So, I was looking at the valve performance for my selection at different flow rates. How I got into this was that I was working on a control valve selection for a condenser water system that has a number of different operating flow rates. But in figuring out how to work around it, I learned some things that will probably be useful, so I thought I would share them. Most of you probably already realize this, and when I noticed the issue, I sort of had a hunch about the reason for it. And for some applications, the digits that were dropped could make the difference between making an accurate prediction from your data and one that was not so good, especially if you multiply them by numbers that have big exponents. In other words, the coefficients presented in the equation are correct, but rounded off. But, in some cases, especially with high power polynomials, your predictions could be way off if you did that because of the compounding of rounding errors. It could be pretty tempting to write a formula that used the trend line equation and assume it was correct. you need to be careful if you use the equation to predict data, especially with higher order polynomials. and have Excel put the equation for the line on the graph …. … experiment with the options to find something that is a reasonable fit …. The bottom line is that if you use Excel’s trend line feature to apply a trend line to a set of data in a graph …. I realized something the other day while doing a curve fit in Excel that I figured was worth sharing.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |