Functions and Engineering Representations

Introduction

Functions provide a mathematical language for describing how one engineering quantity depends on another. In mechatronics, a sensor converts displacement into voltage, a motor converts command input into angular speed, and a controller converts measured error into an actuator signal. Each relationship can be represented as a function, provided that every admissible input produces exactly one output.

A function is written as y=f(x), where x is the independent variable and y is the dependent variable. The notation does not imply that y is always physically determined by a single scalar; multivariable functions such as T=f(θ,I) are also common. This chapter focuses mainly on single-variable functions because they establish the foundations needed for calculus, modeling, control, and system analysis.

Learning Objectives

By the end of this chapter, you should be able to:

1. Define and determine the domain and range of a function.
2. Interpret polynomial, rational, exponential, logarithmic, piecewise, inverse, and composite functions.
3. Sketch and analyze graphs using transformations and parameters.
4. Select function models appropriate to engineering phenomena.
5. Evaluate functions, solve equations, and interpret units and physical restrictions.
6. Recognize modeling limitations and communicate assumptions clearly.

1. Functions, Variables, Domain, and Range

A relation is a collection of ordered pairs (x,y). A function is a relation in which each allowed input x corresponds to exactly one output y. The set of allowed inputs is the domain, and the set of resulting outputs is the range.

For example, the temperature of a component may be modeled by T(t), where t is time. If time is measured from the start of an experiment, the physical domain may be t≥0, even if an algebraic formula is defined for negative values.

The domain is determined by both mathematics and physics. Denominators cannot be zero, the argument of a real logarithm must be positive, and the radicand of an even root must be nonnegative. In addition, a model may restrict variables to nonnegative lengths, positive absolute temperatures, or a specified operating interval.

Function notation and evaluation

If f(x)=2x²-3x+1, then

The expression f(a+h) means that every occurrence of x is replaced by a+h. It is not generally equal to f(a)+h.

Vertical-line test

A graph represents a function of x if every vertical line intersects it at most once. This test is graphical; it does not determine whether a function has an inverse. A horizontal-line test is used to determine whether a function is one-to-one on a stated domain.

2. Graphs and Transformations

The graph of y=f(x) consists of all points (x,f(x)). Graphs reveal zeros, intercepts, increasing and decreasing intervals, extrema, asymptotes, and operating limits.

Important transformations include:

The order of transformations matters. For example, y=2(x−3)²+1 is obtained by shifting y=x² right by 3, stretching vertically by 2, and shifting upward by 1. Its vertex is (3,1), and its range is y≥1.

In engineering graphs, axes must include units. A graph of position versus time has slope with units of velocity, while a graph of voltage versus current has slope with units of resistance when the relationship is linear.

3. Polynomial Functions

A polynomial function has the form

where n is a nonnegative integer and the coefficients a_i are constants, with a_n≠0. The degree is n. Polynomial functions have domain all real numbers and are continuous everywhere.

The leading coefficient and degree determine end behavior. If n is even, both ends have the same direction; if n is odd, the ends have opposite directions. The sign of a_n determines whether the right-hand end rises or falls.

A zero or root is a value r for which p(r)=0. If

then r is a root of multiplicity m. A root of odd multiplicity generally causes the graph to cross the axis; a root of even multiplicity generally causes it to touch and turn. These statements assume the remaining factor is nonzero near the root.

Engineering interpretation

Polynomial models are useful for calibration, trajectory generation, interpolation, and approximation. For example, a cubic position trajectory can satisfy specified initial and final positions and velocities. However, extrapolating a polynomial far outside measured data can produce unrealistic values, so its valid operating interval must be stated.

Worked example: actuator calibration

Suppose a sensor calibration curve is

where d is displacement in millimetres and V is voltage in volts. At d=10\,mm,

The terms have consistent units only if the coefficient units are understood: the coefficient of d² has units V/mm², the coefficient of d has units V/mm, and the constant has units of volts. If the sensor is specified for 0≤d≤12\,mm, that interval—not the full mathematical domain—is the engineering domain.

4. Rational Functions

A rational function is a quotient of polynomials:

Its domain excludes values for which the denominator is zero. If a common factor cancels algebraically, the original excluded value remains a removable discontinuity, often called a hole.

