A fluorescence spectrum-based device and method for rapidly detecting methanol in gasoline

Abstract

The application discloses a device and method for rapidly detecting methanol in gasoline based on fluorescence spectrum. The device comprises a multi-wavelength ultraviolet LED fiber light source system for generating a multi-wavelength light source, a sample chamber, a fluorescence detection system for acquiring the fluorescence spectrum of a to-be-detected oil sample, a fluorescence inner filter correction system for acquiring reference light intensity and transmission light intensity of the to-be-detected oil sample, and a data processing-interaction system. The sample chamber is provided with an opening at the top and three side walls are respectively provided with through holes, the three through holes are respectively a first through hole, a second through hole and a third through hole, the first through hole and the third through hole are oppositely arranged, the opening at the top of the sample chamber is provided with a top cover with a corresponding size to close the opening at the top, and the sample chamber is used for placing the to-be-detected oil sample. The application can realize rapid and accurate detection of the methanol content in gasoline, can weaken the interference of the fluorescence inner filter effect on the fluorescence spectrum detection, can improve the anti-interference ability of the device to scattered light, and has an on-site application advantage.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of gasoline quality detection, and particularly relates to a device and method for rapidly detecting methanol in gasoline based on fluorescence spectrum. BACKGROUND

[0002] With the decrease of oil resources, the increasing number of motor vehicles, and the increasing demand for the environment, countries attach great importance to the improvement of fuel quality. In recent years, in order to improve the octane rating of gasoline and reduce the cost, many unscrupulous businessmen add excessive alcohol additives to gasoline, which seriously threatens the driving safety and environmental safety.

[0003] Since methanol can improve the octane rating of gasoline and is low in price, the phenomenon of artificially adding methanol to gasoline to reduce the cost of gasoline is common in the market. The addition of a small proportion of methanol can improve the economic benefit, but as the amount of methanol increases, the output power of the engine will be insufficient, resulting in insufficient power of the automobile and increased fuel consumption. Methanol has hydrophilicity, and when the temperature is low or the water content increases, the stability of gasoline will decrease, the engine and accessories will be easily corroded, and the service life of the automobile will be shortened. In addition, the harmful exhaust gas emitted by the fuel vehicle containing methanol is higher than that of the ordinary fuel vehicle, which causes great pollution to the environment. Methanol itself is toxic and poses a threat to the health of drivers. Therefore, it is of great significance to detect the content of methanol in the process of transportation, storage, sale and use of finished gasoline to ensure the quality of fuel, the driving safety of vehicles and the safety of the environment.

[0004] At present, the methods commonly used in the petroleum and chemical industry to detect the content of methanol in gasoline include chemical methods and gas chromatography. The chemical method is commonly used for rapid detection on site, and the semi-quantitative detection of the content of methanol is carried out by using a rapid reagent kit for color reaction. This method is not accurate in quantitative results and the waste liquid produced has great harm to the environment. The gas chromatography method can accurately detect the content of methanol, but has the disadvantages of expensive detection equipment, complex operation, long time consumption, unsuitability for on-site detection, easy disturbance by the on-site environment, and lack of a rapid detection device for on-site detection in the methods such as near-infrared spectroscopy and Raman spectroscopy. Therefore, it is necessary to develop a device and a detection method which are fast in detection speed, high in detection accuracy, convenient in operation, small in size and strong in anti-environmental interference. SUMMARY

[0005] The present application provides a device and a method for rapidly detecting methanol in gasoline based on fluorescence spectrum, which are small in size, low in cost, simple in operation and short in detection time, in order to solve the problems of expensive detection equipment, complex operation, long time consumption, unsuitability for on-site detection and easy disturbance by the on-site environment in the existing methods for detecting the content of methanol in gasoline.

[0006] The present application adopts the following technical scheme:

[0007] A fluorescence spectrum-based device for rapid detection of methanol in gasoline, comprising a device shell and a sample chamber, a multi-wavelength ultraviolet LED fiber-optic light source system, a fluorescence detection system, a fluorescence inner filter correction system and a data processing-interaction system arranged in the device shell;

[0008] The sample chamber has an opening at the top and three side walls each having a through hole, each through hole being provided with a fiber-optic collimator, the three through holes being a first through hole, a second through hole and a third through hole, wherein the first through hole and the third through hole are oppositely arranged, the top opening of the sample chamber is provided with a top cover of corresponding size for closing the top opening of the sample chamber, and the sample chamber is used for placing the oil sample to be tested.

[0009] The multi-wavelength ultraviolet LED fiber-optic light source system is installed in the first through hole and can generate multi-wavelength light sources and enter the sample chamber through the first through hole, the multi-wavelength light sources are at least four light sources with different center wavelengths, and the wavelength value range is 250-280 nm with a spacing of not less than 5 nm.

[0010] The fluorescence detection system comprises a fiber-optic spectrometer and a fiber, the fiber-optic spectrometer is connected with the second through hole through the fiber, and is used for acquiring the fluorescence spectrum of the oil sample to be tested.

[0011] The fluorescence inner filter correction system is installed in the third through hole and is used for acquiring the reference light intensity and the transmission light intensity of the oil sample to be tested.

[0012] The data processing-interaction system is used for controlling the experimental process and processing and displaying the experimental data.

[0013] Further, the multi-wavelength ultraviolet LED fiber-optic light source system comprises an ultraviolet LED array and a multi-strand branched fiber, the ultraviolet LED array comprises a PCB and at least four ultraviolet LED diodes with different center wavelengths welded on the PCB, the multi-strand branched fiber comprises at least four input ends and one output end, each input end is connected with the ultraviolet LED diode through a connecting piece, and the one output end is connected with the sample chamber through a first fiber-optic collimator arranged in the first through hole.

[0014] Further, a narrow-band filter corresponding to the center wavelength of the ultraviolet LED diode is arranged in the connecting piece.

[0015] Further, the fluorescence inner filter correction system comprises a collimating mirror, a photodiode and a signal amplification-conversion circuit, the collimating mirror and the photodiode are optically coaxially connected through a fixing piece, and the signal amplification-conversion circuit is connected with the photodiode and the data processing-interaction system, respectively.

