Combustible inversion method based on combustion product marker

Abstract

The invention belongs to the technical field of combustible material type identification and monitoring, and discloses a combustible material inversion method based on combustion product markers. According to the method, the artificial intelligence dynamic inversion algorithm model is constructed to invert the type and the content of the combustible, intelligent analysis and processing of multi-dimensional data are realized, the complex relationship between the combustion product and the combustible can be accurately analyzed, accurate inversion can also be performed in a mixed combustion scene, and the accuracy of combustible identification is improved. According to the method, qualitative analysis and quantitative analysis of the characteristic markers are performed according to original detection data through gas chromatography-mass spectrometry detection analysis, the method is suitable for various organic compounds and part of inorganic compounds, and the research and application range is greatly expanded. The method can be suitable for a smoldering stage or a combustion stage of a fire scene, combustible identification and analysis can be carried out in real time, a basis is provided for restraining fire behavior expansion, and fire loss is reduced to the maximum extent.

Description

Technical Field

[0001] The invention belongs to the technical field of combustible material type identification and monitoring, and relates to a combustible material inversion method based on combustion product markers. Background Art

[0002] When fighting a fire, accurately understanding the combustible conditions in the fire scene is crucial for implementing effective fire-fighting measures. However, the current means of obtaining combustible information in the fire scene are relatively limited, mainly relying on manual observation and questioning of informed people, and there are problems such as inaccurate information. Therefore, the research on combustible inversion based on combustion product markers is of great significance.

[0003] There are also technical solutions for combustible type identification and monitoring in the prior art. For example, a Chinese invention patent application with publication number CN108181422A is for a simulation method and special simulation equipment for combustion in a special ambient atmosphere, which includes: passing a mixture of an oxidant gas and an ambient atmosphere into a burner to facilitate a combustion reaction; then performing combustion gas detection: by analyzing the gas generated after combustion, the gas composition generated by combustion in a special ambient atmosphere can be determined.

[0004] In addition, a Chinese invention patent application with publication number CN105699236A is for a method for detecting liquid combustibles in a fire environment, which comprises: selecting metal materials at the fire scene, placing them in an experimental container after accurate weighing, and determining whether the fire is caused by liquid combustibles by detecting the weight gain of the metal materials after being affected by the fire.

[0005] Although the above two schemes have proposed some solutions for combustible type identification and monitoring, they still have certain limitations: on the one hand, in scheme one, the combustion product information is obtained through a certain method, and then the combustible information in the fire scene is inverted by manual experience. This analysis method has high requirements for the judge, and it is impossible to accurately know the current type, quantity, and degree of pyrolysis of combustibles. On the other hand, in scheme two, the weight gain of metal materials in the fire scene after being affected by the fire is detected, and then it is determined whether the fire is caused by liquid combustibles. This analysis method greatly reduces the timeliness and availability of fire source locating in the fire scene. Summary of the invention

[0006] In view of this, in order to solve the problems of high manual dependence and low combustible material identification accuracy raised in the above-mentioned background technology, the present invention collects the gases produced by the smoldering pyrolysis of combustibles in the air, uses a gas chromatograph to detect and analyze its composition and concentration, and further uses an artificial intelligence algorithm to analyze and process the complex data of combustion product markers in multiple scenarios and multiple types of fire sources, and explores the combustible material change rules hidden behind the combustion product marker composition and concentration data, thereby opening up new inversion paths and improving inversion accuracy. A combustible material inversion method based on combustion product markers is now proposed.

[0007] The purpose of the present invention can be achieved through the following technical scheme: a combustible inversion method based on combustion product markers, comprising: S1, using a filter to capture aerosols in fire smoke, and using an extraction liquid to oscillate and extract the captured aerosols to obtain a test liquid.

[0008] S2. Performing gas chromatography-mass spectrometry detection and analysis on the test liquid to obtain a chromatographic elution curve, and obtaining corresponding original detection data based on the chromatographic elution curve.

[0009] S3. Perform qualitative and quantitative analysis of characteristic markers based on the original detection data to obtain the characteristic marker type of each effective peak and the effective peak area of ​​each characteristic marker, and analyze the mass fraction and volume fraction of each characteristic marker.

[0010] S4. Import the mass fraction and volume fraction of each characteristic marker into the pre-trained artificial intelligence dynamic inversion algorithm model to invert the type and content of combustibles.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention realizes intelligent analysis and processing of multi-dimensional data by constructing an artificial intelligence dynamic inversion algorithm model to invert the type and content of combustibles, accurately analyzes the complex relationship between combustion products and combustibles, and can accurately invert in mixed combustion scenarios, thereby improving the accuracy of combustible identification.

