Method for rapidly predicting surface tension of foam extinguishing agent on site

Abstract

The invention provides a method for rapidly predicting the surface tension of a foam extinguishing agent on site, which comprises the following steps: S1, collecting the surface tension of different types of foam extinguishing agents under different temperature conditions; s2, taking the first surface tension and the temperature measured by the thermocouple as input data, and taking the second surface tension as output data; s3, establishing a machine learning model; s4, after the model parameters are initialized, establishing a mapping relation between the input data and the output data; s5, the first surface tension and the temperature are collected again to serve as input data, and the machine learning model is adopted to predict data of the second surface tension. According to the method, the maximum bubble pressure method is improved based on high-speed camera shooting and image recognition, temperature compensation and a machine learning model are combined to predict the surface tension of the foam extinguishing agent on the fire extinguishing and rescue site, and the requirements that on-site surface tension measurement operation is convenient, the device structure is simple, and the surface tension measurement result is high in accuracy can be met.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of fire fighting, and particularly relates to a method for rapidly predicting surface tension of foam extinguishing agent on site. BACKGROUND

[0002] Foam film-forming property is a key factor determining the fire extinguishing efficiency of water film-forming foam extinguishing agent, and surface tension is an important parameter affecting the film-forming behavior. In the fire rescue site, rapid and accurate detection of the surface tension of water film-forming foam extinguishing agent is of great significance for evaluating its performance quality. Commonly used surface tension measurement methods include capillary rise method, platinum ring method, platinum plate method, spinning drop method, pendant drop method and maximum bubble pressure method. Among them, the maximum bubble pressure method is not sensitive to surface contamination and is suitable for complex high-temperature environments in fire sites. The platinum ring method has higher requirements for the cleanliness of the ring and the sample cell, and is usually limited to laboratory conditions, which is difficult to apply to complex environments such as high temperature and high pressure. Therefore, portable instruments based on the maximum bubble pressure method are often used for surface tension measurement on site.

[0003] According to the search, the following disclosed patents related to the present application are found, and the specific disclosure contents are as follows:

[0004] Patent document CN117054295A discloses a bubble pressure method dynamic surface tension testing device and testing method. It includes a pressure sensor, a gas pump, a control unit, a flow sensor and a needle, and further includes a gas processing unit, the gas processing unit is connected with a three-way valve, one interface of the three-way valve is connected with an exhaust throttle valve, and the other interface is in communication with the gas pump; the patent adds a three-way valve and an exhaust throttle valve at the input air end, and adds a gas storage cavity pressure plate and a silicone rubber pressure reduction film on the gas processing unit, which effectively solves the problems of being easily affected by bubble running pulse, being affected by additional pressure of the gas pump, inaccurate calculation of dynamic surface tension time, and inability to test the equilibrium surface tension of the surfactant in the bubble pressure method.

[0005] Although the above-mentioned patent uses the maximum bubble pressure method to measure the surface tension, the bubble generation rate is not easy to accurately control, and the accuracy of the maximum pressure value read at the moment of bubble detachment cannot be guaranteed, resulting in poor accuracy of the surface tension measurement result. SUMMARY

[0006] Therefore, the present application aims to provide a method for rapidly predicting surface tension of foam extinguishing agent on site, so as to solve the problem of poor accuracy of surface tension measurement result caused by the maximum bubble pressure method in the prior art, which is difficult to accurately control the bubble generation rate, and the accuracy of the maximum pressure value read at the moment of bubble detachment cannot be guaranteed.

[0007] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0008] A method for rapidly predicting the surface tension of foam extinguishing agents on-site includes the following steps:

[0009] S1. Collect surface tension data of different types of foam extinguishing agents under different temperature conditions;

[0010] S2. Set the first surface tension and temperature as input data, set the second surface tension as output data, and preprocess the input and output data;

[0011] S3. Establish a machine learning model for predicting the surface tension of foam extinguishing agents based on temperature compensation;

[0012] S4. After initializing the model parameters, use the first surface tension and temperature as input data and the second surface tension as output data to establish a mapping relationship between the input data and the output data.

