Method for detecting content of microorganisms in aviation fuel

Abstract

The invention relates to a method for detecting the content of microorganisms in aviation fuel. The detection method comprises the following steps: (1) acquiring near infrared spectrum data of an aviation fuel sample by using a near infrared spectrometer; (2) preprocessing the near infrared spectrum data of the aviation fuel sample, and respectively establishing regression models of the near infrared spectrum data and the microorganism content associated data by utilizing the training set through a partial least square method and a support vector machine method to obtain a microorganism content predicted value of the aviation fuel sample; and checking the accuracy of the regression model through the decision coefficient R2 and the root mean square error RMSE. The detection method provided by the invention can realize nondestructive detection of the ATP content in the aviation kerosene sample, and has the characteristics of high detection efficiency, high accuracy and good stability.

Description

Technical Field

[0001] The invention relates to the technical field of microbial content detection, and in particular to a method for detecting the microbial content in aviation fuel. Background Art

[0002] Since 1895, people have gradually realized that microorganisms can use hydrocarbons as metabolic energy, but the issue of microbial contamination of aviation fuel has not been widely concerned until after 1960. Microbial contamination of aviation fuel can cause many problems, such as deterioration of oil quality, fuel quantity measurement failure, clogging and corrosion of fuel lines and filters, etc., which pose a great threat to the safety of aviation fuel system use and also significantly affect the cleanliness of fuel quality. Therefore, predicting the degree of microbial contamination of aviation fuel and monitoring its development trend are the key to solving the problem of microbial contamination of aviation fuel.

[0003] With the development of microbial detection technology, the detection methods of contaminated microorganisms used in aviation fuel systems have been continuously improved, and have undergone changes from traditional culture methods, analytical chemistry methods, and molecular biology methods. At present, the commonly used methods in the civil aviation system include HY-LITE JET A1, MicobMonitor2, FUELSTAT resinae and other detection technologies. Among them, the MicobMonitor2 detection method requires incubation in a reagent bottle at 25°C for 4 days to obtain the final test results. This method has a long detection time and strict requirements, which is not conducive to the rapid detection of aviation kerosene microorganisms in actual production. The FUELSTAT resinae detection method has problems such as false positive reactions, high overall background of the ELISA plate, and high or low absorbance values. The HY-LiTE JET A1 detection method requires 1mL of capture liquid to be added to the sample and shaken to mix. After standing for 5 minutes, the capture liquid containing microorganisms settles to the bottom of the sample bottle and is re-sucked out with a pipette before testing with the HY-LiTE JET A1 pen, and its ATP content is detected by a photometer. This method needs to be carried out under laboratory conditions, and a certain amount of samples need to be collected and mixed with the capture liquid to reflect the data.

[0004] Near infrared spectroscopy is the fastest growing spectral analysis technology since the 1990s. It has the characteristics of fast analysis rate, simple operation, less sample required, no pollution, and no damage to samples. The near infrared region refers to electromagnetic waves with a wavelength of 780-2526nm. Its spectral information comes from the absorption of light energy by the frequency doubling and combined frequency of hydrogen-containing groups such as CH, NH, OH in molecules. The spectra generated by different groups have different absorption peak positions and intensities. It is currently widely used in industries such as petroleum, chemical, agricultural products, pharmaceuticals, and food quality testing.

[0005] CN106770004A discloses a method for detecting the types and contents of microorganisms in dairy products based on near-infrared spectroscopy technology, comprising the following steps: step S1: performing near-infrared spectroscopy scanning on a dairy product sample to obtain a near-infrared spectrum of the sample; step S2: establishing a model after preprocessing the near-infrared spectrum obtained in step S1, using partial least squares discriminant analysis PLS-DA to establish a qualitative detection model, and using partial least squares PLS to establish a quantitative detection model; step S3: using the model established in step S2 to predict whether an unknown sample is infected with bacteria and the type and content of the bacteria at this time. CN104792686A discloses a method for detecting the number of microorganisms and drug content in a semi-solid drug preparation by Fourier transform near infrared spectroscopy, comprising the following steps: recording the information of the reference value of the number of microorganisms and drug content in the sample; measuring the Fourier transform near infrared transmittance and reflectance spectrum of the sample; reasonably dividing the sample into a calibration set and a prediction set according to the items to be tested; selecting the optimal spectral pretreatment method; screening the optimal modeling spectral range; diagnosing and eliminating abnormal values; using chemometric technology to associate the calibration set sample spectrum with the reference value of its corresponding detection item, and respectively establishing a calibration model for predicting the number of microorganisms and drug content; evaluating the performance of the established calibration model; and using the established calibration model to predict the number of microorganisms and drug content in unknown samples. However, the microorganisms in aviation kerosene are not exactly the same as those in the food and pharmaceutical industries, and the accuracy of microbial detection in the petrochemical industry is lower than that in the food and pharmaceutical industry. Therefore, the microbial detection technology in the food industry cannot be directly applied to the microbial detection of aviation kerosene. Summary of the invention

[0006] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method for detecting the content of microorganisms in aviation fuel, which can realize rapid and non-destructive detection of the ATP content in aviation kerosene samples, and has the characteristics of high detection efficiency, high accuracy, good stability and short detection time.

