CN111413324B

CN111413324B - Raman spectrum detection method for trace crude oil in naphtha by using fluorescence background

Info

Publication number: CN111413324B
Application number: CN202010420057.2A
Authority: CN
Inventors: 陈夕松; 王鹤莹; 宋玲政; 梅彬
Original assignee: NANJING RICHISLAND INFORMATION ENGINEERING CO LTD
Current assignee: NANJING RICHISLAND INFORMATION ENGINEERING CO LTD
Priority date: 2020-05-18
Filing date: 2020-05-18
Publication date: 2021-04-06
Anticipated expiration: 2040-05-18
Also published as: CN111413324A

Abstract

The invention discloses a Raman spectrum detection method of trace crude oil in naphtha by using a fluorescence background, which aims at the problem that naphtha is polluted by crude oil leakage in a tube bundle of a heat exchanger at a normal-reduced pressure tower top, fully utilizes the fluorescence background which is often required to be deducted in the Raman spectrum, selects a proper spectrum pretreatment method according to the characteristic that the fluorescence background is changed along with the concentration of the crude oil in the naphtha, retains the fluorescence background, establishes a partial least square model, predicts the trace crude oil leakage in the naphtha, and performs early warning judgment. The method focuses on the relation between the fluorescence background and the concentration of crude oil in naphtha, can rapidly and nondestructively detect the leakage of the crude oil with extremely low concentration, namely trace amount, compared with the traditional method, and has extremely important effects on ensuring the product quality of oil refining enterprises and ensuring the production safety.

Description

Raman spectrum detection method for trace crude oil in naphtha by using fluorescence background

Technical Field

The invention relates to a trace detection method, in particular to a method for predicting trace crude oil leakage by using a crude oil fluorescence background in a Raman spectrum.

Background

In recent years, the amount of high-sulfur high-acid poor crude oil processed in China is increasing, equipment corrosion is serious, and especially the tube bundle of a heat exchanger at the top of a normal pressure reduction tower is the most serious. The internal leakage of the heat exchanger causes crude oil to pollute naphtha, easily causes quality risks, seriously affects the production of the device and brings serious potential safety hazards to the long-period stable operation of the device.

The content of crude oil leaked into naphtha of a heat exchanger is very small, and is generally hundreds or even tens of ppm, so that the crude oil is very difficult to detect. The gas chromatography can analyze the content of hundreds of ppm, and the analysis time is long, which needs about 2 hours. Although the near infrared spectrum detection method is adopted, the detection time is short, the oil product generally has a small absorption coefficient in a near infrared region, the detection limit is usually 0.1%, the oil product can only detect the content of thousands of ppm, and the analysis requirement of trace leakage can not be met. This results in failure to find the problem of heat exchanger internal leakage in time, and brings potential safety hazard to the subsequent reforming and ethylene production using naphtha as raw material. Therefore, a rapid trace detection method is needed, which can find the inner leakage problem in time and ensure the safe, stable and long-period operation of the atmospheric and vacuum distillation unit.

The Raman spectrum is very sensitive to a dark sample, particularly crude oil, and has a strong fluorescence background in the Raman spectrum, while pure naphtha is a colorless sample and has no fluorescence background. The method has the advantages that the strong fluorescence background appears in the Raman spectrum of the naphtha polluted by the crude oil, the fluorescence background is usually used as interference to be deducted in the traditional spectral analysis method, and the trace crude oil in the naphtha is successfully detected by the method by just utilizing the characteristic of the strong fluorescence background of the crude oil in the Raman spectrum.

Disclosure of Invention

The invention aims to provide a Raman spectrum detection method for trace crude oil in naphtha by using a fluorescence background, which has high analysis speed and can accurately detect trace crude oil leakage.

The invention comprises the following steps:

and (3) establishing a trace crude oil Raman spectrum detection model in an off-line manner, and using the model to perform on-line monitoring on the trace crude oil leakage of the heat exchanger at the top of the atmospheric and vacuum tower.

