scenarios, namely summer sunny, cloudy and rainy. On a prospective
monitored day, the timer was set to turn on the air-conditioner
at midnight and turn off at the ensuing midnight. The typical
weather scenarios were based on thirty-year climatic normal data
computed by the local weather bureau [41]. To minimize the effect
of antecedent weather, for a given weather scenario, two days
before it had a similar weather scenario. Thus the monitored sunny
day was preceded by two sunny days, and the same is true
respectively for cloudy and rainy days. Each of the monitored days
was also preceded with a day without air-conditioning to ensure
that it was free from the carry-over effect of the previous day’s
artificial cooling. In other words, the monitored days began with
indoor temperature without the residual effect of antecedent airconditioning.
For a given weather scenario, similar results were
recorded. Instead of calculating average values of the monitored
days, the original data of one monitored day for each weather scenario
was used for detailed analysis in this paper.