Differential Calculus

Introduction

Differential calculus provides a mathematical language for describing instantaneous change. In mechatronics engineering, this language is used to model velocity from position, determine actuator sensitivity, linearize nonlinear sensors, analyze control-system behavior, and optimize mechanical or electrical designs.

The central object is the derivative. Geometrically, it is the slope of a tangent to a graph. Physically, it is an instantaneous rate of change. Computationally, it is the limit of an average rate of change as the interval becomes arbitrarily small. These interpretations are equivalent, but each is useful in a different engineering context.

A function y = f(x) is differentiable at x = a when its derivative exists there. The derivative is defined by

Here, h is an increment in the independent variable. The quotient is an average rate of change over an interval of width h, while the limit gives the instantaneous rate at a. The definition applies when the indicated limit exists and is finite.

Learning objectives

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

1. Interpret derivatives as slopes, rates of change, and sensitivities.
2. Differentiate algebraic, trigonometric, exponential, logarithmic, inverse, implicit, and parametric functions.
3. Apply the product, quotient, and chain rules correctly.
4. Use derivatives for local linearization and engineering approximation.
5. Analyze stationary points, monotonicity, and basic optimization problems.
6. Recognize assumptions, units, domain restrictions, and limitations of derivative models.

1. Rates of change and the derivative

For a function y = f(x), the average rate of change from x = a to x = b is

The numerator and denominator must be expressed in compatible units. For example, if position is measured in metres and time in seconds, the average rate has units of metres per second.

The derivative function is

Alternative notations include y′, dy/dx, and D_x f. The notation dy/dx emphasizes that y changes with respect to x, but it should not automatically be treated as an ordinary fraction in every derivation.

Geometric interpretation

At a point on a curve, the tangent line is the limiting position of secant lines. If f′(a) > 0, the function is increasing locally; if f′(a) < 0, it is decreasing locally; and if f′(a) = 0, the tangent is horizontal. A zero derivative does not by itself prove that a maximum or minimum occurs.

Physical interpretation

If [[MATH{s(t)]] is position in metres as a function of time t in seconds, then

is velocity in metres per second, and

is acceleration in metres per second squared. More generally, a derivative measures sensitivity: dy/dx indicates the approximate change in output per unit change in input near a specified operating point.

Continuity and differentiability

A function is continuous at x = a if

Differentiability at a implies continuity at a, but continuity does not necessarily imply differentiability. The function f(x) = |x| is continuous at zero but has no derivative there because its left and right slopes are −1 and 1. Corners, cusps, vertical tangents, and discontinuities require special attention.

2. Fundamental differentiation rules

The following rules assume that the functions involved are differentiable on the domain under consideration.

Power, sum, and constant rules

For a polynomial

term-by-term differentiation gives

The constant term disappears because a constant has no rate of change.

Negative and fractional powers are included in the power rule, provided the function is defined on the relevant domain. For example,

and

The expression for the square-root derivative is not defined at x = 0 as an ordinary finite derivative.

Product rule

When two varying quantities are multiplied, differentiating each factor separately and multiplying is incorrect. The product rule is required.

For y = x² sin x,

The rule is especially important in mechanical models involving variable force and displacement, or electrical models involving products of time-varying signals.

Quotient rule

For

with x ≠ 2,

A quotient can sometimes be simplified algebraically before differentiating. Such simplification may reduce errors, but domain restrictions from the original expression must still be retained.

Chain rule

The chain rule differentiates a composition of functions. If y = f(u) and u = g(x), then

or equivalently (f∘g)′(x) = f′(g(x))g′(x).

For

let u = 3x² + 1. Then

The chain rule is central to multistage mechanisms, sensor calibration, and control models in which one physical variable depends on another through several transformations.

3. Trigonometric, exponential, and logarithmic functions

Angles in calculus formulas are measured in radians. If angles are measured in degrees, conversion or an additional factor is required. For example, the standard identity d(sin x)/dx = cos x assumes x is in radians.

