daily max Heat Index logic bug? ** solved **

[sevenless]

I noticed that the initial implementation of Heat Index in 2.6a had a few quirks when the temperature was cool, resulting in weird readings of very large negative numbers (e.g. -1834.5 for [actual_th0_heatindex_f]). I assumed this was due to the way the Heat Index formula works and because it is only a valid measurement in warm (>26C/80F) and humid (>40% RH) conditions.

In 2.6b this quirk was apparently fixed, because now the actual_th0_heatindex numbers are much closer to the ambient temperature. However, there still seems to be a bug.

On dry days here (max temp of 75F/24C with low humidity), my [day1_th0_heatindexmax_f] often reports a temperature of 82F/26C with a [day1_th0_heatindexmax_time] of mid-morning (usually in the 8 am hour.) The ambient temperature at this hour is only 55F/13C, so there is apparently a bug!

Perhaps there needs to be some internal logic that checks if the ambient temperature too low to do a standard heat index calculation, and instead skips the calculation and sets th0_heatindex for that time to be equal to ambient temperature?

To see quirk this in action, my data is processed and displayed at this site. Compare the High Heat Index reading on the lower left with the temperature graph at upper right.

(Also if it is of any use in diagnosing the error, my all sensors text output is located at http://www.onsunnyside.com/weather/all_sensors.txt

[admin]

formula has been corrected for the next update.

Will be part of 2.6e