Vertical asymptotes occur at certain denominator zeros that do not cancel. Horizontal or oblique asymptotes describe end behavior. For example,

has a vertical asymptote at x=3. Since numerator and denominator have equal degree, its horizontal asymptote is the ratio of leading coefficients, y=2.

Rational models occur in mechanisms and electronics. A simple steady-state gain model may be G(s)=K/(τs+1), where K is dimensionless gain and τ is a time constant, but such expressions are transfer functions rather than ordinary real-variable functions unless the context defines s. Rational curves can also model calibration saturation, flow resistance, or geometric ratios.

5. Exponential Functions

An exponential function has a variable in the exponent:

Here A is the initial value when t=0, and b is the multiplication factor per unit of t. If b>1, the function grows; if 0<b<1, it decays. The domain is all real numbers and the range is positive when A>0.

A continuous-time model is

where x₀ is the value at t=0, k has units reciprocal to the time unit, and e≈2.71828. The model assumes a constant proportional rate of change:

For decay, k<0. A time constant model is x(t)=x₀e^−t/τ, where τ>0; after one time constant, the remaining fraction is e⁻¹≈0.368.

Example: capacitor charging

For a first-order charging circuit,

where V_s is the source voltage, R is resistance in ohms, C is capacitance in farads, and τ is in seconds. This model assumes an ideal linear resistor and capacitor, a constant source, and zero initial capacitor voltage. At t=τ,

The product RC has units Ω·F=s, confirming dimensional consistency.

6. Logarithmic Functions

The logarithm is the inverse of exponentiation:

The natural logarithm is ln(x)=log_e(x). Its domain is x>0, and its range is all real numbers. Important laws are

These laws require positive arguments in the real-number setting. The change-of-base formula is

Logarithms are useful for decibel scales, pH, signal processing, and solving exponential equations. For example, solving V_s(1−e^−t/τ)=0.9V_s gives

provided V_s≠0 and τ>0.

7. Piecewise Functions and Engineering Constraints

A piecewise function uses different formulas on different parts of its domain:

Piecewise models represent switching, saturation, dead zones, backlash, operating modes, and safety limits. At a boundary, continuity can be tested by comparing the left-hand limit, right-hand limit, and function value. A function is continuous at x=a if these are equal.

A saturation model is

where r is an unsaturated command and u_max>0 is the actuator limit. This model is not differentiable at the switching points, although it is continuous. Such nonsmoothness can affect controller tuning and numerical simulation.

8. Inverse Functions

A function f has an inverse f⁻¹ on a specified domain if it is one-to-one. The defining relationships are

and

for inputs in the appropriate domains. The domain of the inverse is the range of the original function, and its range is the original domain. The graphs of a function and its inverse are reflections in y=x.

Consider f(x)=x². It is not one-to-one on all real numbers because f(2)=f(−2)=4. Restricting the domain to x≥0 gives the inverse f⁻¹(x)=√x. Domain restrictions are therefore essential, not optional.

In engineering, an inverse model may convert sensor output to physical quantity. If a sensor law is V=2d+0.5, then the inverse calibration is d=(V−0.5)/2, with units carefully retained. In practice, noise, hysteresis, quantization, and non-monotonic behavior may prevent a unique inverse.

9. Composite Functions

If f and g are functions, their composition is

The output of g becomes the input of f. The domain consists of values in the domain of g whose outputs lie in the domain of f. Composition is generally not commutative: f∘g and g∘f need not be equal.

Worked example: sensor and actuator chain

Let a displacement sensor produce voltage

and let an actuator command be

where d is in millimetres, v is in volts, and u is in the specified command unit. The complete input-output relationship is

At d=5\,mm, v=0.7\,V and u=3(0.7)²=1.47. The reverse composition (v∘u)(d) represents a different physical sequence and is not an alternative calculation of the same system.

10. Choosing and Evaluating Engineering Models

Model selection should begin with the physical mechanism, measured data, and intended use. A linear model is interpretable and often adequate over a narrow operating range, whereas an exponential model is appropriate when the rate is proportional to the current state. A polynomial may fit calibration data but can behave poorly outside the calibration interval. A piecewise model is appropriate when hardware changes mode.

