In this paper an observer-based controller for Voltage Source Converters (VSCs) feeding constant power loads is introduced. The observer is used for estimating the power load and this estimate is feedforwarded to build a nonlinear controller in order to improve the whole system performance. The proposed observer presents linear error dynamics. The main advantage of our proposal is that linear error dynamics allows to design the observer by using linear techniques. In this way, the observer transient performance (i. e. speed of convergence) is clearly established in the whole range of VSC operation. The feedback controller is based on input-output feedback linearization considering an energy output. In this way, complete input-output linearization is obtained such that linear control strategies can be used in outer loops. Simulation results including performance in presence of commutation frequency, measurement filters delays, and parameters uncertainties are presented to validate the proposed technique.
I. INTRODUCTIONVoltage source converters (VSCs) based on switched electronic devices are being introduced by manufacturers in many industrial applications. These converters present unity power factor, sinusoidal input current, high performance of the DC voltage control, bi-directional power flow, and low total harmonics distortion index. Among others, these features are very useful in AC-DC-AC power conversion, adjustable speed drives, active power filters. In high power applications, VSCs are basic cells of several technologies introducing Flexible AC Transmission Systems (FACTS). Consequently, controllers allowing very high performance VSCs are important in current and future applications.It must be noted that VSC model is nonlinear, MultiInput-Multi-Output and it presents a non-minimum phase feature when DC-Link voltage is considered as output. In order to control VSCs several strategies have been introduced by researchers. Controllers based on Jacobian linearization around an equilibrium point have been introduced in [1], [2]. Among nonlinear approaches, fuzzy logic control [3], [4], passivity-based control [5], Lyapunov-based control [6], backstepping technique [7], and input-output linealization based on Feedback Linearization (FL) [8] have been applied. A nonlinear cascade controller including feedback linearization