|
Tuning the sky temperature correction model
The amount of clouds in the sky is directly related to the infrared radiation emitted by the clouds in the wavelength range of 8-15 µm.
The infrared sensor measures the total downward radiation in this range and it contains several components:
•The radiation emitted by the clouds.
•The radiation emitted by the atmosphere.
•The radiation reflected by the ground.
To setup the CloudWatcher one should define temperature limit below which the sky is considered clear and the temperature limit which distinguishes a cloudy sky from an overcast (very cloudy) sky.
From empirical observation, the limit between clear and cloudy conditions corresponds to a value between -6°C and -3°C (the program default value = -5°C) whereas the limit between cloudy and overcast conditions is a value between 0°C and 2°C (the program default value = 0°C).
However, during warm days / evenings, the measured sky temperature values contain a large component of other atmospheric radiation. In order to cope with this effect a temperature correction model has been introduced (please refer to Sky Temperature Correction Coefficients).
The model is a combination of a linear and an exponential relationships where the linear relationship predominates for ambient temperatures below 20°C whereas the exponential becomes more pronounced for ambient temperatures above this value.
To tune this model, one should observe the Cloud Conditions graph from sunup to sunset for a clear day.
As the ambient temperature changes during the day the Cloud Conditions graph should remain horizontal, provided the sky conditions remain stable and cloudless.
A downward trend in the graph line as the ambient temperature increases means that the calculated sky temperature correction factor must be increased.
If this trend occurs for temperatures below 25°C, the coefficient K1 should be increased by a small amount. Try a value from 33 to 38.
If the trend is more noticeable for values above 30°C, then the coefficients K3, K4 and K5 should be adjusted. Try the following combination - K3 between 4 and 10, K4=100 and K5=100.
If the graph is horizontal but too low, adjust the coefficient K2. This shifts the calculated correction value upwards as the coefficient gets smaller and vice-versa. (Note that negative values are allowed for this coefficient.)
The default values for coefficients K1, K2, K3, K4 and K5 are 33, 0, 0, 0 and 0 and this corresponds to a simple linear relationship.
The following coefficients have also proved to yield good results: K1=33, K2=0, K3=8, K4=100 and K5=100. This combination is more nonlinear for ambient temperatures above 30°C.
|
Tip: When adjusting these coefficients, use the graphical display provided (see – Sky Temperature Correction Coefficients).
|
|
