Dynamics and Optimization of a Quarter Car Model with Power Law Spring

Ritvik Dobriyal1, Pushpendra Kumar1, Niraj Kumar Jha2, Vipul Paliwal1, Narendra Gariya1
Department of Mechanical Engineering
1, Graphic Era Deemed to be University Dehradun India
2, Hari Shankar Singhania Elastomer and Tyre Research Institute, JKTyre, Mysore, India
E-mail: [email protected]

Abstract: Suspension system is an important part of an automobile, which ensures ride comfort and passengers safety. Advanced
sus-pension systems have been developed such as active suspension with feedback systems. However, most of the automobile are
built with traditional passive suspensions because of simple design and easy maintenance. In this paper, a passive suspension
system has been analysed with power law spring because a spring may not behave linearly in specific driving conditions. A quarter
car model is considered to analyse the suspension dynamics. The main contribution of this work is to find the suitable value of the
power index of the spring in order to ensure better suspension performance. Simulation studies are performed to analyse the
performance of the suspension system with varying values of the power index. Majority of results indicate that a spring with a
higher power law index performs better both in terms of ride comfort and road holding for different road speeds and road profiles.
Finally, optimization is performed to find the optimized value of the power index, which verifies the results obtained from
simulation.

Keywords: Genetic Algorithm; Power Law; Quarter Car Model; Ride Comfort

1. Introduction
A suspension system is an essential part of an automobile. Other than isolating the passengers from the road
irregularities a suspension system is also tasked with providing required handling performance. Suspension design
have been extensively studied before using both experimental and theoretical approach [1],[2],[3],[4],[5],[6]. A
detailed theoretical analysis of the suspension system can be carried out by approximating it with three-dimensional
models involving multiple degrees of freedom and non-linear elements. However, two degree of freedom quarter car
model have been extensively used previously to study the suspension behavior (especially for ride and road holding
characteristics) due its simplicity and capability to provide qualitatively correct results. Further, the results obtained
from this simplified model can be analyzed exhaustively using more complex models.
The basic challenge of a suspension system is to provide better ride comfort while at the same time minimizing
suspension travel and providing good road holding. Simultaneously achieving all three objectives is impossible for a
passive suspension [7] and [8]. Turkay and Akc¸ay (2005) [7] studied the compromise between the various
requirements by changing the suspension damping of a linear quarter car model excited by random road profiles.
Optimization of tire stiffness, suspension spring stiffness and damping was performed for a linear quarter car model
excited by random road profile by [9]. Chi et al. (2008) [10] also performed the optimization of a linear quarter car
model using three different techniques namely genetic algorithm, pattern search algorithm and sequential quadratic
programming. Their results indicated that the performance of the quarter car model was significantly improved and
genetic algorithm and pattern search algorithm performed better than the sequential quadratic programming. Verros
et al. (2005) [11] performed the optimization of the suspension spring and damping using quarter car model with
constant rate and dual rate dampers. A nonlinear quarter car model including a driver seat and having quadratic tire
stiffness and cubic suspension stiffness was analyzed and optimized by [12]. Other notable studies on the analysis
and optimization of quarter car model include [13], [14], [15].
Usually the suspension elements i.e. the spring and damping of the quarter car model are considered to be liner, as
this provides for simplicity and yet results in a reasonably accurate representation of the actual suspension system
[16], [3], [17] and [18]. However, the actual elements have considerable nonlinearities and using nonlinear elements
in the suspension system increases the accuracy of the simulations [19].
The coil spring used in the suspension system of commercial vehicles may depart from linear behavior and behave
as a hardening or softening spring. Such behavior or departure from linearity can be adequately described using a
power law. Zhu and Ishitobi (2004) [20] and Zhu and Ishitobi (2006) [21] have studied the dynamics of harmonically
excited half car model and full car model having suspension spring modelled using the power law. However, in the
existing works, the power law index for the suspension spring was kept constant. Therefore, the main contribution of
this paper is to analyze the suspension dynamics with varying power law index for the spring. In the present paper,
the dynamics and optimization of a quarter car model with power law suspension spring is studied by running the
model over localized and distributed geometric irregularities. The power law index is varied over a range such that
the spring changes from a hardening spring to a softening spring.
The organization of the paper is as follows. Section 2 describes the dynamical equations governing the quarter car
model, its parameter values along with the road profiles and calculation procedure. Section 3 describes the various
results for the quarter car model moving over a bump and artificial road profiles. Optimization of the suspension
parameters of the quarter car model is described in section 3.3, while a brief summary and conclusion of the paper is
contained in section 4.