[0012] (2) The present invention uses gas chromatography-mass spectrometry for detection and analysis, and then performs qualitative and quantitative analysis of characteristic markers based on the original detection data. This analysis method is applicable to all kinds of organic compounds and some inorganic compounds. Whether it is a simple combustion product of a single combustible or a complex flue gas produced by the mixed combustion of multiple combustibles, it can be effectively analyzed, greatly expanding the scope of research and application.

[0013] (3) The present invention can be adapted to the smoldering or burning stage of a fire scene, and can identify and analyze combustibles in real time, accurately capture characteristic markers of trace combustion products, and identify potential combustibles, thereby providing a basis for curbing the spread of fire, assisting in the formulation of efficient fire-fighting plans, and minimizing fire losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0015] Figure 1 It is a schematic diagram for implementing the steps of the method of the present invention.

[0016] Figure 2 A schematic diagram of the flow of gas samples for gas chromatography-mass spectrometry detection and analysis corresponding to an embodiment provided by the present invention.

[0017] Figure 3 A logic flow chart of valid time series identification corresponding to an embodiment provided by the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0019] See also Figure 1 As shown, the present invention provides a combustible inversion method based on combustion product markers, including: S1, using a filter to capture aerosols in fire smoke, and using an extraction liquid to oscillate and extract the captured aerosols to obtain a test liquid.

[0020] It should be noted that the purpose of implementing this step is to obtain a test liquid that contains combustion product markers in fire smoke and is suitable for subsequent detection and analysis. The composition of fire smoke is complex, and the aerosol contains key information related to combustibles, such as specific volatile organic compounds and other combustion product markers. However, it is difficult to detect these markers directly. By capturing the aerosol through a filter, the target substances can be preliminarily enriched. By using the extraction liquid for oscillation extraction, the extraction liquid and the aerosol can be fully contacted according to the principle of like dissolves like, so that the combustion product markers in the aerosol are dissolved in the extraction liquid, thereby obtaining a uniform and easy-to-detect test liquid, which provides a suitable sample for the subsequent determination of the combustible component content by gas chromatography-mass spectrometry.

[0021] In a preferred embodiment of the present invention, the specific operation method of oscillating and extracting the captured aerosol is as follows: immersing the filter obtained by using the filter to capture the aerosol in the fire smoke into a pre-prepared extraction liquid, and then performing an oscillation operation based on a preset vibration frequency and vibration duration, thereby obtaining a liquid to be tested.

[0022] It should be noted that the present invention can be adapted to the smoldering stage or the burning stage of a fire scene, can identify and analyze combustibles in real time, accurately capture characteristic markers of trace combustion products, determine potential combustibles, provide a basis for curbing the spread of fire, assist in the formulation of efficient fire-fighting plans, and minimize fire losses.

[0023] S2. Performing gas chromatography-mass spectrometry detection and analysis on the test liquid to obtain a chromatographic elution curve, and obtaining corresponding original detection data based on the chromatographic elution curve.

[0024] It should be noted that the combustion product markers were inverted using gas chromatography. Chromatography, also known as chromatography or chromatographic method, is a physical and chemical separation method that uses the solubility and adsorption properties of substances. The separation principle is based on the difference in the effects of the components in the mixture between the mobile phase and the stationary phase as the basis for separation.

[0025] In a preferred embodiment of the present invention, please refer to Figure 2 As shown, the specific analysis method of the gas chromatography-mass spectrometry detection and analysis is as follows: the test liquid obtained by oscillation extraction is injected into the gas chromatograph, and then a gasification operation is performed to obtain a gas sample, the gas sample is decompressed by a pressure reducing valve, purified by a drying tube, and then flows into the chromatographic column at a preset pressure and a preset flow rate through a pressure stabilizing valve and a flowmeter for separation, and then the separated components are successively flowed into the detector for detection to obtain a chromatographic elution curve.

[0026] It should be noted that the specific implementation methods and corresponding principles of each step are as follows: 1. Injection into a gas chromatograph and gasification: The test liquid after oscillation extraction is injected into a gas chromatograph and gasified because the separation principle of gas chromatography is based on the distribution behavior of substances in the gaseous state. Only when the test liquid is converted into a gas sample can the various components therein be separated in the chromatographic column driven by the carrier gas. After the liquid becomes a gas, the molecular movement is freer, which is conducive to the subsequent separation operation.

