Method and system for determining non-freezing thickness of fire-fighting pipeline thermal insulation material in low-temperature environment

Abstract

The invention provides a method for determining the non-freezing thickness of a fire-fighting pipeline thermal insulation material in a low-temperature environment, and belongs to the field of digital calculation or data processing suitable for specific application, and the method comprises the following steps: S1, collecting experimental data of a fire-fighting pipeline containing a thermal insulation material in the low-temperature environment; s2, establishing a numerical model of the fire-fighting pipeline containing the thermal insulation material, and verifying based on experimental data; s3, a convective heat transfer coefficient formula is obtained; s4, changing the thickness of the thermal insulation material and the environment temperature, and obtaining the non-freezing time of the numerical model under different working conditions; s5, establishing and optimizing a machine learning model; and S6, predicting the thickness of the thermal insulation material of the fire-fighting pipeline in the required non-freezing time. According to the method, the influence of the environment temperature and the thickness of the thermal insulation material on the convective heat transfer coefficient and the deviation between the theoretical convective heat transfer coefficient and the actual value are considered, a correction formula is provided, and the accuracy of numerical simulation is improved; and a method of combining numerical simulation and machine learning is adopted, so that the difficulty of determining the thickness of the thermal insulation material is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of digital computing or data processing suitable for specific applications, and in particular to a method and system for determining the non-freezing thickness of a fire protection pipeline insulation material in a low-temperature environment. Background Art

[0002] Due to the extremely low temperatures in winter, outdoor fire-fighting pipes frequently encounter freezing problems, so anti-freezing and insulation measures must be implemented. The national standard "Technical Specifications for Fire-fighting Water Supply and Fire Hydrant Systems, GB50974-2014" has made corresponding provisions for fire-fighting pipes in low-temperature environments: overhead water-filled pipes should be set in areas where the ambient temperature is not less than 5°C, and anti-freezing measures should be taken when the ambient temperature is below 5°C. However, the relevant regulations do not give detailed insulation material selection and thickness recommendations; at the same time, the experiment of freezing of fire-fighting pipes at low temperatures is not only time-consuming, but also the numerical simulation and its boundary conditions are quite complex, especially the difficulty in determining the heat transfer coefficient, which brings new challenges to determining the non-freezing thickness required for fire-fighting pipe insulation materials in low-temperature environments.

[0003] In order to calculate the thickness of the insulation material under a specific environment, the publication number CN113484360A discloses a method for calculating the anti-condensation thickness of the rubber-plastic insulation material of the metal water supply pipe, which determines the medium temperature in the pipe and the external environment temperature of the pipe by analyzing the composition of the insulation layer and the length of the pipe in the actual project, analyzing the outer surface temperature of the insulation layer of the water supply pipe and the ambient temperature, determining the convective heat exchange between the outer surface of the pipe protective layer and the external environment, analyzing the temperature of the metal pipe wall and the medium temperature, determining the thermal conductivity value of the metal pipe wall, and determining the thickness of the insulation layer according to the convective heat exchange between the outer surface of the pipe protective layer and the external environment and the thermal conductivity value of the metal pipe wall. However, as a measurement method, it is always necessary to obtain various basic parameters and recalculate them for water supply pipes located in different regions or environments to obtain the most suitable thickness of the insulation material, which is difficult and time-consuming. Summary of the invention

[0004] The technical problem to be solved by the present invention is how to quickly and accurately determine the most suitable anti-freezing thickness of the fire protection pipe insulation material in a low temperature environment.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for determining the non-freezing thickness of the heat-insulating material of a fire-fighting pipeline in a low-temperature environment, comprising the following steps:

[0006] S1: Collect experimental data of fire protection pipes containing insulation materials in low temperature environment;

[0007] S2: Establish a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data;

[0008] S3: Obtaining a formula for determining the convective heat transfer coefficient, the thickness of the insulation material, and the temperature difference;

[0009] S4: Change the thickness of the insulation material and the ambient temperature to obtain the freezing-free time of the fire protection pipeline numerical model under different working conditions;

[0010] S5: Build and optimize machine learning models;

[0011] S6: Predict the thickness of the insulation material for fire protection pipes under the required freezing-free time.

[0012] Preferably, in step S1, the experimental data includes pipeline type, insulation material type, thermal conductivity of insulation material, thickness of insulation material, ambient temperature and pipe wall temperature.

[0013] Preferably, in step S2, the specific process of verifying the numerical model based on experimental data is: setting a temperature sensor on the outermost side of the insulation material to obtain the ambient temperature as experimental data, and continuously changing the convective heat transfer coefficient of the outer wall of the insulation material in the numerical model until the outermost temperature of the insulation material in the numerical model is consistent with the experimental data, and determining the convective heat transfer coefficient under the corresponding working conditions.

