First-order systems and their dynamic response

First-order systems K/(τs + 1): how thermometers, tanks and mixing vessels give them, and their step, impulse, ramp and sinusoidal responses.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

A surprising number of process elements behave as first-order systems: a thermometer in a bath, a liquid tank draining through a valve, a stirred mixing tank, a sensor with a thermowell. Knowing how they respond to steps, ramps and sinusoids tells you how long a reading takes to settle, how much a measurement lags a changing process, and how fast a loop can possibly be.

Key ideas

Definition. A first-order system obeys τ·dy/dt + y = K·u, with y and u as deviation variables. Its transfer function is G(s) = Y(s)/U(s) = K/(τs + 1), with a single pole at s = −1/τ.

  • Steady-state gain K: the final change in output per unit change in input (units of output/input).
  • Time constant τ: a measure of speed; it is the product of a capacitance and a resistance (storage ÷ throughput).

Typical derivations.

  • Mercury thermometer: energy balance m·Cp·dT/dt = h·A·(T_bath − T) gives τ = m·Cp/(h·A) and K = 1.
  • Liquid-level tank with linear outlet resistance: A·dh/dt = q_in − h/R gives H(s)/Q_in(s) = R/(A·R·s + 1), so τ = A·R and K = R.
  • Stirred mixing tank: V·dc/dt = q·(c_in − c) gives τ = V/q (the residence time) and K = 1.
  • A non-linear tank (outflow ∝ √h) is first order only after linearisation about the operating point (see the linearisation topic).

Step response. For an input step of size M, y(t) = K·M·(1 − e^(−t/τ)). The response reaches 63.2 % of its final change at t = τ, 86.5 % at 2τ, 95.0 % at 3τ, 98.2 % at 4τ and 99.3 % at 5τ. The initial slope is K·M/τ; if the response kept that slope it would reach the final value at t = τ. There is no overshoot and no inflection.

Impulse response. For an impulse of area M, y(t) = (K·M/τ)·e^(−t/τ): a jump followed by exponential decay.

Ramp response. For u = a·t, y(t) = K·a·(t − τ + τ·e^(−t/τ)). After the transient, the output follows a ramp delayed by τ; with K = 1 the measurement lags the true value by a·τ.

Sinusoidal response. For u = A·sin(ωt), after transients the output is sinusoidal at the same frequency with amplitude ratio K/√(1 + ω²τ²) and phase lag φ = tan⁻¹(ωτ) (between 0 and 90°).

Pure capacity (integrating) process. A tank whose outflow is fixed by a pump gives G(s) = 1/(A·s): the level ramps without limit after a step. It is self-regulating only if the outflow depends on the level.

Formulas

G(s) = K / (τ·s + 1)

  • K steady-state gain (output unit/input unit); τ time constant (s or min).

y(t) = K·M·(1 − e^(−t/τ))

  • Step response to an input step of size M.

t = τ·ln(1/(1 − f))

  • Time to reach fraction f of the final change.

y(t) = K·a·(t − τ + τ·e^(−t/τ))

  • Ramp response, a = input ramp rate (input unit/s). Long-time lag = K·a·τ (for K = 1).

AR = Â_out / (K·Â_in) = 1/√(1 + ω²·τ²), φ = −tan⁻¹(ω·τ)

  • Normalised amplitude ratio and phase angle for a sinusoid of angular frequency ω (rad/s). Time lag of the output wave = |φ|/ω (φ in rad).

τ = m·Cp/(h·A) (thermometer), τ = A·R and K = R (linear tank), τ = V/q (mixing tank)

  • m mass (kg), Cp heat capacity (J/(kg·K)), h film coefficient (W/(m²·K)), A area (m²), R outlet resistance (s/m²), V volume (m³), q flow (m³/s).

Worked examples

Example 1 (standard): tank with linear outlet resistance. A tank of cross-section A = 2 m² drains through a linear resistance R = 0.5 min/m² (so h = R·q_out with h in m, q in m³/min). The inlet flow steps up by 0.4 m³/min. Find the final level change and the level change after 2 min.

  1. τ = A·R = 2·0.5 = 1.0 min; K = R = 0.5 m per (m³/min).
  2. Final change: Δh(∞) = K·M = 0.5·0.4 = 0.20 m.
  3. Δh(t) = 0.20·(1 − e^(−t/1.0)).
  4. At t = 2 min: Δh = 0.20·(1 − e^(−2)) = 0.20·0.8647 = 0.173 m.

Δh(∞) = 0.20 m; Δh(2 min) ≈ 0.173 m

Example 2 (GATE level): thermometer in a ramping bath. A thermometer with τ = 0.5 min is at steady state in a bath at 50 °C. The bath temperature then rises linearly at 2 °C/min. Find the reading and the true bath temperature at t = 2 min, and the long-time lag.

  1. Deviation input u = a·t with a = 2 °C/min; K = 1.
  2. y(t) = a·(t − τ + τ·e^(−t/τ)) = 2·(2 − 0.5 + 0.5·e^(−4)) = 2·(1.5 + 0.00916) = 3.018 °C.
  3. Reading = 50 + 3.018 = 53.02 °C; bath = 50 + 2·2 = 54.00 °C.
  4. Long-time lag = a·τ = 2·0.5 = 1.0 °C (the reading lags by τ = 0.5 min in time).

Reading ≈ 53.02 °C while the bath is at 54.00 °C; the lag tends to 1.0 °C

Example 3 (GATE level): sinusoidal bath temperature. The same thermometer (τ = 0.5 min) sits in a bath oscillating ±2 °C with a period of 2 min.