[0027] 2. Decompression: The gas sample is decompressed through the pressure reducing valve to make the gas pressure reach the range required for the normal operation of the gas chromatograph. Excessive pressure may damage the instrument components and affect the flow state and separation effect of the gas in the instrument. Decompression can ensure the stable operation of the instrument and allow the gas sample to enter the subsequent process at an appropriate pressure.

[0028] 3. Purification: Purification through a drying tube is to remove impurities and moisture from the gas sample. Impurities and moisture may contaminate the chromatographic column, affect the separation effect, and may also damage the detector, reducing the accuracy of the test and the service life of the instrument. The purified gas sample is purer and can ensure the reliability of the test results.

[0029] 4. Stable pressure and flow: The gas sample is allowed to flow into the chromatographic column at a preset pressure and flow rate through a pressure regulating valve and flow meter. This is the key to ensuring the stability and repeatability of the separation effect. Stable pressure and flow rate allow different components to be separated in the chromatographic column according to a fixed rule. If the pressure and flow rate are unstable, the residence time of each component in the chromatographic column will change, resulting in a poor separation effect and inability to accurately obtain information on each component.

[0030] 5. Separation detection and obtaining chromatographic elution curve: Different components have different mobility speeds in the chromatographic column due to different interactions with the stationary phase, thus achieving separation. The separated components flow into the detector for detection one after another. The detector converts the concentration or mass change of each component into an electrical signal, records these electrical signals over time, and finally forms a chromatographic elution curve. The chromatographic elution curve contains information such as the retention time, peak area or peak height of each component, which is an important basis for subsequent qualitative and quantitative analysis. By analyzing the curve, it is possible to determine which components are contained in the test liquid (qualitative analysis) and the content of each component (quantitative analysis).

[0031] In a preferred embodiment of the present invention, please refer to Figure 3 As shown, the specific method of obtaining the corresponding original detection data is as follows: Specifically, the original monitoring data includes the number of effective peaks and the retention time corresponding to each effective peak.

[0032] Extract the chromatographic elution curve, obtain the output electrical signal intensity of each time series corresponding to the chromatographic elution curve, and then compare it with the preset identifiable electrical signal intensity. If the output electrical signal intensity of a time series is greater than the identifiable electrical signal intensity, identify the time series as a valid time series.

[0033] It should be further explained that the chromatographic elution curve is a curve showing the change of the signal detected by the detector over time. Its essence is the change of the output electrical signal intensity over the time series. The recognizable electrical signal intensity is pre-set as a threshold. This threshold is determined based on the detection limit of the instrument and experimental experience. Signals below this threshold may be generated by instrument noise or other interference factors and do not represent the true sample component signal. The output electrical signal intensity of each time series in the chromatographic elution curve is compared with the threshold. When the output electrical signal intensity of a time series is greater than the recognizable electrical signal intensity, it is considered that this time series corresponds to the signal generated by a certain component in the sample, and it is identified as a valid time series. This step is to screen out the signal that truly reflects the sample component.

[0034] Based on the continuity of the time series, each valid time series is separated to obtain each valid time series set, and the output electrical signal strength of each valid time series set corresponding to each valid time series is extracted.

[0035] It should be further explained that, since there may be multiple components in the sample, the effective time series generated by these components may overlap or appear at intervals. Based on the continuity of the time series, adjacent effective time series are merged into a set, so that each effective time series set corresponds to a possible sample component peak. For example, if there is a continuous time series that is greater than the recognizable electrical signal intensity, then this section constitutes an effective time series set. If there is a signal interval below the threshold in the middle, it is divided into different sets. At the same time, the output electrical signal intensity of each effective time series in each effective time series set is extracted. This intensity information can be used to further analyze the characteristics of the peak, such as peak height, peak area, etc., in preparation for quantitative analysis.

[0036] The number of each valid time series set is counted and recorded as the number of valid peaks.

[0037] It should be noted that the value obtained by performing quantitative statistics on the obtained valid time series set is the number of valid peaks. This number directly reflects the number of components with obvious signals detected from the sample under the current detection conditions, that is, the number of different characteristic markers that may exist, providing preliminary data support for the subsequent determination of the types of combustible components in the sample.

[0038] The valid time series corresponding to each valid time series set are extracted, and then arranged in chronological order to obtain the end time series and start time series corresponding to each valid time series set. The difference calculation is performed to obtain the duration corresponding to each valid time series set, which is recorded as the retention time of each valid peak.

