点击次数：

所属单位：自动化学院

发表刊物：IFAC PAPERSONLINE

关键字：Fast modeling Quality prediction Model migration Partial Least Squares (PLS) Steel rolling process

摘要：Modem steel rolling process is commonly designed to produce products with a variety of grade specifications. There exist enormous historical data for the typical products in large batch production; while for small-batch customized products, the lack of sufficient historical data may prevent successful application of traditional data-based process modeling and quality prediction methods. In this paper, a practical and effective strategy is developed for fast modeling and flatness defect prediction of a steel rolling process. The key idea is based on model migration, assuming that high-performance quality prediction models (defined as Base Models) have been available for the typical products. A Principal Component Analysis (PCA) similarity indicator is adopted to measure the difference between new operating modes and the typical operating modes, based on which, new operating modes are divided into non-significantly-changed modes and significantly-changed modes. For the former, the new flatness defect prediction model is developed by screening the similar data in the typical modes and augmenting them into the modeling data set for the new mode. For the latter, the new model is obtained by reconstructing the base model via parameters' shifting and scaling. Case study results can show the validity of the proposed method. (C) 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.

ISSN号：2405-8963

是否译文：否

发表时间：2017-01-01

合写作者：陆宁云,Lu, Jianhua,Zhao, Huiping

通讯作者：邢丽冬