Continuous
Data – Equipment Error
In case we have continuous data, we have separate tests to
check for Equipment Error/Variation
and Operator Error/Variation.
Equipment
Error
To check for
the Error in the measurement contributed by the Equipment, we conduct “Test Retest”.
In Test-Retest we check the Equipment for
its Accuracy and Precision.
Accuracy
The concept of Accuracy says that the Data Points are close
to the Target.
To check Accuracy, we calculate the BIAS of the Equipment (i.e. the difference/distance of the Mean from the Target/True Value).
BIAS = TARGET – AVERAGE
Ideally, the BIAS in the Measurement System (Equipment) should be ZERO/Close to ZERO.
Understanding Accuracy or precision is not easy when
everything is theory. Advance Innovation understands this and that’s why the
Six Sigma training in Delhi is full of examples. To make the understanding
clear of Accuracy we have quoted an example below.
Example: In a
Filling Plant of300 ml. Pepsi Bottles,
the Quality Engineer wants to check if the machine filling Pepsi in bottles is
accurate or not, for that, he takes 20 bottles as Sample and manually measures
the volume of Pepsi in each bottle. The data is as below:
Bottle No.
|
Quantity (ml)
|
1
|
310
|
2
|
305
|
3
|
298
|
4
|
299
|
5
|
295
|
6
|
307
|
7
|
308
|
8
|
305
|
9
|
299
|
10
|
298
|
11
|
309
|
12
|
300
|
13
|
308
|
14
|
306
|
15
|
305
|
16
|
297
|
17
|
299
|
18
|
298
|
19
|
305
|
20
|
309
|
Using the data, he calculates the
Average volume of Pepsi per bottle that he founds to be 303 ml.
The Target volume per bottle was 300 ml.
Now, BIAS = Target – Average
BIAS = 300 – 303
BIAS = -3 ml
As, we can see, the measurement
system/equipment is Biased, due to this biasness, the organization is losing 3
ml. Pepsi per bottle, here, the organization will need to decide whether they
can live with it or not.
Precision
The concept of Precision says that the Data Points are close
to Each other.
To check Precision, we calculate
the Standard Deviation of the
measurement result and compare it with the Equipment tolerance (which is
already present with the equipment, provided by the manufacturer).
In precision, we check if the Thumb Rule is validated or not.
RULE:
“If Standard
Deviation< (1/10)th of the equipment tolerance, then the equipment is said to be Precise.”
Example: Lets say, in the above example,
the QC also calculates the Std Dev. Of the measurement result, which he founds
to be 4.83 ml.
The
Tolerance of the Equipment was +/- 2 ml. which makes the Tolerance as 4 ml.
(Tolerance = Upper Limit – Lower Limit)
So, (1/10)th
of the Tolerance will be .4 ml.
Hence, the
Std. Dev. (4.83 ml.) is more than 1/10th of the tolerance and thus
we can conclude that the Measurement System/Equipment is Not Precise.
Precision
NOTE: An Ideal Data is the data that’s
both Accurate as well as Precise, so we try to ensure that the data points are
close to the target and the variation between the data points is minimum.