Short-Term Reactive Power Forecasting Based on Real Power Demand Using Holt-Winters’ Model Ensemble by Global Flower Pollination Algorithm for Microgrid

Figure 9

Reactive power forecasting: (a) sample-1 with a learning rate of 0.032, (b) sample-2 with a learning rate of 0.067, (c) sample-3 with a learning rate of 0.070, (d) feature index for sample-1, (e) feature index for sample-2, and (f) feature index for sample-3.

(a)

(b)

(c)

(d)

(e)

(f)