This paper proposes a methodology for the synthesis of power system load pr
ofiles by class load study. There are more than 900 customers over various
customer classes are selected by the sampling design for the installation o
f intelligent meters over six districts in Taipower system. The load data o
f all the test customers have been collected and the typical load patterns
of customer classes have been derived for each service district. The billin
g data of all the customers in the Taipower customer information system (CI
S) are retrieved and the typical load patterns derived are applied to solve
the hourly load composition of all customer classes. The actual power cons
umption in each study district, which has been recorded by the Taipower dis
tribution dispatch control system (DDCS), is used to verify the accuracy of
the power profiles derived, By the same way, the power profiles of Taipowe
r system are then synthesized too. It is concluded that the system power pr
ofile can be estimated by the aggregation of class load patterns in a rathe
r precise manner. (C) 2000 Published by Elsevier Science Ltd. All rights re
served.