[0016] Further, the inner wall of the sample chamber and the lower surface of the cover are coated with a black light-absorbing coating with strong ultraviolet light absorption, and the sample chamber is provided with a cuvette placement groove.

[0017] Further, the data processing-interaction system comprises a microcomputer and a touch screen, and the touch screen is provided with a blank background spectrum and absorbance reference value detection window, a fluorescence spectrum and transmitted light intensity detection window, a spectrum pretreatment window, a primary internal filter effect correction window and a quantitative analysis window.

[0018] A method for constructing a methanol content quantitative model based on fluorescence spectrum, based on any of the devices, comprising the following steps:

[0019] Step 1: Select a reference fuel, gasoline, and configure a series of gasoline samples with different methanol volume fractions;

[0020] Step 2: Use the fluorescence spectrum-based rapid methanol detection device to collect the blank background spectrum and reference light intensity under the excitation wavelength of the ultraviolet LED diode;

[0021] Step 3: Use the fluorescence spectrum-based rapid methanol detection device to collect the fluorescence spectrum and transmitted light intensity of the gasoline samples with different methanol volume fractions under the excitation wavelength of the ultraviolet LED diode;

[0022] Step 4: Calculate the primary internal filter effect correction coefficient of the gasoline samples with different methanol volume fractions under the excitation wavelength;

[0023] Step 5: Perform spectrum pretreatment on the fluorescence spectrum obtained in step 3, and then perform primary internal filter effect correction based on the primary internal filter effect correction coefficient obtained in step 4 to obtain the fluorescence spectrum after spectrum pretreatment and primary internal filter effect correction;

[0024] Step 6: Organize the fluorescence spectrum after spectrum pretreatment and primary internal filter effect correction into a k*n-dimensional spectrum matrix, wherein k is the number of ultraviolet LED diodes of the multi-wavelength ultraviolet LED fiber light source system, and n is the dimension of the spectrum data obtained by the fiber spectrometer;

[0025] Step 7: Combine the k*n-dimensional spectrum matrix with the corresponding methanol volume fraction label to form a training data set, use the training data set to train the methanol content quantitative model, obtain the fluorescence spectrum-based methanol content quantitative model, and implant it into the data processing-interaction system.

[0026] Further, in step 4, the calculation process of the primary internal filter effect correction coefficient is as follows:

[0027]

[0028] Wherein, i is the i-th ultraviolet LED diode, i∈[1, k], A i is the absorbance of the oil sample at the i-th ultraviolet LED diode excitation wavelength, I i is the transmittance of the oil sample at the i-th ultraviolet LED diode excitation wavelength, I 0i is the reference light intensity corresponding to the i-th ultraviolet LED diode excitation wavelength, C i is the primary internal filter correction coefficient of the oil sample at the i-th ultraviolet LED diode excitation wavelength.

[0029] In step 5, the spectral pretreatment includes blank background subtraction, scattering removal, and smoothing denoising.

[0030] The blank background subtraction is:

[0031] S afete_i = S i -S blank_i

[0032] Wherein, S i is the fluorescence spectrum of the oil sample at the i-th ultraviolet LED diode excitation wavelength, S blank_i is the blank background spectrum at the i-th ultraviolet LED diode excitation wavelength, S after_i is the fluorescence spectrum of the oil sample after blank background subtraction at the i-th ultraviolet LED diode excitation wavelength.

[0033] The primary internal filter correction is:

[0034] S corr_i = S pre_i × C i ;

[0035] Wherein, S pre_i is the pretreated fluorescence spectrum of the oil sample at the i-th ultraviolet LED diode excitation wavelength; C i is the primary internal filter correction coefficient of the oil sample at the i-th ultraviolet LED diode excitation wavelength, S corr_i is the fluorescence spectrum of the oil sample after primary internal filter correction at the i-th ultraviolet LED diode excitation wavelength.

[0036] A fluorescence spectrum-based rapid detection method for methanol in gasoline, based on any of the devices, the steps include:

[0037] Step 1: Use the fluorescence spectrum-based rapid detection device for methanol in gasoline to collect the blank background spectrum and reference light intensity at the ultraviolet LED diode excitation wavelength;

[0038] Step 2: Put the oil sample to be tested into the sample chamber, and use the fluorescence spectrum-based rapid detection device for methanol in gasoline to collect the fluorescence spectrum and the transmitted light intensity of the oil sample to be tested at the excitation wavelength of the ultraviolet LED diode;

[0039] Step 3: Calculate the primary internal filter effect correction coefficient of the oil sample to be tested at the excitation wavelength;

[0040] Step 4: Perform spectral pretreatment on the fluorescence spectrum obtained in Step 2, and then perform primary internal filter effect correction to obtain the fluorescence spectrum after spectral pretreatment and primary internal filter effect correction;

[0041] Step 5: Arrange the fluorescence spectrum after spectral pretreatment and primary internal filter effect correction into a k*n-dimensional spectral matrix, wherein k is the number of ultraviolet LED diodes of the multi-wavelength ultraviolet LED fiber light source system, and n is the dimension of the spectral data sampled by the fiber spectrometer;

[0042] Step 6: Input the k*n-dimensional spectral matrix into the fluorescence spectrum-based methanol content quantitative model in the data processing-interaction system to obtain the predicted value of the methanol content in gasoline;

[0043] Further, in Step 3, the calculation process of the primary internal filter effect correction coefficient is as follows:

[0044]

[0045] Wherein, A i is the absorbance of the oil sample to be tested at the excitation wavelength of the i-th ultraviolet LED diode, I i is the transmitted light intensity of the oil sample to be tested at the excitation wavelength of the i-th ultraviolet LED diode, I 0i is the reference light intensity corresponding to the excitation wavelength of the i-th ultraviolet LED diode, C i is the primary internal filter effect correction coefficient of the oil sample to be tested at the excitation wavelength of the i-th ultraviolet LED diode;

[0046] In Step 4, the spectral pretreatment includes blank background subtraction, scattering removal, and smoothing denoising;

[0047] The blank background subtraction is:

[0048] S after_i = S i -S blank_i

[0049] Wherein, S i is the fluorescence spectrum of the oil sample to be tested at the excitation wavelength of the i-th ultraviolet LED diode, S blank_i is the blank background spectrum at the excitation wavelength of the i-th ultraviolet LED diode, and S after_iis the fluorescence spectrum of the i-th ultraviolet LED diode excitation wavelength of the oil sample to be tested after deducting the blank background;

[0050] The primary inner filter effect correction is:

[0051] S corr_i is the fluorescence spectrum of the i-th ultraviolet LED diode excitation wavelength of the oil sample to be tested after deducting the blank background; pre_i × C i ;

[0052] Wherein, S pre_i is the pre-processed fluorescence spectrum of the i-th ultraviolet LED diode excitation wavelength of the oil sample to be tested; C i is the primary inner filter effect correction coefficient corresponding to the i-th ultraviolet LED diode excitation wavelength of the oil sample to be tested, S corr_i is the fluorescence spectrum of the oil sample after the primary inner filter effect correction of the i-th ultraviolet LED diode excitation wavelength.

