Load Estimation for Distribution Feeder Monitoring & Management

load modeling;load estimation;distribution feeder;distribution feeder automation

Singh, Urvir ; Mesut E Baran, Committee Chair,Kimberly Weems, Committee Member,Winser E Alexander, Committee Member,Singh, Urvir ; Mesut E Baran ; Committee Chair ; Kimberly Weems ; Committee Member ; Winser E Alexander ; Committee Member

University:North Carolina State University

关键词: load modeling;    load estimation;    distribution feeder;    distribution feeder automation;   

Others  :  https://repository.lib.ncsu.edu/bitstream/handle/1840.16/6306/etd.pdf?sequence=1&isAllowed=y

美国|英语

来源: null

【 摘 要 】

Load Estimation is an indispensable tool for distribution system studies, since knowledge of load profiles along the feeder has direct influence on system planning and operation activities. The main difficulties in the load modeling result from the random behavior of loads, diverse load shapes at customer sites, limitation and uncertainty in the information on loads.This thesis explores a new technique of load modeling and estimation on distribution systems. With the AMI technology on the distribution systems, real-time data about customer loads would be available at the control center, hence an estimate of loads on the distribution feeder can be made. With this estimation and the temperature forecast, a load model predicting the real-time load variations, can be made. This thesis elaborates the statistical approach used to build such a harmonics-based model with auto-correlated errors (a time series model). A time series approach to model and predict the random behavior of distribution feeder loads is explained, by harmonically decomposing the seasonal and daily variation of load consumption. With the historical power data of residential and commercial class available, statistical tools are used to perform load estimation on distribution feeder using SAS (Statistical Analysis System). Various load data sets can be grouped or clustered together, using available ‘clustering analysis’ techniques. The data of a meter whose readings are not available at any time instant can be estimated using the proposed time series method and other available meter readings from its respective cluster.