Adaptive control of convertor fusible steel. (Q556754)

This paper describes an algorithm for adaptive control of a steelmaking process. Ordinarily an operator estimates the state of the metal as the steel is formed using analysis of gas recovery and a variety of other measurements. The authors propose an algorithm for automated adaptive control and stabilization of the steelmaking process. They propose a mathematical model of the process as a system of two nonlinear differential equations \[ \begin{aligned} & \frac{dp}{dt}=Ap+Bw+Eh(p,u)\\ & y=h(p,u)\end{aligned} \] where \(p=[p_{\text{Si}},p_{\text{C}}]^T\) is a state vector describing the concent of silicon and carbon in the metal, \[ A=\left(\begin{matrix} -(k_\theta+k_{\text{Si}}) & 0\\0 & -k_\theta\end{matrix}\right),\quad B=\left(\begin{matrix} 1 & 0 & 1\\ 0 & 1 & 0\end{matrix}\right), \] \(k_\theta\) and \(k_{\text{Si}}\) are nonnegative constants, \(w(t)\) determines the incoming silicon and carbon, \(h(p,u)\) is the speed of decarbonization, \(u\) is the speed of lancing (a controlled parameter), and \(E=[0,-1]^T\). The authors' control algorithm is based on Yakubovich's method of recurrent objective inequalities. They prove that his method stabilizes the solution of the system above. Simulated experiments illustrate the effectiveness of this approach.