Kelvin functions

When familiar tools fall short, science demands new languages to describe the world's complexity. Such is the case with alternating current flowing through a wire, a seemingly simple scenario that conceals a rich physical phenomenon known as the skin effect. Describing this non-uniform current distribution requires moving beyond standard functions into the realm of special functions, giving rise to the elegant and powerful Kelvin functions. This behavior presents a knowledge gap that simple resistance formulas cannot fill, necessitating a more sophisticated mathematical model.

This article serves as your guide to understanding these unique functions. The first chapter, ​​"Principles and Mechanisms"​​, explores their mathematical origins, deriving them from Bessel's equation and uncovering their properties through power series and asymptotic forms. We will see how these functions, ber and bei, represent the coupled real and imaginary parts of a complex physical quantity. Following this, the chapter on ​​"Applications and Interdisciplinary Connections"​​ showcases the remarkable versatility of Kelvin functions, revealing how the same mathematics that governs electrical currents also describes the structural mechanics of a bending plate and connects to other fundamental functions of mathematical physics. Our journey begins with the physical problem that birthed these functions: the intricate dance of an alternating current within a conductor.

Imagine you're trying to describe a dance. You could track the position of one dancer, but that only tells you half the story. The real beauty of a pas de deux lies in the interaction—how one dancer's movement influences the other's. The world of physics is full of such coupled phenomena, and to describe them, we sometimes need a new mathematical language. The Kelvin functions are just such a language, born from the need to describe a seemingly simple situation: alternating current flowing through a wire.

When a direct current (DC) flows through a cylindrical wire, it distributes itself evenly across the wire's cross-section. Simple. But what happens when we use an alternating current (AC)? The story gets far more interesting. The changing magnetic field produced by the current induces swirling electric fields (eddy currents) within the conductor itself, which oppose the main flow of current in the center. The result is that the current density is highest near the surface, or "skin," of the wire and drops off towards the center. This is the famous ​​skin effect​​.

To describe this mathematically, we need to find the current density, which is a complex quantity $y(r)$ that depends on the radial distance $r$ from the center of the wire. This complex function holds information about both the amplitude and the phase of the AC current. It turns out that this function obeys a particular differential equation:

Here, $C$ is a constant that depends on the wire's material and the frequency of the current, and $i$ is the imaginary unit, $i^2 = -1$.

Look at that equation! It resembles the famous Bessel's equation, which describes everything from the vibrations of a drumhead to the propagation of light. But there's a crucial difference: the mischievous little '$i$' sitting in the last term. This single imaginary unit is the key. It tells us that our solution $y(r)$ won't be a simple real-valued function; it lives in the complex plane. This '$i$' is what forces our dancers to move together.

Since the solution $y(r)$ is complex, we can write it in terms of its real and imaginary parts, let's call them $U(r)$ and $V(r)$, so that $y(r) = U(r) + iV(r)$. What happens if we plug this into our differential equation? The equation splits, as if by magic, into two separate equations, one for the real part and one for the imaginary part. By grouping the terms without an $i$ and those with an $i$, we get a coupled system of equations:

Notice the beautiful symmetry here. The equation for $U$ depends on $V$, and the equation for $V$ depends on $-U$. They are inextricably linked. $U(r)$ cannot make a move without $V(r)$ being involved, and vice-versa. These two functions, the physical solutions to this system, are what we call the ​​Kelvin functions​​. Specifically, the solution that is well-behaved at the center of the wire ($r=0$) is written in terms of ​​$\text{ber}_0(x)$​​ and ​​$\text{bei}_0(x)$​​, where $x = Cr$:

The "ber" stands for "Bessel real" and "bei" for "Bessel imaginary." They are the two dancers in our story, forever locked in a mathematical embrace dictated by the physics of electromagnetism.

So we've named these functions, but what do they actually look like? We need a way to calculate them. The secret lies in their connection to a more familiar character in the world of special functions: the ​​modified Bessel function of the first kind​​, $I_0(z)$. The definition is astonishingly compact:

We're evaluating a standard function, $I_0$, but with a strange, complex argument: $x\sqrt{i}$. The entire complexity of the Kelvin functions is bundled into this one maneuver.

