Hits:

Indexed by:期刊论文

Date of Publication:2017-01-01

Journal:Engineering Letters

Included Journals:EI

Volume:25

Issue:3

Page Number:268-276

ISSN No.:1816093X

Abstract:Short-term forecasting in strip thickness of hot rolling is critical to rolling technology, so that dynamic control can be accomplished to increase production and improve product quality. Predicting strip thickness behavior has been always a challenging task due to its complex and non-linear nature. Autoregressive integrated moving average(ARIMA) model has been verified with a better short-term forecasting performance, for the problem of low multi-step prediction accuracy, the rolling strategy is proposed to update model parameters, which develops rolling ARIMA(RARIMA) model. In addition to improve the overall forecasting accuracy of strip thickness, hybrid forecasting of time series data is considered. Hybrid forecasting typically consists of an ARIMA prediction model for the linear component of time series and a nonlinear prediction model for the nonlinear component. In this paper, back propagation neural network(BPNN) is further introduced to forecast the residual of RARIMA model, and rolling ARIMABPNN(RARIMABPNN) continuous forecasting model will be developed, in which rolling forecasting mechanism is used. To the effectiveness of the comprehensive evaluation method, a stability evaluation index is presented, in addition, the proposed method is examined on the two groups of strip thickness data from 620mm strip finishing mill group of hot rolling and the results are compared with some of the basic forecast methods. The results show that the proposed hybrid method could provide a considerable improvement for the forecasting accuracy and stability. © 2017, International Association of Engineers. All rights reserved.