Grid Synchronization Control Schemes for a Three Phase Grid connected Photovoltaic System with Power Quality Disturbances

Abstract

Synchronizing PV system with grid encounters a number of control challenges for maintaining the grid codes. These include various power quality problems such as voltage and current harmonics, voltage sag and swell, and grid frequency fluctuation. In this thesis it is intended to design suitable grid synchronization control scheme for a three-phase single stage grid connected PV system (TPSSGCPVS) considering power quality disturbances. The proposed control schemes are implemented both by simulation in MATLAB/Simulink followed by experimentation on a prototype TPSSGCPVS developed in the laboratory.

Firstly, the thesis focuses on the design of a Self-Tuning Filter-Proportional Integral (STF-PI) grid synchronization control scheme for TPSSGCPVS. The Self-Tuning Filter (STF) extracts the fundamental voltage of the distorted Point of Common Coupling (PCC) voltage and load current without any change in phase and amplitude of the fundamental component. To verify the effectiveness of the proposed STF-PI control scheme, a comparative analysis on its performance with that of the Improved Linear Sinusoidal Tracer-PI (ILST-PI) control scheme is pursued. Form the stability analysis, it is observed that the STF-PI control scheme has a wider range of stability region as compared to the ILST-PI control scheme. Simulations are performed by implementing these control schemes on a PV system considering the power quality disturbances. Subsequently, the proposed STF-PI control scheme is implemented in real-time on a prototype TPSSGCPVS developed in the laboratory. From both the simulation and experimental results obtained, it is verified that the proposed STF-PI control scheme provides effective grid synchronization of the PV system. Along with maximum PV power injection into the grid, the proposed STF-PI control scheme provides efficient harmonics compensation under PCC voltage distortion, load current distortion, load fault, PCC voltage sag and swell. From the obtained results, it is observed that using STF-PI control scheme, the grid current is maintained sinusoidal by reducing the harmonics as compared to ILST-PI control scheme. The current is injected into the grid at Unit Power Factor (UPF) by reducing the reactive current component to almost zero. The grid currents are maintained balanced and sinusoidal with reduced distortion despite load fault. The THD of the grid current is reduced from 26.7% to 4 % using the above STF-PI control scheme, satisfying the limits prescribed by the IEEE 519 grid code. The THD of the grid current is reduced to 4 % using STF-PI control scheme even under PCC voltage sag and swell conditions.

It is observed that, with a fixed cut-off frequency of STF, the THD of the grid current varies in real time as grid frequency varies. In order to further reduce the THD and reduce the THD variation, Extended Kalman Filtering (EKF) and Iterated EKF (IEKF) algorithms are employed for grid synchronization of a PV system. EKF and IEKF are used to estimate the fundamental sinusoidal component of the PCC voltage. IEKF uses an iterative loop to reduce the mean square error and increases the convergence speed of the grid current. From the simulation results it is observed that the grid current reaches the steady state faster using IEKF-PI control scheme than using EKF-PI and STF-PI control schemes. The THD of the grid current in real-time is reduced to the lowest value of 3.5% using the proposed IEKF-PI control scheme than the corresponding values of 3.6% and 4% respectively yielded in case of EKF-PI and STF-PI control schemes. Even by changing the grid frequency, the grid current is maintained sinusoidal using IEKF-PI control scheme. THD variation is minimized using IEKF and EKF-PI control schemes than STF-PI control scheme.

As IEKF uses the Jacobin matrix for linearization, the estimation accuracy is limited to first order approximation of the Taylor series. Unscented transformation is a nonlinear transformation, which propagates the mean and covariance through a nonlinear function. A set of sigma points is chosen to preserve the nonlinear nature of the system. Firstly, an Unscented Kalman (UKF) is proposed to further reduce the THD of the grid current and reduce the THD variation. In UKF, the sigma points are determined by finding the square root of the error covariance, obtained using Cholesky decomposition. In order to apply Cholesky decomposition, the error covariance matrix must be positive semi definite. The loss of the positive definiteness may result in stopping the UKF to run continuously or even cause divergence. To resolve the difficulties encountered in UKF, a Square Root Cubature Kalman Filter (SRCKF)-PI grid synchronization control scheme is proposed. From the obtained results, it is observed that the variation in THD of the grid current is minimized by both UKF-PI and SRCKF-PI control schemes as compared to IEKF-PI control scheme. The THD of the grid current is reduced to a lowest value to 3.2 % using SRCKF-PI control scheme than the corresponding values of 3.3% yielded by UKF-PI control scheme.
Form the obtained results with all the proposed control schemes for grid synchronization of the PV system, it is observed that these are able to maintain the grid codes by reducing the THD below 5%. However, the SRCKF algorithm essentially transmits the square root factors of the predictive and posterior error covariance in order to eliminate the square root operation thus providing the best convergence speed which minimizes the settling time. The estimation of the fundamental component of the PCC voltage using the cubature points in SRCKF algorithm provides increased estimation accuracy. As a result, the THD of the grid current is thus minimized to a value of 3.2 % using SRCKF-PI control scheme. It is thus concluded that amongst all the proposed controllers, SRCKF-PI control scheme exhibits the superior grid synchronization control performance with power quality disturbances.