Comparison of structured residuals design techniques for actuator and sensor fault detection and isolation in a permanent magnet DC motor

Sensors and actuators frequently encounter unexpected deviations from their optimal operating conditions, making their reliability an essential topic in many electric motor drives. In this paper, a comparison of structured residual synthesis methods for actuator and sensor additive fault detection and isolation (FDI) in a permanent magnet (PM) DC motor, applying two different approaches to the design of primary residuals, is presented. The first method involves a standard approach and is based on system transfer matrix synthesis. The second, more convenient approach involves the synthesis of primary residuals based on the analysis of subsystems that describe the observed system. Both methods are applicable for linear and time-invariant (LTI) systems but were successfully applied to a laboratory system that shows nonlinear behavior, time-variant properties, and is affected by a constant disturbance, thanks to the proposed technique of residual translation with the previous design of internal residuals. With the proposed method, all important performance characteristics of the residual generators such as fault sensitivity, reaction speed, and robustness are achieved. The technique can be applied to all mechatronic systems in general, whose behavior can be with satisfactory accuracy described by an LTI model. Given the transfer functions and state-space model of the system in the presence of faults in the Z domain with reduced order, presented technique is computationally efficient for embedded systems. Also, in addition to fault detection and isolation, the system can identify faults by analyzing fault sensitivities, nominal faults, and stationary residual values, making it superior for most FDI applications. A real-time experiment performed on a laboratory setup, which consists of a DC motor, an amplifier designed in the form of a linear electronic circuit, and a Compact RIO 9075 real-time processor is used to develop the FDI system, to generate residuals and confirm advantages of the proposed FDI technique.