Scribd
Upload
57 views16 pages

Chemical Engineering Basics Lec

Using two standard deviations instead of three as the cutoff criterion would lower the limiting value of Y to shut down the process from 27 to 25 batches. This stricter criterion would identify process problems sooner, but could also result in more costly shutdowns when large numbers of bad batches simply reflected normal variation in the process.

Uploaded by

jo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
57 views16 pages

Chemical Engineering Basics Lec

Using two standard deviations instead of three as the cutoff criterion would lower the limiting value of Y to shut down the process from 27 to 25 batches. This stricter criterion would identify process problems sooner, but could also result in more costly shutdowns when large numbers of bad batches simply reflected normal variation in the process.

Uploaded by

jo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 16

Chemical Engineering Basics

Lec (10)

T. Alaa Khalid
Sample Variance of Scattered Data
 Consider two sets of measurements of a random
variable, X-for example, the percentage
conversion in the same batch reactor measured
using two different experimental techniques.

 The sample mean of each set is 70%, but the


measured values scatter over a much narrower
range for the first set (from 68% to 73%) than
for the second set (from 52% to 95 %).
 In each case you would estimate the true value of
X for the given experimental conditions as the
sample mean, 70%, but you would clearly have
more confidence in the estimate for Set (a) than
in that for Set (b).
 Three quantities-the range, the sample
variance, and the sample standard
deviation-are used to express the extent to which
values of a random variable scatter about their
mean value.
 The range is simply the difference between the
highest and lowest values of X in the data set:

 In the first plot the range of X is 5% (73% - 68%)


and in the second plot it is 43% (95% - 52%).
 The range is the crudest measure of scatter: it
involves only two of the measured values and gives
no indication of whether or not most of the values
cluster close to the mean or scatter widely around
it.
 The sample variance is a much better measure.
 To define it we calculate the deviation of each
measured value from the sample mean,
 The degree of scatter may also be expressed in
terms of the sample standard deviation, by
definition the square root of the sample variance:
 The more a measured value (Xj ) deviates from the
mean, either positively or negatively, the greater
the value of (Xj - X)2 and hence the greater the
value of the sample variance and sample standard
deviation.

 If these quantities are calculated for the data sets of


Figure above , for example, relatively small values
are obtained for Set (a) (Sx2 = 0.30, Sx = 0.55) and
large values are obtained for Set (b) (Sx2 = 50, Sx =
7.1).
 For typical random variables, roughly two-thirds
of all measured values fall within one standard
deviation of the mean; about 95% fall within two
standard deviations; and about 99% fall within
three standard deviations.

 Of the 37 measured values of X, 27 fall within one


standard deviation of the mean, 33 within two
standard deviations, and 36 within three standard
deviations.
 Values of measured variables are often reported
with error limits, such as X = 48.2 ± 0.6.

 This statement means that a single measured


value of X is likely to fall between 47.6 and 48.8.

 The midpoint of the range (X = 48.2) is almost


always the mean value of the data set used to
generate this result; however, the significance of
the given error limits (±0.6) is not obvious unless
more information is given.
 The interval between 47.6 and 48.8 may represent
the range of the data set (Xmax - Xmin) or ±0.6
might represent ± Sx, ±2 Sx, or ± 3 Sx .

 If you report a variable value in this manner,


make clear what your error limits mean.
Example: Statistical Quality Control
 Five hundred batches of a pigment are produced
each week. In the plant's quality assurance (QA)
program, each batch is subjected to a precise
color analysis test. If a batch does not pass the
test, it is rejected and sent back for
reformulation.
 Let Y be the number of bad batches produced
per week, and suppose that QA test results for a
12-week base period are as follows:

 The company policy is to regard the process


operation as normal as long as the number of
bad batches produced in a week is no more than
three standard deviations above the mean value
for the base period (i.e., as long as Y ≤ Y +
︠ 3sY ).
 If Y exceeds this value, the process is shut down
for remedial maintenance (a long and costly
procedure).
 Such large deviations from the mean might occur
as part of the normal scatter of the process, but so
infrequently that if it happens the existence of an
abnormal problem in the process is considered the
more likely explanation.
1. How many bad batches in a week would it
take to shut down the process?
2. What would be the limiting value of Y if two
standard deviations instead of three were used
as the cutoff criterion? What would be the
advantage and disadvantage of using this
stricter criterion?
The maximum allowed value of Y
is
 If this criterion were used, 26 bad batches in a
week would be enough to shut down the process.
 The advantage is that: if something has gone
wrong with the process the problem will be
corrected sooner and fewer bad batches will be
made in the long run.
 The disadvantage is that : more costly
shutdowns may take place when nothing is
wrong, the large number of bad batches simply
reflecting normal scatter in 'the process.

You might also like