Actuator's Fault Diagnosis In Control System Based On Data Fusion

As the performance terminating parts of the control system, the actuator iseasier to make faults. The actuator's faults always make direct effect on the controlperformance. So that the actuator's fault is detected in time and accurately isimportant for the control system's operating reliably and stably.In this paper, at first, multi-sensor data fusion theory is discusses;then theoperation principle and the usual faults of the electric actuator are introduced, onthis base, the experiment platform of actuator's fault is build, the typical fault isdemonstrated in the experiment and the fault curves are recorded through theexperiment;at last, two-step fusion method of fault diagnosis is presented, whichfuse the neural networks and D-S evidential theory. When the experiment data areinput into the diagnosis model, the result showed that the diagnosis method candiagnose the actuator's faults accurately.