Machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method

Abstract

The present application relates to a kind of gas dust explosion prevention and control target explosion suppression agent regulation method based on machine learning, belong to explosion prevention and control and material modification technical field.The method is by machine learning model analysis explosion suppression agent under the inhibition effect of composite explosion environment, and based on experimental data, characteristic characterization, to explosion suppression agent composition, concentration and physical structure are targeted optimization.Specific steps include: by literature research screening explosion suppression agent composition;Explosion suppression agent is modified and matched and collects explosion characteristic data;Characteristic importance analysis is carried out using XGBoost model, and the key factor influencing explosion suppression effect is identified;According to model analysis result, optimize explosion suppression agent formula, and verify its effect by experiment.The method greatly improves the inhibition efficiency and adaptability of explosion suppression agent under the condition of composite explosion, realizes the intelligent design of data-driven, and is suitable for a variety of industrial explosion prevention and control scene.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of explosion prevention and control and material modification, and relates to a gas dust explosion prevention and control targeted explosion inhibitor regulation method based on machine learning. BACKGROUND

[0002] In the industrial and scientific research fields, gas dust composite explosion accidents are common, especially in high-risk environments such as mining, petroleum and chemical industry, tunnel construction and dust treatment. The risk of gas and dust composite explosion poses a great threat to personnel safety and facility stability. Traditional prevention and control methods are difficult to fully respond to these complex explosion situations, and the inhibitory effect of powder explosion inhibitor as one of the core prevention and control measures is often limited by the experience-based design and the simplicity of the modification method. The commonly used explosion inhibitors on the market mainly rely on the physical properties of the materials and the simple compounding of the inhibiting components. However, in the face of complex explosion forms such as gas dust composite explosion, traditional explosion inhibitors are insufficient in key factors such as free radical capture, thermal stability and transient diffusion performance. The existing explosion inhibitors often have problems such as thermal instability, uneven dispersion and low free radical capture efficiency in explosion sites, resulting in limited explosion inhibition effect. At the same time, the current explosion inhibitor development lacks systematic data analysis and model prediction means, and the formula optimization mainly relies on experience, which is difficult to adapt to the requirements of variable working conditions and performance under extreme explosion conditions.

[0003] In recent years, with the development of big data and machine learning technology, data-driven material design has gradually emerged in the field of industrial prevention and control. Through machine learning modeling, the composition, concentration and physical structure of the explosion inhibitor can be precisely controlled to improve its inhibitory effect in complex working conditions. However, the existing technology has not been widely applied in the targeted design and optimization of explosion inhibitors, and there is still a lack of systematic research framework and implementation path. For example, the invention with publication number CN202410912979.3 proposes a high polymer material additive ratio optimization method based on big data, which selects the Me element and doping amount constituting the binary magnesium hydride alloy through the optimal prediction model, and constructs a composite hydrogen storage and release material model structure through the obtained alloy; the invention with publication number CN202111360926.8 proposes a composite material hydrogen storage container layer angle design method based on machine learning, which extracts the element information in the finite element model and establishes a database using a normalization function; taking the output of the neural network as the optimization target, the optimal layering angle of the composite material is obtained under the premise that the inner container size and the layering number remain unchanged. The invention with publication number CN202410437061.8 discloses a refractory material firing tunnel kiln temperature intelligent prediction method based on machine learning, which improves the accuracy and reliability of tunnel kiln temperature prediction and completes the intelligent prediction of tunnel kiln temperature. The current data-driven artificial intelligence method has not been used in laboratory explosion tests, and sufficient attention has not been paid to the optimization design of explosion inhibitors.

[0004] The technical problem at present is how to realize the formula optimization of explosion suppression agent under the condition of complex gas dust composite explosion through data driving and intelligent means. The traditional experimental method has problems of high cost, long cycle and low efficiency, and it is difficult to systematically analyze the explosion suppression effect under the combination of multiple components and parameters. In the multi-field coupling situation, the design of explosion suppression agent needs to consider factors such as thermal stability, free radical capture efficiency and physical diffusion characteristics, and these complex influencing factors are difficult to be accurately analyzed and quantified in the traditional method.