[0039] S3. Perform qualitative and quantitative analysis of characteristic markers based on the original detection data to obtain the characteristic marker type of each effective peak and the effective peak area of ​​each characteristic marker, and analyze the mass fraction and volume fraction of each characteristic marker.

[0040] It should be explained that characteristic markers refer to chemical substances produced by specific combustibles during the combustion process that can characterize the existence or characteristics of the combustibles. When different types of combustibles are burned, they will produce different types and proportions of combustion products due to differences in their own chemical composition and structure. The representative components of these products that can be used as a basis for judgment are characteristic markers. For example, the burning of wood will produce some unique volatile organic compounds and furans, which can be used as characteristic markers of wood burning; while the burning of petroleum products will produce specific characteristic markers such as aromatic compounds.

[0041] In a preferred embodiment of the present invention, the specific analysis process of the qualitative analysis is as follows: extract the retention time corresponding to each valid peak, and then perform difference calculation with the pre-saved reference retention time of each characteristic marker, and then take the absolute value to obtain the retention time deviation of each valid peak relative to each characteristic marker, arrange the retention time deviation of each valid peak relative to each characteristic marker in order from small to large, and select the characteristic marker corresponding to the smallest retention time deviation as the characteristic marker type corresponding to the valid peak.

[0042] It should be noted that the following explanations are required for qualitative analysis: 1. Basis for comparison of retention time: In chromatographic analysis, the retention time of different substances under the same chromatographic conditions is unique, just like everyone has a unique fingerprint. The reference retention time of each characteristic marker is saved in advance. These reference values ​​are accurate data obtained by multiple measurements under standard conditions, representing the time from injection to peak elution of each characteristic marker under ideal conditions. The retention time of each effective peak is extracted from the actual detected chromatographic elution curve, which is a record of the time when the actual components in the sample elute.

[0043] 2. Calculate the retention time deviation: Calculate the difference between the retention time of each valid peak and the reference retention time of all pre-saved characteristic markers one by one, and then take the absolute value. The absolute value is taken to unify the measurement direction of the deviation, avoid the positive and negative differences from offsetting each other, and make all deviations reflect the actual time difference. For example, the retention time of a valid peak is 10 seconds, and the reference retention time of a characteristic marker is 8 seconds. The calculated difference is 2 seconds. If the reference retention time of another characteristic marker is 11 seconds, the difference is 3 seconds. In this way, the retention time deviation of each valid peak relative to each characteristic marker is obtained. These deviations reflect the degree of difference between the actual detection peak and the standard peak time of the known characteristic marker.

[0044] 3. Determine the type of characteristic marker: Arrange the retention time deviations of each effective peak relative to each characteristic marker in order from small to large. The smaller the deviation, the closer the retention time of the actual detected peak is to the reference retention time of the characteristic marker, which means that the component corresponding to the effective peak is more likely to be this characteristic marker. Therefore, the characteristic marker corresponding to the smallest deviation is selected as the characteristic marker type corresponding to the effective peak. For example, the retention time deviations of an effective peak relative to characteristic markers A, B, and C are 2 seconds, 3 seconds, and 4 seconds, respectively. Then the characteristic marker A corresponding to the smallest deviation of 2 seconds is identified as the characteristic marker type corresponding to the effective peak, thereby completing the preliminary qualitative judgment of the effective peak component.

[0045] In a preferred embodiment of the present invention, the specific analysis process of the quantitative analysis is as follows: using image processing software to obtain the corresponding peak area in each effective time series concentration is recorded as the effective peak area of ​​each effective peak, and then used as the effective peak area of ​​each characteristic marker.

[0046] Using the formula The effective peak area percentage of each characteristic marker was obtained by analysis ,in Indicates The effective peak area of ​​the characteristic markers, The number of the characteristic marker, , Indicates the number of characteristic markers.

[0047] It is necessary to explain the construction logic of the effective peak area percentage of each characteristic marker: Representative The effective peak area of ​​a characteristic marker is a quantitative reflection of the marker in the detection spectrum. The peak area is related to the content and other properties of the marker and is the key data of a single marker. For all The effective peak areas of the characteristic markers are summed to obtain the total peak area of ​​all markers, which reflects the overall situation. The effective peak area of ​​a single marker is divided by the total peak area of ​​all markers, and then multiplied by , the obtained That is The percentage of effective peak area occupied by each characteristic marker in all markers can be used to intuitively compare the relative importance or content ratio of each characteristic marker in the mixture.

[0048] In a preferred embodiment of the present invention, the specific analysis method of the mass fraction and volume fraction is as follows: Extract the effective peak area percentage of each characteristic marker .

