tunnel fire. Background Art

[0002] Fire is a typical disaster in the operation period of tunnels, with the characteristics of rapid spread, high temperature and difficult rescue. After the fire occurs, the heat radiation and convection of flames and smoke can easily lead to the spread of fire and ignite stranded vehicles, causing mass deaths and injuries to people in the tunnel.

[0003] The dense smoke generated when a tunnel fire occurs makes the tunnel a "black box", which limits the effectiveness of traditional fire monitoring methods. It is difficult to know the direction of fire spread, the distance of spread, the number of fire sources, and the temperature field, and it is impossible to guide fire rescue work. However, most of the current studies on tunnel fire spread are based on the assumption of a single fixed fire source, and fail to consider the spread that may be caused by the fire source during the development process. Although there are a large number of experimental studies on dual fire sources, most of them supply fuel at a fixed rate and do not consider the mutual thermal feedback of the fire sources in the actual spreading fire. At the same time, it is ignored that there is a time delay in the appearance of multiple fire sources in the spreading fire, and they do not appear at the same time and reach the peak HRR. At the same time, in terms of temperature field reconstruction, existing studies often simplify the tunnel into a one-dimensional model, ignoring the lateral spread of the fire, and the current layout of temperature sensors is difficult to capture the temperature field changes caused by the lateral spread of the fire, that is, the amount of data is insufficient, and it is difficult to meet the needs of visual reconstruction of the temperature field.

[0004] Therefore, there is an urgent need for a high-precision tunnel fire time-varying temperature field reconstruction method to study the HRR evolution characteristics under mutual thermal feedback of fire sources, optimize the temperature sensor layout, and predict the fire spread distance and time based on the temperature field reconstruction method; to provide effective theoretical and technical support for fire control and extinguishing, personnel escape and evacuation, fire emergency rescue, etc. in the fire development stage, while also improving the intelligence level of fire rescue and ensuring the safety of people’s lives and property. Summary of the invention

[0005] In view of this, the present invention provides a method and system for reconstructing the time-varying temperature field of a highway tunnel fire in order to solve the problems that the existing research on the spread of tunnel fire has ignored the time delay of the emergence of multiple fire sources and has ignored the lateral spread of fire in the temperature field reconstruction, resulting in difficulty in accurately studying the temperature field changes caused by the spread of fire and poor adaptability.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for reconstructing a time-varying temperature field of a highway tunnel fire comprises the following steps:

[0008] S1: Determine the ignition range of vehicles with different fire source powers in tunnel fires: Based on the critical ignition heat flux of each vehicle component (tire, bumper, wheel arch, fuel tank, fender and hub cover), combined with the fire source heat radiation range formula, smoke heat radiation range formula and smoke temperature attenuation model, the ignition range of the vehicle under different fire source powers is obtained;

[0009] S2: Determine the ignition time of vehicles near the fire source during the development of the tunnel fire: The critical ignition heat flux and ignition time in step S1 are used to obtain the vehicle ignition time TTI within a specific range through a fitting function. i The calculated TTI calculates the expected ignition time ETI, and finds the earliest expected time ETOI that satisfies ignition;

[0010] S3: Calibrate the heat release rate of multiple fire sources in the tunnel fire scale test: Based on the weight loss method and the ignition range determined in step S1 and the ignition time determined in step S2, calibrate the heat release rate (HRR) of multiple fire sources in the tunnel fire scale test;

[0011] S4: Setting the heat release rate curve for numerical simulation of tunnel fire spread: constructing the fire spread in the fire scene based on the result calibrated in step S3, and calculating and analyzing the characteristics of the heat release rate curve under different spacing and longitudinal wind speed conditions;

[0012] S5: Establish a tunnel fire temperature field database: After clarifying the single fire source combustion and multiple fire source heat release rate development curves, obtain more detailed temperature field data and real-time heat release rate data through numerical simulation (setting 49 working conditions including single and double fire sources);

[0013] S6: Design the layout of the sensor array for tunnel fire reconstruction temperature field: Deploy temperature sensors on the left and right lanes of the tunnel respectively, and arrange temperature sensors in a "chessboard" and "triangular network" manner on the top of the tunnel;