[0013] S5. Re-collect the first surface tension and temperature as input data, and use a machine learning model to predict the second surface tension data to obtain a rapid and accurate on-site prediction result of the surface tension of the foam extinguishing agent.

[0014] Furthermore, in step S1, the maximum bubble pressure method involves integrating a small high-speed camera, an LED fiber optic cold light source, a thermocouple, and an embedded microprocessor into the measuring device body. A capillary tube is inserted into the liquid, and gas is blown in to form a bubble. The LED fiber optic cold light source provides illumination to the bubble. The small high-speed camera captures dynamic images of bubble growth. An image recognition method is used to determine whether the bubble radius is consistent with the capillary radius. The internal pressure value of the bubble when they are consistent is extracted, and the surface tension is calculated according to the Laplace formula.

[0015] Furthermore, in step S1, the different types of foam extinguishing agents are 1% AFFF foam extinguishing agent, 3% AFFF foam extinguishing agent, 6% AFFF foam extinguishing agent, 3% AFFF / AR foam extinguishing agent, or 6% AFFF / AR foam extinguishing agent.

[0016] Furthermore, in step S1, the temperature range is 5-50℃.

[0017] Furthermore, in step S2, the surface tension measured by the maximum bubble pressure method and the temperature measured by the thermocouple are used as input data, and the input vector is represented as [γ_mbp, T].

[0018] The surface tension measured by the platinum ring method is the output data, and the output vector is represented as [γ_ring].

[0019] The preprocessing of input and output data includes normalizing the input data and mapping the input vector to the [0,1] region using the Min-Max method.

[0020]

[0021] in, This is the original data; The smallest data point in the original dataset; The largest data point in the original dataset; The data has been normalized.

[0022] The preprocessing of the output data is performed using the following formula:

[0023]

[0024] in, This is the original data; This is the processed data.

[0025] Furthermore, in step S3, the machine learning model for rapidly predicting the surface tension of foam extinguishing agents on-site is an artificial neural network (ANN) model or other machine learning models containing ANN. The multi-layer feedforward network of the artificial neural network model consists of an input layer, a hidden layer, and an output layer, with each unit layer consisting of several artificial neurons. The input layer has 2 artificial neurons, and the input layer data is the surface tension and temperature measured by the maximum bubble pressure method. The output layer has 1 artificial neuron, and the output layer data is the predicted surface tension data measured by the platinum ring method. The hidden layer has 3 hidden layers, and each hidden layer has 6 neurons. The data partitioning mode of the artificial neural network model is that the data samples are randomly divided into training data, validation data, and test data.

[0026] Furthermore, in step S5, a machine learning model is used to predict the data of the second surface tension, and the prediction accuracy is evaluated using the root mean square error (RMSE). The formula for calculating RMSE is as follows.

[0027]

[0028] in, The number of data points; and These are the experimental value and the predicted value, respectively.

[0029] Compared with existing technologies, the method for rapidly predicting the surface tension of foam extinguishing agents on-site according to the present invention has the following beneficial effects:

[0030] (1) The present invention provides a method for rapid on-site prediction of the surface tension of foam extinguishing agent. It uses the maximum bubble pressure method and the platinum ring method to measure surface tension data. The maximum bubble pressure method is improved based on high-speed photography and image recognition. It combines temperature compensation and machine learning models to predict the surface tension of foam extinguishing agent at the fire extinguishing and rescue site. It can meet the requirements of convenient on-site surface tension measurement operation, simple device structure and high accuracy of surface tension measurement results.

