This module is a part of “Engine Carbon Brushes Replacement” project. The purposes of the module are:
a) to explain taken criterions of determination of “start phase” of controlled engine work;
b) to compare “start” phases of engines.

According to allegations of electricians who maintain the engines - there are two kinds of engine start periods- sometimes engine starts “hard” and sometimes “smooth”. In addition to visual observations, this allegation has its confirmation on charts of engine current - at the beginng of work, the current oscillation is noticeably different than the current oscillation of “normal” or “controlled” period of the engine work.

In order to find “abnormal” starts was proposed method based on Nelson rules for detection “out-of-control” conditions. For this purposes, were analyzed all 43961 engine starts of 10 mills, from 2010 til 2018. Were analyzed observations of engine current, which was registered every one minute from engine start. The first point is the first case where the current not equal to zero. The point of beginning of “normal” engine work after the start phase must do not match the following three condition simultaneously:
a) Nine points in a row are on the same side of the mean;
b) Six points in a row are continually increasing or decreasing;
c) Eight points in a row not within 1 standard deviation of the currents mean (for this analyzed period of engine work).

Obviously, that this method allows to analyze engine work periods of more than or equal 8 minutes duration only.

Three rules of determination the end of engine start phase:

The enginies data is saved in plant Pi system and was analyzed by written for this case C# program using Pi SDK libraries.

Quantiles presented in the matrix: 
##      0% 25% 50% 75% 100%
## CM1   0   0   7  35 1202
## CM2   0   0   1  16  941
## CM3   0   0   1  13  292
## CM4   0   0  13  33 1875
## CM10  0   0   5  26 1755
## CM11  0   0   4  32 1073
## CM12  0   0   0  15  354
## RMA   0   0   0   0  107
## RMB   0   0   0   0 1124
## RMC   0   0   0   0   31

According to showing data, the biggest percentiles 50 and 75 for the mill 4, 1, and 10. In opposite the cement mills, on the raw mills, the problem is less pronounced.