Let's see how this works. The function $I_0(z)$ can be written as a power series, which is like giving a recipe to build the function term by term:

Now, let's substitute $z=x\sqrt{i}$. This means $z^2 = x^2(\sqrt{i})^2 = ix^2$. So our series becomes:

The powers of $i$ are the interesting part: $i^0=1$, $i^1=i$, $i^2=-1$, $i^3=-i$, $i^4=1$, and so on, in a cycle of four. When we separate the terms with real coefficients from those with imaginary coefficients, we are finding the series for $\text{ber}_0(x)$ and $\text{bei}_0(x)$, respectively.

For $\text{ber}_0(x)$ (the real part), we only keep the terms where $k$ is even. Let $k=2m$. Then $i^k = i^{2m} = (i^2)^m = (-1)^m$. This gives us:

For $\text{bei}_0(x)$ (the imaginary part), we only keep terms where $k$ is odd. Let $k=2m+1$. Then $i^k = i^{2m+1} = (-1)^m i$. After dividing by $i$, the coefficients are real. This gives us:

This immediately tells us something important about their behavior near the origin ($x=0$). As $x \to 0$, we have $\text{ber}_0(x) \to 1$ and $\text{bei}_0(x) \to \frac{x^2}{4}$. So at the very center of the wire, $\text{ber}_0(0)=1$ and $\text{bei}_0(0)=0$.

Just as Bessel functions come in different orders ($\nu = 0, 1, 2, \dots$) and two "kinds" ($J_\nu$ and $Y_\nu$), so do Kelvin functions. The ​​Kelvin functions of order $\nu$​​, $\text{ber}_\nu(x)$ and $\text{bei}_\nu(x)$, are defined similarly, through the modified Bessel function $I_\nu(z)$. For example, for small arguments, $\text{ber}_1(x)$ behaves like $\frac{\sqrt{2}}{4}x$, showing a different startup behavior than its order-zero cousin.

What about the "second kind"? A second-order differential equation always has two fundamentally different solutions. Our coupled system is no different. The functions $\text{ber}_\nu(x)$ and $\text{bei}_\nu(x)$ are the "first kind" solutions, the well-behaved ones that are finite at the origin. There must be another set of solutions that are not well-behaved at the origin. These are the ​​Kelvin functions of the second kind​​, denoted ​​$\text{ker}_\nu(x)$​​ and ​​$\text{kei}_\nu(x)$​​.

They are defined through the modified Bessel function of the second kind, $K_0(z)$, which is known to have a logarithmic singularity at $z=0$. Following our recipe, we look at $K_0(x\sqrt{i})$. For small $x$, the dominant behavior of $\text{ker}_0(x)$ comes from the logarithm inside $K_0$:

This means that as you approach the center of the wire ($x \to 0$), the value of $\text{ker}_0(x)$ runs off to infinity! This makes it unsuitable for describing the current inside a solid wire, but it's essential for problems involving hollow pipes or regions outside a conductor. Its relationship to the first-kind functions is also intriguingly complex, involving not just series but also logarithmic terms, a hint at its wilder nature.

A truly beautiful feature of the Bessel family of functions is that they are all related. You don't just have a collection of independent functions; you have an interconnected web. Derivatives of one function can be expressed in terms of other functions, often of a different order. These are called ​​recurrence relations​​, and they are the rules of the choreography for our dancers.

For instance, by differentiating the definition $I_1(x e^{i\pi/4}) = \text{ber}_1(x) + i\, \text{bei}_1(x)$ and using known relations for Bessel functions, one can find relationships for the derivatives of Kelvin functions. A delightful example shows that the sum of the derivatives at the origin, $\text{ber}_1'(0) + \text{bei}_1'(0)$, is exactly $\frac{1}{\sqrt{2}}$. These relations aren't just mathematical elegance; they are powerful tools for calculation, allowing us to compute derivatives and functions of higher orders from simpler, known values.

