Online prediction method for product quality in hydrocracking process

Abstract

The present invention provides an online prediction method for product quality in a hydrocracking process. The method comprises: S1 based on sensitive parameters of the product in a time lag window of a process before the product quality offline test time, obtaining a steady- state index of each of the sensitive parameters by using the fitting of the polynomial least squares method; S2 based on the steady-state index of each of the sensitive parameters and the degree of influence of each of the sensitive parameters on the process, constructing the system steady-state exponent by using the Dempster rule, and according to the system steady-state exponent, obtaining steady-state data of the steady-state system in the time lag window of the process before the product quality offline test time; and S3 based on the data mean value of each of the sensitive parameters in the steady-state data in the time lag window of the process, implementing online quality prediction of the product by using the random forest product quality online prediction model. The method provided by the present invention taken full advantage of the variables in the whole hydrocracking process, so that predicted quality can reflect the overall operating condition.

Description

technical field [0001] The invention relates to the field of product quality detection, and more specifically, to an online prediction method for product quality in a hydrocracking process. Background technique [0002] The hydrocracking process is an extremely important secondary processing process for petrochemical enterprises, that is, under the conditions of heating, high hydrogen pressure and catalyst, the heavy oil is cracked and converted into naphtha, gasoline, jet fuel, and diesel And other light, high-quality oil. The hydrocracking process has the advantages of wide processing range of raw materials, strong adaptability of raw materials, high production flexibility, low carbon loss rate, good product quality, and high yield of liquid products. It is one of the flexible and most effective processing methods, and its products are also widely used in automobiles, aviation, ships, heavy machinery, lubricant production, etc. [0003] However, due to changes in feed oi...

AI Technical Summary

Problems solved by technology

However, at present, there are still the following problems in this type of method research: First, the process variables of the whole process of hydrocracking are not fully utilized, and the online prediction models of product quality established basically only consider the process variables of the fractionation system. , which is not conducive to the evaluation of the overall operation status of the hydrocracking process; second, the hydrocracking process has many process variables, and currently it mainly relies on mechanism analysis and plant experience to select auxiliary variables, which is largely subjective; third, the hydrocracking process is a large time-delay The system, even the fractionation system, has a time lag of nearly 1.5 hours. When the external conditions change, the process has a steady state and an unsteady state. The quality online prediction models established in different modes are also different, and need to be distinguished

Images