2. Methodology
Effect of power law spring on the performance of suspension system is evaluated with the help of 2-DOF quarter car model as shown
in fig 1. The dynamics of this model is governed by a set of four ordinary differential equations as given in equation 1 to 4

ks c (1)
v2 = − Sgn( x2 − x1 ) | ( x2 − x1 ) |n − s (v2 − v1 ) − g ,
M M

x2 = v2 , (2)

kt k c
v1 = − ( x1 − y ) + s Sgn( x2 − x1 ) | ( x2 − x1 ) |n + s (v2 − v1 ) − g , (3)
m m m

x1 = v1 , (4)
where 𝑥1 , 𝑥2 , 𝑣1 , and 𝑣2 represent the positions and velocities of the unsprung and sprung mass respectively, 𝑦 denotes the vertical
height of the road surface. The damping of the tire is usually small and is neglected here. Except the suspension spring all other elements
are considered to be linear for the sake of simplicity. The nonlinear suspension spring is modelled using a power law as
F = k s Sgn( x2 − x1 ) | x2 − x1 |n , (5)
where 𝐹 is the spring force, 𝑘𝑠 is the spring stiffness, 𝑥2 − 𝑥1 is the deformation or compression of the spring, 𝑆𝑔𝑛 is the signum
function and 𝑛 is the power law index. Figure 2 shows the spring force vs displacement curves for three different values of 𝑛. For 𝑛
𝑑𝑓
<1 the spring behaves as a softening spring since the derivative of spring force with respect to displacement ( ⁄𝑑𝑥 ) decreases with
displacement, while for 𝑛 >1 the spring behaves as a hardening spring. For 𝑛 =1 the spring behaves as a linear spring.

Fig. 1. Quarter car model under study.

Numerical values for the various parameters of the quarter car model are given in table 1 and are typical of a passenger car.
Table 1. Numerical values for the various parameters of the quarter car model.
Parameter Value
𝑀 500kg
𝑚 30kg
𝑘𝑠 20000N/m
𝑘𝑡 200000N/m
𝑐𝑠 1500N-s/m

The quarter car model is kinematically excited by a prescribed road profile (y in equation 3) by running the model over it at a constant
forward speed. In this study two different kinds of road profiles are used, as shown in fig 3. Figure 3 (a) shows a road profile with a
localized geometric irregularity (a bump), while fig. 3 (b) shows artificial road profiles for smooth and rough roads. These artificial road
profiles are generated using the sinusodial approximation technique given in [22] and [23].
A random road profile can be approximated by a power spectral density function of the form
 −w

() = (0 )   . (6)
 0 
To generate a random road profile of length L a superposition of N sine waves is taken as
N
y ( x) =  Ai sin(i x − i ) (7)
i =1


Ai = (i ) , (8)

with ∆Ω=0.01 and L=628.31m. For generating the smooth and rough road, Φ(Ω0 ) is taken to be 16 and 256 x10-3m3 respectively,
while 𝑤 is taken to be 2 for both roads as given in [22].
Position and velocities of the sprung and unsprung mass can be obtained as a function of time by integrating the set of ordinary
differential equations by using Matlab's ODE 45 solver. This solver is an ordinary differential equation solver based on an explicit
Runge-Kutta formula of [24].
The displacement of sprung and unsprung mass under static condition are calculated as
x1 = −( M + m) g / kt , (9)
and
x2 = −( Mg / k s )1/ n + x1. (10)
Figures showing variation of sprung and unsprung mass displacement with respect to time in the following section are deviations
from the results calculated from eq. 9 and 10.

Fig. 2. Spring force vs displacement for n=0.8, n=1 and n=1.2.