[0049] Using the formula Analyze and obtain the quality score corresponding to each characteristic marker ,in Indicates The relative response factor of each characteristic marker.

[0050] It should be explained that the relative response factor is an important parameter used in chromatographic analysis to correct the difference in the response of different substances on the detector, ensuring that the quantitative analysis results more accurately reflect the true content of each substance in the sample. In chromatographic analysis, different substances have different response sensitivities on the same detector. Even if the actual content of two substances is the same, the peak area or peak height they produce may be different. For example, the content of substance A and substance B in the sample is 1%, but due to the different responses of the detector to them, the peak area of ​​substance A may be twice that of substance B. The relative response factor is a parameter introduced to correct this difference.

[0051] It needs further explanation that the construction idea of ​​the formula is: 1. middle, is the previously calculated The effective peak area percentage of a characteristic marker, which reflects the relative proportion of the marker in the peak area of ​​all markers; It is The relative response factor of a characteristic marker is used to correct the difference in the response of different substances to different detection instruments, so that the peak area is more accurately related to the actual amount of substance. The multiplication of the two comprehensively considers the peak area ratio and response characteristics of the marker.

[0052] 2. It is the sum of the "effective peak area percentage × relative response factor" of all characteristic markers, representing the total amount of all markers combined, and is used to provide a total reference standard.

[0053] 3. Divide the numerator by the denominator and then multiply , obtained That is The mass fraction corresponding to each characteristic marker can more accurately reflect the actual mass proportion of each characteristic marker in the mixture and eliminate the impact caused by the difference in response of different substances.

[0054] In a feasible simulation process, it is assumed that there are three characteristic markers ( ), based on the above formula, data simulation is performed to obtain the corresponding simulation calculation results. Some simulation results can be found in Table 1.

[0055] Table 1. Some simulation data and simulation results

[0056]

[0057] In the above simulation results, the mass fraction of characteristic marker 1 is about 32.4%, which means that after considering the peak area percentage and relative response factor, it accounts for about 32.4% of the mass of the mixture. Although its effective peak area percentage is 0.3 (smaller than characteristic marker 2), the relative response factor of 0.8 makes it not far behind in mass percentage.

[0058] The mass fraction of characteristic marker 2 is about 40.5%, which is the highest among the three markers. Its effective peak area percentage is 0.5 and its relative response factor is 0.6. The combination of the two makes it account for a large proportion in the mixture.

[0059] The mass fraction of characteristic marker 3 is about 27.0%, and the effective peak area percentage and relative response factor are 0.2 and 1.0, respectively. Compared with the other two markers, it has the lowest mass proportion in the mixture.

[0060] Using the formula The volume fraction corresponding to each characteristic marker was obtained by analysis ,in Indicates the pre-saved The density of a characteristic marker.

[0061] In a preferred embodiment of the present invention, the specific analysis method of the relative response factor is: standard samples of each characteristic marker of different concentrations are prepared in advance, and then the standard samples of each characteristic marker of different concentrations are detected by gas chromatograph to obtain the standard effective peak area of ​​each characteristic marker of each concentration.

[0062] Impurity samples of each characteristic marker of a specific concentration are prepared in advance, and then the impurity samples of each characteristic marker of a specific concentration are detected by gas chromatography to obtain the control effective peak area of ​​each characteristic marker of the specific concentration.

[0063] Using the formula The relative response factors of each characteristic marker were obtained by analysis. ,in It represents the effective peak area of ​​each characteristic marker at a specific concentration. Indicates Characteristic marker The standard effective peak area corresponding to the standard sample with the same concentration is Indicates The specific concentration of the impurity sample of the characteristic marker, Indicates Characteristic marker The concentration of the standard sample corresponds to The number indicating the concentration, , A quantity that indicates concentration.

[0064] What needs to be explained is the construction idea of ​​the above formula: 1. The calculation is the The ratio of the effective peak area of ​​the reference of a characteristic marker to the concentration. This ratio reflects the peak area corresponding to the unit concentration of the marker under specific conditions, which is a quantitative measure of the response of the marker under this specific reference.

[0065] 2. It is for characteristic markers For standard samples of different concentrations, the ratio of their standard effective peak area to the corresponding concentration is calculated respectively, and then the average of these ratios is calculated. It represents the average response of the marker at multiple different concentrations.

