NATIONAL INSTITUTE OF TECHNOLOGY, ROURKELA

DEPARTMENT OF MINING ENGINEERING

END SEMESTER EXAMINATIONS 2023-24

SESSION: 2023-24 SEMESTER: 7th COURSE: B Tech

SUBJECT: COMPUTER APPLICATIONS IN MINING SUBJECT CODE: MN-5002
MAX MARKS: 50 EXAM DURATION: 3 HOURS

Instructions:
• Answer all Questions.
• Maximum marks for each question are mentioned on the right side.
• Some questions have internal choices. Take time to read questions properly and address the requirements of the
questions.
• Write precise, to the point answers as per questions’ requirement. In descriptive answers, try and describe the principle
using bulleted sentences if possible. Figures/ charts/ tables always have more impact than text.
• Use of calculators is allowed.
• Draw and label neat figures and sketches in the answer sheets as per requirement.

Q.
Question Marks
No.
1 In reference to the Finite Element Method (FEM), what are the roles of Governing Equations and Basis 2
functions?
2 In reference to FEM, what is the difference between field stress and body force conditions? Mention one 2
example simulation scenario each for both cases.
3 Match the following Simulation methods with its relevant software. 2

i LEM A Phase 2/ RS 2
ii FEM B Slide
iii FVM C PFC
iv DEM D Fluent-CFD
4 What are the four basic purposes of statistical analysis of mining data? 2
5 In reference to Artificial Neural Network (ANN), explain the role of activation functions. Give two 2
examples of activation functions used in ANN with appropriate sketches.
6 What is difference between Continuum and Non-Continuum types of simulation methods? Give one 2
example of each type.
7 What is the requirement of model validation for any simulation analysis? Describe one method for model 2
validation.
8 Match the flow chart components. 2

i Start/Stop A

ii Input B

iii Decision C

iv Process D

9 In a general programming code, what is the difference between a variable and keyword? Explain with 2
relevant example.
10 What do you understand by parametric corelation during a simulation study? What is the advantage of 2
parametric corelation over traditional simulation? Use appropriate example and illustration.
11 In terms of simulation, what is the difference between a landslide and rock-fall? What are the different 2
factors that are required to be considered while modelling a rock-fall scenario?
Q.
Question Marks
No.
12 Using comparisons with a neuron, describe the structure of a regular Artificial Neuron used for ANN 3
analysis. Include the mathematical processing from inputs to output in the artificial neuron. Use appropriate
illustrations.
13 Compare the Finite Difference Method (FDM) with the Finite Element Method (FEM), mentioning 3
simulation principle, advantages, disadvantages, and application of both methods.
OR

Compare the Finite Volume Method (FVM) with the Finite Element Method (FEM), mentioning simulation
principle, advantages, disadvantages, and application of both methods.
14 In regression analysis, what do you understand by under-fitting and over-fitting of data? What are the 3
disadvantages of these types of fitting scenarios? Describe any two methods to check and avoid overfitting?
Explain using appropriate example and illustration.
OR

Explain the working principle of ANFIS using suitable examples and illustrations. What advantages does
ANFIS have over ANN and FIS?
15 Describe the process of Shear strength reduction (SSR) in FEM analysis. Mention the purpose of SSR, role 4
of Strength Reduction Factor(SRF), and the method of calculating critical SRF. Explain using appropriate
example how can it be used for finding FOS of slopes during slope stability analysis.
16 How do you calculate Sum of Absolute Errors (SAE) and Root mean square error (RMSE)? What will the 5
SAE and RMSE be for the following data when fitted with the equation y = (A).x + (B)

x 4 12 18 24 30
y 25 47 63 80 95

Where,
A is the average of all digits of your roll number significant to two decimal positions.
B is the sum of last four digits of your roll number
( For e.g., if your roll number is 522MN1234, A= (7+2+0+1+2+3+4)/7 =19/7 = 2.71; B= 1+2+3+4 = 10 )
17 An ANN is designed that has two inputs, one hidden layer with three 5
neurons, and one output. For a given iteration, the two inputs are
given as 'A' and 'B', while all weights (W1-W7) are having a value of
'C', and all biases (B1-B4) having a value of 'D'.

Assuming all neurons have a Rectified Linear Unit (ReLU)

activation function, find the computed output for the first forward
propagation.

Given that A, B, C, and D are derived from your roll number as follows.
A is the sum of all digits of your roll number
B is the sum of last four digits of your roll number ,
C is the average of all digits of your roll number
D is the average of last four digits of your roll number

For eg. if your roll number is 120MN3456, then

A = (1+2+0+3+4+5+6)=21, B = (3+4+5+6)=18, C =(1+2+0+3+4+5+6)/7 =[21/7]=3.00, and
D= (3+4+5+6)/4 = 18/4=4.5
19 (a) Design and draw an Artificial Neural Network with the following details: 5

No. of Inputs = A (A is the second digit of your roll no.)

No. of Outputs = B (B is one)
No. of Hidden Layers = C (C is the non-decimal portion of the average of all digits in your roll no.)
No. of Neurons per Layer = D (D is the sum of last digit and third last digit of your roll number)

For eg. if your roll number is 120MN1234, then A =2, B = 1, C =[(1+2+0+1+2+3+4)/7] = [(13)/7]= [non-
decimal(1.857)]= 1, and D= (2+4) = 6

(b) Assuming the designed ANN is fully connected (no clustering) and output layer has no bias, find the
number of weights and biases that will be optimized in this ANN.