So, having used a very manual process i was able to identify a fault in the system. It turned out that the fault was in the inverter. Now i appreciate that looking at the ratio of power, voltage or current between the strings (for those who have two different strings of panels) isn't going to help you identify the exact fault as there are many variables like shading, differences in orientation of differenr strings, panel faults or as in my case in one of the mppt chanels in the inverter. But at least it can give you an indication that something is wrong, there were no inverter errors in my case. And there is no one solution here that will work for everyone. In my case the panesl are in two strings, 6 & 8. So, give or take conversion losses etc i was expecting the power ratio to be circa 0.75. For me both strings have the same orientation and no obvious selective shading. But, in different set ups then this may not be quite that simple. But, if you can visualise/analyse your data then you can start to work out what normal looks like. And yes, i know we shouldn't have to.
So, with some help from co-pilot i have a power query (you can use this in Excel or Power bi) which will pull your data from the cloud. I also have a query that then calculates the required ratios. So, i used power bi and with that you can compare different days. So here we have two days about 1 month after the inverter was replaced and two daya from this month. You could have a month's worth of days to scroll though if you wanted. The 2nd image just includes the sollar array power outputs (2nd y axis). And the 3rd image is each of the ratios on their own graph in a bar chart for day 52.
I will pop the scripts and a description of how you can use them into later posts.