[0005] Therefore, the present application provides a gas dust explosion prevention and control targeted explosion suppression agent regulation method based on machine learning. The method realizes the accurate optimization of the formula of the explosion suppression agent through experimental data acquisition, characteristic representation and training and analysis of the machine learning model. Through feature importance analysis, the key factors that significantly affect the explosion suppression effect are selected, so that the composition, concentration and structure parameters of the explosion suppression agent are targeted adjusted to improve its inhibition efficiency and stability in the composite explosion environment. The method uses data-driven prediction and analysis means to significantly shorten the formula optimization cycle and reduce the research and development cost, and provides an intelligent and systematic design idea for the prevention and control of gas dust composite explosion, filling the gap of the existing technology in multi-factor collaborative optimization.

[0006] In summary, the present application realizes intelligent targeted regulation of the explosion suppression agent by introducing machine learning technology under data driving, breaks through the limitations of traditional experience design, provides a scientific basis for safety prevention and control under complex composite explosion conditions, and has important application value and innovative significance. SUMMARY

[0007] Therefore, the present application provides a gas dust explosion prevention and control targeted explosion suppression agent regulation method based on machine learning. The method realizes the accurate optimization of the formula of the explosion suppression agent through experimental data acquisition, characteristic representation and training and analysis of the machine learning model. Through feature importance analysis, the key factors that significantly affect the explosion suppression effect are selected, so that the composition, concentration and structure parameters of the explosion suppression agent are targeted adjusted to improve its inhibition efficiency and stability in the composite explosion environment. The method uses data-driven prediction and analysis means to significantly shorten the formula optimization cycle and reduce the research and development cost, and provides an intelligent and systematic design idea for the prevention and control of gas dust composite explosion, filling the gap of the existing technology in multi-factor collaborative optimization.

[0008] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0009] The gas dust explosion prevention and control targeted explosion suppression agent regulation method based on machine learning comprises the following steps:

[0010] S1: Modification and compounding of explosion suppression agent: according to the screened explosion suppression agent components, the chemical composition and physical structure are optimized, and the inhibition effect in the gas dust composite explosion is improved through modification and compounding treatment;

[0011] S2: Explosion characteristic experiment and data collection: Under controlled conditions, the modified explosion suppressant is used for gas dust composite explosion pre-experiment, and the characteristic data of explosion pressure and explosion temperature are collected to provide a basic data set for the model;

[0012] S3: Characterization of modified explosion suppressant and database construction: The particle size distribution, specific surface area, chemical composition and surface functional groups, and thermal stability of the modified explosion suppressant are characterized, and the explosion characteristic database of gas dust composite explosion suppression is established based on the experimental data;

[0013] S4: Training and feature analysis of machine learning model: Based on the data set in the database, the model is trained using machine learning algorithm, the explosion suppression characteristic prediction model is constructed, and the correlation analysis and importance ranking of the output key features are analyzed;

[0014] S5: Targeted explosion suppressant optimization design: According to the feature importance analysis results of the machine learning model, the composition, concentration and structure parameters of the explosion suppressant are adjusted by chemical modification and regulation method to design and optimize the targeted explosion suppressant formula to improve the explosion suppression effect;

[0015] S6: Experimental verification and effect evaluation: Under the same experimental conditions, the optimized targeted explosion suppressant is verified by experiment, the explosion suppression effect in gas dust composite explosion is evaluated, and the experimental results are compared with the model prediction results to further optimize the explosion suppressant formula or modify the model.

[0016] Further, the machine learning algorithm is XGBoost algorithm.

[0017] Further, the explosion suppressant composition includes fly ash and ammonium polyphosphate.

[0018] Further, the modification method includes acid-base excitation method, co-condensation and amide impregnation.

[0019] Further, the characterization method includes dynamic light scattering method, specific surface area analysis, energy scattering spectrum, Fourier infrared spectrum and thermal gravimetric analysis.

