Abstract:
After having reviewed, all the existing deterministic and stochastic models,
being used currently, all over the world, to generate data for non-
conventional energy systems, a need was felt for appropriate mathematical
models to generate solar radiation and wind speed for realistic operating
conditions for energy systems, by mathematical equations characterizing the
deterministic and the stochastic components of the long term measured
records, specially from the point of view of our country, Pakistan.
Stochastic modeling of hourly global solar radiation; Markov transition
matrix (MTM) model; triangulation method (modified Angstrom model) for
mean monthly daily global solar radiations; autoregressive moving average
(ARMA) and autoregressive integrated moving average (ARIMA) models for
wind; discrete state stochastic simulation of daily averaged wind speed;
simulation of wind energy for different wind turbines are the models, which
we fitted and tested their validity for solar and wind data of different cities of
Pakistan. The generation of synthetic sequences using these models was also
done and the results obtained by these models were compared.
Using stochastic modeling of hourly global solar radiation, AR(1) models
give the best results showing good agreement between generated and
observed solar data. MTM approach to generate hourly averaged global
radiation led to a reliable simulator. The method of triangulation was found
better than any single regression equation methods. For generation of wind
data, ARMA(2, 0); ARMA(2, 2); ARMA(1, 0) and ARMA(2, 0) are the best
fitted models for winter, spring, summer and autumn respectively. The time
series model, which we used
to generate wind
power for different turbines
suggested MOD2 turbines, the best suited for the coastal area considered.