[0014] S7: Propose an interpolation method for reconstructing the temperature field of a tunnel fire: By writing Python code, the interpolation area is gridded, and the grid points are interpolated using three two-dimensional interpolation methods: Kriging interpolation, inverse distance weighted interpolation, and spline interpolation, to achieve real-time drawing of the temperature cloud map;

[0015] S8: Generate a real-time temperature field cloud map of the tunnel fire temperature field: The tunnel ceiling temperature attenuation physical model is integrated with the interpolation algorithm of step S7. When the fire source is not directly below the temperature sensor, a virtual point with the highest temperature is created directly above the fire source. The temperature attenuation model is used to calculate the temperature of the virtual point and a weight is assigned. The remaining parts are calculated using only the interpolation algorithm.

[0016] S9: Draw the reconstructed temperature field cloud map of different sensor array layouts and interpolation method combinations: Use the cubic spline interpolation function in step S7 to calculate the temperature value of the known coordinate point on the top of the tunnel, and use the coolwarm in the matplotlib library to map it to a preset color scale, so that the color of each grid point represents the temperature at its location, so as to form a two-dimensional temperature cloud map that intuitively shows the temperature gradient change, and compare it with the actual temperature cloud map;

[0017] S10: Evaluate the interpolation accuracy of each layout for reconstructing the tunnel fire temperature field: Take the ceiling temperature data of a certain working condition, use the three interpolation methods in step S7, and cross-validate the results; evaluate the accuracy through the mean relative error (MRE), the lower the MRE value, the higher the accuracy of the interpolation method; evaluate the robustness through the two indicators of mean absolute error (MAE) and root mean square error (RMSE); the lower the MAE and RMSE values, the smaller the error of the interpolation result, the higher the accuracy, and the better the robustness; thus, an interpolation method suitable for generating tunnel fire temperature cloud maps is obtained.

[0018] Furthermore, the formula for the heat radiation range of the fire source in S1 is:

[0019] (1)

[0020] Where: Q rf is the radiant heat received by adjacent vehicles, kW / m 2 ; is the transmittance of air to thermal radiation, generally taken as 1; r is the distance between the target object and the virtual point source, m; θ is the angle between the connecting line and the normal direction of the target object; Q r is the external radiation power of the virtual point source, MW; Q r =x r Q, χ r is the flame radiation fraction, ranging from 0.2 to 0.32. The fire source fuel in this study is anhydrous ethanol and is taken as 0.25. Q is the heat release rate of the fire source.

[0021] Furthermore, the flue gas heat radiation range formula in step S1 is:

[0022] (2)

[0023] Where: Q rs Indicates the smoke radiation heat received by adjacent vehicles, W / m 2 ; ε r represents the flue gas emissivity, which is 0.8; T represents the flue gas temperature, K.

[0024] Further, the fitting function relationship in step S2 is:

[0025] (3)

[0026] Where: t is the ignition time, s; is the radiation heat flux, kW / m 2 ; Goodness of fit (R 2 ) is 0.99, and the two are in a power exponential relationship, indicating that the ignition time decreases according to a power law with the increase of radiation heat flux.

[0027] The expected ignition time is:

[0028] (4)

[0029] The actual ignition time is:

[0030] (5)

[0031] Where: A time collection representing the fire development growth period and the stable period.

[0032] Furthermore, FDS was used for numerical simulation in S5, and the FDS acquisition frequency was set to 3 times / s; 1050 rows of data including temperature, wind speed, and heat release rate (HRR) were obtained for each working condition to expand the data set.

[0033] Further, in step S6, the temperature sensors at both ends of the tunnel top are 0.5 m away from the left and right side walls, and the distance between adjacent temperature sensors is 5-10 m.

[0034] Furthermore, the pandas library is used in S7 to read the csv table data storing the temperature numbers and sensor coordinate data, and matplotlib.pyplot is used for drawing.

