lectricity load demands was proposed in both campus and building
level. The performance of the proposed model was evaluated using
the CCHP plant data collected from the UCI campus. The results
revealed that the proposed model was able to provide high quality
forecasts for both cooling and electricity load demands. Coefficient
of determination R2 and adjusted coefficient of determination R2
adj
for forecasting cooling demand were 88.4% and 88.3%, implying that
the proposed model could explain more than 88% of the total variability
within testing data. These indices were 70.8% and 70% for
electricity load demand in weekdays and 43% and 40% in weekends
respectively. The weekend cooling demand forecasts could signifi-
cantly improve by using more stable weekend data.
The proposed model is now running in the campus and is
forecasting both cooling and electricity load demands as a part of
an integrated CCHP optimization platform. Further information
from exogenous factors such as occupancy can improve the performance
of the proposed model.