Generator and Electrical System - Small Wind Turbines: Analysis, Design, and Application - page 237

have the controller programmed to use a look-up table of optimum power
versus frequency, derived from the maximum C p (for power) and corresponding
k (for frequency) from the performance curve. The power trajectory shown in
Fig. 7.1 was constructed in this manner. At any time the actual power and
frequency can be compared to the optimum values and the duty cycle altered if
there is a significant departure. It is held that this strategy is more robust than the
perturb and observe scheme, however it has at least one major shortcoming: the
optimal power curve varies with air density so a turbine in a cold climate or on a
mountain may need reprogramming. Furthermore, the optimal power curve is
derived from the steady performance curve whereas the wind is nearly always
unsteady. Given that small turbines have considerably less inertia than large ones—
recall Sect. 1.8 —there may be considerable benefit in developing unsteady MPPT
algorithms. This would require significant improvements in our knowledge of
unsteady blade aerodynamics and wake behaviour.
One aspect of MPPT that has yet to be discussed is the rapid reduction in
efficiency at part load that is shown in Fig. 7.1 which is typical of small generators,
e.g. Di Tommaso et al. [ 13 ]. Many small wind turbines operate for considerable
time at low generator efficiency, and this is often overlooked in developing MPPT
algorithms. Trade-offs may well be profitable in varying the frequency away from
the maximum aerodynamic efficiency if there is a correspondingly larger increase
in generator efficiency.
An advantage of implementing MPPT in the boost converter rather than the
inverter is that battery charging turbines are thereby able to produce more power.
However, further power electronics are required to control the battery charging if
V out varies and/or the batteries are reaching full charge.
AC output power requires an inverter to convert the DC voltage into a fixed
frequency and voltage waveform. A simple single phase ''bridge'' inverter is
shown in Fig. 11.9 where V bat is either the DC voltage from the battery bank or
V out from the boost converter in Fig. 11.8 . The term ''bridge'' refers to the output
of the inverter which ''bridges'' the ''legs'' of the inverter (devices 1 and 3 form
one leg while devices 2 and 4 form the second leg of the inverter). No filter is
employed, thus the load voltage waveform, V load , is not sinusoidal, Fig. 11.10 .
Nonetheless, this approach is sometimes taken to make cheap inverters for low
Fig. 11.9 A simple single
phase full-bridge inverter