Finding the Range of PHEV Controller Parameters Using the Confidence Interval (CI) Estimation from Prediction Model

Abstract

The serious problem of wind energy sources always occurs in an isolated small power system is a serious frequency deviation problem which is a result of the intermittent nature of wind power. Alternative way to relive this frequency deviation is applying plug-in hybrid electric vehicle (PHEV) to control power in the system. However, improper setting PHEV power deviation controller cannot manage the real power unbalance deviation in the isolated small power system and cause worse frequency control to follow. In order to avoid unsuitable setting PHEV power charging control a machine learning method, an artificial neural network (ANN) was used to find suitable values set of PHEV controller parameters. In selection the PHEV controller parameters using confidence interval (CI) from the suitable values set for optimizing the PHEV controller parameters. The results show the superior frequency control effects of the proposed PI controllers.