[0031] (2) The present invention provides a method for rapidly predicting the surface tension of foam extinguishing agents on site. This method utilizes the advantages of the maximum bubble pressure method, which is highly resistant to the environment and suitable for on-site use. Combined with the high-precision data of the laboratory standard method (platinum ring method), a machine learning dataset and prediction model based on temperature compensation are constructed. This method can quickly estimate the surface tension closer to the true value in complex and high-temperature fire extinguishing environments, effectively solving the problem that traditional methods are difficult to balance between ease of operation and measurement accuracy under on-site conditions.

[0032] (3) The method for rapid on-site prediction of surface tension of foam extinguishing agent described in this invention achieves real-time and efficient prediction of surface tension by establishing a machine learning model represented by artificial neural network. It is easy to operate, does not require complex and expensive equipment, can meet the needs of rapid on-site detection, and provides reliable data support for the quality judgment of fire extinguishing agent and the evaluation of fire extinguishing effectiveness at fire scene. Attached Figure Description

[0033] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0034] Figure 1 This is a schematic flowchart of a method for rapidly predicting the surface tension of foam extinguishing agents in the field, as described in an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of a method for rapidly predicting the surface tension of foam extinguishing agents on-site, as described in an embodiment of the present invention.

[0036] Figure 3 The surface tension prediction result A is the artificial neural network training data described in this embodiment of the invention.

[0037] Figure 4 The surface tension prediction result B is the artificial neural network verification data described in this embodiment of the invention.

[0038] Figure 5 C represents the surface tension prediction result of the artificial neural network test data described in this embodiment of the invention. Detailed Implementation

[0039] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0040] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0041] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0042] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0043] A method for rapidly predicting the surface tension of foam extinguishing agents on-site includes the following steps:

[0044] S1. Thermocouples are placed inside the container of the liquid to be tested. The surface tension of different types of foam extinguishing agents under different temperature conditions is collected in a laboratory environment using the maximum bubble pressure method and the platinum ring method.

[0045] S2. The surface tension measured by the maximum bubble pressure method is the first surface tension. The first surface tension and the temperature measured by the thermocouple are used as input data. The surface tension measured by the platinum ring method is the second surface tension. The second surface tension is used as output data. The input data and output data are preprocessed.

[0046] S3. Establish a machine learning model for predicting the surface tension of foam extinguishing agents based on temperature compensation;

[0047] S4. After initializing the model parameters, use the first surface tension and temperature as input data and the second surface tension as output data to establish a mapping relationship between the input data and the output data.

[0048] S5. Re-collect the first surface tension and temperature as input data, and use a machine learning model to predict the second surface tension data to obtain a rapid and accurate on-site prediction result of the surface tension of the foam extinguishing agent.

[0049] In step S1, the maximum bubble pressure method integrates a small high-speed camera, an LED fiber optic cold light source, a thermocouple, and an embedded microprocessor into the measuring device body. In this embodiment, the measuring device body is a K100C fully automatic surface tension meter, and the small high-speed camera, LED fiber optic cold light source, thermocouple, and embedded microprocessor are all existing technologies. A capillary tube is inserted into the liquid and gas is blown in to form a bubble. The LED fiber optic cold light source provides illumination to the bubble, and the small high-speed camera captures dynamic images of bubble growth. An image recognition method is used to determine whether the bubble radius is consistent with the capillary radius. The internal pressure value of the bubble when they are consistent is extracted, and the surface tension is calculated according to the Laplace formula.

[0050] This paper proposes a method to measure surface tension using the maximum bubble pressure method and the platinum ring method. The maximum bubble pressure method is improved based on high-speed photography and image recognition. Combined with temperature compensation and a machine learning model, this method predicts the surface tension of foam extinguishing agents at fire-fighting and rescue sites. It meets the requirements of convenient operation, simple device structure, and high accuracy in surface tension measurement. This method leverages the advantages of the maximum bubble pressure method—its strong environmental tolerance and suitability for field use—and combines high-precision data from the standard laboratory method (platinum ring method) to construct a temperature-compensated machine learning dataset and prediction model. This allows for the rapid estimation of surface tension values ​​closer to the true values ​​in complex, high-temperature fire-fighting environments, effectively solving the problem of traditional methods struggling to balance operational convenience and measurement accuracy under field conditions.

