In order to eliminate the coupling between the loops for control in the system of scrap copper smelting, we
propose the methods to built the dynamic models of inverter-fan-furnace pressure loop and natural gas and combustion
air flow-air fuel ratio-furnace temperature loop based on data-driven, established the thought of multi-variable control
model with the amount of scrap copper, gas flow and speed of fan as input, temperature and pressure of furnace as output,
then use the method of PID neural network to decouple. Simulation results show that the control system be with the
features of fast response, small overshoot and without static error, and also multi-variable PID neural network adjusts the
connection weights based on the effect produced by the changes of object parameters, achieve the decoupling of the
coupling variables effectively; as with reference to the PID control requirements, making the whole system be simple
and standard.
|