Best method to test simulated data?

zjwright — Fri, 09 Jun 2023 01:00:47 GMT

Hello,

I am working to validate if a washer is needed on an assembly. We have a screw which will be threading through 2 surfaces. We want to validate if a washer is needed to prevent the screw from bottoming out. We have dimensions & spec limits for all 3 components.

I've started in Excel by generating 10,000 random samples for each of the components, but other than checking if screw length is less than the sum of the 2 surfaces, and plotting yes or no, there's not much I can do in Excel.

Using JMP what would be a good method to test this? Any/all advice is appreciated

For reference, I have JMP Pro 17

Re: Best method to test simulated data?

statman — Tue, 18 Apr 2023 16:49:12 GMT

First, welcome to the community.

I have a few questions for you before I can suggest analysis. Also the more context provided the better the advice (e.g., what is the assemble intended to do?)

1. Where are you getting the data for the 10,000 random samples? Have you studied the measurement systems? If you are simply using specs, that is pretty much worthless.

2. Do you know the consistency of the screw dimensions? How have you studied this? If so, have you calculated a mean and standard deviation?

3. Do you know if washer thickness is consistent? How have you studied this? If so, do you have mean and standard deviation?

4. Do you know if the depth of the hole is consistent? It appears that the depth is a function of two surfaces. Are those thicknesses consistent...How have you studied this? If so, do you have mean and standard deviation?

5. Is there a torque that is required for the screw? Is the torque consistent? How have you studied this? If so, do you have mean and standard deviation?

How analysis proceeds is a function of what are you trying to do and what questions you are trying to answer? Do you want to understand the probability associated with Bottoming out? Or the probability of the screw not having enough torque on it? Are you trying to design an assembly that doesn't require a washer?

topic Best method to test simulated data? in Discussions

Best method to test simulated data?

Re: Best method to test simulated data?