[0014] Preferably, in step S3, the formula for determining the convective heat transfer coefficient, the thickness of the thermal insulation material, and the temperature difference is: Where h is the convective heat transfer coefficient, α is determined by the temperature drop data of the insulation material, k is the thermal conductivity of the ambient air, R is the outer diameter of the fire protection pipe, D is the thickness of the insulation material, Pr and Gr are the Prandtl number and Grashof number of the ambient air respectively, and C and n are constants.

[0015] The invention takes into account the deviation between the theoretical convection heat transfer coefficient and the actual value, proposes a correction formula, and improves the accuracy of numerical simulation.

[0016] Preferably, in step S4, the specific process of obtaining the non-freezing time is: a liquid fraction sensor is set in the outermost layer of the fire-fighting pipe to obtain the liquid fraction, and the time from the beginning to the liquid fraction of the outermost layer of the fire-fighting pipe becomes 0.02 is taken as the non-freezing time.

[0017] Preferably, in step S5, the machine learning model is a neural network model.

[0018] Preferably, the input of the neural network model is the characteristic parameters of the thermal insulation material, the ambient temperature and the time to keep free of freezing, and the output is the thickness of the thermal insulation material.

[0019] Preferably, in step S5, the process of optimizing the machine learning model is: optimizing the number of hidden layers, training function and transfer function thereof by trial and error method, and determining the number of hidden layers, training function and transfer function of the machine learning model based on the mean square error between the predicted value and the true value when the mean square error is minimized.

[0020] Preferably, the formula for determining the number of hidden layers is: Where m is the number of hidden layers, a and b are the number of input layers and output layers respectively, and t is an integer;

[0021] The training functions include trainlm, traingd, and traingda;

[0022] The transfer functions include tansig and logsig.

[0023] The present invention adopts the trial and error method to optimize the number of hidden layers, training functions and transfer functions, thereby improving the accuracy of the machine learning model and being able to more accurately predict the thickness of the insulation material of the fire-fighting pipe under the required non-freezing time.

[0024] Corresponding to the above method, the present invention also provides a system for determining the non-freezing thickness of the fire protection pipeline insulation material in a low temperature environment, comprising the following modules:

[0025] Experimental data collection module: used to collect experimental data of fire protection pipes containing insulation materials in low temperature environments;

[0026] Numerical model building module: used to build a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data;

[0027] Module for obtaining the formula of convection heat transfer coefficient: used to obtain the formula for determining the convection heat transfer coefficient, the thickness of the insulation material, and the temperature difference;

[0028] Obtaining non-freezing time module: used to obtain the non-freezing time of the fire protection pipeline numerical model under different working conditions when the thickness of the insulation material and the ambient temperature are changed;

[0029] Machine learning module: used to build and optimize machine learning models;

[0030] The module for obtaining the thickness of thermal insulation materials is used to predict the thickness of thermal insulation materials for fire protection pipes under the required time to remain free of freezing.

[0031] The advantages of the present invention are:

[0032] (1) The influence of ambient temperature and insulation material thickness on the convective heat transfer coefficient was considered, and the deviation between the theoretical convective heat transfer coefficient and the actual value was considered. A correction formula was proposed to improve the accuracy of numerical simulation;

[0033] (2) A method combining numerical simulation and machine learning was used to reduce the difficulty of determining the thickness of the insulation material;

[0034] (3) It can be easily applied to various scenarios where it is necessary to determine the thickness of the pipeline insulation material without repeated calculations. After the machine learning model is determined, it is only necessary to input the characteristic parameters of the insulation material, the ambient temperature, and the required anti-freezing time to determine the thickness of the insulation material. This method is simple, fast, and highly applicable. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of Embodiment 1 of the present invention;

[0036] Figure 2 is a schematic diagram of a numerical model in Example 1 of the present invention;

[0037] Figure 3 It is a fitting diagram of the formula for determining the convective heat transfer coefficient of Example 1 of the present invention;

[0038] Figure 4 This is a performance graph of the machine learning model constructed in Example 1 of the present invention on the test set. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Example 1

[0041] like Figure 1 This embodiment provides a method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment, and the specific process is as follows:

[0042] Step 1: Collect experimental data of fire protection pipes containing insulation materials in a low temperature environment, wherein the experimental data mainly include pipe type, insulation material type, thermal conductivity of insulation material, thickness of insulation material, and experimental environment temperature and pipe wall temperature data that change with time;

