Research On Generalized Predictive Control In Temperature Control Of Resistance Furnace

The heat-treatment resistance furnace is a kind of critical equipment for thermal treatment. Because of huge bulk, high treatment temperature and complex working condition, the character such as MIMO, great coupling, time-varying, the time of large lag and non-determinacy appear in the furnace temperature, this leads to the difficulty to set up the precise mathematical model. So its temperature is very difficult to be controlled precisely and satisfactorily using the traditional control theory. Therefore, how to control temperature precisely becomes the key problem in the heat-treatment process.First, the paper comprehensively presents the situation, development trends about the heat-treatment industry and the application of temperature control of resistance furnace. Second, according to the problems system of MIMO, great coupling, time-varying, the time of large lag and non-determinacy. The system identification method of CARIMA model with RLS is given, and then the temperature with Multivariable Generalized Predictive Control is controlled. On the basis of this, deduce the MGPC algorithm and direct algorithm for its coefficients of diagonal CARIMA model, avoids solving Diophantine equations, gains the purpose of simplification MGPC.According to the request of heat-treatment about the HSR-1600-8 bogie-hearth resistance furnace, the simulation of the furnace system is done with the historical data. The simulation demonstrates this algorithm has good control function, well robust, and energy-saving, it is a new algorithm with high theory and practice value and will be widely applied.