(a) (b)

Fig. 3. Figure (a) shows the road profile with a parabolic bump, while figure (b) shows artificial road profiles for smooth and
rough roads
3. Results
3.1 Response under excitation from isolated irregularity
The response of the quarter car model for motion over the bump (shown in fig. 3 (a)) at a constant forward speed of 10km/h is shown in
fig. 4. Figure 4 (a) shows the displacement of sprung mass as a function of time. The displacement rises sharply at nearly 3.5 seconds
when the quarter car model moves up the bump and then falls sharply as the model climbs down. Following this the oscillations persist
for a while and eventually die down due to the damping in the suspension system. Similar behaviour is seen in fig. 4. (b) and (c), also
rapid fluctuations in sprung mass acceleration and vertical load on the tire can be observed as the quarter car model hits the bump. For
the suspension spring travel, initially the graph falls rapidly due to the compression of the spring as the tire climbs up the bump, this is
followed by a rapid rise in the graph due to the extension of the spring as the tire climbs down the bump. Beyond this the behaviour of
the graph is qualitatively similar to the previous three figures.
From fig. 4. (a) it can be seen that the greatest sprung mass travel is for n=0.8. Since providing a smooth level ride to the passengers is
a key requirement of a suspension system, therefore it is worth investigating the variation of total sprung mass travel with index n. Figure
6 (a) shows the variation of total sprung mass displacement with n for different forward speeds. Low values of speeds are chosen since
they are typical of a vehicle moving over a bump. From the figure it can be seen that the higher values of n are helpful in reducing the
displacement of sprung mass, further, this effect is more observable for lower speeds than for higher speeds.
One more point to observe from fig. 4. (a) is that the signal for n=0.8 i.e. the red curve leads the blue curve (n=1) which leads the black
curve (n=1.2). Frequency of oscillation of the sprung mass decreases with increasing n. This is indeed the case as can be seen in fig.5
which shows the time signal of fig. 4. (a) in frequency domain. The reason suspension system with lower n can oscillate with a higher
𝑑𝑓
frequency is because the local spring stiffness i.e. ( ⁄𝑑𝑥 ) is higher for lower n. For n varying from 0.8 to 1.2 the initial compression
𝑑𝑓
of the suspension spring varies from 0.17 to 0.3m. Under such condition ( ⁄𝑑𝑥 ) is higher for lower n, thus due to higher local spring
stiffness the oscillation frequency of the suspension system with lower n is higher.
For motion over a localized geometric irregularity such as a bump or pothole the maximum value of absolute acceleration of sprung
mass can serve as an indication of severity of shock [25], [26] and [27]. Figure 6 (b) shows variation of maximum absolute acceleration
with n for different forward speeds. For all forward speeds under consideration the maximum acceleration of sprung mass decreases
with increasing n. For speed of 10km/h and less the maximum value of acceleration occurs in the downward direction while for higher
speeds the maximum value of acceleration occurs in the upward direction and the variation with n is weaker as compared to lower
speeds.
For a tire to provide adequate traction there should be sufficient vertical load on it. During motion of a vehicle over a road irregularity,
vertical load on the tire fluctuates and can severely reduce its road holding capacity. For this reason, it is desirable to have a higher value
for the minimum vertical load on tire. Variation of minimum vertical load with n for various values of forward speeds is given in fig. 6
(c). It can be seen that the minimum vertical load rises gently with n. This rise reduces slightly with increasing speed until it is negligible
for forward speed of 14km/h.
Other than providing a level ride and cushioning the shock, it is advantageous to have the suspension spring travel as less as possible
so that the suspension system can fit in minimum space. Therefore, it is worthwhile to investigate the variation of suspension spring
travel with n. Figure 6 (d) shows the variation of total spring travel with n for different forward speeds. For speeds of 8 and 10 km/h a
slight decrease in spring travel can be observed, however for higher speeds an overall slight increase can be observed.

(a) (b)
 (c) (d)

Fig. 4. Figure (a), (b), (c) and (d) show the sprung mass displacement, sprung mass acceleration, vertical load on tire, and travel
for suspension spring for three different values of n. The red curve is for n=0.8, blue for n=1 and black for n=1.2.

Fig. 5. Figure shows the data in fig. 4 (a) in frequency domain. The red curve is for n=0.8, blue for n=1 and black for n=1.2.