[0043] Step 2: Establish a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data. The specific process is as follows:

[0044] S201: According to the actual size parameters and physical parameters of the fire protection pipeline, establish Figure 2The numerical model of the fire protection pipe with insulation material shown, where D is the thickness of the insulation material and R is the outer diameter of the fire protection pipe;

[0045] The numerical model parameters in this embodiment are shown in the following table:

[0046]

[0047] S202: A temperature sensor is set at the outermost side of the insulation material to obtain the ambient temperature as experimental data. The parameters of the above-mentioned experimental conditions are used to continuously change the convective heat transfer coefficient of the outer wall of the insulation material in the numerical model. The trial and error method is used to make the outermost temperature of the insulation material in the numerical model consistent with the experimental data, and the convective heat transfer coefficient under the corresponding conditions is determined.

[0048] Step 3: By continuously determining the convective heat transfer coefficient under different working conditions, the formula for determining the convective heat transfer coefficient, the thickness of the insulation material, and the temperature difference is obtained: Wherein h is the convective heat transfer coefficient, k is the thermal conductivity of the ambient air, R is the outer diameter of the fire pipe, and D is the thickness of the insulation material; Pr and Gr are the Prandtl number and Grashof number of the ambient air, respectively; α is a parameter to be determined, which is specifically determined according to the temperature drop data of each insulation material; C and n are constants, and preferably, C and n are 0.48 and 0.25, respectively;

[0049] Example: The fitting results of the convective heat transfer coefficient of the aluminum silicate shell as the insulation material are as follows: Figure 3 As shown, the fitting formula is:

[0050]

[0051] Step 4: Change the thickness of the insulation material and the ambient temperature to obtain the freezing-free time of the fire protection pipe numerical model under different working conditions. The specific process is as follows:

[0052] S401: changing the thickness of the insulation material and the ambient temperature, and obtaining the convective heat transfer coefficient of the numerical model under different working conditions according to the convective heat transfer coefficient determination formula determined in step 3. The different working conditions of the numerical simulation in this embodiment are shown in the following table:

[0053]

[0054] S402: A liquid fraction sensor is set at the outermost layer of the pipeline to obtain the liquid fraction. The time from the beginning to the liquid fraction of the outermost node becoming 0.02 is defined as the non-freezing time, and the non-freezing time under different working conditions is obtained.

[0055] Step 5: According to the non-freezing time obtained under different working conditions in step 4, with the characteristic parameters of the insulation material, ambient temperature, and non-freezing time as input, and the thickness of the insulation material as output, a data set is constructed, and the data set is divided into a training set and a test set in a ratio of 7:3. A machine learning model is established, and the mean square error (MSE) between the predicted value and the true value is used as the basis for optimizing the machine learning model.

[0056] The machine learning model in this embodiment adopts a neural network model, and the basis for optimizing the model is as follows:

[0057] S501: Optimize the number of hidden layers: The formula for determining the number of hidden layers is: Where m is the number of hidden layers, a and b are the number of input layers (two in this embodiment, representing the ambient temperature and the thickness of the insulation material) and the number of output layers (one in this embodiment, representing the time to keep the temperature from freezing), respectively, and t is a constant, taking an integer of 2-10;

[0058] The number of hidden layers is optimized by trial and error method. The optimization results of the number of hidden layers in this embodiment are shown in the following table:

[0059]

[0060] It is found from the above table that the average MSE is the smallest when the number of hidden layers is 12, that is, the best number of hidden layers in this embodiment is 12;

[0061] S502: Optimize the training function, the optimization function includes trainlm, traingd, and traingda; use the trial and error method to optimize the training function. The optimization results of the training function in this embodiment are shown in the following table:

[0062]

[0063] It is found from the above table that the average MSE is the smallest when the training function is trainlm, that is, trainlm is the best training function in this embodiment;

[0064] S503: Optimizing the transfer function, the transfer function includes tansig and logsig; using a trial and error method to optimize the transfer function, the optimization results of the transfer function in this embodiment are shown in the following table:

[0065]

[0066] It is found from the above table that the average MSE is the smallest when the transfer function is tansig, that is, tansig is the best transfer function in this embodiment.

[0067] In particular, this embodiment performs prediction verification on the neural network model on the training set. The prediction results on the test set are compared with the simulation results to obtain the performance diagram of the neural network model on the test set. Figure 4 As shown, the symbols are the correspondence between the predicted value and the simulated value, and the line is the theoretical reference line. It is observed that the regression values ​​are all near the theoretical reference line, indicating that the prediction accuracy of the neural network model in this embodiment is relatively high.

