Roark's Table 24, 1a - Circular plate of Constant Thickness

[athomas-2]

I am having difficulties with case 1a - 1d in table 24. Things start to fall apart when I enter dimensions for a circular plate that has radii smaller than 1 inch. The deflection at the inner and outer radius can't be calculated.

Am I doing something wrong? Is this a limitation based on the assumptions? i have attached screenshots for your reference.

[Werner_E]

Its hard to debug a picture. Please post your worksheet.

BTW, its a good idea to limit global assignments to an absolute minimum. It looks like you are using them in place of the normal assignment.

[athomas-2]

Sorry about that. Worksheet is attached. I copied and pasted directly from the Roark e-book and then modified the input values.

[Werner_E · Accepted solution]

I don't like the way the range r with non-integer values and unit is used as an index for the array Y. I'd rather turn r into a vector and use an appropriate range running from 0 with step 1 as first matrix index.

But to make as few changes as possible - here is a solution to your problem.

Replace the definition of Y by

or even better by

you divide by 0.1*b because this is the stepwidth of the range r.

[Fred_Kohlhepp]

I agree with Werner that the range definition is the problem.

For Y and S and M (and any place else it occurs) replace (r-b)*100/in with (r-b)*100/b

And the sheet will work.

I have a lot of problems with the Roark e-book and the use of global defines and the use/creation of vectors for solutions and plotting.

The basic definitions are usually right; the implementations are crappy!

[Werner_E]

Fred Kohlhepp wrote:

I agree with Werner that the range definition is the problem.

For Y and S and M (and any place else it occurs) replace (r-b)*100/in with (r-b)*100/b

And the sheet will work.

It may work that way, but you would only use 10% of the matrix Y, setting the remaining 90% to zero.

It should rather be (r-b)*10/b or one of the version suggested by me in my last post.

[PhilipOakley]

The problem is (I believe) in the creation of the "Y" matrix just before the red error.

It's the scaling factor 100/in (a problem that occurs in a number of the Roarks sheets), thus the first (row) index is not a set of integers.

In V15 M030 the error is wrongly reported as being on the 'i' (colum) index, rather than it's row index in the calculation of Y.

Simply rework the scaling so that you get 100 points (or whatever number of sample points you want on the graph) as you span "r".

You may need to redefine 'r' as well if it steps too far past the end of the plate.

Do watch out for the same error style whenever you see a scaling in 'inches' in the Roark's sheets.

Philip

[athomas-2]

Thanks for all of the help. All input was very useful. Everything is working now.