[0053] The beneficial effects of the present application are:

[0054] (1) The fluorescence inner filter correction system of the device can detect the absorbance at each ultraviolet LED diode excitation wavelength, and the data processing-interaction system can calculate the primary inner filter effect correction coefficient at each excitation wavelength according to the parameter, realize the primary inner filter effect correction of the fluorescence spectrum of the oil sample, enhance the fluorescence signal of methanol, and improve the detection effect of the subsequent methanol content quantitative model based on the fluorescence spectrum, which helps to solve the inner filter effect interference problem of the fluorescence spectrum detection technology in the field of gasoline quality detection, and helps the on-site application of the device.

[0055] (2) The sample chamber is coupled with the fluorescence spectrum detection light path and the inner filter effect correction light path, which can detect the fluorescence spectrum and the transmission light intensity at the same time, detect the blank background spectrum and the reference light intensity at the same time, save the detection time, and at the same time, avoid the noise interference that may be introduced by detecting the fluorescence spectrum and the absorbance at different times; the inner wall of the sample chamber is coated with an ultraviolet light absorbing coating, which can further reduce the noise interference of excitation light scattering on detection.

[0056] (3) The design of the multi-wavelength ultraviolet LED fiber optic light source system can realize fluorescence spectrum detection at multiple excitation wavelengths, and the data processing-interaction system control program can freely select different ultraviolet LED diodes to emit excitation light sources for fluorescence spectrum detection at specified excitation wavelengths; at the same time, the combination of LED array and multiple branched optical fibers in the multi-wavelength ultraviolet LED fiber optic light source system ensures that different excitation light sources share the same incident light path, avoiding the instrument noise that may be introduced by different ultraviolet LED diodes emitting LED light sources due to different light paths.

[0057] (4) The fluorescence spectrum-based rapid detection device for methanol in gasoline is small in size, low in cost, simple in operation and short in detection time; through the cooperation of the anti-interference design inside the device and the correction method in the detection method, the instrument noise, the environmental noise and the fluorescence inner filter interference are reduced, and the quantitative detection precision of methanol is improved; the rapid detection device and method involved in the application can be applied to the on-site rapid detection of gasoline in multiple links such as transportation, storage, sales and use, and can be applied to the detection of gas stations, petroleum and chemical enterprises, market supervision departments and the like, and has important significance for the on-site application of the fluorescence spectrum technology in the petroleum and chemical field. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 is the appearance schematic view of the fluorescence spectrum-based rapid detection device for methanol in gasoline involved in the application;

[0059] Figure 2 is the internal structure section view of the fluorescence spectrum-based rapid detection device for methanol in gasoline involved in the application;

[0060] Figure 3 is the optical path system structure view and the section view of the fluorescence spectrum-based rapid detection device for methanol in gasoline involved in the application; a is the optical path system structure view, and b is the optical path system section view;

[0061] Figure 4 is the flow chart of the fluorescence spectrum-based rapid detection method for methanol in gasoline involved in the application;

[0062] Figure 5 is the fluorescence spectrum of the 92# gasoline sample with the added methanol volume fraction of 0.5% after different data processing operations of the application; a is the fluorescence spectrum waterfall chart after the spectrum pretreatment, b is the fluorescence spectrum waterfall chart after the spectrum pretreatment and the primary inner filter effect correction, c is the fluorescence spectrum detail chart after the spectrum pretreatment, and d is the fluorescence spectrum detail chart after the spectrum pretreatment and the primary inner filter effect correction;

[0063] Figure 6 is the fluorescence spectrum of the 92# gasoline sample with the added methanol volume fraction of 1% after different data processing operations of the application; a is the fluorescence spectrum waterfall chart after the spectrum pretreatment, b is the fluorescence spectrum waterfall chart after the spectrum pretreatment and the primary inner filter effect correction, c is the fluorescence spectrum detail chart after the spectrum pretreatment, and d is the fluorescence spectrum detail chart after the spectrum pretreatment and the primary inner filter effect correction;

[0064] Reference signs: 1 - device housing; 2 - PCB board; 3 - ultraviolet LED diode; 4 - narrow-band filter; 5 - first through-hole; 6 - multi-strand bifurcated optical fiber; 7 - first optical fiber collimator; 8 - sample chamber; 9 - collimating mirror; 10 - photodiode; 11 - second optical fiber collimator; 12 - optical fiber; 13 - optical fiber spectrometer; 14 - second through-hole; 15 - third through-hole; 16 - fluorescence internal filter correction system; 17 - touch screen; 18 - device switch; 19 - top cover. Specific implementation method

[0065] The specific implementation of the present application will be further described below in combination with the drawings.

[0066] Since the existing detection equipment is expensive, complicated to operate, time-consuming, not suitable for on-site detection, and susceptible to interference from the on-site environment, the present application provides a fluorescence spectrum-based rapid detection device for methanol in gasoline, which comprises a device housing 1 and a sample chamber 8, a multi-wavelength ultraviolet LED optical fiber light source system, a fluorescence detection system, a fluorescence internal filter correction system 16, and a data processing-interaction system arranged inside the device housing 1.

[0067] The sample chamber 8 has an opening at the top and three side walls each having a through-hole, and an optical fiber collimator is arranged on each through-hole. The three through-holes are a first through-hole 5, a second through-hole 14, and a third through-hole 15. The first through-hole 5 and the third through-hole 15 are arranged opposite to each other. The opening at the top of the sample chamber 8 is provided with a top cover 19 of corresponding size for closing the sample chamber 8. The sample chamber 8 is used for placing the oil sample to be tested.

