Passive Fault Tolerant Control System Using Feed-forward Neural Network for Two-Tank Interacting Conical Level Control System Against Partial Actuator Failures and Disturbances

摘要

A novel approach for passive fault-tolerant control (PFTC) system design against actuator faults is proposed. The scheme is based on Feed-forward Neural Network (FFNN) plus conventional PI controller, during the fault occurred into system FFNN will give additional control output to the system according to fault magnitude. The FFNN will trained using back-propagation algorithm. To eliminate the steady-state tracking error, the PI controller is also incorporated. The following fault type and input signals are considered: abrupt, step, sine, and trapezoidal trajectory inputs. The effectiveness and the superiority of the proposed approach are demonstrated using Two-Tank Interacting Conical Level Control System (TTICLCS) example. The simulation performed in MATLAB Simulink platform, also different integral errors like IAE, ISE and IATE are presented to validate the proposed approach.