An Improved Droop-Controlled Microgrid Using Intelligent Variable Droop Coefficient Estimation

An intelligent variable droop coefficient estimation is proposed in this study for a microgrid operated in islanded mode to improve the transient performance under sudden load variation. Owing to the constant droop coefficient of the active power/frequency droop characteristic, traditional droop-controlled microgrid has some disadvantages, such as slow transient response and poor disturbance rejection, which will result in poor power quality. Hence, the conventional droop-controlled microgrid using the constant droop coefficient is unable to provide the load support instantly and has been distinguished as the tardy load support issue. Therefore, a novel variable droop coefficient estimation is proposed in this study to overcome the shortcomings of the traditional droop-controlled microgrid under sudden load variation. Moreover, a recurrent polynomial probabilistic fuzzy neural network (RPPFNN) regarded as an angular frequency controller is developed for the intelligent variable droop coefficient estimation to rapidly estimate a suitable droop coefficient for the droop-controlled microgrid. Finally, some experimental results are given to certify the effectiveness of the droop-controlled microgrid using the proposed variable droop coefficient estimation with the RPPFNN angular frequency controller for the suppression of power and frequency fluctuations and the reduction of rate-of-change-of-frequency (RoCoF).