PUBLISHED BY

World's largest Science,

Technology & Medicine
Open Access book publisher

108,000+
3,300+ INTERNATIONAL 114+ MILLION
OPEN ACCESS BOOKS AUTHORS AND EDITORS DOWNLOADS

AUTHORS AMONG
BOOKS TOP 1% 12.2%
DELIVERED TO AUTHORS AND EDITORS
MOST CITED SCIENTIST FROM TOP 500 UNIVERSITIES
151 COUNTRIES

Selection of our books indexed in the

Book Citation Index in Web of Science™
Core Collection (BKCI)

Chapter from the book Statistical Approaches With Emphasis on Design of

Experiments Applied to Chemical Processes
Downloaded from: http://www.intechopen.com/books/statistical-approaches-with-
emphasis-on-design-of-experiments-applied-to-chemical-processes

Interested in publishing with IntechOpen?

Contact us at [email protected]
 DOI: 10.5772/intechopen.69501

Provisional chapter
Chapter 9

Application
Application of
of Taguchi-Based
Taguchi-Based Design
Design of
of Experiments
Experiments
for
for Industrial
Industrial Chemical
Chemical Processes
Processes

Rahul Davis and Pretesh John

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/intechopen.69501

Abstract

Design of experiment is the method, which is used at a very large scale to study the
experimentations of industrial processes. It is a statically approach where we develop
the mathematical models through experimental trial runs to predict the possible output
on the basis of the given input data or parameters. The aim of this chapter is to stimulate
the engineering community to apply Taguchi technique to experimentation, the design
of experiments, and to tackle quality problems in industrial chemical processes that they
deal with. Based on years of research and applications, Dr. G. Taguchi has standard-
ized the methods for each of these DOE application steps. Thus, DOE using Taguchi
approach has become a much more attractive tool to practicing engineers and scientists.
And since the last four decades, there were limitations when conventional experimental
design techniques were applied to industrial experimentation. And Taguchi, also known
as orthogonal array design, adds a new dimension to conventional experimental design.
Taguchi method is a broadly accepted method of DOE, which has proven in producing
high-quality products at subsequently low cost.

Keywords: DOE, Taguchi method, industrial chemical processes, parameter

optimization, ANOVA

1. Introduction

Industries are engaged in a variety of activities such as developing new products, improv-
ing previous designs, maintenance, controlling and improving ongoing processes and some
more. Experimentation is a frequent task in these activities to measure and analyse the output,
and for this purpose engineers/researchers use many tools like statistics, analytical models,
etc., regardless of their background in it [1].

© 2018 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons
©Attribution License (http://creativecommons.org/licenses/by/3.0),
2016 The Author(s). which
Licensee InTech. This chapter is distributed under permits
the terms ofunrestricted
the Creative use,
Commons
distribution,
Attribution and reproduction
License in any medium, provided the original
(http://creativecommons.org/licenses/by/3.0), work
which is properly
permits cited. use, distribution,
unrestricted
and reproduction in any medium, provided the original work is properly cited.
138 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

Montgomery [2] writes,

“Experiments are performed in almost any field of enquiry and are used to study the performance of pro-
cesses and systems. The process is a combination of machines, methods, people and other resources that
transforms some input into an output that has one or more observable responses. Some of the process
variables are controllable, whereas other variables are uncontrollable, although they may be controllable
for the purpose of a test. The objectives of the experiment include: determining which variables are most
influential on the response, determining where to set the influential controllable variables so that the
response is almost always near the desired optimal value, so that the variability in the response is small,
so that the effect of uncontrollable variables are minimized.”

In today’s era, the purpose of experiments in industries is essentially optimization and robust
design analysis (RDA, which is used to make the system less sensitive to variations in uncon-
trollable noise factors or in other words to make the system robust). DOE, or experimental
design, is the name given to the techniques used for guiding the choice of the experiments
to be performed in an efficient way. In a general way, the process analysis can be expressed
as the study of the cause-effect relationships which may be carried out by drawing infer-
ences from a finite number of samples. And one of the most important purposes of it is to
design sampling experiments that are productive and cost-effective and provide a sufficient
data base in a qualitative sense [3]. Design of experiments has been applied successfully in
diverse fields such as agriculture (improved crop yields have created grain surpluses), the
petrochemical industry (for highly efficient oil refineries) and Japanese automobile manufac-
turing (giving them a large market share for their vehicles), and still its implementation area
is spreading and providing the optimized results. These developments are due in part to the
successful implementation of design of experiments. The reason to use design of experiments
is to implement valid and efficient experiments that will produce quantitative results and
support sound decision-making [4].

1.1. Brief history

Statistical experimental design, together with the basic ideas underlying DOE, was born in
the 1920s from the work of Sir Ronald Aylmer Fisher [5]. Fisher was the statistician who cre-
ated the foundations for modern statistical science. The second era for statistical experimental
design began in 1951 with the work of Box and Wilson [6], who applied the idea to indus-
trial experiments and developed the response surface methodology (RSM), which is used to
find out the relationships between various process parameters and one or more responses.
The work of Dr. Genichi Taguchi in the 1980s [7], despite having been very controversial
(described briefly in heading 2.4), had a significant impact in making statistical experimental
design popular and stressed the importance it can have in terms of quality improvement.

