Since the 1980s, the ideas of robust design advocated by the Japanese quality consultant Dr. Genichi Taguchi have been gaining acceptance. The principle behind the Taguchi approach is that as an engineering task, it is considerably easier to adjust the mean level of some response variable to a target value than it is to reduce the variation of the response variable. Consequently, Taguchi techniques focus priority on the reduction of variation by paying considerable attention to how the variability of the response changes as the factor levels change.

In recent years many advances have been made in the vast intertwined subjects of experimental design and analysis of variance. New methods are constantly being developed as their need arises in experimentation.

Typical reasons why an experiment may be performed include the following:

- to determine the levels or settings of those factors which can be controlled that give rise to a maximum or minimum response,
- to determine the main causes of variation in a response,
- to compare the responses achieved at different settings of controllable factors.

Observations collected from experiments can be analysed allowing conclusions to be drawn about the underlying model.

This is demonstrated by the following example: consider an experimenter who intends to carry out an experiment to determine the effect of 3 factors A, B and C, say, on a response. Suppose A, B and C each have 2 possible levels such as high and low. One possible experiment to study this situation could be to keep the levels of A and B constant and to take 4 observations with C at a high level and 4 observations with C at a low level. Analysing this data would only give information on the effect of different levels of factor C on the response at the particular combination of levels of factors A and B chosen. Other experiments would then have to be carried out; for each experiment, the set of observations obtained would arise from keeping 2 factors constant and varying the third. A much better experiment would be to take a single observation at each of the 8 possible combinations of levels of factors A, B and C (for example A low, B high, C low is one of the 8 possible combinations). Such an experiment is termed a factorial design and can be analysed to determine the effect of each of the 3 factors A, B and C separately and in addition it can provide information about interaction between any pair of factors and about interaction involving all 3 factors.

When designing or planning an experiment, careful consideration should be given to the subsequent analysis that will be carried out. The efficiency of the analysis will depend upon the particular experimental design that is used to generate the data.

In robust design, factors included in experimentation are divided into two sets:

- control factors – these are factors that are easy and inexpensive to control in the design of the product,
- noise factors – these are factors which may affect the response of interest but which are difficult to control when the product is being manufactured.

A particular combination of levels of the control factors is termed robust if variation in the response is small despite uncontrolled variation in the levels of the noise factors. Orthogonal arrays are employed to construct a fractional factorial design with treatment combinations as combinations of levels of both control and noise factors.

QM&T’s Experimental Design and Taguchi Technique courses are:

- Foundation Course on Experimental Design – One Day
- Advanced Course on Experimental Design- One Day

Each course is of one day duration, which allows the delegate to choose their most appropriate level. Alternatively, just completing the first day would give an excellent insight into the world of Experimental Design. For delegates who already have some knowledge of experimental design, attending the second course would provide an introduction to Taguchi’s methods. Attending the third course would provide an extensive description of Taguchi methods. Post course support has been successfully provided. This is when the delegate has a specific project in mind, where they would like guidance and support from QM&T’s statistical specialist in practically applying the lessons learnt. The specialist is skilled and experienced in the practical application of these techniques. This support includes; establishing the initial scope, design and completion of the experiment, analysis of results and determination of appropriate solution.

1. Foundation Course on Experimental Design

As a consequence of attending this course, delegates will:

- Have a general understanding of the role of experimental design in process improvement;
- Have an awareness of the importance of factor and level selection;
- Understand the importance of choosing appropriate responses;
- Be aware of the reason for randomisation in experimental design;
- Be able to plan, carry out and analyse an experiment involving a single factor; Identify possible blocking factors;
- Be able to plan, carry out and analyse a block design.

Items covered include:

- Experiment planning strategy;
- Response selection;
- Replication and randomisation;
- Experiments involving a single factor;
- Blocking criteria and block designs;
- Analysis of experiments using analysis of variance.

2. Advanced Course on Experimental Design

As a consequence of attending this course, delegates will be able to build on the understanding and skills gained from attending the foundation course. Delegates will be:

- Able to appreciate the necessity for careful factor selection, including selection of the number of levels appropriate for each factor;
- Understand the importance of determining which factor interactions are likely to be important;
- Appreciate the use and advantages of orthogonal arrays;
- Able to plan a full factorial design by the use of an orthogonal array;
- Analyse a full factorial design;
- Plan a fractional factorial experiment and analyse the experiment;
- Work towards a robust design strategy.

Items covered include:

- Overview of planning strategy used in experimental design;
- Choices regarding factors, levels and interactions;
- Full factorial experiments;
- Analysis of full factorial experiments;
- Fractional factorial experiments;
- Confounding;
- Aliasing;
- Analysis of fractional factorial experiments.

The lecturing team is comprised of professionals who have been especially selected for their recognised knowledge and experience in the field of statistics, Experimental design and Taguchi Technique. See QM&T Team specifically QM&Ts Statistical Specialist

See Training Courses for price list and order form

- Quick Links
- - 8D PROBLEM SOLVING
- - 5 WHY'S
- - PROJECT QA
- - PROCESS MAPPING
- - POKA-YOKE
- - APQP
- - MUDA
- - AS 9100 FOUNDATION
- - AS 9100 AUDIT
- - BUSINESS PROCESS ANALYSIS
- - CUSTOMER SATISFACTION
- - FAILURE MODE & EFFECTS
- - FAULT TREE ANALYSIS
- - ISO TS 16949 FOUNDATION
- - ISO TS 16949 AUDIT
- - ISO 9001 INTERNAL AUDIT
- - ISO 20000 FOUNDATION
- - ISO 20000 AUDIT
- - INSPECTION TECHNIQUES
- - KAIZEN
- - MEASURING TECHNIQUES
- - PRINCIPLES & PRACTICES
- - PPAP
- - ROOT CAUSE ANALYSIS
- - SIX SIGMA
- - SOFTWARE QA
- - STATISTICAL PROCESS CONTROL
- - QUALITY PLANNING
- - SUPPLIER QUALITY ASSURANCE
- - TAGUCHI TECHNIQUE
- - VALUE STREAM MAPPING