No abstract

In an off-grid hybrid wind-diesel-battery system, the diesel generator is often not utilised efficiently, therefore compromising its lifetime. In particular, the general rule of thumb of running the diesel generator at more than 40% of its rated capacity is often unmet. This is due to the variation in power demand and wind speed which needs to be supplied by the diesel generator. In addition, the frequent start-stop of the diesel generator leads to additional mechanical wear and fuel wastage. This research paper proposes a novel control algorithm which optimises the operation of a diesel generator, using genetic algorithm. With a given day-ahead forecast of local renewable energy resource and load demand, it is possible to optimise the operation of a diesel generator, subjected to other pre-defined constraints. Thus, the utilisation of the renewable energy sources to supply electricity can be maximised. Usually, the optimisation studies of a hybrid system are being conducted through simple analytical modelling, coupled with a selected optimisation algorithm to seek the optimised solution. The obtained solution is not verified using a more realistic system model, for instance the physical modelling approach. This often led to the question of the applicability of such optimised operation being used in reality. In order to take a step further, model-based design using Simulink is employed in this research to perform a comparison through a physical modelling approach. The Simulink model has the capability to incorporate the electrical and mechanical (Simscape) physical characteristics into the simulation, which are often neglected by other authors when performing such study. Therefore, hybrid system simulation models are built according to the system proposed in the work. Finally, sensitivity analyses are performed as a mean of testing the designed hybrid system's robustness against wind and load forecast errors.

This paper explores the impact of EMS discrepancies that a microgrid may experience due to different time horizon and implementation environment in higher-level energy managements systems as compared to lower level control frameworks. These include time-shifting, magnitude deviation, and averaging effect of the renewables and load demand profiles. We use the resilient microgrid concept as a case study considering that the reliability of such system is crucial, especially during islanding mode. We demonstrate through experiment that the non-ideal effects naturally take place in a microgrid system, resulting a discrepancy between the measured and expected ideal battery state-of-charge. Accordingly, the resiliency of the microgrid may be affected if unplanned load shedding took place in order to not violate the lower limit of the battery state-of-charge. Allowing the battery to discharge at a higher depth may temporarily solve the problem, however, this comes at the expense of a higher rate of battery degradation. Instead, we proposed a power sharing scheme and by interconnecting the microgrids, the resiliency may be improved.

This paper presents the modeling and characterization of the tower shadow effects using a wind turbine emulator in a laboratory environment. In particular, the downwind wind turbines are considered here as their tower shadow effects are more significant compared to the upwind counterpart. Simulation and experimental results have shown that the wind speed deficit due to this non-ideal effect is significant. In addition, the tower shadow effects occur typically two to three times per revolution, depending on the number of blades. The modeling of the tower shadow profiles for tubular and four-leg tower configurations are presented. Typically, these towers are used in small wind turbine applications. The tower shadow profiles are emulated experimentally using a wind turbine emulator with its characteristics being explained. The limitations of emulating the tower shadow effects using a wind turbine emulator are demonstrated through the frequency response test performed in this work. In this research, the wind turbine emulator is connected to an isolated grid which is formed by three single-phase inverters. Finally, the paper concludes with a sensitivity analysis of the power oscillations for different widths and magnitudes of the tower shadow profile.

No abstract