mathematics//differential equations//PDE

A PDE (partial differential equation) is a differential equation with several independent variables, typically space and time, whose unknown is a field such as a concentration \(c(x,t)\) or a temperature \(T(x,y,z,t)\), and it is used to describe how such a field evolves. Derivatives are partial, one per direction, and the equation relates how the field changes in time to how it varies in space.

A PDE (partial differential equation) is a differential equation with several independent variables, typically space and time, whose unknown is a field such as a concentration c(x,t)c(x,t)c(x,t) or a temperature T(x,y,z,t)T(x,y,z,t)T(x,y,z,t), and it is used to describe how such a field evolves. Derivatives are partial, one per direction, and the equation relates how the field changes in time to how it varies in space.

The canonical example is the diffusion equation, where the time rate at a point depends on the curvature of the field around it; waves, heat and transport are the other classics. A PDE needs an initial profile, not just a number, and boundary conditions in space (initial condition). Chop space into nodes and the PDE becomes a large ODE system, one equation per node, which is why the two share the same stability vocabulary; that is also where the graph Laplacian comes from, and how PDEs are simulated (discretization).

Diffusion flattens the profile, the sharp bump first: the same profile at five times, the earliest at the top.

Play the same bump under each equation: diffusion flattens it, the wave equation splits it into two pulses that bounce off the fixed ends, and advection carries it along unchanged; the map below keeps the whole history, space across and time down.

Chopping space into nodes is what engineers do in practice. An oven has, strictly, a temperature at every point; its controller and its estimator use a handful of representative nodes (wall, load, air), each a state of an ODE model. A finite-element thermal model of a motor does the same with thousands of nodes, and the price of fine meshes is stiffness: small cells exchange heat in milliseconds while the whole machine warms over minutes, so explicit integrators crawl and implicit ones are used instead.