On-line soft measuring method of regular first-line oil flash point

Abstract

The invention relates to an on-line soft measuring method of a regular first-line oil flash point. The method comprises selecting and constructing key variables such as a regular first-line distillation temperature, a regular top temperature, regular top pressure, a regular second-line stripping steam flow and a regular first-line extraction-feeding ratio total amount as neural network model inputvariables for influencing the regular first-line oil flash point, building a regular first-line oil flash point soft meter based on an artificial neural network model through the regular first-line oil flash point as a corresponding output, and correcting a model output through a manual analysis value to realize the real-time monitoring and control of the quality of the regular first-line productin the atmospheric tower. The method can guide an operator to timely adjust the operation process. The flash point estimated value provided by the soft meter can provide basis for the process control, optimize the operation process and improve the crude oil extraction rate of the atmospheric and vacuum distillation unit.

Description

technical field [0001] The invention belongs to the intersecting field of petroleum refining and intelligent control, and relates to an online soft measurement method for the flash point of normal-line oil, a product of an atmospheric tower in an atmospheric and vacuum device in the petroleum refining process. Background technique [0002] As the main equipment of the atmospheric and vacuum unit, the atmospheric column processes a large amount of crude oil from the primary distillation column and consumes a large amount of energy. Therefore, ensuring the efficient and stable operation of the atmospheric tower has become the key to saving energy and reducing consumption of atmospheric and vacuum devices, and improving the extraction rate of crude oil. [0003] The flash point of constant first-line oil is the main quality control index of constant first-line products, which mainly reflects the weight of the oil produced, so the quality of its control is not only related to th...

AI Technical Summary

Problems solved by technology

[0006] (1) Due to frequent changes in the nature of crude oil and operating conditions, the calculation accuracy of the built model gradually decreases with time, so the effect of the soft sensor model is limited

[0007] (2) Atmospheric and vacuum devices are highly nonlinear systems with serious variable coupling. There are often many variables selected in soft sensor modeling, and they are interrelated, and there are many redundant variables, which instead cause the output stability of the model to decline.

Images