Hi,
I'm having trouble finding a symbolic solution I can use in an excel sheet for later use.
I am aware that it's mabye impossible to find a symbolic solution.
But I was hoping it was, And that you guys could help me?
All the values is >0
Erik
I have attached the file now. And I came to think of it, that alle the values er >0.
Yeah it's 25deg.
I need to put the expression for c.uk into an Excel sheet, because i'm working on an general excel program. The simple program shall be used by my colleagues to determining the soil strength based on a given foundation size, force, depth and some of the other variables in the formula. Which is given based on the foundations size (the s.i factors).
It's looking something like this (in Danish).
So all the variables have a known value depending on the foundations size, depth etc. So manually I could check the result for each 0,1 meters, but that will take for ever..
I hope that makes sense?
First, let's observe that this is an unnecessarily complex equation.
Define a variable
X =
which can then become
s = (A+B+C)*X or X = s/(A+B+C)
So now you have a much simpler equation to solve. If the RHS is all known quantities this becomes some basic high school algebra!
There's no symbolic solution, the order of c.ud in your equation is too high:
Try to numerically solve the equation, using the root() function in Prime.
Success!
Luc
@LucMeekes wrote:
There's no symbolic solution, the order of c.ud in your equation is too high:
Try to numerically solve the equation, using the root() function in Prime.
Success!
Luc
I agree that the final equation cannot be solved symbolically by Prime - looks like we would end up at an equation of order five.
Unfortunately using "root" in Prime seems not to be an option as the goal is to provide an Excel sheet for users not using Prime.
The Excel solver sure could do the job but I fear its not an option, too, because its not an automatic one click operation (at least as far as i know) and might be too much for an average Excel user. Not sure if the Excel solver could be made a one-click experience using some scripting.
So one remaining option is to implement/program a numerical solve algorithm in Excel - I guess there might be some ready made toolboxes available, but I can't recommend anything here as I only seldom use Excel and then only the features it provides out of the box.
Another option could be to use an approximation (Taylor series, second or third order) and its exact solution.
I have not tried and it may be that the expressions given by the symbolic "solve" are so large that Prime would refuse to display them.
Furthermore you would need to heavily test the results in Prime against the "exact" ones provided by the "root" function using a lot of typical test values to see, if the quality of the approximation is good enough at all.