[0068] The multi-wavelength ultraviolet LED optical fiber light source system is installed in the first through-hole 5 and can generate multi-wavelength light sources and enter the sample chamber 8 through the first through-hole 5. The multi-wavelength light sources are at least four light sources with different center wavelengths, and the wavelength value range is 250-280 nm with a spacing of not less than 5 nm.

[0069] The fluorescence detection system comprises an optical fiber spectrometer 13 and an optical fiber 12. The optical fiber spectrometer 13 is connected with the second through-hole 14 through the optical fiber 12 and is used for acquiring the fluorescence spectrum of the oil sample to be tested.

[0070] The fluorescence internal filter correction system 16 is installed in the third through-hole 15 and is used for acquiring the reference light intensity and the transmission light intensity of the oil sample to be tested.

[0071] The data processing-interaction system is used for controlling the detection process and processing and displaying the detection data.

[0072] Example 1:

[0073] As Figures 1-3As shown, the present application provides a fluorescence spectrum-based rapid detection device for methanol in gasoline, which comprises a multi-wavelength ultraviolet LED fiber light source system, a sample chamber 8, a fluorescence detection system, a fluorescence inner filter correction system 16, a data processing-interaction system, a power supply system, and a device shell 1.

[0074] Figure 1 The present application relates to the appearance of a fluorescence spectrum-based rapid detection device for methanol in gasoline. The top cover 19 of the device can be opened and closed, and the sample chamber 8 is adapted to detect a standard 1cm*1cm quartz cuvette. When the top cover 19 is opened, the sample chamber 8 can be placed inside for detection. The device realizes detection process control and methanol concentration display through the touch screen 17 in the data processing-interaction system, and controls the start and stop of the entire device through the device switch 18.

[0075] As shown in Figure 2 , Figure 3 , the multi-wavelength ultraviolet LED fiber light source system comprises an ultraviolet LED array, a multi-strand bifurcated optical fiber 6, and an LED light source control circuit. The ultraviolet LED array and the multi-strand bifurcated optical fiber 6 are packaged in the device shell 1. The ultraviolet LED array comprises four ultraviolet LED diodes 3 with different center wavelengths, which are soldered on the PCB board 2. The multi-strand bifurcated optical fiber 6 comprises four input ends and one output end, and the four input ends are connected to the four ultraviolet LED diodes 3 through connecting pieces, and the one output end is connected to a sample chamber 8 through a No. 1 optical fiber collimator 7. The connecting pieces between the four input ends and the four ultraviolet LED diodes 3 are provided with narrow-band optical filters 4 corresponding to the center wavelengths of the ultraviolet LED diodes 3. The LED light source control circuit comprises four LED control modules, each of which has a constant current drive module and a relay connected to form each LED control module. The output end of the constant current drive module of each LED control module is connected to one ultraviolet LED diode 3 on the PCB board 2, and the relay of each LED control module is connected to the data processing-interaction system.

[0076] As shown in Figure 3 , the fluorescence inner filter correction system 16 comprises a collimating mirror 9, a photodiode 10, and a signal amplification-conversion circuit. The signal amplification-conversion circuit comprises a signal amplification circuit and an A / D conversion circuit. The collimating mirror 9 and the photodiode 10 are connected in optical coaxial connection through a fixing piece. The photodiode 10 is connected to the signal amplification circuit and the A / D conversion circuit, and the signal output end of the A / D conversion circuit is connected to the data processing-interaction system.

[0077] As shown in Figure 3As shown in (b), through holes are respectively opened in the centers of the three side walls of the sample chamber 8, which are respectively marked as through hole No. 1 5, through hole No. 2 14, and through hole No. 3 15, wherein through hole No. 1 5 is connected to the multi-wavelength ultraviolet LED fiber optic light source system, through hole No. 2 14 is connected to the fluorescence inner filter correction system 16, and through hole No. 3 15 is connected to the fluorescence detection system; the through hole No. 1 5 and the through hole No. 2 14 are opened on two parallel side walls, and the through hole No. 3 15 is opened on the side wall perpendicular to the planes of the through holes No. 1 and No. 2; the top cover 19 is composed of an openable and closable magnetic top cover; the inner surface of the sample chamber 8 and the lower surface of the top cover 19 are coated with a black light-absorbing coating that strongly absorbs ultraviolet light.

[0078] Through the above technical solution, the sample chamber 8 opens three through holes to connect the three systems respectively, and the fluorescence spectrum detection light path and the excitation light absorbance detection light path are coupled together through the sample chamber 8, which can realize the synchronous detection of the fluorescence spectrum and the excitation light absorbance, save detection time, and reduce environmental interference; at the same time, since the directionality of the light source emitted by the ultraviolet LED diode 3 itself is not good enough, after being collimated by the No. 1 optical fiber collimator 7, it may still be scattered in the narrow space of the sample chamber 8, interfering with the fluorescence spectrum detection. In addition to performing spectral preprocessing in data processing, the ultraviolet light absorbing coating on the inner surface of the sample chamber 8 and the lower surface of the top cover 19 can absorb the scattered excitation light to reduce noise interference.

[0079] like Figure 3 As shown in (b), the fluorescence detection system includes a second optical fiber collimator 11, an optical fiber 12, and a fiber spectrometer 13. The fiber spectrometer 13 is connected to the third through hole 15 through the optical fiber 12 and the optical fiber collimator 11, and is connected to the data processing-interaction system through a data line.

[0080] Depend on Figure 2 、 3 It can be seen that the optical paths of the multi-wavelength ultraviolet LED fiber optic light source system and the fluorescence inner filter correction system 16 are 180 degrees, and the optical paths of the multi-wavelength ultraviolet LED fiber optic light source system and the fluorescence detection system are 90 degrees. They are coupled together through the sample chamber 8, and the fluorescence spectrum and absorbance parameters of the sample can be detected simultaneously, saving time and reducing noise interference.

[0081] The detection wavelength range of the optical fiber spectrometer 13 includes 240nm-400nm.

[0082] In some embodiments, since different small fiber optic spectrometers have different detection wavelength ranges, the detection range of the small fiber optic spectrometer is 180-650 nm.