Usually, data subject to experimental error (noise) are involved, and the results can be sig-
nificantly affected by noise. Thus, it is better to analyse the data with appropriate statistical
methods. The basic principles of statistical methods in experimental design are replication,
randomization and blocking. Replication is the repetition of the experiment to obtain a more
precise result (sample mean value) and to estimate the experimental error (sample standard
deviation). Randomization refers to the random order in which the runs of the experiment
are to be performed. In this way, the conditions in one run neither depend on the conditions
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 139
http://dx.doi.org/10.5772/intechopen.69501

of the previous run nor predict the conditions in the subsequent runs. Blocking aims at isolat-
ing a known systematic bias effect and prevents it from obscuring the main effects [8]. This
is achieved by arranging the experiments in groups that are like one another. In this way, the
sources of variability are reduced, and the precision is improved.

The design of experiments (DOE) is explained by Lye [9], as a methodology for systematically
applying statistics to experimentation. In DOE, a sequence of tests is designed in which pur-
poseful vary the input parameters (factors) of a product or process to examine the reasons of
the variation in the output response [10]. By the end of the twentieth century, DOE was no lon-
ger viewed as merely a stand-alone tool, because it was packaged together with a structured
initiative for business improvement known as Six Sigma. Moreover, an increased emphasis on
DOE took place during this period in Six Sigma literature [11]. DOE is a good tool to under-
stand and optimize products or process parameters. It is quick as well as cost-effective.

1.2. Advantages of DOE

With real engineering examples, Czitrom [12] listed the following advantages of DOE:

• A good amount of data can be obtained with lesser resources (experiments, time, material,
etc.).

• The estimates of the effect of each factor (variable) on the response are more precise.

• It is a systematic way to estimate the interactions between the process factors.

• There is experimental information in a larger region of the factor space.

1.3. DOE techniques

A survey was carried out within the industry which identifies the needs of using an efficient
and practical technique for the experimentation. It was surveyed that 76% of industries con-
sider themselves in need of a methodology [13]. So here are listing some of the techniques
that are in use in Industries. The list of the techniques considered is far from being complete
since the aim of the section is just to introduce the reader into the topic showing the main
techniques which are used in practice [14].

• Randomized complete block design

• Latin square

• Full factorial
• Fractional factorial

• Central composite

• Box-Behnken [15]

• Plackett-Burman [16]

• Taguchi [7]
140 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

• Random

• Halton, Faure and Sobol sequences

• Latin hypercube

• Optimal design

• Response surface design

Several DOE techniques are available to the experimental designer. However, as it always
happens in optimization, there is no best choice. The correct DOE technique selection depends
on the problem to be investigated and on the aim of the experimentation.

M. Cavzzuti [14] concluded that items to be considered are:

a. The number of experiments N which can be afforded. In determining the number of

experiments, an important issue is the time required for a single experiment. There is a
lot of difference between whether the response variable is extracted from a quick simula-
tion in which a number is computed or taken from a spreadsheet or it involves the set-
ting up of a complex laboratory experiment. In the former case, it could take a fraction
of a second to obtain a response, and in the latter one, each experiment could take days.

b. The number of parameters k of the experiment. For many DOE techniques, the number of
experiments required grows exponentially with the number of parameters (Figure 1). Not

Figure 1. Number of experiments required by the DOE techniques [14].

 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 141
http://dx.doi.org/10.5772/intechopen.69501

necessarily to use a cheap technique is the best choice, because a cheap technique means
imprecise results and insufficient design space exploration. Unless the number of experi-
ments which can be afforded is high, it is important to limit the number of parameters as
much as possible to reduce the size of the problem and the effort required to solve it. Of
course, the choice of the parameters to be discarded can be a particularly delicate issue.
This could have done by applying a cheap technique (like Plackett-Burman etc.) as a pre-
liminary study for estimating the main effects.

2. Introduction to Taguchi method

2.1. Brief history

After the Second World War, allied forces observed some of the major drawbacks of the Japanese
telephone system, that is, extremely poor quality and unsuitability for long-term communica-
tion purposes. To overcome these drawbacks, an improved system was required, for this the
allied command recommended establishing research facilities to develop a state-of-the-art com-
munication system. At that time, the electrical communication laboratories (ECL) were came
on the stage with Dr. Genichi Taguchi (Figure 2) in charge of improving the R&D productivity
and enhancing product quality. It was observed that the ratio of the time and money expended
on engineering experimentation and testing is very high than the efforts given to the process of
creative brainstorming to minimize the expenditure of resources. He noticed that the process
of inspection, screening and salvaging cannot improve poor quality. The inspection process is
done to check the quality but it can’t increase the quality by itself. Therefore, he believed that
quality concepts should be based upon, and developed around, the philosophy of prevention.

This moved Taguchi to develop new optimizing methods of the processes of engineering
experimentation. He believed that the best way to improve quality was to design and build it

Figure 2. Dr. G. Taguchi [17].

142 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

into the product. He quoted that [18], “Cost is more important than quality but quality is the best
way to reduce cost.”
He developed the techniques which are now known as Taguchi methods (TM). His main
contribution lies not in the mathematical formulation of the design of experiments, but rather
in the accompanying philosophy. Taguchi method is different from the traditional techniques
because of Taguchi’s concepts of design. By his methods, he developed Robust Manufacturing
systems that are insensitive to daily or seasonal variations of environment, wears or other
noise factors. His philosophy had far reaching consequences, yet it is founded on three very
simple concepts [19].

Taguchi’s new technique consist of three concepts about quality, these are:

1. Quality should be designed into the product and not inspected into it.

2. Quality is better achieved by minimizing the deviation from a target. The product should
be so designed that it is immune to uncontrollable environmental factors.

