PracticeBack to top

Pomodoro

Pomodoro timer is idle

Second Order ODEs

GitHub Changelog

Suggest edit Practice

1. What counts as a second-order ODE

A second-order ordinary differential equation is an equation involving an unknown function \(y(x)\) and its derivatives up to order two:

$$ F(x, y, y', y'') = 0 $$

The order matters because it tells you how many independent initial conditions are typically needed to determine a solution.

Second-order ODEs show up naturally in:

  • Newtonian mechanics

  • vibration and wave models

  • electrical circuits

  • beam deflection and structural models

  • heat and diffusion problems after separation of variables

The central tasks are usually:

  • solve the equation exactly when possible

  • classify the motion or response

  • apply initial or boundary conditions

  • interpret the physical behavior

2. Classification and standard forms

Linear vs nonlinear

A second-order ODE is linear if it can be written as

$$ a_2(x) y'' + a_1(x) y' + a_0(x) y = g(x) $$

where \(y\), \(y'\), and \(y''\) appear only to the first power and are not multiplied together.

It is nonlinear if, for example, it contains terms such as:

$$ y^2,\quad y y',\quad \sin(y),\quad (y')^2 $$

Homogeneous vs nonhomogeneous

The linear equation is homogeneous when

$$ g(x) = 0 $$

and nonhomogeneous when \(g(x) \neq 0\).

Common standard forms

General linear form

$$ a_2(x) y'' + a_1(x) y' + a_0(x) y = g(x) $$

Normalized form

If \(a_2(x) \neq 0\), divide through by \(a_2(x)\):

$$ y'' + p(x) y' + q(x) y = r(x) $$

This form is often best for theory and method selection.

Constant-coefficient form

$$ ay'' + by' + cy = g(x) $$

with constants \(a \neq 0\), \(b\), and \(c\).

3. Existence, uniqueness, and IVPs

For the initial value problem

$$ y'' = f(x, y, y'),\qquad y(x_0) = y_0,\qquad y'(x_0) = v_0 $$

the existence and uniqueness theorem says that if \(f\) and its relevant partial derivatives are continuous near \((x_0, y_0, v_0)\), then a unique local solution exists.

For linear equations in normal form

$$ y'' + p(x) y' + q(x) y = r(x), $$

continuity of \(p(x)\), \(q(x)\), and \(r(x)\) on an interval is enough to guarantee a unique solution for any specified pair of initial conditions in that interval.

Why two initial conditions?

Integrating a second-order equation effectively introduces two constants of integration. That is why you usually need:

$$ y(x_0) = y_0,\qquad y'(x_0) = v_0 $$

to pin down one specific solution.

4. Homogeneous linear equations with constant coefficients

The foundational case is

$$ ay'' + by' + cy = 0 $$

Try a solution of the form \(y = e^{rx}\). Substitution gives the characteristic equation

$$ ar^2 + br + c = 0 $$

The roots of this quadratic determine the solution form.

Case 1: two distinct real roots

If \(r_1 \neq r_2\), then

$$ y(x) = C_1 e^{r_1 x} + C_2 e^{r_2 x} $$

Case 2: repeated real root

If \(r_1 = r_2 = r\), then

$$ y(x) = (C_1 + C_2 x)e^{rx} $$

The extra factor of \(x\) gives a second independent solution.

Case 3: complex conjugate roots

If

$$ r = \alpha \pm i\beta $$

then

$$ y(x) = e^{\alpha x}\bigl(C_1 \cos(\beta x) + C_2 \sin(\beta x)\bigr) $$

This is the most common form for oscillatory motion.

Example

Solve

$$ y'' - 3y' + 2y = 0 $$

Characteristic equation:

$$ r^2 - 3r + 2 = 0 = (r-1)(r-2) $$

So

$$ y(x) = C_1 e^x + C_2 e^{2x} $$

5. Forcing and particular solutions

For a nonhomogeneous linear equation

$$ ay'' + by' + cy = g(x), $$

the general solution has the form

$$ y = y_h + y_p $$

where:

  • \(y_h\) solves the homogeneous equation

  • \(y_p\) is any one particular solution of the full equation

The structure of \(g(x)\) usually determines the method.

