Loading...
2 results
Search Results
Now showing 1 - 2 of 2
- A Fault Diagnosis Scheme Based on the Normalized Indexes of the Images eccentricity for a Multilevel Converter of a Switched Reluctance Motor DrivePublication . Amaral, Tito G.; Pires, Vitor; Foito, Daniel José Medronho; Cordeiro, Armando; Rocha, José Inácio Pinto Rosado; Pires, A. J.; Martins, J. F.This paper addresses the fault detection and diagnosis of a fault in the switches of the Switched Reluctance Machine (SRM) power electronic converter. Due to the advantages of using multilevel converters with these machines, a fault detection and diagnosis algorithm is proposed for this converter. The topology under consideration is the asymmetric Neutral Point Clamped (ANPC), and the algorithm was developed to detect open and short circuit faults. The proposed algorithm is based on an approach that discriminates eccentricity of the images formed by the converter voltages. This discrimination is realized through the development of normalized indexes based on the entropy theory. Besides the different fault type the algorithm is also able to detect the transistor under fault. The possibility to implement the proposed approach will be verified through simulation tests.
- Fault detection and diagnosis technique for a SRM drive based on a multilevel converter using a machine learning approachPublication . Amaral, Tito G.; Pires, Vítor; Foito, Daniel José Medronho; Pires, A. J.; Martins, J. F.SRM drives based on multilevel converters is now a solution well accepted due to their interesting features like extended voltage range and capability to fault tolerance. However, one aspect that is fundamental to ensure fault tolerance or preventive maintenance is the fault detection and diagnosis of failures in power semiconductors. In this way, in this paper it is presented a new diagnostic method for the failure of those semiconductors in asymmetric neutral point clamped converters. The proposed method will be based on the development of specific patterns that are associated to each semiconductor and fault type. The procedures presented here are based on the image identification of the currents patterns in the multilevel converter that allow the identification of distinct fault type. The pattern recognition system uses visual-based efficient invariants features for continuous monitoring of multilevel converter The proposed method will be verified through several tests in which were used a simulation tool and an experimental prototype.