Noise variance estimate for blast furnace temperature of hot metal based on Autoregressive model in presence of noise

Abstract

Noise variance is an important variable for data filtering. In order to estimate the noise variance of hot iron temperature in process of blast furnace (BF) ironmaking, this work will study parameter estimate of AutoRegressive (AR) process in presence of noise based on BF observed data. Furthermore, a given instrumental variable choosing method and recursive least squares algorithm will be delivered in this paper. The proposed method requires loose assumptions, which are more close to the data fact in blast furnace ironmaking process. Finally, noise variance estimate results are shown by simulation tests.