ABSTRACT
This work aims to enhance the ability of a direct current (DC) microgrid to guarantee the power supply without interruptions by considering the dynamic efficiency of each power converter in the power balance. Previous works show that the converter efficiency varies according to the instant power.
If the variable efficiency of the converters in the microgrid is not considered, some extra power must be considered to compensate the losses in the power balance. However, this leads to a waste of available energy and unnecessary load shedding. The work presented here includes the power converters dynamic efficiencies in the control of a DC microgrid to improve its performance. MATLAB/Simulink simulations were carried out and the results show that the dynamic efficiency can reduce the load shedding and improve the total DC microgrid efficiency.
EFFICIENCY OF THE POWER CONVERTERS IN DC MICROGRID
In a DC microgrid two kinds of power converters are involved: the DC/DC converter and the DC/AC converter. Their main mission is to interface the sources with the DC bus and to flexibly control the exchanged power. Although a lot of different power converters have been conceived for this purpose, this work studies only the half-bridge DC/DC converter, the full-bridge DC/AC converter and the basic boost converter as illustrated in Figure 2a,b.
Figure 2. (a) Half bridge DC/DC converter; (b) Full bridge DC/AC converter (c) Basic Boost DC/DC converter
SAFETY MARGIN AND DYNAMIC EFFICIENCY
As the solar irradiation is always varying unpredictably and the characteristics of PV panels are nonlinear, a MPPT algorithm is generally used to maximize the PV power generation. Thus, in Figure1 the PV output power, PPV, is constantly varying; similarly for the constrained load power, PL_C, is varying randomly. Hence, the powers of the auxiliary sources, i.e., the storage power, PS, and the grid power, PG, are necessary to keep the power balance.
SIMULATION RESULTS
Three Power Balancing Algorithms Comparison
In this first subsection, three algorithms are compared for the same PV MPPT power profile P pv recorded on the 6 November 2014. This day is a typical case of solar irradiation evolution in northern France, which includes relatively high irradiation and the variations caused by clouds. The load power demand profile PL_D is a simplified one to make the analysis easier. These powers are presented in Figure 5.
Figure 5. PV power profile of 6 November 2014 and load power profile
Taking into account the variation of converter efficiencies during the studied period, it is obvious that the global efficiency of the DC microgrid has also some fluctuation. Figure 11 shows the evolution of instant global efficiency, defined as η = pL_C/(pPV + pS + pG), for the DCE method. It can be seen clearly that the global efficiency does not have a direct relation with PV profile.
The fluctuation of the blue curve was evidently caused by the fluctuation of PV power, but the highest PV power on the 12 March 2015 did not correspond to the highest global efficiency. On the contrary, during some period, this day has the lowest global efficiency. This implies that the improvement of the global efficiency may need optimization of power flow dispatching in the microgrid with prediction of PV power generation and load power demand.
Figure 11. Comparison of the DC microgrid global efficiencies of the three days for DCE method
DYNAMIC CONVERTER EFFICIENCY IN POWER BALANCING AND POWER CONTROL
The process of the convergence will lead to more bus voltage oscillations, thus the power quality is degraded. In order to accelerate the convergence and improve the power quality of the DC microgrid, a feed-forward method based on estimation of efficiency can be introduced in the control loop. Thus, the following loop is implemented instead of the one above (Figure 13).
Figure 13. Improved control loop with converter efficiency estimation
CONCLUSIONS
To ensure the power quality of DC microgrid, the power balance method including load shedding is essential. Different power balancing methods are proposed and tested in simulation. It is proved that the power balancing cannot work without an appropriate PSM. The dynamic efficiency estimation enables the system to know better the power losses in the microgrid thus making more flexible and precise load shedding. The simulation tests based on different real PV profiles prove it.
The dynamic efficiency estimation is also proposed to help treat the converter power loss by changing the power control loop. According to the simulations this method helps reduce the bus voltage fluctuation but increases energy costs on cloudy days. The power range of the simulation can be easily scaled up to a few 100 kW to show more obvious economic benefits, but the conclusions will not be changed compared to the chosen low power range. In future work progress can be made by optimizing the powerflow dispatching in the microgrid with consideration of power losses to improve the global efficiency.
Source: Sorbonne University
Authors: Hongwei Wu Id | Manuela Sechilariu | Fabrice Locment