Dear ChE 272 students,

Here's a suggestion regarding problem 3.5, in response to a student question.

> Hey professor Greenfield ... I have a question regarding problem 3.5 in > the homework due this thursday. We did it on Matlab and I was wondering > what you wanted from us. Whether you wanted us to write down all the > error answers we got from the program, whether you wanted us to print > out what our matlab code gave us ( all the errors for the 20 terms) or > print out the code itself. If we print out all the errors we solved for > using matlab it equals out to 8 pages worth of errors if we get a > separate error for each series with terms from 1 to 20.

A raw series of numbers isn't so useful for showing how big the errors are. Instead, I would suggest a graph that shows the error (y axis) as a function of the number of terms (x axis). This should let you show all that information in a much more economical way. Showing both curves on the same plot is even more informative.

In all of the assignments for 272, focus on how you can provide your answers in ways that suggest meaning the most easily.

****************

Also, there were questions regarding the last problem. The idea in the last problem is to determine a formula for the relative error

d(y_i P) / (y_i P)

as a function of the uncertainty in mole fraction, d (x_i), and temperature, d T.

We did a problem similar to this in class on Thursday 1/25. You start with the formula for y_i P, take the ln of each side, and then calculate the differential. If you aren't sure about what the differential means, take the derivative with respect to a new variable (call it r, for example), assuming that y_i P, x_i, and T all depend on r. Then cross off each dr at the end.

Prof. Greenfield

Return to E-mail Archive

Return to CHE 272 home page