Application Research of On-line Quality Control Method to Metallurgical Products

A novel methodology of on-line monitoring, diagnosis and optimization for process parameters is proposed to improve the quality of steel products and reduce the mass quality problems caused by the inability to timely adjust the production process parameters. First, the control limit is calculated by the SVDD method with normal samples. Then, the on-line diagnosing method is proposed through the contribution chart of the process parameters. Finally, the normal samples which are adjacent to abnormal samples are selected as target samples and used to construct the optimization rules. The effectiveness of the proposed method is verified using a dataset from a steel production process. The results indicate that the methodology has a good applicability in complex production processes with nonlinear and strong correlation characteristics.