[0083] The data processing-interaction system includes a microcomputer and a touch screen 17, and in this embodiment, the microcomputer adopts Raspberry Pi. The touch screen 17 is provided with a blank background spectrum and absorbance reference value detection window, a fluorescence spectrum and transmitted light intensity detection window, a spectrum pretreatment window, a primary internal filter effect correction window and a quantitative analysis window.

[0084] A fluorescence spectrum-based rapid methanol detection device for gasoline opens and closes through a device switch 18 on a device housing 1.

[0085] The inside bottom surface of the sample chamber 8 is provided with a cuvette placement groove, which is suitable in size for a standard 1cm*1cm quartz cuvette.

[0086] The central wavelengths of the four ultraviolet LED diodes 3 of the ultraviolet LED array are 250nm, 255nm, 260nm and 275nm respectively, and the central wavelengths of the narrowband filters 4 arranged in the connection between the ultraviolet LED diodes 3 and the input end are 250nm, 255nm, 260nm and 275nm respectively.

[0087] By adopting the above technical solution, the combination of the ultraviolet LED array, the narrowband filter 4 and the multi-strand bifurcated optical fiber 6 simplifies the existing wide-band light source (such as a xenon lamp) and a light splitting system; the ultraviolet LED diodes 3 improve monochromaticity through the narrowband filter 4; the combination of the LED array and the multi-strand bifurcated optical fiber 6 enables the LED array multi-wavelength excitation light to share one excitation light transmission channel, saves device space and reduces fluorescence spectrum detection noise caused by different light paths of multiple ultraviolet LED diodes 3; through the data processing-interaction system control of the switch of each ultraviolet LED diode 3 in the ultraviolet LED array light source, single-wavelength excitation light in a single time period can be input into one input end of the multi-strand bifurcated optical fiber 6 after passing through the narrowband filter 4, and then output through the output end of the multi-strand bifurcated optical fiber 6 and enter the sample chamber 8 through the No.1 optical fiber collimator 7. Sequential control of the switch of the ultraviolet LED diode 3 can realize the multi-wavelength fluorescence spectrum and absorbance detection process. Petroleum products such as gasoline are composed of various alkane compounds, have strong fluorescence reaction and strong absorbance characteristics, and the emission spectrum of methanol overlaps with the absorption spectrum of gasoline. Therefore, there is a relatively obvious fluorescence internal filter effect in the gasoline system, which can cause the fluorescence spectrum intensity to weaken and the shape to distort, greatly interfering with the quantitative detection of methanol in gasoline. Without diluting the oil sample, the absorbance at the excitation wavelength point is detected, and the primary internal filter effect correction of the oil sample fluorescence spectrum is performed using the absorbance, which can restore the fluorescence intensity of each spectrum to a certain extent and improve the linear relationship. By adopting the above technical solution, i.e., adding a fluorescence detection system and a fluorescence internal filter correction system 16, the absorbance of excitation light is detected simultaneously when the fluorescence spectrum at a certain excitation wavelength is detected, and the primary internal filter effect correction is realized.

[0088] Example 2:

[0089] The present invention is based on the device provided in Example 1 to carry out a method for constructing a quantitative model for methanol content based on fluorescence spectroscopy and a method for rapid detection of methanol in gasoline based on fluorescence spectroscopy, comprising the following steps:

[0090] Step 1: Select a benchmark fuel gasoline and prepare a series of gasoline samples with different methanol volume fractions.

[0091] The specific process for preparing gasoline samples with different methanol volume fractions is as follows: A series of gasoline samples with methanol volume fractions ranging from 0.5% to 90% are prepared using a base gasoline fuel. The methanol volume fraction in the mixed samples varies: within the range of 0% to 20%, the methanol volume fraction gradient is 0.5%; within the range of 21% to 30%, the methanol volume fraction gradient is 1%; within the range of 35% to 60%, the methanol volume fraction gradient is 5%; and within the range of 70% to 90%, the methanol volume fraction gradient is 10%.

[0092] Calculate the required volume of anhydrous methanol based on the methanol volume fraction of the oil sample to be prepared and the preparation capacity. Use a pipette to transfer a certain volume of methanol into a container. Then, adjust the volume with gasoline and shake to mix thoroughly. Inject the resulting sample into a cuvette and move the cuvette into the sample chamber 8. During the preparation process, minimize the exposure time of the anhydrous methanol reagent to air and avoid the incorporation of water into the oil sample.

[0093] Step 2: Use the rapid detection device for methanol in gasoline based on fluorescence spectroscopy to collect blank background spectra and reference light intensity at four excitation wavelength points, recorded as S blank_i , I 0i ; where i is the i-th UV LED diode 3, i∈[1,4].

[0094] The specific process is as follows: do not put the cuvette in and ensure that the top cover 19 is closed; start the blank background spectrum and absorbance reference value detection process through the interactive system; the Raspberry Pi controls the control module relay of the No. 1 ultraviolet LED diode 3 in the multi-wavelength ultraviolet LED fiber optic light source system to close, and the photodiode 10 converts the excitation light signal after the collimation mirror 9 into an electrical signal. The Raspberry Pi controls the sampling to obtain the reference light intensity of the No. 1 ultraviolet LED diode 3 at the corresponding excitation wavelength, which is recorded as I 01 At the same time, the fiber optic spectrometer 13 samples the blank background spectrum under the control of the Raspberry Pi, which is recorded as S blank_1 .

[0095] The control module relays of the UV LED diodes 2-4 are controlled to be closed in sequence, and the other relays are disconnected, so as to obtain the reference light intensity and blank background spectrum at the corresponding excitation wavelength of the UV LED diodes 2-4.

[0096] Step 3: Collect the fluorescence spectra and transmission intensity of the gasoline samples with different methanol volume fractions at four excitation wavelengths, respectively denoted as S i , I i .

[0097] The specific detection process is as follows: the cuvette filled with the oil sample is put in, and the fluorescence spectrum and transmission intensity detection process are started through the interactive system; the Raspberry Pi controls the No. 1 ultraviolet LED diode 3 to emit excitation light, and other light sources are turned off; the Raspberry Pi controls the fiber spectrometer 13 to sample the fluorescence spectrum S1 at the excitation wavelength of 250 nm, and at the same time, the fluorescence internal filter correction system 16 samples the transmission intensity I1 at the excitation wavelength of 250 nm; the fluorescence spectrum and transmission intensity at the excitation wavelength corresponding to the No. 2-4 ultraviolet LED diode 3 are obtained in turn.

