Minimum-Loss Control Strategy for a Dual-VSI DFIG DC System

This article addresses the minimum-loss control of the dual voltage-source inverter (VSI) and doubly fed induction generator (DFIG) system connected to a dc link. The minimum-loss operating conditions for field-oriented control based on the airgap flux are obtained analytically using Lagrange multipliers and validated with numerical optimization. As the main contribution of this article, the analysis accounts for core and VSI losses, providing the optimal stator frequency law and rotor/stator d-axis current split ratio, and an implicit expression for the optimal flux trajectory formulated as equality between suitable d-axis and q-axis loss functions. In the proposed implementation, this implicit condition is enforced by using a proportional-integral controller and avoiding look-up tables. Furthermore, the stator and rotor VSI controls are implemented in two independent digital signal processors with no communication, which may ease the use of off-the-shelf VSI units. The optimal conditions and control strategy are fully validated by simulations and experiments on a prototype. The main scope of application is wind-energy dc-grid technology.

I. Introduction

In The last two decades, the doubly fed induction generator (DFIG) has been a popular choice for wind energy conversion systems (WECSs) interfaced with the ac mains, due to the significant savings in the power converter—usually rated to 30% of the overall WECS power [1]. With the growing interest towards dc transmission and distribution systems—especially for wind farm interconnection [2] and dc microgrids [3], [4], researchers have started looking at the options and benefits of integrating a DFIG into a dc power network. A simple solution to minimize the number of controlled converters and their power rating is the so-called DFIG-dc system. It adopts only a single, de-rated voltage-source inverter (VSI) on the rotor side and a fully rated diode bridge on the stator, both connected to the same dc-link or dc grid [5]. This scheme effectively implements the cheapest power electronics while allowing high dynamic control and some degree of freedom to optimize losses, but suffers from some severe drawbacks, such as harmonic-related extra losses and torque-ripple. Although the torque-ripple can be mitigated at the control level [6], torque-ripple and harmonic compensation cannot be achieved simultaneously [7], without adding extra hardware such as active filters or multipulse rectifiers. In order to improve the DFIG-dc system, Nian et al. [8] proposed to replace the diode-bridge with a stator-side VSI and operate the DFIG with a slip equal to −1, then realizing a dual-VSI DFIG-dc system where both the VSIs convert half of the total power. The dual-VSI topology for wound-rotor induction machines (WRIM) was explored in the past for high-power motor drive applications in order to achieve a 2-p.u. speed range and using converters rated to 1 p.u. [9]–[13]. This inherent modularity is attractive for multi-MW wind turbines, where a larger speed range can also improve the amount of yearly energy harvested.