(a) (b)

(c) (d)

Fig. 6. Figure (a), (b), (c) and (d) show the variation of sprung mass displacement, sprung mass acceleration, vertical load on tire,
and travel for suspension spring with n for different speeds.
3.2 Response under excitation from artificially generated road
For motion of a vehicle over a road profile with nearly uniformly distributed geometric irregularities, rms value of acceleration of the
sprung mass can serve as an indicator of ride comfort [28], [29] and [12]. Figure 7 shows the time series of the sprung mass acceleration
for motion with a constant forward speed of 50km/h over the smooth road profile for three different values of n. Careful examination of
the figure reveals that acceleration peaks higher for n=0.8 and is least for n=1.2. The variation of rms acceleration of the sprung mass of
the quarter car model with n for different values of forward speed is presented in fig. 8. The values of the speeds are chosen such that
they are representative of a vehicle moving on a straight smooth road. Figure 8 shows that rms acceleration of the sprung mass reduces
with increasing n for both smooth and rough roads. This effect is more pronounced for higher speeds than lower ones.

Fig. 7. Sprung mass acceleration of the quarter car model for three different values of n for a constant forward speed of 50km/h driven
over the smooth road profile

(a) (b)
Fig. 8. Figure (a) and (b) show the variation of rms acceleration of the sprung mass with n for different forward speeds for
smooth and rough road respectively.

Although rms acceleration of the sprung mass shows a decrease with n, however only those vibrations which are within the frequency
range of 0.5 to 20 Hz are considered important for ride. Vibrations below this frequency range are termed quasi-static, while higher
frequency oscillations lie in the acoustic range [30]. In [31] the ride frequency range is estimated to be from 0.25 to 25Hz.
Figure 9 shows the spectra of the sprung mass acceleration driven at a forward speed of 50km/h over the smooth and rough road. It
can be seen that most of the power lies within the frequency range associated with ride comfort. Spectral Power is also higher for n=0.8
as expected from the previous results, especially for frequency above 1 Hz. Further, it can also be observed that the response is stronger
at higher frequency for lower n and vice versa. This is similar to the result seen in fig. 5. Spectra calculated over a speed range of 40 to
80km/h were found to be qualitatively similar.
Other than the rms acceleration, the ride comfort in this study is also evaluated using weighted rms acceleration calculated as per
section 6 equation 9 of ISO 2631-1 [25]. The results are given in fig. 10. The results are similar to fig. 8 with weighted rms acceleration
marginally decreasing with n. Similar to the previous results, the observed effect is more prominent for smooth road as compared to the
rough.
Maintaining effective contact with the road surface and minimising suspension spring travel are also desirable features of a
suspension system. For motion over road profiles, variation in vertical load on the tire and suspension spring travel are qu antified
using their rms values. Negligible deviation in rms vertical load with both forward speed and n is observed as given in fig. 11. For
compression of suspension spring the rms values increase with vehicle speed and a slight increment with n is also observed.
 (a)

(b)
Fig. 9. Figure (a) and (b) show the spectrum of acceleration of the sprung mass for different n at a forward speed of 50km/h for
smooth and rough roads respectively. The dashed vertical lines indicate the range of frequency from 0.25 to 25 Hz.

(a) (b)

Fig. 10. Figure (a) and (b) show the variation of weighted rms acceleration of the sprung mass with n for different forward speeds
for smooth and rough road respectively.

(a) (b)
Fig. 11. Figure (a) and (b) show the variation of rms vertical load on the tire with n for different forward speeds for smooth and
rough road respectively

(a) (b)
Fig. 12. Figure (a) and (b) show the variation of rms compression of suspension spring with n for different forward speeds for
smooth and rough road respectively

3.3 Optimization
The suspension elements of the quarter car model are optimized with respect to ride comfort and suspension spring
travel. Suspension characteristic i.e. 𝑘𝑠 , 𝑐𝑠 and n are optimized over a range of driving speeds and different road
profiles as discussed in the previous section using genetic algorithm. Genetic Algorithm (GA) is chosen for
optimization, which is an advanced tool for finding best solution in optimization of desired variables of an objective
function. Robustness of solution is the main objective of GA with the help of balance between efficiency and
effectiveness of the solution. Although, there exist many other optimization tools, but in this study, GA has been
selected due to its ability of global search and easy implementation with good results [32].
The performance index for the suspension system is based on the normalized values of modulus of sprung mass
acceleration and suspension spring travel as given in eq. 11
J = w1 J1 + w2 J 2 , (11)

where 𝑤1 and 𝑤2 are the weighing coefficients and 𝐽1 and 𝐽2 are given as
T
1
gT 0
J1 = | x2 | dt , (12)

