CO2_2016 - page 31

29
Chimica Oggi - Chemistry Today
- vol. 34(2) March/April 2016
The best setting of temperature at 1100 minutes is 530
o
C
which resulted in a yield of just under 60%. Figure 4 is a plot of
temperature against time using a colouring gradient for the
value of yield at each experimental point. It shows the highest
yield of 59.3% at 1100 minutes and 530
o
C. It also indicates a
lot of white space that should be of concern.
This white space indicates we have no data to reliably infer
what yield to expect when:
-
-
Time is less than 1000 minutes and temperature is greater
than 530
o
C
-
-
Time is less than 1000 minutes and temperature is less than
510
o
C
-
-
Time is greater than 1200 minutes and temperature is
greater than 530
o
C
-
-
Time is greater than 1200 minutes and temperature is less
than 510
o
C
which is the reason why many designs are still coded on a
-1 to +1 scale.
-
-
In the 1940’s, Finney whilst at Rothampstead, introduced
the fractional factorial design which allowed many
factors to be investigated at half or even quarter the cost
of Fisher’s factorial design.
-
-
In the 1950’s, George Box and others, deployed DOE in
chemistry and industrialised the application of DOE while
at Imperial Chemical Industry Laboratories.
-
-
Each decade since has seen improvements in the DOE
method until we get to the 2010’s when the Definitive
Screening Design was invented by Brad Jones DOE
Architect for JMP Statistical Discovery from SAS and Chris
Nachtsheim, Carlson School of Management, University
of Minnesota.
-
-
DOE has growing acceptance in Design, Development,
Scale-Up, and Manufacturing areas to ensure key questions
are answered predictably and with lowest cost. It helps
optimize products and processes, and to do so predictably,
reducing risk to your business. With DOE we can transfer
products and processes with a better knowledge of how
they work, reducing the occasions we need to improve or
“fire fight” products or processes when in production, which
makes room for more R&D time to be spent in innovation,
new product and process development.
THE TRADITIONAL EXPERIMENTAL APPROACH
A common alternative to DOE is to experiment by varying one
factor at a time (OFAT). The OFAT approach is illustrated using
two factors – temperature and time – to investigate their effect
on one of the response – yield. Prior knowledge indicates that it
is best to investigate temperature over the range 500 to 550
o
C
and to vary time over the range 500 to 1300 minutes.
The first sequence of experiments were performed by keeping
temperature fixed at 520
o
C and varying time within the
range of 500 to 1300 minutes in increments of 100 minutes. This
resulted in the yield curve in figure 2 and a suggested time of
1100 to 1200 minutes is required to maximise yield.
The next set of experiments were performed by varying
temperature from 500
o
C to 550
o
C in increments of 10
o
C
while keeping time fixed at 1100 minutes as shown in figure 4
- 1100 minutes was favoured to a slightly higher value in order
to keep the cost of running the process lower.
Figure 1.
DOE has played a role in increasing agricultural yields.
Figure 2.
Relationship
between yield
and time at
520
o
C.
Figure 3.
Relationship
between
yield and
temperature at
1100 minutes.
Figure 4.
Time vs Temperature with experimental points coloured
by yield.
1...,21,22,23,24,25,26,27,28,29,30 32,33,34,35,36,37,38,39,40,41,...68
Powered by FlippingBook