Skip to main content
Mathematics LibreTexts

9.9: The Convolution Theorem

  • Page ID
    90978
  • \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

    Finally, we consider the convolution of two functions. Often we are faced with having the product of two Laplace transforms that we know and we seek the inverse transform of the product. For example, let’s say we have obtained \(Y(s)=\frac{1}{(s-1)(s-2)}\) while trying to solve an initial value problem. In this case we could find a partial fraction decomposition. But, are other ways to find the inverse transform, especially if we cannot perform a partial fraction decomposition. We could use the Convolution Theorem for Laplace transforms or we could compute the inverse transform directly. We will look into these methods in the next two sections. We begin with defining the convolution.

    We define the convolution of two functions defined on \([0, \infty)\) much the same way as we had done for the Fourier transform. The convolution \(f * g\) is defined as \[(f * g)(t)=\int_{0}^{t} f(u) g(t-u) d u .\label{eq:1}\] Note that the convolution integral has finite limits as opposed to the Fourier transform case.

    The convolution operation has two important properties:

    1. The convolution is commutative: \(f * g=g * f\)
      Proof. The key is to make a substitution \(y=t-u\) in the integral. This makes \(f\) a simple function of the integration variable. \[\begin{align} (g * f)(t) &=\int_{0}^{t} g(u) f(t-u) d u\nonumber \\ &=-\int_{t}^{0} g(t-y) f(y) d y\nonumber \\ &=\int_{0}^{t} f(y) g(t-y) d y\nonumber \\ &=(f * g)(t) .\label{eq:2} \end{align}\]
    2. The Convolution Theorem: The Laplace transform of a convolution is the product of the Laplace transforms of the individual functions: \[\mathcal{L}[f * g]=F(s) G(s)\nonumber \] Proof. Proving this theorem takes a bit more work. We will make some assumptions that will work in many cases. First, we assume that the functions are causal, \(f(t)=0\) and \(g(t)=0\) for \(t<0\). Secondly, we will assume that we can interchange integrals, which needs more rigorous attention than will be provided here. The first assumption will allow us to write the finite integral as an infinite integral. Then a change of variables will allow us to split the integral into the product of two integrals that are recognized as a product of two Laplace transforms.
      Carrying out the computation, we have \[\begin{align} \mathcal{L}[f * g] &=\int_{0}^{\infty}\left(\int_{0}^{t} f(u) g(t-u) d u\right) e^{-s t} d t \nonumber \\ &=\int_{0}^{\infty}\left(\int_{0}^{\infty} f(u) g(t-u) d u\right) e^{-s t} d t\nonumber \\ &=\int_{0}^{\infty} f(u)\left(\int_{0}^{\infty} g(t-u) e^{-s t} d t\right) d u\label{eq:3} \end{align}\] Now, make the substitution \(\tau=t-u\). We note that \(\operatorname{int}_{0}^{\infty} f(u)\left(\int_{0}^{\infty} g(t-u) e^{-s t} d t\right) d u=\int_{0}^{\infty} f(u)\left(\int_{-u}^{\infty} g(\tau) e^{-s(\tau+u)} d \tau\right) d u\) However, since \(g(\tau)\) is a causal function, we have that it vanishes for \(\tau<0\) and we can change the integration interval to \([0, \infty)\). So, after a little rearranging, we can proceed to the result. \[\begin{align} \mathcal{L}[f * g] &=\int_{0}^{\infty} f(u)\left(\int_{0}^{\infty} g(\tau) e^{-s(\tau+u)} d \tau\right) d u\nonumber \\ &=\int_{0}^{\infty} f(u) e^{-s u}\left(\int_{0}^{\infty} g(\tau) e^{-s \tau} d \tau\right) d u\nonumber \\ &=\left(\int_{0}^{\infty} f(u) e^{-s u} d u\right)\left(\int_{0}^{\infty} g(\tau) e^{-s \tau} d \tau\right)\nonumber \\ &=F(s) G(s) .\label{eq:4} \end{align}\]

    We make use of the Convolution Theorem to do the following examples.

    Example \(\PageIndex{1}\)

    Find \(y(t)=\mathcal{L}^{-1}\left[\frac{1}{(s-1)(s-2)}\right]\).

    Solution

    We note that this is a product of two functions \[Y(s)=\frac{1}{(s-1)(s-2)}=\frac{1}{s-1} \frac{1}{s-2}=F(s) G(s) \text {. }\nonumber \] We know the inverse transforms of the factors: \(f(t)=e^{t}\) and \(g(t)=e^{2 t}\).

    Using the Convolution Theorem, we find \(y(t)=(f * g)(t)\). We compute the convolution: \[\begin{align} y(t)&=\int_{0}^{t} f(u) g(t-u) d u\nonumber \\ &=\int_{0}^{t} e^{u} e^{2(t-u)} d u\nonumber \\ &=e^{2 t} \int_{0}^{t} e^{-u} d u\nonumber \\ &=e^{2 t}\left[-e^{t}+1\right]=e^{2 t}-e^{t} .\label{eq:5} \end{align}\] One can also confirm this by carrying out a partial fraction decomposition.

