ODE: Step Functions and Piecewise Forcing

Differential equations model dynamic systems, but many real-world inputs—a switch flipping in a circuit, a valve opening in a process control system, or an impact force in mechanics—are not continuous. They turn on or off abruptly. To model these discontinuous or piecewise-defined forcing functions efficiently, engineers rely on the Heaviside step function and the power of the Laplace transform. Mastering these tools transforms a cumbersome, case-by-case ODE solution into an elegant, unified algebraic process, which is essential for analyzing systems with switching behavior.

The Heaviside Step Function: The Mathematical Switch

The foundational tool for modeling abrupt changes is the Heaviside step function, denoted $u_c(t)$ or $u(t-c)$. It is defined as: $$u_c(t)=\{\begin{array}{ll}0,& t<c\\ 1,& t\ge c\end{array}$$ You can think of $u_c(t)$ as a perfect, instantaneous switch that turns "on" at time $t=c$. The value $c$ is called the shift or delay. When $c=0$, we often write simply $u(t)$. This function is the building block for representing any function that exhibits piecewise behavior, allowing you to "activate" parts of an expression at specific times.

A related, useful function is the window function or pulse, created by combining two step functions. For example, a function that is 1 between $t=a$ and $t=b$ and zero elsewhere can be written as $u_a(t)-u_b(t)$. This ability to "turn on" and "turn off" terms is what makes the step function so powerful for constructing piecewise descriptions.

Representing Piecewise Functions with Step Functions

Any piecewise-defined function can be rewritten compactly using linear combinations of Heaviside functions. The strategy is to express the function as a sum, where each term is the desired expression for a given time interval, multiplied by a step function that activates it at the correct starting time.

Consider a classic piecewise forcing function: $$f(t)=\{\begin{array}{ll}g(t),& 0\le t<c\\ h(t),& t\ge c\end{array}$$ This can be rewritten using step functions as: $$f(t)=g(t)\cdot [1-u_c(t)]+h(t)\cdot u_c(t)$$ A more compact and commonly used form is: $$f(t)=g(t)+[h(t)-g(t)]u_c(t)$$ Here, $g(t)$ is the "baseline" function active from $t=0$. At $t=c$, the term $[h(t)-g(t)]u_c(t)$ switches on, adding the necessary adjustment to change the function from $g(t)$ to $h(t)$. This representation is crucial because it allows us to cleanly apply the Laplace transform to the entire piecewise system at once.

Laplace Transform of Step Functions and the Second Shifting Theorem

The Laplace transform of the basic Heaviside function is a standard result: $$\mathcal{L}\{u_c(t)\}=\frac{e^{-cs}}{s},s>0$$ However, the real power comes from the Second Shifting Theorem (or Time-Shifting Property). This theorem states that if the Laplace transform of $f(t)$ is $F(s)$, then the transform of a function shifted and turned on by a step function is: $$\mathcal{L}\{f(t-c)u_c(t)\}=e^{-cs}F(s)$$ This is the operational workhorse. It means that to find the Laplace transform of an expression like $t^2{u}_3(t)$, you must first rewrite it as a function of $(t-3)$. Observe: $t^2{u}_3(t)$ is not of the form $f(t-3)u_3(t)$. To apply the theorem, we let $f(t)=(t+3{)}^2$, because then $f(t-3)=((t-3)+3{)}^2=t^2$. Therefore, $$\mathcal{L}\{t^2{u}_3(t)\}=\mathcal{L}\{f(t-3)u_3(t)\}=e^{-3s}\mathcal{L}\{f(t)\}=e^{-3s}\mathcal{L}\{(t+3{)}^2\}$$ You then expand $(t+3{)}^2$ and take its transform term-by-term. This "rephrasing" step is the most critical part of correctly applying the Second Shifting Theorem.

Solving ODEs with Piecewise Forcing: A Complete Workflow

Let's synthesize these concepts to solve a typical engineering ODE with a switched forcing function. Consider the initial value problem modeling a damped system with a delayed force: $$y^{\prime \prime }+3y^{\prime }+2y=f(t),y(0)=0,y^{\prime }(0)=0,\text{where }f(t)=\{\begin{array}{ll}0,& 0\le t<2\\ 4,& t\ge 2\end{array}$$

Step 1: Express the forcing function using $u_c(t)$. We have $f(t)=4u_2(t)$.