Trigonometric derivatives

The last formula applies where cos x ≠ 0. The reciprocal functions satisfy

By the chain rule,

Exponential functions

For the natural exponential function,

More generally,

For a positive constant a,

where ln a is the natural logarithm of a. Thus, if y = 5^2x−1,

Logarithmic functions

For x > 0,

The chain rule gives

where g(x) > 0 on the real domain. For example,

For g(x) ≠ 0, logarithmic differentiation is often useful:

The absolute value extends the real formula to negative values of g(x), provided g(x) is nonzero.

4. Inverse functions and implicit differentiation

If a function has an inverse f⁻¹, then the derivative of the inverse is

This requires that the inverse exists locally and that the denominator is nonzero.

Important inverse-trigonometric derivatives include

and

The notation arcsin x means the inverse function of sine, not the reciprocal 1/sin x.

Implicit differentiation

An equation may define y in terms of x without explicitly isolating y. Differentiate both sides with respect to x, treating y as a function of x. Every derivative involving y therefore includes dy/dx.

For the circle

implicit differentiation gives

so

At the point (3,4), the tangent slope is −3/4. The formula is not applicable at the top and bottom points where y = 0; there the tangent is vertical.

For a more involved relation,

we obtain

and hence

provided x + 3y² ≠ 0.

5. Parametric differentiation

In a parametric description, both coordinates depend on a parameter t:

If dx/dt ≠ 0, then

This follows from the chain rule and is useful for robot trajectories, cam profiles, and motion paths.

For

we have

Therefore,

At t = 1, the point is (2,−1) and the slope is 1/2.

The second derivative for a parametric curve is not simply d²y/dt² ÷ d²x/dt². Instead,

provided the required derivatives exist and dx/dt ≠ 0.

6. Local linearization and differentials

A differentiable function is approximately linear near a chosen operating point x = a. Its linearization is

This is the tangent-line approximation. It is useful when a nonlinear sensor, actuator, or mechanical relation is operated over a sufficiently small range.

If ∆x is small, then

The differential notation is

where dx is a small input change and dy is the corresponding first-order output estimate. The approximation becomes more accurate as the input change approaches zero, assuming the function is differentiable.

Worked example: square-root approximation

Estimate √4.1 using linearization about a = 4. Let f(x) = √x. Then

Thus

At x = 4.1,

The actual value is approximately 2.0249, so the approximation is accurate to about 0.0001. The units are consistent: the derivative has units of output per input, and multiplying by an input change gives output units.

Engineering sensitivity

Suppose a sensor output is V = f(T), where T is temperature and V is voltage. The sensitivity at an operating point is

with units such as volts per kelvin. A first-order calibration model is

where V₀ = f(T₀). This model is local; using it far from T₀ can create significant error if the sensor is nonlinear.

7. Derivatives in analysis and design

Critical points and monotonicity

A critical number occurs where f′(x) = 0 or where f′(x) does not exist, provided x lies in the domain. To identify intervals of increase and decrease, examine the sign of f′.

For f(x) = x³ − 3x,

The critical points are x = −1 and x = 1. A sign chart shows that the function increases on (−∞,−1) and (1,∞), and decreases on (−1,1). Therefore, x = −1 is a local maximum and x = 1 is a local minimum.

The second-derivative test can assist classification. If f′(a) = 0 and f″(a) > 0, then a is a local minimum; if f″(a) < 0, it is a local maximum. If f″(a) = 0, the test is inconclusive.

Optimization

Engineering optimization commonly seeks a maximum performance or minimum cost subject to physical constraints. A reliable procedure is:

1. Define the design variable and feasible domain.
2. Express the objective as a function of that variable.
3. Find interior critical points and inspect endpoints.
4. Compare objective values and check physical feasibility.
5. Report units, assumptions, and sensitivity to parameters.

Related rates

In a related-rates problem, quantities depend on a common time variable. Differentiate the governing relation with respect to time before substituting numerical values.

For a circular actuator whose radius changes with time, area satisfies A = πr². Therefore,

If r = 0.20 m and dr/dt = 0.01 m/s, then

The units confirm the result: metres multiplied by metres per second gives square metres per second.