[0098] Step 4: Calculate the primary internal filter correction coefficient of the gasoline sample with different methanol volume fractions at four excitation wavelengths, denoted as C i ; wherein i is the i-th ultraviolet LED diode 3. The calculation process is as follows:

[0099]

[0100] Wherein, A i is the absorbance of the oil sample at the excitation wavelength of the i-th ultraviolet LED diode 3, I i is the transmission intensity of the oil sample at the excitation wavelength of the i-th ultraviolet LED diode 3, I 0i is the reference light intensity corresponding to the excitation wavelength of the i-th ultraviolet LED diode 3, C i is the primary internal filter correction coefficient of the oil sample at the excitation wavelength of the i-th ultraviolet LED diode 3.

[0101] Step 5: The fluorescence spectra of the gasoline samples with different methanol volume fractions obtained in step 3 are preprocessed, including blank background subtraction, scattering removal, and smoothing denoising, to obtain the preprocessed fluorescence spectra; and the primary internal filter effect of each oil sample fluorescence spectrum is corrected using the primary internal filter correction coefficient of each oil sample calculated in step 4. The primary internal filter correction method is as follows:

[0102] S corr_i = S pre_i × C i

[0103] Wherein, S pre_i is the preprocessed fluorescence spectrum of the oil sample at the excitation wavelength of the i-th ultraviolet LED diode 3; C i is the primary internal filter correction coefficient of the oil sample at the excitation wavelength of the i-th ultraviolet LED diode 3, S corr_iis the fluorescence spectrum of the oil sample after the primary inner filter effect correction of the oil sample at the i-th excitation wavelength of the ultraviolet LED diode 3.

[0104] The specific process is: the fluorescence spectra S1, S2, S3, S4 of each oil sample at 4 excitation wavelengths are deducted from the blank background spectrum signal; then, Rayleigh scattering removal and Savitzky-Golay filtering and smoothing are performed on the fluorescence spectrum to obtain S pre_i ; the 4 fluorescence spectra of each oil sample after spectral pretreatment are multiplied by the primary inner filter effect correction coefficient at the corresponding excitation wavelength.

[0105] Step 6: The fluorescence spectrum data of each gasoline oil sample after pretreatment and primary inner filter effect correction in step 5 is arranged into a k*n-dimensional spectral matrix, where k is the number of ultraviolet LED diodes 3 of the multi-wavelength ultraviolet LED fiber light source system, k is 4 here, and n is the dimension of the spectral data sampled by the fiber spectrometer.

[0106] The specific process of spectral data arrangement: the fluorescence spectrum data of each gasoline oil sample obtained in step 6 is arranged into a matrix form, that is, the fluorescence spectra at 4 excitation wavelengths are combined in the row direction, and arranged into a 4*n-dimensional spectral matrix, where n is the dimension of the spectral data sampled by the fiber spectrometer, that is, the number of sampling emission wavelength points. For a single sampling, the fluorescence spectrum dimension obtained by the fiber spectrometer is 1*n.

[0107] Step 7: The 4*n-dimensional spectral matrix in step 6 and the methanol content label are correspondingly composed into a training data set, and after training the methanol content quantitative model using the training data set, a fluorescence spectrum-based methanol content quantitative model is obtained, which is transplanted on the Raspberry Pi.

[0108] Step 8: For the oil sample to be tested with unknown methanol content, according to steps 2-6, the fluorescence spectrum and absorbance at 4 excitation wavelengths are obtained, and spectral pretreatment, primary inner filter effect correction and data arrangement are performed in the Raspberry Pi. The fluorescence spectrum of the oil sample to be tested is analyzed by using the quantitative model established in step 7 to predict the methanol content of the oil sample to be tested, which is displayed on the touch screen 17.

[0109] The blank background deduction process in the spectral pretreatment process in step 5 is as follows:

[0110] S after_i = S i -S blank_i

[0111] Wherein, S i is the fluorescence spectrum of the oil sample at the i-th excitation wavelength of the ultraviolet LED diode 3, S blank_i is the blank background spectrum at the i-th excitation wavelength of the ultraviolet LED diode 3, and Safter_i is the fluorescence spectrum of the oil sample after deducting the blank background at the i-th ultraviolet LED diode 3 excitation wavelength.

[0112] The fluorescence spectrum based gasoline methanol rapid detection device and method of the present application can detect the fluorescence spectrum at multiple excitation wavelengths and the excitation light absorbance, realize the primary internal filter effect correction of the fluorescence spectrum of the oil sample, enhance the fluorescence signal of methanol, make the effect of the subsequent quantitative detection model better, and help the on-site application of the device.

[0113] The primary internal filter effect correction and on-site application advantages of the fluorescence spectrum based gasoline methanol rapid detection device and method of the present application are illustrated by Examples 3 and 4:

[0114] Example 3: 92# gasoline is used as the reference fuel, a mixed oil sample with a methanol volume fraction of 0.5% is configured, and the fluorescence spectrum and absorbance of the oil sample at four excitation wavelengths are collected according to steps 2-6 in Example 2, the primary internal filter effect correction coefficient is calculated, and the spectrum is pretreated and corrected, and the correction coefficient calculation results are shown in Table 1.

[0115] Table 1: Absorbance at different excitation wavelengths and primary internal filter effect correction coefficient

[0116]

[0117] Example 4: 92# gasoline is used as the reference fuel, a mixed oil sample with a methanol volume fraction of 1% is configured, and the fluorescence spectrum and absorbance of the oil sample at four excitation wavelengths are collected according to steps 2-6 in Example 2, the primary internal filter effect correction coefficient is calculated, and the spectrum is pretreated and corrected, and the correction coefficient calculation results are shown in Table 2.