T
1 1
( Mg / ks ) T 0
and J2 = 1/ n
| x2 − x1 | dt , (13)

and T is the time taken for the quarter car model to run over the road profile. The weighing coefficients 𝑤1 and 𝑤2 are
chosen to be 0.7 and 0.3 respectively. Greater weight-age is given to sprung mass acceleration because ride comfort
is given higher priority in our study.
For performing the optimization, the range of suspension stiffness and damping was specified from 15000 to
25000 N/m and 1000 to 2500 N-s/m respectively, while the range of power law index was specified from 0.7 to 1.5.
The results of optimization of the suspension system or minimization of the performance index as given in eq. 11 at
different forward speeds are given in fig. 13 and 14. It can be seen that the optimized values of the suspension
parameters vary little with the forward velocity. The optimization technique (GA) selects lower spring stiffness,
higher damping and a slightly hardening suspension spring, both for motion over a bump and artificially generated
smooth road. Optimization results for artificially generated rough road were similar to the results obtained for the
artificially generated smooth road.
Selection of lower spring stiffness is in accord with the results of [9], where the standard deviation of sprung mass
acceleration was found to be directly proportional to the square of the suspension spring stiffness for motion over a
random road profile (see equation 13 and fig. 3 of the reference). Selection of higher damping is likely due to the fact
that higher damping will quickly dampen the oscillations and thus result in lower values of sprung mass acceleration
and suspension spring travel.
The power law index n is selected to be somewhat greater than one (between 1.1 and 1.3). This selection can be
understood based on the results presented in the previous sections. From fig. 4 and 8 it can be seen that both the
maximum acceleration (in case of motion over a bump) and the rms acceleration of the sprung mass decreases with
increasing n. Therefore, opting for a lower value of n is not preferable, however, opting for a very high value of n is
not beneficial since with increasing n the decrease in sprung mass acceleration with n is insignificant, besides it may
result in higher spring travel (see fig.12).

(a) (b)

Fig. 13. Figure (a) and (b) show the optimized values of 𝒌𝒔 , 𝒄𝒔 and n for different forward speeds for motion over bump.

(a) (b)
Fig. 14. Figure (a) and (b) show the optimized values of 𝒌𝒔 , 𝒄𝒔 and n for different forward speeds for motion over smooth road
profile.
4. Conclusion
In the present work, suspension dynamics have been analysed considering the quarter car model with a power law spring. The main
contribution of this work is to analyse the performance of the suspension system with varying values of the power law index. The
developed model has been analysed through simulation over the following road profiles: i) road with a localized bump and ii) random
road with smooth and rough profiles. From the obtained results for a bump, it is observed that the amplitudes of sprung mass
displacement and acceleration decrease with higher value of power index (n=1.2), which indicates comparatively increased drive
comfort. Moreover, it is found that the minimum vertical load on tire rises with higher value of power index ensuring better road holding
capability. It is also observed that the total spring travel decreases slightly with higher value of power index, which can result in reduction
of required space for the suspension system.
Subsequently, simulation studies are conducted for random road with smooth and rough profiles. From the obtained results, it is
observed that the rms acceleration of the sprung mass decreases with increasing values of power index and this observation holds for
different forward speeds over smooth and rough road profiles. In addition, negligible deviations are observed in vertical load and total
spring travel with variation of the power index for both the smooth and rough road profiles. In addition, for lower values of n the sprung
mass acceleration is dominant in the higher frequencies for both motion over the bump and road profile. Therefore, a lower value of n
can possibly result in a coarse ride.
Finally, optimization is performed in order to verify the obtained simulation results. Genetic algorithm is selected to optimize the
values of power index, spring stiffness and damping coefficient; while minimizing the objective function depending on the sprung mass
acceleration and spring compression. The optimized value of the power index is found in the range of 1.15 to 1.23 for roads with bump
and smooth profiles and for different speeds. The optimization results support the simulation studies that higher value of the power index
(n>1) will lead to better suspension performance.
References
1. Borowiec M, Sen A K, Litak G, Hunicz J, Koszałka G & Niewczas A, Forschung im Ingenieurwesen, 74(2) (2010) 99-
109.
2. Satyanarayana V S V, Sateesh B & Rao N M, International Journal of Dynamics and Control, 8(1) (2020) 218-228.
3. Sharp R S & Crolla D A, Vehicle system dynamics, 16(3) (1987) 167-192.
4. Hrovat D, Automatica, 33(10) (1997) 1781-1817.
5. Cao D, Song X & Ahmadian M, Vehicle system dynamics, 49(1-2) (2011) 3-28.
6. Wedig W V, (2016), International Journal of Dynamics and Control, 4(2) (2016) 213-220.
7. Türkay S & Akçay H, Journal of sound and vibration, 282(1-2) (2005) 111-124.
8. Karnopp D, Vehicle System Dynamics, 15(1) (1986) 41-54.
9. Gobbi M & Mastinu G, Journal of sound and vibration, 245(3) (2001) 457-481.
10. Chi Z, He Y & Naterer G F, Transactions of the Canadian Society for Mechanical Engineering, 32(2) (2008) 297-312.
11. Verros G, Natsiavas S & Papadimitriou C, Journal of Vibration and Control, 11(5) (2005) 581-606.
12. Nagarkar M P, Patil G J V & Patil R N Z, Journal of advanced research, 7(6) (2016) 991-1007.
13. Kuznetsov A, Mammadov M, Sultan I & Hajilarov E, Journal of Sound and Vibration, 330(12) (2011) 2937-2946.
14. Gündoğdu Ö, International Journal of Industrial Ergonomics, 37(4) (2007) 327-332.
15. Georgiou G, Verros G & Natsiavas S, Vehicle System Dynamics, 45(1) (2007) 77-92.
16. Nakhaie Jazar G, Alkhatib R & Golnaraghi M F, Vehicle System Dynamics, 44(06) (2006) 477-512.
17. Kim C & Ro P I, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile
Engineering, 214(8) (2000) 851-864.
18. Kim C, Ro P I & Kim H, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile
Engineering, 213(5) (1999) 457-470.
19. Maher D & Young P, Vehicle system dynamics, 49(3) (2011) 463-480.
20. Zhu Q I N & Ishitobi M, Journal of Sound and Vibration, 275(3-5) (2004)1136-1146.
21. Zhu Q & Ishitobi M, International Journal of Solids and Structures, 43(3-4) (2006) 747-759.
22. Zhao D, Applied Mechanics and Materials, 110 (2012) 1150-1154.
23. Loprencipe G & Zoccali P, Journal of Modern Transportation, 25(1) (2017) 24-33.