We've looked at the functions up close, near the origin. What happens far away, when $x$ is large? This corresponds to looking at the current distribution in a very thick wire, or at very high frequencies. This is the domain of ​​asymptotics​​.

Here, the Kelvin functions reveal their most dramatic behavior. The analysis is a bit involved, but the result is wonderfully intuitive. For large $x$, the derivative of $\text{ber}_0(x)$, for example, looks something like this:

Let's dissect this. There are three parts:

1. A slow decay: $\frac{1}{\sqrt{2\pi x}}$, which makes the amplitude slowly decrease.
2. A rapid growth: $\exp(\frac{x}{\sqrt{2}})$, an exponential term.
3. An oscillation: $\cos(\dots)$, a wiggling cosine term.

The combination of exponential and oscillatory behavior is characteristic of waves that are being damped. In the context of our wire, this asymptotic form is the mathematical signature of the skin effect! It tells us that far from the center (large $x$, but remember $r$ is radius, so we are thinking about scaling), the current amplitude changes in a way that is both oscillatory and rapidly decaying as we move inward from the skin. The beauty of the mathematics is that it hands us the physics on a silver platter.

As a final thought, let's ask a technical question. Are $\text{ber}_0(x)$ and $\text{bei}_0(x)$ truly independent functions? In the theory of differential equations, we have a tool for this called the ​​Wronskian​​. For two functions $f(x)$ and $g(x)$, it's defined as $W(x) = f(x)g'(x) - f'(x)g(x)$. For two independent solutions to a standard second-order ODE, the Wronskian is typically a simple function, like $A/x$.

What about our Kelvin functions? If we compute their Wronskian, we find something curious. For small $x$, the Wronskian of $\text{ber}_0(x)$ and $\text{bei}_0(x)$ is approximately $x/2$. It's not a constant, nor is it $A/x$. It starts at zero and grows linearly. This is a subtle but profound mathematical fingerprint. It's a reminder that $\text{ber}_0$ and $\text{bei}_0$ are not solutions to a single, simple, real differential equation. They are solutions to a coupled system, born from one elegant complex equation. They are two sides of the same complex coin, forever linked in a dance that describes the hidden flow of energy in the world around us.

Now that we have taken apart the elegant machinery of the Kelvin functions, exploring their gears and springs, it's time for the real fun. It's time to see what this machine does. Learning the principles of physics and mathematics without seeing their application is like memorizing the rules of chess without ever playing a game. The true beauty of a concept is revealed not in its abstract definition, but in the surprising and profound ways it shows up in the world, describing the dance of electricity, the bending of a steel plate, and even the hidden relationships between seemingly distant mathematical ideas.

So, let's go on a journey and see where these peculiar functions, born from a simple differential equation, make their home.

Perhaps the most classic and visceral application of Kelvin functions is in the world of electromagnetism, particularly when alternating currents (AC) are involved. When you send a direct current (DC) through a thick copper wire, it happily spreads out, flowing uniformly through the entire cross-section. But alternating current is more fickle. As the current rapidly switches direction, it induces changing magnetic fields within the wire itself. These fields, in turn, induce swirling "eddy currents" that oppose the flow in the center of the wire. The net effect is that the current is pushed outwards, crowding into a thin layer near the surface. This is the famous ​​skin effect​​.

But how thick is this "skin"? And how does this crowding affect the wire's properties, like its resistance and inductance? To answer this precisely, you have to solve the diffusion equation for the electromagnetic fields inside the cylindrical conductor. The solutions that naturally emerge are Bessel functions with a complex argument—which is precisely the definition of Kelvin functions! They describe how the current density decays from the surface inward. The ber function is related to the in-phase part of the current, and the bei function is related to the out-of-phase part. In the high-frequency limit, a practical concern for radio-frequency circuits, this analysis allows engineers to calculate the effective internal inductance of a wire, a crucial parameter in designing high-speed electronics.