8. Applied case: linearizing a robotic position sensor

Background

The following is an illustrative engineering case. A rotary position sensor produces an output voltage modeled by

where θ is measured in radians and V is measured in volts. The sensor is used near the operating position θ₀ = π/6.

Problem

Estimate the voltage change caused by a small angular displacement of ∆θ = 0.02 rad, and determine the local sensitivity.

Analysis and solution

Differentiate the sensor model:

At θ₀ = π/6, since cos(π/6) = √3/2,

Thus the first-order voltage change is

The linearized sensor equation is

Since V(π/6) = 2 + 0.8(1/2) = 2.4 V, the predicted output at θ = π/6 + 0.02 is approximately 2.41386 V.

Outcome and interpretation

The sensitivity is approximately 0.693 V/rad at the operating point. Sensitivity changes with position because the cosine factor changes. Near θ = π/2, the first-order sensitivity is zero, so a linear voltage-based position estimate becomes poorly conditioned. This illustrates why calibration and operating-point selection matter in mechatronic systems.

Learning points

The derivative provides both a local model and a quantitative measure of sensor responsiveness. The approximation is valid for sufficiently small angular changes, uses radians, and ignores higher-order terms. A control engineer should therefore validate the linear model over the actual operating range rather than assume that one sensitivity applies globally.

9. Limitations, responsible use, and best practice

Differentiation is a model-based operation. Its result is meaningful only when the underlying function is appropriate and the variables are measured consistently. Real sensor data may contain noise, quantization, delays, and discontinuities. Numerical differentiation can amplify high-frequency noise, so filtering, smoothing, or model-based estimation may be required. Filtering itself can introduce phase delay and should not be applied without considering control stability.

A derivative at a mathematical point may not represent a useful physical rate if the system has finite sampling time or abrupt switching. Piecewise models require one-sided derivatives at boundaries. Domain restrictions must be preserved after algebraic manipulation. Trigonometric differentiation requires radian measure, and all engineering quantities must retain consistent units.

In software and AI-enabled engineering tools, symbolic algebra and automatic differentiation can reduce routine errors, but they do not replace interpretation. Engineers remain responsible for checking domains, singularities, units, numerical conditioning, and whether the computed derivative corresponds to the physical variable of interest. Derivatives used in optimization or machine learning should also be examined for sensitivity to noisy or biased data.

Chapter Summary

Differential calculus describes instantaneous change through the limit of an average rate of change. The derivative is simultaneously a tangent slope, a physical rate, and a local sensitivity. Standard rules include the constant, power, sum, product, quotient, and chain rules. Trigonometric, exponential, logarithmic, and inverse-function derivatives extend these rules to important engineering models.

Implicit differentiation handles relations in which the dependent variable is not isolated, while parametric differentiation gives dy/dx = (dy/dt)/(dx/dt). Local linearization uses L(x) = f(a) + f′(a)(x − a) to approximate nonlinear behavior near an operating point. Derivatives also support monotonicity analysis, optimization, related-rate calculations, sensor calibration, and robotic motion analysis. Correct application requires attention to domains, differentiability, units, radians, singularities, approximation error, and the limitations of mathematical and measured models.

Key Terms

Derivative: The limit of an average rate of change as the interval approaches zero.

Differentiability: The property of having a finite derivative at a point or throughout a domain.

Instantaneous rate of change: The derivative evaluated at a particular input.

Tangent line: The limiting line that touches a curve at a specified point.

Continuity: The property that the limiting function value equals the function value at a point.

Product rule: A rule for differentiating the product of two differentiable functions.

Quotient rule: A rule for differentiating the quotient of two differentiable functions.

Chain rule: A rule for differentiating a composite function.

Implicit differentiation: Differentiating a relation while treating the dependent variable as a function of the independent variable.

Parametric function: A function represented by coordinates depending on a common parameter.

Sensitivity: The rate at which an output changes with respect to an input.

Linearization: A first-order tangent-line approximation near an operating point.

Differential: A first-order change estimate such as dy = f′(x)dx.

Critical number: A domain value where the derivative is zero or does not exist.

Stationary point: A point where the derivative is zero.