    Example \(\PageIndex{2}\)

    Consider the initial value problem, \(y^{\prime \prime}+9 y=2 \sin 3 t, y(0)=1\), \(y^{\prime}(0)=0\).

    Solution

    The Laplace transform of this problem is given by \[\left(s^{2}+9\right) Y-s=\frac{6}{s^{2}+9} .\nonumber \] Solving for \(Y(s)\), we obtain \[Y(s)=\frac{6}{\left(s^{2}+9\right)^{2}}+\frac{s}{s^{2}+9} .\nonumber \] The inverse Laplace transform of the second term is easily found as \(\cos (3 t)\); however, the first term is more complicated.

    We can use the Convolution Theorem to find the Laplace transform of the first term. We note that \[\frac{6}{\left(s^{2}+9\right)^{2}}=\frac{2}{3} \frac{3}{\left(s^{2}+9\right)} \frac{3}{\left(s^{2}+9\right)}\nonumber \] is a product of two Laplace transforms (up to the constant factor). Thus, \[\mathcal{L}^{-1}\left[\frac{6}{\left(s^{2}+9\right)^{2}}\right]=\frac{2}{3}(f * g)(t),\nonumber \] where \(f(t)=g(t)=\sin 3 t\). Evaluating this convolution product, we have \[\begin{align} \mathcal{L}^{-1}\left[\frac{6}{\left(s^{2}+9\right)^{2}}\right] &=\frac{2}{3}(f * g)(t)\nonumber \\ &=\frac{2}{3} \int_{0}^{t} \sin 3 u \sin 3(t-u) d u\nonumber \\ &=\frac{1}{3} \int_{0}^{t}[\cos 3(2 u-t)-\cos 3 t] d u\nonumber \\ &=\frac{1}{3}\left[\frac{1}{6} \sin (6 u-3 t)-u \cos 3 t\right]_{0}^{t}\nonumber \\ &=\frac{1}{9} \sin 3 t-\frac{1}{3} t \cos 3 t .\label{eq:6} \end{align}\]

    Combining this with the inverse transform of the second term of \(Y(s)\), the solution to the initial value problem is \[y(t)=-\frac{1}{3} t \cos 3 t+\frac{1}{9} \sin 3 t+\cos 3 t .\nonumber \] Note that the amplitude of the solution will grow in time from the first term. You can see this in Figure \(\PageIndex{1}\). This is known as a resonance.

    clipboard_e33e5a2149ff2410bbeda5f8ffb57fa01.png
    Figure \(\PageIndex{1}\): Plot of the solution to Example \(\PageIndex{2}\) showing a resonance.

    Example \(\PageIndex{3}\)

    Find \(\mathcal{L}^{-1}\left[\frac{6}{\left(s^{2}+9\right)^{2}}\right]\) using partial fraction decomposition.

    Solution

    If we look at Table 9.7.2, we see that the Laplace transform pairs with the denominator \(\left(s^{2}+\omega^{2}\right)^{2}\) are \[\mathcal{L}[t \sin \omega t]=\frac{2 \omega s}{\left(s^{2}+\omega^{2}\right)^{2}}\nonumber \] and \[\mathcal{L}[t \cos \omega t]=\frac{s^{2}-\omega^{2}}{\left(s^{2}+\omega^{2}\right)^{2}} .\nonumber \] So, we might consider rewriting a partial fraction decomposition as \[\frac{6}{\left(s^{2}+9\right)^{2}}=\frac{A 6 s}{\left(s^{2}+9\right)^{2}}+\frac{B\left(s^{2}-9\right)}{\left(s^{2}+9\right)^{2}}+\frac{C s+D}{s^{2}+9} .\nonumber \] Combining the terms on the right over a common denominator, we find \[6=6 A s+B\left(s^{2}-9\right)+(C s+D)\left(s^{2}+9\right) .\nonumber \] Collecting like powers of \(s\), we have \[C s^{3}+(D+B) s^{2}+6 A s+(D-B)=6 .\nonumber \] Therefore, \(C=0, A=0, D+B=0\), and \(D-B=\frac{2}{3}\). Solving the last two equations, we find \(D=-B=\frac{1}{3}\).

    Using these results, we find \[\frac{6}{\left(s^{2}+9\right)^{2}}=-\frac{1}{3} \frac{\left(s^{2}-9\right)}{\left(s^{2}+9\right)^{2}}+\frac{1}{3} \frac{1}{s^{2}+9} .\nonumber \] This is the result we had obtained in the last example using the Convolution Theorem.


    This page titled 9.9: The Convolution Theorem is shared under a CC BY-NC-SA 3.0 license and was authored, remixed, and/or curated by Russell Herman via source content that was edited to the style and standards of the LibreTexts platform; a detailed edit history is available upon request.