POWER-SYSTEM DYNAMIC LOAD MODELING USING ARTIFICIAL NEURAL NETWORKS

The dynamic characteristics of power system loads are critical to obta ining quality operating point prediction and stability calculations. T he composition of components at a load bus makes the aggregated behavi or too complicated to be expressed by a simple form. Armed with the th eorems recently developed on the approximation capability of artificia l neural networks we devise a load model to describe the complex dynam ic behavior of loads. Real field data are used to train and test this model. The results verify that this model can emulate load dynamics we ll and should therefore be suitable as a representation of load for st ability analysis.