Multirate output feedback control for complex industrial processes in double-layer network environment with RBF performance index

This paper investigates the output feedback control problem for a class of complex industrial processes in double-layer (upper operation layer and local device layer) network environment with radial basis function (RBF) performance index. First, sampled-data local plants at the device layer with sampling period T
d
are controlled by local output feedback proportional integral (PI) controllers. Then, the outputs and inputs of local plants are sampled at operation layer with sampling period T
u
to form the index prediction by using RBF neural networks. In addition, consider the effect of Ethernet between operation and device layer, a Bernoulli random binary distribution is employed to model the packet dropout phenomenon during data transmission. Furthermore, considering the nonlinearity of RBF neural networks, the Lagrange multipliers approach is utilized to solve the real time optimization (RTO) problem at operation layer and an output feedback compensator is designed to regulate the dynamic setpoints. Finally, a rougher flotation process model is employed to demonstrate the effectiveness of the proposed method.