Artificial Intelligence-Based Descriptive, Predictive, and Prescriptive Coating Weight Control Model for Continuous Galvanizing Line

Zinc wiping is a phenomenon used to control zinc-coating thickness on steel substrate during hot dip gal-vanizing by equipment called air knife. Uniformity of zinc coating weight in length and width profile alongwith surface quality are most critical quality parameters of galvanized steel. Deviation from tolerance levelof coating thickness causes issues like overcoating (excess consumption of costly zinc) or undercoatingleading to rejections due to non-compliance of customer requirement. Main contributor of deviation fromtarget coating weight is dynamic change in air knives equipment setup when thickness, width, and type ofsubstrate changes. Additionally, cold coating measurement gauge measure coating weight after solidifica-tion but are installed down the line from air knife resulting in delayed feedback. This study presents a coat-ing weight control model (Galvantage) predicting critical air knife parameters air pressure, knife distancefrom strip and line speed for coating control. A reverse engineering approach is adopted to design a pre-dictive, prescriptive, and descriptive model recommending air knife setups that estimate air knife distanceand expected coating weight in real time. Implementation of this model eliminates feedback lag experi-enced due to location of coating gauge and achieving setup without trial-error by operator. KCI Citation Count: 0