[0068] Step 6: Use the optimized neural network model to predict the thickness of the insulation material for the fire-fighting pipe under the required non-freezing time. That is, input the characteristic parameters of the insulation material, ambient temperature, and non-freezing time into the optimized neural network model, and the thickness of the insulation material for the fire-fighting pipe under the required non-freezing time can be output.

[0069] This embodiment takes into account the impact of ambient temperature and changes in the thickness of the insulation material on the convective heat transfer coefficient, takes into account the deviation between the theoretical convective heat transfer coefficient and the actual value, proposes a correction formula, and improves the accuracy of the numerical simulation; a method combining numerical simulation and machine learning is adopted to reduce the difficulty of determining the thickness of the insulation material; at the same time, the prediction speed of the machine learning model is extremely fast, which not only ensures the prediction accuracy, but also greatly reduces the time and difficulty of the prediction.

[0070] Example 2

[0071] Corresponding to Embodiment 1 of the present invention, this embodiment provides a system for determining the non-freezing thickness of a fire protection pipe insulation material in a low temperature environment, comprising the following modules:

[0072] The experimental data collection module is used to collect experimental data of fire protection pipes containing insulation materials in low temperature environments. The experimental data mainly include pipe type, insulation material type, thermal conductivity of insulation material, thickness of insulation material, and experimental environment temperature and pipe wall temperature data that change with time.

[0073] Numerical model building module: used to build a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data. It specifically includes the following units:

[0074] Establishing numerical model unit: used to establish numerical model of fire protection pipeline containing thermal insulation material according to actual size parameters and physical property parameters of fire protection pipeline;

[0075] Verification of numerical model unit: It is used to determine the convective heat transfer coefficient under the corresponding working conditions by trial and error method while continuously changing the convective heat transfer coefficient of the outer wall of the insulation material in the numerical model using the parameters of the experimental conditions until the outermost temperature of the insulation material in the numerical model is consistent with the ambient temperature obtained by the outermost temperature sensor of the insulation material.

[0076] Module for obtaining the formula for the convective heat transfer coefficient: used to obtain the formula for determining the convective heat transfer coefficient, the thickness of the insulation material, and the temperature difference. The specific formula obtained is: Where h is the convective heat transfer coefficient, k is the thermal conductivity of the ambient air, R is the outer diameter of the fire protection pipe, and D is the thickness of the insulation layer; Pr and Gr are the Prandtl number and Grashof number of the ambient air respectively, α is the parameter to be determined, which is determined according to the temperature drop data of each insulation material, and C and n are constants.

[0077] Obtaining the non-freezing time module: It is used to obtain the non-freezing time of the fire protection pipe numerical model under different working conditions when the thickness of the insulation material and the ambient temperature are changed.

[0078] Machine Learning Module: used to build and optimize machine learning models, including the following units:

[0079] Establishing a machine learning model unit: It is used to construct a data set based on the non-freezing time obtained under different working conditions in the non-freezing time acquisition module, with the characteristic parameters of the insulation material, the ambient temperature, and the non-freezing time as inputs and the thickness of the insulation material as outputs, and divide the data set into a training set and a test set at a ratio of 7:3 to establish a machine learning model;

[0080] Optimizing the number of hidden layers: This unit is used to optimize the number of hidden layers using a trial-and-error method. The formula for determining the number of hidden layers is: Where m is the number of hidden layers, a and b are the number of input layers and output layers respectively, t is a constant, which is an integer between 2 and 10. The number of hidden layers with the minimum MSE is taken as the optimization result.

[0081] Optimizing training function unit: used to optimize the training function by trial and error method, wherein the optimization functions include trainlm, traingd and traingda, and the training function with the minimum MSE is taken as the optimization result;

[0082] Optimizing transfer function unit: used to optimize the transfer function by trial and error method, the transfer function includes tansig and logsig, and the transfer function with the minimum MSE is taken as the optimization result.

[0083] The module for obtaining the thickness of thermal insulation materials is used to predict the thickness of thermal insulation materials for fire protection pipes under the required time to remain free of freezing.

