Electrical Distribution Network Measurement Data Pre-processing for Knowledge Discovery Preliminary Results

R. Pinheiro and Z.A. Vale (Portugal)


Distribution Networks; Data Mining; Measurement Data; Knowledge Discovery; Network Operation


Aiming to develop a Data Mining architecture that based on distribution networks own history, helps controlling, operating and planning them; this paper describes primary results on processing distribution networks real measurement data. Portuguese northern distribution network is described as well as the data collecting system and data itself. The existing SCADA system that allows monitoring and operating the entire network from one control center, also stores events and measures from network. Measurement data from the whole distribution is stored every 15 minutes, and therefore constitutes a very large dataset. Although this dataset contains information about the network past history, it is not used for control, operation and planning purposes, due to its largeness and complexity. On this paper one perform one first look into this database, trying to process the data in order to make it easier to understand. Data cleaning and preprocessing process is discussed, preparing future deeper Data Mining analysis. Meanwhile, just some simple data processing allows information contained on the huge distribution measurement database to come to light Some interesting network situations are then identified and discussed. Primary results are plotted and conclusions are withdrawn.

Important Links:

Go Back