METHODS IA1

1. Introduction
The Steel Dragon 2000, located in Nagashima, Japan takes the crown for the longest roller coaster in
the entire world. Sitting at 1.5 miles long, the passenger has an enthralling 4 minutes of experiencing
different forms of acceleration. Using inspiration from the Steel Dragon 2000, Dreamworld, located
in Gold Coast of Queensland, Australia, plans to introduce a new roller coaster to institute a new
level of excitement. The objective of this report is to develop a roller coaster for Dreamworld to
grant passengers high levels of adrenaline. The report will create a mathematical model to represent
the roller coaster’s 2D outline (not accounting for any external tilt, invert or veer left or right) and
will be consist of four different functions (exponential base e, logarithmic, quadratic, cubic, etc.). The
model will also comply with further specifications provided by Dreamworld to fit their design
criteria.

2. Considerations
2.1 Observations
 It was observed that approximately 80% of Dreamworld’s visitors come during the
summer months of Australia (Similarweb, 2024), therefore must account for a
design that takes cooling as a factor. Consequently, the mathematical model will
ensure that around 60% of the track consists of somewhat steep descents
(elaborated further in assumptions) to maximise airflow thus exemplifying rate of
cooling for riders.
 It was observed that the Gold Coast’s average summer temperature reaches around
28-30° C, with humidity levels consistently overtaking 70% (Gold Coast Weather,
2024). Therefore, the mathematical model will justify cooling via steep descents as
well as constant changes in steepness via the changing of gradients to increase
adrenaline as well as cooling

2.2 Assumptions
 It was assumed that the roller coaster would be built on the start platform (5 units above the
x-axis) and will therefore be attached to two poles perpendicular to the ground (represented
≈ 335.5 units below y-axis).
 It was assumed that the start platform is built 338.5m above the ground and will therefore
be on the 2nd level of the roller coaster venue.
 It was assumed that the start platform extends 4m before starting the first phase of the
roller coaster ride.
 It was assumed that the final brake run extends 50m at the ground.
 Another assumption is that once the car finished the track, there would be brakes on the
track that would slowly stop the car.
 It was assumed that the roller coaster will not, at any point, exceed a steepness of 75° , as
constrained by Dreamworld’s design conditions.
 It was assumed that the roller coaster will steep below the starting platform twice as
instructed by Dreamworld’s design conditions.
 It was assumed that the roller coaster does not account for any external tilt, invert or veer
left or right.
 2.3 Mathematical Concepts and Techniques
Every little roller coaster segment will be precisely calculated by using differential calculus, a
mathematical technique that helps determine the derivative of functions to detect the gradient at
any given point. To represent the profile of the roller coaster, a piecewise function will be
constructed on Desmos. Once the model of the graph has been plotted, the adrenaline factor of the
roller coaster will be calculated via the steepness-factor of each segment of the roller coaster. The
mathematical model will ensure that this value does not exceed 75° . Differentiation will be used to
find the steepest gradient of the curve. To construct the actual path of the roller coaster, regression,
sliders, domain and range will be used. Simultaneous equations will be used to ensure that the path
may follow 2 specific points or to equate a segment to match two other segments. Of course,
function notation and best practices will be demonstrated in the making of the path.

3. Designing a Roller Coaster

3.1 Piece-wise function and Functional Notation
A piece-wise function is a function consisting of multiple different functions partitioned by multiple
domains.

What follows is the presentation of a design where the roller coaster will model in a piecewise
function, hence incorporating exponential functions in base e, logarithmic, quadratic, and cubic in
nature for various sections of the track.

Each section will take on a specific domain to ensure that continuity and smoothness are well taken
care of in ensuring that the ride is seamless. The functions will then be chosen and adapted
according to given constraints that say no steeper than 75° and have maximum height of 100m by
Dreamworld.

Graphing utilities, such as Desmos, will be used to visualise and verify the overall design to ensure
sections of the model join smoothly. The final product should be a coherent 2D model that fulfills
safety and excitement criteria as briefed by Dreamworld.

3.2 Steepness Factor

An individual steepness-factor is the value of the angle (degrees) of the steepest point of a segment.
The derivative of a function determines the rate of change of said function. This can be used to
calculate the steepness-factor of the function.

