Design of Experiments and Statistical Quality ControlLaajuus (3 ECTS)

Objective

The student masters the basics of statistical design of experiments (DOE), and is able to design efficient series of experiments in research and development projects. The student is able to analyze the results obtained from designed experiments in order to meet the objectives of quality control, optimization or of designing systems.

Content

1. Basic concepts of DOE; model and variable types.

2. The most common designs related to qualitative and quantitative variables.

3. Analyzing experimental results using analysis of variance or regression analysis.

4. Experimental optimization and response surfaces (the Box-Wilson strategy).

5. Visualising the results of statistical analyses by families of curves and response surfaces.

6. Using Excel or statistical software in DOE.

Qualifications

Basics in Engineering Mathematics, Equations and Matrices, Functions and Derivatives, Differential and Integral Calculus, Basic Course in Statistics

Assessment criteria, satisfactory (1)

1. The student can list the basic principles of design of experiments; the student can explain the difference between empirical and mechanistic models, and he or she is able to formulate linear models of several variables.

2. The student is able to create factorial designs, and especially 2N factorial designs.

3. The student is able to carry out multiple linear regression analyses.

4. The student is able to describe the Box-Wilson strategy.

5. The student is able to visualize effects of variables in linear models.

6. The student takes part into the computer labs and carries out the given exercises acceptably.

Assessment criteria, good (3)

1. The student is able to apply the basic principles of DOE in simple practical cases; the student is able to formulate models used in analysis of variance.

2. The student is able to create CC designs and fractional 2N designs.

3. The student is able to choose and carry out analysis variance in a given simple problem.

4. The student is able to conclude when a 2N design requires supplementary experiments.

5. The student is able to create response surfaces.

6. The student is able to use his or her computer skills in practical cases of experimental design.

Assessment criteria, excellent (5)

1. The student is able to classify categorical factors and, based on that, choose on an appropriate empirical model.

2. The student is able to choose a design of an appropriate resolution for a given case.

3. The student is able to assess the reliability of empirical models in several different ways.

4. The student is able to calculate gradients and design new experiments in their direction.

5. The student is able to use graphical visualization techniques in empirical optimization.