The establishment process of the detection model is as follows:

(1) preparing crude petroleum naphtha mixed samples with different concentrations of 100ppm, 75ppm, 50ppm, 30ppm and 25ppm off line, and preparing pure petroleum naphtha samples without crude petroleum;

(2) measuring the Raman spectrum of each sample, wherein the laser power is 300mw, the integration time is 300ms, the spectrum average frequency is 20 times, and the spectrum range is 3-3400 cm^-1；

(3) Smoothing the original spectral curve of the sample by adopting a moving average smoothing method, removing burr parts in the spectral curve, improving the signal-to-noise ratio of the spectral, wherein the smoothing window is 5;

(4) correlating the spectrum with different crude oil concentration data, and modeling by adopting a partial least square method, wherein the wave number range of the modeling is 50-2500 cm^-1Performing spectrum preprocessing before modeling, wherein the method is vector normalization, and finally evaluating the model by using a leave-one interactive verification method;

after the detection model is established, the model is used for carrying out on-line monitoring on the production process, and the steps are as follows:

(1) collecting the Raman spectrum of naphtha on line, wherein the laser power is 300mw, the integration time is 300ms, the average frequency of the spectrum is 20 times, and the spectrum range is 3-3400 cm^-1；

(2) For the spectrum to be measured 3-3400 cm^-1The range is processed by a moving average smoothing method, the smoothing window is 5, and 50-2500 cm is intercepted^-1Processing the spectrogram by vector normalization after the range;

(3) and (3) predicting the crude oil concentration by using the established partial least square model, and starting early warning if the predicted concentration of half of continuous 10 moments or continuous 5 moments exceeds a set early warning limit.

Has the advantages that:

the invention discloses a Raman spectrum detection method for trace crude oil in naphtha by using a fluorescence background. Compared with the traditional detection method, the method has the advantages of rapidness and no damage, can detect trace crude oil leakage, and has very important significance for finding out the problem of leakage in the atmospheric and vacuum distillation unit in time and ensuring the safe production of the unit.

Drawings

FIG. 1 is a structural diagram of an on-line Raman analysis system of a heat exchanger at the top of a normal and reduced pressure tower

FIG. 2 is a flow chart of the operation of the on-line Raman analysis system of the heat exchanger at the top of the atmospheric and vacuum tower

FIG. 3(a) is a Raman spectrum of a sample in a prediction result chart of an embodiment of the present invention

FIG. 3(b) is a spectrum of FIG. 3(a) after smoothing

FIG. 3(c) is a Raman spectrum of a purified naphtha sample according to an embodiment of the present invention

FIG. 3(d) is a spectrum of different crude oil concentrations of batch A sample according to an embodiment of the present invention

FIG. 3(e) is a model diagram of partial least squares modeling of all spectra of samples from batch A and batch B according to an embodiment of the present invention

FIG. 3(F) is a graph showing the results of the verification of samples of lots C, D and F according to an embodiment of the present invention

Detailed description of the preferred embodiment

The following describes the effect of the method in analyzing the trace amount of crude oil in naphtha by a specific operation flow with reference to the drawings and specific examples. The present embodiment is implemented on the premise of the technical solution of the present invention, but the scope of the present invention is not limited to the following examples.

The invention takes the analysis process in a laboratory as an example, and is also suitable for online detection in the industrial process. The system structure during online detection is shown in fig. 1, and the online monitoring process is shown in fig. 2. The mixed samples of crude oil and naphtha with different concentrations are configured in a laboratory, Raman spectra are scanned, a partial least square model is established according to the Raman spectra, and the feasibility of detecting the leakage of the trace crude oil is proved by verifying the samples.

The method specifically comprises the following steps:

(1) 5 batches of crude oil and naphtha mixed samples were prepared separately:

batch a samples: 0ppm, 25ppm, 30ppm, 50ppm, 75ppm, 100 ppm;

sample batch B: 0ppm, 25ppm, 30ppm, 50ppm, 75ppm, 100 ppm;

sample batch C: 0ppm, 25ppm, 30ppm, 50ppm, 75ppm, 100 ppm;

and D, batch of samples: 0 ppm;

batch F of samples: 0 ppm;

(2) aiming at the samples, scanning by using a Raman spectrometer to obtain 3-3400 cm of each sample^-1Raman spectra in the range, 3 spectra were scanned per sample. A specific raman spectrum is shown in fig. 3 (a);

(3) performing 5-point smoothing treatment on each tensiman spectrum by using a moving average filtering method, wherein the smoothed spectrogram is shown in fig. 3 (b);

