Title :
Acceptability of Four Transformer Top-Oil Thermal Models—Part II: Comparing Metrics
Author :
Jauregui-Rivera, Lida ; Tylavsky, Daniel J.
Author_Institution :
Arizona State Univ., Tempe
fDate :
4/1/2008 12:00:00 AM
Abstract :
The acceptability of four transformer top-oil thermal models is examined vis-a-vis training with measured data. Acceptability is defined as having the qualities of adequacy, consistency, and accuracy. Metrics are used to characterize the likely deficiencies of these models. It is shown that the classical IEEE/ANSI standard model is unacceptable for model identification purposes. The linear top-oil model is acceptable for FOFA transformers but not NOFA. The models by Susa et al. [7] and Swift et al. [5], [6], may be useful for FOFA transformers and NOFA transformers. Further research with larger training data sets is warranted.
Keywords :
matrix algebra; power transformers; thermal analysis; transformer oil; FOFA transformers; IEEE/ANSI standard; NOFA transformers; thermal model; training data sets; transformer top-oil; ANSI C57.91; system identification; top-oil temperature; transformer; transformer thermal modeling;
Journal_Title :
Power Delivery, IEEE Transactions on
DOI :
10.1109/TPWRD.2007.905576