The same physics governs how an external magnetic field penetrates a conductor. Imagine a solid metal cylinder placed in a magnetic field that oscillates back and forth. The field doesn't just appear instantly inside the cylinder. It has to diffuse inward, battling the eddy currents it creates along the way. Kelvin functions perfectly describe this process. They tell us two things: how much the field is attenuated (weakened) as it moves toward the center, and by how much its phase lags behind the field at the surface. At the very center of the cylinder, the phase lag $\delta$ is given by a beautifully simple relationship between the ber and bei functions: $\tan(\delta) = \mathrm{bei}_0 / \mathrm{ber}_0$. This isn't just a mathematical curiosity; it's a physical reality, a time-delay written in the language of special functions.

One of the most thrilling moments in science is when you discover that a tool you developed for one problem unexpectedly solves a completely different one. So it is with Kelvin functions. Let's leave the world of electricity and step into the realm of solid mechanics.

Imagine a large, thin elastic plate—say, a sheet of steel—resting on a uniform elastic foundation, like a giant block of foam. Now, what happens if you apply a concentrated force at a single point, as if poking it with an infinitely sharp needle? The plate will deform, creating a dimple that ripples outwards. The mathematical description of this shape is called the "fundamental solution" or "Green's function" for the system. It's the most basic response from which all other deformation patterns can be built.

The governing equation for this physical system is $(\Delta^2 + \lambda^4)E = \delta(\mathbf{x})$, where $\Delta$ is the Laplacian operator, $\lambda$ is a parameter related to the stiffness of the plate and foundation, and $\delta(\mathbf{x})$ represents the point-like poke. When you solve this equation, a remarkable thing happens: the shape of the deflection, $E(r)$, is described precisely by the Kelvin function of the second kind, $\mathrm{kei}_0(r)$. Specifically, $E(r) \propto -\mathrm{kei}_0(\lambda r)$. Think about that for a moment. The very same mathematical DNA that governs how AC current avoids the center of a wire also describes how a metal plate sinks into a foam bed. This is a stunning example of the unity of physics and engineering, where disparate phenomena are discovered to obey the same underlying mathematical laws.

Beyond direct physical applications, Kelvin functions serve as a crucial node in a vast, interconnected web of mathematical physics. They are part of a grand family of "special functions," and understanding their relationships is like discovering a hidden genealogy that links the pillars of scientific calculation.

For instance, consider the ​​Legendre polynomials​​, $P_n(x)$. These are the workhorses of fields with spherical symmetry, from calculating planetary gravitational fields to describing the probability clouds of electrons in an atom. They seem, on the surface, to have little to do with the cylindrical world of Kelvin functions. Yet, if you look at the Legendre polynomials in a very specific asymptotic limit—for large order $n$ and an argument just slightly displaced into the complex plane—they magically transform into Kelvin functions. It's a beautiful surprise, showing how one set of solutions can morph into another under the right conditions.

This web extends further. The ​​Airy function​​, $\mathrm{Ai}(x)$, is another celebrity in the world of mathematical physics. It famously describes the rainbow's fringe, the wavelike patterns of light near a caustic (like the bright curve of light inside a coffee cup), and the quantum state of a particle in a uniform gravitational field. It solves a completely different differential equation ($y'' - xy = 0$). And yet, it too is related to Kelvin functions. There exists a deep integral connection between them, where integrating the product of an Airy function and a Kelvin function over all space yields a simple constant. They are not the same, but they 'know' about each other; they are tied together in the abstract space of all possible functions.

Finally, Kelvin functions even make surprising appearances in the bridge between the discrete and the continuous. In fields like condensed matter physics, one often needs to calculate the total effect of an infinite lattice of atoms, which involves summing up a series. The brilliant Abel-Plana summation formula provides a way to convert such a discrete sum into a continuous integral. When this formula is applied to sums involving the modified Bessel function $K_0$—which itself is closely related to Kelvin functions—the resulting integral expression features the Kelvin function $\mathrm{kei}_0$ explicitly. This provides a powerful tool for analytically handling sums that arise in the study of crystal lattices and other periodic structures.

From the hum of electricity in a wire, to the dimple on a steel plate, to the hidden kinship with the functions of rainbows and atoms, the Kelvin functions are far more than a textbook exercise. They are a testament to the unifying power of mathematics—a versatile language that nature uses, again and again, to write its stories.