Why another evolution equation?
A displaced string does not generally settle to zero like a temperature distribution: in an ideal model it oscillates. Its displacement and its velocity at the initial time are separate pieces of data. The one-dimensional wave equation models this distinction and gives a useful setting in which to see how boundary conditions select a Fourier basis.
Let and consider a twice continuously differentiable real-valued function on satisfying
Here is the wave speed, is the initial displacement and the initial velocity. The ends are held fixed. In this classical setting continuity at the corners requires . For arbitrary data these conditions and further smoothness or compatibility can matter; an series alone does not automatically give a twice continuously differentiable solution up to the corners.
Separation and the two initial conditions
Try . Apart from zero factors, separation gives . With , a nonzero solution exists precisely for
Indeed, gives a linear function and a combination of hyperbolic functions; the two endpoint conditions force zero in either case. For the endpoint at zero removes the cosine and the endpoint at requires . The corresponding time equation is , with . Thus each mode has both a cosine and a sine in time.
Theorem 2.5.1
(Finite sine data give a classical solution)
Suppose and . Then
is the unique twice continuously differentiable solution on of the stated problem.
Proof
(Check the modes, then compare by energy)
The finite sum may be differentiated term by term. Each summand obeys because , and vanishes at both ends. At its displacement coefficient is ; differentiating in time gives velocity coefficient . To prove uniqueness, take the difference of any two classical solutions. It has zero initial data and fixed ends. Its energy, defined below, is constant and initially zero. Hence everywhere, by continuity and nonnegativity of the energy integrand. Since , . The argument does not require the competing solution to be a finite sum.
Energy is conserved, not each energy component
Definition 2.5.2
(Energy of the string)
For a classical solution , the energy at time is
The first term is kinetic energy and the second strain energy.
Theorem 2.5.3
(Energy conservation with fixed ends)
If is twice continuously differentiable on , satisfies and , then is constant on .
Proof
(Differentiate and integrate by parts)
Differentiating under the integral, substituting the equation and integrating by parts yield
The fixed endpoint condition implies at each end.
Proof X-ray
Where the uniqueness argument uses each hypothesis
Continuity of second derivatives allows differentiation under the integral and the integration by parts. The PDE is used once, to turn into , after which the interior terms combine into the exact derivative . The boundary enters subtly: what kills the flux is not but its time derivative , which follows because for every . In the uniqueness proof, zero energy and a continuous nonnegative integrand force pointwise. That alone gives only a constant; the zero boundary value, or equally the zero initial displacement, fixes the constant as zero.
Without the fixed ends the boundary term represents energy transfer through the ends and cannot simply be dropped. The separate kinetic and strain terms can change even though their sum is constant: a mode at maximal displacement has zero velocity but can have nonzero slope.
For the finite solution, sine and cosine orthogonality give the more explicit identity
This is useful for checking a calculation: energy depends on both initial data. It is also a stability statement. If two solutions differ initially by finite sine coefficients , their difference keeps exactly this energy at every time. It does not say that their pointwise displacement error remains equal to its initial value.
A complete two-mode calculation
Worked example 2.5.4
(Displacement and velocity excite the same mode differently)
Take , , and . The frequencies are and , so the solution is
The factor multiplying the initial velocity coefficient is essential: differentiating produces . The conserved energy is
At , the displacement is and the velocity is . Their energies still sum to .
Common mistake
(Conserving kinetic and strain energy separately)
Only the total is conserved. In the example above, at the kinetic part is and the strain part is . At the kinetic part is and the strain part is . Each part changed; the sum stayed .
Common mistake
(Using instead of )
Writing gives initial velocity coefficients , not . The error is invisible in the initial displacement and appears only when is checked.
Worked example 2.5.5
(Data that is a sine sum in disguise)
Take , , and . The identity shows that is a finite sine sum with and , and . The theorem gives
The mode formula gives . As a check, compute directly: , and , so .
Worked example 2.5.6
(A tiny high-mode velocity error)
With , , perturb the initial velocity by and leave the displacement unchanged. The difference of the two solutions is the single mode , whose displacement never exceeds . Its energy is at every time. In the opposite edge case, a displacement error with zero velocity error carries energy , a hundred times larger: energy weights displacement errors by , so it measures slopes, not heights.
Misconceptions and limitations
Counterexample mode
A smooth initial profile can still conflict with fixed ends
The function is smooth on , but it cannot be the initial trace of a solution continuous on the closed space-time rectangle with : continuity at would force both and . A weaker solution may be constructed in a different function-space setting; this contradiction only rules out the stated continuous classical solution.
Unlike heat flow, the time factors in the wave series do not damp high frequencies. Consequently an infinite series requires its own convergence and regularity analysis; differentiating it twice is not justified merely by writing down its coefficients. Similarly, conserved energy does not prove that each spatial coefficient is constant.
Practice with solutions
Quick check
(For , , and , find and its energy.)
Here .
Solution · Answer
Here , so and .
Quick check
(With and , the initial displacement is and the initial velocity is zero. Find the first positive time at which the displacement is identically zero. Is the energy then zero?)
Look at the velocity at that time.
Solution · Answer
The solution is ; the first positive zero is . Its velocity is then , so the energy remains , not zero.
Quick check
(Explain how the uniqueness proof changes if both endpoint values are zero but the initial velocities of two solutions differ.)
Compute the initial energy of the difference.
Solution · Answer
Their difference still has constant energy, but its initial energy need not be zero. Thus the argument bounds the difference in energy and no longer proves the two solutions equal.