Till innehåll på sidan
Till KTH:s startsida Till KTH:s startsida

Condition Monitoring of Stator Windings with a Networked Electric Drive

Tid: Fr 2022-06-03 kl 10.00

Plats: Kollegiesalen, Brinellvägen 8, Stockholm

Videolänk: https://kth-se.zoom.us/j/63027069979

Språk: Engelska

Ämnesområde: Elektro- och systemteknik

Respondent: Giovanni Zanuso , Elkraftteknik

Opponent: Senior Lecturer Antonio Griffo,

Handledare: Professor Viktória Fodor, Nätverk och systemteknik; Professor Luca Peretti, Elkraftteknik

Exportera till kalender

QC 20220510


Electric drives are widely used in industry, where they are also part of the plant communication architecture. This architecture is challenged by the Industry 4.0 initiative that aims to make the devices in industrial plants more interconnected and with additional functionalities. These changes heavily affect electric drives, and thus their future role in industrial networks should be investigated.

The first part of this work analyzes two examples of additional functionalities for electric drives from a network system perspective: condition monitoring and multi-drive systems. The suitability of the industrial communication protocols is evaluated for both application cases. Condition monitoring and multi-drive systems are further analyzed considering EtherCAT and CAN networks. A performance model is proposed to control multi-drive systems with EtherCAT, where condition monitoring data is also considered. The transmission of bulk data originated by condition monitoring methods is considered in the traditional industrial fieldbus CAN, and an extended schedulability analysis is proposed.

The second part of this work deals with the implementation of condition monitoring algorithms for the stator winding insulation in electric machines. Initially, interturn short-circuit faults in induction motors are investigated. An analytical and a finite-element model are developed and experimentally validated by means of a motor prototype with tapped windings, which can emulate the interturn faults. Fault detection methods based on the negative-sequence current and the rotor slot harmonics are analyzed both theoretically and experimentally. The stator winding insulation condition, including the groundwall insulation, is also considered for condition monitoring utilizing the MHz-range oscillations in the stator currents after switching transitions. Such oscillations depend on the parasitic capacitances of the stator winding, which in turn relate to the insulation condition. In order to quantify the variations in the current oscillations, and thus the insulation change, two metrics are proposed and analyzed. The variations of the insulation condition are emulated by adding additional capacitors to the stator winding taps, and then induced through an accelerated aging procedure applied to the whole motor. All the experiments are conducted with a custom converter that can simultaneously perform the drive control algorithm, the interturn fault detection methods, the communication with external devices, and the MHz-range sampling.

This work shows that condition monitoring and multi-drive system control can be implemented in electric drives using existing industrial communication protocols, such as EtherCAT and CAN. This work proves that industrial converters can perform online both the detection of interturn short-circuit faults and the monitoring of the stator insulation.