JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Reading the Solution: Transient and Steady State

You have the integrating-factor formula in hand — now learn to read what it says. Every constant-coefficient linear solution splits into a fading transient that remembers where you started and a lasting steady state set entirely by the forcing.

From a formula to a story

The previous guides handed you a machine: put a first-order linear equation into standard form y' + p(x) y = q(x), build the integrating factor, integrate once, and out comes the general solution. That machine is reliable, but a formula full of symbols is not yet understanding. This final rung is about reading the answer — looking at the two pieces it always produces and asking what each one physically *does*. When the equation describes something settling down over time, those two pieces have names you will meet everywhere in science: the transient and the steady state.

To keep the picture sharp, switch the independent variable to time t and focus on the case that matters most for this story: a constant coefficient. Take y' + a y = q(t) with a a fixed positive number. From guide 3 you already know the answer splits as a homogeneous part plus a particular part, y = y_h + y_p. Here y_h = C e^(-a t) and y_p(t) is one function that follows the forcing q(t). The whole insight of this guide is reading what happens to each of those two terms as t grows large.

The transient: the part that forgets

Look hard at y_h = C e^(-a t). Because a is positive, e^(-a t) shrinks toward zero as time runs on — slowly if a is small, briskly if a is large, but always heading to nothing. This is the transient term: it is loud at the start and then dies away. Notice where the arbitrary constant C lives — entirely inside this fading piece. C is set by your initial condition, so the transient is precisely the part of the solution that *remembers where you started*. And since it decays, that memory is temporary: give it enough time and the system forgets its initial state completely.

How fast it forgets is captured by one number. Write a = 1/tau; then the transient is C e^(-t/tau), and tau is called the time constant — the time for the transient to shrink to about 37% (that is, 1/e) of its size. After about three or four time constants the transient is practically gone. A large a means a small tau means fast forgetting; a small a means a sluggish system that clings to its past for a long while. This single number tells you, at a glance, *how long the start-up wobble lasts*.

The steady state: the part that lasts

Once the transient has faded, only y_p(t) remains. This is the steady-state term — the long-run behaviour the system settles into, and the crucial thing about it is that *it carries no arbitrary constant*. It does not depend on where you started; it is dictated entirely by the forcing q(t). Two runs of the same system from wildly different initial states differ only in their transients, and once those have died away the two solutions become indistinguishable, riding the very same steady state. The forcing has the last word.

The *shape* of the steady state mirrors the shape of the forcing. If q(t) is a constant Q, the steady state is a constant level y_p = Q/a — a flat line the solution approaches. If q(t) is a sinusoid, the steady state is a sinusoid of the same frequency, generally shifted in phase and scaled in amplitude (this lag is the seed of the frequency-response ideas you will meet in second-order forced oscillations). The system, in the long run, sings back in the key it is being driven — it just takes a transient's worth of time to tune in.

y' + a y = q(t),  a > 0

y(t)  =     y_p(t)      +     C e^(-a t)
            --------           ----------
          STEADY STATE          TRANSIENT
        set by forcing q       set by start C
        lasts as t -> inf      fades as t -> inf

constant forcing q = Q   ->   steady state = Q / a
time constant tau = 1/a  ->   ~37% left after t = tau
One card: the same solution, read as two terms with two fates.

Seeing it in the world

This is not an abstraction; it is the shape of everyday settling. Pour hot coffee into a cool room and Newton's law of cooling gives T' = -k(T - T_room), a linear equation whose steady state is exactly T_room — the temperature the forcing (the room) imposes — while the transient C e^(-k t) is the initial heat draining away. The coffee starts hot (a big transient), then forgets its starting temperature and relaxes to the room. You have *watched* a transient decay every time a drink cooled.

The same two-piece story plays out in an RC or RL circuit: throw a switch and the current surges, then settles to the level the source demands. The brief surge is the transient, the settled level is the steady state, and the circuit's time constant tau = R C (or L/R) fixes how long the surge lasts — engineers tune exactly this. Mixing tanks, charging capacitors, drug levels in the bloodstream, a thermostat catching up to a new setting: all wear the same y = transient + steady-state coat, because all are governed by a stable first-order linear equation.

A worked reading, and honest edges

Put it all together on y' + 2 y = 6 with y(0) = 5. The integrating factor is e^(2t); running the recipe gives the general solution y = 3 + C e^(-2t). Apply y(0) = 5: 5 = 3 + C, so C = 2, and y(t) = 3 + 2 e^(-2t). Now *read* it. The steady state is 3 (which is just Q/a = 6/2, the level where y' = 0). The transient is 2 e^(-2t), starting at 2 and vanishing with time constant tau = 1/2. So the solution begins at 5, then slides down to 3 and stays — your initial 'extra 2' is exactly the transient, draining away.

  1. Read off the steady state: set y' = 0 in y' + a y = q to find the level the forcing imposes (here 6/2 = 3).
  2. Identify the transient: it is the homogeneous part C e^(-a t), the piece holding the arbitrary constant.
  3. Pin C with the initial condition, so the transient measures exactly the gap between your start and the steady state.
  4. Read the time constant tau = 1/a to know how long the transient lasts — a few tau and it is gone.

Two honest cautions before you climb on. First, this clean split is the gift of constant-coefficient *linearity* and decay: it relies on superposition to add the pieces and on a > 0 for the transient to actually die. If a varies in sign over the interval, or the equation is nonlinear, 'transient plus steady state' can stop being meaningful — a nonlinear equation may approach a moving target, oscillate forever, or blow up instead of settling. Second, 'steady state' means *settled*, not necessarily *constant*: under a sinusoidal drive the steady state is a perpetual oscillation. It is steady in the sense that the system has stopped remembering its start, not in the sense of standing still.

Where this lands you

You have now closed the first-order arc: the separable family, the lost-solution pitfall, the linear family, the integrating factor, and — here — how to *read* the answer it produces. The transient-and-steady-state decomposition is more than a first-order trick; it is your first glimpse of a theme that organizes nearly all of linear ODEs. One level up, in second-order forced oscillations, the very same words return: the homogeneous solution is again a transient that decays, and the particular solution is again the steady state set by the forcing — only now with the richer possibility of resonance.

So carry this reading habit upward. Whenever you solve a stable linear equation, do not stop at the formula — split it, name the transient, find the steady state, and read the time constant. That habit turns a wall of algebra into a short, honest sentence about behaviour: *the system surges, forgets where it began, and relaxes to whatever the forcing demands.* That sentence is what the mathematics was for.