[0118] Table 2: Absorbance at different excitation wavelengths and primary internal filter effect correction coefficient

[0119]

[0120] By Figure 5 comparing (a), (b) and Figure 6 comparing (a), (b), it can be found that the present application can correct the primary internal filter effect of the fluorescence spectrum of the gasoline oil sample, and restore the intensity of the fluorescence signal; by Figure 5 comparing (c), (d) and Figure 6 comparing (c), (d), the primary internal filter effect correction effect of the fluorescence spectrum of the present application is more prominent in the characteristic peak range of the emission spectrum at different excitation wavelengths, and for quantitative analysis of mixtures, the fluorescence characteristic peak signal of the substance has a greater influence on the prediction result; therefore, the present application can correct the internal filter effect in the on-site detection of methanol in gasoline to a certain extent, and has on-site application advantages.

[0121] The input data of the quantitative model described in the method for rapidly detecting methanol in gasoline based on fluorescence spectrum is the fluorescence spectrum of the oil sample after spectrum pretreatment and primary internal filter effect correction, and the output prediction result is the methanol content in the oil sample. The specific implementation method thereof can be the combined application of various spectrum feature extraction methods and correction methods, such as the combination of the principal component analysis (PCA) in the feature extraction method, the uninformative variable elimination method (UVE), and the partial least squares regression (PLS) in the correction method, and the support vector regression (SVR).

[0122] The following illustrates the specific implementation mode of the quantitative model training:

[0123] Example 5: After obtaining the training data set according to steps 1-7 using No. 95 gasoline as the reference fuel; the variable projection importance method is used for variable selection of the wavelength points of the fluorescence spectrum, and the wavelength points with a VIP value greater than 1 are reserved as the characteristic variables to screen out variables with low importance for methanol content regression; according to the variable selection result, the characteristic variables of each sample are extracted, and the spectrum characteristic variable data are arranged in a vector format; and the partial least squares regression method (PLS) is used to train the methanol content regression model.

[0124] In this embodiment, the trained variable selection model and the regression model jointly constitute the quantitative model of the methanol content based on the fluorescence spectrum. The quantitative model is arranged on the Raspberry Pi of the device. For the oil sample to be measured, after the spectrum data and the absorbance parameters are detected using the device according to step 8, the data processing (spectrum pretreatment, primary internal filter effect correction) and analysis (quantitative model prediction) are realized on the Raspberry Pi, and finally the methanol content prediction result is output on the touch screen 17.

[0125] The above-described embodiments have described the technical solutions and beneficial effects of the present application in detail, and it should be understood that the above-described embodiments are only specific embodiments of the present application and are not used to limit the present application. Any modification, supplement, and equivalent replacement, etc. made within the principle range of the present application shall be included in the protection range of the present application.

Claims

1. A rapid detection device for methanol in gasoline based on fluorescence spectroscopy, characterized in that: The device comprises a housing (1) and a sample chamber (8) arranged therein, a multi-wavelength ultraviolet LED fiber optic light source system, a fluorescence detection system, a fluorescence internal filter correction system (16), and a data processing-interaction system; The sample chamber (8) is open at the top and has through holes on its three side walls, each of which is provided with a fiber collimator. The three through holes are respectively a No. 1 through hole (5), a No. 2 through hole (14), and a No. 3 through hole (15), wherein the No. 1 through hole (5) and the No. 3 through hole (15) are arranged opposite to each other. The top opening of the sample chamber (8) is provided with a top cover (19) of corresponding size for closing the top opening of the sample chamber (8), and the sample chamber (8) is used for placing an oil sample to be tested; The multi-wavelength ultraviolet LED fiber optic light source system is installed in the first through hole (5), and is capable of generating a multi-wavelength light source and entering the sample chamber (8) through the first through hole (5), wherein the multi-wavelength light source is a light source with at least four different central wavelengths, the wavelength range is 250-280nm, and the spacing is not less than 5nm; The fluorescence detection system includes a fiber optic spectrometer (13) and an optical fiber (12), wherein the fiber optic spectrometer (13) is connected to the second through hole (14) via the optical fiber (12) and is used to obtain the fluorescence spectrum of the oil sample to be tested; The fluorescence inner filter correction system (16) is installed in the third through hole (15) and is used to obtain the reference light intensity and the transmitted light intensity of the oil sample to be tested; The data processing-interaction system is used for controlling the detection process and processing and displaying the detection data.

2. The device for rapid detection of methanol in gasoline based on fluorescence spectroscopy according to claim 1, characterized in that: The multi-wavelength ultraviolet LED fiber optic light source system comprises an ultraviolet LED array and multiple bifurcated optical fibers (6), wherein the ultraviolet LED array comprises a PCB board (2) and at least four ultraviolet LED diodes (3) with different central wavelengths welded on the PCB board (2); the multiple bifurcated optical fibers (6) comprise at least four input ends and one output end, each input end is connected to one of the ultraviolet LED diodes (3) via a connector, and the output end is connected to a sample chamber (8) via a first optical fiber collimator (7) provided in the first through hole (5).

3. The device for rapid detection of methanol in gasoline based on fluorescence spectroscopy according to claim 2, characterized in that: A narrowband filter (4) corresponding to the central wavelength of the ultraviolet LED diode (3) connected thereto is arranged in the connecting piece.

4. The device for rapid detection of methanol in gasoline based on fluorescence spectroscopy according to claim 1, characterized in that: The fluorescence inner filter correction system (16) comprises a collimator (9), a photodiode (10) and a signal amplification-conversion circuit, wherein the collimator (9) and the photodiode (10) are optically coaxially connected via a fixing member, and the signal amplification-conversion circuit is respectively connected to the photodiode (10) and the data processing-interaction system.

5. The device for rapid detection of methanol in gasoline based on fluorescence spectroscopy according to claim 1, characterized in that: The inner wall of the sample chamber (8) and the lower surface of the top cover (19) are coated with a black light-absorbing coating that strongly absorbs ultraviolet light, and a cuvette placement groove is provided in the sample chamber (8).

6. The device for rapid detection of methanol in gasoline based on fluorescence spectroscopy according to claim 1, characterized in that: The data processing-interaction system comprises: a microcomputer and a touch screen (17), wherein the touch screen (17) is provided with a blank background spectrum and absorbance reference value detection window, a fluorescence spectrum and transmitted light intensity detection window, a spectrum preprocessing window, a primary inner filter effect correction window and a quantitative analysis window.

