I never heard of design of experiments ! until yesterday my professor told me to employ it in my major project. so i did experimentation for all the possible combinations i.e 27 cases But according to taguchi only 9 will suffice so 18 cases are useless ? i cannot put those 18 cases? if i use taguchisir?
Hallo, Taguchi is a robust design method that can be used to save time and cut down on experimental costs. If running full factorial tests is possible, easy, and inexpensive (in this case you have already done all the tests) you should use the actual test results. Because Taguchi runs fewer tests and predicts (not facts) the best scenario. Taguchi is also not recommended if your test results are very sensitive and required very high accuracy. Another way you can use all test results is to verify Taguchi, develop your Taguchi test matrix for your case, pick the test results, and find an optimum solution. Next, see if your Taguchi result agrees with the full test outcome.
Hi , What about a combination where there are let's say 3 parameters but 2 of them get 3 levels and the third has just 2 levels. How should we deal with such combination ? How can we reduce the number of experiments? Thanaks
Maya, this is called mixed-level design, one way is, you can use a Dummy level for the third one making it balanced; the Dummey level could be a repetition of any of the existing two levels (you can tag it as 1" (this means you are repeating the level 1 as dummy level 3) or 2". Choose whichever (between 1st or 2nd level) you think would be more representative. Due to this, the design will be orthogonal but will not be balanced, you need to be careful about variance.
@@MohammadShafinulHaque Hi Mohammad , thank you for your fast response.Hope you are well these days. Is there any other way except dummy level ? What application Can i fill in my parameters & levels and get as output the minimal recomanded experiments ? Is it avaliable for free ? Thanks again Maya
@@mayamosseri2164 Hi Maya, if you need it once or twice you can use the trial version of Minitab or XLSTAT, which gives you 30 days and 14 days free use.
Check the video for around 17.58 minutes, and distribute your parameter and levels as shown in the table (it does not matter which one goes where once you create the table stick to that format), then reduce the test number to four based on the figure and table below.
Very useful video. Very effectively presented Taguchi method for DOE . Keep it up 👍👍. Can you suggest good videos for learning Confirmation Test & Followup test Plz ?
Hello Mohammad, First of all, where are you from? very good English pronunciation you have!! 1.Can you share that presentation? 2. Do you have more videos about Taguchi on your channel? I searched by without sucess...lol kkk
A short answer is that is how Taguchi approach selects experiment minimizing the number of tests, see the rule. And you can find the Taguchi experiment tables in the web.
Now Taguchi is crystal clear for me ( at least for low levels), Thank you very much Dear Prof. Mohammad.
You're welcome Unal Sevim
Dear prof Mohammad, thank you for sharing this presentation with us. Very useful!
You are welcome. I am glad that it was helpful.
Thank you
You're welcome
Excellent work sir
Thank you Rana
NICELY EXPLAINED..THAN YOU..
You are welcome.
You explained very nicely
Thank you Raspreet. I am glad that it was helpful 😊
I never heard of design of experiments ! until yesterday my professor told me to employ it in my major project.
so i did experimentation for all the possible combinations i.e 27 cases
But according to taguchi only 9 will suffice
so 18 cases are useless ?
i cannot put those 18 cases? if i use taguchisir?
Hallo, Taguchi is a robust design method that can be used to save time and cut down on experimental costs. If running full factorial tests is possible, easy, and inexpensive (in this case you have already done all the tests) you should use the actual test results. Because Taguchi runs fewer tests and predicts (not facts) the best scenario. Taguchi is also not recommended if your test results are very sensitive and required very high accuracy. Another way you can use all test results is to verify Taguchi, develop your Taguchi test matrix for your case, pick the test results, and find an optimum solution. Next, see if your Taguchi result agrees with the full test outcome.
Hi , What about a combination where there are let's say 3 parameters but 2 of them get 3 levels and the third has just 2 levels. How should we deal with such combination ? How can we reduce the number of experiments?
Thanaks
Maya, this is called mixed-level design, one way is, you can use a Dummy level for the third one making it balanced; the Dummey level could be a repetition of any of the existing two levels (you can tag it as 1" (this means you are repeating the level 1 as dummy level 3) or 2". Choose whichever (between 1st or 2nd level) you think would be more representative. Due to this, the design will be orthogonal but will not be balanced, you need to be careful about variance.
@@MohammadShafinulHaque Hi Mohammad , thank you for your fast response.Hope you are well these days.
Is there any other way except dummy level ?
What application Can i fill in my parameters & levels and get as output the minimal recomanded experiments ?
Is it avaliable for free ?
Thanks again
Maya
@@mayamosseri2164 Hi Maya, if you need it once or twice you can use the trial version of Minitab or XLSTAT, which gives you 30 days and 14 days free use.
hello, thank you for your helpful video. I want to ask how you selected the four(4) tests out of the eight(8). How is the selection done? Thank you
Check the video for around 17.58 minutes, and distribute your parameter and levels as shown in the table (it does not matter which one goes where once you create the table stick to that format), then reduce the test number to four based on the figure and table below.
Thank you sir
Really helpful
You’re welcome
Thank you so much
You're most welcome
Very useful video. Very effectively presented Taguchi method for DOE . Keep it up 👍👍. Can you suggest good videos for learning Confirmation Test & Followup test Plz ?
Hello Mohammad,
First of all, where are you from?
very good English pronunciation you have!!
1.Can you share that presentation?
2. Do you have more videos about Taguchi on your channel?
I searched by without sucess...lol kkk
Hi Tiago, thank you! I don't have any more video for Taguchi right now. May be in future 🙂
Nice explanation
Thank you, Rajesh.
Amazin explanation! thank you , question how do you go from 8 experiments to 4? I thin ki will need to google it.... Thanks anyway
A short answer is that is how Taguchi approach selects experiment minimizing the number of tests, see the rule. And you can find the Taguchi experiment tables in the web.
Gotta watch this kc ito ung Naka toka sakin na report 😰
Thank you
You talk too much, speed it up!
You can speed up from TH-cam play options