Chapter 06.

04
Nonlinear Models for Regression-More Examples
Chemical Engineering
Example 1
Below is given the FT-IR (Fourier Transform Infra Red) data of a 1:1 (by weight)
mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC). Absorbance 𝑃 is
given as a function of wavenumber, m.
Table 1 Absorbance as a function of wavenumber.

Wavenumber, 𝑚 Absorbance, 𝑃
(cm−1 ) (arbitrary unit)
804.184 0.1591
827.326 0.0439
846.611 0.0050
869.753 0.0073
889.038 0.0448
892.895 0.0649
900.609 0.1204

Regress the above data to a second order polynomial

𝑃 = 𝑎0 + 𝑎1 𝑚 + 𝑎2 𝑚2
Find the absorbance at 𝑚 = 1000 cm−1
Solution
Table 2 shows the summations needed for the calculations of the constants of the
regression model.

06.04.1
 06.04.2 Chapter 06.04

Table 2 Summations for calculating constants of model.

Absorban
Wavenumber,
ce, 𝑃
𝑖 𝑚 𝑚2 𝑚3 𝑚4 𝑚×𝑃 𝑚2 × 𝑃
(arbitrar
(cm−1 )
y unit)
1 804.18 0.1591 6.4671× 105 5.2008× 108 4.1824× 1011 127.95 1.0289× 105
2 827.33 0.0439 6.8447× 105 5.6628× 108 4.6849× 1011 36.319 3.0048× 104
3 846.61 0.0050 7.1675× 105 6.0681× 108 5.1373× 1011 4.233 3.583× 103
4 869.75 0.0073 7.5647× 105 6.5794× 108 5.7225× 1011 6.349 5.522× 103
5 889.04 0.0448 7.9039× 105 7.0269× 108 6.2471× 1011 39.828 3.5409× 104
6 892.90 0.0649 7.9726× 105 7.1187× 108 6.3563× 1011 57.948 5.1742× 104
7 900.61 0.1204 8.1110× 105 7.3048× 108 6.5787× 1011 108.43 9.7655× 104
7
∑ 6030.4 0.4454 5.2031× 106 4.4961× 109 3.8909× 1012 381.06 3.2685× 105
𝑖=1

𝑃 = 𝑎0 + 𝑎1 𝑚 + 𝑎2 𝑚2 is the quadratic relationship between the absorbance and the

wavenumber where the coefficients 𝑎0 , 𝑎1 , 𝑎2 are found as follows
𝑛 𝑛 𝑛

𝑛 (∑ 𝑚𝑖 ) (∑ 𝑚𝑖2 ) ∑ 𝑃𝑖
𝑖=1 𝑖=1 𝑖=1
𝑛 𝑛 𝑛 𝑎0 𝑛

(∑ 𝑚𝑖 ) (∑ 𝑚𝑖2 ) (∑ 𝑚𝑖3 ) [𝑎1 ] = ∑ 𝑚𝑖 𝑃𝑖

𝑖=1 𝑖=1 𝑖=1 𝑎2 𝑖=1
𝑛 𝑛 𝑛 𝑛

(∑ 𝑚𝑖2 ) (∑ 𝑚𝑖3 ) (∑ 𝑚𝑖4 ) ∑ 𝑚𝑖2 𝑃𝑖

[ 𝑖=1 𝑖=1 𝑖=1 ] [ 𝑖=1 ]
𝑛=7
∑7𝑖=1 𝑚𝑖 = 6.0304 × 103
∑7𝑖=1 𝑚2𝑖 = 5.2031 × 106
∑7𝑖=1 𝑚3𝑖 = 4.4961 × 109
∑7𝑖=1 𝑚4𝑖 = 3.8909 × 1012
∑7𝑖=1 𝑃𝑖 = 0.4454
∑7𝑖=1 𝑚𝑖 𝑃𝑖 = 381.06
∑7𝑖=1 𝑚2𝑖 𝑃𝑖 = 3.2685 × 105
We have
7.0000 6.0304 × 103 5.2031 × 106 𝑎0 0.4454
[6.0304 × 10 5.2031 × 106
3 9 ] [𝑎1 ] = [381.06 ]
4.4961 × 10
12 𝑎2 5
5.2031 × 106 4.4961 × 109 3.8909 × 10 3.2685 × 10

Solve the above system of simultaneous linear equations, we get

Nonlinear Models for Regression-More Examples: Chemical Engineering
06.04.3

𝑎0 43.985
[𝑎1 ] = [−0.10268 ]
𝑎2 −5.9922 × 10−5

The polynomial regression model is

𝑃 = 𝑎 0 + 𝑎 1 𝑚 + 𝑎 2 𝑚2
= 43.985 − 0.10268𝑚 − 5.9922 × 10−5 𝑚2

Figure 1 Second order polynomial regression model for absorbance as a function of

wavenumber.