Monotonicity: The increasing or decreasing behavior of a function over an interval.

Related rates: Rates of change of variables linked by a common equation and time variable.

Local approximation: An approximation intended to be accurate near a selected point.

Domain restriction: A limitation on allowable input values imposed by a formula or physical model.

Automatic differentiation: A computational method that evaluates derivatives by systematically applying the chain rule.

Review Questions

Short Answer Questions

1. Define the derivative using a limit and explain its geometric meaning.
2. State the product rule and the chain rule.
3. Why must the standard derivatives of sine and cosine use radians?
4. What is the difference between continuity and differentiability?
5. State the formula for the derivative of a function defined parametrically by x(t) and y(t).

Long Answer Questions

1. Derive the derivative of f(x) = x² from the limit definition and interpret the result geometrically.
2. Explain how implicit differentiation is applied to x² + y² = r², including the domain condition for the resulting slope formula.
3. Explain local linearization and discuss why it is useful in sensor calibration.
4. Compare explicit, implicit, and parametric representations of a curve, giving one engineering use for each.
5. Describe a complete derivative-based procedure for solving a constrained one-variable engineering optimization problem.

Analytical / Application Questions

1. A motor-shaft position is θ(t) = 0.5t³ − 2t² + 3t radians. Determine angular velocity and angular acceleration, and evaluate both at t = 2 s.
2. A mechanism follows x = 3 cos t and y = 2 sin t. Find dy/dx and determine the tangent slope at t = π/4.
3. A nonlinear sensor obeys V(T) = 1.5 + 0.02T² volts, with temperature T in degrees Celsius. Linearize the model at T = 10 °C and estimate the output change for a temperature increase of 0.5 °C.

Multiple Choice Questions with Answers

Question 1

What does f′(a) represent geometrically?

A. The area under the curve at a

B. The slope of the tangent line at a

C. The average value of the function over all inputs

D. The second derivative at a

Correct Answer: B. The slope of the tangent line at a

Question 2

If f(x) = 3x⁴ − 2x, what is f′(x)?

A. 12x³ − 2

B. 3x³ − 2

C. 12x⁴ − 2

D. 4x³ − 2

Correct Answer: A. 12x³ − 2

Question 3

What is the derivative of y = (2x + 1)³?

A. 3(2x + 1)²

B. 6(2x + 1)²

C. (2x + 1)²

D. 6(2x + 1)³

Correct Answer: B. 6(2x + 1)²

Question 4

For y = x² sin x, which derivative is correct?

A. 2x cos x + x² sin x

B. 2x sin x + x² cos x

C. 2x sin x + 2x cos x

D. x² cos x

Correct Answer: B. 2x sin x + x² cos x

Question 5

The derivative d(ln x)/dx = 1/x is valid over which real domain?

A. x < 0

B. x ≠ 0

C. x > 0

D. All real x

Correct Answer: C. x > 0

Question 6

For x² + y² = 25, what is dy/dx where y ≠ 0?

A. x/y

B. −y/x

C. −x/y

D. 2x/2y

Correct Answer: C. −x/y

Question 7

If x = t² + 1 and y = t³ − 2t, what is dy/dx at t = 1?

A. 1/2

B. 2

C. 1

D. −1/2

Correct Answer: A. 1/2

Question 8

What is the linearization of a differentiable function f at x = a?

A. L(x) = f′(a) + f(a)x

B. L(x) = f(a) + f′(a)(x − a)

C. L(x) = f(x) + f′(x)a

D. L(x) = f(a)(x − a)

Correct Answer: B. L(x) = f(a) + f′(a)(x − a)

Question 9

A position function is measured in metres and time in seconds. What are the units of its second derivative with respect to time?

A. Metres

B. Metres per second

C. Seconds per metre

D. Metres per second squared

Correct Answer: D. Metres per second squared

Question 10

For f(x) = x³ − 3x, which values are stationary points?

A. x = 0 only

B. x = −3 and x = 3

C. x = −1 and x = 1

D. x = 1/3 only

Correct Answer: C. x = −1 and x = 1