3. The cost quality should be measured as a function of deviation from the standard and the
losses should be measured system-wide.

In Taguchi’s thinking, the quality improvement is an ongoing effort. He endeavoured con-

tinually to reduce the variation around the target value. Right population selection as near
to the target value or desired value is the first stair step of the quality improvement. And to
accomplish this, Taguchi designed experiments using especially constructed tables known as
“orthogonal arrays” (OA). It makes the design of experiments very easy and consistent.

Taguchi’s two most important contributions to quality engineering are as follows:

i) The use of Gauss’s quadratic loss function to quantify quality.

ii) The development of robust designs (parameter and tolerance design).

Since the early 1980s when applications to different industries began in western hemisphere,
the Taguchi method is evaluated in different platforms like books, articles, panel discus-
sions, etc. Taguchi methods have touched most of the manufacturing processes to optimize
the process in such a way so that noise factors do not affect the output. Several reports [20–
29] evaluated Taguchi methods from a statistical standpoint. In these reports, the parameter
design received the most attention. These reports confirm that Dr. Taguchi made important
contributions to quality engineering; however, without some basic statistical knowledge, it
is hard to apply his technique. Specifically, the use of signal-to-noise ratios in identifying
the nearly best factor levels to minimize quality losses may not be efficient [30].

2.2. Taguchi method

In Taguchi method, we assume that we are designing an engineering system—it might be a

machine that performs some intended function or it might be a production process. We use the
knowledge of fundamental about the system and process parameters for efficient ­experimentation.
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 143
http://dx.doi.org/10.5772/intechopen.69501

We can skip all the extra effort that might have gone in to investigating interactions. By this,
we can decrease the number of factors. Taguchi categorize the factors in two sets:

1. Control factors, which are under our control.

2. Noise factors, which are not under our control, except during experiments in the laboratories.

In the 1920, Sir R. A. Fisher first proposed the DOE with multiple factors known as Factorial
Design of Experiments. In the full factorial design, we work on the all possible combina-
tions extruded from the preselected set of factors. Mostly all industrial experiments depends
upon on the number of factors and if we consider all possible combination then it becomes
harder and also time consuming to execute a large sequence of experiments. The full frac-
tional design consists of kn experiments, where “k” is the number of levels of factors and “n”
is the number of factors (factors are the different variables, which determine the functionality
or performance of a product or system). After that, partial factorial method came into the exis-
tence. In this, the number of experiments is reduced, and only a small set from all possibilities
is selected, which produces the most information.

Taguchi’s approach complements the following two important areas:

• Taguchi constructed a special set of orthogonal arrays (OA) to lay out his experiments.
Taguchi prepared a new set of standard OAs which could be used for a number of experi-
mental situations.

• He also devised a standard method for analysis of the results.

As mentioned, the full factorial design requires many experiments to be carried out. It
becomes laborious and complex, if the number of factors increases. To overcome this prob-
lem, Taguchi suggested a specially designed method called the use of orthogonal array. In
this, we can study the more factors or parameter space with lesser number of experiments.
Opposite to full factorial analysis, the Taguchi method reduces the number of experimental
runs to a reasonable one, in terms of cost and time, using orthogonal arrays [31]. For example,
if there are three factors called A, B and C, all are examined with two levels called “1” and
“2” (in general, they are referred as “1” and “−1”). Then, according to the full fractional, the
number of experiments should be 23 = 8. In Table 1, the full fractional array is shown; such
experiments can find all main and two and three factor interactions.

But at the same time, Taguchi’s L8 array can deal with seven factors and their two levels as
shown in Table 2, such experiments can find all seven main factors effects.

So, it can be seen that the full fractional method is dealing with three factors, and orthogonal
array is dealing with seven factors in same eight experimental runs.

Taguchi method is based on mixed levels, highly fractional factorial designs and other orthog-
onal designs. It distinguishes between control variables and noise variables. In this, we choose
two sets of parameters, that is, controlled and noise parameters or variables. Respectively, we
choose orthogonal designs also. The design chosen for the controlled variable is known as
inner array and for the noise variables is known as outer array. The combination of the inner
144 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

A B C AB BC AC ABC
1 1 1 1 1 1 1

2 1 1 2 1 2 2

1 2 1 2 2 1 2

2 2 1 1 2 2 1

1 1 2 1 2 2 2

2 1 2 2 2 1 1

1 2 2 2 1 2 1

2 2 2 1 1 1 2

Table 1. Full fractional factor assignments to experimental array columns.

A B C D E F G
1 1 1 1 1 1 1

2 1 1 2 1 2 2

1 2 1 2 2 1 2

2 2 1 1 2 2 1

1 1 2 1 2 2 2

2 1 2 2 2 1 1

1 2 2 2 1 2 1

2 2 2 1 1 1 2

Table 2. Orthogonal array factor assignments to experimental array columns.

and the outer arrays gives the crossed array, which is the list of all the samples scheduled by
the Taguchi method. It means that for each sample in the inner array, the full set of experi-
ments of the outer array is performed. The advantage of this cross combination is that it pro-
vides the information about the relation between the parameters which plays very important
for the robust system designing.