A disciplined modeling workflow is:

1. Define inputs, outputs, units, and the operating interval.
2. Identify physical restrictions and expected qualitative behavior.
3. Select a candidate function family.
4. Estimate parameters from theory or data.
5. Check dimensions, limiting cases, and boundary behavior.
6. Compare predictions with measurements.
7. State uncertainty, assumptions, and validity limits.

Parameters are not merely curve-fitting numbers. For example, in x(t)=x₀e^−t/τ, x₀ describes an initial condition and τ describes a dynamic time scale. Parameter changes should have an interpretable physical meaning.

Applied Case: Modeling a Conveyor-Based Sorting Cell

Background

Consider an illustrative mechatronics cell in which a conveyor transports parts past a sensor and a pneumatic actuator. The controller must estimate part position and issue a timed command.

Problem

The conveyor speed is approximately constant at v=0.40\,m/s. A sensor is located 0.60\,m before the actuator. The sensor has a calibration law

where x is distance in metres and V is voltage. The controller accepts a normalized command based on the sensor voltage, but commands are limited to the interval 0≤u≤1.

Analysis and approach

The travel time from the sensor to the actuator is modeled by

where L is distance and v is speed. Thus,

If the command law is u=2(V−0.10), then composition with the sensor model gives

before saturation. The implemented command is therefore the piecewise function

The model is valid only over the calibrated range and assumes constant conveyor speed, negligible sensor delay, and no actuator response time. If the actuator has response delay t_a, the command must be scheduled earlier, approximately at t_d−t_a, provided this quantity is nonnegative.

Outcome and learning points

The analysis combines a linear calibration, composition, a piecewise saturation function, and a time-of-flight calculation. It also demonstrates why units and timing assumptions matter. A controller designed from the unsaturated equation alone could request an impossible command. In a real system, speed variation would require encoder feedback or a more sophisticated position estimate.

Limitations, Safety, and Contemporary Practice

Functions are simplified representations. Measurement noise means that a fitted function is not exact; parameter estimates should be accompanied by uncertainty or error measures. A model may be structurally wrong even when it fits existing data. Extrapolation is especially risky for polynomials and empirical calibrations.

In mechatronics, function models also support digital twins, simulation, and data-driven control. Machine-learning models may approximate complicated functions, but they do not remove the need for domain restrictions, validation, explainability, and safety limits. An AI-based estimator should be tested across operating conditions, monitored for distribution changes, and prevented from commanding unsafe states. Physical interlocks and verified limits must not be replaced solely by a learned function.

Good engineering practice requires version-controlled calibration data, traceable units, documented parameter estimates, test cases at boundaries, and independent verification of safety-critical behavior. When a model informs a decision affecting people or equipment, its assumptions and failure modes must be communicated clearly.

Chapter Summary

Functions describe input-output relationships and provide the foundation for mathematical engineering models. Domain and range must be determined from both algebraic restrictions and physical operating conditions. Polynomial functions are continuous and useful for calibration and trajectories; rational functions describe ratios and may contain discontinuities or asymptotes. Exponential functions model proportional growth or decay, while logarithms invert exponentials and support scale conversion and parameter estimation. Piecewise functions represent switching, saturation, and operating limits. Inverse functions recover inputs only when the original function is one-to-one on an appropriate domain. Composite functions model cascaded sensors, actuators, and subsystems.

Graph transformations reveal how parameters affect behavior. Reliable engineering representation requires dimensional consistency, meaningful parameter interpretation, boundary checks, validation against measurements, and explicit limitations. These skills support later study of limits, differentiation, integration, differential equations, control, and system modeling.

Key Terms

Function: A relation assigning exactly one output to each allowed input.

Domain: The set of permitted input values.

Range: The set of output values produced by a function.

Independent variable: The input variable selected or controlled in a model.

Dependent variable: The output variable determined by the input.

Polynomial function: A finite sum of nonnegative integer powers of a variable.

Degree: The highest power with a nonzero coefficient in a polynomial.

Root: An input at which a function equals zero.

Rational function: A quotient of two polynomials with a nonzero denominator.

Asymptote: A line or curve that describes limiting graph behavior.

