Sei sulla pagina 1di 3

INDIAN INSTITUTE OF TECHNOLOGY – MADRAS

Department of Chemical Engineering


CH-3050: Computational Techniques
Home Assignment 04
Due on: Tuesday, 28th September 2010

Problem 1: Reaction Kinetics using Linear Regression

The rate of an enzymatic reaction is given by the following expression:


k[S]
r= .
Km + [S]
The problem of estimating k and Km can be converted to linear regression by inverting the above
1
expression and defining x = [S] and y = 1r . The following data was obtained in the lab:

[S] 1.233 0.540 0.442 0.258 0.198 0.162 0.130 0.128


r 5.970 3.319 2.253 2.547 1.493 1.182 1.095 0.869

1. Obtain y = a0 + a1 x and hence find the values of k and Km

2. Now, take x as an independent variable and obtain x = b0 + b1 y and hence find the values of
k and Km

Compare the two values and explain your results.

Problem 2: Multi-Linear Regression

Two different methods were shown in the lectures for linear regression of the form y = a0 +a1 x+a2 u.
The first method was to minimize the sum of square errors,
N 
X  2
I= yi − a0 + a1 xi + a2 ui
i=1

by equating
∂I ∂I ∂I
= = =0
∂a0 ∂a1 ∂a2
and solving for a0 , a1 , a2 .

An alternate method was the matrix method, where we created data matrices to represent the data
−1
in the form Y = Aφ + E, and obtained the solution as φ = AT A AT Y .

1
CH-3050: Computational Techniques
August–November 2010

1. Give the expression to compute a0 , a1 , a2 using the first method

2. Show that for the problem involving one dependent (y) and two independent (x, u) variables,
the matrix method leads to the same result as that obtained using the first method.

3. For the following data, obtain the fit using both the methods described above.

y 4.5 5.9 6.25 6.75 7.7 8.85 10.75 11.85


x 4 5 6 7 11 13 18 19
u 0.4 0.8 0.85 0.95 1 1.25 1.55 1.85

Problem 3: Polynomial Interpolation

The following data is available for saturation pressure of acetone as a function of temperature:

T (K) 259 273 290 320 390 446 470 508


psat 0.0427 0.0945 0.2153 0.7445 5.655 17.68 26.63 47.0

We can fit polynomials up to order 7 to the above data. Please complete the following:

• Fit polynomials of all orders from 1 to 7 to the above data.

• Compute the correlation coefficient for each of them.

• Comment on what order polynomial would best represent the data and why.

• An additional experiment was conducted and the saturation pressure at 351 K was 2.016 bar.
How well does your chosen polynomial predict the saturation pressure?

Problem 4: Lagrange and Divided Difference Methods

Repeat the previous problem using (1) Lagrange and (2) Newton’s interpolating polynomials.

2 Home Assignment 04
CH-3050: Computational Techniques
August–November 2010

Problem 5: Bonus Problem (not graded with the rest of the assignment)

We now solve a nonlinear regression problem. You will solve the problem by two methods: (i) using
your own program, and (ii) using an in-build MATLAB function.

3 Home Assignment 04

Potrebbero piacerti anche