Undetermined coefficients

Use this when \(g(x)\) is built from exponentials, polynomials, sines, cosines, or products of these.

Typical guesses:

Forcing \(g(x)\)Trial form for \(y_p\)
\(e^{ax}\)\(Ae^{ax}\)
\(P_n(x)\)polynomial of degree \(n\)
\(\sin bx\) or \(\cos bx\)\(A\cos bx + B\sin bx\)
\(e^{ax}P_n(x)\)\(e^{ax}\) times a polynomial of degree \(n\)
\(e^{ax}\sin bx\), \(e^{ax}\cos bx\)\(e^{ax}(A\cos bx + B\sin bx)\)

Resonance rule

If your trial solution duplicates part of \(y_h\), multiply the trial by \(x\) enough times to make it linearly independent.

Example of resonance

Solve

$$ y'' - 2y' + y = e^x $$

Since the homogeneous equation has characteristic \((r-1)^2 = 0\), the homogeneous solution already contains \(e^x\) and \(xe^x\). A correct trial is

$$ y_p = Ax^2 e^x $$

because the repeated root forces two powers of \(x\).

Variation of parameters

Use this when undetermined coefficients is not convenient.

If \(y_1\) and \(y_2\) are independent solutions of the homogeneous equation

$$ y'' + p(x)y' + q(x)y = r(x), $$

then a particular solution can be built as

$$ y_p = u_1(x)y_1(x) + u_2(x)y_2(x) $$

where \(u_1\) and \(u_2\) are determined from a system involving the Wronskian.

6. Variable-coefficient special cases

Euler-Cauchy equations

The Euler-Cauchy, or equidimensional, form is

$$ x^2 y'' + a x y' + b y = 0 $$

Try \(y = x^m\). Then:

$$ y' = mx^{m-1},\qquad y'' = m(m-1)x^{m-2} $$

Substitution gives the auxiliary equation

$$ m(m-1) + am + b = 0 $$

So the same root patterns appear as in constant-coefficient problems.

Reduction by substitution

Sometimes a change of variables converts a variable-coefficient equation into a constant-coefficient one. For Euler-Cauchy equations, the substitution

$$ x = e^t $$

is a common route.

7. Reduction of order and variation of parameters

Reduction of order

If one nonzero solution \(y_1\) of the homogeneous equation is known, try

$$ y = v(x)y_1(x) $$

to reduce the order by one. After substitution, the equation for \(v\) is lower order and can often be integrated.

This is especially useful when:

  • one solution is obvious

  • the equation is linear but not constant-coefficient

  • you are given one solution and asked to find the general solution

Wronskian

The Wronskian of two functions \(y_1\) and \(y_2\) is

$$ W(y_1,y_2) = \begin{vmatrix} y_1 & y_2 \\ y_1' & y_2' \end{vmatrix} = y_1 y_2' - y_1' y_2 $$

If \(W \neq 0\) on an interval, the functions are linearly independent there.

For

$$ y'' + p(x)y' + q(x)y = r(x), $$

variation of parameters uses the Wronskian to build \(u_1\) and \(u_2\).

8. Initial and boundary value problems

Initial value problems

An IVP specifies values at a single point:

$$ y(x_0) = y_0,\qquad y'(x_0) = v_0 $$

Steps:

  1. Solve the differential equation.

  2. Differentiate if needed.

  3. Apply the initial conditions.

  4. Solve for the constants.

Boundary value problems

A BVP specifies conditions at different points, for example:

$$ y(0) = 0,\qquad y(L) = 0 $$

These occur frequently in:

  • beam bending

  • heat conduction

  • vibrating strings

  • Sturm-Liouville problems

Boundary conditions can sometimes produce:

  • a unique solution

  • no solution

  • infinitely many solutions

depending on the forcing and the eigenstructure of the operator.

