I have 3 different factors - DOE help

[marcusja2002]

It's been several years since I conducted a DOE so I'm in need of some help taking the rust off.

I currently have access to excel and google, but no SPC software (Minitab, MATLAB, etc.)

I have 3 different factors
Torque level (1-10), RPMs (350-3500), and the measurements I'll be taking are Actual Torque readings (Ft-lbs).

If I took each option I would have over 1260 different combinations I'd want to take an actual torque value from. Since torque has many factors involved I'd be checking 5-10 bolts per setting (6300, to 12600 bolts).

This is where my brain runs into a wall. While I believe, everything should be linear as the RPMs and torque levels go up so should the actual torque but I'm not 100% sure.

Is there a program or process to help me organize this into a feasible experiment?

Thanks in advance.

 

[Statistical Steven]

Re: DOE help

You only mentioned two factors (Torque level and RPMs). Your response is Actual Torque readings. There are several designs possible depending on your stage of work. If you want to see if your two factors impact your response, then a simple 2-level design with a center point would suffice.

Torque Level 1 and 10
RPMs 350 and 3500
Center Point 5.5 and 1925 (Torque and RPM respectively).

This is 5 runs x 6 Bolts per run would be 30 observations.

If you want to build a model, you might look into Response Surface Methods.

It's been several years since I conducted a DOE so I'm in need of some help taking the rust off.

I currently have access to excel and google, but no SPC software (Minitab, MATLAB, etc.)

I have 3 different factors
Torque level (1-10), RPMs (350-3500), and the measurements I'll be taking are Actual Torque readings (Ft-lbs).

If I took each option I would have over 1260 different combinations I'd want to take an actual torque value from. Since torque has many factors involved I'd be checking 5-10 bolts per setting (6300, to 12600 bolts).

This is where my brain runs into a wall. While I believe, everything should be linear as the RPMs and torque levels go up so should the actual torque but I'm not 100% sure.

Is there a program or process to help me organize this into a feasible experiment?

Thanks in advance.

 

[marcusja2002]

Re: DOE help

I know once I do all of this I'm supposed to be able to produce a formula that if I put in a specific level and rpm then it will pop out a specific torque value. Or I reverse it and if I choose a torque value I want it will tell me the level or rpm.

How do I turn the data into that formula or where can I read up on that step?

 

[JeantheBigone]

Since you only have two factors, maybe this will help?

Most Practical DOE Explained (with Template)