Cavity Sequences in Continuously Cast Billets: Part II. Stochastic Models

A stream of voids is often encountered in the core of continuously cast steel billets. This article presents mathematical simulations of the cavity process. Parameters of the models are deduced from the data described in Part I of this article. A Markov simulation represents the chain as a binary sequence; an autoregressive (AR[2]) model and two other related simulations produce series of interval lengths and void fractions. A bridge model uses two independent random numbers to simulate cavity events. The performance of the models is checked against empirical findings. Dynamics of the cavity process are, finally, discussed.