To find 𝑃 where 𝑚 = 1000 cm−1 :

𝑃 = 𝑎 0 + 𝑎 1 𝑚 + 𝑎 2 𝑚2
= 43.985 − 0.10268𝑚 − 5.9922 × 10−5 𝑚2
= 43.9855 − 0.10268 × (1000) − 5.9922 × 10−5 × (1000)2
= 1.2221

Example 2
The mechanism of polymer degradation reaction kinetics is suspected to follow Avrami
or random nucleation reaction,
(𝑇−𝑇0 ) − 𝐸
𝑓(α) = 𝐴 𝑒 𝑅𝑇
𝑏
06.04.4 Chapter 06.04

where 𝑓(α) = −ln(1 − α), T is the absolute temperature (K), 𝑏 is the heating rate in
K/min, 𝐴 is the frequency factor with units of rate constant, 𝑅 is the gas constant (8.314
kJ/kmol-K) and 𝑇0 is the activation temperature. Given that 𝑇0 = 338.75 K, 𝑏 = 10
K/min and conversion, α, at different temperatures are as given in table 3. Use the
method of least squares to determine the values of 𝐴and𝐸.

Table 3 Conversion at given different temperatures

Temp (K) 360 370 380 390 400 410 420 430 440
Conversion, 0.105 0.201 0.342 0.514 0.675 0.802 0.892 0.954
1.00
α 5 0 5 6 7 6 4 4

Solution
To set-up the table, we must re-write equation
(𝑇−𝑇0 ) − 𝐸
−ln(1 − α) = 𝐴 𝑒 𝑅𝑇
𝑏
as
𝑏ln(1 − α) 𝐸
− = 𝐴𝑒 −𝑅𝑇
(𝑇 − 𝑇0 )

Taking natural log of both sides of the above equation, we obtain

𝑏ln(1−α) 𝐸
ln [− (𝑇−𝑇 ) ] = ln(𝐴) − 𝑅𝑇
0

so that the equation is in the form 𝑦 = β0 + β1 𝑥 where

𝑏ln(1−α)
𝑦 = ln [− (𝑇−𝑇 ) ]
0
β0 = ln(𝐴)
𝐸
β1 = − 𝑅
1
𝑥=𝑇
𝑛 ∑𝑛 𝑛 𝑛
𝑖=1 𝑥𝑖 𝑦𝑖 −∑𝑖=1 𝑥𝑖 ∑𝑖=1 𝑦𝑖
β1 = 2
𝑛 ∑𝑛 2 𝑛
𝑖=1 𝑥1 −(∑𝑖=1 𝑥𝑖 )

∑𝑛
𝑖=1 𝑦𝑖 ∑𝑛
𝑖=1 𝑥𝑖
β0 = ( ) − β1 ( )
𝑛 𝑛
Nonlinear Models for Regression-More Examples: Chemical Engineering
06.04.5

Table 4 Example on nonlinear exponential problem.

𝑖 𝑇 α 𝑥 𝑦 𝑥2 𝑥×𝑦
1 360 7.7160×
0.1055 2.7778× 10−3 −2.9476 10−6 −8.1877× 10−3
2 370 7.3046×
0.2010 2.7027× 10−3 −2.6338 10−6 −7.1183× 10−3
3 380 6.9252×
0.3425 2.6316× 10−3 −2.2862 10−6 −6.0163× 10−3
4 390 6.5746×
0.5146 2.5641× 10−3 −1.9588 10−6 −5.0225× 10−3
5 400 6.2500×
0.6757 2.5000× 10−3 −1.6936 10−6 −4.2341× 10−3
6 410 5.9488×
0.8026 2.4390× 10−3 −1.4796 10−6 −3.6088× 10−3
7 420 5.6689×
0.8924 2.3810× 10−3 −1.2932 10−6 −3.0791× 10−3
8 430 5.4083×
0.9544 2.3256× 10−3 −1.0835 10−6 −2.5199× 10−3
∑ −1.5376× 5.1797×
2.0322× 10−2 101 10−5 −3.9787× 10−2

