Objective Accurate detection of quality indicators and mixed oil concentration at the mixed oil interface during the batch transportation of various media (gasoline, diesel, methanol, etc.) in refined oil pipelines is essential for identifying the mixed oil interface, determining the mixed oil cutting point, and facilitating high-precision, intelligent pipeline operations. Currently, no online detection system in the industry can simultaneously detect both quality indicators and mixed oil concentration for pipeline transportation.

Methods Using a domestic Fourier transform near-infrared spectrometer, an online detection system was designed to directly collect near-infrared spectra of media transported in high-pressure pipelines, such as refined oil and methanol. At the inlet of the Lanzhou Terminal of the Urumqi-Lanzhou Refined Oil Pipeline, samples of 0^# diesel and 92^# gasoline from different refineries and batches were collected. Mixed-oil samples of diesel-gasoline, methanol-diesel, and methanol-gasoline were then prepared, and their near-infrared spectra were acquired. Key quality indicators (flash point of diesel and final boiling point of gasoline) were measured for 0^# diesel, diesel-gasoline mixed oil, 92^# gasoline, and gasoline-diesel mixed oil as reference values. Chemometric multivariate calibration methods were then employed to establish near-infrared spectral calibration databases for the key quality indicators of 0^# diesel, 92^# gasoline, and their mixed oils, as well as for the concentration of gasoline-diesel and methanol-refined oil mixed oils. The online detection system was verified online for gasoline, diesel, and their mixed oils at the inlet of the Lanzhou Terminal of Urumqi-Lanzhou Refined Oil Pipeline.

Results The root-mean-square errors between the near-infrared predicted values and the reference values for the flash point of 0^# diesel, the final boiling point of 92^# gasoline, and the flash point of diesel-gasoline mixed oil were 2.3 °C, 2.4 °C, and 2.8 °C, respectively, meeting the error requirements for the reproducibility of standard methods. The absolute error between the near-infrared predicted value and the reference value for the gasoline volume fraction in gasoline-diesel mixed oil samples did not exceed 0.6％, while the absolute errors for the methanol volume fraction in methanol-diesel and methanol-gasoline mixed oils were also within 0.6％.

Conclusion The online near-infrared spectral detection system effectively addresses the technical challenge of simultaneously detecting key quality indicators and the concentration of mixed oil. This capability is crucial for accurately monitoring the batch transportation of refined oil and methanol, as well as interfaces of their mixed oils, thereby enhancing production process intelligence.