Enhanced control approaches for harmonic compensation and reactive power support for improving grid power quality

Industrial loads are predominantly composed of electrical loads of nonlinear behaviour such as frequency converters, motor drives and power converters. These are the primary sources of harmonic pollution in grids, resulting in network loss contributing to poor efficiency, a degraded power factor and damage to equipment connected to the grid. Thus, the power quality of the electrical grid is severely reduced, raising concerns for utility and consumers. Active power filters (APF) have evolved as the best solution to perform harmonics compensation in power systems. Another power quality concern that is expected to dominate future distribution grids is overvoltage issues arising from high PV penetration. Numerous studies have shown that reactive power support (RPS) from grid-interfaced power converters (DC-AC) can be an effective solution for regulating grid voltage within the nominal range. In this regard, the primary focus of this thesis is to explore the scope for improving APF performances through advanced control techniques and suppressing overvoltage problems through RPS by using end-user reactive capable devices.

Harmonic extraction is the key aspect of APF control algorithms to generate a reference current for harmonic compensation. These are generally based on utilizing Phase-Locked-Loop (PLL) algorithms. However, PLL performance is prone to grid-voltage scenarios and requires additional structures (filters) to make it robust against abnormalities in grid-voltage. This increases the complexity in implementation and consumes excessive microcontroller resources. Another aspect is the performance of inner loop current control, which is responsible for generating the output current close to the reference current with minimum error. An accurate mathematical model that can accurately predict the dynamics of an APF is essential for designing optimal control parameters. However, the understanding of APF dynamics and controller design are hindered because of complex mathematical models of a higher order.

Multiple reactive capable devices are present in low voltage distribution systems, such as residential photovoltaic systems (PV), Electric vehicles (EV), and Home appliances (HA) have been suggested as solutions for overvoltage problems by extracting RPS from them. Research studies have focussed on only one device at a time for grid RPS, and there is a lack of research on coordination among multiple reactive capable devices while they participate in working towards the same common goal.

This thesis presents solutions to the limitations mentioned above as follows: The first contribution of this thesis is to present a novel harmonic extraction approach based on the Trigonometric Orthogonal Principle (TOP) and a self-tuning filter (STF). The TOP method provides a simple and fast approach to extracting the reference current, while STF provides a simplified structure for generating the required synchronization signal. Overall, the proposed method can extract the reference current with the fewest calculations, which can be implemented in low-cost microcontrollers. The proposed method executes ten times faster than the conventional DQ method. The effectiveness and advantage of the proposed structure are verified through simulation and experimental results.

The second contribution is the development of a simplified mathematical model of a three-level neutral-point clamped (NPC) inverter-based APF for medium-voltage high-power applications. Firstly, a novel and simple approach based on state-space averaging is proposed for deriving a mathematical model of a three-level NPC inverter. Secondly, the derived model is used for an optimized controller parameter design for the inner current control loop and the outer voltage control loop. Finally, the accuracy and performance of the derived model are verified through simulation and experimental results. The proposed approach greatly simplified the controller parameter design procedure.

The third contribution of this thesis relates to the development and implementation of a coordinated reactive power support methodology that utilizes the demand-side flexibilities of end-user reactive capable devices for mitigating overvoltage problem. Two device prioritization strategies are proposed that consider the reliability of reactive power capable consumer devices and management complexity, thereby allowing consumers to either enhance the candidate devices’ lifetime or reduce management complexity while participating in grid support. A cloud-based architecture is implemented to optimally coordinate consumers’ reactive power capable demand-side resources such as EVs, solar PV systems, flexible home appliances, etc. considering their varying characteristics, ratings, and purposes.