Then, it is recommended by Dr. Taguchi to use the quality loss function to measure the per-
formance characteristics. The quality loss function is a continuous function that is defined in
terms of the deviation of a design parameter from an ideal or target value. The value of this
loss function is further transformed into signal-to-noise (S/N) ratio. Performance characteris-
tics are available in three categories to determine the S/N ratio:
• Nominal-the-Best

• Larger-the-Best
• Smaller-the-Best
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 145
http://dx.doi.org/10.5772/intechopen.69501

The Taguchi method has four basic phases in the optimization process, these are as follows:

1. First phase is to timely thinking about the quality characteristics and determining the pa-
rameters which important to the product or process.

2. In second phase the experiments sequence is designed and executed accordingly.

3. In third phase of the optimization process the statistical analysis is done to determine the
optimum conditions.

4. Finally in the fourth phase the confirmation test is run with the optimum conditions.

Barrado et al. [32] expanded above-mentioned four phases into the following steps for imple-
menting Taguchi experimental design:

Step 1. Selection of the output or target parameters.

Step 2. Identification of the input parameters and their levels.

Step 3. Determining the suitable orthogonal array (OA).

Step 4. Assign factors and interactions to the columns of the array.

Step 5. Conduct the experiments.

Step 6. Statistical analysis and the signal-to-noise ratio and determining the optimum setting
of the factor levels.

Step 7. Perform confirmatory experiment (if necessary).

First of all, the input parameters and the respective operation levels and output or responses
are selected. After this choice, the best fitted or economical matrix experiment or orthogonal
array is selected [33, 34]. Here, in Table 3, the array selector is shown.

After performing the experiments as per the chosen array, we choose the desired signal-to-
noise ratio function (smaller-the-better, larger-the-better and nominal-the-better). The S/N
ratio is a logarithmic function which can also be defined as an inverse of variance. Generally
in the optimization of the process or product design and in minimizing the variability we
use it. If we maximize the S/N ratio, we reduce the variability of the process against undesir-
able changes in noise factors. Because S/N ration and variance are inversely proportional,
so the chosen factors should produce maximum value of S/N so that we get the minimum
variability.

Three types of common problems and respective signal-to-ratio function are presented in Table 4.

In Table 4, yi denotes the nth observations of response variable and μ2 denotes the square of
mean and σ2 the variance of the observations of the replicated response values.

For analysing the data, one of the most common methods used is ANOVA. ANOVA is a sta-
tistical technique that assesses potential differences in a scale-level dependent variable by a
nominal-level variable having two or more categories [35]. The ANOVA is developed by Sir
146 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

Numbers of Parameters (P) Number of levels

2 3 4 5

2 L4 L9 L16 L25

3 L4 L9 L16 L25

4 L8 L9 L16 L25

5 L8 L18 L16 L25

6 L8 L18 L32 L25

7 L8 L18 L32 L50

8 L12 L18 L32 L50

9 L12 L18 L32 L50

10 L12 L27 L32 L50

11 L12 L27 L50

12 L16 L27 L50

13 L16 L27

14 L16 L36

15 L16 L36

16 L32 L36

17 L32 L36

18 L32 L36

19 L32 L36

20 L32 L36

21 L32 L36

22 L32 L36

23 L32 L36

24 L32

25 L32

26 L32

27 L32

28 L32

29 L32

30 L32

31 L32

Table 3. Taguchi orthogonal array selector.

 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 147
http://dx.doi.org/10.5772/intechopen.69501

Choose… S/N ratio formulas Use when the goal is to…

Smaller the better ​S / N = − 10 log​[​ n_ ​  ​∑  ​   ​yi​ ​ ​ 2​]​

1 n
Minimize the response
i=1

Nominal the better ​μ​ 2​ Target the response and you want to base the S/N
​S / N = 10 log ​__
 ​σ​ 2 ​​ 
ratio on means and standard deviations

​S / N = − 10 log​[_ ​ ​y1​ ​  ​ 2 ​​]  ​
n
Larger the better ​ n1 ​ ​∑ ​​   _ Maximize the response
i=1 i

Table 4. Types of problems and respective signal-to-ratio function.

Ronald Fisher in 1918. It is the extended version of T test and Z test. The reason of its origin
was the limitations of the T & Z tests which have the problem of only allowing the nominal
level variable to have just two categories. This analysis method is also famous with the title
“The Fisher Analysis of Variance.”

The use of ANOVA depends on the research design. Commonly, ANOVAs are used in three
ways:

a. One-way ANOVA

b. Two-way ANOVA
c. N-way multivariate ANOVA.

2.3. Applications of Taguchi method in industrial chemical processes

The Taguchi method is used whenever the settings of interest parameters are necessary, not
only for manufacturing processes. Therefore, the Taguchi approach is used in many domains
such as environmental sciences [36, 37], agricultural sciences [38], physics [39], statistics [40],
management and business [41], medicine [42] and in chemical processes as well [43]. Here,
some of the literature reviews are given, which clearly show the application of Taguchi method
in the chemical processes of various industries.