To determine this, 2 major steps were taken:

Movable Point within Segment

To emulate the steepness factor practically, a point was assigned to each segment of the graph and
by using sliders (coming up), the point could be used to determine the steepness of any part of the
segment.
Figure 1: Derivative function (dotted) of segment function (solid and above)

This point consisted of the x and y coordinates of the derivative function of segment 3. This is
because the report is not trying to monitor the function of segment 3 itself, but rather the rate of
change of the function of segment 3. The slider is utilised to move the point through the derivative
function of segment 3.

Figure 2: Movable point within segment 3

Finding the Angle
Finding the angle is straightforward as the report utilised the tan−1 function for the derivative of the
function to evaluate the steepness in radians. Generally (in the case of the roller coaster) the tan−1
dy
−1 ' dy
follows the format tan ( ), but since the derivative of a function f ( x )= computes the change
dx dx
−1
already, all that needs to be done is a simple tan (f ' (x))

Figure 3

180
Converting to degrees was just a simple product of the result and . Note how the value is a
π
negative due to a depreciating slope. This does not matter however as long as the absolute value
does not exceed 75 ° .

3.2 Non-Linear Regression and matching gradients

Regression is a statistical technique that relates a dependent variable to one or more independent
variables. A regression model is able to show whether changes observed in the dependent variable
are associated with changes in one or more of the independent variables. In this case the dependant
variable is simply f (x) and the independent variable simply x . It does this by essentially determining
a best-fit line and seeing how the data is dispersed around this line (Beers & Rubin, n.d.).

In order to ensure that each segment of the roller coaster joined up smoothly, the report employed
regression in order to construct the formulas itself. Specifically, since report wants to calculate the
best fit “curve” instead of line, non-linear regression is employed. This is due to the higher-degree
polynomial properties of the curve.

Furthermore, as two segments of the graph have to connect smoothly, its gradients have to be the
same. In Desmos, this was done simply by feeding in values of two extremely close points to a
regression model and it calculating a best-fit line. For example, suppose the report is trying to
calculate f 04 using linear regression and ensuring it connects smoothly with f 03, it would:

Figure 4: context of previous function

 Figure 5: Regression methods

 Calculate the y values at the end of the domain of the previous function
 Plot the points on a regression table
 Based on the type of formula the report wants the next segment to be, the formula is
written as standard form
 The table is referenced to the formula via the notation of ~ and the appropriate x and y
values and Desmos automatically plots the curve of best fit, which happens to
 
Figure 6: segment 4 connects with segment 3 smoothly

Figure 7:segment 4 function

3.3 Simultaneous Equations

Simultaneous equations are equations involving two or more unknowns that are to have the same
values in each equation (Oxford Languages and Google - English | Oxford Languages, n.d.).

Simultaneous equations were employed in the case of the roller coaster more so to check that the
gradients of two functions are the same value (gradient matching) AND to also check whether the
functions actually intersect (continuity). For example:

Making sure that f 02∧f 03 join smoothly:

Given the functions:

2
f 02( x)=−0.0373832 x +0.669782 x +0 {9 ≤ x ≤15 }

f 03 (x)=−1.3876 × 10 ex +2.08911 {14.99 ≤ x ≤ 16.8 }

−7

Since both functions intersect at x=15 based on the domain, that is going to be the primary x
value.

Continuity check:

f 02 ( x )=f 03 (x)
2
f 02 ( 15 ) =−0.0373832⋅15 + 0.669782⋅15 ≈ 1.63551
 −7 15
f 03 ( 15 ) =−1.3876 ⋅10 ⋅e +2.08911 ≈1.63550
The values are not the exact same, meaning, they do not intersect 100% perfect, but this is negligible
due to the extremely small margin of error.

Gradient matching check:

' '
f 02 ( x )=f 03 (x)
'
f 02 ( 15 ) =−0.0747664 ⋅15+0.669782 ≈−0.45171
' −7 15
f 03 ( 15 ) =−1.3876 ⋅10 ⋅e ≈−0.45361
Again, the values do not equate to be 100% perfect, but this is negligible due to the small margin of
error.

Therefore, it can be concluded that f 02∧f 03 join smoothly via the verification of the continuity and
gradient matching check.

3.4 Sliders
Sliders were used mainly hand-in-hand with piecewise functions as well as determining the
steepness. The sliders can have different values for increments, minimum or maximum values, but
for all functions, the increment was set to 0.1 and minimum and maximum values were the
respective domains of the functions. In the case of finding the angle of steepness, the sliders were
assigned to the points in order to ensure the steepest point of the segment does not exceed 75 ° .

