Voltage control strategies for loss minimzation in autonomous microgrids
AdvisorTrzynadlowski, Andrzej M.
Electrical and Biomedical Engineering
AltmetricsView Usage Statistics
This dissertation investigates the novel idea of flexible-voltage autonomous microgrids (MG), employing several interconnectable dc buses operating in a minimum-voltage mode. In comparison with the traditional fixed-voltage MGs, the proposed MGs reduce losses to gain significant enhancement in efficiency. It is widely believed that energy systems of the future will heavily depend on MGs rich in power electronics converters (PECs). This dissertation is focused on MGs with a high degree of self-sufficiency, without precluding sporadic links with the power grid. Potential applications of those MGs include: (a) distributed generation power systems, (b) ships, land vehicles, aircraft, and spacecraft, (c) users in need of power supply impervious to vulnerabilities of the grid, and (d) localities lacking an access to a grid.Modern pulse-width modulated PECs allow rapid and wide-range changes of voltages and currents. High switching frequencies result in high power quality and fast dynamic response, but each switching event causes energy loss related to the magnitudes of input voltage and output current. In the existing MGs, the bus voltages are maintained at a fixed level. However, many heavy loads, such as electric drives, operate most of the time with a reduced voltage, which is adjusted by decreasing the voltage gain of the feeding converter. This makes the voltage pulses high and narrow. If instead the pulses were made wide and low, then with the current unchanged the conduction losses would remain unchanged, but the switching losses would greatly decrease. This observation leads to the main idea of the dissertation, namely MGs whose dc-bus voltages are allowed to fluctuate and which are maintained at the lowest possible level. Loss minimization, apart from energy savings, may be critical for autonomous MGs with a tight balance of power.In this dissertation, two methods are proposed for calculating the minimum (optimum) required dc voltage level. In the first method, a central control unit allocates the minimum required dc voltages to individual buses by employing the information obtained from control systems of the adjustable voltage loads. For example, most of the variable-speed ac motors employ the so-called constant volts per hertz strategy, in which the relation between frequency and voltage is clearly specified. In the more sophisticated high-performance drives, the instantaneous values of the desired speed, torque, and current are available, allowing the required voltage estimation from the equation of power balance.In the second method, the problem of determining the optimal dc voltage and power settings is formulated as an optimization problem with the objective function of minimizing the converter losses. Genetic algorithm is utilized in solving the optimization problem. Due to limited available power from renewables, reducing the converter losses will enhance the survivability of the microgrid and ease the cooling requirements, resulting in a more compact system. A model of a 20-bus microgrid with the dc distribution network is employed to verify the effectiveness of the proposed methods.