The identification and incorporation of quality costs and robustness criteria are becoming a
critical issue while addressing chemical process design problems under uncertainty. Fernando
P. Bernardo et al. [44] used Taguchi loss functions along with other robustness criteria and
presented a systematic design framework. They conducted their study within a single-level
stochastic optimization formulation, and an efficient cubature technique is used for the estima-
tion of the expected values. An optimal design was discovered, together with a robust operat-
ing policy. It was observed that these parameters maximizes average process performance.
148 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

Kundu et al. [45] investigated the optimal operation conditions to prepare activated carbon (AC)
using palm kernel shell (PKS). They choose four control factors for their research which were irra-
diation time, microwave power, concentration of impregnation substance which was phosphoric
acid and impregnation ratio between acid and PKS aided by Taguchi optimization method. After
successful implementation of Taguchi method, the optimal settings were found and the optimal
levels were microwave power of 800 W, irradiation time of 17 min, impregnation ratio of 2 and
concentration of acid 85% (undiluted). After the confirmation test with optimal settings, acti-
vated carbon with high BET surface area of 1473.55 m2 g−1 and high porosity was obtained.
Zolfaghari et al. [46] presented a systematic optimization approach by using the Taguchi
method for removal of lead (Pb) and mercury (Hg) by a nanostructure, zinc oxide-modified
mesoporous carbon CMK-3 denoted as Zn-OCMK-3. CMK-3 was synthesized by using SBA-
15 and then oxidized by nitric acid. Their investigation using Zn-OCMK-3 in the frame of
the Taguchi method brought forth the optimum conditions for Pb and Hg adsorption from
water. The determined optimum conditions for removal of Pb (II) and Hg (II) were the agi-
tation time of 120 min, the initial concentration of 10 mg/l, the temperature of 35 ◦C, the
dose of 0.7 g/l, and the pH of 6. Based on the confirmation test under optimum conditions
for Zn-OCMK-3, it was observed that this nanoporous carbon is very effective in removing
the lead and mercury with a high pollutant removal efficiency (PRE) i.e. 97.25% for Pb (II)
and 99% for Hg (II). Removal of lead and mercury were highly concentration dependent.
Number of Pb and Hg ions highly increase from initial concentration of 10–400 mg/l. It is
also observed that a lot of ions cannot be adsorbed at high concentration, which reduce the
removal efficiency.

Venkata Mohan et al. [47] applied design of experimental methodology (DOE) by Taguchi
approach to find out the effects of selected factors on the H2 production with the final aim of
optimizing the process. They selected four factors for their research study, that is, inlet pH,
inoculum type, feed composition and inoculum pre-treatment methods. Here, also Taguchi
method enhanced the process. Results showed significant variation and process efficiency.
Among the factors studied with respect to H2 production, feed composition showed stronger
influence followed by inlet pH, pre-treatment method and origin of the inoculum. Taguchi
approach also permits process optimization of system by a set of independent factors (over a
specific region of interest (levels) by identifying the influence of individual factors, establish-
ing the relationship between variables and operational conditions and finally estimate the
performance at optimum levels obtained. By using optimized conditions of the factors, the
rate of H2 production can be enhanced by almost threefold (0.376–1.166 mmol/day), same
positive enhanced results were recorded for substrate degradation also.

Taguchi-based DOE methodology provides a systematic and efficient mathematical approach

to understand complex process of fermentative H2 production and substrate degradation
for the optimization of the near optimum design parameters, only with a few well-defined
experimental sets [48].

Messias Borges Silva et al. [49] applied Taguchi method in optimization of the treatment con-
ditions of polyester-resin effluent by means of Advanced Oxidative Processes (AOPs). The
output parameter were Chemical oxygen demand (COD). Their study consist of two type of
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 149
http://dx.doi.org/10.5772/intechopen.69501

evaluation. First one is, the influence individual independent parameter and the second one
is the respective interactions effect. This allowed determining the most statistically significant
variables and the optimal conditions. The treatment by means of advanced oxidative process
provided an approximate 35% reduction in chemical oxygen demand of the polyester-indus-
try wastewater. However, when compared to studies describing the treatment of this effluent
by Advanced Oxidative Processes, it was seen that the results were relevant. This investiga-
tion confirms the effectiveness of Taguchi Method in the industrial chemical processes.
Liao et al. [50] applied the Taguchi method and designs of experiments (DOE) approach to
optimize parameters for chemical mechanical polishing (CMP) processes in wafer manufac-
turing. Planning of experiments was based on a Taguchi orthogonal array table to determine
an optimal setting, and significant results were found.
S. V. Mohan et al. [51] applied the Taguchi robust experimental design (DOE) methodology on
a dynamic anaerobic process treating complex wastewater by an anaerobic sequencing batch
biofilm reactor (AnSBBR). Their work was to optimize the process and also at the other hand
to evaluate the influence of distinct factors on the process. They considered the uncontrollable
(noise) factors. This is the first kind of study of anaerobic process evaluation and process opti-
mization by using the Taguchi methodology adopting dynamic approach.

The biological oxidation of ferrous ion by iron-oxidizing bacteria is potentially a useful indus-
trial process for removal of H2S from industrial gases, desulphurization of coal, removal of
sulfur dioxide from flue gas, treatment of acid mine drainage and regeneration of an oxidant
agent in hydrometallurgical leaching operations that pH of feed solution has the most contribu-
tion in the biooxidation rate of ferrous ion. When the parameters were set according to Taguchi
approach, the biological reaction rate was obtained. Mousavi et al. [52] studied and find out the
optimum values of the process parameters on the ferrous biooxidation rate by immobilization
of a native Sulfobacillus species on the surface of low-density polyethylene (LDPE) particles in
a packed-bed bioreactor using Taguchi method. L16 array was used with five factors and their
four levels. Temperature, initial pH of feed solution, dilution rate and initial concentration
of Fe3+ and aeration rate are considered as input parameters in Taguchi technique. Analysis
of variance (ANOVA) was used to determine the optimum conditions and most significant
process parameters affecting the reaction rate. Results indicated that pH of feed solution has
the most contribution in the biooxidation rate of ferrous ion. When the parameters were set
according to Taguchi approach, the biological reaction rate was obtained 8.4 g L−1 h−1.
Taguchi robust design method with L9 orthogonal array was implemented to optimize experi-
mental conditions for the preparation of nano-sized silver particles using chemical reduction
method [53]. The parameters for chemical-mechanical polishing (CMP) in an ultra-large-scale
integrated (ULSI) planarization process are explored using the Taguchi method [54]. So, it can
be seen easily the popularity of the Taguchi method in industrial chemical processes.