[0020] Further, the targeted optimization design includes increasing the specific surface area, surface modification, adjusting the particle size distribution, optimizing the inhibitor mass fraction, increasing the proportion of irregular particles, adding a thermal stability enhancer, surface chemical modification and introducing a surface modifier.

[0021] Further, the experimental verification includes measuring the pressure and temperature parameters during the explosion process.

[0022] Further, the experimental verification results are compared with the model prediction results for further optimization of the explosion suppressant formula or modification of the model.

[0023] The beneficial effects of the present application are:

[0024] (1) Reducing cost and experimental period: through the combination of model prediction and experimental verification, the present application reduces the trial and error cost and the number of experiments, making the design process of the explosion inhibitor more efficient and reliable, and greatly shortening the research and development period.

[0025] (2) High efficiency in inhibiting composite explosion reaction: through chemical modification of the explosion inhibitor composition and optimization design of the physical structure, the present application significantly improves the inhibition efficiency of the explosion inhibitor in gas-dust composite explosion, effectively blocks the explosion reaction chain, reduces the reactant generation, delays or prevents the explosion propagation, thereby improving the industrial safety.

[0026] (3) Strong adaptability: the present application adopts multi-level optimization of the explosion inhibitor characteristics, including particle size, morphology, surface functional group and thermal stability, etc., to ensure that the explosion inhibitor has stronger stability and diffusion performance under extreme conditions such as high temperature and high pressure, and is suitable for various composite explosion scenes, with wide adaptability.

[0027] (4) Data-driven targeted optimization: based on the feature analysis and optimization process of machine learning, the present application can identify the key factors that significantly affect the explosion inhibition effect through data-driven method, so as to perform targeted design and dynamic optimization on the explosion inhibitor, and realize precise adaptation to different explosion conditions.

[0028] (5) Improving the safety and reliability of industrial application: the targeted explosion inhibitor formula optimization method of the present application has higher explosion inhibition efficiency and stability in complex industrial environment, significantly improves the safety prevention and control effect in gas-dust composite explosion environment, and provides reliable technical support for high-risk scenes such as mines, chemical industry and tunnels.

[0029] Other advantages, objects and features of the present application will be described in the following specification to some extent, and will be apparent to those skilled in the art based on the study of the following text, or can be taught from the practice of the present application. The objects and other advantages of the present application can be achieved and obtained by the following specification. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to make the objects, technical solutions and advantages of the present application clearer, the preferred detailed description of the present application will be made below in combination with the drawings, in which:

[0031] Figure 1 is the overall flowchart of the present application;

[0032] Figure 2 is the physical and chemical property characterization of the modified explosion inhibitor;

[0033] Figure 3a graph of experimental data collection results for explosion characteristics;

[0034] Figure 4 a graph of prediction data analysis of the XGBoost model;

[0035] Figure 5 a graph of feature correlation analysis of the machine learning model;

[0036] Figure 6 a graph of feature importance analysis of the machine learning model;

[0037] Figure 7 a graph of prediction data analysis of the XGBoost model after updating the database. DETAILED DESCRIPTION

[0038] The present application can be implemented or applied in other different specific embodiments, and various modifications or changes can be made to the details based on different views and applications without departing from the spirit of the present application. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present application in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0039] The accompanying drawings are only used for exemplary illustration, and the representation is only a schematic diagram, not a physical diagram, and should not be understood as a limitation on the present application; in order to better illustrate the embodiments of the present application, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual product size; it is understandable to those skilled in the art that some well-known structures and their descriptions in the drawings may be omitted.

