Substation day-ahead automated Volt/VAR optimization scheme

This paper proposes a new control algorithm to run a Volt/VAR optimization (VVO) scheme. The expected benefit of the proposed VVO control algorithm is to increase effectiveness of the VVO scheme by identifying optimal times when VVO scheme needs to be turned On and Off. The VVO effectiveness is measured in terms of saved kWh. A NN (Neural Network) based prediction model is used to identify optimal VVO strategy. A NN is designed to provide day-ahead hourly energy prediction at each substation with VVO scheme by using a number of predictors, such as hourly ambient conditions, day of the week, time of the day, etc. A NN is used for its ability to model nonlinear and complex interactions of predictors to provide accurate day-ahead predictions. The proposed approach is illustrated using an actual case study.