24. Dormand J R & Prince P J, Journal of computational and applied mathematics, 6(1) (1980) 19-26.
25. ISO, Mechanical vibration and shock-evaluation of human exposure to whole-body vibration-Part 1: General
requirements. Geneva, Switzerland: ISO, (1997). 2631-1.
26. Sharp R S (1999, Pauwelussen, JP, Swets & Zeitlinger Publishers), Lisse, the Netherlands, 87-96.
27. Griffin M J, Vehicle System Dynamics, 45(7-8) (2007) 679-698.
28. Özcan D, Sönmez Ü & Güvenç L, International Journal of Vehicular Technology, 2013.
29. Molina-Cristobal A, Papageorgiou C, Parks G T, Smith M C & Clarkson P J, Multi-objective controller design:
evolutionary algorithms and bilinear matrix inequalities for a passive suspension, presented at In 13th IFAC Workshop
on Control Applications of Optimization, Cachan-Paris, France, 2006, 386-391.
30. Cossalter V, Doria A, Garbin S & Lot R, Vehicle System Dynamics, 44(4) (2006) 339-355.
31. Guiggiani M, In The Science of Vehicle Dynamics, Springer, Cham., (2018) 417-459).
32. Goldberg D E, Genetic algorithms. Pearson Education, 2006, India.
We thank the reviewers for their valuable comments which have significantly improved the quality of our manuscript.
We have tried to answer each and every comment given by the reviewer to the best of our efforts.
1. Revise the abstract and include results also.
The results are now included in the abstract.
2. Some of the grammatical errors are there like...write performed instead of
per-formed in abstract. Some where []) brackets are used. Use square brackets
[].
Grammatical errors in the abstract have been removed. Everywhere in the manuscript square brackets have been used.
3. In section 2 methodology, write Figure 2 instead of figFigure 2.
The required corrections have been made in section 2.
4. Write Figure 1 citation first in the text instead of Figure 2.
Figure 1 and 2 have been exchanged such that figure 1 citation now appears first in the text.
5. Error! Reference source not found.w???? Correct this.
This error has been corrected in section 3.

6. Figure 6 is missing.
Figure 6 has been added to the manuscript.