[0066] 3. Divide the numerator by the denominator to get , i.e., the relative response factor, is used to measure the The relative relationship between the response of a characteristic marker at a specific concentration and the average response at multiple concentrations. Through this factor, the response differences of different characteristic markers due to their own differences in properties can be corrected during detection, so as to more accurately analyze the content of each marker and other information in the future.

[0067] In a feasible simulation process, it is assumed that there are two characteristic markers ( ), each marker has 3 different concentrations of standard samples ( ), based on the above formula, data simulation is performed to obtain the corresponding simulation calculation results. Some simulation results can be referred to Table 2.

[0068] Table 2. Partial simulation data and simulation calculation results

[0069]

[0070] The above simulation results show that for characteristic marker 1, the relative response factor of the peak area ratio per unit concentration at a specific concentration compared with the average peak area ratio per unit concentration at multiple concentrations is about 0.54. This shows that in this detection system, the response at a specific concentration is relatively weak compared with the average response at multiple concentrations of the detection system.

[0071] The situation of characteristic marker 2 is similar. The relative response factor of the peak area ratio per unit concentration at a specific concentration is also about 0.54 compared with the average peak area ratio per unit concentration. This shows that the response degree of characteristic marker 2 at a specific concentration is also weaker than its average response degree at multiple concentrations, and is similar to the relative response of characteristic marker 1. The relative response factor can be used for subsequent correction calculations of information such as the characteristic marker content to make the analysis results more accurate.

[0072] It should be noted that the present invention uses gas chromatography-mass spectrometry for detection and analysis, and then performs qualitative and quantitative analysis of characteristic markers based on the original detection data. This analysis method is applicable to all kinds of organic compounds and some inorganic compounds. Whether it is a simple combustion product of a single combustible or a complex flue gas produced by the mixed combustion of multiple combustibles, it can be effectively analyzed, which greatly expands the scope of research and application.

[0073] S4. Import the mass fraction and volume fraction of each characteristic marker into the pre-trained artificial intelligence dynamic inversion algorithm model to invert the type and content of combustibles.

[0074] In a preferred embodiment of the present invention, the specific training method of the artificial intelligence dynamic inversion algorithm model is as follows: extract a large amount of historically stored combustible combustion data, specifically including the type and content of combustibles, and the corresponding post-combustion characteristic marker types, mass fractions of each characteristic marker, and volume fractions.

[0075] The artificial intelligence dynamic inversion algorithm model is obtained by continuous training of the artificial intelligence algorithm based on the above data.

[0076] It is necessary to explain that the training method of the artificial intelligence dynamic inversion algorithm model is to allow the model to learn the intrinsic relationship between the type and content of combustibles and the characteristic markers of combustion products, so as to achieve the goal of accurately inverting combustibles based on combustion products. For example, the combustion product data generated by various combustibles such as wood, plastic, and fabric under different combustion conditions are collected to form a rich data set.

[0077] In a preferred embodiment of the present invention, the specific method of inverting the type and content of combustibles is as follows: extract the mass fraction and volume fraction of each characteristic marker, and input them into the artificial intelligence dynamic inversion algorithm model respectively.

[0078] Obtain the pre-saved combustion records of each combustible material, including the type and content of the combustible material, reference characteristic markers, and reference mass fractions and reference volume fractions of each reference characteristic marker.

[0079] The reference characteristic marker corresponding to each combustible combustion record is compared with the input characteristic marker. If the type of the reference characteristic marker corresponding to a combustible combustion record is consistent with the type of the input characteristic marker, the combustible combustion record is used as the reference combustible combustion record.

[0080] The reference mass fraction and reference volume fraction of each reference characteristic marker corresponding to each reference combustible combustion record are extracted, and then compared with the mass fraction and volume fraction of each input characteristic marker to obtain the mass fraction similarity and volume fraction similarity of each characteristic marker.

[0081] In a feasible embodiment, the specific analysis method of the mass score similarity of each characteristic marker is as follows: the reference mass score of each reference characteristic marker corresponding to each reference combustible combustion record is subtracted from the mass score of each input characteristic marker, and then the absolute value is taken, and then the ratio is calculated with the mass score of each input characteristic marker to obtain the mass score similarity of each characteristic marker.

[0082] In a feasible embodiment, the specific analysis method of the mass fraction similarity of each characteristic marker may refer to the specific analysis method of the volume fraction similarity of each characteristic marker.

[0083] The mass fraction similarity and volume fraction similarity of each characteristic marker are summed up according to the weight to obtain the similarity of each characteristic marker, and then the mean is calculated to obtain the reference evaluation index of each reference combustible combustion record.

