RSM Part-2 Central Composite Design with Illutrated Application Example

Dear Friends, We are happy to share this second video on Response Surface Design. In this video, Hemant Urdhwareshe, Fellow of ASQ and Six Sigma Master Black Belt explains the structure of Response Surface Design with focus on Central Composite Design (CCD). Hemant has taken lot of efforts to illustrate this little complex concept with lot of graphics and animations so that viewers can understand it clearly and quickly. Hemant has also illustrated how to create, analyse and interpret results of Central Composite Design on Minitab with a virtual catapult. We recommend that viewers should watch our previous videos:
(1) RSM Part-1 • Response Surface Desig...
(2) Fractional Factorial Designs and Resolution Code: • DOE-5: Fractional Fact...

[@Sandeep-zw3gs]

Great Explanation & Great knowledge given . Thankyou.

[@instituteofqualityandrelia7902]

You are most welcome!

[@abhishekpariyar884]

Well explained sir, thank you

[@instituteofqualityandrelia7902]

Welcome! Appreciate your feedback!

[@cyrilcnwakaji]

Thank you sir

[@uhemant1]

Welcome!

[@jeejin2686]

Good video sir, I have a question, if P value lack of fit less than 0.05 but r square more than 90, So I can use this model to predict or not

[@instituteofqualityandrelia7902]

Good question! If p-value of LOF is low, the model is not fitting well. You can try transformations such as box cox. If you still get a similar p-value, you will need to decide whether some more factors need to be added, or change levels or use the model with caution.

[@jeejin2686]

@@instituteofqualityandrelia7902 thank you very much sir

[@amirsafari7140]

hi, I'm working on laser-assisted turning and currently experimenting with five parameters: depth of cut, feed rate, spindle speed, laser power, and laser angle (the distance between the laser spot and cutting edge). Initially, I opted for a face-centered central composite design. However, I've noticed that when the laser power is set to 0 (indicating the laser is off), the laser angle becomes irrelevant. I'm uncertain about the appropriate experimental design in this situation. Could you please suggest a method or name for this scenario, considering the disregard for laser angle when laser power is zero? Thank you for your insights!

[@instituteofqualityandrelia7902]

Apologize for my late response. I need to know more details. Please write to me on [e-mail]

[@VenkataSivaRaVadlamani]

Thanks for your explanation! I have a limitation of having only 20 experimental runs for 3 factors design in Minitab. Since my 3 factors CCD needs 29 runs, can you suggest any adjustments or alternatives I can use with this constraint in mind? Your insights would be much appreciated!

[@instituteofqualityandrelia7902]

Thank you for your keen interest! Consider Box Behnken Design. For details see www.itl.nist.gov/div898/handbook/pri/section3/pri3362.htm. See if this design fits on your factor levels considering the constraints.

[@VenkataSivaRaVadlamani]

@@instituteofqualityandrelia7902 I apologize if my previous question was unclear. To rephrase, I need to perform a Central Composite Design (CCD) with 3 factors using 29 experimental runs in Minitab. However, the standard CCD for 3 factors would only involve 20 experiments. I was advised to use 29 experiments for greater precision. My current plan is to consider an additional 9 center points, making it a total of 29 runs Could you please advise how to create response surface design for 29 runs in Minitab? Your guidance would be greatly appreciated.

[@alfredafedzi7557]

Thank you for the video. I have a question: 1. The example you illustrated was 2 factors, two levels using CCD, and it contains only low and high? How can I use the CCD for 2 factors, 3 levels? or what design would you recommend for 2 factors, 3 level experiment to find optimal condition? .

[@uhemant1]

Hello Alfred, RSM already has five levels. Corner points, center point, and axial points. And there is a relationship between these. If you have only three levels, and two factors, RSM is not possible (If I understand your question).

[@alfredafedzi7557]

@Hemant Urdhwareshe ok. But the video showed was for only 2 levels, why was that possible and 3 levels cannot?

[@uhemant1]

The video shows two levels and a center point which makes it three levels.

[@alfredafedzi7557]

@@uhemant1 So is it right to say it's two factors and 3 level design?

[@haymarnhnin-gt9ww]

If I have 2 Numeric factors ( 3 levels ) and 1 Categoric factor (only 2 levels), can I use Face centered central composite design? Could you please give me a suggestion?

[@instituteofqualityandrelia7902]

The face centred design will also have five levels. Two for corner points, 1 centre point, and two for face centred axial points. You can add categorical factor levels which means design will be replicated as per levels of the categorical factor.

[@divyang089]

for 3 factorial number of center points should not be 9? what you are showing 6 center points.

[@instituteofqualityandrelia7902]

Number of center points in a 3-factor RSM (CCD) will be 6. Not 9.