Fault Detection of Electromechanical Actuators via Automatic Generation of a Fuzzy Index

Abstract

Regardless of dimensions, complexity, or purpose of a mechatronic system, fault detection (FD) represents a critical element to achieve robustness and safety, while reducing risks and maintenance costs. Important requisites for an FD component include flexibility, performance, platform independence, and interpretability. To pursue these goals, this article presents a framework for the FD on electromechanical systems based on a fuzzy inference system. Fuzzy logic (FL) is a powerful modeling tool whose applicability to complex systems, however, can be limited by the effort required to specify the fuzzy inference rules, whose number grows rapidly with the number of input variables and membership functions (MFs). To this purpose, we propose the Fuzzy INDices (F-IND) framework as the foundation for an FD system; F-IND automatically generates the fuzzy rules from the specification of qualitative best and worst cases on the MFs of each input variable, which greatly reduces the burden on model design. This article presents the mathematical model of F-IND, its implementation, and its application to the specific case of FD of electric motors, evaluating its computational performance and the accuracy in both simulated and experimental settings.