Yes indeed, changing the bin arrangement such that the centers are at simulated runs, helped. But still, there is some difference I found in averaged extreme value (calculated) between MExtremes and Bladed.
So, for example, MExtremes calculated the averaged extreme value as 109.9 kNm and the closest/higher simulated extreme to this was 113.05. Now Bladed only tabulates the closest/highest extreme and this was 109.12 kNm which means its calculated average extreme was less than this value. Hence a difference of 3.5 % in loads is seen here.
I would think averaging of extremes should be straightforward in both softwares. Would ‘peakfinder’ makes a difference? Currently, I have this option switched off. Could you think of any differences? Maybe the binning is still a bit off.
And yes, I use the output files generated by FAST and read in Bladed to perform post processing. Hence, there is no difference in the output time series.
Which of the two results looks more correct based on the time-series data you are post-processing? I don’t know much about GH Bladed’s post-processor, but does GH Bladed give you any other information that could help you debug the source of the problem?
The peak finder (FindPeaks) option in MExtremes fits a quadratic curve to the peak to more accurate identify the peak; I would expect this option to only make a difference if the time series is too coarse (and the maximum peak would get higher, not lower).
Perhaps it would help to identify the source of the problem by simplifying the analysis by only post-processing one or a few time series to start.
We seem to have found the reason of the difference. A colleague of mine checked the FAST and Bladed simulations and he found Bladed gives/tabulates the ultimate loads that is "closest to averaged maximum " while FAST gives/tabulates the value that is " closest and higher than averaged maximum ".
For eg. when the averaged maximum value of loads is 1.4, then Bladed could give 1.38 while MExtremes could give 1.45. Hence, in general one could say MExtremes would give conservative results.
Do you agree to it or have any thoughts on the two methods? I think IEC 61400-1 ed. 3 is not really clear on it.
Interesting. NREL implemented MExtremes following our understanding of the standard requirements. It seems like the developers of GH Bladed have a slightly different understanding. I agree that the approach in MExtremes is a bit more conservative. I think this question should be raised to the standards committee or certification bodies responsible for ensuring compliance to the standard.