2.4. Advantages and disadvantages of Taguchi method

The foundation of the DOE in Taguchi method (TM) is orthogonal array design that is very
simple method for analysing the outputs. His work was controversial and criticized by some
150 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

researchers for being inefficient and ineffective in some cases [55]. But still the simplicity of
Taguchi method increased its use in manufacturing industries. Here, some of the advantages
and disadvantages are discussed briefly which clears its popularity even in the clouds of con-
troversies. A survey shows that 51% of respondents are familiar with TM, although only 14%
of them apply it [13].

The key step of the TM is to increasing the quality level with less affecting the cost factor.
TM provides the optimal settings for the processes which can improve quality, and these set-
tings attained from TM are also insensitive to the variation of noise factors. Basically, classical
process parameter design is complex and not easy-to-use, and at the other hand, Taguchi
technique is user friendly [56].
Another advantage of the Taguchi method is that it emphasizes a mean performance charac-
teristic value close to the target value rather than a value within certain specification limits,
thus improving the product quality. Also, it is straightforward and easy to apply to many
engineering situations, this property makes it a powerful yet simple tool for industries. It can
be used to quickly narrow the scope of a research project or to identify problems in a manu-
facturing process from data already in existence [57].
It is probably unfortunate that the important concepts advocated by Taguchi have been
overshadowed by controversy associated with his approach to modelling and data analy-
sis. There have been several research papers and books which explain, review, or raise
questions on the Taguchi’s ideas [55, 58]. One of its demerit is that the results obtained from
it are only relative. It is unable to indicate exactly which parameters have the highest effect
on the performance or response.

Another debating topic is that when orthogonal array do not tests all the possible combina-
tions of the factors, then this method should not be adopted with all relationships between all
factors. The Taguchi method has been criticized for its difficulty in accounting for interactions
between parameters.
Another limitation is that the Taguchi methods are offline. So in the case of dynamically
changing processes for example a simulation study, the TM is not appropriate. And also, the
Taguchi method is most effectively applied at the designing stage of a product development,
it cannot help after the process is started. Because TM deals with designing quality rather than
correcting for poor quality [59].

3. Conclusion

Industries are in the need of the method of conducting experiments which optimize processes
and increase the quality of products. Same is desired in chemical industrial processes also. For
this, design of experiments is the basic step of quality improvement via optimized processes.
And this requires proper planning and layout of the experiments and accurate analysis of
the results. And Dr. G. Taguchi studied these issues very well and developed his method.
Thus, DOE using Taguchi approach has become a much more attractive tool to practicing
engineers and scientists. Since the conventional experimental design techniques were applied
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 151
http://dx.doi.org/10.5772/intechopen.69501

to industrial applications or processes, these conventional methods were always covered with
drawbacks and limitations. And Taguchi array, also known as orthogonal array design, adds
a new dimension to conventional experimental design. Taguchi method is a broadly accepted
method of DOE which has proven in producing high-quality products at subsequently low
cost. In most of the industrial applications or processes, the researchers and scientists used
Taguchi method with other analytical tools in their research works, and in industrial chemical
processes, it is also showing great results in optimization of the processes. A very important
and fundamental part of Taguchi’s method is to make sure that the product performs well
even in noise; it helps in making the product long lasting. Taguchi’s method can be applied in
a very short amount of time; it does not take a lot of effort and improves the processes dramat-
ically. The manufacturing industries can improve their processes very quickly and efficiently
by applying the Taguchi’s method. Around Industrial chemical processes, Taguchi method is
showing the outstanding performance by optimizing the process parameters and reducing
the number of experiments via orthogonal arrays. Taguchi method gives a new height to the
processes by making them cost-effective and quick with improved quality.

Author details

Rahul Davis* and Pretesh John

*Address all correspondence to: [email protected]

Department of Mechanical Engineering, Shepherd Institution of Engineering and Technology,

SHUATS, Allahabad, India

References

[1] Bisgaard S. Teaching statistics to engineers. The American Statistician. 1991;45(4):274-

283. DOI: 10.2307/2684452
[2] Montgomery DC. Design and Analysis of Experiments. 5th ed. New York: Wiley; 2000
[3] Toutenburg H, Shalabh. Statistical Analysis of Designed Experiments. 3rd ed. Heidelberg:
Springer Texts in Statistics; 2009. DOI 10.1007/978-1-4419-1148-3_1
[4] Telford JK. A brief introduction to design of experiments. Johns Hopkins APL Technical
Digest. 2009;27(3):224-232
[5] Fisher RA. Statistical Methods for Research Workers. Edinburgh: Oliver and Boyd; 1925
[6] Box GEP, Wilson KB. Experimental attainment of optimum conditions. Journal of the
Royal Statistical Society. 1951;13:1-45
[7] Taguchi G, Wu Y. Introduction to Off-line Quality Control. Nagoya: Central Japan
Quality Control Association; 1980
[8] NIST/SEMATECH. NIST/SEMATECH e-handbook of statistical methods. 2006. Retrieved
from http://www.itl.nist.gov/div898/handbook/
152 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

