IDENTIFICATION OF IMPACTING FACTORS OF SURFACE DEFECTS IN HOT ROLLING PROCESSES USING MULTI-LEVEL REGRESSION ANALYSIS

Severe competition in US steel industry urges quality improvements in hot rolling processes. Surface defects have been a long-standing troubling issue in hot rolling processes due to the ineffectiveness of existing detection methods. This paper presents an advanced statistical analysis method to identify the impacting factors in surface defects of hot rolling processes. The surface defects on the steel is measured by a new sensing system, the “HotEye” imaging system. The process variables considered in this paper include the heat number, strand number, and billet-location. Due to the structural characteristic of the data, multilevel analysis is presented to help identify the relationships between the process variables and the number of surface defects. A detailed case study is presented to illustrate the effectiveness of this method. The result obtained can provide guidelines for root cause identification and quality improvement of hot rolling processes.