This module is a part of “Engine Carbon Brushes Replacement” project. The purpose of the module is a classification of the Cement Mill 2 engine start phases.

Trying to clustering the quality of the start phases was concluded, that is important to clusterize whole working periods because the start phase is a part working period and the difference between start phases is determined by the difference between work periods. It is better way to clusterize whole work period using start phase as one of the classifiers.

As classifiers of work periods were taken four parameters:
a) average current;
b) standard deviation of current;
c) duration of start phase;
d) period of time between previous and analyzed work period.

Before running classification algorithms let’s check whether there are any correlations between chosen classifiers.

There is no any correlation between classifiers.

Classification methods

Classification method will be chosen from hierarchical clustering “ward.D”, “ward.D2”, “single”, “complete”, “average”, “mcquitty”, “centroid” methods and K-Mean clustering method.

The best clustering method and the optimal number of clusters are evaluated by Calinski-Harabasz Index.
It is better to take as minimum quantity of groups as possible (no more than 4 or 5) but to see a general picture let’s research from 2 to 20 clusters. Results are shown in the chart.
(R- code is based on functions published on
https://github.com/ethen8181/machine-learning/blob/master/clustering_old/clustering/clustering_functions.R)

According to the chart, the best result brings K-Mean method divides data for 18 groups.
The biggest CH index k=18 of k-Mean line on the “CH-index” chart. Clusterization all observations by 18 clusters is redundant, I intend to divide data no more than 5 clusters. On “wss” chart, the line of K-Meam represents smallest values and dramatic decline of wss changed by a relatively smooth decline on k=5. CHIndex chart gives the best result on k=5 for a k-mean method. For clusterization will be used a k-mean method with k=5.

Let’s represent the CH Index data in the matrix view:

##     k ward.D ward.D2 single complete average mcquitty centroid  kmeans
## 1   2 462.50  574.87 305.83   433.43  305.83   295.91   305.83  643.65
## 2   3 361.71  627.03 331.30   408.75  331.30   331.30   331.30  692.31
## 3   4 440.71  713.47 292.29   431.31  367.20   292.29   287.06  800.10
## 4   5 561.06  804.80 274.56   364.89  370.36   359.35   299.70  993.54
## 5   6 486.12  840.61 259.01   434.65  381.49   428.84   354.88  934.51
## 6   7 599.54  901.08 285.15   373.83  328.42   547.34   305.71  960.88
## 7   8 550.00  967.97 256.09   397.22  346.65   495.52   339.99 1006.17
## 8   9 494.52  968.57 225.17   361.87  401.28   456.19   395.10 1117.70
## 9  10 456.74  981.74 200.35   401.35  440.12   422.18   356.53 1116.19
## 10 11 532.78 1013.29 180.47   662.65  399.47   384.30   322.36 1161.12
## 11 12 631.34 1043.43 170.21   703.74  367.32   354.74   296.38 1218.33
## 12 13 682.27 1080.95 167.82   747.55  338.22   408.80   336.70 1195.10
## 13 14 643.80 1105.06 158.56   697.54  313.67   379.10   311.21 1198.25
## 14 15 616.61 1126.99 156.12   710.89  294.46   353.73   290.28 1277.70
## 15 16 588.02 1149.39 147.76   674.60  280.99   397.81   271.17 1274.09
## 16 17 562.55 1146.15 140.12   644.61  267.56   378.08   254.42 1257.55
## 17 18 541.83 1140.07 132.58   610.45  252.16   360.67   242.27 1347.55
## 18 19 627.62 1140.50 127.30   704.21  238.59   342.07   230.29 1343.81
## 19 20 696.07 1143.12 120.61   678.88  269.75   444.52   222.74 1332.24

Let’s apply K-Mean clustering method with k=5. The algorithm divided all engine runnings into 5 clusters by following way:

Clusters research

By comparison information gain of all classifiers against clusters, we can see relative weight of all of them (classifiers ) in clustering: 

##                    VarName     weight
## 1               avgCurrent 0.85017803
## 2 hoursFromPreviousRunning 0.06018505
## 3               startPhase 0.05858842
## 4               stdCurrent 0.01042305

The most significant influence in clustering has average current, the less significant influence has standard deviation of current.

With the help of radar chart, let’s see distribution of chosen parameters between the groups. Two additional parameters was added to the chart: average and standard deviation of current during the start phase (they are not used as classifiers in clusterization).

The situation somehow reminds situation on CM1: There are three major groups with the only difference in average current.
One of the small group “4” is different from others by a significant standard deviation of the current (both start phase and whole work period), and the long time interval from the previous running.
The group “2” main difference is a significantly long start phase.

Let’s see the average current in the groups compared with the average current of start phases of those groups.

Major groups 1, 3, 5 have a relatively narrow range of current oscillation and the averages current of start phases (white point) are close to general averages.

In opposed, small group 4 has a wide range of the current fluctuation.

Group “2” is kind of in-between type of stable major groups and small unstable group “4”.

Now we have a question:
Maybe some processes or events connected to the mill have had a temporary influence and belong to the specific time period and now have no more importance?

According to the chart, obviously, there is no any connection between division by groups and special time period. All groups represented during the whole analyzed time period.

Actually, cement mill 2 produces only OPC cement.

As already explained, the average current is the main factor of group division. The next chart shows the density of values of current within the clusters.


According to the radar chart, cluster “2” has a different average duration of start phase. Cluster “2” includes the runnings with longest start phases (next chart):

Group “2” includes the runnings with longest start phases, and group “4” includes the most unstable runnings.