Study on load classification of power distribution transformer based on data stream management

Traditional load classification of power distribution transformer can not reflect to the characteristics of power utilization according the fact, consideration on the characteristics of time-varying can make related work be fair. Using k-mean clustering algorithm, we classify the load of actual power distribution transformer customers according to similarity of trend in a period of time, and put forward a method of dynamic load classification based on sliding window model and DFT synopsis data structure in data stream management techniques. Experiment results show that this method is effective. (6 pages)