A Hybrid Algorithm for Parameter Estimation (HAPE) for Diesel Generator Sets

Abstract

In order to simulate the dynamic phenomena of devices within a microgrid, suitable simulation models are needed. Historically, parameterization approaches have been explored for large generator sets rather than for smaller generator sets, which would be more suitable in a microgrid. Additionally, non-invasive experimental methods are often preferred over invasive methods when collecting data. Further, if there are many uncertain parameters, then it is more imperative to employ a larger computing platform. In this work, a hybrid algorithm for parameter estimation (HAPE) is utilized to find a fitting simulation model for a diesel genset. A Sobol parameter sensitivity analysis is conducted to inform the HAPE of the more influential parameters. Finally, the HAPE is developed for a supercomputing platform so that the HAPE can be executed in a massively parallel fashion.