Design Of Experiments - Tool

Design of experiments

Design of experiments

Design of experiments (DOE) is a valuable tool to:Optimize product and process designs Accelerate the development cycle Reduce development costs Improve the transition of products from research and development to manufacturing Effectively trouble shoot manufacturing problems. Today, Design of Experiments is viewed as a quality technology to achieve product excellence at lowest possible overall cost.

Objectives of Experimentation

■ ■ ■ ■

■ ■

The following are some of the objectives of experimentation in an industry : Improving efficiency or yield Finding optimum process settings Locating sources of variability Correlating process variables with product characteristics Comparing different processes, machines, materials etc Designing new processes and products.

Various terms used in Experimentation

In the context of discussion on experimental designs, the common frequently used terms are : ■ ■ ■ ■ ■ ■

factor Level Treatment combination Response Effect Interaction

Traditional approach

The traditional approach to product and process optimization is to conduct one variable at-a-time experiments. This approach though simple to plan and execute, suffers from several drawbacks. For instance, varying factor ‘A’ from its nominal value ‘A1’ to some other value ‘A2’ may produce a given change in the quality of the product, when other factor ‘B’ is at a value ‘B1’. However a different change in the quality of the product will result, when factor ‘B’ is at a value ‘B2’. This effect known as interaction effect, can not be detected under traditional approach..

Statistically designed experiments

A statistically designed experiment permits simultaneous consideration of all the possible factors that are suspected to have bearing on the quality problem under investigation and as such even if interactions effect exist, a valid evaluation of the main effect can be made. Scanning a large number of variables is one of the ready and simpler objectives that a statistically designed experiment would fulfill in many problem situations.

Statistically designed experiments

Even a limited number of experiments would enable the experimenter to uncover the vital factors as which further trials would yield useful results. The approach has number of merits, it is quick, reliable and efficient.

Basic principles of experimentation

■ ■ ■

Basic principles of Experimentation are : Randomization Replication Local Control

Basic principles of experimentation

■

Randomization : is described as an insurance against extraneous factors.

■

Replication : increases the sensitivity of the experiment i.e power of detecting differences between treatments.

■

Local Control : reduces the effect of natural variability of materials, environmental conditions etc.

Planning for experimentation

It is known widely that a properly well planned experiment helps to achieve better efficiency and hence certain amount of thinking must be done before deciding to carry out the experiment and actually conducting the experiment.

Planning for experimentation

     

The various steps to be followed in this direction are listed below : Selection of area of study : Pareto analysis Proof of the need for experimentation Brain storming and Cause & Effect diagram : To list all the possible factors Classification of factors Interactions to be studied Response and type of model for analysis Note : Pareto Analysis, Brain storming and Cause & Effect diagrams have already been covered in the previous slides.

Proof of the need for experimentation

After having selected the area for experimentation we have to ensure that the problem is of ‘Break through’ or ‘Improvement’ nature and not a problem of ‘control’ nature. For this purpose past data should be suitably analyzed and plotted on some process control chart to check whether the process is within statistical control or not. If the analysis shows lack of control or statistical instability, then it is a problem of ‘control’ nature and experimentation may not be needed.

Proof of the need for experimentation

However if the problem is of chronic nature and there is stability in the process, then it establishes the need for experimentation. Before deciding to carry out experimentation the need for experimentation must be established.

Classification of factors

Tools like brainstorming and cause & effect diagrams helps in identification of factors and preparing a complete list of the factors involved in any experiment. Factors listed can be classified into three categories : 1. Experimental Factors 2. Control Factors 3. Error or Noise Factors

Classification of factors

1. Experimental factors are those which we really experiment with by varying them at various levels. 2. Control Factors are those which are kept at a constant (controlled) level throughout experimentation.

Classification of factors

3. Error or Noise factors are those which can neither be changed at our will nor can be fixed at one particular level. Effect of these factors causes the error component in the experiment and as such these factors are termed as error or noise factors.

Note : At the planning stage itself all the factors viz. Experimental, Control and error should be recognized. This will help to tackle them appropriately during experimentation.

Response and type of model for analysis

■

■

The ultimate observations or data generated by the experiment is known as the response. The response may be : Continuous or measurement type and follows a normal distribution Continuous or measurement type but does not follow normal distribution