Title :
Missing data treatment of the load profiles in distribution networks
Author :
Grigoras, G. ; Cartina, G. ; Bobric, E.C. ; Barbulescu, C.
Author_Institution :
Dept. of Power Syst., Gh. Asachi Tech. Univ., Iasi, Romania
Abstract :
In distribution systems, the determination of the load is relatively simple when measurements are available. Frequently, due to various causes such as metering and transmission equipment failures, data are missing for part or all of a day. In these cases, estimation a ldquocorrectedrdquo value must be made. The paper presents two methods (k-Nearest Neighbors, (kNNs) and Clustering methods) for treatment of missing data problems in electric distribution networks. The numerical results indicate that the methods are efficient in the estimation of the missing load values for electric distribution stations.
Keywords :
distribution networks; learning (artificial intelligence); load flow; parameter estimation; clustering methods; distribution networks; electric distribution networks; k-nearest neighbors; load determination; load profiles; missing data treatment; Clustering methods; Data analysis; Equipment failure; Failure analysis; Information retrieval; Load modeling; Monitoring; State estimation; Telemetry; clustering method; distribution networks; fuzzy kNNs method; load profile; missing data;
Conference_Titel :
PowerTech, 2009 IEEE Bucharest
Conference_Location :
Bucharest
Print_ISBN :
978-1-4244-2234-0
Electronic_ISBN :
978-1-4244-2235-7
DOI :
10.1109/PTC.2009.5282021
