Scribd
Upload
30 views

Linear Regression Fitting of A Nonlinear Expression (Example 1)

This document describes using linear regression to fit nonlinear rate data from a first-order chemical reaction to determine the rate constant k. It provides sample concentration and time data, transforms the concentration data to linearize it, performs linear regression via two methods to determine the slope and intercept, and calculates k as the negative slope. It also outputs the standard error and correlation coefficient of the regression line.

Uploaded by

XheikhKaleem
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
30 views

Linear Regression Fitting of A Nonlinear Expression (Example 1)

This document describes using linear regression to fit nonlinear rate data from a first-order chemical reaction to determine the rate constant k. It provides sample concentration and time data, transforms the concentration data to linearize it, performs linear regression via two methods to determine the slope and intercept, and calculates k as the negative slope. It also outputs the standard error and correlation coefficient of the regression line.

Uploaded by

XheikhKaleem
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 6

Linear Regression Fitting of a

Nonlinear Expression (Example 1)

Consider the first order rate expression for a chemical


reaction:
r  kC A

For a constant volume batch reactor the mass balance is:

dCA
  kC A
n

dt
Rearranging the expression and then integrating from
an initial concentration of CA0 to a final concentration CA
gives:

CA
dCA t


CA0 CA
   kdt
0

CA
ln  kt or C A  C A 0e  kt

C A0
Based on the data below find the rate constant k for the first order reaction:

CA (gmole/liter) Time (sec)


0.100 0
0.0892 500
0.0776 1000
0.0705 1500
0.0603 2000
0.0542 2500
0.0471 3000
Ca=[0.1; 0.0892; 0.0776; 0.0705; 0.0603; 0.0542; 0.0471];
t=0:500:3000
t=t'
n=7; %data points
y=log(Ca./Ca(1));%transform y data
x=t

a1=(n*sum(x.*y)-sum(x)*sum(y))/(n*sum(x.^2)-sum(x)^2);
a0=sum(y)/n-a1*sum(x)/n;

fprintf('\nThe slope of the regression line equals %.6f\n',a1);


fprintf('\nThe intercept of the regression line equals %.6f\n',a0);
fprintf('\nThus k equals %.7f\n',-a1);
S=sum((a0+a1.*x-y).^2);
stderror=sqrt(S/(n-2));
fprintf('\nThe standard error of the regression line is %.6f\n',stderror);

Sb=sum((y-sum(y)/n).^2);
r=sqrt((Sb-S)/Sb);
fprintf('\nThe correlation coefficient of the regression line equals %.6f\n\n',r);
We could also solve example using polyfit

Ca=[0.1; 0.0892; 0.0776; 0.0705; 0.0603; 0.0542; 0.0471];


t=0:500:3000
t=t'
n=7; %data points
y=log(Ca./Ca(1));%transform y data
x=t
fprintf('\nComparing the results with the use of the polyfit function \n');

coeff=polyfit(x, y, 1);
fprintf('\nThe slope of the regression line equals %.6f\n',coeff(1));
fprintf('\nThe intercept of the regression line equals %.6f\n',coeff(2));
fprintf('\nThus k equals %.7f\n',-coeff(1));

You might also like