Why Does A Constant-Outlet Liquid-Level Pilot Plant Eliminate Set-Point Offset But Not Disturbance Offset?

The difference comes from where the integrating behavior appears in the control loop. For a liquid-level process with constant outlet flow, the tank level is the integral of the inlet-flow imbalance, giving (G_p(s)=1/(A s)). Under proportional control, this process integrator drives the level to the set point after a set-point change, but a sustained load disturbance requires a sustained controller output and therefore a nonzero error; the result is a steady-state offset.

A process integrator can eliminate set-point offset under proportional control because any persistent level error continues changing the level until the error disappears. A load disturbance is different: the controller must maintain a permanently changed flow, and proportional control can produce that changed output only by maintaining a nonzero error.

Why the Liquid-Level Process Is Integrating

The tank stores material

The level changes according to the difference between inlet and outlet flow:

[ A\frac{dh}{dt}=q_{\text{in}}-q_{\text{out}} ]

where (A) is the tank cross-sectional area and (h) is the liquid level.

If the outlet flow is constant, deviations in level are caused by the accumulated difference between the inlet and outlet flows. In deviation-variable form:

[ \frac{H(s)}{Q_{\text{in}}(s)}=\frac{1}{A s} ]

The (1/s) term is the mathematical representation of integrating storage.

A persistent flow imbalance cannot remain unnoticed

If the inlet flow exceeds the constant outlet flow, the tank level continues to rise. If the inlet flow is lower, the level continues to fall.

Therefore, a sustained input-flow change produces a continuing level change unless the controller adjusts the inlet flow to restore the material balance.

Why Proportional Control Removes Set-Point Offset

The controller output depends on level error

For proportional control,

[ u=K_c e ]

where (u) is the manipulated input, (K_c) is the controller gain, and (e) is the set-point error.

A nonzero error produces a controller action, and that action changes the inlet flow. Because the process itself integrates flow imbalance, the level keeps moving while the error remains nonzero.

The process integrator drives the error to zero

After a set-point step, the level initially differs from the new target. Proportional control changes the inlet flow, and the integrating process continues moving the level toward the target.

At steady state, the level must stop changing. Since

[ \frac{dh}{dt}=0 ]

the inlet and outlet flows must be equal. For a set-point change with no sustained load disturbance, the required steady-state manipulated-input change is zero relative to the new equilibrium flow. Consequently, the proportional controller can have zero output deviation with zero error.

The level therefore settles at the set point, although the transient response may include oscillation or other dynamic effects depending on the gain and process conditions.

The distinction from an ordinary self-regulating process

In a self-regulating process, such as a process with a flow-dependent outlet, a changed steady-state input usually requires a changed steady-state controller output. Proportional control must then maintain a nonzero error to generate that output.

The constant-outlet liquid-level process is different because its process gain contains an integrator. That integrator can produce the required level change without requiring a permanent control error for a set-point change.

Why Load Disturbances Still Cause Offset

A disturbance changes the required flow balance

Suppose an inlet-flow disturbance changes the material balance. For example, an additional outlet flow may suddenly remove more liquid from the tank.

To restore a constant level, the manipulated inlet flow must increase by exactly the amount needed to compensate for that disturbance.

Proportional control needs error to hold a new output

A proportional controller can increase the inlet flow only when it sees an error:

[ \Delta u=K_c\Delta e ]

To maintain the increased inlet-valve opening after the disturbance has been compensated, the controller must continue receiving a nonzero error signal. If the level error returned to zero, the proportional controller output would return to its original value, and the flow balance would be lost again.

The system therefore reaches a new equilibrium with the correct flow balance but with the level displaced from its set point.

The offset is inversely related to controller gain

For a unit step load disturbance, the steady-state error magnitude is commonly expressed as

[ |e_{\text{ss}}|=\frac{1}{K_c} ]

with the sign determined by the disturbance and the chosen error convention. Increasing (K_c) reduces the offset, but finite proportional gain cannot make it exactly zero.

This is the central contrast:

Set-point change: the process integrator moves the level until the error can return to zero.
Load disturbance: a permanent change in manipulated flow is required, so proportional control must retain a permanent error.

Understanding the Trade-Offs

Higher proportional gain reduces, but does not eliminate, offset

A larger (K_c) produces a larger manipulated-flow correction for a given level error. This reduces the disturbance offset according to the inverse relationship with gain.

However, very high gain can make the loop oscillatory or unstable because the controller reacts more aggressively to level deviations and measurement dynamics.

Integral action eliminates disturbance offset

Adding integral action gives the controller memory of accumulated error:

[ G_c(s)=\frac{K_c}{\tau_I s} ]

The integral term continues changing the controller output as long as error persists. It can therefore increase the inlet flow to compensate for the disturbance and continue adjusting until the level error reaches zero.

Integral action changes the dynamics

Integral control removes the steady-state offset, but it can make the response slower and can increase overshoot or oscillation if the reset time is too short or the gain is too high.

The pilot plant is valuable because it isolates this control principle clearly: a process integrator may provide zero set-point offset with P control, while controller integral action is still needed for zero load-disturbance offset.

Making the Right Choice for Your Goal

The appropriate controller depends on whether the dominant requirement is set-point tracking, disturbance rejection, or both.

If your primary focus is set-point tracking: Proportional control can achieve zero steady-state level offset in a constant-outlet integrating process, provided the loop remains stable.
If your primary focus is load-disturbance rejection: Add integral action, because proportional control alone requires a persistent level error to sustain the compensating flow.
If your primary focus is fast, well-damped operation: Tune proportional and integral action together, balancing reduced offset against overshoot, oscillation, and sluggishness.
If your primary focus is understanding the pilot plant: Separate set-point and disturbance tests, because the same P controller can show zero offset in one test and finite offset in the other.

The key is to distinguish the process integrator’s ability to remove set-point error from the controller integrator’s ability to remove disturbance error.

Summary Table:

Control scenario	Process behavior	Proportional-control result	Recommended approach
Set-point change	The tank integrates flow imbalance until the level reaches the target.	Zero steady-state offset is possible.	Use P control if the loop remains stable.
Load disturbance	A permanent flow correction is required to restore material balance.	Finite offset remains because error is needed to sustain controller output.	Add integral action for zero offset.
Higher proportional gain	A given level error produces a larger flow correction.	Offset decreases, but may cause oscillation or instability.	Increase gain carefully and verify damping.
Integral control	The controller accumulates error and continues adjusting output.	Load-disturbance offset is eliminated.	Tune gain and reset time to limit overshoot.

Need to demonstrate integrating-process control in a teaching or research environment? Contact LABPARK to discuss an Educational or Vocational Unit Operations Pilot Plant tailored to your needs. LABPARK provides liquid-level and other chemical engineering pilot plants, along with bioprocess and biotech and environmental and water-treatment systems, for universities, research institutes, and enterprises. Our systems help teams teach process dynamics, compare proportional and integral control, conduct repeatable set-point and disturbance tests, and build practical operator experience on equipment designed for training, research, and industrial application.