This is a method that, although designed for a climate like that of the British Isles, gives approximate results correlation techniques of order 1 in our climate.
group is based on the year in different periods of 60 to 65 days around the solstices, and apply an equation where the radiation to be calculated for the nth day, depends on values \u200b\u200bB and Cn = C * n, that every kind of weather are fixed. The sum of these two values, called trend. These values \u200b\u200bare given for each of the four periods into which the year. A trend we add a variable, consisting of what we call statistical fluctuation in L (Q (n-1) plus a random number "r (n)", which follows a normal distribution, such that its variance is calculated by variance given in the table to the trend within each period, L is the parameter value autocorrelation, also known.
is true that V (r) = V (Q) * (1-L ^ 2).
Based on available data in the main equation, we can calculate G (n):
G (n) = B + C * n + L * Q (n-1) + r (n)
but we must realize that each Q (n) depends on the previous day reached in the Q (n-1), by the following exppresión:
Q (n) = LQ (n-1) + r (n-1)
This year the first value, we will Q (n) = 0
To generate the number random, we will:
= NORMINV (RAND (), 0, ROOT (V (Q) * (1-L ^ 2))
The results obtained with the data B (MJ / m ^ 2), C (MJ / m ^ 2), V (Q) (MJ / m ^ 2) and L, for each period of 60 or 65 days, are shown in the chart attached.
0 comments:
Post a Comment