[0051] In a laboratory environment, the surface tension values ​​(γ-ring) of different types of foam extinguishing agents under different temperature conditions were collected using the platinum ring method. The platinum ring method was performed using the SITA DynoTester+, and the different types of foam extinguishing agents mentioned in step S1 were 1% AFFF foam extinguishing agent, 3% AFFF foam extinguishing agent, 6% AFFF foam extinguishing agent, 3% AFFF / AR foam extinguishing agent, or 6% AFFF / AR foam extinguishing agent. The temperature range for data collection was 5-50℃.

[0052] The input data consists of surface tension measured by the maximum bubble pressure method and temperature measured by thermocouples, denoted as [γ_mbp, T]. The output data consists of surface tension measured by the platinum ring method, denoted as [γ_ring].

[0053] The input and output data are preprocessed, including normalizing the input data and mapping the input vector to the [0,1] region using the Min-Max method.

[0054]

[0055] in, This is the original data; The smallest data point in the original dataset; The largest data point in the original dataset; This is the data after normalization.

[0056] The preprocessing of the output data is performed using the following formula:

[0057]

[0058] in, This is the original data; This is the processed data.

[0059] Before evaluating the predictive performance, the data needs to be inversely transformed to remap the prediction results back to the original data range. The prediction performance of the machine learning is then evaluated using the inversely transformed prediction results.

[0060] A machine learning model, represented by an artificial neural network, was established for rapid on-site prediction of the surface tension of foam extinguishing agents. Model parameter initialization included randomly setting initial values ​​for weights and corresponding biases. The multi-layered feedforward network of the artificial neural network model consisted of an input layer, hidden layers, and an output layer, with each unit layer composed of several artificial neurons. The input layer had 2 artificial neurons, and the input data consisted of surface tension and temperature measured by the maximum bubble pressure method; the output layer had 1 artificial neuron, and the output data consisted of predicted surface tension measured by the platinum ring method; the hidden layers had 3 layers, each with 6 neurons. The data partitioning pattern for the artificial neural network model involved randomly dividing the data samples into training data, validation data, and test data. 70% of the data was used for training (30 sets of experimental data), 15% for validation (7 sets of experimental data), and 15% for testing (7 sets of experimental data). The model was trained using the training data, and the surface tension prediction result A of the artificial neural network training data is shown below. Figure 3As shown in Table 1, the experimental and predicted values ​​of the training data are presented. The model was evaluated using validation data, and the surface tension prediction result B of the artificial neural network validation data is shown in the table. Figure 4 As shown in Table 2, the experimental and predicted values ​​of the validation data are presented. The surface tension prediction result C of the artificial neural network test data is shown in Table 2. Figure 5 As shown in Table 3, the experimental and predicted values ​​of the test data are presented.

[0061] Table 1. Experimental and predicted values ​​of surface tension from artificial neural network training data.

[0062]

[0063] Table 2. Experimental and predicted values ​​of surface tension from artificial neural network validation data.

[0064]

[0065] Table 3. Experimental and predicted values ​​of surface tension from artificial neural network training data.

[0066]

[0067] Table 4. Surface tension data measured by the maximum bubble pressure method and the platinum ring method.

[0068]

[0069] The trained and validated machine learning model was used to predict the surface tension test data. The prediction accuracy was evaluated using the root mean square error (RMSE), which is calculated using the following formula.

[0070]

[0071] in, The number of data points; and These are the experimental value and the predicted value, respectively.

[0072] The root mean square errors (RMSE) of the constructed machine learning model for rapid on-site prediction of foam extinguishing agent surface tension were 0.405, 0.83, and 0.93 for the training, validation, and test data, respectively. According to Table 4, the RMSE of the surface tension not predicted by the machine learning model (measured by the maximum bubble pressure method) was 1.68. The prediction results of the machine learning model for rapid on-site prediction of foam extinguishing agent surface tension constructed in this invention show good agreement with the experimental data.