[0007] In order to achieve the purpose of the invention, the present invention adopts the following technical solutions:

[0008] The present invention provides a method for detecting the content of microorganisms in aviation fuel, the detection method comprising the following steps:

[0009] (1) Collect near infrared spectral data of aviation fuel samples using a near infrared spectrometer;

[0010] (2) The near infrared spectral data of aviation fuel samples were preprocessed, and then the regression models of the near infrared spectral data and the microbial content correlation data were established using the training set by partial least squares method and support vector machine method to obtain the predicted value of the microbial content of aviation fuel samples. 2The accuracy of the regression model was tested using the root mean square error (RMSE).

[0011] The method for detecting the microbial content in aviation fuel provided by the present invention collects near-infrared spectrum data of aviation fuel samples, establishes a regression model in combination with a chemometric method, and predicts the ATP content in unknown samples through the regression model, which further reflects the number of microorganisms. The present invention realizes rapid and non-destructive detection of the ATP content in aviation fuel samples, and can also perform rapid online detection of microorganisms. It has high detection efficiency, high accuracy, good stability, and greatly shortened detection time. At the same time, it does not require sample pretreatment, does not consume reagents, is green and environmentally friendly, has low cost, is easy to operate, and is suitable for large-scale promotion.

[0012] Preferably, before collecting the near-infrared spectrum data of the aviation fuel sample in step (1), the content of microorganisms in the aviation fuel sample is measured using an aviation fuel detection pen as a standard method measurement value.

[0013] Preferably, the determination of the content of microorganisms in the aviation fuel sample using an aviation fuel detection pen is to determine the ATP content in the aviation fuel sample, and the determination results are divided into three ranges, namely: light contamination range: 0-1000RLU / L; moderate contamination range: 1000-5000RLU / L; severe contamination range: greater than 5000RLU / L.

[0014] If the aviation fuel sample volume used is not 1L, the measured RLU value should be divided by the actual sample volume before comparison with the above indicators.

[0015] Preferably, the near-infrared spectral data in step (1) is collected by transmission sampling, and each spectrum is the average result of 1-1000 scans, with a wavelength range of 700-2700nm and a wavelength increment of 0.5-20nm.

[0016] The number of scans is 1-1000 times, for example, 1 time, 100 times, 300 times, 600 times or 1000 times, but is not limited to the listed values, and other unlisted values ​​within the numerical range are also applicable, preferably 300 times.

[0017] The wavelength range is 700-2700 nm, and the wavelength may be, for example, 700 nm, 900 nm, 1500 nm, 2000 nm or 2700 nm, but is not limited to the listed values, and other values ​​not listed within the numerical range are also applicable, preferably 1500-2000 nm.

[0018] The wavelength increment is 0.5-20 nm, and the wavelength may be, for example, 0.5 nm, 1 nm, 2 nm, 10 nm or 20 nm, but is not limited to the listed values, and other values ​​not listed within the numerical range are also applicable, preferably 1-10 nm.

[0019] Preferably, before the near-infrared spectrum data is collected in step (1), the aviation fuel sample is placed in a sample box, the aviation fuel sample is scraped flat with a box cover, and the sample is placed on a bracket together with the box cover. The probe of the near-infrared spectrometer is clamped in the circular hole of the sample box cover and clamped vertically, and then the near-infrared spectrum is collected by a transmission sampling method.

[0020] Preferably, the preprocessing step in step (2) includes: performing first-order differential 9-point smoothing on the near-infrared spectrum data of the aviation fuel sample, and then importing it into the analysis software.

[0021] Preferably, when the regression model is established by partial least squares regression (PLS) in step (2), the peak intensities of the pre-processed near-infrared spectral data are composed into an X matrix, the property indicators measured by the standard method are composed into a Y matrix, and the Y matrix and the X matrix are associated by partial least squares to establish the regression model.

[0022] Preferably, when the regression model is established by the support vector machine method (SVM) in step (2), a radial basis kernel function is selected as the support vector machine kernel function, and a grid optimization algorithm is used to determine the penalty function and the kernel function parameters.

[0023] It should be noted that due to the uncertainty of any detection method for microorganisms in aviation fuel, the measurement of microorganisms is not represented by quantitative values, but rather by multiple sampling tests to show the growth trend of microorganisms to judge the growth of microorganisms. Its value is different from the physical and chemical properties of aviation fuel, and its volatility is much higher than the value of physical and chemical properties. The repeatability error between 5-25% is within the normal range.

