Multi-step-ahead prediction of cold rolling chatter state based on the combination of Functional Data Analysis and General Autoregression Model

Abstract Mill chatter seriously restricts the improvement of production efficiency and development new products. Thus, a method cold rolling monitoring early warning is proposed based on combination Functional Data Analysis (FDA) General Autoregression Model (GAM). Firstly, multi-source data are preprocessed by FDA to realize smooth fitting unequally sampled sample space constructed mechanism. Then, multi-step-ahead prediction executed through different machine learning algorithms GAM, effects compared evaluated. Finally, maximum step defined select optimal algorithm, conclusion drawn that Extra Tree Regression (ETR) has best effect. Therefore, for can effectively solve problems delay, false alarm distortion in prediction.