CSTR fault positioning method based on Xgboost regression model

Abstract

The invention discloses a CSTR fault positioning method based on an Xgboost regression model and discloses a CSTR fault positioning method based on Xgboost regression. The method comprises the following steps: 1) collecting normal data generated by a sensor in a CSTR and unknown offline data; and 2) establishing a monitoring model of the normal data acquired in the step 1), and freely selecting different monitoring models according to the requirements of different occasions; (3) establishing a monitoring model through the step (2), substituting the offline unknown data acquired in the step (1)into the monitoring model, extracting sample statistics to carry out fault detection, and screening out fault data. The method has the advantages that the variable importance of the Xgboost regression model measures the influence of the variables on the output prediction precision, the calculation of the metric value of each variable is mutually independent, and compared with the prior art, the method does not contain other variable acting components, so that the influence of the trailing effect is eliminated;

Description

technical field [0001] The invention relates to the field of CSTRs, in particular to a CSTR fault location method based on an Xgboost regression model. Background technique [0002] The Continuous Stirred Tank Reactor (CSTR) is a very important reaction equipment in chemical production and is widely used. In the production of the three major synthetic materials of chemical fiber, plastic, and synthetic rubber, CSTR accounts for more than 90% of the synthetic production reactors. In addition, it is also widely used in the fields of pharmaceuticals, pesticides, and fuels. In view of the wide application of CSTR in the actual production process, it is of great research value to ensure the stability and safety of its operation. [0003] With the continuous scale and complexity of modern chemical production, when the faults in production cannot be accurately identified and recovered in time, huge losses are often caused. With the continuous generation of a large amount of data ...

AI Technical Summary

Problems solved by technology

At present, the common fault location methods based on multivariate statistical analysis mainly include contribution graph method, reconstruction method and reconstruction contribution method (RBC), but these methods are prone to the influence of smearing effect, which may lead to misdiagnosis in practical applications

A

Images