[9] Lye LM. Tools and toys for teaching design of experiments methodology. In: 33rd
Annual General Conference of the Canadian Society for Civil Engineering. Toronto,
Ontario, Canada. 2005
[10] Montgomery DC. Design and Analysis of Experiments. 6th ed. John Wiley & Sons, Inc;
Hoboken, NJ. 2005
[11] Brady JE, Allen TT. Six Sigma literature: A review and agenda for future research.
Quality and Reliability Engineering International. 2006;22:335-367
[12] Czitrom V. One factor at a time versus designed experiments. The American Statistician.
1991;53(2):126-131
[13] Tanco M, Viles E, Ilzarbe L, Álvarez MJ. Manufacturing Industries Need Design of Experi­
ments (DOE). In Proceedings of the World Congress on Engineering. 2007;2:1108-1112.
London: Newswood Limited. ISBN:978-988-98671-2-6
[14] Cavazzuti M. Design of Experiments. In Optimization Methods: From Theory to Design.
Springer-Verlag Berlin Heidelberg. 2013, pp. 13-42. DOI:10.1007/978-3-642-31187-1_2
[15] Box GEP, Behnken D. Some new three level designs for the study of quantitative vari-
ables. Technometrics. 1960;2:455-475
[16] Plackett RL, Burman JP. The design of optimum multifactorial experiments. Biometrika.
1946;33(4):305-325
[17] http://wiki.mbalib.com/w/images/5/5a/%E7%94%B0%E5%8F%A3%E7%8E%84%E
4%B8%80.jpg
[18] AZquotes. Taguchi’s quote. Retrieved from http://www.azquotes.com/quote/641638.
Accessed on January 2017
[19] Ranjit R. A Primer on the Taguchi Method (Competitive Manufacturing Series), Van
Nostrand Reinhold. ISBN-10: 0442237294, ISBN-13: 978-0442237295
[20] Box GEP, Fung CA. Minimizing Transmitted Variation by Parameter Design. Center
for Quality and Productivity Improvement, University of Wisconsin-Madison, (Report
no. 8). 1986
[21] Box GEP, Bisgaard S, Fung C. An explanation and critique of Taguchi’s contributions to
quality engineering. Quality and Reliability Engineering International. 1988;4(2):123-131
[22] Box GEP, Jones S. Designing products that are robust to the environment. Total Quality
Management. March 1992;3(3):265-282
[23] Bisgaard S. Quality Engineering and Taguchi Methods: A Perspective. The Center for
Quality and Productivity Improvement of University of Wisconsin, Madison (CQPI
Report no. 40). 1990
[24] Czitrom V. An Application of Taguchi’s Methods Reconsidered. ASA Meeting,
Washington, DC. 1946
[25] Bisgaard S, Diamond NT. An Analysis of Taguchi’s Method of Confirmatory Trials.
CQPI Reports (Report no. 60). 1990
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 153
http://dx.doi.org/10.5772/intechopen.69501

[26] Bisgaard S. Process Optimization- going Beyond Taguchi Methods. CQPI Reports (Report
no. 70). 1991
[27] Bisgaard S. A comparative analyses of the performance of Taguchi’s linear graphs. CQPI
Reports (Report no. 82). 1992
[28] Bisgaard S, Ankenman B. Analytic Parameter Design. CQPI Reports (Report no. 103). 1993
[29] Steinberg D, Burnsztyn D. Noise Factors, Dispersion Effects and Robust Design. CQPI
Reports (Report no. 107). 1946
[30] Maghsoodloo S, Ozdemir G, Jordan V, Huang C H. Strengths and limitations of Taguchi’s
contributions to quality, manufacturing, and process engineering. Journal of Manufacturing
Systems. 2004;23(2):73-126. DOI: 10.1016/S0278-6125(05)00004-X
[31] Taguchi G, Jugulum R, Taguchi S. Computer-based Robust Engineering: Essentials for
DFSS. Milwoukee, WI: ASQ Quality Press; 2004
[32] Barrado E, Vega M, Grande P, Del Valle JL. Optimization of a purification method for
metal-containing wastewater by use of a Taguchi experimental design. Water Research.
1996;30(10):2309-2314. DOI: 10.1016/0043-1354(96)00119-4
[33] Padke MS. Quality Engineering Using Robust Design. Englewood Cliffs: Prentice Hall; 1989
[34] Taguchi G, Konishi S. Taguchi Methods Orthogonal Arrays and Linear Graphs: Tools for
Quality Engineering. Dearborn, Michigan: American Supplier Institute; 1987
[35] Statistics Solutions. ANOVA [WWW Document]. 2013. Retrieved from http://www.statis-
ticssolutions.com/academic-solutions/resources/directory-of-statistical-analyses/anova/
[36] Daneshvar N, Khataee AR, Rasoulifard MH, Pourhassan M. Biodegradation of dye solu-
tion containing Malachite green: Optimization of effective parameters using Taguchi
method. Journal of Hazardous Materials. 2007;143(1-2):214-219
[37] Du Plessis BJ, De Villiers GH. The application of the Taguchi method in the evaluation of
mechanical flotation in waste activated sludge thickening. Resources, Conservation and
Recycling. 2007;50(2):202-210. DOI: 10.1016/j.resconrec.2006.06.014
[38] Tasirin SM, Kamarudin SK, Ghani JA, Lee KF. Optimization of drying parameters of
bird's eye chilli in a fluidized bed dryer. Journal of Food Engineering. 2007;80(2):695-700
[39] Wu CH, Chen WS. Injection moulding and injection compression moulding of three-
beam grating of DVD pickup lens. Sensors and Actuators A: Physical. 2006;125(2):367-375
[40] Villafranca RR, Zúnica L, Zúnica RR. Ds-optimal experimental plans for robust param-
eter design. Journal of Statistical Planning and Inference. 2007;137(4):1488-1495
[41] Elshennawy AK. Quality in the new age and the body of knowledge for quality engi-
neers. Total Quality Management and Business Excellence. 2004;15(5-6):603-614
[42] Ng EYK, Ng WK. Parametric study of the biopotential equation for breast tumour iden-
tification using ANOVA and Taguchi method. Medical and Biological Engineering and
Computing. 2006;44(1-2):131-139
154 Statistical Approaches With Emphasis on Design of Experiments Applied to Chemical Processes