[0084] It should be noted that different characteristic markers may have different importance in judging the type and content of combustibles, so corresponding weights are assigned to mass fraction similarity and volume fraction similarity respectively. The determination of weights is usually based on professional knowledge, experience and analysis of a large amount of experimental data. For example, some characteristic markers have a stronger indicative effect on a specific type of combustible, and their corresponding weights will be greater. Exemplarily, the corresponding weight for mass fraction similarity is 0.6, and the corresponding weight for volume fraction similarity is 0.4.

[0085] The reference evaluation indexes of the reference combustible combustion records are compared, the reference combustible combustion record corresponding to the maximum reference evaluation index is selected as the target combustible combustion record, and the combustible type and content corresponding to the target combustible combustion record are output as the combustible type and content.

[0086] It should be noted that the reference evaluation index comprehensively reflects the similarity between the characteristic markers of the currently detected combustion products and the combustion records of each reference combustible. In actual fire scenes, the characteristic markers of combustion products contain key information about combustibles. The combustion of different combustibles will produce characteristic markers with unique proportions and types. The reference evaluation index of each reference combustible combustion record is calculated through the previous steps. By comparing these indexes, the reference record closest to the current combustion product characteristics can be found. For example, there are three reference combustible combustion records A, B, and C, and the corresponding reference evaluation indexes are 0.8, 0.6, and 0.7, respectively. The higher the index, the more similar it is to the current combustion product characteristics, so the record of A has the highest match with the current situation.

[0087] It should be noted that the present invention realizes intelligent analysis and processing of multi-dimensional data by constructing an artificial intelligence dynamic inversion algorithm model to invert the type and content of combustibles. It can accurately analyze the complex relationship between combustion products and combustibles, and can accurately invert in mixed combustion scenarios, thereby improving the accuracy of combustible identification.

[0088] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A combustible inversion method based on combustion product markers, characterized in that: include: S1. Using a filter to capture aerosols in fire smoke, and using an extraction liquid to oscillate and extract the captured aerosols to obtain a test liquid; S2, performing gas chromatography-mass spectrometry detection and analysis on the liquid to be tested to obtain a chromatographic elution curve, and obtaining corresponding original detection data based on the chromatographic elution curve; S3. Performing qualitative and quantitative analysis of characteristic markers based on the original detection data to obtain the characteristic marker type of each effective peak and the effective peak area of ​​each characteristic marker, and analyzing the mass fraction and volume fraction of each characteristic marker; S4. Import the mass fraction and volume fraction of each characteristic marker into the pre-trained artificial intelligence dynamic inversion algorithm model to invert the type and content of combustibles.

2. The combustible material inversion method based on combustion product markers according to claim 1, characterized in that: The specific operation method of oscillating and extracting the captured aerosol is as follows: The filter disc obtained by using the filter disc to capture the aerosol in the fire smoke is immersed in the pre-prepared extraction liquid, and then an oscillation operation is performed based on the pre-set vibration frequency and vibration duration to obtain the test liquid.

3. The combustible material inversion method based on combustion product markers according to claim 2, characterized in that: The specific analysis method of the gas chromatography-mass spectrometry detection analysis is as follows: The test liquid obtained by oscillation extraction is injected into the gas chromatograph, and then a gasification operation is performed to obtain a gas sample. The gas sample is decompressed by a pressure reducing valve, purified by a drying tube, and then flows into the chromatographic column at a preset pressure and preset flow rate through a pressure stabilizing valve and a flowmeter for separation. The separated components are then successively flowed into the detector for detection to obtain a chromatographic elution curve.

4. The combustible material inversion method based on combustion product markers according to claim 1, characterized in that: The specific method of obtaining the corresponding original detection data is as follows: Specifically, the original monitoring data includes the number of effective peaks and the retention time corresponding to each effective peak; Extract the chromatographic elution curve, obtain the output electrical signal intensity of each time series corresponding to the chromatographic elution curve, and then compare it with the preset identifiable electrical signal intensity. If the output electrical signal intensity of a time series is greater than the identifiable electrical signal intensity, identify the time series as a valid time series; Based on the continuity of the time series, each valid time series is separated to obtain each valid time series set, and the output electrical signal strength of each valid time series set corresponding to each valid time series is extracted; Count the number of each valid time series set and record it as the number of valid peaks; The valid time series corresponding to each valid time series set are extracted, and then arranged in chronological order to obtain the end time series and start time series corresponding to each valid time series set. The difference calculation is performed to obtain the duration corresponding to each valid time series set, which is recorded as the retention time of each valid peak.

