This paper deals with discrete-time statespace current controllers for three-phase grid converters equipped with an LCL filter. The integral action in the controller can be implemented either using an integrator or a disturbance observer. The results show that the disturbance-observer-based and integrator-based controllers become mathematically equal if the feedforward gains are selected to be equal, the feedforward zero is placed to cancel the pole originating from the integral action, and the closed-loop poles are placed identically. The equivalent performance in both designs is verified by means of analyses and experiments. The equivalence is also shown for double-frequency current controllers.

The stability of the converter-grid interconnection can be studied by analyzing the product of the converter output admittance and the grid impedance. For reliable stability analysis, it has been of interest to obtain accurate converter output admittance models for a wide range of frequencies, ideally also around and above the Nyquist frequency of the converter system. This paper presents a modeling method for the output admittance of power converters defined in the Laplace domain that takes into account the discrete nature of the control system. The modeling method is based on analyzing the intersample behavior of sampled-data systems, a class of systems which includes the modern digitally-controlled power converters. The proposed method is compared to conventional admittance modeling methods, and its accuracy is validated by means of simulations and experiments.

This paper presents a real-time identification method for LCL filters used with three-phase grid converters. The method can be applied to identify both the inductance and capacitance values of the filter and the series resistance seen by the converter. As a side-product, an estimate of the grid inductance seen from the point of connection is also obtained. A wideband excitation signal is added to the converter voltage reference. During the excitation, converter current and converter voltage reference samples are used for identification. The samples are preprocessed in real time by removing DC biases and significant grid-frequency harmonics. Parameters of two discrete-time models are estimated at each sampling instant with a recursive estimation algorithm. Depending on the estimated model, the model parameter estimates are translated to either the resistance or the inductance and capacitance values of the system. The method can be embedded to a control system of pulse-width-modulation (PWM) based converters in a plug-in manner. Only the DC-link voltage and converter currents need to be measured. Simulation and experimental results are presented for a 12.5-kVA grid converter system to evaluate the proposed method.

This paper presents a real-time algorithm for identifying the inductance and capacitance values of LCL filters used with grid converters. As a side product, the grid inductance seen from the point of common coupling is also estimated. A wideband excitation signal is added to the converter voltage reference. During the excitation, the converter currents and the converter voltage reference are sampled. The samples are preprocessed in real time by removing DC biases and significant grid-frequency harmonics. Parameters of a discrete-time model are estimated at each sampling instant with a recursive estimation algorithm. The model parameter estimates are translated into inductance and capacitance values. The method can be embedded to a control system of PWM-based converters in a plug-in manner. Only the DC-link voltage and converter currents need to be measured. Simulation and experimental results are presented for a 12.5-kVA grid converter system to evaluate the proposed method.

This paper deals with discrete-time state-space current control of three-phase converters equipped with an LCL filter. Either the converter or grid current is measured and the unknown states are estimated using a reduced-order observer. The stability and dynamic performance of the control designs based on these two current measurement options are compared by means of analysis and experiments at different sampling frequencies and under varying grid conditions, ranging from strong to very weak. Equal reference-tracking performance under nominal conditions is used as a basis for comparison between these two options. If a strong grid is assumed in the control tuning, the controller based on the grid current measurement (GCM) is found to be more robust against varying grid conditions in a wide range of sampling frequencies than the controller based on the converter current measurement (CCM). The CCM leads to better dynamic performance as compared to the GCM if the resonance frequency of the system falls below the critical resonance frequency.