[0035] Furthermore, the expression of Kriging interpolation in the two-dimensional interpolation method is:

[0036] (6)

[0037] Where: z(x) is the unknown position The estimated value of; n is the number of known observation points; Z i For a known point x i The actual observation value at; λi is the interpolation weight, which is calculated by the Kriging criterion according to the positional relationship between each known point and the target unknown point in space and the predefined covariance function;

[0038] The inverse distance weighted interpolation expression is:

[0039] (7)

[0040] Where: is the number of known observation points; It is The attribute value of each observation point; For the The weight of each observation point on P, ; Indicates the point to be interpolated With Observation points The Euclidean distance between is the weight exponent, usually greater than zero, used to control the speed at which the weight decays with distance;

[0041] The expression of spline interpolation is:

[0042] Every cubic polynomial It is expressed as:

[0043] (8)

[0044] Where: a i , b i , c i and d i are the coefficients of the polynomial, which need to be determined by the interpolation conditions and other constraints. In x i Passing through point x i ,y i .have , which means The smoothness conditions for cubic spline interpolation include continuity, first-order derivative continuity, and second-order derivative continuity:

[0045] (9)

[0046] These conditions can be transformed into a series of linear equations to form a linear equation system. By solving this linear equation system, we can get , and The value of , thus obtaining the complete cubic spline interpolation function.

[0047] Furthermore, the physical model of tunnel ceiling temperature attenuation in S8 is specifically:

[0048] (10)

[0049] Where: is the temperature at the reference position; The distance from the reference position Temperature at m; is the initial ambient temperature; K is the attenuation coefficient; the temperature unit is K.

[0050] Furthermore, the average relative error in S10 is specifically:

[0051] (11)

[0052] The mean absolute error is:

[0053] (12)

[0054] The root mean square error is specifically:

[0055] (13)

[0056] Where: Indicates the temperature value of each temperature measuring point; n is the number of samples; is the estimated value of temperature.

[0057] A system based on a method for reconstructing a time-varying temperature field of a highway tunnel fire, characterized by comprising:

[0058] At least one processor; and a memory in communication with the at least one processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to implement the above-mentioned method for reconstructing the time-varying temperature field of a highway tunnel fire.

[0059] The beneficial effects of the present invention are:

[0060] The present invention discloses a method for reconstructing the time-varying temperature field of a highway tunnel fire. Compared with the traditional method, the method has good reliability and feasibility, can be used to optimize the layout of temperature sensors, and predict the fire spread distance and time according to the temperature field reconstruction method; it provides effective theoretical and technical support for fire control and extinguishing, personnel escape and evacuation, fire emergency rescue, etc. in the fire development stage, and also improves the intelligence level of fire rescue, and maximizes the protection of people's lives and property safety.

[0061] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0063] Figure 1 A flow chart for establishing a tunnel fire temperature field database in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0064] Figure 2 (a) is a chessboard-type display layout 1 in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0065] Figure 2 (b) is a second chessboard-type display layout in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0066] Figure 2 (c) is a diagram for evaluating the accuracy of chessboard interpolation points in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0067] Figure 3 (a) is a triangulated network layout 1 in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0068] Figure 3 (b) is a second triangulated network layout in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0069] Figure 3 (c) is a diagram for evaluating the accuracy of chessboard interpolation points in a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0070] Figure 4 A flow chart for creating a target plane grid in S7 of a method for reconstructing a time-varying temperature field of a highway tunnel fire according to the present invention;

[0071] Figure 5 A flowchart of real-time generation of temperature cloud map in S7 of a method for reconstructing time-varying temperature field of a highway tunnel fire according to the present invention;

[0072] Figure 6 The present invention is a method for reconstructing the time-varying temperature field of a highway tunnel fire. In S9, the Kriging interpolation method is used. The sensors are arranged in a chessboard pattern, with a sensor arranged every 5m. The temperature field is reconstructed into a two-dimensional temperature cloud. Figure 1 ;

[0073] Figure 7 The invention discloses a method for reconstructing the time-varying temperature field of a highway tunnel fire. In S9, the inverse distance weighted average interpolation method is used. The sensors are arranged in a chessboard pattern, with a sensor arranged every 5 m. The temperature field is reconstructed into a two-dimensional temperature cloud. Figure 2 ;

[0074] Figure 8The present invention is a method for reconstructing the time-varying temperature field of a highway tunnel fire. In S9, a cubic spline interpolation method is used. The sensors are arranged in a chessboard pattern, with a sensor arranged every 5m. The temperature field is reconstructed into a two-dimensional temperature cloud. Figure 3 ;

