Originally posted by bcroe
View Post
For example, I've for 2 strings of 8 ea. 327's and a spec sheet for them that says the Voltage drop per panel per deg. C. of panel temp. change is -0.1766 V/deg. C, making a spec sheet Voltage change of (0.1766)*(8) = ~1.413 v/deg. C. per string. After a lot of quasi steady state measurements of near simultaneous string Voltage and measurements of all 16 individual panel temps. under clear skies at daily min. incidence angle both summer and winter, and recording as many variables as I've got the means to measure (roof ambient air temp. GHI, wind vector, dew point and other stuff) , I calc'd and use a voltage drop/deg. C constant of 1.478 V/deg. C. per string. That's got some uncertainty in it (mean = 1.478 V/deg. C, std. dev. 0.123 V/deg. C, N = 60), but it seems to produce consistent results when applied to other conditions that are useful in my attempts to measure array fouling, and other things such as possible measurements of panel deterioration as f(time), and other stuff. All that leads me to the confirming conclusion that, as a practical matter, array mpp Voltage is pretty much only f(array temp.) as most all the cogent and technically accepted literature seems to say.
Science is about precision and more about absolutes. There's no variation in Avogadro's Number for example. Most of engineering is about approximations and looking for explanations about real world variations, and ultimately accepting that for all our attempts at accuracy, in the end, and in spite of our best efforts, everything we measure is no better than an approximation.
Bottom line for engineering: Is the damn thing safe ? And does it accomplish the required task for a reasonable price ? The rest of the application decisions involve economics and politics, which, in the real world, will always govern.
Comment