Principal component analysis of vibration signal in ladle metallurgy

The process of bottom inert gas stirring in ladle metallurgy is dominated by manual control. However, this type of control has considerable limitations in process optimization. Various researchers have used vibration, sound and image signals to predict the degree of stirring inside the vessel. In literature, there is limited treatment of detecting stirring at low flow rates and interpreting vibration signals. Principal component analysis of triaxial vibration data collected from a cold model shows that low flow rate stirring can be measured using signals from vibration sensors if the frequency ranges considered is carefully chosen.