How to determine the average baseline for the following data in SPC

hi to all,

i have a question regarding how to determine the average baseline for the following data in SPC. Hope you could help me in this one. We are actually having this OEE data since 2009 for a certain lines. Can i use SPC in analyzing these OEE data using x chart? I planned to use SPC in looking into the OEE of all lines..

from my readings i came to understand that in establishing the baseline i need to remove the outliers and that it should be in compliance with the MW rule. When I was looking at the data i think there wer many many trends like those days with many data below the average line. Now my question is how will i determine my average baseline for the data attached.

Thanks a lot

I do not see the calculation for the average of 35 in the spreadsheet. I get 25 when I calculate the average for the same time range that you used for the standard deviation. So, I would say that 35 is the incorrect value to use for the baseline average.

i used 35 because this number is a cumulative result, these numbers were OEE rating for the day i times it with 100 to get a whole number. then I summed up all the loading hrs for that months, and all downtime for the months to get the availability, then also i summed up all the operating hrs t0 get the theoritical number of units produced and then also the total units produced in order to get the performance, and then i got the total of all good units in order also to have my quality rate. so from these numbers i got 35...can i use 35?...by the way if i use 25 can i use all the data, or will there be data that will be removed? thanks steve..i use your excel worksheet in SPC the one you posted here in this forum.

If the data in your table is some sort of ratio, it is best to track the numerator and denominators separately, summing them up for the baseline average and then dividing.

It looked to me like 35 was pretty far off the apparent center line of the data. I'd suggest trying 25, and then see if you have any trend signals in the data. By the way, the "MW" rule of mine is just to tell if you have enough data for the basis for a baseline. Looks like you have plenty of data. But see how 25 in conjustion with the standard deviation you calculated works, and then subdivide the data if needed.