Online controller for cold circulating fluidized bed

A new approach for modeling and control of cold flow circulating fluidized bed (CFB) using diagonal recurrent neural network (DRNN) has been proposed. The drawbacks of the usual feed forward network (FNN) are that it is a static mapping and requires a large number of neurons and takes a long training time. The usual fixed learning rate, which is based on empirical trail, and error scheme is slow and does not guarantee convergence. The architecture of DRNN is a modified model of the fully connected recurrent neural network with one hidden layer and the hidden layer is comprised of self recurrent neurons. The control system consists of diagonal recurrent neuro-identifier (DRNI) and diagonal recurrent neuro-controller (DRNC). A controlled plant is identified by the DRNI, which provides the sensitivity information of the plant to the DRNC, where the control action takes place. To ensure faster convergence, both the DRNI and DRNC are trained using a dynamic back-propagation (DBP) algorithm.