[0040] The same or similar reference numerals in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it should be understood that if the terms "upper", "lower", "left", "right", "front", "back" and the like indicate the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationship in the drawings are only used for exemplary illustration, and should not be understood as a limitation on the present application, for those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0041] Figure 1 is a general flowchart of the present application, showing the steps of the gas dust explosion prevention and control targeting explosion inhibitor regulation method based on machine learning. Figure 2To characterize the physical and chemical properties of the modified explosion suppressant, including particle size distribution, specific surface area, thermogravimetric, adsorption test and other important characteristic data. Figure 3 To collect experimental data of explosion characteristics, the results show the dynamic changes of pressure, temperature and other parameters of different component explosion suppressants in composite explosion. Figure 4 To analyze the prediction data of the XGBoost model, the model's prediction effect on explosion characteristics is shown. Figure 5 To analyze the feature correlation of the machine learning model, the Pearson correlation analysis shows the positive and negative influence of different explosion suppressant features on the explosion suppression effect. Figure 6 To analyze the feature importance of the machine learning model, the ranking shows the influence of different explosion suppressant features on the explosion suppression effect. Figure 7 To analyze the prediction data of the XGBoost model after updating the database, the actual suppression effect of the designed target explosion suppressant is matched with the model prediction results.

[0042] The present application provides a gas and dust explosion prevention and control targeted explosion suppressant regulation method based on machine learning. The method is a gas and dust explosion prevention and control targeted explosion suppressant regulation method based on machine learning. Specifically, the molecular simulation method comprises the following specific steps:

[0043] The specific steps of the embodiment are as follows:

[0044] S1 Literature research and theoretical analysis: based on the theory of explosion inhibitor material science and explosion dynamics, literature research is carried out, and potential modified explosion suppressant components and modification regulation methods for composite explosion are preliminarily screened to provide theoretical basis for modification and compounding

[0045] Firstly, literature research and theoretical analysis are carried out, based on the existing explosion inhibitor material science and composite explosion dynamics theory, the chemical components and physical properties suitable for gas and coal dust composite explosion inhibitor are screened out. Combined with relevant theoretical knowledge, the characteristics of different materials in chain reaction hindering, free radical trapping, thermal stability and surface functional group are analyzed, and the candidate explosion suppressant components with high explosion suppression potential are determined. The screened components will be used for subsequent experiments and model analysis to provide scientific basis for modification and compounding.

[0046] S2 Modification and compounding of explosion suppressant: according to the screened explosion suppressant components, the chemical composition and physical structure are optimized, and the inhibition effect in gas and dust composite explosion is improved through modification and compounding treatment

[0047] According to the explosion suppressant components screened in S1, the chemical composition and physical structure are optimized through modification and compounding treatment to improve the inhibition effect in gas and dust composite explosion environment. The specific operation includes:

[0048] With modified fly ash (FAC) as the inner shell carrier, after carbon residue treatment, the fly ash activity is excited by acid-base excitation method to obtain nitrogen-containing modified MFAC1 and MFAC2, and on this basis, ammonium polyphosphate is impregnated by co-condensation and amidation to obtain phosphorus-containing modified PMFAC1 and PMFAC2

[0049] S3 Explosion characteristic experiment and data collection: Under controlled conditions, the modified explosion inhibitor is used for gas dust composite explosion pre-experiment, and the explosion pressure is collected as characteristic data to provide a basic data set for the model

[0050] Under the standard experimental conditions of a 20-L explosion sphere, the modified fly ash explosion inhibitor is used for gas dust composite explosion pre-experiment, and key characteristic data during the explosion process are collected, including explosion pressure, explosion temperature (such as Figure 3 The experimental data serve as a basic data set for the machine learning model, ensuring that the model can accurately capture the influence of explosion inhibitor composition and modification conditions on composite explosion inhibition effect, facilitating subsequent model training and feature analysis

[0051] S4 Characterization of modified explosion inhibitor and database construction: The particle size distribution, specific surface area, chemical composition and surface functional groups, and thermal stability of the modified explosion inhibitor are characterized, and a gas dust composite explosion inhibition characteristic database is established in combination with the explosion characteristic data obtained by experiment