[43] Houng JY, Liao JH, Wu JY, Shen SC, Hsu HF. Enhancement of asymmetric bio reduction
of ethyl 4-chloro acetoacetate by the design of composition of culture medium and reac-
tion conditions. Process Biochemistry. 2007;42(1):1-7

[44] Bernardo FP, Pistikopoulos EN, Saraiva PM. Quality costs and robustness ­criteria in chem-
ical process design optimization. Computers and Chemical Engineering. 2001;25:27-4

[45] Kundu A, Gupta BS, Hashim MA, Redzwan G. Taguchi optimization approach for pro-
duction of activated carbon from phosphoric acid impregnated palm kernel shell by
microwave heating. Journal of Cleaner Production. 2015;105:420-427. DOI: 10.1016/j.
jclepro.2014.06.093

[46] Zolfaghari G, Sari AE, Anbia M, Younesi H, Amirmahmoodi S, Nazari AG. Taguchi opti-
mization approach for Pb(II) and Hg(II) removal from aqueous solutions using modified
mesoporous carbon. Journal of Hazardous Materials. 2011;192:1046-1055

[47] Mohan SV, Raghavulu SV, Mohanakrishna G, Srikanth S, Sarma PN. Optimization and
evaluation of fermentative hydrogen production and wastewater treatment processes
using data enveloping analysis (DEA) and Taguchi design of experimental (DOE) meth-
odology. International journal of hydrogen energy. 2009;34:216-226. DOI: 10.1016/j.
ijhydene.2008.09.044

[48] Mohan SV, Reddy BP, Sarma PN. Ex-situ slurry phase bioremediation of chrysene
contaminated soil with the function of metabolic function: process evaluation by data
enveloping analysis (DEA) and Taguchi design of experimental methodology (DOE).
Bioresource Technology. 2009;100(1):164-72. DOI: 10.1016/j.biortech.2008.06.020

[49] Silva MB, Carneiro LM, Silva JPA, Oliveira IS, Filho HJI, Almeida CRO. An application of
the Taguchi method (robust design) to environmental engineering: Evaluating advanced
oxidative processes in polyester-resin wastewater treatment. American Journal of
Analytical Chemistry. 2014;5:828-837. DOI: 10.4236/ajac.2014.513092

[50] Liao HT, Shie JR, Yang YK. Applications of Taguchi and design of experiments methods in
optimization of chemical mechanical polishing process parameters. International Journal
of Advanced Manufacturing Technology. 2008;38:674-682. DOI: 10.1007/s00170-007-1124-7

[51] Mohan SV, Rao NC, Prasad KK, Muralikrishna P, Rao RS, Sarma PN. Anaerobic treat-
ment of complex chemical wastewater in a sequencing batch biofilm reactor: Process
optimization and evaluation of factors interaction using the Taguchi dynamic DOE
methodology. Biotechnology and Bioengineering. 2005;90(6):732-745

[52] Mousavi SM, Yaghmaei S, Jafari A, Vossoughi M, Ghobadi Z. Optimization of ferrous bio-
oxidation rate in a packed bed bioreactor using Taguchi approach. Chemical Engineering
and Processing: Process Intensification. 2007;46(10):935-940. DOI: 10.1016/j.cep.2007.06.010

[53] Kim KD, Han DN, Kim HT. Optimization of experimental conditions based on the
Taguchi robust design for the formation of nano-sized silver particles by chemical
reduction method. Chemical Engineering Journal. 2004;104(1-3):55-61. DOI: 10.1016/j.
cej.2004.08.003
 Application of Taguchi-Based Design of Experiments for Industrial Chemical Processes 155
http://dx.doi.org/10.5772/intechopen.69501

[54] Ho CY, Lin ZC. Analysis and application of grey relation and ANOVA in chemical–­
mechanical polishing process parameters. International Journal of Advanced
Manufacturing Technology. 2003;21(1):10-14. DOI: 10.1007/s001700300001

[55] Nair VN. Taguchi's parameter design: A panel discussion. Technometrics. 1992;31(2):
127-161

[56] Montgomery DC. Design and Analysis of Experiments. Singapore: Wiley. 1991

[57] Fraley S, Oom M, Terrien B, Date JZ. Design of Experiments via Taguchi Methods:
Orthogonal Arrays: The Michigan Chemical Process Dynamic and Controls Open Text
Book. USA. 2006. 678-698

[58] Parks J. On stochastic optimization: Taguchi methods demystified, its limitations and
fallacy clarified. Probabilistic Engineering Mechanics. 2006;16(1):87-101

[59] Unitek Miyachi Group. Welding material control. Technical Application Brief. 1999;2:1-5