(4) when the crude oil concentration of 0ppm in A, B, C, D, F batches is observed, namely the spectrum of the pure naphtha sample is shown in a graph (c) in FIG. 3, it can be seen that the Raman spectrum of the naphtha of each batch has little change, 50-2500 cm^-1The regions do not drift significantly;

(5) the spectrum of the sample of batch A with different crude oil concentrations was observed, as shown in FIG. 3 (d). The method can obviously show that the spectrogram in the range of 50-2500 cm < -1 > gradually inclines upwards along with the increase of the concentration of the crude oil, which is caused by the enhancement of the fluorescence background, and the higher the concentration of the crude oil is, the higher the fluorescence intensity is;

(6) intercepting at 50-2500 cm^-1The spectrum section of the inner spectrum is subjected to vector normalization processing;

(7) performing partial least square modeling by adopting all spectra of samples of batch A and batch B, verifying by adopting leave-one-interaction verification method, and determining coefficient R²At 0.9358, leaving a cross-validation root mean square error RMSECV of 8.33ppm, the model was constructed as shown in FIG. 3(e), and the predicted results for each sample are shown in Table 1;

TABLE 1A, B batch sample prediction results

Note: the unit of content is ppm

(8) The results of the samples from lots C, D and F were used as the verification samples and are shown in FIG. 3(F), and the predicted results of each sample are shown in Table 2, and all the samples without crude oil were correctly identified. Verification of sample R²The RMSECV was 0.9705, 7.95 ppm.

Table 2C, D, F batch validation sample prediction results

Note: the unit of content is ppm

From the above analysis, it was found that the raman spectroscopy can quantitatively detect a trace amount of crude oil in naphtha, and if the detection limit is set to 3 times RMSECV, the raman spectroscopy can detect a crude oil content at a concentration of at least 30 ppm. Therefore, the method can quickly and nondestructively detect the trace crude oil leaked from the naphtha so as to avoid the product quality reduction caused by the leakage of the crude oil into the naphtha product in the industrial production process and even cause the threat to the safe production of the subsequent device.

Claims

1. A method for detecting trace crude oil in naphtha by utilizing a Raman spectrum fluorescence background is characterized by being used for detecting whether trace crude oil internal leakage occurs in an overhead heat exchanger of an atmospheric and vacuum distillation unit and comprising the following steps of:

(1) preparing crude oil naphtha mixed samples with different crude oil concentrations in an off-line manner, and preparing pure naphtha samples which are not doped with crude oil;

(2) measuring the Raman spectrum of each sample, wherein the spectral range is 3-3400 cm^-1；

(3) Smoothing the original spectrum curve of the sample by adopting a moving average filtering method, and removing a burr part in the spectrum curve; then, carrying out spectrogram interception, and intercepting 50-2500 cm^-1An inner spectral band; then carrying out vector normalization processing;

(4) correlating the spectrum with different crude oil concentration data, and modeling by adopting a partial least square method, wherein the wave number range of the modeling is 50-2500 cm^-1(ii) a Evaluating the model by adopting a leave-one-out interactive verification method after modeling;

(5) the Raman spectrum of naphtha is measured on line, and the spectrum range is 3-3400 cm^-1；

(6) For the spectrum to be measured 3-3400 cm^-1Processing the range by a moving average filtering method, and intercepting 50-2500 cm^-1Processing the spectrogram by vector normalization after the range;

(7) using the established partial least square model to predict the original oil concentration in naphtha if m is continuous₁The predicted concentration of half of the time within each time exceeds the set early warning limit or continues for m₂And if the predicted concentration at each moment exceeds the set early warning limit, starting early warning.

2. The method of claim 1, wherein the queue length of the moving average filtering method is selected to be 5.

3. The method for detecting trace crude oil in naphtha using fluorescence background of Raman spectrum as claimed in claim 1, wherein m is₁＝10，m₂＝5。

4. The method as claimed in claim 1, wherein the crude oil concentration in the mixed sample of crude oil and naphtha is 100ppm, 75ppm, 50ppm, 30ppm, 25 ppm.

5. The method as claimed in claim 1, wherein the Raman spectrum is measured by setting the laser power at 300mw, the integration time at 300ms, and the number of spectrum averaging times at 20.