Multivariable Control of a Travelling-Load Furnace

Application of an adaptive multivariable technique to the on-line digital computer control of a pilot-scale electrically-heated travelling-load furnace is considered. The method employs a linear model which is recursively identified and employed to compute control variables which minimise predicted errors occuring at the next sampling instance. The perfonnance of the multivariable scheme is assessed and compared with that obtained using conventional single-loop control employing PID control algorithms.