| Adding ethers, alcohols and other additives to gasoline can promote combustion of gasoline, increase octane number of gasoline, but additives may have bad effects on automobile engines, and may also have bad effects on environment. At present, the standard measurement method of additives in gasoline is by gas chromatography. Gas chromatography have some defects, for example, took long time for analysis, high cost for equipment maintenance. Therefore, gas chromatography cannot meet the demand of supervision of quality of gasoline in fast and accurate quality. As a rapid analytical technique, Infrared spectroscopy has many superiority:simple operation simple, rapid analysis, etc. It has been widely used in oil refining industry. In tradition, infrared spectroscopy combined with chemometrics methods can achieve rapid detection in many fields. But the quality and suitability of the model affected by many factors, for example, types of samples, quantity and distribution of samples. Besides, the workload of establishing a good model is heavy. To meet the requirement of rapid analysis of additives in gasoline, a new rapid quantitative method for the determination of oxygenates and the compounds not included in the national standard in gasoline using near-infrared spectroscopy is raised by this paper.At first, we need to prepare various samples with various concentrations of different types of additives. Separate spectrum signal of additives from gasoline solution by oblique projection algorithm, and then establish a standard curve with spectrum signal and their concentrations by one linear regression. According to the standard curve, we can calculate the content of the test samples. Studies show that the extracted spectrum signal by oblique projection algorithm is consistent with the spectrum of pure compounds. And the correlation between the spectral signals’response with their concentrations in gasoline is good. The correlation coefficient of Mid-infrared spectroscopy is above 0.99, the deviation of real concentrations and predicted concentrations of various types of additives is less than 0.47%, the relative deviation is less than 7.43%; The correlation coefficient of Near Infrared Spectroscopy is above 0.95, the deviation is less than 0.66%, the relative deviation is less than 13.68%.In this paper, based on oblique projection algorithm, we presented a new method for determination of methanol and ethanol in gasoline simultaneously. The area of characteristic peak of hydroxyl group have relationship with their molar content in sample. By use of this relationship, we can obtain the content of hydroxyl group is gasoline, that is the total content of methanol and ethanol. Because both methanol and ethanol only have one hydroxyl. By use of the characteristic absorption peak of ethanol, we can got the content of ethanol. Except the content of ethanol from the total content. the content of methanol is obtained. In this method, the deviation of both methanol and ethanol is less than 0.73%, the relative deviation is less than 10.66%. Compared the results with PLS models’, we can find that this method was more accurate and effective to reduce the error due to interaction between methanol and ethanol by their similar nature brings.Finally, this paper also explored a new method to determine the total insoluble of diesel, that is, combine ATR-FTIR with partial least squares calibration method (PLS) to establish a model. the calibration standard deviation (SEC) of the model is 0.544 mg/100ml, and the predicted standard deviation (SEP) of the model is 0.747 mg/100ml. The predicted results meet national standard, and precision and accuracy of the method was basically consistent with the GB standard. |
PDF Full Text Request