[0024] Preferably, the determination coefficient R in step (2) is 2 The calculation formulas for the root mean square error RMSE are as follows:

[0025]

[0026]

[0027] Where: y i,actual is the measured value of the i-th sample measured by the standard method; i,predicted is the predicted value obtained by using the regression model; It is the average value of the measured values ​​of all samples in the calibration set or validation set using the standard method; n is the number of samples in the calibration set or validation set.

[0028] Preferably, the microorganisms in the microbial content of the aviation fuel sample in step (2) include bacteria and / or fungi.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The method for detecting the microbial content in aviation fuel provided by the present invention collects near-infrared spectrum data of aviation fuel samples, establishes a regression model in combination with a chemometric method, and predicts the ATP content in unknown samples through the regression model, which further reflects the number of microorganisms. The present invention realizes rapid and non-destructive detection of the ATP content in aviation fuel samples, and can also perform rapid online detection of microorganisms. It has high detection efficiency, high accuracy, good stability, and greatly shortened detection time. At the same time, it does not require sample pretreatment, does not consume reagents, is green and environmentally friendly, has low cost, is easy to operate, and is suitable for large-scale promotion. DETAILED DESCRIPTION

[0031] The technical solution of the present invention is further described below by specific implementation methods. It should be understood by those skilled in the art that the embodiments are only to help understand the present invention and should not be regarded as specific limitations of the present invention.

[0032] Example 1

[0033] This embodiment provides a method for detecting the content of microorganisms in aviation fuel, and the detection method comprises the following steps:

[0034] (1) Four samples were selected from the upper, middle and lower layers of the oil tank in the aviation fuel storage and transportation system, four samples were selected from the bottom of the recovery tank, and four samples were selected from the filter outlet. A total of 20 samples were selected and numbered 1-20. Within 24 hours after sampling, the HY-LiTE JET A1 aviation fuel detection pen was used to measure the content of microorganisms in each sample, that is, the ATP content, as the standard method determination value. The results are shown in Table 1.

[0035] Table 1

[0036]

[0037] Then put the aviation fuel sample in the sample box, scrape it flat with the box cover, and put it on the bracket together with the box cover. The probe of the near-infrared spectrometer is stuck in the round hole of the sample box cover, clamped vertically, and then the near-infrared spectrum data is collected by transmission sampling. Each spectrum is the average result of 300 scans, with a wavelength range of 1500-2000nm and a wavelength increment of 2nm. Each sample is scanned 5 spectra continuously, and a total of 100 near-infrared spectra are obtained.

[0038] (2) The near infrared spectrum data of aviation fuel samples were divided into 5 spectral intervals, and a first-order differential 9-point smoothing process was performed for each interval. The processed spectral peak intensities were composed into an X matrix, and the ATP content measured by the standard method was composed into a Y matrix. A local prediction model was established for each spectral interval, and the root mean square error (RMSE) was used as the evaluation standard. The spectral intervals participating in the modeling were increased from the interval with the smallest cross-validation root mean square error. The 2-5 spectral intervals with the smallest root mean square error were selected as the preset spectral intervals for constructing the analysis model. Based on the preset spectral intervals, the regression model was established using the partial least squares method. In The Unscramber software, the spectral data of each sample was matched with the data of ATP content one by one. The regression model of ATP was established using the partial least squares method and the method of full cross-validation. When the regression model was established using the support vector machine method, the radial basis kernel function was selected as the support vector machine kernel function, and the penalty function and kernel function parameters were determined using the grid optimization algorithm. The comparison results of the predicted values ​​and the true values ​​of the models obtained by different data processing methods, modeling spectral areas and modeling methods are shown in Table 2.

[0039] Table 2

[0040]

[0041]

[0042] Then the coefficient of determination R 2 The accuracy of the regression model was tested by the root mean square error (RMSE). The test parameters of the models obtained by different data processing methods, modeling spectral regions and modeling methods are shown in Table 3.

[0043] Table 3

[0044]

[0045] It can be concluded from Table 3 that the detection method provided by the present invention can be used to detect the content of microorganisms in aviation fuel, with high detection efficiency and good stability, and can achieve rapid and non-destructive detection of the ATP content in aviation fuel samples.

[0046] Example 2

[0047] This embodiment provides a method for detecting the microbial content in aviation fuel. The difference from Example 1 is that the aviation fuel sample is replaced with dirty oil with obvious turbidity taken from a dirty oil tank. The same pretreatment optimization procedure is used to complete the modeling and prediction of the ATP content. The SVM model parameters under the optimal conditions are shown in Table 4.