[0075] Fig. 9 This is the actual temperature cloud map in the method S9 for reconstructing the time-varying temperature field of a highway tunnel fire according to the present invention. DETAILED DESCRIPTION

[0076] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0077] like Figure 1-Figure 9 A method for reconstructing a time-varying temperature field of a highway tunnel fire is shown, comprising the following steps:

[0078] S1: Based on the critical ignition heat flux of each vehicle component (tires, bumpers, wheel arches, fuel tanks, fenders and hub covers), combined with the fire source heat radiation range formula, smoke heat radiation range formula and smoke temperature attenuation model, the ignition range of the vehicle under different fire source powers is obtained.

[0079] Among them, the formula for the heat radiation range of the fire source is:

[0080] (1)

[0081] Where: Q rf is the radiant heat received by adjacent vehicles, kW / m 2 ; is the transmittance of air to thermal radiation, generally taken as 1; r is the distance between the target object and the virtual point source, m; θ is the angle between the connecting line and the normal direction of the target object; Q r is the external radiation power of the virtual point source (MW), Q r =x r Q, χ r is the flame radiation fraction, ranging from 0.2 to 0.32. The fire source fuel in this study is anhydrous ethanol and is taken as 0.25. Q is the heat release rate of the fire source.

[0082] The formula for flue gas heat radiation range is:

[0083] (2)

[0084] Where: Q rs Indicates the smoke radiation heat received by adjacent vehicles, W / m2 ; ε r represents the flue gas emissivity, which is 0.8; T represents the flue gas temperature, K.

[0085] S2: The critical ignition heat flux and ignition time are fitted to obtain the vehicle ignition time (Time To Ignition, TTI) within a specific range, and at each moment t i The calculated TTI calculates the expected time of ignition (ETI) and finds the earliest expected time of ignition (ETOI). This method is compared with the traditional ignition time prediction method FTP (flux-time product) to verify the reliability of the method.

[0086] The fitting function relationship is:

[0087] (3)

[0088] Where: t is the ignition time, s; is the radiation heat flux kW / m 2 ; Goodness of fit (R 2 ) is 0.99, and the two are in a power exponential relationship, indicating that the ignition time decreases according to a power law with the increase of radiation heat flux.

[0089] The expected ignition time is:

[0090] (4)

[0091] The actual ignition time is:

[0092] (5)

[0093] Where: A time collection representing the fire development growth period and the stable period.

[0094] S3: Calibration of heat release rates from multiple fire sources in tunnel fire scale test:

[0095] Based on the weight loss method and the calculated ignition range and ignition time, the heat release rate of multiple fire sources in the tunnel fire in the scaled test was calibrated (18 working conditions). The initial fire source with a larger peak HRR was called fire source A, and the new fire source with a smaller peak HRR that developed over time was called fire source B.

[0096] S4: Setting the heat release rate curve for tunnel fire spread numerical simulation:

[0097] Based on the calibration results of multiple fire sources, the fire spread situation in the fire scene is constructed, the fire spread conditions under the mutual influence of multiple fire sources are discussed, and the heat release rate curves under different spacings and longitudinal wind speeds are calculated and analyzed.

[0098] S5: Reference Figure 1 , establish a tunnel fire temperature field database:

[0099] After clarifying the development curves of single-fire source combustion and multiple-fire source heat release rates, more detailed temperature field data and real-time HRR data were obtained through numerical simulation (setting 49 working conditions including single and double fire sources). The FDS acquisition frequency was set to 3 times / s, and 1050 rows of data including temperature, wind speed, and HRR were obtained for each working condition to expand the data set.

[0100] S6: Reference Figure 2-Figure 3 , sensor array layout for tunnel fire reconstruction temperature field:

[0101] To improve the accuracy and coverage of monitoring, a new row of temperature sensors was added to the left and right lanes of the tunnel, forming a wider temperature sensing array as shown in the figure below. At the same time, to capture the spread of fire in the tunnel cross section, a row of temperature sensors was added at the top of the tunnel, 0.5m away from the left and right side walls. They were arranged in a "chessboard" and "triangular network" manner.