[0052] The physical and chemical characteristics of the modified explosion inhibitor are systematically characterized by dynamic light scattering, specific surface area analysis, energy scattering spectroscopy, and thermogravimetric analysis, including particle size distribution, specific surface area, chemical composition, surface functional groups, and thermal stability (see Figure 2). The experimental data is integrated with the characterization results to establish a composite explosion suppressant characteristic database, including: x1: specific surface area, x2: particle size D50 (pm), x3: particle size D90 (pm), x4: surface area average particle size (pm), x5: volume average particle size (pm), x6: aspect ratio S50, x7: convexity S50, x8: elongation S50, x9: roundness, x10: sphericity S50, x11: peak value of 3410 in FTIR, x12: peak value of 3430 in FTIR, x13: peak value of 3220 in FTIR, x14: peak value of 2920 in FTIR, x15: FTIR peak value 2130, x16: peak value 1630 in FTIR, x17: FTIR 1690 peak value, x18: FTIR 1430 peak value, x19: peak value of 1250 in FTIR, x20: peak value of 1090 in FTIR, x21: peak value of 1020 in FTIR, x22: peak value of 788 in FTIR, x23: peak value of 100°C in thermal gravimetric curve, x24: peak value of 300 in TG curve, x25: peak value of 550 in TG curve, x26: peak value of 700 in thermal gravimetric curve, x27: time, x28: mass fraction of inhibitor, to provide system input features for the machine learning model, ensuring that the model can effectively identify and extract key factors that have a significant impact on the explosion suppression effect.

[0053] S5 Training of machine learning model and feature analysis: based on the data set in the database, the model is trained using machine learning algorithms, a prediction model of explosion suppression characteristics is constructed, and the correlation analysis and importance ranking of output key features are analyzed

[0054] Based on the database constructed in S4, the data set is trained using the XGBoost machine learning algorithm to construct a prediction model of explosion suppression characteristics, and the prediction results are as shown in Figure 4 The correlation between the physical and chemical characteristics of the explosion suppressant can be observed through feature correlation analysis (as shown in Figure 5 The results of feature importance analysis are as shown in Figure 6 The key factors that significantly affect the explosion suppression effect are determined, including time (x27), mass fraction of inhibitor (x28), specific surface area (x1), particle size D50 (x2), particle size D90 (x3), etc. These key characteristics play a dominant role in the prediction of the explosion suppression effect by the model, providing a scientific basis for the formulation optimization of explosion suppressants. The model training process combines cross-validation and accuracy evaluation to ensure the adaptability and prediction accuracy of the model under different experimental conditions. The feature analysis results of machine learning in this step provide a scientific basis for the formulation optimization of explosion suppressants.

[0055] Optimized design of S6 targeted explosion suppressant: Based on the feature importance analysis results of the machine learning model, the composition, concentration, and structural parameters of the explosion suppressant were adjusted through chemical modification and control methods to design an optimized targeted explosion suppressant formulation to improve the explosion suppression effect.

[0056] Based on the feature importance analysis results output by the machine learning model, targeted optimization design is achieved by adjusting the composition, concentration, and structural parameters of the explosion suppressant. Specifically: 1. Select nanoscale materials or further refine the explosion suppressant particles through grinding to increase the specific surface area; 2. Modify the surface of the material, such as by adding functional groups, to further improve the reactivity of the explosion suppressant surface; 3. Appropriately increase the small particle size component (e.g., reduce D50 to below 750 μm) to improve the rapid diffusion ability of the explosion suppressant in the initial stage of the explosion, thereby accelerating the suppression of the explosion reaction chain; 4. Optimize the mass fraction of the inhibitor according to the dust concentration and scale of the explosion environment to ensure that it reaches an effective suppression concentration in the initial stage of the explosion; 5. Increase the proportion of irregularly shaped explosion suppressant particles (appropriately reduce roundness) to allow them to suspend more effectively in the explosion environment and improve spatial coverage; 6. Add thermal stability enhancers (such as metal oxide nanoparticles) to improve the material's resistance to high temperatures, thereby enhancing its effectiveness in resisting high-temperature explosion environments. 7. Surface chemical modification of the explosion suppressant, using plasma surface treatment or surface coating with functional polymers to increase the number of functional groups such as hydroxyl and carbonyl groups, thereby improving its free radical scavenging ability; 8. Introduction of surface modifiers to enhance the stability and activity of functional groups, ensuring a strong suppressive effect in high-temperature, high-pressure explosive environments, or wet chemical oxidation treatment. The optimized formulation will possess higher composite explosion suppression efficiency, effectively improving the prevention and control effect.

