Parameter estimation for measurement-based load modeling using the Levenberg-Marquardt algorithm

Accurate load models are extremely important for the operation planning and power system control. Load characteristics critically influence the power system behavior especially during events that frequently occur in the power system. The introduction of synchronized measurements helps to collect high quality data for more reliable load models. This paper presents an effective method for modeling and identification of dynamic loads. Using synchronized measurements a parameter estimation of a dynamic load model during various events is performed. For more realistic results a well-studied and accepted model, the exponential recovery load model, is used. The effectiveness of the parameter estimation method is examined through simulations applied on the 39-bus New England system.