𝑛=8
8
−2
∑ 𝑥𝑖 = 2.0322 × 10
𝑖=1
∑8𝑖=1 𝑦𝑖 = −1.5376 × 101
∑8𝑖=1 𝑥𝑖 𝑦𝑖 = −3.9787 × 10−2
∑8𝑖=1 𝑥2𝑖 = 5.1797.10−5
8(−3.9787×10−2 )−(2.0322×10−2 )(−1.5376×10−1 )
β1 = 8(5.1797×10−5 )−(2.0322×10−2 )2
3
= −4.1561 × 10

−1.5376 × 101 2.0322 × 10−2

β0 = − (−4.1561 × 103 )
8 8
= 8.6352
𝐴 = 𝑒 β0
= 𝑒 8.6352
= 5.6264 × 103
𝐸 = −β1 𝑅
= −(−4.1561.103 ) × 8.3140
= 3.4553 × 104
This gives the model as
06.04.6 Chapter 06.04

(𝑇−338.75) 4.1561×103
−ln(1 − α) = 5.6264 × 103 × 𝑒− 𝑇
10

Figure 2 Polymer degradation reaction kinetics rate as a function of temperature.

Nonlinear Models for Regression-More Examples: Chemical Engineering
06.04.7

Example 3
The progress of a homogeneous chemical reaction is followed and it is desired to
evaluate the rate constant and the order of the reaction. The rate law expression for the
reaction is known to follow the power function form
−𝑟 = 𝑘𝐶 𝑛 . (1)
Use the data provided in the table to obtain 𝑛 and𝑘.
Table 11 Chemical kinetics
𝐶𝐴 (gmol/l) 4 2.25 1.45 1.0 0.65 0.25 0.06
gmol⁄
−𝑟𝐴 ls 0.398 0.298 0.238 0.198 0.158 0.098 0.048

Solution
Taking natural log of both sides of Equation (1), we obtain
ln(−𝑟 ) = ln(𝑘) + 𝑛ln(𝐶 )
Let
ln(−𝑟 ) = 𝑧
ln(𝐶 ) = 𝑤
ln(𝑘) = 𝑎0 from which 𝑘 = 𝑒 𝑎0 (2)
𝑛 = 𝑎1 (3)
We get
𝑧 = 𝑎0 + 𝑎1 𝑤
This is a linear relation between 𝑧 and 𝑤, where
∑𝑚 𝑚 𝑚
𝑖=1 𝑤𝑖 𝑧𝑖 −∑𝑖=1 𝑤𝑖 ∑𝑖=1 𝑧𝑖
𝑎1 = 2
𝑚 ∑𝑚 2 𝑚
𝑖=1 𝑤𝑖 −(∑𝑖=1 𝑤𝑖 )

∑𝑚
𝑖=1 𝑧𝑖 ∑𝑚
𝑖=1 𝑤𝑖
𝑎0 = 𝑚
− 𝑎1 𝑚
(4a,b)

Table 6 Kinetics rate law using power function.

𝑖 𝐶 −𝑟 𝑤 𝑧 𝑤×𝑧 𝑤2
1 4 0.398 1.3863 −0.92130 −1.2772 1.9218
2 2.25 0.298 0.8109 −1.2107 −0.9818 0.65761
3 1.45 0.238 0.3716 −1.4355 −0.5334 0.13806
4 1 0.198 0.0000 −1.6195 0.0000 0.00000
5 0.65 0.158 −0.4308 −1.8452 0.7949 0.18557
6 0.25 0.098 −1.3863 −2.3228 3.2201 1.9218
7 0.06 0.048 −2.8134 −3.0366 8.5431 7.9153
7
∑ −2.0617 −12.391 9.7657 12.7401
𝑖=1
06.04.8 Chapter 06.04

𝑚=7
∑7𝑖=1 𝑤𝑖 = −2.0617
∑7𝑖=1 𝑧𝑖 = −12.391
∑7𝑖=1 𝑤𝑖 𝑧𝑖 = 9.7657
∑7𝑖=1 𝑤2𝑖 = 12.7401

From Equation (4a, b)

7×(9.7657)−(−2.0617)×(−12.391)
𝑎1 = 7×12.7401−(−2.0617)2
= 0.50408
−12.391 −2.0617
𝑎0 = − (−0.50408)
7 7
= −1.6217
From Equations (2) and (3), we obtain
𝑘 = 𝑒 −1.6217
= 0.19755
𝑛 = 𝑎1
= 0.50408
Finally, the model of progress of that chemical reaction is
−𝑟 = 0.19755 × 𝐶 0.50408
Nonlinear Models for Regression-More Examples: Chemical Engineering
06.04.9

Figure 3 Kinetic chemical reaction rate as a function of concentration.