[0048] Table 4

[0049] SVM <![CDATA[Coefficient of determination R 2 > 0.9 Root mean square error RMSE 56.75

[0050] It can be concluded from Table 4 that the detection method provided by the present invention can measure the microbial content in aviation fuel with high accuracy, high detection efficiency and good stability, and can realize rapid and non-destructive detection of ATP content in aviation fuel samples.

[0051] Example 3

[0052] This embodiment provides a method for detecting the microbial content in aviation fuel. The difference from Example 1 is that the aviation fuel sample is replaced with dirty oil with obvious turbidity taken from a dirty oil tank. The same pretreatment optimization procedure is used to complete the modeling and prediction of the ATP content. The PLS model parameters under the optimal conditions are shown in Table 5.

[0053] Table 5

[0054] PLS <![CDATA[Coefficient of determination R 2 > 0.89 Root mean square error RMSE 67.88

[0055] It can be concluded from Table 5 that the detection method provided by the present invention can measure the microbial content in aviation fuel with high accuracy, high detection efficiency and good stability, and can realize rapid and non-destructive detection of ATP content in aviation fuel samples.

[0056] In summary, the detection method of microbial content in aviation fuel provided by the present invention collects near-infrared spectrum data of aviation fuel samples, establishes a regression model in combination with chemometric methods, and predicts the content of ATP in unknown samples through the regression model, which further reflects the number of microorganisms. The present invention realizes rapid and non-destructive detection of ATP content in aviation fuel samples, and can also perform rapid detection of online microorganisms. It has high detection efficiency, high accuracy, good stability, and greatly shortened detection time. At the same time, it does not require sample pretreatment, does not consume reagents, is green and environmentally friendly, has low cost, is easy to operate, and is suitable for large-scale promotion.

[0057] The applicant declares that the above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention shall fall within the protection scope and disclosure scope of the present invention.

Claims

1. A method for detecting the content of microorganisms in aviation fuel, characterized in that: The detection method comprises the following steps: (1) Collect near infrared spectral data of aviation fuel samples using a near infrared spectrometer; (2) The near infrared spectral data of aviation fuel samples were preprocessed, and then the regression models of the near infrared spectral data and the microbial content correlation data were established using the training set by partial least squares method and support vector machine method to obtain the predicted value of the microbial content of aviation fuel samples. 2 The accuracy of the regression model was tested using the root mean square error (RMSE).

2. The detection method according to claim 1, characterized in that: Before collecting the near-infrared spectrum data of the aviation fuel sample in step (1), the content of microorganisms in the aviation fuel sample is measured using an aviation fuel detection pen as a standard method measurement value.

3. The detection method according to claim 2, characterized in that: The method of using an aviation fuel test pen to determine the content of microorganisms in an aviation fuel sample is to determine the ATP content in the aviation fuel sample. The measurement results are divided into three ranges, namely: light pollution range: 0-1000RLU / L; moderate pollution range: 1000-5000RLU / L; severe pollution range: greater than 5000RLU / L.

4. The detection method according to any one of claims 1 to 3, characterized in that: The near-infrared spectral data of step (1) are collected by transmission sampling, and each spectrum is the average result of 1-1000 scans, with a wavelength range of 700-2700nm and a wavelength increment of 0.5-20nm.

5. The detection method according to claim 4, characterized in that: The near-infrared spectral data of step (1) are collected by transmission sampling, and each spectrum is the average result of 300 scans, with a wavelength range of 1500-2000nm and a wavelength increment of 1-10nm.

6. The detection method according to any one of claims 1 to 5, characterized in that: The pre-processing step in step (2) includes: performing first-order differential 9-point smoothing on the near-infrared spectrum data of the aviation fuel sample, and then importing it into the analysis software.

7. The detection method according to claim 2, characterized in that: When the regression model is established by partial least squares method in step (2), the peak intensities of the pre-processed near infrared spectral data are composed into an X matrix, the property indicators measured by the standard method are composed into a Y matrix, and the Y matrix and the X matrix are associated by partial least squares method to establish the regression model.

8. The detection method according to any one of claims 1 to 7, characterized in that: When the regression model is established by the support vector machine method in step (2), the radial basis kernel function is selected as the support vector machine kernel function, and the penalty function and the kernel function parameters are determined by the grid optimization algorithm.

9. The detection method according to claim 2, characterized in that: The determination coefficient R in step (2) 2 The calculation formulas for the root mean square error RMSE are as follows: Where: y i,actual is the measured value of the i-th sample measured by the standard method; i,predicted is the predicted value obtained by using the regression model; It is the average value of the measured values ​​of all samples in the calibration set or validation set using the standard method; n is the number of samples in the calibration set or validation set.

10. The detection method according to any one of claims 1 to 9, characterized in that: The microorganisms in the microbial content of the aviation fuel sample in step (2) include bacteria and / or fungi.

Images