Exponential function: A function in which the variable occurs in an exponent.

Logarithm: The inverse operation of exponentiation.

Piecewise function: A function defined by different formulas on specified subdomains.

Inverse function: A function that reverses a one-to-one function on stated domains.

Composition: The function formed by applying one function to the output of another.

Parameter: A constant that determines a model’s behavior or physical characteristics.

Saturation: Limitation of an output to a prescribed maximum or minimum.

Calibration: Establishing a quantitative relationship between a measured signal and a physical quantity.

Model domain: The operating interval over which a mathematical representation is considered valid.

Review Questions

Short Answer Questions

1. Define the domain and range of a function and explain why physical restrictions may reduce an algebraic domain.
2. State the domain restriction for a real logarithm and for a rational function.
3. Distinguish exponential growth from exponential decay using the sign of the continuous-time parameter k.
4. Why must a function be one-to-one on a stated domain before it can have an inverse there?
5. Explain the meaning of the composite function (f∘g)(x).

Long Answer Questions

1. Describe the principal graph transformations and illustrate how they affect y=f(x).
2. Compare polynomial, rational, exponential, and piecewise models in terms of domain, continuity, and engineering applications.
3. Derive the time required for a first-order charging model to reach a specified fraction of its final value.
4. Explain how calibration, units, and operating limits should be incorporated into an engineering function model.
5. Discuss why extrapolation can be dangerous even when a function fits measured data accurately.

Analytical / Application Questions

1. A sensor obeys V(d)=0.04d²+0.2d+0.5 for 0≤d≤10\,mm. Calculate the voltage at d=6\,mm, identify the units required for each coefficient, and explain why the stated interval matters.
2. An actuator command is u(r)=min(1,max(0,2r)). Express this as a piecewise function and determine the command for r=−0.2, r=0.3, and r=0.8.
3. A temperature difference follows θ(t)=θ₀e^−t/τ, with θ₀=80\,°C and τ=12\,s. Determine the time at which the difference falls to 10\,°C and explain the assumptions behind the calculation.

Multiple Choice Questions with Answers

Question 1

Which condition is required for a real rational function p(x)/q(x)?

A. p(x) must never be zero

B. q(x) must be zero for at least one input

C. q(x)≠0 for every input in the domain

D. The numerator and denominator must have equal degree

Correct Answer: C. q(x)≠0 for every input in the domain

Question 2

What is the range of f(x)=x²+3 over the real numbers?

A. y≤3

B. y≥3

C. All real numbers

D. y>3

Correct Answer: B. y≥3

Question 3

For x(t)=x₀e^−t/τ with τ>0, what fraction of x₀ remains at t=τ?

A. Approximately 0.135

B. Approximately 0.368

C. Approximately 0.632

D. Approximately 1.000

Correct Answer: B. Approximately 0.368

Question 4

Which transformation changes y=f(x) into y=f(x−4)?

A. Shift left by 4

B. Shift right by 4

C. Shift upward by 4

D. Reflect in the y-axis

Correct Answer: B. Shift right by 4

Question 5

What is log₁₀(1000)?

A. 1

B. 2

C. 3

D. 10

Correct Answer: C. 3

Question 6

Why is f(x)=x² not invertible on all real numbers?

A. It has no range

B. It is not one-to-one

C. It is discontinuous

D. Its domain is empty

Correct Answer: B. It is not one-to-one

Question 7

If f(x)=2x+1 and g(x)=x², what is (f∘g)(3)?

A. 7

B. 19

C. 37

D. 49

Correct Answer: B. 19

Question 8

A saturation function limits a command to which kind of range?

A. An unbounded range

B. A prescribed minimum and maximum

C. Only negative values

D. Only irrational values

Correct Answer: B. A prescribed minimum and maximum

Question 9

For R(x)=(2x+1)/(x−3), what is the vertical asymptote?

A. x=−1/2

B. x=0

C. x=2

D. x=3

Correct Answer: D. x=3

Question 10

A sensor has V(d)=0.5d+0.2. What displacement corresponds to V=2.2\,V?

A. 1\,mm

B. 2\,mm

C. 4\,mm

D. 5\,mm

Correct Answer: C. 4\,mm