7. A method for constructing a quantitative model of methanol content based on fluorescence spectroscopy, characterized in that: The device according to any one of claims 1 to 6 is obtained by the following steps: Step 1: Select the benchmark fuel gasoline and prepare a series of gasoline samples with different methanol volume fractions; Step 2: using the rapid detection device for methanol in gasoline based on fluorescence spectroscopy to collect a blank background spectrum and a reference light intensity at the excitation wavelength of the ultraviolet LED diode (3); Step 3: using a rapid methanol detection device in gasoline based on fluorescence spectroscopy, respectively collecting the fluorescence spectra and transmitted light intensities of the gasoline samples with different methanol volume fractions at the excitation wavelength of the ultraviolet LED diode (3); Step 4: Calculate the primary inner filter effect correction coefficient of gasoline samples with different methanol volume fractions at the excitation wavelength; Step 5: performing spectral preprocessing on the fluorescence spectrum obtained in step 3, and then performing primary inner filter effect correction based on the primary inner filter effect correction coefficient obtained in step 4, to obtain a fluorescence spectrum after spectral preprocessing and primary inner filter effect correction; Step 6: Arranging the fluorescence spectrum after spectral preprocessing and correction of the primary inner filter effect into a k*n dimensional spectrum matrix, wherein k is the number of ultraviolet LED diodes (3) of the multi-wavelength ultraviolet LED fiber optic light source system, and n is the dimension of the spectrum data sampled by the fiber optic spectrometer (13); Step 7: The k*n dimensional spectral matrix and the corresponding methanol volume fraction label are combined into a training data set. After using the training data set to train the methanol content quantitative model, a methanol content quantitative model based on fluorescence spectroscopy is obtained and implanted into the data processing-interaction system.

8. The method for constructing a methanol content quantitative model according to claim 7, wherein: In step 4, the calculation process of the primary inner filter effect correction coefficient is as follows: Where i is the i-th UV LED diode, i∈[1,k], A i is the absorbance of the oil sample at the excitation wavelength of the i-th UV LED diode, I i is the transmitted light intensity of the oil sample at the excitation wavelength of the i-th UV LED diode, I 0i is the reference light intensity corresponding to the excitation wavelength of the i-th UV LED diode, C i is the correction coefficient of the primary inner filter effect at the excitation wavelength of the i-th UV LED diode of the oil sample; In step 5, the spectrum preprocessing includes blank background subtraction, scattering removal, and smoothing and denoising; The blank background subtraction is: S after_i =S i -S blank_i Among them, S i is the fluorescence spectrum of the oil sample at the excitation wavelength of the i-th UV LED diode, S blank_i is the blank background spectrum at the excitation wavelength of the i-th UV LED diode, S after_i is the fluorescence spectrum of the oil sample at the excitation wavelength of the i-th UV LED diode after deducting the blank background; The primary inner filter effect correction is: S corr_i =S pre_i ×C i ; Among them, S pre_i is the pre-processed fluorescence spectrum of the oil sample at the excitation wavelength of the i-th UV LED diode; C i is the correction coefficient of the primary inner filter effect corresponding to the excitation wavelength of the i-th UV LED diode of the oil sample, S corr_i is the fluorescence spectrum of the oil sample after correction of the primary inner filter effect at the excitation wavelength of the i-th UV LED diode of the oil sample.

9. A method for rapid detection of methanol in gasoline based on fluorescence spectroscopy, characterized in that: Based on the device according to any one of claims 1 to 6, the steps include: Step 1: using the rapid detection device for methanol in gasoline based on fluorescence spectroscopy to collect a blank background spectrum and a reference light intensity at the excitation wavelength of the ultraviolet LED diode (3); Step 2: placing the oil sample to be tested into the sample chamber (8), and using a rapid methanol detection device in gasoline based on fluorescence spectroscopy to collect the fluorescence spectrum and transmitted light intensity of the oil sample to be tested at the excitation wavelength of the ultraviolet LED diode (3); Step 3: Calculate the primary inner filter effect correction coefficient of the oil sample to be tested at the excitation wavelength; Step 4: performing spectral preprocessing on the fluorescence spectrum obtained in step 2, and then performing primary inner filter effect correction to obtain a fluorescence spectrum after spectral preprocessing and primary inner filter effect correction; Step 5: Arrange the fluorescence spectrum after spectral preprocessing and primary inner filter effect correction into a k*n dimensional spectral matrix, where k is the number of UV LED diodes in the multi-wavelength UV LED fiber optic light source system, and n is the dimension of the spectral data sampled by the fiber optic spectrometer; Step 6: Input the k*n dimensional spectral matrix into the methanol content quantitative model based on fluorescence spectrum in the data processing-interaction system to obtain the predicted value of methanol content in gasoline.

10. The detection method according to claim 9, characterized in that: In step 3, the calculation process of the primary inner filter effect correction coefficient is as follows: Among them, A i is the absorbance of the oil sample to be tested at the excitation wavelength of the i-th UV LED diode, I i is the transmitted light intensity of the oil sample under the excitation wavelength of the i-th UV LED diode, I 0i is the reference light intensity corresponding to the excitation wavelength of the i-th UV LED diode, C i is the correction coefficient of the primary inner filter effect at the excitation wavelength of the i-th UV LED diode of the oil sample to be tested; In step 4, the spectrum preprocessing includes blank background subtraction, scattering removal, and smoothing and denoising; The blank background subtraction is: S after_i =S i -S blank_i Among them, S i is the fluorescence spectrum of the oil sample to be tested under the excitation wavelength of the i-th UV LED diode, S blank_i is the blank background spectrum at the excitation wavelength of the i-th UV LED diode, S after_i is the fluorescence spectrum of the oil sample to be tested at the excitation wavelength of the i-th UV LED diode after deducting the blank background; The primary inner filter effect correction is: S corr_i =S pre_i ×C i ; Among them, S pre _ i is the pre-processed fluorescence spectrum of the oil sample at the excitation wavelength of the i-th UV LED diode; C i is the correction coefficient of the primary inner filter effect corresponding to the excitation wavelength of the i-th UV LED diode of the oil sample to be tested, S corr_i is the fluorescence spectrum of the oil sample after correction of the primary inner filter effect at the excitation wavelength of the i-th UV LED diode of the oil sample to be tested.

Images