Piecewise

Steepness

3.4 Domain
. The domain restricts the x value of the functions, thus goes hand in hand in the making of
piecewise functions. The domain goes by the format a ≤ x ≤ b where a∧b are constants, thus
limiting the function to only a certain area of the graph. The domain helps greatly with determining
the continuity and gradient matching of two functions via simultaneous equations and also with
non-linear regression as the closer to the domain the points are, the much more accurate and
smooth the resultant function would be.
Domain and its application with non-linear regression

Figure 8: Domain: non-linear regression

Figure 9: domain of said function

Domain and its application in making piece-wise functions

Figure 10: domain and piecewise functions

3.4 All Functions and ensuring steepness (will finish up

later)
Function Steepest point Steepness of Function (° )
f 0 ( x )=5 {−4 ≤ x ≤ 0 } (0,5) 0

( )π
2 sin x +5 {0 ≤ x ≤ 9 }
6

4. Evaluate, Verify and Improve

4.1 Evaluating and Verifying
The roller coaster design was evaluated to ensure it met the Dreamworld constraints and
requirements. Each segment was modelled as part of a piecewise function and has successfully
ticked off the criteria regarding the four different functions being exponential, quadratic, logarithmic
and cubic functions, which were all confirmed continuously and smoothly. The functions were made
using non-linear regression by plotting two points from the previous function extremely close with
one another such that the regression model calculates a continuous and smooth function of the
report’s format. Using differential calculus, gradients were verified by checking the derivatives at
transition points with simultaneous equations. Usage of domain was used to ensure continuity and
gradient matching as part of the simultaneous equations and also for non-linear regression as the
closer the point, the more accurate the model. While minor discrepancies were noted, they were
negligible and confirmed smooth transitions. Steepness under 75 ° was verified using the
implementation of the steepness factor and the usage of the derivatives of the plotted segment’s
functions. As for steepness, sliders were used to check steepness dynamically and ensuring that the
design was constrained to under 75 ° following Dreamworld’s requirements.
 4.2 Strengths and Limitations
The design adhered to Dreamworld’s requirements and ensured smooth transitions between
segments and ensures the safety of all passengers. Maximisation of cooling and airflow was
employed via the steep ascends and even steeper descends without compromising safety
restrictions( ¿ 75 °). The curves were precise and continuous and did not concede any major gaps
within the 2D roller coaster design. As for safety, the report calculated the steepest point and then
used the steepness factor in order to find the maximum steepness angle and ensuring that it was
always below 75 ° , furthermore confirming passenger safety.

However, limitations include minor discrepancies in gradient matching, and although the difference
is negligible, still can be refined further. The use of the Desmos graphing tool meant the reports
design is based mostly on its accuracy and limit of precision of floating-point numbers because
Desmos does not go further than 17 significant digits of precision. Therefore, this could cause a slight
error when calculating, which may have also been the reason for gradient matching discrepancies.

5. Conclusion.
6. Reference List
7. Appendices
7.1 Segment 0 and 11 – Linear Functions

f 0 ( x )=5 {−4 ≤ x ≤ 0 }

Figure 11: Perusal period at start platform level

f 11 ( x ) =−335.47249 {500 ≤ x ≤550 }

Figure 12: Final brake run at ground level

The start and the end of the roller coaster track was chosen to be a linear function. This is for the
passenger to better understand when the first and last phases of the roller coaster are. The ride
starts at the flat track at y=5 (start platform) and the linear function allows for a 4-metre perusal
period before the ride begins to build up suspense while ensuring it intersects with the next function
for a continuous ride. Another linear function was placed after the final descent to signal the
passengers the end of the ride. (See calculations of intersection points in Appendices)

Segment 1 – Sine Function

A sin [ B ( x +c ) ] + D 2 sin ( π6 x)+5 {0 ≤ x ≤ 9 }

The function represents the initial ascent of the roller coaster. This sine function creates a gradual
suspense up to 2m at y=7 above the starting platform and descends 2m below at y = 3m, giving the
passengers an initial feel for the ride before the big segments. The amplitude was tweaked via the A
variable and finally was set to 2 for the reasons given before. To give it a shorter period, the report
π
tweaked the period value to and there was no horizontal shift thus the C=0. The report had set
6
D=5 as the starting platform is set at y=5 . This function did not use regression.
Sample Calculations
Steepest point of cosine function:

Figure 13: Derivative of f_01: Cosine function's steepest point

Figure 14: calculating the steepness of the steepest point