DESIGN OF A DISTRIBUTION DATABASE WITH CUSTOMER DAILY LOAD PATTERNS

A distribution automation system requires a complete and capable database to solve the operation and analysis problems in the distribution system. In this paper, a generic database for distribution automation has been developed and a distribution database of an actual Taipower feeder which serves a mixture of different types of customers has been created to support various functions of distribution system analysis. To demonstrate the applications of the distribution database a three-phase load flow program retrieves data from the database and solves the performance of a distribution feeder over a daily 24-hour period during the summer season. Since the load characteristic of a distribution system varies with time, a regression technique was applied to model load patterns for each type of customer according to the field test data. By the load patterns derived, the load behaviors of a distribution system can be represented more precisely for the computer simulation. The proposed load patterns are stored as an attribute of the customer record in the database and applied to the studies of distribution feeders.