[0057] S7 Experimental Verification and Effect Evaluation: Under the same experimental conditions, the optimized targeted explosion suppressant was experimentally verified to evaluate its suppression effect in gas-dust combined explosions. The experimental results were compared with the model prediction results to further optimize the explosion suppressant formulation or modify the model.

[0058] Under the same experimental conditions as S3, the targeted explosion suppressor designed in S6 was experimentally verified, and key data on the pressure explosion characteristics during the explosion process were measured. The experimental results were compared and analyzed with the prediction results of the machine learning model, such as... Figure 7 As shown, the accuracy of the model predictions is verified. If there are deviations between the experimental results and the predictions, the explosion suppressant formulation can be further optimized or the model database can be updated to ensure the stability and reliability of the explosion suppressant in practical applications, and to achieve targeted optimization of explosion suppressants for gas and coal dust combined explosions.

[0059] Finally, it is to be explained that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the purpose and scope of the technical solutions, and all should be covered in the scope of the claims of the present application.

Claims

1. A machine learning-based method for targeted suppression agent regulation in gas dust explosion prevention, characterized in that: The method comprises the following steps: S1: modification of the explosion inhibitor: according to the screened explosion inhibitor components, the chemical composition and physical structure are optimized, and the inhibition effect in the gas-dust composite explosion is improved through modification and compounding treatment; wherein the explosion inhibitor components include fly ash and ammonium polyphosphate; S2: explosion characteristic experiment and data collection: under the control condition, the modified explosion inhibitor is used for gas-dust composite explosion pre-experiment, the characteristic data of explosion pressure and explosion temperature are collected, and the basic data set is provided for the model; S3: characterization of the modified explosion inhibitor and database construction: the particle size distribution, specific surface area, chemical composition and surface functional group, and thermal stability of the modified explosion inhibitor are characterized, the explosion characteristic data obtained by the experiment are combined, and a gas-dust composite explosion inhibition characteristic database is established; S4: training and feature analysis of machine learning model: based on the data set in the database, the model is trained by using XGBoost algorithm, the explosion inhibition characteristic prediction model is constructed, and the correlation analysis and importance ranking of the output key features are analyzed; S5: optimization design of targeted explosion inhibitor: according to the feature importance analysis result of the XGBoost model, the composition, concentration and structure parameters of the explosion inhibitor are adjusted by chemical modification and regulation method, the optimized targeted explosion inhibitor formula is designed to improve the explosion inhibition effect; S6: experimental verification and effect evaluation: under the same experimental conditions, the optimized targeted explosion inhibitor is verified by experiment, the inhibition effect in the gas-dust composite explosion is evaluated, and the experimental results are compared with the model prediction results, so as to further optimize the explosion inhibitor formula or modify the model.

2. The machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method according to claim 1, characterized in that: The modification method comprises acid-alkali excitation method, co-condensation and amide impregnation.

3. The machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method according to claim 1, characterized in that: The characterization method comprises dynamic light scattering method, specific surface area analysis, energy scattering spectrum, Fourier infrared spectrum and thermogravimetric analysis.

4. The machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method according to claim 1, characterized in that: The targeted optimization design includes increasing the specific surface area, surface modification, adjusting the particle size distribution, optimizing the inhibitor mass fraction, increasing the proportion of irregular particles, adding a thermal stability enhancer, surface chemical modification and introducing a surface modifier.

5. The machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method according to claim 1, characterized in that: The experimental verification includes measuring the pressure and temperature parameters in the explosion process.

6. The machine learning-based gas dust explosion prevention and control targeting explosion inhibitor regulation method according to claim 1, characterized in that: The experimental verification results are compared with the model prediction results for further optimization of the explosion inhibitor formula or modification of the model.

Images