PROJECT TITLE:

Neural Network-Based Model Predictive Control: Fault Tolerance and Stability - 2015

ABSTRACT:

This brief deals with nonlinear model predictive control designed for a tank unit. The predictive controller is realized by suggests that of a recurrent neural network, which acts as a one-step ahead predictor. Then, based on the neural predictor, the control law comes solving an optimization drawback. An vital issue in management theory is stability of the management system. During this transient, this problem is investigated by showing that a price operate is monotonically decreasing with respect to time. The derived stability conditions are then used to redefine a constrained optimization problem in order to calculate a management signal. As the automated management system can prevent faults from being observed, the management system is supplied with a fault diagnosis block. It is realized by suggests that of a multivalued diagnostic matrix, which is determined on the premise of residuals calculated employing a set of partial models. Each partial model is designed in the shape of a recurrent neural network. This temporary proposes also a technique of compensating sensor, actuator, and method faults. When a sensor fault is isolated, the system estimates its size and, based mostly on this info, the controller is fed with a determined, close to real, tank level price. Actuator and process faults will be compensated thanks to application of an unmeasured disturbance model.