Polynomial help

rodders

rodders

rodders

Soldato
Joined
14 Dec 2004
Posts
3,697
Location
South
Maths, bleh.

I have some dyno output data from a master bench and data taken at the same points on a second bench.

I need to make the output of the second bench, line up with the master.

Master bench data, RPM v Torque

2000 168.663
4000 154.17
6000 140.05
8000 124.223
10000 104.61
12000 79.13

Second bench

2000 183.586
4000 169.37
6000 154.41
8000 136.463
10000 113.29
12000 82.65

I have set a trend line on each and taken the polynomials:

Master y = -3E-07x2 - 0.0039x + 176.87

Second y = -5E-07x2 - 0.0027x + 189.89

I have the answer given to me, of +0.001×Cd*^2+0.6077×Cd*+22.093

So if I take the dyno torque output of the second bench and apply that to it, the result will be inline with the master.

But how do you work out that correcting polynomial?
 
How about making a new plot by subtracting the torque values from one another at each RPM and then fit the data in the most appropriate way (linear. polynomial etc). You can then use the equation of the fitted data as an offset/correction.
 
You don't need polys for the individual sets. You have a set x (second) that you'd like to transform into set y (master), i.e. y = f(x) where f(.) is some polynomial.

This is easy here since the rpm points are the same for each set.

So

x = [168.663, ..., 79.13]
y = [183.586, ..., 82.65]

Then use something like MATLAB's polyfit(x, y, 2) (i.e. 2nd order), or numpy.polyfit in Python.

The resulting 3 coefficients are the order 0, 1 and 2 elements of the correcting polynomial :)
 
You must log in or register to reply here.
Back
Top Bottom