Difference between annual energy production and sum of all months

[EC]

Hi,
Please, can someone explain me why is there difference between annual energy production and sum of energy production of all months?
You must see image below:

http:// postimg.org/image/mrosqnacj/full/

Best regards,
Eurico Cordeiro

[Duncan]

Can you explain what you have done here?

Do you have continuous data from February 2012 to July 2013? (18 months?)

Do you have no data from October, November and December 2012? (15 months?), or were these omitted from your screenshot?

Are you calculating separate tab files for each month and then making separate GWCs/Atlases? Are you making per-month tabs and atlases (so you include data from March 2012 and March 2013, for example)?

Are you applying weighting according to the different number of days in each month?

[EC]

Hi,

October, November and December 2012 were omitted during upload. Sorry!
http:// img812.imageshack.us/img812/930/fpu9.png

Yes, I made separate tab files for each month. When I have data from the same month in more than one year I join that in a single tab file.
I made calculation for each month and in this case I am not applying weighting according to the different number of days, I am dividing per 12.

Thank you!
Eurico

[Duncan]

Hello again,

If you apply weights to the monthly annualised AEP figures you have calculated, then the weighted sum is about 1059K: almost a match for your 1060K value.

I think that the six months for which you have two sets of data (Feb-Jul) should be doubly-weighted before summation.


Month Annualised AEP /12 Data months Weight Monthly weighted
Jan 85580 1 0,666666667 57053,33333
Feb 83446 2 1,333333333 111261,3333
Mar 76854 2 1,333333333 102472
Apr 76684 2 1,333333333 102245,3333
May 79044 2 1,333333333 105392
Jun 81841 2 1,333333333 109121,3333
Jul 99106 2 1,333333333 132141,3333
Aug 119908 1 0,666666667 79938,66667
Sep 114360 1 0,666666667 76240
Oct 111491 1 0,666666667 74327,33333
Nov 87186 1 0,666666667 58124
Dec 75446 1 0,666666667 50297,33333

Sum 1090946 18 12 1058614


If you also correct for the different months having different numbers of days, the match improves slightly more.

Effectively, when you are using a tab created from the entire 18 month time series, you're double-weighting any tendencies which are related to the summer season. So to make the monthly numbers match, you need to replicate this bias.

I think the rule is that you should only take 12 complete months when calculating an input climate for WAsP.

Of the two 18-month results you've produced, I guess that the "sum of months" one is less wrong. Using an 18 month time series to calculate AEP is definitely not correct. Maybe one of the WAsP scientists can clarify this?

[EC]

Hello Duncan,

Thank you very much for your help and your advices.

Best regards,
Eurico