Step 2: Apply the Laplace transform to the entire ODE. Using linearity and derivative properties: $$[s^{2}Y(s)-sy(0)-y^{\prime }(0)]+3[sY(s)-y(0)]+2Y(s)=\mathcal{L}\{4u_2(t)\}$$ Substituting initial conditions gives: $$(s^2+3s+2)Y(s)=4\cdot \frac{e^{-2s}}{s}$$ Thus, $Y(s)=\frac{4e^{-2s}}{s(s^2+3s+2)}=\frac{4e^{-2s}}{s(s+1)(s+2)}$.

Step 3: Handle the shifted term using the Second Shifting Theorem in reverse. We have $Y(s)=e^{-2s}\cdot F(s)$, where $F(s)=\frac{4}{s(s+1)(s+2)}$. We find the inverse Laplace transform of $F(s)$ first via partial fractions: $$\frac{4}{s(s+1)(s+2)}=\frac{A}{s}+\frac{B}{s+1}+\frac{C}{s+2}$$ Solving yields $A=2$, $B=-4$, $C=2$. Therefore, $f(t)={\mathcal{L}}^{-1}\{F(s)\}=2-4e^{-t}+2e^{-2t}$.

Step 4: Apply the inverse of the Second Shifting Theorem. Since $Y(s)=e^{-2s}F(s)$, the inverse transform is: $$y(t)={\mathcal{L}}^{-1}\{e^{-2s}F(s)\}=f(t-2)\cdot u_2(t)$$ Therefore, the final solution is: $$y(t)=[2-4e^{-(t-2)}+2e^{-2(t-2)}]u_2(t)$$ This solution is elegantly piecewise: it is exactly zero for $t<2$ (as the step function is off), and for $t\ge 2$, it follows the response of the system to a constant force that started at $t=2$.

Modeling Physical Systems with Switching Behavior

This formalism is directly applicable to core engineering domains. In circuit analysis, a voltage source that is connected to an RLC circuit at time $t=t_0$ is modeled as $V_s{u}_{t_0}(t)$. Solving the resulting integro-differential equation with the Laplace transform automatically incorporates the switching event and provides the complete transient and steady-state response.

In mechanical systems, a constant force that begins to act on a mass-spring-dashpot system after a delay—like a motor engaging—is represented as $F_0{u}_c(t)$. The solution procedure is identical, yielding the displacement $x(t)$ before and after the force is applied. This approach is vastly superior to solving two separate ODEs and matching initial conditions at $t=c$, as it handles the discontinuity intrinsically through the algebraic machinery of the transform.

Common Pitfalls

1. Misapplying the Second Shifting Theorem for transforms. The most frequent error is attempting to write $\mathcal{L}\{g(t)u_c(t)\}$ directly as $e^{-cs}G(s)$. This is only true if $g(t)$ is already written as $f(t-c)$. Correction: Always re-express $g(t)u_c(t)$ in the form $f(t-c)u_c(t)$ by substituting and solving for $f(t)$. For example, for $\cos (t)u_{\pi }(t)$, set $f(t)=\cos (t+\pi )=-\cos (t)$, so that $f(t-\pi )=\cos (t)$.

1. Forgetting the step function in the inverse transform. When you have a term like $e^{-cs}F(s)$, its inverse Laplace is $f(t-c)u_c(t)$, not just $f(t-c)$. Omitting the $u_c(t)$ means your solution is incorrectly defined for $t<c$. Correction: The step function is an integral part of the inverse; it ensures the solution respects causality and is zero before the switching event.

1. Incorrectly representing piecewise functions. A common mistake is misidentifying the "baseline" function $g(t)$ in the representation $f(t)=g(t)+[h(t)-g(t)]u_c(t)$. Correction: Always let $g(t)$ be the function valid on the first interval, starting at $t=0$. The step function then encodes the change at $t=c$.

Summary

• The Heaviside step function $u_c(t)$ is the essential mathematical tool for modeling instantaneous switching or piecewise-defined inputs in dynamic systems.
• Any piecewise function can be rewritten as a sum of terms, each activated at a specific time by a step function, creating a single expression valid for all $t\ge 0$.
• The Laplace transform of $u_c(t)$ is $e^{-cs}/s$, and the Second Shifting Theorem, $\mathcal{L}\{f(t-c)u_c(t)\}=e^{-cs}F(s)$, is the key to transforming and inverting expressions involving delays.
• Solving ODEs with piecewise forcing involves expressing the force with step functions, transforming the ODE, solving algebraically for $Y(s)$, and carefully applying the inverse Second Shifting Theorem to obtain a solution that inherently incorporates the switching behavior.
• This methodology is directly applicable to modeling real engineering systems in circuits and mechanics, providing a unified, efficient solution framework superior to solving multiple initial value problems.