9. Mechanical applications

Many second-order ODEs come from Newton’s second law:

$$ m y'' = \sum F $$

Mass-spring system

With no damping and no external forcing:

$$ m y'' + k y = 0 $$

The solution is harmonic:

$$ y(x) = C_1 \cos(\omega x) + C_2 \sin(\omega x), \qquad \omega = \sqrt{\frac{k}{m}} $$

Damped vibration

With viscous damping:

$$ m y'' + c y' + k y = 0 $$

Divide by \(m\):

$$ y'' + 2\zeta \omega_n y' + \omega_n^2 y = 0 $$

where

$$ \omega_n = \sqrt{\frac{k}{m}},\qquad \zeta = \frac{c}{2\sqrt{mk}} $$

Damping regimes

RegimeConditionBehavior
Underdamped\(\zeta < 1\)Oscillatory with decaying amplitude
Critically damped\(\zeta = 1\)Fastest return without oscillation
Overdamped\(\zeta > 1\)Non-oscillatory, slower return

Interactive visual

Damped oscillator

Tweak damping and frequency to see how the oscillation fades and tightens over time.

Period 3.14
Decay slow

Forced vibration

With external forcing \(F(t)\):

$$ m y'' + c y' + k y = F(t) $$

The solution is usually split into:

$$ y(t) = y_{\text{transient}}(t) + y_{\text{steady}}(t) $$

If \(F(t)\) is periodic, resonance and near-resonance can produce large response amplitudes.

10. Qualitative behavior

Even when an exact formula is available, it helps to understand the shape of the solution.

Stability of equilibrium

For the autonomous equation

$$ y'' + f(y, y') = 0 $$

equilibria correspond to constant solutions. Linearization near equilibrium often predicts local behavior.

Phase-plane viewpoint

Rewrite the second-order ODE as a first-order system:

$$ u = y,\qquad v = y' $$

Then:

$$ u' = v,\qquad v' = \text{expression in }u,v,x $$

This is useful for:

  • understanding oscillations

  • identifying spirals, nodes, and saddles

  • checking the effect of damping

Long-term behavior

For linear constant-coefficient equations, the real parts of the characteristic roots control growth or decay.

  • Negative real part: decay

  • Zero real part: sustained oscillation or neutral behavior

  • Positive real part: growth and instability

11. Problem-solving workflow

  1. Put the equation in standard form.

  2. Determine whether it is linear or nonlinear.

  3. Identify whether it is homogeneous.

  4. Check if coefficients are constant, Euler-Cauchy, or variable.

  5. Solve the homogeneous equation first.

  6. Choose a method for a particular solution if forcing is present.

  7. Apply initial or boundary conditions.

  8. Sanity-check the result against the original equation and the expected behavior.

Fast method selection

Equation typeBest first method
Linear constant coefficients, homogeneousCharacteristic equation
Linear constant coefficients, polynomial/exponential/trig forcingUndetermined coefficients
Linear variable coefficients with known solutionReduction of order
General linear forcingVariation of parameters
Euler-CauchyPower trial \(y=x^m\)
Physical mass-spring modelCharacteristic equation with damping interpretation

Common pitfalls

  • Forgetting two integration constants for a second-order homogeneous solution

  • Using an undetermined-coefficients guess that duplicates the homogeneous solution

  • Dropping the \(x\)-factor needed for repeated roots

  • Mixing up \(y'\) and \(y''\) when applying initial conditions

  • Applying a method that assumes constant coefficients to a variable-coefficient equation

  • Treating a boundary value problem like an initial value problem

12. Formula summary

Standard linear form

$$ a_2(x)y'' + a_1(x)y' + a_0(x)y = g(x) $$

Normalized form

$$ y'' + p(x)y' + q(x)y = r(x) $$

Constant-coefficient homogeneous equation

$$ ay'' + by' + cy = 0 $$

Characteristic equation:

$$ ar^2 + br + c = 0 $$

Solution templates

Distinct real roots:

$$ y = C_1 e^{r_1 x} + C_2 e^{r_2 x} $$

Repeated root:

$$ y = (C_1 + C_2 x)e^{rx} $$

Complex roots:

$$ y = e^{\alpha x}\bigl(C_1\cos(\beta x) + C_2\sin(\beta x)\bigr) $$

Euler-Cauchy trial

$$ y = x^m $$

leading to

$$ m(m-1) + am + b = 0 $$

General linear solution

$$ y = y_h + y_p $$

Wronskian

$$ W = y_1 y_2' - y_1' y_2 $$

Mass-spring frequency

$$ \omega_n = \sqrt{\frac{k}{m}} $$

Damping ratio

$$ \zeta = \frac{c}{2\sqrt{mk}} $$

Sources