5. The combustible material inversion method based on combustion product markers according to claim 4 is characterized in that: The specific analysis process of the qualitative analysis is as follows: The retention time corresponding to each valid peak is extracted, and then the difference between the retention time and the pre-saved reference retention time of each characteristic marker is calculated, and the absolute value is taken to obtain the retention time deviation of each valid peak relative to each characteristic marker. The retention time deviations of each valid peak relative to each characteristic marker are arranged in order from small to large, and the characteristic marker corresponding to the smallest retention time deviation is selected as the characteristic marker type corresponding to the valid peak.

6. The combustible material inversion method based on combustion product markers according to claim 5, characterized in that: The specific analysis process of the quantitative analysis is as follows: The corresponding peak area of ​​each effective time series is obtained by using image processing software and recorded as the effective peak area of ​​each effective peak, which is then used as the effective peak area of ​​each characteristic marker; Using the formula The effective peak area percentage of each characteristic marker was obtained by analysis ,in Indicates The effective peak area of ​​the characteristic markers, The number of the characteristic marker, , Indicates the number of characteristic markers.

7. The combustible material inversion method based on combustion product markers according to claim 6, characterized in that: The specific analysis method of the mass fraction and volume fraction is as follows: Extract the effective peak area percentage of each characteristic marker ; Using the formula Analyze and obtain the quality score corresponding to each characteristic marker ,in Indicates The relative response factor of each characteristic marker; Using the formula The volume fraction corresponding to each characteristic marker was obtained by analysis ,in Indicates the pre-saved The density of a characteristic marker.

8. The combustible material inversion method based on combustion product markers according to claim 7, characterized in that: The specific analysis method of the relative response factor is: Preparing standard samples of each characteristic marker of different concentrations in advance, and then using a gas chromatograph to detect the standard samples of each characteristic marker of different concentrations to obtain the standard effective peak area of ​​each characteristic marker of each concentration; Preparing impurity samples of each characteristic marker at a specific concentration in advance, and then using a gas chromatograph to detect the impurity samples of each characteristic marker at the specific concentration to obtain a control effective peak area of ​​each characteristic marker at the specific concentration; Using the formula The relative response factors of each characteristic marker were obtained by analysis. ,in It represents the effective peak area of ​​each characteristic marker at a specific concentration. Indicates Characteristic marker The standard effective peak area corresponding to the standard sample with the same concentration is Indicates The specific concentration of the impurity sample of the characteristic marker, Indicates Characteristic marker The concentration of the standard sample corresponds to The number indicating the concentration, , A quantity that indicates concentration.

9. The combustible material inversion method based on combustion product markers according to claim 1, characterized in that: The specific training method of the artificial intelligence dynamic inversion algorithm model is as follows: Extract a large amount of historically stored combustible combustion data, including the type and content of combustibles, as well as the corresponding characteristic marker types after combustion, and the mass fraction and volume fraction of each characteristic marker; The artificial intelligence dynamic inversion algorithm model is obtained by continuous training of the artificial intelligence algorithm based on the above data.

10. The combustible material inversion method based on combustion product markers according to claim 1, characterized in that: The specific method of inverting the type and content of combustibles is as follows: The mass fraction and volume fraction of each characteristic marker are extracted and input into the artificial intelligence dynamic inversion algorithm model respectively; Obtaining pre-saved combustion records of various combustibles, specifically including the type and content of the combustibles, reference characteristic markers, and reference mass fractions and reference volume fractions of each reference characteristic marker; Compare the reference characteristic marker corresponding to each combustible combustion record with the input characteristic marker. If the type of the reference characteristic marker corresponding to a combustible combustion record is consistent with the type of the input characteristic marker, the combustible combustion record is used as the reference combustible combustion record. Extract the reference mass fraction and reference volume fraction of each reference characteristic marker corresponding to each reference combustible combustion record, and then compare them with the mass fraction and volume fraction of each input characteristic marker to obtain the mass fraction similarity and volume fraction similarity of each characteristic marker; The mass fraction similarity and volume fraction similarity of each characteristic marker are summed up according to the weight to obtain the similarity of each characteristic marker, and then the mean is calculated to obtain the reference evaluation index of each reference combustible combustion record; The reference evaluation indexes of the reference combustible combustion records are compared, the reference combustible combustion record corresponding to the maximum reference evaluation index is selected as the target combustible combustion record, and the combustible type and content corresponding to the target combustible combustion record are output as the combustible type and content.

Images