TE benchmark

Citation Author(s):
Yuchen
He
Submitted by:
Yuchen He
Last updated:
Fri, 01/27/2023 - 10:00
DOI:
10.21227/30jf-h712
License:
Creative Commons Attribution
82 Views
Categories:
Machine Learning
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

The liquid product H in stream 9 is considered as quality variable while 22 process variables XMEAS (1) – XMEAS (22) are selected as process variables. 960 samples are collected as training samples, and another 960 samples are collected as the test dataset in the same working condition.

Instructions: 

TE benchmark is a chemical simulation platform, which is widely used to evaluate the process monitoring performance of different soft sensing models. The process is mainly composed of five units: reactor, condenser, compressor, separator, and stripper, involving 41 measurement variables and 12 control variables. In TE process, different units organically cooperate with each other. Since various control strategies are implemented, a part of the variables can be autocorrelated while others are not.

Dataset Files

    Files have not been uploaded for this dataset

     

    QUESTIONS?