Open AccessImproving accuracy of wind resource assessment through feedback loops of operational performance data: A South African case study
Author(s) -
David Pullinger,
Abdelfatah K. Ali,
M. Zhang,
Nigel Hill,
T. Crutchley
Publication year - 2019
Publication title -
journal of energy in southern africa
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.24
H-Index - 20
eISSN - 2413-3051
pISSN - 1021-447X
DOI - 10.17159/2413-3051/2019/v30i3a5669
Subject(s) - wind resource assessment , benchmarking , wind power , resource (disambiguation) , grid , yield (engineering) , environmental resource management , environmental science , wind speed , operations research , environmental economics , statistics , offshore wind power , engineering , meteorology , computer science , geography , business , economics , mathematics , computer network , materials science , electrical engineering , geodesy , marketing , metallurgy
This study addresses two key objectives using operational performance data from most of the Round 1 wind farms connected to the grid in South Africa: benchmarking of wind farm performance and validation of the pre-construction energy yield assessments. These wind farms were found to perform in line with internationally reported levels of wind farm availability, with a mean energy-based availability of 97.8% during the first two years of operation. The pre-construction yield assessments used for financing in 2012 were found to over-predict project yield (P50) by 4.9%. This was consistent with other validation studies for Europe and North America. It was also noted that all projects exceed the pre-construction P90 estimate. The reasons for this discrepancy were identified, with the largest cause of error being wind flow and wake-modelling errors. Following a reassessment using up to date methodologies from 2018, the mean bias in pre-construction predictions was 1.4%.