Sensorless control of grid-connected brushless doubly fed reluctance wind generators

Aghakashkooli, Mohammadreza (2024) Sensorless control of grid-connected brushless doubly fed reluctance wind generators. Doctoral thesis, Northumbria University.

Text (Doctoral thesis) aghakashkooli.mohammadreza_phd(19009067).pdf - Submitted Version Restricted to Repository staff only until 28 September 2024. Download (30MB) | Request a copy

Abstract

The brushless doubly-fed reluctance generator (BDFRG) is a prospective, medium-speed substitute to a conventional doubly fed induction generator (DFIG) for wind power applications. Whilst retaining the costeffectiveness of a fractional converter rating, the absence of slip-rings and brushes makes it more reliable than the DFIG. The reluctance rotor design of the BDFRG, with half the total number of the stator poles, provides magnetic coupling between the stator windings with different numbers of poles and applied frequencies. Suchunusual structure and operating principles bring another important reliability and cost advantage over DFIG: half the ‘natural’ synchronous speed for the same number of rotor poles and line frequency. Hence, a simpler two-stage gearbox can be used in place of a failures prone three-stage one for DFIG wind turbines. The existing vector control strategies for the BDFRG require the information of rotor position and speed obtained by an electro-mechanical speed/position sensor. Therefore, this study presents different sensorless control methods based on model reference adaptive system (MRAS) to further improve the reliability of the BDFRG-based wind turbine. The parameter sensitivity analyses of these methods have resulted in proposing innovative error compensation techniques and, most importantly, the development of a new, viable, dual MRAS observer which raises a truly machine parameter independent estimation of the rotor speed and position by estimating the inductance ratio on-line. The proposed approach has clearly demonstrated the ability to accurately estimate the rotor angular velocity and position. As a result, the controller and observer stability with decoupled real and reactive power responses of the grid-connected winding have been guaranteed over the entire operating speed range, and without any prior knowledge of the machine parameters and their off-line testing procedures.The effectiveness of the developed methods has been verified by realistic computer simulations and real-time hardware-in-the-loop studies of a custom, large-scale BDFRG design, and further supported by experiments on a 1.5 kW proof-of-concept laboratory prototype.

Actions (login required)

View Item

Downloads