[0102] S7: Interpolation method for reconstruction of tunnel fire temperature field:

[0103] By writing Python code, the interpolation area is Figure 4 , the grid operation shown in the figure interpolates the grid points to achieve real-time drawing of the temperature cloud map. Among them, the pandas library is used to read the csv table data storing the temperature number and sensor coordinate data, and matplotlib.pyplot is used for drawing.

[0104] The temperature distribution of the tunnel ceiling can be regarded as a continuous surface, which is a typical two-dimensional interpolation problem. Common two-dimensional interpolation methods include Kriging interpolation, inverse distance weighted interpolation, spline interpolation, etc.

[0105] The core expression of Kriging Interpolation (KI) is:

[0106] (6)

[0107] Where: z(x) is the unknown position The estimated value of; n is the number of known observation points; Z i For a known point x iλi is the actual observation value at ; λi is the interpolation weight, which is calculated by the Kriging criterion according to the positional relationship between each known point and the target unknown point in space and the predefined covariance function.

[0108] The expression formula of Inverse Distance Weighted (IDW) is:

[0109] (7)

[0110] Where: is the number of known observation points; It is The attribute value of each observation point; For the The weight of an observation point on P is generally defined as: ; Indicates the point to be interpolated With Observation points The Euclidean distance between is the weight exponent, usually greater than zero, and is used to control how fast the weight decays with distance.

[0111] Cubic Spline Interpolation (CSI):

[0112] Every cubic polynomial It can be expressed as:

[0113] (8)

[0114] Where: a i , b i , c i and d i are the coefficients of the polynomial, which need to be determined by the interpolation conditions and other constraints. In x i Passing through point x i ,y i .have , which means The smoothness conditions for cubic spline interpolation include continuity, first-order derivative continuity, and second-order derivative continuity:

[0115] (9)

[0116] These conditions can be transformed into a series of linear equations to form a linear equation system. By solving this linear equation system, we can get , and The value of , thus obtaining the complete cubic spline interpolation function.

[0117] S8: Real-time temperature cloud map generation of tunnel fire temperature field:

[0118] The tunnel ceiling temperature attenuation physical model shown in Formula 10 is integrated with the interpolation algorithm. When the fire source is not directly below the temperature sensor, a virtual point with the highest temperature is created directly above the fire source. The temperature attenuation model is used to calculate the temperature of the virtual point and a larger weight is assigned. The rest of the points are calculated using only the interpolation algorithm. When the location of the fire source is known, the temperature attenuation model can be used for calculation and a more accurate temperature above the fire source can be obtained.

[0119] The physical model of tunnel ceiling temperature attenuation is as follows:

[0120] (10)

[0121] Where: is the temperature at the reference position; The distance from the reference position Temperature at m; is the initial ambient temperature; K is the attenuation coefficient; the temperature unit is K.

[0122] Reference Figure 5 The real-time generation process of the temperature cloud map shown in the figure, except for the construction of the interpolation function in the gray box, the rest of the process is cycled once each time the temperature sensor array data is updated to generate the real-time temperature cloud map.

[0123] S9: Reconstructed temperature field rendering of different sensor array layouts and interpolation method combinations:

[0124] Reference Figure 6-Figure 9 The temperature values ​​of the known coordinate points on the top of the tunnel are calculated using the cubic spline interpolation function, and mapped to the preset color scale using coolwarm in the matplotlib library. Finally, the color of each grid point represents the temperature at its location, forming a two-dimensional temperature cloud map that intuitively shows the change in temperature gradient. The temperature cloud map is compared with the actual temperature cloud map. All three interpolation methods can better reflect the overall temperature distribution of the tunnel ceiling.

[0125] S10: Evaluation of interpolation accuracy of various layouts for reconstruction of tunnel fire temperature field:

[0126] Evaluation of the interpolation accuracy of each layout for reconstruction of the tunnel fire temperature field: The ceiling temperature data of a certain working condition was taken, and three interpolation methods were used to cross-validate the results; the accuracy was evaluated by the mean relative error (MRE), and the lower the MRE value, the higher the accuracy of the interpolation method; the robustness was evaluated by the mean absolute error (MAE) and root mean square error (RMSE); the lower the MAE and RMSE values, the smaller the error of the interpolation result, the higher the accuracy, and the better the robustness; thus, an interpolation method suitable for generating tunnel fire temperature cloud maps was obtained.

