Document Type

Thesis - Open Access

Award Date


Degree Name

Master of Science (MS)

Department / School

Electrical Engineering and Computer Science

First Advisor

Reinaldo Tonkoski


Energy Management System, Hidden Markov Model, Intra-Day Forecasting, Microgrids


Accurate solar irradiance forecasting is the key to accurate estimation of solar power output at any given time. The accuracy of this information is especially crucial in diesel-PV based remote microgrids with batteries to determine the set points of the batteries and generators for their optimal dispatch. This, in turn, is related directly to the overall operating cost because both an overestimation and an underestimation of the irradiance means additional operating costs for either suddenly ramping up the backup resources or causing under-utilization of the available PV power output. Accurately predicting the solar irradiance is not an easy task because of the sporadic nature of the irradiance that is received at the solar panel surfaces. Handling the dynamic nature of the irradiance pattern requires a strong and flexible model that can precisely capture the irradiance trend in any given location at a given time. Usually, such a robust model requires a lot of input variables like weather data including humidity, temperature, pressure, wind speed, wind direction, etc. and/or large inventory of satellite images of clouds over a long period of time. The expensive sensors and database tools for collecting and storing such huge information may not be installed in remote locations. Therefore, this thesis prioritizes on developing a simple method requiring a minimum input to accurately forecast the solar irradiance for remote microgrids.

Library of Congress Subject Headings

Solar radiation -- Forecasting.
Distributed generation of electric power.
Microgrids (Smart power grids)
Markov processes.



Number of Pages



South Dakota State University



Rights Statement

In Copyright