[0084] This embodiment first uses the experimental data collection module to collect experimental data of fire-fighting pipes containing insulation materials under low temperature environment; then uses the numerical model establishment module to establish a numerical model of fire-fighting pipes containing insulation materials, and verifies the model based on experimental data; then obtains the determination formula of convective heat transfer coefficient, insulation material thickness, and temperature difference through the convective heat transfer coefficient formula module; considers the deviation between the theoretical convective heat transfer coefficient and the actual value to make the simulation more accurate; uses the non-freezing time acquisition module to obtain the non-freezing time of the fire-fighting pipe numerical model under different working conditions; also uses the machine learning module to establish and optimize the machine learning model, and uses the trial and error method to optimize the number of hidden layers, training functions, and transfer numbers of the machine learning model, so that the accuracy of the machine learning model prediction is greatly improved; finally, uses the insulation material thickness acquisition module to predict the insulation material thickness of the fire-fighting pipe under the required non-freezing time. This system is simple and easy to implement, and has the advantages of accurate and fast prediction, simplifies the selection process, and opens up a new way for the selection of non-freezing thickness of fire-fighting pipe materials under low temperature environment.

[0085] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining the non-freezing thickness of fire protection pipe insulation materials in a low temperature environment, characterized in that: The following steps are involved: S1: Collect experimental data of fire protection pipes containing insulation materials in low temperature environment; S2: Establish a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data; S3: Obtaining a formula for determining the convective heat transfer coefficient, the thickness of the insulation material, and the temperature difference; S4: Change the thickness of the insulation material and the ambient temperature to obtain the freezing-free time of the fire protection pipeline numerical model under different working conditions; S5: Build and optimize machine learning models; S6: Predict the thickness of the insulation material for fire protection pipes under the required freezing-free time.

2. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 1 is characterized in that: In step S1, the experimental data includes pipeline type, insulation material type, insulation material thermal conductivity, insulation material thickness, ambient temperature and pipe wall temperature.

3. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 1, characterized in that: In step S2, the specific process of verifying the numerical model based on experimental data is: setting a temperature sensor on the outermost side of the insulation material to obtain the ambient temperature as experimental data, and continuously changing the convective heat transfer coefficient of the outer wall of the insulation material in the numerical model until the outermost temperature of the insulation material in the numerical model is consistent with the experimental data, and determining the convective heat transfer coefficient under the corresponding working conditions.

4. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 3 is characterized in that: In step S3, the formula for determining the convective heat transfer coefficient, the thickness of the insulation material, and the temperature difference is: Where h is the convective heat transfer coefficient, α is determined by the temperature drop data of the insulation material, k is the thermal conductivity of the ambient air, R is the outer diameter of the fire protection pipe, D is the thickness of the insulation material, Pr and Gr are the Prandtl number and Grashof number of the ambient air respectively, and C and n are constants.

5. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 1, characterized in that: In step S4, the specific process of obtaining the non-freezing time of the fire protection pipe numerical model under different working conditions is as follows: a liquid fraction sensor is set along the diameter direction of the fire protection pipe to obtain the liquid fraction of different nodes, and the time from the beginning to the time when the liquid fraction of the outermost layer of the fire protection pipe becomes 0.02 is taken as the non-freezing time.

6. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 1, characterized in that: In step S5, the machine learning model is a neural network model.

7. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 6, characterized in that: The input of the neural network model is the characteristic parameters of the thermal insulation material, the ambient temperature and the time to keep free of freezing, and the output is the thickness of the thermal insulation material.

8. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 7, characterized in that: In step S5, the process of optimizing the machine learning model is: using the trial and error method to optimize the number of hidden layers, training function and transfer function, based on the mean square error between the predicted value and the true value, when the mean square error is minimized, the number of hidden layers, training function and transfer function of the machine learning model are determined.

9. The method for determining the non-freezing thickness of the fire protection pipe insulation material in a low temperature environment according to claim 8, characterized in that: The formula for determining the number of hidden layers is: Where m is the number of hidden layers, a and b are the number of input layers and output layers respectively, and t is an integer; The training functions include trainlm, traingd, and traingda; The transfer functions include tansig and logsig.

10. A system for determining the non-freezing thickness of fire protection pipe insulation materials in a low temperature environment, characterized in that: Includes the following modules: Experimental data collection module: used to collect experimental data of fire protection pipes containing insulation materials in low temperature environments; Numerical model building module: used to build a numerical model of fire protection pipes containing insulation materials and verify the model based on experimental data; Module for obtaining the formula of convection heat transfer coefficient: used to obtain the formula for determining the convection heat transfer coefficient, the thickness of the insulation material, and the temperature difference; Obtaining non-freezing time module: used to obtain the non-freezing time of the fire protection pipeline numerical model under different working conditions when the thickness of the insulation material and the ambient temperature are changed; Machine learning module: used to build and optimize machine learning models; The module for obtaining the thickness of thermal insulation materials is used to predict the thickness of thermal insulation materials for fire protection pipes under the required time to remain free of freezing.

Images