[0073] By establishing machine learning models represented by artificial neural networks, real-time and efficient prediction of surface tension can be achieved. The operation is simple and does not require complex and expensive equipment. It can meet the needs of rapid on-site detection and provides reliable data support for judging the quality of fire extinguishing agents and evaluating fire extinguishing effectiveness at fire scenes.

[0074] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for rapidly predicting the surface tension of foam extinguishing agents on-site, characterized in that: Includes the following steps, S1. Collect surface tension data of different types of foam extinguishing agents under different temperature conditions; S2. Set the first surface tension and temperature as input data, set the second surface tension as output data, and preprocess the input and output data; S3. Establish a machine learning model for predicting the surface tension of foam extinguishing agents based on temperature compensation; S4. After initializing the model parameters, the first surface tension and temperature in step S2 are used as input data, and the second surface tension is used as output data to establish a mapping relationship between input data and output data. S5. Re-collect the first surface tension and temperature as input data, and use a machine learning model to predict the second surface tension data to obtain a rapid and accurate on-site prediction result of the surface tension of the foam extinguishing agent.

2. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S1, a thermocouple needs to be placed inside the container holding the liquid to be tested. The surface tension is collected in a laboratory environment using the maximum bubble pressure method and the platinum ring method. The maximum bubble pressure method integrates a small high-speed camera, an LED fiber optic cold light source, a thermocouple, and an embedded microprocessor into the measuring device body. A capillary tube is inserted into the liquid and gas is blown in to form a bubble. The LED fiber optic cold light source provides illumination to the bubble. The small high-speed camera collects dynamic images of bubble growth. The image recognition method is used to determine whether the bubble radius is consistent with the capillary radius. The internal pressure value of the bubble when the two are consistent is extracted, and the surface tension is calculated according to the Laplace formula.

3. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S1, different types of foam extinguishing agents are AFFF foam extinguishing agents or AFFF / AR foam extinguishing agents with different contents.

4. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S1, the temperature range is 5-50℃.

5. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S2, the first surface tension is the surface tension measured by the maximum bubble pressure method, and the first surface tension and the temperature measured by the thermocouple are used as input data, and the input vector is represented as [γ_mbp, T]. The second surface tension is the surface tension measured by the platinum ring method. The surface tension measured by the platinum ring method is the output data, and the output vector is represented as [γ_ring]. The preprocessing of input and output data includes normalizing the input data and mapping the input vector to the [0,1] region using the Min-Max method. in, This is the original data; This is the smallest data point in the original dataset; The largest data point in the original dataset; The data has been normalized. The preprocessing of the output data is performed using the following formula: in, This is the original data; This is the processed data.

6. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S3, the machine learning model for rapidly predicting the surface tension of foam extinguishing agents on-site includes at least an artificial neural network model. The multi-layer feedforward network of the artificial neural network model consists of an input layer, a hidden layer, and an output layer, with each unit layer consisting of several artificial neurons. The input layer has 2 artificial neurons, and the input layer data is the surface tension and temperature measured by the maximum bubble pressure method. The output layer has 1 artificial neuron, and the output layer data is the predicted surface tension data measured by the platinum ring method. The hidden layer has 3 hidden layers, and each hidden layer has 6 neurons. The data partitioning mode of the artificial neural network model is that the data samples are randomly divided into training data, validation data, and test data.

7. The method for rapidly predicting the surface tension of foam extinguishing agents on-site according to claim 1, characterized in that: In step S5, the machine learning model is used to predict the data of the second surface tension. The prediction accuracy is evaluated by the root mean square error. The formula for calculating the root mean square error is as follows. in, The number of data points; and These are the experimental value and the predicted value, respectively.

Images