A Material Proportioning Controller based on Rough Neural Network

C.N. Zhang and Y. Wu (Canada)


Controller, Neural Network, Rough Set, Rough Membership Function, Rough Neuron, Material Proportioning


A material-proportioning controller based on rough membership function neural network is presented in this paper. Two types of rough neurons have been used in the design of the rough neural classification system: approximation neurons and decision-based decider neuron. The outputs of a neural network are proportions (%) of raw materials that should be fed into the raw mill and the assessment of the closeness between the output and the target values. Data obtained from simulation are used for neuron implementation and testing. The simulation results show that the compositions are very close to the target values. Network performs good control on the composition throughout the test.

Important Links:

Go Back