We are planning to conduct full factorial experiments. I found in most of the literature that the analysis associated with Full Factorial experiments is ANOVA based on coefficients.
I wan to know can we use Signal-to-Noise ratio analysis for the data obtained through Full factorial experiment.
Dear Sir
When you use factorial experiments the aim is to determin the effect of each factor, the most important effects, the interractions effects to use them to find a linear model
So this factors are calculated using mathematic formula i don't thing that using S/N analysis will be useful
Using SN will cause that you will lose degrees of freedom of an experiment, what can lead to incorrect conclusions.
e.g. if you are conducting full factorial experiment 33 (three inputs on three levels) it means that you will have 27 combinations of inputs levels. If you will have replication e.g. 2 (two measurements per combination) it will give you 54 DOF (degrees of freedom). But if you will use SN as an output then replications do not count and you have only 27 DOF. So it will be probably better to use mean value of measurements as an output (54 DOF).
It's possible. But as per my opinion, S/N ratio can be well suited for Taguchi method only. It is better to use ANOVA technique for your factorial design based conducted experiments. You can directly use the data to obtain the most influencing parameter. For more detail, use the book "Applied design of experiments and taguchi method" by k.krishnaiah and sahabudden. It is well written book for this issue
@Dr. Thangaraj Muthuramalingam
Is the book "Applied design of experiments and taguchi method" by k.krishnaiah and sahabudden readily available in the market?
If not, where do i get it?
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Material Characterization