Abstract: The production data is the data with the widest source, longest-span, and largest-scale in modern refineries. It is of great significance for optimizing production, improving quality and efficiency, and building smart refinery to fully handle and make good use of production big data. The general standard process software package SmartPec was specially designed for processing the production data of the smart refinery, and its application in catalytic reforming data processing were introduced. The mean square error of the generalized linear model was reduced by 76.9% and the mean square error of the neural network model was reduced by 88.3% by data pre-processing and categorical modeling. Four key factors affecting the yield of reforming aromatics were identified accurately, and the scheme of maximizing the yield of reforming aromatics was provided. The results show that SmartPec can greatly improve the accuracy of data modeling and prediction of catalytic reforming and various units in refinery plants, with a wide range of applications and strong versatility; it can be used to detect differential features between multiple types of samples, with flexible application mode and strong stability; it can be used for a variety of optimization tasks, easy to use, convenient and fast.

Key words: smart refinry, big data, catalytic reforming, data processing, analysis process, software package