Multivariate statistical model-based system for monitoring the operation of a continuous caster and detecting the onset of impending breakouts

A method for monitoring the operation of a continuous caster operating at a predetermined caster speed, the method including the following steps:

• acquiring off-line measurements of process parameters selecting training data from said off-line measurements of process parameters to represent normal operation of a continuous caster;
• development of a multivariate statistical model corresponding to normal operation of the continuous caster with input from said training data;
• generation of detection thresholds from said multivariate statistical model and from said off-line measurements of process parameters;
• acquiring on-line measurements of process parameters during operation of the continuous caster;
• and
• determining whether said on-line measurements of process parameters are consistent with normal operation of the continuous caster according to said multivariate statistical model, characterized in that
• said multivariate statistical model is used to calculate an output vector of scores with input from said on-line process parameters;
• test statistics are calculated from said vector of scores including calculating at least one univariate test statistic on one of said scores, and a multivariate test statistic on a selected number of said scores;
• and
• said at least one univariate test statistic and multivariate test statistic are compared to said detection thresholds to generate a detection signal, said detection signal being indicative of whether the continuous caster is operating normally and whether corrective action to control the caster speed is required.