A multi-grade chemical process soft sensor modeling method automatically generating samples

Abstract

A multi-grade chemical process soft sensor modeling method automatically generating samples comprises the following steps: (1) dividing the data collected from the multi-grade chemical process into atraining set and a test set as raw data; (2) establishing a generative antagonistic network AGAN based on gradient penalty and Wasserstein distance, and inputting the divided training set into the generative antagonistic network to train the network; (3) using the trained AGAN to generate virtual samples, and forming a new training set together with the original training set; (4) using the new training set as the driving data, training the soft sensor modeling, adjusting the parameters of the soft sensor model to adapt to the new training set, and using the trained soft sensor model to predictthe key quality variables of the multi-brand chemical process. The invention utilizes automatic generation of antagonistic network to generate data to make up the deficiency of data quantity, and improves the prediction accuracy of the soft sensor model based on data driving.

Description

technical field [0001] The invention relates to the field of soft sensor modeling of chemical process, in particular to a method for soft sensor modeling of multi-brand chemical process based on generative confrontation network. Background technique [0002] In the chemical process, the estimation of the key quality variables of the process plays an important role in the continuous and stable operation of the production device, the guarantee of product quality and the full use of the production capacity of the device. However, some important quality variables are difficult to measure directly with online sensors, such as the activity of catalysts, the melt index of polymers, certain quality indicators of petroleum and petrochemical products, and the concentration of bacteria and products in fermentation processes, so soft sensors are required modeling methods to predict it. [0003] In the case of a deep understanding of chemical process technology and known mechanism, mech...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of high data acquisition cost and low prediction accuracy of the existing multi-brand chemical process soft-sensing method, the present invention provides a multi-brand chemical process soft-sensing modeling method that automatically generates samples, using gradient penalty and Wasserstein The Automatically Generated Adversarial Network (AGAN for short) of the distance generates data to make up for the lack of data volume and improve the prediction accuracy based on the data-driven soft sensor model

Images