With some help of a reader from Bishop-Hill’s web site, some discrepancies have been noted.
Appears that the number used by DECC and National Grid for mean wind power/energy for the full year 2009 are different by almost three times.
As explained below, the difference could be related to DECC using a combination of actual and estimate (based on “typical load factor or the design”) generation numbers. I do not have, and probably can not get the source data from DECC (per their email this morning).
I therefore conclude, pending errors in my thinking, so far with the data I have that the National Grid numbers, which appear to be “actual” are more accurate than DECC’s actual/estimated generation numbers. Further, DECC’s computed load factors may have some issues with circular logic.
Would welcome some “peer review” here to spot any errors.
JK on 1 November 2010 at 1:44 p.m. said at Bishop-Hill:
The energy statistics section on the Decc web site publishes the information on installed capacity and actual generation from renewable sources, which suggest that for onshore wind the generation is in the range 20-25% of installed capacity.
This DECC table is reproduced here and I’ve highlighted the two areas of interest. I’ve looked at nothing else so far.
For the only year where DECC data and the source 30-minute interval power generation data from National Grid is 2009. The National Grid data is presented in GW as power. Comparing the two for the full year 2009:
The GWh (Gigawatt Hours) column above is from the DECC data in DUKES&-4-1.xls. Assuming (24 h/day x 365 days/year) 8,760 hours per year this means an average power output of 9,304 GWh / 8760 h=1.06 GW.
The mean annual power from the National Grid data is obtained by averaging the 17,520 30-minute interval data points for power produced 2009. This worked out to be 0.379 GW. The histogram of the data along with a fitted probability distribution is as shown below:
Digging further into the DECC data see Note 7 (highlighted above) where they say for the wind data:
(7) Actual generation figures are given where available, but otherwise are estimated using a typical load factor or the design load factor, where known.
Since I don’t have access to the source data I don’t (yet) know how much of their data is actual and how much is estimated for 2009. Further, if they are using “typical” and/or “design” load factors to estimate actuals, then this will certainly impact their computation of “actual” load factor. Seem like a bit of circular logic here.
Would welcome any thoughts here in comments about what else could be wrong here.