Directing against features of the coal-pulverizing system equiped with ball mills

such as multi-variable

strong coupling

non-linear

and time-variation etc

a PID adaptive decoupling control system with modified single neuron for the ball mills has been proposed.In this control system

the weights of single neuron are learned on-line by the supervisory Hebb learning algarithm

the proportional gain and learing rate of single neuron adopt on-line adjustment;the structure of single neuron can be changed by integral separated PID method

the results of Matlab/Simulink emulation show that the proposed control method has better quality than that of the traditional PID control method

having advantages of fast tracking ability

strong robustness

and good decouping capability

it can effectively solve the time-varying and coupling problems in the ball mill system.