Design of Experiments (DoE) simply explained
ฝัง
- เผยแพร่เมื่อ 17 พ.ค. 2024
- In this video, we discuss what Design of Experiments (DoE) is. We go through the most important process steps in a DoE project and discuss how a DoE helps you to reduce the number of experiments. We then discuss how you can estimate the number of experiments needed and we go through the most common experimental designs: Full factorial design, Fractional factorial design, Plackett-Burman Design, Box-Behnken Design, Central Composite Design.
► DoE Calculator
datatab.net/statistics-calcul...
► EBOOK
datatab.net/statistics-book
0:00 What is design of experiments?
3:12 Steps of DOE project
5:56 Types of Designs
6:26 Why design of experiments and why do you need statistics?
6:47 How are the number of experiments in a DoE estimated?
9:26 How can DoE reduce the number of runs?
10:09 What is a full factorial design?
12:04 What is a fractional factorial design?
15:27 What is the resolution of a fractional factorial design?
21:54 What is a Plackett-Burman design?
22:46 What is a Box-Behnken design?
24:00 What is a Central Composite Design?
24:34 Creating a DoE online
Back in the days I had to pay a full fee just to attend a DOE classes however, I didn't get to understand a simple thing. The reason is that, they couldn't address the DOE principle as simple as this.
Words can't express my sincere gratitude for you at DataLab.
Keep it up guys, sharing knowledge is caring for everyone.
❤❤❤❤❤❤❤❤
Hi, many many thanks for your nice feedback!!!! Yes, of course we will continue : ) Again thanks and Regards, Hannah and Mathias
Please is there a way you can put me through sir, I need it for my research but I don't understand it at all.
Yeh I agree. I sat in a DOE masters class for three days - this short video did it in 20 minutes!!
@@timwatson9413 Thanks : )
Mam,am from India (Tamilnadu -chennai) super explanation
Many many thanks : )
Excellent explanation with practical example
Many thanks : )
Highly appreciated, how in easy steps DOE explained.
Many thanks : )
Thank you!
You're welcome! Thanks for your feedback!
Hello, thank you very much for this wonderful video.
I have a question, for the equation that is used to estimate the number of runs needed that depends on standard deviation and the effect that is relevant to us).
Where do I get the standard deviation? Do I need to make a random number of runs first and then determine the standard deviation then use it in the equation?
Incredible video with such amazing clarity! Could you please also make some videos about screening and optimization, please?
Many thanks for your nice feedback! I will put it on my To Do List!
This is so helpful and useful for my research
Thanks you : )
10.04 16 effects for lubrication and 16 runs for temperature.... this makes it a total of 32 runs, but explained as 24 runs... could you please help to clarify..??
👍
Thanks : )
Nice video! I was wondering: At 9:26, N= 2,4 so you would do 2 +2 runs, but later N = 16 so you do 8 + 8 runs? What is the difference between both instances?
Hello, thank you very much for your feedback! This is because you have to round up, so 2.4 would be 3 and you can't divide 3 by 2, so you need 4 attempts in total! Regards Hannah
@@datatab Ah I understand now! Thank you!
@@lianne199 You are welcome : )
Hi there! incredible content here.. but i do have a question regarding case example at 19:55, i noted that there is a third factor, C, which was not discussed when introducing response analysis to determine if there are any interaction between A and B. How can we then interpret if there is an interaction of C with A and C with B to the response variable?
Hi man ythanks for you comment! Oh, I'm sorry if we have explained this in a misleading way! Of course the response must be measured taking C into account!
Nice graphic
Many thanks : )
Man I got lost half way in the video. Try re-watching with not luck.
In German we say: "aller guten Dinge sind drei" : ) Regards Hannah
@@datatab i like that! in english we say "third time's the charm"
@@cvspvr : )
Mistake at 10:00, a total of 32 runs would be required, not 24 runs.
I think the point was that 3 different variations will be brought, i.e., first keeping oil constant, 8 runs will be tested at low temperature, 8 runs at high temperature. This way the temperature effect is monitored. Then to evaluate the lubrication effect, temperature was kept at low and only lubricant was changed i.e., from oil to grease.
What about lubricant & high temperature?
@@nda9921 that is only for Full Factorial Design case where all possible interactions are evaluated.