Identification of static distribution load parameters using general regression neural networks

This paper explains the motivation for and use of a general regression neural network to map temporal load class distribution data into static LOADSYN load parameters. Simulated data generated by LOADSYN is used as a training set. A general regression neural network (GRNN) is trained to achieve LOADSYN functionality, and a method is outlined for further associating the load parameters with temperature, time of day, day of week, and customer type.