  1. ω = 2π/P = 2π/2 = 3.1416 rad/min; ωτ = 1.571.
  2. AR = 1/√(1 + 1.571²) = 1/√3.467 = 0.537.
  3. Reading amplitude = 0.537·2 = ±1.07 °C.
  4. φ = −tan⁻¹(1.571) = −57.5°; time lag = 1.004 rad / 3.1416 rad/min = 0.32 min.

Common mistakes

  • Reading τ as the time to reach the final value; at t = τ only 63.2 % of the change has occurred.
  • Forgetting to multiply by the input step size M or by the gain K.
  • Using absolute instead of deviation variables, so the initial steady value is lost or double-counted.
  • Using the period P instead of ω = 2π/P in the amplitude ratio.
  • Treating a tank with a pumped outlet as first-order; it is a pure integrator with no steady state.
  • Calling a first-order system "overdamped": damping ratio applies to second-order systems.

For GATE CH

This is one of the most frequently examined topics. Expect: time to reach a given fraction of the final value, the reading of a thermometer at a given time, the time constant from a test, ramp lag, and amplitude ratio or phase lag of a sinusoidal input. Derivations of τ and K for tanks and thermometers also appear, often combined with linearisation. Memorise the 63.2/86.5/95/98.2 % markers.

Quick check

  1. A first-order system with τ = 5 s receives a step. How long to reach 90 % of the final change?
  2. What is the gain of G(s) = 6/(3s + 1), and its time constant?
  3. A thermometer (τ = 0.2 min) follows a bath rising at 5 °C/min. What is the long-time lag?
  4. At ωτ = 1, what are the amplitude ratio and phase lag?

Answers: 1. 5·ln 10 = 11.5 s. 2. K = 6, τ = 3 time units. 3. 1.0 °C. 4. AR = 0.707, phase lag 45°.

Try answering each one aloud before you open it.

  1. 1.What is a first-order system in process control?Concept

    A first-order system in process control is a dynamic system that can be described by a first-order differential equation. It has a single energy storage element and its response to a step input is characterized by an exponential rise or decay. The system's behavior is defined by its time constant, which indicates how quickly the system responds to changes.

  2. 2.Explain the concept of time constant in a first-order system.Concept

    The time constant, often denoted by τ (tau), is a measure of the speed of response of a first-order system. It is the time required for the system's response to reach approximately 63.2% of its final value after a step input is applied. A smaller time constant indicates a faster response, while a larger time constant indicates a slower response.

  3. 3.How does a first-order system respond to a step input?Concept

    When a first-order system is subjected to a step input, its output response is an exponential function that gradually approaches a new steady-state value. The response is characterized by an initial rapid change followed by a slower approach to the final value, with the rate of change determined by the system's time constant.

  4. 4.Why are first-order systems commonly used to model chemical processes?Application

    First-order systems are commonly used to model chemical processes because many processes can be approximated as having a single dominant energy storage element, such as a tank or a heat exchanger. This simplification allows for easier analysis and control design, as the mathematical models are less complex and can still provide accurate predictions of system behavior.

  5. 5.What happens to the response of a first-order system if the time constant is doubled?Application

    If the time constant of a first-order system is doubled, the system's response to a step input will become slower. Specifically, it will take twice as long for the system to reach approximately 63.2% of its final value. This means the system will take longer to settle into its new steady-state after a change.

  6. 6.How can you determine the time constant of a first-order system from its step response graph?Application

    To determine the time constant from a step response graph, identify the point at which the response reaches 63.2% of its final steady-state value. The time at this point is the time constant τ. This method assumes the system starts from rest and the input is a step change.

  7. 7.What is the effect of feedback on the dynamic response of a first-order system?Application

    With proportional feedback control, a first-order process K/(τs + 1) gives a closed-loop response that is still first order, with time constant τ/(1 + Kc·K) and gain Kc·K/(1 + Kc·K). So negative feedback makes the response faster and reduces sensitivity to disturbances, but proportional action alone leaves an offset of 1/(1 + Kc·K) for a unit set-point step. A first-order process under P control can never become unstable; instability needs extra lags or dead time in the loop.

  8. 8.Calculate the time constant of a first-order system if its step response reaches 95% of its final value in 10 seconds.Numerical

    The step response is y/y∞ = 1 − e^(−t/τ). Setting 0.95 = 1 − e^(−10/τ) gives e^(−10/τ) = 0.05, so τ = 10/ln(20) = 10/2.996 ≈ 3.34 s. A useful check is that 95 % is reached at about 3τ.

  9. 9.A first-order system has a time constant of 5 seconds. How long will it take to reach 90% of its final value after a step input?Numerical

    From 0.90 = 1 − e^(−t/5), e^(−t/5) = 0.10, so t = 5·ln(10) = 5 × 2.303 ≈ 11.5 s, which is about 2.3 time constants.

  10. 10.Explain why the Laplace transform is useful in analyzing first-order systems.Concept

    Transforming τ·dy/dt + y = K·u with zero deviation initial conditions gives the transfer function K/(τs + 1), so the response to any input is just G(s)·U(s) inverted from a table. For a step of size M, Y = K·M/(s(τs + 1)), which inverts directly to K·M(1 − e^(−t/τ)). The single pole at s = −1/τ immediately shows that the system is stable and how fast it responds, and substituting s = jω gives the amplitude ratio and phase lag for sinusoidal inputs.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?