Grid Monitoring and Advanced Control of Distributed Power Generation Systems

The movement towards a clean technology for energy production and the constraints in reducing the CO2 emissions are some factors facilitating the growth of distributed power generation systems based on renewable energy resources. Consequently, large penetration of distributed generators has been reported in some countries creating concerns about power system stability. This leads to a continuous evolution of grid interconnection requirements towards a better controllability of generated power and an enhanced contribution of distributed power generation
systems to power system stability. As an example, the latest published grid codes stress the ability of distributed generators, especially wind turbines, to stay connected during short grid disturbances and in addition to provide active/reactive power control at the point of common coupling.
Based on the above facts, the need for improving and adding more features to the control of distributed power generation systems (DPGS) arises. As a consequence, this thesis focuses on grid monitoring methods and possible approaches in control in order to obtain a more reliable and  exible power generation system during normal and faulty grid conditions.
First part of the thesis investigates possible algorithms for fast and accurate identi cation of utility network variables such as voltage amplitude, frequency, phase angle and line impedance. Special attention has been paid to grid synchronization algorithm in terms of accurate estimation of grid voltage phase angle. Has been found that identi cation of positive and negative sequence components and in addition, the capability of the algorithm to follow only the positive sequence component, plays a crucial role in providing a clean synchronization angle even during severe
voltage unbalance caused by grid faults. Several algorithms for phase angle detection have been investigated and two novel algorithms that prove to be very robust against grid voltage disturbances have been developed.
The proposed synchronization algorithms have been further developed in order to estimate also the amplitude of grid voltage and the frequency of utility network. As a result, fast and accurate identi cation of both variables has been achieved. In addition, positive and negative sequence components of grid voltage can also be calculated. Simple, yet powerful ltering techniques, based on second order generalized integrator (SOGI) and delay signal cancellation
(DSC) have been used to separate the sequence components. Simulation and experimental results attest the accuracy and e ectiveness of the developed algorithms in identifying the frequency, phase angle and magnitude of grid voltages during severe distortions of utility network. Methods for identi cation of grid impedance have also been investigated in this thesis. The state of the art methodologies for assessing the value of line impedance have been studied, leading
to the development of two new methods for identi cation of line impedance. First method uses variations in both active and reactive power at the point of common coupling (PCC) to obtain current and voltage
uctuation in two operating points, based on which the algorithm is extracting the values for both resistive and inductive part of line impedance. Because this methodology can create voltage  fluctuation at the connection point, it can be subject for flicker emission. As a consequence, a second method, which uses a grid voltage control loop to set the reference for reactive power has been developed. If in the rst method the variations in power are made consecutively ( rst variation in active power is made then variation in reactive power follows), in the second method the variation of reactive power occurs at the same time when variation in active power happens, in order to cancel out the voltage
uctuation at the point of common coupling. The results presented con rm the accuracy and e ectiveness of both methods developed.
A considerable part of this thesis is dealing with control of grid side converter of a distribution system during normal and faulty grid conditions. The state of the art control structures for grid tied power converters have been initially identi ed and di erent types of controllers have been studied and compared. The possibility of using the information about grid variables into the control structure in order to improve the control of DPGS has also been investigated. As
a consequence, improved behavior of resonant controller has been noticed if grid frequency information is forwarded to its internal model. Additionally, controllers such as dead beat and hysteresis controller improve their robustness to parameter mismatch if the identi ed value of grid impedance is passed to the controller. Moreover, several control strategies to provide exible active and reactive power control during grid faults have been developed using the information of positive and negative sequence components. Simulation and experimental results are presented
in order to validate the above studies.
The research done in this thesis makes it possible to assess the behavior and control abilities of a grid connected power converter when running on faulty grid conditions. Furthermore, this study can serve as a good basis for deeper investigation in some particular areas and also for further research on control of DPGS during grid disturbances.