[0127] The average relative error is:

[0128] (11)

[0129] The mean absolute error is:

[0130] (12)

[0131] The root mean square error is specifically:

[0132] (13)

[0133] Where: Indicates the temperature value of each temperature measuring point; n is the number of samples; is the estimated value of temperature.

[0134] Figure 2 (c) and Figure 3 The red points in (c) are temperature measurement points, whose data are known. The gray points are feature points selected for interpolation accuracy evaluation, whose data are generated by the interpolation model, and whose coordinates remain unchanged in different grid layout schemes.

[0135]

[0136] The evaluation of the average relative error of each interpolation combination shows that the average error of group Z7 is the smallest, which is 0.16%. That is, when the sensors are arranged according to layout 3, the interpolation result has the highest accuracy and is more suitable for practical applications. The average error of group Z10 is the largest, which is 6.66%. That is, when the sensors are arranged according to layout 2, the interpolation result has the lowest progress. In practical applications, this combination should be avoided. In summary, the layout interval of temperature sensors in most tunnels is 10m. If the temperature sensors are arranged at a density of 10m / piece, according to Table 1, the Z4 interpolation combination has the lowest error rate (3.01%), which means that the triangulated network layout and Kriging interpolation model should be used to reconstruct the temperature field.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A method for reconstructing the time-varying temperature field of a highway tunnel fire, characterized in that: The following steps are involved: S1: Determine the ignition range of vehicles with different fire source powers in tunnel fires: According to the critical ignition heat flux of each vehicle component, combined with the fire source heat radiation range formula, smoke heat radiation range formula and smoke temperature attenuation model, the ignition range of the vehicle under different fire source powers is obtained; S2: Determine the ignition time of vehicles near the fire source during the development of the tunnel fire: The critical ignition heat flux and ignition time in step S1 are used to obtain the vehicle ignition time TTI within a specific range through a fitting function. i The calculated TTI calculates the expected ignition time ETI, and finds the earliest expected time ETOI that satisfies ignition; S3: Calibrate the heat release rate of multiple fire sources in the tunnel fire scale test: calibrate the heat release rate of multiple fire sources in the tunnel fire scale test based on the weight loss method and the ignition range determined in step S1 and the ignition time determined in step S2; S4: Setting the heat release rate curve for numerical simulation of tunnel fire spread: constructing the fire spread in the fire scene based on the result calibrated in step S3, and calculating and analyzing the characteristics of the heat release rate curve under different spacing and longitudinal wind speed conditions; S5: Establish a tunnel fire temperature field database: After clarifying the single fire source combustion and multiple fire source heat release rate development curves, obtain more detailed temperature field data and real-time heat release rate data through numerical simulation; S6: Design the layout of the sensor array for temperature field reconstruction in tunnel fires: temperature sensors are placed on the left and right lanes of the tunnel respectively, and temperature sensors are arranged in a "chessboard" and "triangular network" manner on the top of the tunnel; S7: An interpolation method for reconstructing the temperature field of a tunnel fire is proposed: the interpolation area is gridded, and the grid points are interpolated using three two-dimensional interpolation methods: Kriging interpolation, inverse distance weighted interpolation, and spline interpolation, to achieve real-time drawing of the temperature cloud map; S8: Generate a real-time temperature cloud map of the tunnel fire temperature field: The tunnel ceiling temperature attenuation physical model is integrated with the interpolation algorithm of step S7. When the fire source is not directly below the temperature sensor, a virtual point with the highest temperature is created directly above the fire source. The temperature attenuation model is used to calculate the temperature of the virtual point and a weight is assigned. The remaining parts are calculated using the interpolation algorithm. S9: Draw the reconstructed temperature field cloud map of different sensor array layouts and interpolation method combinations: Use the interpolation function in step S7 to calculate the temperature value of the known coordinate point on the top of the tunnel, and use the coolwarm in the matplotlib library to map it to the preset color scale, so that the color of each grid point represents the temperature at its location, so as to form a two-dimensional temperature cloud map that intuitively shows the temperature gradient change, and compare it with the actual temperature cloud map; S10: Evaluate the interpolation accuracy of each layout of tunnel fire temperature field reconstruction: Take the ceiling temperature data of a certain working condition, use the three interpolation methods in step S7, and cross-validate the results; evaluate the accuracy through the mean relative error (MRE); evaluate the robustness through the two indicators of mean absolute error (MAE) and root mean square error (RMSE); and derive an interpolation method suitable for generating tunnel fire temperature cloud maps.

2. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: The formula for the heat radiation range of the fire source in step S1 is: （1） Where: Q rf is the radiant heat received by adjacent vehicles, kW / m 2 ; is the transmittance of air to thermal radiation, which is 1; r is the distance between the target object and the virtual point source, m; θ is the angle between the connecting line and the normal direction of the target object; Q r is the external radiation power of the virtual point source, MW; Q r =x r Q, χ r is the flame radiation fraction, ranging from 0.2 to 0.32, the fire source fuel is anhydrous ethanol and is 0.25, and Q is the heat release rate of the fire source.

3. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 2, characterized in that: The formula for the flue gas heat radiation range in step S1 is: （2） Where: Q rs Indicates the smoke radiation heat received by adjacent vehicles, W / m 2 ; ε r represents the flue gas emissivity, which is 0.8; T represents the flue gas temperature, K.

4. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: The fitting function relationship in step S2 is: （3） Where: t is the ignition time, s; is the radiation heat flux kW / m 2 ; Goodness of fit (R 2 ) is 0.99, and the two are in a power exponential relationship. The ignition time decreases according to the power law with the increase of radiation heat flux; The expected ignition time is: （4） The actual ignition time is: （5） Where: A time collection representing the fire development growth period and the stable period.

5. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: Based on step S5, the numerical simulation adopts FDS, and the FDS acquisition frequency is set to 3 times / s; 1050 rows of data including temperature, wind speed, and heat release rate (HRR) can be obtained for each working condition to expand the data set.

6. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: In step S6, the temperature sensors at both ends of the tunnel top are 0.5 m away from the left and right side walls, and the distance between adjacent temperature sensors is 5-10 m.

7. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: The Kriging interpolation expression in the two-dimensional interpolation method in step S7 is: （6） Where: z(x) is the unknown position The estimated value of; n is the number of known observation points; Z i For a known point x i The actual observation value at; λi is the interpolation weight; The inverse distance weighted interpolation expression is: （7） Where: is the number of known observation points; It is The attribute value of each observation point; For the The weight of each observation point on P, ; Indicates the point to be interpolated With Observation points The Euclidean distance between is the weight exponent, which is used to control the speed at which the weight decays with distance; The expression of spline interpolation is: Every cubic polynomial It is expressed as: （8） Where: a i , b i , c i and d i are the coefficients of the polynomial; to ensure In x i Passing through point x i ,y i ,have ,mean ; The smoothness conditions for cubic spline interpolation include continuity, first-order derivative continuity, and second-order derivative continuity: （9） These conditions are transformed into a series of linear equations, thus forming a linear equation system. By solving this linear equation system, we can get , and The value of , thus obtaining the complete cubic spline interpolation function.

8. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: The physical model of tunnel ceiling temperature decay in step S8 is specifically: （10） Where: is the temperature at the reference position; The distance from the reference position Temperature at m; is the initial ambient temperature; K is the attenuation coefficient; the temperature unit is K.

9. A method for reconstructing a time-varying temperature field of a highway tunnel fire according to claim 1, characterized in that: The average relative error in step S10 is specifically: （11） The mean absolute error is: （12） The root mean square error is specifically: （13） Where: Indicates the temperature value of each temperature measuring point; n is the number of samples; is the estimated value of temperature.

10. A system based on a method for reconstructing time-varying temperature fields in highway tunnel fires, characterized in that: include: at least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, the instructions being used to be executed by the processor to implement the method for reconstructing the time-varying temperature field of a highway tunnel fire as described in any one of claims 1 to 9.

Images